The Natural Radiation Environment VII, Volume 7: Seventh International Symposium on the Natural Radiation Environment (NRE-VII) Rhodes, Greece, 20-24 May 2002 (Radioactivity in the Environment)

THE NATURAL RADIATION ENVIRONMENT VII Seventh International Symposium on The Natural Radiation Environment (NRE-VII) ...

Author: J.P. McLaughlin | E.S. Simopoulos | F. SteinhÃ¤usler

159 downloads 2046 Views 14MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

THE NATURAL RADIATION ENVIRONMENT VII Seventh International Symposium on The Natural Radiation Environment (NRE-VII)

RADIOACTIVITY IN THE ENVIRONMENT A companion series to the Journal of Environmental Radioactivity Series Editor M.S. Baxter Ampﬁeld House Clachan Seil Argyll, Scotland, UK Volume 1: Plutonium in the Environment (A. Kudo, Editor) Volume 2: Interactions of Microorganisms with Radionuclides (F.R. Livens and M. Keith-Roach, Editors) Volume 3: Radioactive Fallout after Nuclear Explosions and Accidents (Yu.A. Izrael, Author) Volume 4: Modelling Radioactivity in the Environment (E.M. Scott, Editor) Volume 5: Sedimentary Processes: Quantiﬁcation Using Radionuclides (J. Carroll and I. Lerche, Authors) Volume 6: Marine Radioactivity (H.D. Livingston, Editor) Volume 7: The Natural Radiation Environment VII (J.P. McLaughlin, S.E. Simopoulos and F. Steinhäusler, Editors)

THE NATURAL RADIATION ENVIRONMENT VII Seventh International Symposium on The Natural Radiation Environment (NRE-VII) Rhodes, Greece, 20–24 May 2002

Editors J.P. McLaughlin University College Dublin, Ireland

S.E. Simopoulos National Technical University of Athens, Greece

F. Steinhäusler University of Salzburg, Austria

2005 AMSTERDAM – BOSTON – HEIDELBERG – LONDON – NEW YORK – PARIS SAN DIEGO – SAN FRANCISCO – SINGAPORE – SYDNEY – TOKYO

ELSEVIER B.V. Radarweg 29 P.O. Box 211, 1000 AE Amsterdam The Netherlands

ELSEVIER Inc. 525 B Street, Suite 1900 San Diego, CA 92101-4495 USA

ELSEVIER Ltd The Boulevard, Langford Lane Kidlington, Oxford OX5 1GB UK

ELSEVIER Ltd 84 Theobalds Road London WC1X 8RR UK

© 2005 Elsevier Ltd. All rights reserved.

This work is protected under copyright by Elsevier, and the following terms and conditions apply to its use: Photocopying Single photocopies of single chapters may be made for personal use as allowed by national copyright laws. Permission of the Publisher and payment of a fee is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. Special rates are available for educational institutions that wish to make photocopies for non-proﬁt educational classroom use. Permissions may be sought directly from Elsevier’s Rights Department in Oxford, UK: phone: (+44) 1865 843830; fax: (+44) 1865 853333; e-mail: [email protected]. Requests may also be completed on-line via the Elsevier homepage (http://www.elsevier.com/locate/permissions). In the USA, users may clear permissions and make payments through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; phone: (+1) (978) 7508400; fax: (+1) (978) 7504744; and in the UK through the Copyright Licensing Agency Rapid Clearance Service (CLARCS), 90 Tottenham Court Road, London WlP 0LP, UK; phone: (+44) 20 7631 5555; fax: (+44) 20 7631 5500. Other countries may have a local reprographic rights agency for payments. Derivative Works Tables of contents may be reproduced for internal circulation, but permission of Elsevier Science is required for external resale or distribution of such material. Permission of the Publisher is required for all other derivative works, including compilations and translations. Electronic Storage or Usage Permission of the Publisher is required to store or use electronically any material contained in this work, including any chapter or part of a chapter. Except as outlined above, no part of this work may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission of the Publisher. Address permissions requests to: Elsevier’s Rights Department, at the phone, fax and e-mail addresses noted above. Notice No responsibility is assumed by the Publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent veriﬁcation of diagnoses and drug dosages should be made.

First edition 2005 Library of Congress Cataloging in Publication Data A catalog record from the Library of Congress British Library Cataloguing in Publication Data A catalogue record from the British Library

ISBN: 0-08-044137-8 ISSN: 1569-4860 ∞ The paper used in this publication meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Printed in The Netherlands.

Seventh International Symposium NATURAL RADIATION ENVIRONMENT (NRE-VII) 20–24 M a y 2 0 0 2 , R h o d e s , G r e e c e

Organisers J.P. Mc Laughlin, University College Dublin, Ireland S.E. Simopoulos, National Technical University Athens, Greece F. Steinhäusler, University of Salzburg, Austria

Scientific Secretariat M.J. Anagnostakis, D.J. Karangelos, E.P. Hinis, N.P. Petropoulos, P.K. Rouni

Symposium Secretary V.A. Griva

NATIONAL TECHNICAL UNIVERSITY OF ATHENS NUCLEAR ENGINEERING SECTION

15780 Athens, Greece • tel: (30) 210772 2911 & 2912, fax: (30) 210772 2914 e-mail: [email protected] • http://nreVII.nuclear.ntua.gr

T . his Page Intentionally Left Blank

vii

Invited Foreword It seems to me that, of all Conference series in the ﬁelds of environmental radioactivity/nuclear geochemistry/radioecology/radiological protection, the Natural Radiation Environment (NRE) series is pre-eminent. Its history, traditions, breadth and sustained high standards distinguish it from and set it above the others. Major international participation and strict peer-review of its papers have helped to ensure that the NRE books have been not just historical records of large meetings, with the normal range of quality variation and research completeness that conferences always entail, but are, in fact, comprehensive, selected, reviewed and revised compilations of current cutting-edge knowledge on a very important topic. And it is indeed a signiﬁcant, wide-ranging and expanding subject area, fundamental to so many powerful applications across so many disciplines of science and arguably the most important of all in the environmental radioactivity ﬁeld. As founder of both the Journal of Environmental Radioactivity (JER) and this Radioactivity in The Environment book series, I believe that together these two publication series should provide a comprehensive service to the specialist scientiﬁc community. The book series provides a complementary and parallel forum to the peer-reviewed research functions of JER by producing quality textbooks for students and researchers worldwide. As such, the core of this series is provided by a set of speciﬁcally written volumes on the key scientiﬁc issues, e.g. separate books on marine, terrestrial and atmospheric radioactivity, on health effects on man and biota, on modelling, on tracers, etc., etc. These books, some still in production, are largely multi-authored, being organised and produced by guest editors. However, there is scope and place also within the series for a limited conference proceedings sub-series, but many provisos must be attached to inclusion of the latter. First, of course, is the necessity that the topic is of particular interest internationally and that a great deal of selection and full peer-review should be involved in the book production process. Not many conference proceedings satisfy these requirements, but certainly NRE VII does. First, the topic is hugely relevant and of wide general impact. Secondly, the organisers and editors of the book have been fastidious in their preparatory work, rejecting many conference papers after peer-review and having most papers revised and improved relative to the original oral presentations. I thank them for carrying out the huge workload of arranging review, revising, rejecting, accepting, selecting, compiling, coordinating, correcting and submitting this volume so efﬁciently and punctually. Most of my communications on this book have been with Professor Simos E. Simopoulos and Dr. Nick Petropoulos of the National Technical University of Athens and all I can say about that aspect is that it was both efﬁcient and a pleasure. Through their editorial efforts, I am conﬁdent that this compilation combines the qualities of a research journal with the size and breadth associated with a textbook.

viii

Invited foreword

This book is written by a sizeable and key section of that global community of scientists who devote themselves to studying the consequences of the fundamental and signiﬁcant fact that this planet is radioactive not at all mainly by man’s doing but, if you like, by God’s doing. That the man-in-the-street believes that the planet is radioactive because of man’s recent experimentations with nuclear weapons, nuclear energy and nuclear applications is sad, made all the more ironic by the fact that natural radioactivity and nuclear processes were so intimately involved in the very creation of the universe, of the planet and of life as we know it. But blame for this lack of common understanding falls at least partly if not largely on our own doorstep since we, the scientiﬁc community, have essentially failed to communicate objectively and effectively. Well, here is another database to provide stimulation and resource for such communication in future – an encyclopedic collection of papers on the state of NRE art and science. Let us use it as a source of material not just for research but also for teaching both within academia and beyond. NRE I was, in fact, the ﬁrst book that I ever used as a research support in my own career for I used it as a reference for my ﬁrst experimental research project as an undergraduate – on thorium-230 uptake in the oceans by marine sediments – 40 years ago. NRE provided “a bible” for me then. I truly hope that this new volume will be similarly important as an information source to some new young scientists in our exciting ﬁeld. For what can be more exciting than some of the subjects herein? – the origin of the planet and elements, the age of the Earth and geological timescales, isotopes as tracers of dynamic processes on sea, land and air, the relationship between radiation exposure and cancer, radiation exposures from building materials or from radon in home and workplace, increasing awareness of the occurrence of technological enhancement of the natural radiation environment through non-nuclear industries, dosimetry to future astronauts during space travel, doses from depleted uranium weaponry, health effects of living in areas of high background radiation and so on and so forth. I was recently invited to present a summary of the possible main research needs and challenges in this ﬁeld (Environmental Radioactivity at 2000: Status and Priorities. In: Inaba, J., Hisamatsu, S., Ohtsuka, Y. (Eds.), Distribution and Speciation of Radionuclides in the Environment. Institute of Environmental Sciences Publishers, Aomori, Japan, 2000, ISBN 4-9980604-3-0 C3040). In reviewing this book, I am happy to ﬁnd within it invaluable responses and contributions in each of the priority areas of need that I described then, such as for (a) increased use of radiotracers, particularly natural radionuclides, to further understanding of natural rates and mechanisms, (b) scientiﬁc and socioeconomic study of the remediation and restoration of contaminated sites, (c) development of radiological protection criteria and frameworks for ﬂora and fauna as well as for humans, (d) improvement of predictive models of radionuclide distributions in contaminated and natural environments, (e) improved comparative assessment of contaminant effects via better understanding of the ecotoxicology of radioactive and non-radiological contaminants, (f) improved experimental investigation of the speciation of radionuclides, their coordination chemistry, kinetics and microbiology and, ﬁnally, (g) the need to enhance public and political understanding by improving communications.

Invited foreword

ix

In short, I am convinced that the science contained herein moves the subject forward to a truly signiﬁcant extent. In conclusion, the science of the natural radioactivity of the Earth is, in my view, so much more interesting than that of anthropogenic radionuclides – with so many more powerful applications across so many more interdisciplinary boundaries and with generally greater dosimetric signiﬁcance – that it is a true honour to be instrumental in bringing this book to you. I hope that it will inform and inspire. And, ﬁnally again, I thank the volume editors for their great work in bringing this book to you in such very good shape. Murdoch S. Baxter Series Editor e-mail: [email protected]

T . his Page Intentionally Left Blank

xi

Preface The Natural Radiation Environment Symposium (NRE VII), the Seventh in the NRE series, which commenced forty years ago in 1963 at Rice University Texas, was convened in Rhodes (Greece) in May 2002. During the intervening four decades, the research work presented at these NRE Symposia has, in no small way, contributed to a deeper understanding of natural radiation and, in particular, of its contribution to human radiation exposures. It is clear from the quality and diversity of the 143 papers in this special volume of Elsevier’s Radioactivity in the Environment book series that the study of the natural radiation environment is an active and continually expanding ﬁeld of research. The papers in this volume fall into a number of main and topical research areas namely: – – – –

the measurement and behaviour of natural radionuclides in the environment; cosmic radiation measurement and dosimetry; the external penetrating radiation ﬁeld at ground level; TENR (Technologically Enhanced Natural Radiation) and NORM (Naturally Occurring Radioactive Materials) studies; – assessment of the health effects of radon; – regulatory aspects of natural radiation exposures. In these papers, the results of many new surveys of natural radionuclide levels in the environment and of improved methods of detection are described. While some of the natural radiation sources investigated are unmodiﬁed by human activity, many accounts are given here of exposures to natural sources which have been enhanced by technology. Such TENR and NORM exposures are shown to range from activities such as mining, oil and gas exploitation, the use of industrial by-products as building materials, to space travel, to name but a few. In several cases, quite high doses to some individuals are shown to occur. Accounts are given here of methods to prevent and reduce exposures to such sources. Excluding doses from radiotherapy and radiation accidents, natural radiation exposures are now accepted as being the major contributors to the annual average radiation dose received by humans under normal conditions. In particular, in view of the large range of doses that are due to radon exposure, the common practice in the past of referring to natural radiation as “background radiation”, with its connotation of insigniﬁcance, is now clearly redundant. In the past, for a variety of reasons, which were not always scientiﬁcally based, undue attention was often directed at individual annual doses of a few μSv from artiﬁcial radiation sources while ignoring some tens of mSv from natural sources. In contrast to this ethically questionable and unsustainable position, there is now a growing recognition and acceptance by the contemporary radiation protection community of the need to reduce exposures to elevated

xii

Preface

levels of natural radiation. This is clearly evidenced by the emergence in recent years – both at various national and international levels – of a large body of recommendations and regulations targeted at the control of natural radiation exposures. These range from the protection of aircrew and astronauts from cosmic radiation and solar ﬂares, to controls on the level of radon in the workplace. A number of accounts are given in this volume of some of these regulations and aspects of their practical implementation. While the dose limits used in radiation protection are largely based on a no-threshold linear hypothesis, extrapolating from observed health effects at high doses, the actual shape of the dose–response curve at the level of radiation protection dose limits is still uncertain. Studies on the health effects of natural radiation exposures may help to remove some of this uncertainty. To date, epidemiological studies of populations exposed to elevated external penetrating radiation of natural origin have been largely inconclusive in this regard. On the other hand, notwithstanding confounding effects and other problems, recent case-control residential radon epidemiological studies have demonstrated clear evidence of a non-negligible lifetime risk of lung cancer from indoor radon at concentrations comparable to the radon action levels set by various agencies. As the risk from the radon series arises primarily from alpha radiation, these studies assist in quantifying the radiation-weighting factor for alpha radiation. Some recent studies on the assessment of the health risks from exposures to natural radiation are presented in this publication. The NRE VII Symposium, at which the peer-reviewed papers in this volume were presented, was attended by over 260 scientists from many disciplines. They overwhelmingly expressed the need to improve the coordination of NRE-related research and regulatory activities worldwide. Therefore it was agreed to found the Natural Radiation Environment Association (NREA) to assist in the planning and implementation of research projects, as well as in the conduct of dedicated training courses and intercalibration exercises. Furthermore, it is intended that the NREA will establish a dedicated electronic database on NRE, TENR and NORM global data. Currently, the organisational details (management, secretariat, funding, membership requirements) are being discussed. In summary, NRE VII demonstrated clearly the vitality of NRE-related research, the need for strengthening further international cooperation on this truly global issue, and the high value of NRE-based information for radiation protection in general. It is hoped that this will provide sufﬁcient encouragement for the scientiﬁc and regulatory community to continue this valuable conference series also in the future. We wish to gratefully acknowledge all organisations that supported the Symposium, primarily the US Environmental Protection Agency, which was the main sponsor of the Symposium. We also wish to express our appreciation to our colleagues, who voluntarily assisted in reviewing the Symposium manuscripts. Our appreciation is extended to Dr. N.P. Petropoulos for the coordination of the whole process, which led to the publication of this volume. December 2003 J.P. McLaughlin, University College, Dublin, Ireland S.E. Simopoulos, National Technical University of Athens, Greece F. Steinhäusler, University of Salzburg, Austria

xiii

Contents

Organisers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

v

Invited Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

vii

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xi

1. Invited

1

1.

2.

3.

Historical development of the Natural Radiation Environment Symposia by A.C. George . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3

Understanding health risks from low doses of ionizing radiation by D.G. Thomassen & A. Patrinos . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

12

The signiﬁcance of the natural environment in radiation protection by F. Steinhäusler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

18

2. Keynote 4.

27

The theory of cosmic-ray and high-energy solar-particle transport in the atmosphere by K. O’Brien . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

29

Capabilities and limitations of high-resolution gamma spectrometry in environmental research by J. Uyttenhove . . . . . . . . . . . . . . . . . . . . . . .

45

6.

Radon decay products in outdoor air by J. Porstendörfer & M. Gründel . . .

56

7.

Contribution of animal experimental data for the risk assessment of exposure to radon decay products by G. Monchaux . . . . . . . . . . . . . . . . . . . .

66

On the exposure circumstances and some further risk estimates regarding leukemia in ages 0–19 years and exposure to ionizing radiation in homes of uranium-containing alum shale-based concrete by G. Åkerblom, L. Hardell, M. Fredrikson & O. Axelson . . . . . . . . . . . . . . . . . . . . . . . . . . .

77

5.

8.

xiv

Contents

9. The implementation in national legislation of Title VII of the Council Directive 96/29/Euratom: Some general remarks and the case of Italy by S. Risica, F. Bochicchio & C. Nuccetelli . . . . . . . . . . . . . . . . . . . . . . . . . .

85

10. Exposure potential of depleted uranium ammunition in the environment – a review by J.P. Mc Laughlin . . . . . . . . . . . . . . . . . . . . . . . . . . .

95

3. Radioactivity measurements, releases and dosimetry

105

11. Changes in terrestrial natural radiation levels over the history of life by P.A. Karam & S.A. Leslie . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

107

12. Gamma ray direction ﬁnder by K. Fujimoto & Y. Noda . . . . . . . . . . . .

118

13. Continuous measurement of environmental gamma radiation using a Ge semiconductor detector and 222 Rn concentration in air by T. Ichiji & T. Hattori .

126

14. Uranium concentrations and isotopic ratios in Austrian drinking waters by M. Gegner & K. Irlweck . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

135

15. Natural radionuclides in radium-rich soils in North-East Estonia by E. Realo & K. Realo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

140

16. Depleted uranium determination in soil samples by I. Bikit, J. Slivka, ˇ ´ ci´c & D. Mrdja . . . . . M. Krmar, M. Veskovi´c, L. Conki´ c, E. Varga, S. Curˇ

150

17. Measurement of snow cover based on external radiation by J. Paatero, E. Kyrö, J. Hatakka, V. Aaltonen & Y. Viisanen . . . . . . . . . . . . . . . . . . . . .

155

18. Ra and U isotopes determination in phosphogypsum leachates by alphaparticle spectrometry by J.L. Aguado, J.P. Bolívar, E.G. San Miguel & R. García-Tenorio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

160

19. Efﬁciency calibration for 210 Pb gamma-spectrometric determinations in sediment samples by E.G. San Miguel, J.P. Pérez-Moreno, J.P. Bolívar, J.L. Aguado & R. García-Tenorio . . . . . . . . . . . . . . . . . . . . . . . .

166

20.

238 U

and its daughter products in Greek surface soils by M.J. Anagnostakis, E.P. Hinis & S.E. Simopoulos . . . . . . . . . . . . . . . . . . . . . . . . . .

175

21. Data leading to the investigation of a relation between seismic activity and airborne radon decay product concentrations outdoors by D.J. Karangelos, N.P. Petropoulos, M.J. Anagnostakis, E.P. Hinis & S.E. Simopoulos . . . .

187

Contents

22.

23.

Calibration of a HPGe detector for in-situ gamma spectrometry: a comparison between a Monte Carlo based code and an experimental method by F. D’Alberti & M. Forte . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

198

A new technique for accurate measurements of Ra-226 with γ-spectroscopy in voluminous samples by M. Manolopoulou, S. Stoulos, D. Mironaki & C. Papastefanou . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

207

4. Radon and thoron 24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

xv

213

Radon concentration in the tunnels of a hydroelectric power station under construction in Italy, a case study by S. Verdelocco, D. Walker & P. Turkowsky

215

Determination of the radon potential of a building by a controlled depressurisation technique (RACODE) by W. Ringer, H. Kaineder, F.-J. Maringer & P. Kindl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

221

Population exposure to inhaled radon and thoron progeny by O. Iacob, C. Grecea & E. Botezatu . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

232

Preferential radon transport through highly permeable channels in soils by R.B. Mosley . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

238

A new approach to increasing the uptake of radon remediation in England by S. Scivyer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

260

User-friendly computer programs that calculate radon cancer risks, progeny accumulation and gamma exposure by S. Rydell . . . . . . . . . . . . . . . .

265

Testing of radon-reducing measures under strictly controlled conditions in a laboratory house by P. de Jong & W. van Dijk . . . . . . . . . . . . . . . . .

276

Looping variation of correlation between radon progeny concentration and dose rate in outdoor air by N. Fujinami, T. Watanabe & T. Tsutsui . . . . . .

284

Comparative dosimetry in homes and mines: estimation of K-factors by J.W. Marsh, A. Birchall & K. Davis . . . . . . . . . . . . . . . . . . . . . . .

290

The principle of a passive, on/off, alpha track, closed radon detector by J. Andru . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

299

Intercomparison exercise of calibration facilities for radon gas activity concentration by A. Röttger, A. Honig, G. Butterweck, Ch. Schuler, V. Schmidt, H. Buchröder, A. Rox, J.C.H. Miles, I. Burian, N. Michielsen, V. Voisin, F.-J. Maringer & A. Vargas . . . . . . . . . . . . . . . . . . . . . . . . . . . .

306

xvi

Contents

35. In-vivo measurement of deposition and absorption of unattached radon progeny by G. Butterweck, Ch. Schuler, G. Vezzù, J.W. Marsh & A. Birchall

314

36. Assessing radiation dose for tour guides in Australian show caves from radon monitor measurements by S.B. Solomon, J. Peggie & R. Langroo . . . . . .

326

37. Inﬂuence of variable stress on underground radon concentrations by A. Kies, F. Massen & Z. Tosheva . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

334

38. Measurement of the unattached radon decay products with an annular diffusion channel battery by N. Michielsen, V. Voisin & G. Tymen . . . . . . . .

339

39. Intercomparison of the radiosensitivity of different measuring devices for environmental levels of radon by A.C. George & N. Bredhoff . . . . . . . . . .

346

40. Integrating measurements of radon and thoron and their deposition fractions in the respiratory tract by W. Zhuo, S. Tokonami, H. Yonehara & Y. Yamada

352

41. Determination of 218 Po nanometer size distribution in a controlled environment by two new systems by A. Vargas, N. Michielsen, C. Le Moing, M. Rio, G. Tymen & X. Ortega . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

361

42. Measurements and modelling of combined diffusive and advective radon transport in porous building materials by M. van der Pal, W.H. van der Spoel, R.J. de Meijer, N.A. Hendriks & E.R. van der Graaf . . . . . . . . . . . . . .

371

43. Czech study on lung cancer risk from residential radon by A. Heribanová & L. Tomášek . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

381

44. Czech study on uranium miners – evaluation of temporal factors in a 50 year follow-up by L. Tomášek . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

389

45. Activity concentrations of the thoron and radon progenies Pb-212 and Pb210 in the healing gallery of Badgastein, Austria by G. Wallner, P. Pany & S. Ayromlou . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

397

46. Characterization of thoron and radon ﬂow-through sources by R. Rolle, M. Gründel, R. Schulz & J. Porstendörfer . . . . . . . . . . . . . . . . . . . .

404

47. Some aspects of the radon problem in Kazakhstan by V.N. Sevostyanov . .

409

48. Studies on the short-lived radon decay products: The inﬂuence of the unattached fraction on the measurement of the activity size distribution by M. Gründel, A. Reineking & J. Porstendörfer . . . . . . . . . . . . . . . . .

420

Contents

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

xvii

Radon exposure of the Greek population by D. Nikolopoulos, A. Louizi, A. Serefoglou & J. Malamitsi . . . . . . . . . . . . . . . . . . . . . . . . . . .

425

Preliminary study of two high radon areas in Greece by A. Louizi, D. Nikolopoulos, A. Serefoglou & J. Malamitsi . . . . . . . . . . . . . . . .

431

Year-to-year variations in radon levels in a sample of UK houses with the same occupants by N. Hunter, C.B. Howarth, J.C.H. Miles & C.R. Muirhead

438

The charged fraction of the 218 Po ions in air under environmental conditions by P. Pagelkopf, M. Gründel & J. Porstendörfer . . . . . . . . . . . . . . . .

448

Activity size distribution in outdoor air: Short-lived (214 Po, 214 Bi/214 Po) and long-lived (210 Pb, 210 Po) radon and thoron (212 Pb, 212 Po) decay products and 7 Be by M. Gründel, A. Reineking & J. Porstendörfer . . . . . . . . . . . . .

454

A comparison of the precision of (quasi-) continuous methods of measuring radon/thoron and decay product concentrations in air by R. Rolle . . . . . .

459

Correlation between radon concentration and geological structure of the Kraków area by K. Kozak, J. Swako´n, M. Paszkowski, R. Gradzi´nski, J. Łoskiewicz, M. Janik, J. Mazur, J. Bogacz, T. Horwacik & P. Olko . . . .

464

Application of a portable LS counter for measurements of radon progeny in air by S. Chalupnik & A. Kies . . . . . . . . . . . . . . . . . . . . . . . . . .

470

Calibration and quality assurance of radon measurements in Sweden. Activities in 1990–2000 by N. Hagberg . . . . . . . . . . . . . . . . . . . . . . . .

478

The diurnal change in the vertical distribution of atmospheric 222 Rn due to the growth and rise of the stable stratiﬁcation height in the atmospheric boundary layer by K. Yoshioka & T. Iida . . . . . . . . . . . . . . . . . . . . . . . . . .

489

A new method for supplying low radon air by using a hollow ﬁber module by T. Iida & T. Kato . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

497

Dose evaluation of indoor thoron progeny in some areas in China by Q. Guo, J. Cheng & Y. Chen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

506

EU-Concerted Action for a survey on radon exhalation rate measurements for building materials and soils by L. Roelofs, R. Wiegers, K.-H. Puch & G. Keller

512

Radon concentration survey in schools of the Friuli-Venezia Giulia region, North-East Italy by C. Giovani, C. Cappelletto, M. Garavaglia & R. Villalta

520

xviii

63.

Contents 210 Po

implanted in glass surfaces: Calibration and improved performance for retrospective radon reconstruction by D.J. Steck, J.A. Berglund & R.W. Field

528

64. Development of a method for estimating the airborne concentration of radon progeny, using an imaging plate by T. Iimoto, T. Kosako, N. Sugiura & K. Kawashima . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

535

65. Concentration, distribution and transportation of 222 Rn and its decay products in the environment of coastal Karnataka and Kaiga in southwest India by H.M. Mahesh, D.N. Avadhani & K. Siddappa . . . . . . . . . . . . . . . . .

542

66. Natural radiation levels in Tamil Nadu and Kerala, India by S. Tokonami, H. Yonehara, S. Akiba, M.V. Thampi, W. Zhuo, Y. Narazaki & Y. Yamada .

554

67. Natural radiation exposures for cave residents in China by S. Tokonami, Q. Sun, S. Akiba, T. Ishikawa, M. Furukawa, W. Zhuo, C. Hou, S. Zhang, Y. Narazaki, H. Yonehara & Y. Yamada . . . . . . . . . . . . . . . . . . . . .

560

68. A study of atmospheric radon transport as a tracer of pollutants over the Japan Sea by H. Aoshima, Y. Hashiguchi, J. Moriizumi, K. Yoshioka, Y.S. Kim & T. Iida . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

567

69. Non-destructive evaluation of concrete condition using radon exhalation monitoring: a feasibility study by E.R. van der Graaf & R.J. de Meijer . . . . . .

573

70. Study of Radon-222 exhalation of phosphogypsum blocks used as building materials. Comparison with modeling by F. Fournier, J.E. Groetz, F. Jacob, H. Lettner, A. Chambaudet & J.M. Crolet . . . . . . . . . . . . . . . . . . . .

582

71. 15 years repeated investigations of the radon levels in 105 remediated Swedish dwellings, mostly single-family buildings by B. Clavensjö . . . . . . . . . .

590

72. Comparison between long-term and short-term measurements for indoor radon risk mapping by F. Tondeur & I. Gerardy . . . . . . . . . . . . . . . .

598

73. Cancer risks from radon in indoor air and drinking water in Sweden. The Swedish Radiation Protection Authority’s risk assessment by L. Mjönes & R. Falk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

604

74. Review of seasonal variation in residential indoor radon concentrations by H. Arvela . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

612

75. Developments in radon-safe building in Finland by H. Arvela, J. Bergman, R. Yrjölä, J. Kurnitski, M. Matilainen & P. Järvinen . . . . . . . . . . . . . .

618

Contents

76.

77.

78.

79.

80.

81.

82.

83.

84.

85.

86.

87.

88.

89.

90.

xix

Stochastic radon lung dosimetry – modeling variability of bronchial cellular doses by R. Winkler-Heil & W. Hofmann . . . . . . . . . . . . . . . . . . . .

624

Stochastic state-vector model of radiation carcinogenesis applied to radoninduced lung cancer risk by D.J. Crawford-Brown & W. Hofmann . . . . . .

632

Biophysical mechanisms and radiation doses in radon therapy by H. Tempfer, A. Schober, W. Hofmann, H. Lettner & F. Steger . . . . . . . . . . . . . . .

640

Bioaerosols as carriers of radon progeny by S. Kagerer, T. Rettenmoser, W. Hofmann, A. Falkensteiner & F. Steger . . . . . . . . . . . . . . . . . . .

649

Radon in Finnish mines – regular monitoring since 1972 by M. Annanmäki, E. Oksanen, E. Venelampi & M. Markkanen . . . . . . . . . . . . . . . . . .

657

Technical guidance on safe installation of a GAC ﬁlter used for removing radon from water by M. Markkanen . . . . . . . . . . . . . . . . . . . . . . .

665

Outdoor radon and thoron in the USA, Canada, Finland and Thailand by N.H. Harley, P. Chittaporn, M. Heikkinen, R. Merrill & R. Medora . . . . .

670

Early thoron concentration levels in Guarapari: Dosimetric implications by A.S. Paschoa & J. Pohl-Rülling . . . . . . . . . . . . . . . . . . . . . . . . .

678

Indoor occupancy and radon exposure in Finland by I. Mäkeläinen, S. Moisio, H. Reisbacka & T. Turtiainen . . . . . . . . . . . . . . . . . . . . . . . . . . .

687

Indoor radon: controlling factors, deﬁnition of the radon potential and its geographical distribution over Austria by P. Bossew & H. Lettner . . . . . . .

694

Simultaneous measurement of radon and thoron exhalation rate from soil and building materials by C. Cosma, O. Cozar, T. Jurcut, C. Baciu & I. Pop . . .

699

Exposure to radon in the northern part of the Republic of Kyrgyzstan by M. Zhukovsky & R. Termechikova . . . . . . . . . . . . . . . . . . . . . . .

706

Stratospheric radon measurements in three North American locations by I.M. Fisenne, L. Machta & N.H. Harley . . . . . . . . . . . . . . . . . . . . .

715

Determination of soil-gas radon concentration in low permeability soils by Martin Neznal & Matej Neznal . . . . . . . . . . . . . . . . . . . . . . . . . .

722

Indoor radon long-term variation assessment by I.V. Yarmoshenko, Z.S. Zunic, J.P. Mc Laughlin, J. Paridaens, I.A. Kirdin & K. Kelleher . . . . . . . . . . .

726

xx

Contents

91. Quality assurance of individual radon measurements by T.R. Beck, J. Schwedt & P. Hamel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

731

92. Sensitivity analysis of an improved method for measuring the radon diffusion coefﬁcient of porous materials by W.H. van der Spoel & M. van der Pal . .

740

93. Dose to the fetus from 222 Rn in maternal drinking water by E.S. Robbins & N.H. Harley . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

749

94. Prenatal exposure due to ingestion of natural radionuclides by S. Risica & C. Nuccetelli . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

756

95. Meta-analysis of twenty radon and lung cancer case–control studies by I.V. Yarmoshenko, I.A. Kirdin, M.V. Zhukovsky & S.Y. Astrakhantseva . .

762

96. A new Austrian recommendation guide for radon prevention in the design and construction of new buildings in areas with highly elevated radon levels by F.-J. Maringer, P. Schillfahrt, T. Auer & R. Pecina . . . . . . . . . . . . .

772

97. The new radon programme in England by B.M.R. Green & L. Davey . . . .

779

98. Idea of assessment of the annual dose when non-continuous data exist by D. Kluszczy´nski . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

788

99. A passive technique to measure radon progeny surface deposition variations in rooms and chambers by J.P. Mc Laughlin & C. Walsh . . . . . . . . . . .

794

100. Radon on selected underground tourist routes in Poland by J. Olszewski, W. Chru´scielewski, J. Kacprzyk, D. Kluszczy´nski & Z. Kami´nski . . . . . .

803

101. Indoor radon in the karst region of Lithuania by L. Pilkyt˙e, G. Mork¯unas & G. Åkerblom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

807

102. The behaviour of Rn-222 decay products at the air–glass interface and its implication for retrospective radon exposure estimates by B. Roos & C. Samuelsson . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

813

103. Radon transfer from ground to houses and prediction of indoor radon in Germany based on geological information by J. Kemski, R. Klingel, A. Siehl & R. Stegemann . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

820

104. Checking the “10 point system” for an evaluation of the soil radon potential by D. Bleile & J. Wiegand . . . . . . . . . . . . . . . . . . . . . . . . . . . .

833

105. Radon-in-water secondary standard preparation by D.J. Karangelos, N.P. Petropoulos, E.P. Hinis & S.E. Simopoulos . . . . . . . . . . . . . . . . . . .

842

Contents

xxi

5. Cosmic radiation

849

106. Sequential measurements of energy spectrum and intensity for cosmic ray neutrons by T. Nakamura, T. Nunomiya, N. Hirabayashi, H. Suzuki, Y. Sato & D.A.H. Rasolonjatovo . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

851

107. Cosmic radiation at aircraft altitudes calculated by CARI-6 and its comparison with measurements by K. Fujitaka, M. Okano, Y. Uchihori, T. Koi & H. Kitamura . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

858

108. Factors and processes controlling the concentration of the cosmogenic radionuclide 7 Be at high-altitude Alpine stations by E. Gerasopoulos, P. Zanis, C.S. Zerefos, C. Papastefanou, W. Ringer, H.W. Gäggeler, L. Tobler & H.J. Kanter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

863

109. Aircrew exposure assessment by means of a Si-diode spectrometer by F. Spurný & T. Dachev . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

871

110. Practical considerations for implementation of the EU Directive in the ﬁeld of cosmic radiation exposure by D. Irvine & D.J.C. Flower . . . . . . . . . . .

876

111. Monitoring of the cosmic radiation on IBERIA commercial ﬂights: One year’s experience of in-ﬂight measurements by J.C. Saez Vergara, A.M. Romero Gutiérrez, R. Rodriguez Jiménez, R. Dominguez-Mompell Román, P. Ortiz García & F. Merelo de Barberá . . . . . . . . . . . . . . . . . . . . . . . . . .

885

112. Health aspects of radiation exposure on a simulated mission to Mars by W. Friedberg, K. Copeland, F.E. Duke, K. O’Brien & E.B. Darden Jr. . . . .

894

113. Observation of dose-rate increases during winter thunderstorms and MonteCarlo simulation of bremsstrahlung generation by T. Torii, M. Takeishi & K. Okubo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

902

114. Measurements of cosmic radiation doses during air travel by a handy method with an electronic personal dosemeter by M. Furukawa . . . . . . . . . . . .

909

115. Twenty years of TLD measurements on board space vehicles by the Hungarian “Pille” system by S. Deme, I. Fehér, I. Apáthy, G. Reitz & Yu. Akatov .

916

116. Variation of radiation exposure as a function of altitude between 8 and 30 km by B.J. Lewis, L.G.I. Bennett, A.R. Green, A. Butler, M.J. McCall & B. Ellaschuk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

926

117. Neutron dosimetry onboard aircraft using superheated emulsions by M. Hajek, T. Berger, N. Vana & B. Mukherjee . . . . . . . . . . . . . . . . . . . . . . .

941

xxii

Contents

118. Measurements and calculations of the radiation exposure of aircrew personnel on different ﬂight routes by M. Hajek, T. Berger, L. Summerer, W. Schöner & N. Vana . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

948

119. New designs of optical multilayer ﬁlters for extending the detection of the highest-energy cosmic rays by atmospheric ﬂuorescence optical telescopes in the presence of crescent moonlight by E. Fokitis, S. Maltezos, P. Moyssides & A. Geranios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

955

6. Technologically enhanced natural radiation

965

120. Radiological impact of ﬂue gas puriﬁcation in a coal-ﬁred power plant in Belgium by J. Paridaens & H. Vanmarcke . . . . . . . . . . . . . . . . . . . . . .

967

121. Towards the identiﬁcation of work activities involving NORM in Italy by F. Trotti, S. Bucci, B. Dalzocchio, C. Zampieri, M. Lanciai, C. Innocenti, S. Maggiolo, L. Gaidolﬁ & M. Belli . . . . . . . . . . . . . . . . . . . . . . .

973

122. Impact of radium-bearing mine waters on the natural environment by S. Chalupnik . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

985

123. Naturally occurring radioactivity in industrial by-products from coal-ﬁred power plants, from municipal waste incineration and from the iron- and steelindustry by K.-H. Puch, R. Bialucha & G. Keller . . . . . . . . . . . . . . . .

996

124. An assessment of the radiological consequences of using phosphorus slag in concrete foundation poles by E.R. van der Graaf & R.J. de Meijer . . . . . .

1009

125. Orphan sources of NORM/TENORM throughout the World by A.S. Paschoa & F. Steinhäusler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1017

126. Radioenvironmental survey of the Megalopolis power plants ﬂy ash deposits by D.J. Karangelos, P.K. Rouni, N.P. Petropoulos, M.J. Anagnostakis, E.P. Hinis & S.E. Simopoulos . . . . . . . . . . . . . . . . . . . . . . . . . .

1025

127. The radiological impact of naturally occurring radionuclides in the oil and gas industry on the UK population by S. Warner Jones & K. Smith . . . . . . . .

1030

128. Depleted uranium: some remarks on radiation protection by M. Grandolfo, A. Mele, C. Nuccetelli & S. Risica . . . . . . . . . . . . . . . . . . . . . . . .

1040

129. The impact on man and the environment from the military use of depleted uranium: worst-case scenario by F. Steinhäusler & A.S. Paschoa . . . . . . .

1047

Contents

xxiii

130. Experimental results on the environmental samples collected around sites in South Serbia, Kosovo and Montenegro where DU weapons were deployed in 1999 by P. Gaca, Z.S. Žunic, J.W. Mietelski, E. Tomankiewicz & M.P.R. Waligórski . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1056

131. Developments in the management of exposures from radon in natural gas in the UK by D.W. Dixon & C.K. Wilson . . . . . . . . . . . . . . . . . . . . .

1064

132. Remediation case study of a coal ﬁred power plant tailings pond by P. Szerbin, L. Juhász, I. Csige, A. Várhegyi, J. Vincze, T. Szabó & F.-J. Maringer . . .

1071

133. The release of radium from scales produced in the North Sea oil ﬁelds by S. Ghose & B. Heaton . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1081

134. The radium concentration in groundwater at a waste disposal site in Brazil. Is it naturally occurring or a contaminant? by M.H. Magalhães, R. Zenaro & D.C. Lauria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1090

7. Internal and external exposure

1097

135. Exposure of the population through mineral water consumption by E. Botezatu, O. Iacob, G. Elisei & O. Capitanu . . . . . . . . . . . . . . . . . . . . . . . .

1099

136. Lung cancer risk in humans and rats: single vs. multiple exposures by H. Fakir, W. Hofmann, I. Aubineau-Laniece, R.S. Caswell, J.R. Jourdain & A. Sabir

1108

137. Quantiﬁcation of radon-progeny deposition on the skin in underwater radontherapy by H. Lettner, W. Hofmann, H. Tempfer & A. Schober . . . . . . . .

1116

138. A tentative method to evaluate the building material contribution to indoor gamma dose rate by C. Nuccetelli, C. Bolzan, F. Bochicchio & SETIL Working Group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1123

8. High background areas

1129

139. Distribution and behaviour of natural radionuclides in soil samples of Goa on the southwest coast of India by D.N. Avadhani, H.M. Mahesh, N. Karunakara, Y. Narayana, H.M. Somashekarappa & K. Siddappa . . . . . . . . . . . . . .

1131

140. Apparent lack of radiation susceptibility among residents of the high background radiation area in Ramsar, Iran: can we relax our standards? by S.M.J. Mortazavi & P.A. Karam . . . . . . . . . . . . . . . . . . . . . . . . .

1141

141. Natural radioactivity and radiation exposure in the high background area of the Chhatrapur beach placer deposits of Orissa, India by D. Sengupta, A.K. Mohanty, S.K. Das & S.K. Saha . . . . . . . . . . . . . . . . . . . . . .

1148

xxiv

Contents

9. Effects on biota and ecosystems

1153

142. Radiocesium gamma dose rates in a Greek pine forest: measurements and Monte Carlo computations by A. Clouvas, S. Xanthos, M. AntonopoulosDomis & D.A. Alifragis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1155

143. Distribution coefﬁcients and concentration factors of 226 Ra and 228 Th in the Greek marine environment by G. Trabidou, H. Florou, P. Kritidis, Ch. Chaloulou & Ch. Lykomitrou . . . . . . . . . . . . . . . . . . . . . . . .

1167

List of Attendees . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1177

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1189

1. Invited

T . his Page Intentionally Left Blank

3

Historical development of the Natural Radiation Environment Symposia A.C. George Radon Testing Corporation of America, 2 Hayes Street, Elmsford, NY 10523, USA

The idea for a Natural Radiation Environment Symposium was born by Wayne Lowder of the Health and Safety Laboratory (HASL) of the US Atomic Energy Commission (USAEC) and Professor John Adams of the Geology Department of Rice University Houston Texas. Wayne’s curiosity about an article that appeared on page one of The New York Times on December 3, 1961 triggered a telephone conversation with John Adams. The Times article was based on a paper that appeared in the proceedings of the National Academy of Sciences (NAS) in November 1961. The NAS original paper was about the thorium content of New Hampshire granite and how it might provide for an abundant energy source in a breeder reactor. Both John and Wayne studied the geological anomaly of New Hampshire from different perspectives and with different types of ﬁeld instruments. In the summer of 1962, they had the opportunity to intercompare their respective ﬁeld gamma spectrometers in New Hampshire. That ﬁeld exercise was probably the ﬁrst of its type and demonstrated the need for more intercomparisons of instruments and techniques in the future. Field exercises have now become an important feature of most NRE Symposia and workshops. The New York Times article presented an opportunity to study and utilize the New Hampshire survey data from an area of unusually high natural gamma levels and exposure, resulting from the high thorium content of Conway granite bedrock. When the New Hampshire data became available, HASL collaborated with the Harvard School of Public Health to measure the population exposure to natural radiation in the high area (New Hampshire) and in a low background area in nearby Vermont. The results from these two localities were published in Nature and later in the proceedings of NRE II. During the interaction between the two pioneers, the USAEC had a great interest in studying natural radiation background and its variation to provide some perspective on the signiﬁcance of the high levels of fallout in the environment. The Health and Safety Laboratory (HASL) of the AEC at that time, was measuring fallout gamma levels, mostly from 95 Zr–Nb and 137 Cs, which in some instances were comparable to outdoor natural gamma radiation levels. In fact, Lowder ﬁrst became involved with natural radiation when, in response to a reRADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07001-9

© 2005 Elsevier Ltd. All rights reserved.

4

A.C. George

quest from Willard Libby (a Nobel-prize winner) and then an AEC commissioner, he prepared a very useful bibliography of previous publications in the ﬁeld [1]. Another important intercomparison was between ion chambers and ground-based and airborne gamma spectrometers for environmental radiation and radionuclide measurements. The beneﬁts of the intercomparisons were tremendous. For example, the ﬁeld intercomparisons of ion chambers showed that gamma measurements in Europe and the USA, were quite compatible, but there were signiﬁcant problems in ion chamber calibrations with respect to cosmic radiation, which were later, resolved. After long discussions, John and Wayne came to the conclusion that it would be desirable and beneﬁcial to standardize methods and techniques to obtain more accurate and detailed knowledge of the natural radiation base line of the human exposure. The two pioneers also realized that important data regarding natural radiation were scattered throughout the literature of a number of science disciplines making access difﬁcult. They decided that, by bringing together scientists from different disciplines working on natural radiation, the end result would be a uniﬁed source of information that could provide the tools and data to assess quantitatively the signiﬁcance of additional radiation exposure from man-made sources. They saw the need for an International Forum that would bring some coherence to the scattered research and data. The exchange of more ideas between the two pioneers bore fruit by organizing the First International Symposium on the Natural Environment, April 10–13, 1963. I started working at HASL in October 1963 and I regretfully missed the ﬁrst NRE. I am happy to say that I participated in all the others including the present NRE VII on the beautiful Island of Rhodes.

1. Why measure natural radiation During his address at the banquet of NRE II [2] in Houston in 1972, Merril Eisenbud of the USAEC, talked about the Natural versus the Unnatural, maintaining that “the naturally radioactive environment is an important subject not only as a fascinating area of scientiﬁc inquiry but because the information that evolves from such an investigation provides the perspective needed to evaluate the signiﬁcance of the various levels and types of ionizing radiation exposure encountered when using nuclear energy.”

2. When it all began Our knowledge of natural radiation goes back to early 1896, when Antoine Henri Becquerel noted the fogging of a photographic plate by potassium uranyl sulfate. Within the next ﬁve years, the radioactivity of thorium, polonium, radium and radon had been established. Some of the ﬁnest original work in the history of science was done in the ﬁeld of Natural Radioactivity. The early pioneers of physical sciences between 1900–1920, such as Soddy, Rutherford, Lord Strutt and the Curies, were giants in the early history of sciences and they provided the foundation for much of what we know today about the Natural Radiation Environment. They were all recipients of the Nobel Prize in Physics or Chemistry.

Historical development of the Natural Radiation Environment Symposia

5

In 1910–1912, on a series of balloon ﬂights, Victor Hess (a Nobel prize-winner) and others discovered cosmic radiation. By 1914, both cosmic and terrestrial radiations had been measured and their most important natural radioactive elements were identiﬁed. With the development of nuclear weapons and technology following World War II, a lot of attention was given to uranium and thorium prospecting, including their concentrations in different geological materials. In the mid-1950s, concern about exposure from the fallout of nuclear weapons testing, stimulated a new interest in background radiation. Over subsequent decades, extensive surveys were conducted in the ﬁeld and bibliographies were published about every ﬁve years in the various reports of the United Nations Scientiﬁc Committee on the effects of Atomic Radiation (UNSCEAR), as well as in several reports of the National Council on Radiation Protection and Measurements (NCRP). In the late 1960s HASL of the USAEC, embarked on a measurement strategy to deﬁne the extent of human inhalation exposure from radon and radon decay products, the main natural radiation contributors to human radiation exposure. The characterization of uranium mine atmospheres in the late 1960s and early 1970s [3], provided data that was used in lung dosimetry to assess the health risk from radon and radon decay products. Also a variety of radon and radon decay measurement methods and techniques were developed, which with some modiﬁcation are still in use today to characterize the indoor radon environment. Until 1984 research on radon in the USA, was done by USAEC, USERDA and then by USDOE, mostly investigating the conditions in uranium mines, uranium mills and tailings and at the USAEC excessed sites. Measurements of natural radiation were made in a number of sites with technologically enhanced radiation exposures to occupational workers as well as the general public. I was personally involved in some of these site investigations from 1974 to 1978. The uranium mine and mill tailings and the waste from the phosphate industry are two examples of enhanced radiation exposure. I believe once this was discovered, we focused not only on gamma radiation levels but the measurement and concentration of alpha radiation emitters from radon, thoron and their decay products. In Europe, many countries were aware of the radon problem and began indoor radon surveys to map out the areas of high radon levels. I recall the visit of some Swedish scientists at HASL in New York in the late 1970s, to discuss mutual concerns about indoor radon and exchange information on instrumentation and testing protocols. I believe every country in Europe has conducted a national radon survey by now. With the discovery of high radon concentrations in US homes in 1984, the US EPA recognized that the US had a serious radon problem with some of the highest indoor radon levels in the world. The EPA began to develop and regulate a program to address the problem and to assess the health risk of the American public from indoor radon. In the late 1980s in the USA, a radon industry evolved with more than 1000 radon measurement ﬁrms and several hundred radon mitigation companies. At the same time, EPA began to develop a proﬁciency program to oversee the quality of measurements. As of this writing about 18 million measurements have been made in the USA with an estimated 6 million homes found to have indoor radon concentration levels above 150 Bq m−3 . The concern about indoor radon became global, and the bulk of the papers in the Symposia and workshops on the Natural Radiation Environment were related to radon and radon decay

6

A.C. George

products. Since 1963, after the ﬁrst NRE Symposium, thousands of papers on radon were published in many scientiﬁc journals and laboratory reports. The USDOE in cooperation with CEC, IAEA, the USEPA and some universities addressed the radon problem and exchanged information through special meetings, workshops and Symposia such as the Anacapri meeting in 1983, the Lisbon NRE IV in 1987, the Salzburg NRE V in 1991, the workshop at Rimini in 1993, the NRE VI in Montreal in 1995 and the Athens workshop in 1999. The list of Symposia, workshops and seminars that I am about to present is not meant to be complete, but will include all the previous NRE Symposia and a few other major seminars and workshops on Natural Radiation. The First NRE Symposium was held at Rice University Houston Texas, on April 10–13, 1963 [4]. It coincided with the 50th anniversary of Rice University. The organizers and hosts were Rice University Department of Geology (key person, John Adams) and the Health and Safety Laboratory (HASL) of the US Atomic Energy Commission (USAEC) (key person, Wayne Lowder). The Division of Radiological Health of the US Public Health Service also provided ﬁnancial support. About 200 scientists participated in the 1963 NRE Symposium. This is a large number considering this was the ﬁrst of its kind. About half of the scientists were authors or co-authors of the presented papers. The ﬁrst day of the symposium was devoted to intercalibrations and intercomparisons of the various instruments used by the different groups to study natural radioactivity. There were ﬁve participants from the US and one from the UK. Three locations in and about Houston were used for the measurements. Sixty-one papers on a variety of subjects representing all aspects of the Natural Radiation Environment were presented during the remaining three days. Only ten papers were related to radon and radon decay products. The proceedings were published in a very impressive hard cover book, 1070 pages long. At the conclusion of the symposium, the participants very enthusiastically expressed the desire to plan and convene another such symposium within some reasonable time because the ﬁst gathering proved to be very fruitful and useful in exchanging information on a subject covered previously by many disciplines. The Second International Symposium on the Natural Radiation Environment was held at Houston again, August 7–11, 1972 [2]. The organizers were Rice University (John Adams) and the University of Texas School of Public Health (Thomas Gesell), and HASL, USAEC (Wayne Lowder). About 110 scientists from 20 countries attended the various sessions and 55 papers were presented. About a dozen papers were related to radon. The proceedings were published in two volumes totaling 960 pages. The proceedings were dedicated to Professor Yasuo Miyake of Japan and Professor K.Z. Morgan of the US Oak Ridge National Laboratory. At the conclusion of the symposium, a decision was made to hold large-scale symposia on NRE at four to ﬁve year intervals and more limited meetings and workshops during the intervening periods. The ﬁrst and second informal workshops on the Natural Radiation Environment were organized by HASL, USAEC in New York in March 1972 [5] and February 1974 [6], respectively. Both workshops provided a forum to identify areas of further research and set the stage for the development of cooperative research programs for information exchange. Most US laboratories conducting studies in environmental radiation were represented. The proceedings of the HASL workshop were published as USAEC reports.

Historical development of the Natural Radiation Environment Symposia

7

The First Special International Mini-Symposium on the Natural Radiation Environment was held at Pocos de Caldas, Brazil, June 16–20, 1975 [7]. The topic was “Areas of High Natural Radioactivity and their Biological Implications.” Ninety-three scientists attended the conference from 16 countries and discussed a wide range of geological radioactive anomalies. The proceedings were published by the Brazilian Academy of Sciences incorporating 14 presentations by eminent scientists and 10 additional abstracts totaling 190 pages of text. The objectives of the mini-Symposium were: (1) to inventory the places with radioactive anomalies; (2) to review the status of the investigations being conducted, and (3) to assess the information on the effects of low-level exposure to ionizing radiation. At the conclusion of the meeting, an anomalous radioactive environment was deﬁned as one that is characterized by the following. (a) The exposure rate from internal sources is > 200 mrem yr−1 (> 2 mSv yr−1 ). (b) The long-lived alpha activity ingested from the local water and diet is > 50 pCi day−1 (1.9 Bq day−1 ). (c) The 222 Rn concentration in drinking water is > 5000 pCi L−1 (> 185 Bq L−1 ). (d) The 220 Rn and 222 Rn in atmospheric air is > 1 pCi L−1 (> 37 Bq m−3 ). The Tenth Midyear Topical Symposium of the Health Physics Society on the Natural Radiation Environment was held at Saratoga Springs, New York, October 11–13, 1976. This coincided with the Bicentennial Celebrations. The program committee was Robert Ryan of Rensselaer Polytechnic Institute representing the Northeastern New York Local chapter of the Health Physics Society, John Matuszek, New York State Department of Health and James McLaughlin, HASL USERDA. A total of 50 papers were presented ranging from altered radiation environments, dose evaluation, environmental radiation ﬁelds, instrumentation and problems of radiological monitoring. The published proceedings totaled 575 pages [8]. A Workshop on Methods for Measuring Radiation in and Around Uranium Mills was held May 23–26, 1977, in Albuquerque, New Mexico [9]. It was organized by USERDA, the American Nuclear Society, the Health Physics Society and the Atomic Industrial Forum. There were 205 participants and 33 papers were presented on the environmental aspects of uranium mill operations, gamma radiation surveys around the mills and radon and radon decay product measurements. The published proceedings totaled 410 pages. The Third NRE Symposium was held on April 23–28, 1978 in Houston, Texas, the same site as for the previous symposia in the series in 1963 and 1972 [10]. The symposium was sponsored by the US Department of Energy (USDOE), Ofﬁce of Health and Environmental Research (key person Wayne Lowder) and the University of Texas School of Public Health (key person Thomas Gesell). The growing interest in NRE was reﬂected in the number of papers as compared with the two previous symposia. Present were 184 scientists from 20 countries. At the end of the meeting an informal intercomparison of instruments and techniques was held near Burnet Texas. Eleven laboratories from 5 countries participated in measuring radon and radon decay product concentrations as well as external penetrating radiation. The beneﬁt of these intercomparisons was obvious in providing an opportunity for new instruments and techniques to be compared. Of the 108 presentations, about one half were related to radon. The proceedings were published in two large volumes totaling 1736 pages.

8

A.C. George

A Second Special Mini-Symposium on NRE was held January 19–23, 1981 at the Bhabha Atomic Research Center in Bombay, India [11]. The hosts were the Bhabha Atomic Research Center and the Atomic Energy Government of India. Eighty-ﬁve abstracts were published of which about 35 were related to radon. The published proceedings were 177 pages long. The Le Vesinet Seminar on the Radiological Burden of Man from Natural Radioactivity in the Countries of the European Community, organized by the Commission of European Community (CEC), was held in Paris, December 1979. The papers related to radon represented 20% of the total. The proceedings were published by CEC, Luxembourg An International Seminar on Indoor Exposure to Natural Radiation and Associated Risk Assessment was held at Anacapri on the island of Capri, Italy, October 3–7, 1983 [12]. The seminar was organized by CEC (key person, Jacques Sinnaeve) and by ENEA, Italy (key person, Gian Clemente). The seminar covered the implications of exposure to low doses arising from indoor radon, thoron and their decay products and from indoor gamma exposure. I believe, that at that time, the concern for radon dominated the subject of natural radiation and brought together many radon rangers to the island of Capri to share their experiences on the newly emerging indoor air pollutant. All of the 85 presentations were related to radon, a ﬁrst of its kind, with emphasis on radon in the indoor environment. The proceedings of the seminar were published in the Journal of Radiation Protection Dosimetry, totaling 435 pages. At the conclusion of the seminar, three National Laboratories intercompared their methods for measuring radon and radon decay products at NRPB, Oxford, UK. A Seminar on the Exposure to Enhanced Natural Radiation and its Regulatory Implications was held in Maastricht, The Netherlands, March 23–27, 1985 [13]. It was sponsored by the Dutch Administration. The proceedings were published in the Science of the Total Environment 45 (1985). The papers related to radon represented 55% of the total. The Fourth International Symposium on the Natural Radiation Environment was held in Lisbon Portugal December 7–11, 1987 [14]. It was organized by the Commission of European Community (CEC), the National Laboratory of Engineering Industrial Technology of Portugal and the US Department of Energy, Ofﬁce of Health and Environmental Research. More than 200 scientists from 25 countries and six continents participated. A total of 120 papers were presented of which 65% were related to radon, covering radon behavior, metrology, surveys, dose and risk assessment and mitigation techniques. The published proceedings totaled 554 pages. The Fifth International Symposium on the Natural Radiation Environment was held in Salzburg, Austria, September 22–28, 1991 [15]. It was organized by the Commission of the European community (CEC), USDOE, the International Atomic Energy Agency (IAEA) and the University of Salzburg. The Symposium covered many important topics such as radon and thoron, the radiological impact of non-nuclear industrial releases and the exposure of workers and the general public to natural sources of radiation in non-domestic environments. There were 163 paper presentations of which 70% were on radon and thoron. The proceedings were published in the Journal of Radiation Protection Dosimetry, totaling 774 pages. After the conclusion of the Symposium many groups participated in an intercomparison exercise at Bad Gastein, Austria. The intercomparisons conducted inside the spa cavern, were for the methods and techniques used by the participants in their respective laboratories to measure radon and radon decay products. Some participants intercompared their respective radon ﬂux measurement techniques on nearby soils.

Historical development of the Natural Radiation Environment Symposia

9

The First International Workshop on Indoor Radon Remedial Action – The Scientiﬁc Basis and the Practical Implications, was held in Rimini, Italy [16], June 27–July 2, 1993. It was organized by the Commission of European Community (CEC), USDOE, USEPA, the Italian ENTEA, the International Center for Theoretical Physics and the International Center for Theoretical and Applied Ecology with contribution by the city of Rimini. Because of the increased awareness of the potential health impact of indoor radon, the scope of the workshop was enlarged to include health effects (biological as well as physical), surveys and policy matters. There were 211 participants from many countries. In all, there were 81 presentations, of which 78 dealt with radon issues and 3 were related to thoron. The proceedings totaling 375 pages were published in the Journal of Radiation Protection Dosimetry. The Sixth International Symposium on the Natural Radiation Environment was held In Montreal, Canada, June 5–9, 1995 [17]. The organizers and ﬁnancial supporters were Clarkson University Potsdam, NY (key person, Philip Hopke), USDOE, USEPA, CEC, the Atomic Energy Control Board of Canada and the Health and Welfare of Canada. About 300 scientists attended and 150 papers were presented. One hundred and ﬁve papers were related to radon and another ﬁve to thoron. The topics included natural radioactivity sources including enhancement of the natural radioactivity by human activities, concentrations and transfer pathways, environmental impacts of natural radiation, health effects, remediation technology, exposure dose, and risk assessment and risk management. The proceedings were published in a special issue of the Journal of Environment International, totaling 1153 pages. The week following NRE VI, June 12–15, 1995, EML DOE, hosted the Sixth International Radon Metrology and Intercomparison Workshop in New York [18]. Thirty participants from different institutions and from 11 countries attended. The laboratory exercises consisted of 220 Rn and 222 Rn concentration measurements and exhalation measurements from a radium spiked concrete slab. Field exercises included soil gas-radon measurements and radon exhalation in New Jersey soils. The most recent International Workshop, Radon in the Living Environment was held in Athens, Greece, April 19–23, 1999 [19]. It was organized by the National Technical University of Athens (key person Prof. Simopoulos), the European Commission and the European Union Concerted Actions, ERRICA (European Research into Radon in Construction). The workshop coordinators were the UK Building Research Establishment and the Federal Ofﬁce for Radiation Protection, Berlin, Germany. More than 215 research and radon specialists from 32 countries participated. A total of 170 papers were presented. At this workshop, there were many young researchers from Eastern Europe, an encouraging sign that radon research in Europe is alive and well. The papers, all related to radon, covered many topics such as geology, metrology retrospective techniques, health effects, radon resistant new construction, remediation, modeling and radon surveys. The key persons at the Athens Workshop are responsible for the planning of the current Seventh Symposium on the Natural Radiation Environment on the beautiful Island of Rhodes. In this, the NRE VII Symposium, more than 260 scientists from 42 countries are participating. A total of about 300 abstracts have been accepted for presentation either as platform (oral) papers or as posters. In recent years, there has been a great interest in cosmic radiation and the potentially signiﬁcant exposures of aircraft crew and frequent ﬂiers. In addition to papers at the various NRE Symposia, there were several special meetings on this subject including a CEC semi-

10

A.C. George

nar in Luxembourg in 1989, an FAA sponsored meeting in Oklahoma in 1992 and the NCRP Annual Meeting in 1998. Many of the techniques developed and used for background radiation and radon and radon daughter measurements have been applied to critical environmental measurements at nuclear facilities after accidents or unusual releases (TMI, Chernobyl, Mayak) or during decommissioning. The evaluation of contamination at and near nuclear facilities was accomplished using techniques and instrumentation developed for natural radiation background measurements. A good example is Kevin Miller’s work with ion chambers and ﬁeld gamma spectrometry at various facilities [20].

3. Conclusions We have come a long way since the First Symposium on the Natural Radiation Environment in 1963. In the last 40 years, thousands of pages of good science and good information were shared by two generations of scientists and researchers. It is my hope that this trend will continue in the future in this constantly changing world. When you leave this current gathering and go back to your laboratories or place of work, think about new projects around the Natural Radiation Environment and keep the series going for another 40 years and beyond. I hope that in 4–5 years from now, NRE VIII will become a reality in another exciting place such as the one we are in today. For now, I wish you a great meeting and good luck.

References [1] W.M. Lowder, L.R. Solon, Background radiation: A literature search, USAEC Report NYO-4712, 1956. [2] M. Eisenbud, in: The Natural Radiation Environment II, USERDA, Oak Ridge, TN, 1974, pp. 941–947, CONF720805, vol. 2. [3] A.C. George, I. Hinchcliffe, R. Sladowski, Size distributions of radon daughter particles in uranium mine atmospheres, Am. Ind. Hyg. Assoc. J. 36 (1975) 484–490. [4] J.A. Adams, W.M. Lowder (Eds.), The Natural Radiation Environment, University of Chicago Press, Chicago, 1964. [5] J.E. McLaughlin (Ed.), USAEC Report HASL-269, 1972. [6] W.M. Lowder (Ed.), USAEC Report HASL-287, 1974. [7] International Symposium on Areas of High Natural Radioactivity, Pocos de Caldas, Brazil, June 16–20, 1975. [8] Tenth Midyear Topical Symposium of the Health Physics Society, Northeastern New York Chapter, Rensselaer Polytechnic Institute, Troy, NY 12181, 1976. [9] Workshop on Methods for Measuring Radiation in and Around Uranium Mills, Atomic Industrial Forum 3 (9) (1977). [10] T.F. Gesell, W.M. Lowder (Eds.), Natural Radiation Environment III, DOE Symposium Series, vol. 51, USDOE, Oak Ridge, TN, 1980, Report CONF-780422. [11] Second Special Symposium on the Natural Radiation Environment. Book of Abstracts, Bhabha Atomic Research Center, Bombay, India, 1981. [12] Proceeding of the International Seminar on Indoor Exposure to Natural Radiation and Related Risk Assessment, Anacapri, Italy, October 3–5, 1983, Radiat. Prot. Dosim. 7 (1–4) (1984). [13] Proceedings of the Seminar on Exposure to Enhanced Natural Radiation and its Regulatory Implications, Maastricht, The Netherlands, March 25–27, 1985, Sci. Total Environ. 45 (1985). [14] Proceedings of the Fourth International Symposium on the Natural Radiation Environment, Lisbon, Portugal December 7–11, 1987, Radiat. Prot. Dosim. 24 (1–4) (1988).

Historical development of the Natural Radiation Environment Symposia

11

[15] Proceedings of the Fifth International Symposium on the Natural Radiation Environment, Salzburg, Austria, September 22–28, 1991, Radiat. Prot. Dosim. 45 (1–4) (1992). [16] Proceedings of the First International Workshop on Indoor Radon Remedial Action. The Scientiﬁc Basis and Practical Implications, Rimini, Italy, June 27–July 2, 1993, Radiat. Prot. Dosim. 56 (1–4) (1994). [17] The Natural Radiation Environment VI, Montreal, Canada, June 5–6, 1995, Environ. Int. 22 (Suppl. 1) (1996). [18] A.R. Hutter, E.O. Knutson, Health Phys. J. 74 (1998) 108–114. [19] Proceedings of a Workshop on Radon in the Living Environment, Athens, Greece, April 19–23, 1999, Sci. Total Environ. 272 (1–3) (2001). [20] K.M. Miller, R. Larsen, The development of ﬁeld-based measurement methods for radioactive fallout assessment, Health Phys. J. 82 (2002) 609–625.

12

Understanding health risks from low doses of ionizing radiation∗ D.G. Thomassen, A. Patrinos Ofﬁce of Biological and Environmental Research, SC-72 Ofﬁce of Science, US Department of Energy, 19901 Germantown Road, Germantown, MD 21701-1290, USA

Epidemiological and toxicological research has long been used to characterize health responses of populations and individuals to high radiation doses. This information is used to set exposure standards to protect the public and the workforce. Standards are set using models that predict unmeasurable cancer frequencies at low radiation doses by extrapolating from the number of cancers observed following high dose exposures. Recently, new techniques and instruments have been developed that enable direct measurements of the biological changes induced by low doses of radiation. These research tools provide new information to help deﬁne radiation’s measurable effects on cells and molecules at environmentally relevant exposures. Such research will underpin the development of future science-based radiation risk regulatory policy. Microbeams are being developed that can expose individual cells, or speciﬁc parts of cells such as the nucleus, the cytoplasm, or even speciﬁc regions of the nucleus or cytoplasm, to a wide range of radiation energies (from heavy ions to electrons) and doses (including the ultimate low dose – a single ion). These instruments will enable critical questions about the biological effects of very low doses of radiation to be addressed. Rapid progress in the Human Genome Project has enabled studies on the effects of radiation at the level of the individual gene. The biological effects of high, carcinogenic doses of radiation can now be compared directly with the effects of low doses for which there is only sparse and conﬂicting data as to whether risks for selected cancers increase or decrease. However, critical questions remain. Are the cellular effects induced by high and low radiation doses identical at the molecular level, differing only quantitatively in proportion to dose? Do cells recognize changes induced by low doses of radiation the same way they recognize changes induced by high doses? Do cells, tissues and whole organisms each respond in a qualitatively similar way to high and low doses of radiation? * The research summaries described here are not the result of primary scientiﬁc research by the authors but were reported by the principal investigators at a meeting of scientists funded by the DOE Low Dose Radiation Research Program held March 25–27, 2002. For meeting abstracts, see the program web site at www.lowdose.org.

RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07002-0

Published by Elsevier Ltd.

Understanding health risks from low doses of ionizing radiation

13

Observations such as the by-stander effect demonstrate that unirradiated neighbors of cells exposed to low doses of radiation can exhibit biological responses as if they had been irradiated. The biological mechanisms responsible for this and other phenomena such as the adaptive response and genomic instability remain to be determined, but they all suggest that the effects of low doses of radiation cannot be understood by simply extrapolating from the effects seen at high doses. Using emerging biological data to help predict health risks from low doses of radiation is itself a difﬁcult but crucial challenge. For this research to have an impact on radiation risk regulatory policy, the new data need to be incorporated into biologically plausible mathematical models that predict risks from low doses of radiation. The US Department of Energy’s Low Dose Radiation Research Program (www.lowdose.org) is speciﬁcally supporting research to determine health risks from exposures to low levels of radiation. Together with the results from complementary research programs around the world, the new scientiﬁc information generated in this research program is critical to adequately and appropriately protect people from radiation and to make the most effective use of national resources.

1. Introduction Epidemiological and toxicological research has long been used to characterize health responses of populations and individuals to high radiation doses. This information is used to set exposure standards to protect the public and the workforce. Standards are set using models that predict unmeasurable cancer frequencies at low radiation doses by extrapolating from the number of cancers observed following high dose exposures. The US Department of Energy’s Low Dose Radiation Research Program (www.lowdose.org) was established in 1997, with encouragement from Senator Pete Domenici (Republican Senator from New Mexico), to improve the scientiﬁc basis for the development of risk protection policies for ionizing radiation that adequately and appropriately protect both workers and the public while at the same time making the best use of public funds. A driving reality behind this research program is the fact that we regulate exposure to low levels of ionizing radiation using the linear no-threshold model, the premise of which is that there is no safe level of exposure and for which there is also no direct data to prove or disprove the appropriateness of the model. The difﬁculty and challenge, from both a policy and a scientiﬁc perspective, was emphasized at a 1999 meeting of international radiation biologists and policy makers.1 In a consensus statement for the executive summary of the conference report, the conferees agreed that “the lowest dose at which a statistically signiﬁcant radiation risk has been shown is ∼ 100 mSv (∼ 10 rads of X-rays),” yet we regulate radiation exposures to both workers and the public at much lower levels. Several key questions need to be answered. Is human health risk from radiation exposure always proportional to dose? Can any amount of dose increase that risk? Can a single radiation ionization cause cancer? In the past it has been all but impossible to directly measure the biological effects of radiation doses of less than 10 rads to say nothing of determining their 1 Published in the Executive Summary from Bridging Radiation Policy and Science, an international meeting of experts held at Airlie House Conference Center, December 1–5, 1999.

14

D.G. Thomassen, A. Patrinos

potential health effects. Today, we have new techniques and instruments that enable direct measurements of the biological changes induced by low doses of radiation. Advances in instrumentation and in research, especially in genomics, provide new information to help deﬁne radiation’s measurable effects on cells and molecules at environmentally relevant exposures. Such research will, hopefully, underpin the development of future science-based radiation risk regulatory policy. Traditionally, the biological effects of radiation exposures have been viewed as one-hit phenomena, that is, only the hit cell was affected. Now, it is widely accepted that the interaction of ionizing radiation with living systems is far more complex involving higher order biological effects such as communication between irradiated and unirradiated cells and tissue effects. For this reason, a signiﬁcant challenge for scientists today is to conduct research on the biological effects of radiation at all levels of biological organization – from molecules to cells to tissues to organisms. The DOE Low Dose Radiation Research Program focuses its efforts in four general areas: • Low dose radiation versus normal metabolic oxidation – are the damage and biological responses the same or different? • Are there thresholds for the biological effects of low doses of radiation? • Are there genetic factors that affect an individual’s risk from or sensitivity to low doses of ionizing radiation? • How should we communicate the research results? 2. Low dose radiation versus normal metabolic oxidation An interesting quandary is whether low doses of ionizing radiation simply increase, in a strictly quantitative manner, the amount of damage induced in cells as a result of normal oxidative processes and the biochemistry of life. If this is the case, it could be proposed that normal cellular mechanisms that recognize and efﬁciently repair this type of damage might be able to do the same for the additional damage induced by low doses of ionizing radiation thereby making some yet- to-be-deﬁned low levels of ionizing radiation no greater a biological risk than normal cellular processes. Alternatively, low doses of ionizing radiation may induce damage that is qualitatively different from that normally found in cells and some of this damage might not be as efﬁciently recognized or repaired. If this later possibility is shown to be the case, each increment of low dose ionizing radiation might increase the risk of radiation-induced biological effects. Research reported at the Low Dose Radiation Research Program grantee and contractor workshop (Betsy Sutherland, Principal Investigator) suggests that low doses of ionizing radiation may not simply increase the types of damage induced by normal metabolic processes. Cells irradiated with as little as 3 rads of radiation showed clustered DNA lesions not seen in unirradiated controls. This difference may either reﬂect a threshold of normal damage recognition and repair processes or may be evidence for qualitative differences in the types of damage induced by low doses of ionizing radiation compared to the damage induced by normal cellular processes. Either way, these results represent the types of studies that are needed if we are to actually study the biological effects of low doses of ionizing radiation. Further, the relevance of this observation for human health risk also needs to be determined.

Understanding health risks from low doses of ionizing radiation

15

3. Biological responses to low doses of radiation New technologies and instruments are also playing an important role in studies of the biological effects of ionizing radiation. Microbeam irradiators are important tools being developed and used in a number of laboratories around the world. These instruments can be used to expose individual cells, or speciﬁc parts of cells such as the nucleus, the cytoplasm, or even speciﬁc regions of the nucleus or cytoplasm, to a wide range of radiation energies (from heavy ions to electrons) and doses (including the ultimate low dose – a single ion). These instruments are enabling critical questions about the biological effects of very low doses of radiation to be addressed. For example, microbeam irradiators can be used to help identify and deﬁne the target size for radiation – the nucleus? the cytoplasm? They can help scientists determine the role of cell-to-cell communication between irradiated and unirradiated cells, information that is needed help deﬁne the effective biological target and target size for radiation. They can also help deﬁne the effects of normal tissues surrounding an individual or small number of irradiated cells on the ultimate biological response. In recent years, a number of biological responses to radiation have challenged the traditional view of radiation effects. For example, observations such as the by-stander effect, made in a number of cell culture systems, demonstrate that unirradiated neighbors of cells exposed to low doses of radiation can exhibit biological responses as if they had been irradiated. Now, there is evidence that by stander effects can also be observed in intact tissues (Kevin Prise, Principal Investigator). The biological mechanisms responsible for this phenomenon, whether in vitro or in vivo, remains to be determined, but the observations suggest that the effects of low doses of radiation cannot be understood by simply extrapolating from the effects seen at high doses. Interestingly, a number of studies reported in the literature over many years on clastogenic factors in people or animals exposed to radiation may also represent examples of bystander-related phenomena in vivo. A key impact of the bystander phenomenon is that it alters the “traditional” target size for radiation-induced damage, changing our view, not only of how individual cells respond to radiation but, possibly, how tissues respond as well. A fascinating merger of traditional radiation biology with a set of tools emerging from current capabilities in DNA sequencing and genomics is the analysis of the individual genes that are expressed in response to radiation exposures. While it is not surprising that radiation exposures induce the expression of unique sets of genes, what is surprising are the responses induced by high (200 rads) and low (10 rads) doses of radiation. While common sets of genes are induced in different cell and tissue systems in response to high and low doses of radiation, there are an equal, if not greater, number of genes whose expression is uniquely induced by high and low doses of radiation. In other words, low doses of radiation induce the expression of a substantial number of genes not induced by high doses of radiation and vice versa (Andy Wyrobeck, Al Fornace, David Chen, Principal Investigators). Clearly, with respect to the induction of gene expression, high doses of radiation are not simply more of a low dose of radiation – an observation with potentially dramatic implications for the extrapolation of high dose effects to low doses of radiation, though the relationship between this observation and biological risk clearly needs to be determined. New genetic probes have also been used to better characterize the detailed nature of radiation-induced chromosomal aberrations, a traditional indicator of radiation damage. Advanced chromosome “painting” techniques now demonstrate that radiation causes highly com-

16

D.G. Thomassen, A. Patrinos

plex chromosomal aberrations that can not be accounted for by “simple hit theory” (Bill Morgan, Principal Investigator). Other recent results also suggest differences in the biological responses at high and low doses of radiation. Low doses of radiation (< 3 rads) have been shown, under certain conditions, to reduce the frequency cell transformation below the frequency in unirradiated controls (Les Redpath, Principal Investigator). Low doses of radiation (0.5–4 rads) have also been shown to be better at killing cells than high doses (Peter Johnson, Principal Investigator). As with the other phenomena described here, the relationship between this observation and biological risk needs to be determined; however, these observations also demonstrate that there is not a clear linear relationship between the biological effects of radiation at high and low doses. Research is also underway in a number of areas for which it is still too early to tell what the outcomes will be with respect to the effects of radiation in the 1 to 10 rad, or below, region. Radiation has been shown to induce genomic instability in both isolated cells and intact animals, a process in which radiation exposure results in the generation of chromosomal instability (seen as chromosomal aberrations) many generations after the initial exposure. Genomic instability is also a feature of many cancers. If it is possible to establish a clear relationship between exposures to low doses of radiation and the induction of genomic instability, it will be an important link between a traditional high radiation dose endpoint, a cancer-related phenotype, and low doses of radiation.

4. Genetic susceptibility to low doses of radiation Rapid progress in the human genome project has also made it more likely that we will be able to determine if there are speciﬁc genes/phenotypes that make some individuals more susceptible, or even more resistant, to the adverse effects of low doses of radiation. A number of radiation sensitive phenotypes, e.g., Fanconi’s anemia and xeroderma pigmentosum, have previously been described based on the dramatic phenotypes of individuals homozygous for speciﬁc genes. Cohorts of individuals who received high doses of radiation for medical treatment are now being reexamined for variable radiation sensitivity and, eventually, using human DNA sequence data to see if there are additional genes that make individuals more or less sensitive to radiation. While important information scientiﬁcally, the long-term results of these studies also have important and difﬁcult ethical and societal implications. Should individuals who are genetically at greater risk for adverse effects of radiation be prevented from having certain types of jobs? Should employers even have access to this information if it is available or obtainable? Would all individuals chose to know this information even about themselves? These are difﬁcult questions with equally difﬁcult and uncertain answers. Using emerging biological data to help predict health risks from low doses of radiation is itself a difﬁcult but crucial challenge. For this research to have an impact on radiation risk regulatory policy, the new data needs to be incorporated into biologically plausible mathematical models that predict risks from low doses of radiation. Developing such models is a complex task in itself. The path to the generation of such models is far from certain. Current “biologically based” risk models are “informed” by biology but generally only use the most

Understanding health risks from low doses of ionizing radiation

17

rudimentary biological information such as cell proliferation, differentiation, and the general concepts about the role of mutations and variable numbers of biological “steps” that are involved in the development of cancer. The different types of biological information described here will not be equally useful, if useful at all, in the development of an entirely new generation of biologically based risk models. However, it is clear that new models are needed to serve as a key interface or interpreter between the detailed biological effects of radiation and the development of science-based risk policies that adequately and appropriately protect people from the adverse effects of radiation.

5. Communicating risks of low dose radiation As if the science of low dose radiation biology was not a great enough challenge, an even greater challenge is the open and clear communication of this new science and its biological and health risk implications to the public. For many reasons, both good and bad, reasonable and unreasonable, the public generally has a fear and distrust of radiation that far exceeds what current scientiﬁc wisdom tells us is reasonable. Many studies of risk perception have shown that people generally rank familiar risks from things over which they have a choice, such as driving a car or even high dose medical therapy, as less of a threat than things like radiation from medical waste or from living next to a nuclear power plant, over which they have little or no choice and for which they generally have no real-life experiences on which to base their perception. While the public generally has a more positive than negative view of science, the public perceptions or understanding of the risks of low doses of radiation, be they greater or lesser than currently predicted by the linear no-threshold models, will not likely be changed in the near future simply from the results of new scientiﬁc research. Changing the perceptions of the public about radiation from a more emotion-based reaction to a more science-based understanding will take many years, likely decades, of rebuilding trust between the public and radiation scientists, radiation policy makers, and public and private organizations with responsibility for radiation – from medical uses to energy production to waste disposal. The US Department of Energy’s Low Dose Radiation Research Program is speciﬁcally supporting research, such as that described here, to determine health risks from exposures to low levels of radiation. Together with the results from complementary research programs around the world, the new scientiﬁc information generated in this research program is critical to adequately and appropriately protect people from radiation and to make the most effective use of national resources.

18

The signiﬁcance of the natural environment in radiation protection F. Steinhäusler Institute of Physics and Biophysics, University of Salzburg, Hellbrunnerstr. 34, A 5020 Salzburg, Austria

The comparative assessment of individual and collective doses demonstrates clearly the predominant role of the exposure to the natural radiation environment (NRE), such as: NRE contributes almost four-times the combined contributions from nuclear diagnostic medicine and nuclear power production to the collective dose of mankind; the ten-year global use of coal results in approximately the same collective dose as the collective dose due to the Chernobyl accident; the annual collective dose of all persons occupationally exposed at various technical sites (mining, industry, medicine, nuclear fuel cycle) is equivalent to about 10 hours of global exposure to the NRE. A new concept (INRE Scale) is proposed to communicate radiation exposure scenarios to the public, using the bandwidth of normal NRE exposure as a reference value and linking it to the Basic Safety Standards. 1. The ultimate criterion: the dose The ﬁrst report of the then newly founded United Nations Scientiﬁc Committee on the Effects of Atomic Radiation in 1958 stated clearly the four major concerns at the time [1]: (1) the available knowledge is insufﬁcient to evaluate with any degree of precision the possible effects of low-level radiation exposure on man; (2) accumulation of radioactive contamination may increase the probability for somatic damage of an exposed individual; (3) society is required to use uttermost caution in order to avoid any underestimation of the radiation associated risk; (4) it cannot be excluded that the current approach of assessing the risks associated with chronic exposure to low-level radiation may be an overestimation. Forty-four years later the radiation protection community is still faced with a broad spectrum of interpretations of the above caveats, ranging from an almost psychotic fear of anything radioactive by large segments of the public, to the euphoric embracing of the hormesis concept by a select group of health physicists as the apparent absolution of having to deal with RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07003-2

© 2005 Elsevier Ltd. All rights reserved.

The signiﬁcance of the natural environment in radiation protection

19

linear dose–effect relationship and low dose/low dose rate problems. The only common denominator – but frequently missing parameter – in this ongoing discussion is the physical unit of dose. Therefore this study will use the radiation dose as the mutually acceptable basis for the discussion between all parties involved in order to compare radiation exposure scenarios. Depending on the issue, either individual doses or collective doses will be used to characterize different population groups. The global view the twelve UNSCEAR reports published hitherto, representing almost half a century of international data assembled on radioactivity, forms the basis for a comparative assessment of the exposure of man to the natural radiation environment (NRE) versus man-made sources [1–12].

2. Natural radiation exposure: rising? From the moment of fertilization, all through the gestation period of the fetus and throughout life, every man everywhere is exposed to the natural radiation environment. Based on the currently best value for the global average individual dose1 of 2.4 mSv a−1 , the total mean dose received over a lifetime is about 180 mSv. It should be noted though that the corresponding value for some persons, residing in natural high background areas, could be as high as 2 Sv. Under these conditions even the average person is impacted every second by radiation due to the NRE to the extent of > 5500 gamma rays, 110 secondary cosmic rays, 28 neutrons from the cosmic rays, 4200 K-40 atoms, 8 radon/decay product atoms, and 2 uranium atoms. Between the 1960s and the 1980s the knowledge about the importance of the dose contribution by radon (Rn 222) and its decay products (Rn-d) increased signiﬁcantly. The arrival of low-cost integrating radon detectors improved global data on indoor Rn exposure. Together with progress in lung dosimetry, this resulted in an increase of the Rn-d contribution to the total dose. Figure 1 reﬂects this trend: the average dose due to the NRE has increased by

Fig. 1. Average annual radiation dose as assessed by UNSCEAR for adults due to natural radiation during the period 1962–2000. 1 Unless indicated otherwise, dose is used for the term effective dose.

20

F. Steinhäusler

Fig. 2. Annual individual radiation dose due to cosmic radiation for residents and air travellers.

almost 120% since 1972, largely due to the increased Rn-d contributions representing about 54% of the total average NRE dose for an adult. For the past decade this value has remained unchanged, i.e. the value of 2.4 mSv a−1 has been corroborated by all recent studies and can be considered robust. By comparison, the global database on the unattached Rn-d and for thoron (Rn-220) decay products (Tn-d) is comparatively weaker and still in need of further strengthening. Whilst this global average describes the overall order of magnitude of the dose man receives from the NRE, three examples will demonstrate the wide range of NRE-related exposures possible: – individual doses due to cosmic rays can exceed 2 mSv a−1 for residents in high altitudes (e.g. La Paz), as well as for occupationally exposed persons (astronauts, crew members of supersonic aircraft) (Fig. 2); – in some high background areas the gamma dose rate due to terrestrial radiation outdoors can reach up to 30 μGy h−1 (Fig. 3); – dwellings have been found with maximum Rn concentration 85 000 Bq m−3 (Fig. 4). Dose values due to Rn-d can reach tens of mSv a−1 for residents in such radon-prone areas, particularly when the buildings are located in mining areas (specially uranium and phosphate rock). A comparison of the annual collective dose to the global population due to man-made sources and the NRE shows the following (Fig. 5): – The contribution by the politically disputed nuclear fuel cycle equals about 0.1% of the global collective dose contributed by nuclear medicine. The latter, however, is not only the second largest overall contributor (2.5 × 106 man-Sv a−1 ), it is also fully accepted by the

The signiﬁcance of the natural environment in radiation protection

Fig. 3. Terrestrial radiation outdoors.

Fig. 4. Maximum radon (Rn-222) concentration indoors in different countries.

Fig. 5. Global collective effective dose in year 2000 from natural and man-made sources.

21

22

F. Steinhäusler

Fig. 6. Annual global radiation dose due to the fertilizer industry as compared to the impact by the Chernobyl-fallout.

public because of the clear identity of beneﬁciary and exposed individual and frequent lack of knowledge by the patient about the dose received.2 – The NRE contributes almost four-times the combined contributions from nuclear diagnostic medicine and nuclear power production to the collective dose of mankind. Technologically enhanced natural radiation (TENR) plays an increasingly important role in radiation protection. Certain industrial production methods, occupational activities, or changes in life-style – frequently without any primary intention to change the natural conditions with regard to the NRE – can lead to an increased NRE exposure. Examples are: phosphate fertilizer industry and its wastes; rare earth industry and mineral processing; reduction of the room air ventilation rate in order to conserve energy; the use of peat, coal and thermal energy as alternative sources of energy. Figure 6 shows for the fertilizer industry that – even under the conservative assumption of only 5% recycling of phosphor gypsum as construction material – the residents of such buildings will receive a collective dose, which will be equivalent to the collective global dose due to Chernobyl fallout within two years.

3. Energy production and occupational radiation exposure In Fig. 7 the annual collective doses resulting from the routine operation of the different methods of producing energy are compared. Also shown is the collective dose due to the complete nuclear fuel cycle for 40 years: – the annual global dose due to the use of coal as energy source is more than 100 times higher than that from oil, gas and geothermal combined; 2 It is questionable whether large numbers of the public even associate “radiation exposure” with some of the modern medical exposure scenarios, e.g. computer-tomography is generally not associated with radiation exposure.

The signiﬁcance of the natural environment in radiation protection

23

Fig. 7. Global annual dose due to non-nuclear energy production as compared to the total (1950–1989) dose due to nuclear power production (excluding accidental releases).

Fig. 8. Occupational annual collective doses for the nuclear fuel cycle and other workers.

– every year the global collective dose due to the use of coal is on average about ten-times higher than the corresponding value resulting from 40 years of industrial application of nuclear energy; – the ten-year global use of coal as energy source results in approximately the same collective dose as the global collective dose due to the Chernobyl accident; – the contribution of the non-nuclear energy production to the global collective dose represents < 2% of the corresponding NRE value. Globally the annual collective dose of all persons occupationally exposed at various technical sites (mining, industry, medicine, nuclear fuel cycle) amounts to about 14 500 man-Sv (Fig. 8). This value is equivalent to about 10 hours of global exposure to the NRE. The largest contribution is provided by the TENR radiation exposure of workers in above-ground work-

24

F. Steinhäusler

places due to Rn-d inhalation (6000 man-Sv), followed by the mining and mineral processing industries coal miners (about 5100 man-Sv). In view of the comparatively small annual collective dose by the workers in the nuclear fuel cycle (including uranium mining: 1440 man-Sv), the disproportionately high amount of ﬁnancial resources used for their protection against radiation is noticeable. This discrepancy is also noticeable with regard to individual doses: some miners in Indonesia, Malaysia and Thailand can receive up to 180 mSv a−1 , which would be considered unacceptable for workers in the nuclear industry [13].

4. INRE: a new approach to communicating radiation exposure based on the NRE The scientiﬁc community has been largely successful in communicating risks due to natural phenomena, such as wind speeds (Beaufort Scale 1–17) or earthquakes (Mercalli Scale 1–12, Richter Scale 1–no upper value). Similarly, the severity of accidental incidents at nuclear installations is described in a qualitative manner using the International Nuclear Event Scale (INES Scale 1–7). It is proposed to develop an additional scale for communicating radiation exposure qualitatively, based on the exposure to the NRE, the so-called International Natural Radiation Exposure (INRE) Scale. The INRE Scale could be used for illustrative purposes of normal operational procedures as well as accidental situations. For example, for communicating to the public the magnitude of an occupational exposure due to a novel industrial process resulting in a radiation dose to the workers; or: the dose to members of the public due to depleted uranium ammunition during a post-conﬂict situation. By combining the INRE information with the concept underlying the Basic Safety Standards [14], this could also be applied to illustrate whether intervention by regulatory authorities is deemed necessary or not. The reference base line of the INRE Scale is the global average dose due to NRE exposure, i.e. INRE Scale 1 refers to 2.4 mSv a−1 and the normal bandwidth of global NRE dose values. In addition, a reference bandwidth is also deﬁned for NRE-related gamma dose rate, reﬂecting upper values of typical circumstances. Such scenarios can be found, e.g. in high background areas, or in dwellings with elevated Rn levels indoors, but still below any Action Level as deﬁned by the authorities. In circumstances such as these no intervention is thought to be needed at INRE Scale 1. INRE Scale 2 encompasses exposure scenarios, which are borderline cases between the “normal” bandwidth of the NRE and signiﬁcantly elevated levels, typically an order of magnitude above the upper values of “normal” NRE exposure. This includes exposure scenarios, which can be found in mining industries or other TENR-related activities. It will have to be decided on a case-by-case whether it is justiﬁed to intervene. INRE Scale 3 means that the exposure is unacceptably high in comparison to the “normal” NRE bandwidth, e.g. the Rn concentration indoors exceeds signiﬁcantly Action Levels. Intervention is almost always required. This approach may assist to communicate exposure scenarios, which might be difﬁcult to explain to the public otherwise by categorizing them in accordance with the proposed deﬁnition of the INRE values 1–3 as shown in Table 1.

The signiﬁcance of the natural environment in radiation protection

25

Table 1 International Natural Radiation Exposure (INRE) Scale for communicating radiation exposure scenarios to members of the public INRE Scale 1

INRE Scale 2

INRE Scale 3

Normal exposure to the natural radiation environment: • reference range of the gamma dose rate: < 2 μGy h−1 ; or: • reference range of the annual dose: < 10 mSv • intervention unlikely to be justiﬁable Elevated exposure as compared to the typical natural radiation environment: • reference range of the gamma dose rate: 2–20 μGy h−1 ; or: • reference range of annual doses: 10–100 mSv • intervention may be necessary Unacceptably high exposure as compared to the typical natural radiation environment: • reference range of the gamma dose rate: > 20 μGy h−1 • reference annual doses: > 100 mSv • intervention should be almost always justiﬁed

5. Outlook Worldwide the awareness increases about the negative environmental impact of using coal as the main source of energy for electricity production. Together with the fast growing populations and economies in Asia and the political instability in the oil producing countries in the Middle East, this will put renewed pressure on political decision makers to reconsider the use of nuclear power. At the same time, risk aversion is becoming a dominant feature of industrialized societies, aiming for zero risk as the ultimate goal. However, even in the richest societies there are only limited means available to control and minimize risks, i.e. it will be necessary to optimize the use of the available ﬁnancial resources to practice an integrated risk management. Such an undertaking will not only have to address the media-favorite risks (e.g. any radiation-related risk, irrespective of its magnitude), but apply a weighted approach to set priorities for all radiological and non-radiological risks man is exposed to. With regard to radiation protection, this revision of the current strategy will necessitate that regulatory priorities are set in accordance with the individual and the collective risk. The ultimate goal should be the optimization of the means available for managing the total radiation-induced risk, placing less weight on the source of the radiation exposure but rather on its magnitude and focusing on its cost-effective controllability. The vast database assembled worldwide on the NRE and TENR could provide a useful tool in putting exposures resulting from man-made radiation sources into perspective.

References [1] UNSCEAR (United Nations Scientiﬁc Committee on the Effects of Atomic Radiation), Report to the General Assembly, Thirteenth Session, United Nations, New York, 1958, Suppl. No. 17 (A/3838). [2] UNSCEAR (United Nations Scientiﬁc Committee on the Effects of Atomic Radiation), Report to the General Assembly, Seventeenth Session, United Nations, New York, 1962, Suppl. No. 16 (A/5216). [3] UNSCEAR (United Nations Scientiﬁc Committee on the Effects of Atomic Radiation), Report to the General Assembly, Nineteenth Session, United Nations, New York, 1964, Suppl. No. 14 (A/5814).

26

F. Steinhäusler

[4] UNSCEAR (United Nations Scientiﬁc Committee on the Effects of Atomic Radiation), Report to the General Assembly, Twenty-ﬁrst Session, United Nations, New York, 1966, Suppl. No. 14 (A/6314 and Corr. 1). [5] UNSCEAR (United Nations Scientiﬁc Committee on the Effects of Atomic Radiation), Report to the General Assembly, Twenty-fourth Session, United Nations, New York, 1969, Suppl. No. 13 (A/7613 and Corr. 1). [6] UNSCEAR (United Nations Scientiﬁc Committee on the Effects of Atomic Radiation), Report to the General Assembly, Twenty-seventh Session, United Nations, New York, 1972, Suppl. No. 25 (A/8725 and Corr. 1). [7] UNSCEAR (United Nations Scientiﬁc Committee on the Effects of Atomic Radiation), Report to the General Assembly, Thirty-second Session, United Nations, New York, 1974, Suppl. No. 40 (A/32/40). [8] UNSCEAR (United Nations Scientiﬁc Committee on the Effects of Atomic Radiation), Report to the General Assembly, Thirty-seventh Session, United Nations, New York, 1982, Suppl. No. 45 (A/37/45). [9] UNSCEAR (United Nations Scientiﬁc Committee on the Effects of Atomic Radiation), Report to the General Assembly, Forty-ﬁrst Session, United Nations, New York, 1986, Suppl. No. 16 (A/41/46). [10] UNSCEAR (United Nations Scientiﬁc Committee on the Effects of Atomic Radiation), Report to the General Assembly, Forty-third Session, United Nations, New York, 1988, Suppl. No. 45 (A/43/45). [11] UNSCEAR (United Nations Scientiﬁc Committee on the Effects of Atomic Radiation), Report to the General Assembly, Forty-eight Session, United Nations, New York, 1993, Suppl. No. 46 (A/48/46). [12] UNSCEAR (United Nations Scientiﬁc Committee on the Effects of Atomic Radiation), Report to the General Assembly, Fifty-ﬁfth Session, United Nations, New York, 2000, Suppl. No. 46 (A/55/46). [13] F. Steinhäusler, Management of radiation protection in the mining, milling and downstream processing of mineral sands, in: Minesafe International from Principles to Practice, Proc. Int. Conf., Session 2, Perth, Western Australia, 22–26 March 1993. [14] Council Directive 96/29/Euratom of May 13, 1996 laying down basic safety standards for the protection of health of workers and the general public against the dangers arising from ionising radiation, Ofﬁcial J. Eur. Commun. Ser. L 159 (29.6.1996).

2. Keynote

T . his Page Intentionally Left Blank

29

The theory of cosmic-ray and high-energy solar-particle transport in the atmosphere K. O’Brien Northern Arizona University, Flagstaff, AR 86011-6010, USA

An essentially analytical theory of the transport of high-energy radiation through the Earth’s atmosphere is presented here. The transport of primary and secondary particles is described by a solution of the Boltzmann transport equation. The transport of secondary particles is based on a solution of the Boltzmann equation separable into longitudinal and transverse components, applicable to high-energy hadron–nucleus collisions, and based on work by Passow [1] and reported by Alsmiller [2], and on Elliott [3] and Williams [4]. All secondary particles other than hadrons are mediated by meson production and decay. The breakup of primary nuclei as a result of collisions with atoms of air is treated by means of a generalized Rudstam [5] formula. The theory described here has also been applied to the atmospheres of other solar system bodies. 1. Introduction This is a necessarily abbreviated account (due to the limitations of space) of theoretical developments in atmospheric cosmic-ray transport that have spanned some decades since the ﬁrst publication in 1970 [6]. These developments have allowed the calculation of the propagation and attenuation of primary cosmic rays and the generation and propagation of secondary cosmic rays in the Earth’s atmosphere and the atmospheres of other solar system bodies [7]. These results have been also been generalized to treat the propagation of energetic solar particles in the Earth’s atmosphere. 2. Cosmic-ray and energetic solar-particle radiation propagation through the atmosphere 2.1. Atmospheric structure The atmospheric structure, its pressure, temperature and density as a function of elevation was taken from Beranek [8]. Its composition was taken from CIRA [9], yielding an average RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07004-4

© 2005 Elsevier Ltd. All rights reserved.

30

K. O’Brien

atomic number of 7.22, an average atomic weight of 14.485 and an ionization potential of 93 eV. 2.2. The transport equation The Boltzmann equation is an integro-differential equation describing the behavior of a dilute assemblage of corpuscles. It was derived by Ludwig Boltzmann in 1872 to study the properties of gases. It applies equally to the behavior of radiation. Boltzmann’s equation is a continuity equation in phase space in terms of the angular ﬂux t), the number of particles of a given type (nucleons, leptons, mesons, etc.) at a x , E, Ω, φi ( at a time t. It is given by location x , with an energy E, a direction Ω, t = Qij , Bˆ i φi r, E, Ω, j

• ∇ + σi + di − (∂/∂Ei )Si , Bˆ i = Ω ∞ dΩ dEB σij EB → E, Ω → Ω φj r, EB , Ω, t , Qij = j

di =

4π

E

1 − β 2 (τi cβ),

(1)

where Bˆ i is the Boltzmann operator; σi is the absorption cross section for particles of type i; di is the decay probability per unit ﬂight path of radioactive particles (such as muons or mesons) of type i; Si is the stopping power for charged particles of type i (assumed to be zero t) is the particle ﬂux of type-j particles at location x , x , EB , Ω, for uncharged particles); φi ( energy E, direction Ω and time t; Qij is the “scattering-down” integral, the production rate an energy E at a location x , by collisions with nuclei of particles of type-i with a direction Ω, at a higher energy EB ; σij is the doublyor decay of type-j particles having a direction Ω differential inclusive cross section for the production of type-i particles with energy E and from nuclear collisions or decay of type-j particles with an energy EB and a a direction Ω ; βi is the speed of a particle of type i with respect to the speed of light (= ν/c); direction Ω τi is the mean life of a radioactive particle of type i in the rest frame; and c is the speed of light in vacuo. The most signiﬁcant of the reactions that result from cosmic-ray and solar-particle collisions with the nuclei that comprise the Earth’s atmosphere are: Ai + Air → N Z n n + N Z p p + N Z π π + N Z K K + N Z A Aj , p + Air → N p n n + N p p p + N p π π + N p K K, n + Air → N n n n + N n p p + N n π π + N n K K, π± → μ± + ν, π0 → 2γ → electromagnetic showers, μ± → e± + 2ν → electromagnetic showers.

(2)

The theory of cosmic-ray and high-energy solar-particle transport in the atmosphere

31

Equation (2) implies that atmospheric cosmic rays propagate by means of the nucleonic cascade (the ﬁrst three lines are recursive), and that all other secondaries ultimately result from nucleon-nucleus collisions. High-energy nuclei and nucleons collide with nuclei, producing other high-energy nucleons, which in turn collide with other nuclei, losing energy to secondary mesons and eventually to leptons and photons. Ai and Aj are primary and secondary cosmic-ray nuclei, and the N i j are the multiplicities of j -type particles resulting from the collisions of nuclei, protons and neutrons with nuclei of air.

3. The hadronic component The solution to equation (1) is assumed to be separable into a longitudinal component and a transverse component: = Li (x, E)Ti (y, z, E). φi x, E, Ω (3) 3.1. The longitudinal component of the hadronic transport equation The theory of the longitudinal component depends most critically on 3 terms in equation (1): (1) the total collision cross section; (2) the partial inelasticities of each of the emitted hadrons; (3) the multiplicity of secondary hadrons and nuclei emitted from a collision. 3.1.1. The total collision cross section The total collision cross section, σ , is constant and geometric. As, at high energies, elastic scattering is primarily forward, it is set to zero. An elastic scattering that results in the colliding particle traveling undeviated from its original direction involves no exchange of energy and hence does not affect its transport through the medium. The total collision cross section is given by σ = ξ πri2 + πrt2 L/At cm2 g−1 (4) where 1/3 ri = 1.28 Ai − 1 fm,

1/3 rt = 1.28At

fm

and ξ =

1, At 4, 2, At > 4,

(5)

L is Avogadro’s number, A is the atomic weight, and the subscripts i and t correspond to the incident and target nucleus, respectively. 3.1.2. The partial inelasticities The partial inelasticities, Kj , are deﬁned as EB → Ω dE (E + mj )Fij EB → E, Ω Ki =

(6)

η

where mj is the rest mass of the j -type particle and η is the lowest energy an i-type particle can be emitted from a nucleus struck by a j -type particle.

32

K. O’Brien Table 1 Partial inelasticities for nucleon–air collisions Hadron

p-nucleus

n-nucleus

P N π+ π0 π− K+ K0 K−

Ki 0.211 0.211 0.180 0.180 0.112 0.034 0.034 0.022

Ki 0.211 0.211 0.112 0.180 0.180 0.022 0.034 0.034

The analytical form for the production spectrum required by the theory is −Ω , Fij = αi EBl /E (l+1) δ Ω

(7)

where δ is Dirac’s improper function, implying that the secondary particles emitted from a collision proceed in the same direction as the incident primary. Substituting equation (7) in equation (6), integrating, and taking the limit as EB goes to inﬁnity yields Ki = αi /(1 − l)

(8)

and therefore

−Ω . Fij = (1 − l)Ki EBl /E (l+1) δ Ω

(9)

The values of the partial inelasticities to be used are taken from O’Brien [10]. These values are based on both accelerator and cosmic-ray data and are exhibited in Table 1. 3.1.3. The secondary particle multiplicity The secondary particle multiplicity is deﬁned thus: EB → Ω dE. Fij EB → E, Ω nj (EB ) =

(10)

η

Representing Fij by equation (10), the index l can be determined by ﬁtting equation (10) to the shower-particle data of Meyer, Teucher and Lohrmann [11]. The stationary solution φ for secondary hadrons for a unit incident ﬂux along the x-coordinate with an energy EB [1,2]: Li (x, EB → E) = Ai σ (1 − l)Kj EBl /E l+1 U (EB − ηi ) x/B(EB , E)

(11) × I1 2 rB(EB , E) ,

(1 − l)Ki ln EB − ln EU (E − ηj ) + ηi U (ηj − E) , B(EB , E) = σ (12)

U (x) =

j =n,p

1, 0,

x > 0, x < 0.

The theory of cosmic-ray and high-energy solar-particle transport in the atmosphere

33

The Heaviside function, U , shuts off secondary particle production below a bombarding energy ηj to take into account the effect of charged-particle slowing down on secondary particle production by a charged incident particle, such as a proton. I1 is the hyperbolic Bessel function of the ﬁrst kind. Ai is a correction factor that accounts for secondary meson decay and charged particle slowing down. The solution of the Boltzmann equation for a unit incident ﬂux of energy EB and atomic number Z at a depth x1 > x0 in a direction given by some zenith angle Ω is 0 EZ 1 0 1 σZ /SZ (E ) dE φZ EZ , x1 = SZ EZ , x0 /SZ EZ , x1 exp − (13) EZ1

where x=

E0 E1

dE /SZ (E )

(14)

is the distance in the medium where the incident nucleus at an energy E 0 has been reduced to → Ω) in equation (1) is given an energy E 1 . For secondary nuclei, the term σij (EB → E, Ω by a modiﬁed Rudstam CDMD equation. 3.2. The transverse component of the hadronic transport equation The cosmic-ray ﬂux is incident on the upper boundary of the atmosphere, which is treated as a free surface with a radius of curvature of 6.378 × 108 cm. The curvature is introduced by means of the Chapman Function. The unit of r is g cm−2 , which simpliﬁes the format of the Boltzmann equation, though, of course, the detailed density structure of the atmosphere must be used to properly account for pion and muon decay and allows the speciﬁcation of a free surface, which would otherwise be impossible, as the atmosphere has an exponential density dependence. Elliott [3] and Williams [4] have, by applying a Fourier transform to the transport equation, shown how the linear, one-dimensional form might be applied to two- or three-dimensional problems. It has been applied to account for the spreading of the cosmic-ray beam and thus to provide an improved boundary condition for the cosmic-ray problem. Equation (1) can be rewritten in Cartesian coordinates, where x has the same meaning as above:

Bˆ i = ∂/∂x + (1 − μ2 ) cos ϕ(∂/∂y) + sin ϕ(∂/∂z) + σ + di − ∂Si (E)/∂E, (15) t = Qij , Bˆ i φi r, E, Ω, j

with the ﬂux entering at some point x1 and traveling along x. The upper free surface is located along z at some point z0 . Taking the Fourier transform of equation (15) gives ∞ ∞ exp(iBy y + Bz z), dy dz φi r, E, Ω φi (B) = (16) −∞

−∞

34

K. O’Brien

which yields φi B = ξ φi Li (r , E) exp(iBy + iBz ),

(17)

where ξ is chosen so that ∞ ∞ dy dz φi B = Li

(18)

−∞

−∞

and Li is the longitudinal component of the solution as given by equation (7). The transverse momentum distribution from a high-energy hadron–nucleus collision is given by [12]: √ Ni (θ ) = Cpi /π exp −Cpi2 θ 2 (19) where pi = E 2 + 2mi E is the momentum of the secondary, mi is its mass and C = 3.11 × 10−6 . The mean angle of emission from the cascade as a function of angle can be gotten from the transverse momentum distribution by multiplying Ni by the Jacobian that converts ﬂuxes per unit momentum to ﬂuxes per unit energy (J = p/W ) where W is the total energy. Thus the mean angle of emission is ∞ θ¯i = 2π (20) θ J Ni dθ = π/C (Ei + m), 0

and since this is a nucleonic cascade, m is the nucleonic mass. Instead of inverting the Fourier transform, −σtr,i is substituted for iB yielding φi (E, r) = ξ Li exp −σtr,i y 2 + z2 = Li (E, x)Ti (E, y, z) + φp (x, EB )

(21)

where

σtr,i = σ 1 − cos(π/2 − θ¯i )

(22)

is the transport cross section. For a component of the cosmic-ray ﬂux entering the atmosphere at some zenith angle greater than zero, the contribution to the ﬂux at the origin, at a depth z0 in the atmosphere is given by the integral ∞ z0 dy dz φi (E, r) φi (E, x) = (23) −∞

−∞

or

φi (E, x) = Li 2 − (1 + σtr,i z0 ) exp(−σtr,i z0 ) /2 + φp (EB , x).

(24)

Equation (19) takes into account the existence of the upper free surface on the radiation intensity in the atmosphere, a boundary above which there is no contribution to the atmospheric cosmic-ray ﬂuxes except from the primary ﬂuxes; that is, no secondary particles above the free surface contribute to the ﬂux at x.

The theory of cosmic-ray and high-energy solar-particle transport in the atmosphere

35

3.3. The low-energy neutron component Equation (24) is valid for high-energy hadrons. Charged and radioactive hadrons are largely restricted to high energies because of the effects of charged-particle slowing down and radioactive decay at low energies. Low-energy hadron transport is therefore low-energy neutron transport. Observing from the work of Armstrong [13] and Hess et al. [14] that the shape of the low-energy cosmic-ray neutron spectrum changes but slowly with depth, the neutron spectrum of Rösler [15] was applied to the neutron spectrum calculated using equation (19) between 500 MeV and 0.5 eV. Below 0.5 eV, the Hess et al. [14] spectrum was made piecewise continuous with that neutron spectrum.

4. The lepton and photon component 4.1. The muon component The source for the muon ﬂux, Qπμ is

Qπμ = (mπ± /2πmμ )φπ (mπ /mμ )Ej dπ .

(25)

The ﬂux at a depth x1 > x0 in a direction given by some zenith angle Ω is, in the continuous-slowing-down approximation, 1 0 1 φμ Eμ , x1 = Qπμ Sμ Eμ , x0 /Sμ Eμ , x1 exp −

0 Eμ

1 Eμ

1 dπ /Sμ E dE .

(26)

The relationship between E 0 , E 1 and x = x0 − x1 is given by x =

E0 E1

dE /Sμ (E )

(27)

4.2. The electromagnetic component Muons and pions are radioactive. Muons have a mean life of 2.197 μs, charged pions, a mean life of 26 ns and neutral pions a mean life of 84 as. Muons decay into two neutrinos and an electron, and neutral pions into two photons (cf. equation (2)). Charged pions give rise to the muon component of cosmic rays. In this approach, the energy of the two photons from neutral pion decay is deposited where the pion is produced and decays. The energy from a decaying muon is equally shared among the three light particles, so that the electron gets one third, and this energy is also deposited at the point where the electron is produced. Secondary negaton, positon and photon scalar spectra were calculated using CASCADE [16] and coupled to the locally deposited energy.

36

K. O’Brien

5. Cosmic rays 5.1. Composition Peters [17] represents the integral cosmic-ray spectrum by log Φ = a − 0.0495[11.9 + log(1.7 + E)]2

(28)

where Φ is the number of particles with energies greater than E GeV, per (m2 s sr). The differential spectrum is therefore log ϕ = a − 0.0495 11.9 + log 1.7 + E 2

+ log 0.0990 11.9 + log(1.7 + E) /(1.7 + E) (29) where ϕ is now the ﬂux per (GeV m2 s sr) per nucleon. The constant a governs the magnitude and intensity of ϕ. Gaisser and Stanev [18] (1998) gave a table of relative particle intensities at 10.6 GeV normalized to the oxygen ﬂux (≡ 1). The oxygen ﬂux per nucleon at that energy is 3.26 × 10−6 per (cm2 s sr GeV). Their data appear in Table 2, along with absolute intensities obtained by multiplying their data in column 3 by the oxygen ﬂux. Equation (29) is solved for each of these components. 5.2. The primary spectrum The primary cosmic-ray spectrum used in the calculations described below is divided into twelve groups: • the protons in the hydrogen ﬂux, the unbound or free protons, and • eleven groups of primary nuclei. Table 2 Composition of cosmic rays at 10.6 GeV per nucleon Z

Element

Relative abundances

1 2 3–5 6–8 9–10 11–12 13–14 15–16 17–18 19–20 21–25 26–28

H He Li–Be C–O F–Ne Na–Mg Al–Si P–S Cl–Ar K–Ca Sc–Mn Fe–Ni

730 34 0.4 2.20 0.3 0.22 0.19 0.03 0.01 0.02 0.05 0.12

The theory of cosmic-ray and high-energy solar-particle transport in the atmosphere

37

The elements above helium in the periodic table are represented here by the astronomer’s term, “metals”, rather than the traditional cosmic-ray physicist’s term, HZEs. The metals are, after all, generated from the circumstellar material in the vicinity of a Type II supernova and are proper material for study by astronomers, and HZE is so clumsy, not to say ugly, a term. Cosmic-ray proton spectra below 10 GeV are represented by the following equation [19]: −2.65 , ϕ = 9.9 × 104 E + 780 exp −2.5 × 10−4 (30) where E is in MeV/nucleon. The eleven groups of nuclei are obtained by multiplying equation (30) by the values in column three of Table 2. The cosmic-ray spectra above 10 GeV are represented by Peters’ model, equation (29). 5.3. The geomagnetic ﬁeld A particle entering the Earth’s magnetic ﬁeld must have sufﬁcient momentum per unit charge, or rigidity to penetrate the Earth’s magnetic ﬁeld and collide with air nuclei and produce atmospheric cosmic rays. That rigidity, a function of the particle’s charge, and the zenith and azimuthal angles it makes with the Earth’s surface is called the “cutoff” rigidity and is expressed in GV (GeV c−1 per unit charge). Shea and Smart [20–23] have calculated vertical cutoff rigidities for a number of magnetic epochs using numerical methods. These are effective cutoffs, taking into account the shielding of the solid Earth and the penumbra. Rösler [15] has normalized Störmer’s equation to the vertical cutoff given by the Shea– Smart calculations. Thus the details of the moments of the geomagnetic ﬁeld are carried by the vertical cutoff distribution, but otherwise the ﬁeld is locally dipole. In this theory, the cutoff rigidity is treated like a high-pass ﬁlter, though for certain angles and locations this may not be absolutely correct. 5.4. Solar modulation Gleeson and Axford [24] have shown theoretically that the effect on the galactic cosmic-ray spectrum of passage through the interplanetary medium is approximately the same as would be produced by a heliocentric potential with a magnitude at the Earth’s orbit equal to the energy lost per unit charge to that point by interacting with the solar wind. The energy spectrum at the Earth’s orbit is then obtained from the unmodulated energy spectrum outside the heliopause:

2 ϕi (E) dE = ϕi (T ) dE r(E)/p(T ) , T = E + ZV (31) where r is the rigidity, and V is the voltage. As ϕi is isotropic outside the heliosphere, it is, by Liouville’s theorem, isotropic at the Earth’s orbit. 5.5. Determining the heliocentric potential Neher [25] measured the cosmic-ray proton spectrum during a series of 31 balloon ﬂights between July 31 and August 4, 1965. Half were made from shipboard going north from Peru

38

K. O’Brien

to Greenland. The other half were simultaneous ﬂights made from Bismarck, North Dakota. The former were used to analyze the cosmic-ray spectrum by virtue of its interaction with the Earth’s magnetic ﬁeld and the latter to correct for changes in its intensity. The importance of this experiment to the determination of the heliocentric potential is that the timing of each of these balloon ﬂights is precisely known and could be related to the Deep River neutron monitor counting rate at the time of the ﬂight. Equation (31) ﬁts the measured spectrum with V = 500 MV. The further assumption was made that the NM-64 monitors respond in the same way to cosmic-ray ﬂuxes as the Deep River and Goose Bay neutron monitors, that purely local effects are negligible, and that small changes in the cosmic-ray spectra with modulation and with changes in ground elevation are also negligible, enabling the use of other high-latitude monitors. This has become important, as the Deep River monitor has been taken down.

6. Comparison with experiment 6.1. The cosmic-ray ionization proﬁle A comparison with ionization measurements over Durham, New Hampshire in 1969(26) is shown in Fig. 1. The ionization, or ion-pair production rate, is given as I , ion-pairs per (cm3 s) of air at NTP. Agreement is seen to be quite good. 6.2. The cosmic-ray neutron proﬁle Atmospheric neutron ﬂuxes were calculated for Aire sur l’Adour, France, for the geomagnetic and solar conditions of 1966 and compared with a variety of measurements [15,27,33]. For convenience, the experimental results were characterized in terms of the geomagnetic latitude.

Fig. 1. Calculated and measured [26] cosmic-ray ionization over Durham, New Hampshire in 1969.

The theory of cosmic-ray and high-energy solar-particle transport in the atmosphere

39

Fig. 2. Calculated and measured [27–33] cosmic-ray neutron proﬁle at 42◦ geomagnetic latitude.

Despite the scatter in the measurement results, agreement is good (Fig. 2). The conditions of the calculation were chosen to correspond to the measurements of Boella et al. [27], which were made at the highest altitude and, therefore, most sensitive to geomagnetic and solar effects.

7. Energetic solar particles High-energy solar particles, produced in association with solar ﬂares and coronal mass ejections, occasionally bombard the Earth’s atmosphere, resulting in radiation intensities additional to the already-present cosmic radiation. Access of these particles to the Earth’s vicinity during times of geomagnetic disturbance are not adequately described by using static geomagnetic ﬁeld models. These solar ﬂuxes are also often distributed non-uniformly in space, so that ﬂuxes measured by satellites at great distances from the Earth and which sample large volumes of space around the Earth do not accurately predict ﬂuxes locally at the Earth’s surface. A method is described here, which uses the ground-level neutron monitor counting rates as adjoint sources of the ﬂux in the atmosphere immediately above them to obtain solar-particle ionization rates as a function of position over the Earth’s surface. This approach has been applied to the large September 29–30, 1989 event (GLE 42). It was applied to determine the magnitude and distribution of the solar-particle ionization from this atypically large event. 7.1. The acceleration mechanism Solar ﬂares and coronal mass ejections (CMEs) can accelerate hydrogen (and some helium and heavier nuclei) ions to high energies by means of the shock-acceleration mechanism. The shock-acceleration mechanism yields a spectrum, which is a power-law per unit of rigidity or (since protons alone were measured and therefore alone will be considered) momentum, and can be converted to the ﬂux per unit energy by multiplying, again, by the appro-

40

K. O’Brien

priate Jacobian: ϕ(r) = a(t)r −γ (t) ,

φ(E) = ϕ(r)r/(E + m).

(32)

When the Earth is intercepted by a shock or other interplanetary disturbance, the induced current in the magnetosphere affects the geomagnetic ﬁeld in a complex manner, changing the distribution of cutoff rigidities and usually reducing them. If the Earth is in the right position, it may be intercepted by the plasma accelerated by a prior shock. This plasma will affect the geomagnetic ﬁeld in a complex manner, changing the distribution of cutoff rigidities and usually reducing them. Further, the ﬂux may be distributed non-uniformly over the Earth’s surface. These factors make a straightforward calculation of the resulting radiation distributions from satellite data impossible. However, since there are a number of cosmic-ray neutron monitors distributed over the land surface of the Earth, they may be used to obtain a(t) whereas γ (t) can be obtained from satellite data. 7.2. The computational method In principle, one could use the ground-level neutron monitor data as adjoint sources and solve the adjoint form of the transport equation, applying the satellite spectra as boundary conditions. However, since many radiation components contribute to the response of a neutron monitor, an equal number of adjoint calculations would have to be calculated for each value of γ . A simpler and more straightforward approach is to execute a forward calculation of the neutron monitor response for a range of values of γ and interpolate among them. The value of a(t) in the equation for the ﬂux is determined by setting the integral ﬂux above 100 MeV to unity for each value of γ (t). 7.3. Neutron monitor data The neutron monitor stations that were in existence and had useful data during ground-level event 42 (GLE 42, September 29–30, 1989) and which were used to obtain the necessary adjoint source data are listed in Table 3. These data and the associated counting rates were obtained from WDC-C2 for Cosmic Rays at Ibaraki University by ftp [34]. 7.4. Satellite particle spectra The satellite particle energy spectra used here were derived from data obtained by particle detectors aboard the GOES-7 satellite maintained by the NOAA Space Environment Center. These detectors measure the ﬂux of energetic protons at geostationary orbit from energies of 600 keV to greater than 700 MeV (or momenta of 330 MeV c−1 to greater than 1300 MeV c−1 ) in 11 discrete channels. The observations of protons of greater than about 9 MeV (130 MeV c−1 ) are representative of those that would be obtained outside the magnetosphere in that protons of higher energy have full access to geostationary orbit (or alternatively, in that the geomagnetic cutoff at geostationary orbit has never been observed to exceed 9 MeV (130 MeV c−1 )) [35].

The theory of cosmic-ray and high-energy solar-particle transport in the atmosphere

41

Table 3 Neutron monitor stations used in the analysis of GLE 42 Latitude −78.3 −75.5 −63.5 −57.4 −33.2 −27.3 −18.0 −0.4 25.3 28.4 33.3 36.2 40.9 42.6 44.2 47.9 48.0 48.2 50.5 50.8 51.3 52.0 54.8 54.9 57.6 58.3 60.3 61.8 62.4 62.7 64.5 70.4 88.2

Longitude (degrees geomagnetic) 0.0 230.9 43.8 127.7 80.3 90.1 82.6 354.0 205.5 184.5 150.7 122.1 174.4 92.1 157.8 89.5 102.4 315.4 210.2 193.7 351.7 87.9 357.7 95.6 348.9 301.8 191.3 117.1 225.7 125.6 12.6 264.1 2.0

Station South Pole Terre Sanae8 Kerguelen Hermanus Potchefstroom Tsumeb Huancayo Tokyo Beijing Alma Ata Tbilisi Irkutsk Rome Novosibirsk Jungfraujoch Lomnicky Stit Climax Magadan Yakutsk Newark Dourben Durham Kiel Deep River Calgary Tixie Oulu Cape Schmidt Apatity Goose Bay Inuvik Thule

The principal correction that had been applied to those data were a correction for the HEPAD response to backward ﬂuxes through it, and subtraction of the background counting rates in each channel due primarily to galactic cosmic rays, their progeny, and to a lesser extent, instrument noise. 7.5. Results Ionization calculations were carried out for 35 hours following the event, for elevations of sea level: 30 000; 40 000; 50 000; 60 000; 70 000 and 80 000 feet for each of the sites using the theory described above but designed for solar-particle events. Figure 3 exhibits the hourly-

42

K. O’Brien

Fig. 3. Cosmic-ray ionization at a depth of 100 g cm−2 in the atmosphere at the peak of GLE 42, 0130 hrs, September 30, 1989.

averaged distribution of ionization due to the solar-particle event at the period of maximum intensity of the event at an atmospheric depth of 100 g cm−2 [35]. Any irregularities in the apparent solar particle ﬂux distributions are presumably due to the approximations described above and the existence of irregularities and/or gradients in the interplanetary medium as well as to variations in the access of these particles to the magnetosphere. It is demonstrable that signiﬁcant perturbations of magnetospheric access can occur (at least at energies of the order of 500 MeV or less) as a result of the interference of interplanetary shocks and/or magnetic structures with the uniform propagation of solar energetic particles into the magnetosphere.The ionization distribution from this large event has a maximum at the highest latitudes at an atmospheric depth of 100 g cm−2 of almost 2000 I .

8. Conclusions A two-component analytical solution to the Boltzmann equation has been successfully applied to the propagation of cosmic rays in the Earth’s atmosphere. Its adequacy has been demonstrated by comparison with experimental data. This equation has also been applied to the propagation of high-energy solar particles. Unfortunately there are no data, to the best of the author’s knowledge, with which these calculations can be compared. The solar-particle calculations are more difﬁcult and more approximate than the cosmic-ray calculations due to magnetospheric effects, time dependence and the sparseness of the data base on which these calculations are based. Nonetheless the approach and the results are felt to be reasonable.

The theory of cosmic-ray and high-energy solar-particle transport in the atmosphere

43

Acknowledgements I wish to acknowledge many useful discussions with Ernst Felsberger, Wolfgang Heinrich, Brent Lewis and Herbert H. Sauer. Stefan Rösler generously supplied me with his code for calculating non-vertical cutoffs, and with his low-energy neutron spectrum, both of which play important parts in these calculations. I wish especially to acknowledge M.A. Shea and D.F. Smart for generously giving me their calculations of the 1995 Epoch of the vertical cutoff distribution in advance of its publication.

References [1] C. Passow, Phenomenologische Theorie zur Berechnung einer Kaskade aus schweren Teilchen (Nukleonenkaskade) in der Materie (Phenomenological theory for the calculation of a cascade of heavy particles (nucleonic cascade) in matter), Notiz A 285, Deut. Elektron. Synchrotron, Hamburg, 1962. [2] F.S. Alsmiller, A general category of soluble nucleon–meson cascade equations, ORNL-3746, 1965. [3] J.P. Elliott, Proc. Roy. Soc. London Ser. A 228 (1955) 424. [4] M.M.R. Williams, Nucleonik 9 (1966) 305. [5] G. Rudstam, Z. Naturforsch. A 21 (1966) 1027. [6] K. O’Brien, J. Geophys. Res. 75 (1970) 4357. [7] G.J. Molina-Cuberos, J.J. Lopez-Moreno, R. Rodrigo, L.M. Lara, K. O’Brien, Planet. Space Sci. 47 (1999) 1347. [8] L.L. Beranek, Acoustic properties of gases, in: D.E. Gray (Ed.), American Institute of Physics Handbook, McGraw–Hill, 1957. [9] R. Kallmann-Bijl, L.F. Boyd, H. Lagow, S.M. Poloskov, W. Priester, COSPAR International Reference Atmosphere, North-Holland, Amsterdam, 1986. [10] K. O’Brien, A semi-empirical model of inclusive nucleon and pion production from proton–nucleus collisions, Rep. HASL-261, US Atomic Energy Comm., New York, 1972. [11] H. Meyer, M.W. Teucher, E. Lohrmann, Nuovo Cimento 28 (1963) 1399. [12] J. Ranft, Nucl. Instrum. Methods 48 (1967) 133. [13] T.W. Armstrong, K.C. Chandler, J. Barish, J. Geophys. Res. 78 (1973) 2715. [14] W.N. Hess, E.H. Canﬁeld, R.E. Lingenfelter, J. Geophys. Res. 66 (1961) 665. [15] S. Roesler, W. Heinrich, H. Schraube, Radiat. Res. 149 (1998) 87. [16] H.L. Beck, Nucl. Sci. Eng. 39 (1970) 120. [17] B. Peters, Cosmic rays, in: Condon, Odishaw (Eds.), Handbook of Physics, McGraw–Hill, New York, 1958, pp. 9-201–9-244. [18] T.K. Gaisser, T. Stanev, Cosmic rays, European Phys. J. C 3 (1998) 132. [19] M. Garcia-Muñoz, G.M. Mason, J.A. Simpson, Astrophys. J. 202 (1975) 265. [20] M.A. Shea, D.F. Smart, A ﬁve by ﬁfteen degree world grid of calculated cosmic-ray vertical cutoff rigidities for 1965 and 1975, in: 14th International Cosmic Ray Conference, Conference Papers 4415, 1979; Available from Max-Planck Institut für Extraterrestriche Research, Munich, Germany. [21] M.A. Shea, D.F. Smart, A world grid of calculated cosmic ray vertical cutoff rigidities for 1980.0, in: 18th International Cosmic Ray Conference, Conference Papers 3:415, 1983. Available from Tata Institute of Fundamental Research, Bombay, India. [22] M.A. Shea, D.F. Smart, J.R. McCall, Can. J. Phys. 46 (1988) S1098. [23] D.F. Smart, M.A. Shea, Galactic cosmic radiation and solar energetic particles, in: A.S. Jursa (Ed.), Handbook of Geophysics and the Space Environment, National Technical Information Service, Springﬁeld, VA, 1985, pp. 6-1–6-29. [24] L.J. Gleeson, W.I. Axford, Astrophys. J. 149L (1967) 116. [25] H.V. Neher, J. Geophys. Res. 72 (1967) 1527.

44

K. O’Brien

[26] M. Lowder, D. Raft, L. Beck, Experimental determination of cosmic-ray charged particle intensity proﬁles in the atmosphere, in: E.A. Warman (Ed.), Proceedings of the National Symposium on Natural and Manmade Radiation in Space, Las Vegas, Nevada, March 1–5, 1971, in: National Aeronautics and Space Administration Report NASA TM X-2440, 1971, pp. 908–913. [27] G. Boella, G. Degli Antoni, C. Dilworth, G. Giannelli, E. Rocca, L. Scarsi, D. Shapiro, Nuovo Cimento 29 (1963) 103. [28] G. Boella, C. Dilworth, M. Panetti, L. Scarsi, Earth Planet. Sci. Lett. 4 (1968) 393. [29] F. Hajnal, J.E, McLaughlin, M.W. Weinstein, K. O’Brien, Sea-level cosmic-ray neutron measurements, USAEC Report HASL-241, 1971. [30] J.E. Hewitt, L. Hughes, J.W. Baum, J.B. McCaslin, A. Rindi, R. Smith, L.D. Stephens, R.H. Thomas, R.V. Grifﬁth, C.G. Welles, Health Phys. 34 (1978) 375. [31] R.A.R. Kent, Cosmic ray neutron measurements, USAEC Report HW-SA-2870, 1963. [32] M. Yamashita, L.D. Stephens, H.W. Patterson, J. Geophys. Res. 71 (1966) 3817. [33] M. Florek, I. Szarka, K. Holy, D. Nikodemova, H. Hrabovcova, Radiat. Prot. Dosim. 67 (1996) 187. [34] T. Watanabe, ftp://ftp.env.sci.ibaraki.ac.jp, 1998. [35] K. O’Brien, H.H. Sauer, Technology 7 (2000) 449.

45

Capabilities and limitations of high-resolution gamma spectrometry in environmental research J. Uyttenhove Ghent University, Radiation Physics Laboratory, Krijgslaan 281 (S-12), B-9000 Gent, Belgium

The availability of large intrinsic germanium detectors has opened new perspectives for fast, accurate and reliable environmental measurements. Many important isotopes in the natural and the technologically enhanced radiation environment have some suitable gamma lines, allowing quantitative determination of the radionuclides by high-resolution gamma spectrometry. In this paper, the capabilities and the limitations of a typical gamma spectrometry set-up that can be used for in-situ applications and laboratory measurements on soil samples are discussed and emphasis is laid on the sensitivity of laboratory measurements as well as on the applications for systematic surveys. The results of a survey for post-Chernobyl 137 Cs and natural radiation (40 K, U- and Th-series) in Belgium, obtained using a portable Ge-detector, clearly illustrate the advantages of high-resolution spectroscopy for in-situ applications. Such techniques can also be employed for measuring depleted uranium (DU) in Kosovo. The question of determining the Minimum Detectable Activity (MDA) of the set-up is always very important. A usual approach deﬁnes MDA as the maximum amount of activity for a particular nuclide that could be present and remain undetected in a real spectrum, with a given fractional error (e.g. 20%). The MDA for DU in soil samples (100–150 grams) was extensively investigated for the given set-up. The calculated 20% MDA value for DU is as low as 15 Bq for a 18 h long measuring time. Illustrating the capabilities of the set-up, the results of a 50 sampling point survey in Kosovo are presented. A total of 150 soil samples were measured for natural radionuclides, fallout isotopes and DU content. In none of the 150 measured samples was a signiﬁcant trace of the 1001 keV line found. The mean values for the natural isotopes in Kosovo soil are quite similar to those found in the south of Belgium. The mean value of 137 Cs concentration for the top Kosovo soil layers investigated is 53 Bq kg−1 . 1. Introduction The availability of large intrinsic germanium detectors has opened new perspectives for fast, accurate and reliable environmental measurements. Many important isotopes in the natural RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07005-6

© 2005 Elsevier Ltd. All rights reserved.

46

J. Uyttenhove

and the technologically enhanced radiation environment have some suitable gamma lines, allowing quantitative determination of the radionuclides by high-resolution gamma spectrometry. In this paper, we discuss the capabilities and the limitations of a typical gamma spectrometry set-up that can be used for in-situ applications and for laboratory measurements on soil samples. After a short review of post-Chernobyl in-situ results for Belgium, we lay emphasis on the sensitivity of laboratory measurements on soil samples and its applications in a systematic survey for depleted uranium (DU) in Kosovo. All our results are obtained with a portable Ge-detector (34% efﬁciency, 1.75 keV resolution fwhm) mounted in a multi-attitude cryostat (7.5 litre dewar) and coupled to an InSpector2000 (all from Canberra Ind.) with DSP and trapezoidal pulse-shaping (12 μs rise time, 0.8 μs ﬂat top). The measuring chain is calibrated with NIST sources SRM 4276 (mixed-radionuclide 125 Sb–125m Te, 154,155 Eu) and SRM 4353 (Rocky Flats soil containing 234,238 U, 228,230,232 Th, 228 Ac, 226 Ra, 137 Cs, 40 K). In a 4096 channel spectrum (energy range from 80 to 1500 keV), we can measure γ-emitting isotopes from the natural U and Th series, 40 K, depleted uranium (DU), and 137 Cs. The most important natural contributions are from: • The 238 U-series (half-life 4.5 billion years): 214 Pb (242.0, 295.2 and 351.9 keV lines) and 214 Bi (609.3 and 1120.3 keV lines) • The 232 Th-series (half-life 14 billion years): 228 Ac (338.3, 911.2 and 969.0 keV lines), 212 Pb (238.6 keV line) and 208 Tl (583.2 keV) • Potassium (40 K) (half-life 1.3 billion years): 40 K (1460.8 keV line) Other important nuclides are: 234m Pa (DU) (1001.0 keV); 137 Cs (661.6 keV). All those nuclides can be measured via relatively high-energy gamma lines (238 to 1461 keV), without problems with self-absorption; no special source preparation is required for the laboratory measurements with “survey” accuracy (15–20%).

2. Review of in-situ measurements The results of a survey for post-Chernobyl 137 Cs contamination and natural radiation (40 K, Uand Th-series) in Belgium clearly illustrate the advantages of high-resolution spectroscopy for in-situ applications. As most of these results are published earlier [1,2], only a short review of the method and the results is given here. The in-situ technique provides a means for the complete characterisation of the gamma radiation ﬁeld at a given location. It allows a fast, accurate and sensitive determination of radionuclides in the soil. A better soil representativity can be achieved compared with sample collection and subsequent laboratory analysis (representativeness of the sample and hot spots problems). An unshielded detector placed one meter above the ground (Fig. 1) detects gamma rays from an area of approximately 20 meter radius, representing a huge volume of soil. Figure 2 gives a typical spectrum of a measurement in a region in the extreme south of Belgium; one can easily see the 662 keV 137 Cs line together with the 1461 keV 40 K line and the typical lines of the natural U and Th series. A measuring time between 30 and 60 minutes allows accuracy better than 15% at the actual remaining 137 Cs concentrations from Chernobyl

Capabilities and limitations of high-resolution gamma spectrometry in environmental research

47

Fig. 1. In-situ detector set-up.

Fig. 2. Typical in-situ gamma spectrum.

in Belgium. In the calculations of the 137 Cs concentration, we use an exponential depth proﬁle with relaxation parameter α = 20 m−1 (equivalent with a relaxation length of 5 cm) which is applicable to aged fallout [3]. In Tintigny (see the spectrum in Fig. 2) a national maximum value for 137 Cs (6600 Bq m−2 ) is measured in the summer of 1997. About 60 measurements,

48

J. Uyttenhove

Fig. 3. Results for 137 Cs.

Fig. 4. 40 K results.

Capabilities and limitations of high-resolution gamma spectrometry in environmental research

49

equally distributed across the Belgian territory (30 507 km2 ) were executed in 1995–1997. The results are displayed as a post map in Fig. 3. Typical 137 Cs values range between 0.5 and 6.6 kBq m−2 . A conversion factor of 1 μSv y−1 per kBq m−2 [4] gives a maximal annual extra dose of 6.6 μSv; it is negligible in comparison with the normal annual dose of 3.5 mSv. The higher concentrations in the eastern part of Belgium can be explained if we look at the rainfall data on those critical days in May 1986 (rainfall map in [1]). For natural isotopes, a homogeneous distribution in the soil is assumed (α = 0 m−1 ). In the map of Fig. 4, the 40 K results are displayed, based on the 1461 keV line. As on Fig. 3, the numbers on the axes are Lambert coordinates as used on military maps in Belgium.

3. Laboratory measurements: spectrometry on Kosovo soil samples A systematic DU study on soil samples from Kosovo clearly illustrates the capabilities and limitations of high-resolution gamma spectrometry in environmental research, even in the case of isotopes with unfavourable γ-spectroscopic data like 238 U or depleted uranium. 3.1. Experimental conditions The same intrinsic germanium detector (P-type 34%, 1.75 keV fwhm) is used in special old lead and copper shielding. We also use the same DSP InSpector2000, USB coupled to a PC with Genie2000 software for data analysis. Soil samples are collected in 100 ml PE vials with masses between 120 and 160 gram. A good shielding of the detector is essential for the sensitivity, as discussed later in the MDA (minimal detectable activity) paragraph. 3.2. Some data about natural uranium and DU DU is a by-product of nuclear fuel production; when the ﬁssile 235 U is separated from the natural uranium during the enrichment process, a resulting waste product is DU, containing app. 99.74% 238 U. Chemically, it behaves like natural U, but its activity is less (only 60 to 75% of the natural U, depending on the enrichment process). It is also possible that DU is originating from the reprocessing of burned nuclear fuel elements; in that case some traces (0.01–0.1%) of 236 U can be expected [5]. Speculations on the presence of relevant amounts of Pu in that type of DU have until now never been conﬁrmed. Uranium (and DU) is a heavy metal with chemical toxicity, like many other heavy metals (Pb, Hg, etc.). Typical activity for natural uranium (without daughter products) is 26 Bq mg−1 ; DU (of type 1) has 19.4 Bq mg−1 (following the data in the CEA-report [5]). Due to its very high density, DU is used as a counterweight in ships, planes and rockets. For military weapon, DU-penetrators used in bullets give very high impact forces for piercing armour plates. DU powder is also pyrophoric; it can ignite spontaneously at temperatures above 600◦ . The combination of those two properties makes DU-ammunition very lethal for tanks and soldiers inside.

50

J. Uyttenhove

3.3. Spectroscopic data on DU The Gamma ray spectrum catalogue by R.L. Heath [6] – available on the Internet – gives a clear overview of the situation. In the decay of natural uranium (99.27% 238 U, 0.718% 235 U), we have an equilibrium between the U-mother (half-life 4.47 × 109 y) and the Pb and Bi daughter isotopes far down the decay process; the typical gamma radiations of 214 Pb (242, 295 and 352 keV) and 214 Bi (609 and 1120 keV) have good and well-known intensities and can easily be measured with a germanium detector. In the case of DU, however, the decay process starts with pure 238 U, without any daughter isotopes (removed in the enrichment process). The long half-life of 234 U (246 000 y) is a real barrier for the formation of 214 Pb and 214 Bi because any DU production started only after 1940! Only the 234 Pa and 234 Th isotopes are in equilibrium with 238 U after a few months. Fortunately, we have a typical but weak 1001.03 keV line in the decay of 234m Pa that can be used for the selective determination of DU. Absorption and detector efﬁciency problems are the reason why the low-energy lines (98 keV from the 234m Pa decay and 62 and 92 keV lines from the 234 Th decay) are not really useful for the determination of DU with our detector type. As one can see in the spectrum in Fig. 5, lower trace, the mean background of our set-up in the 1000 keV region is very low (2.2 counts/channel in 10 000 s). The only drawback is the small emission probability of the 1001 keV-line; we use the value 0.835(2)%, based on the review article by Nzuruba [7]. One should be very careful about the 234m Pa intensities used in different catalogues and software packs; old values of 0.59 for this 1001 keV line can be found! Also in the case of the second, very weak 766.36 keV line from the 234m Pa decay the situation is not clear: one can ﬁnd values between 0.294% (in the R.L. Heath catalogue [6] and in the Canberra NuChart-based software) and 0.315(3)% in Nzuruba [7]. The background in the 766 keV region is also much higher than in the 1001 keV region; consequently, we use the 766 keV line only for control and conﬁrmation purposes and all our calculations are based on the 1001.03 keV line. One should always remember also that the natural uranium in the soil gives a very small 1001 keV line. Strictly speaking, our DU measurements are based on the excess 234m Pa activity in the 1001 keV line. However, with the normal uranium

Fig. 5. High-energy region of 2 spectra (measuring time 40 000 s). Lower trace: a typical spectrum of a Kosovo soil sample (#52A, 133 g); no DU detected. Upper trace: a calibration spectrum with 2.4 milligramme DU added to a similar sample (see Section 3.4).

Capabilities and limitations of high-resolution gamma spectrometry in environmental research

51

concentrations in the 25 Bq kg−1 region, this “natural” contribution falls into the background and no 1001 keV line is detected in the absence of DU (as one can see in the lower trace of Fig. 5). In a recent publication, another approach for discriminating natural U and DU, also based on γ-ray spectrometry, is published by Miki Shoji et al. [8]. The three samples collected at each location must give coherent data for the uranium series, the thorium series and 40 K (nearly homogeneous distribution in the soil); the 661.6 keV line from 137 Cs is mainly expected in the top layer sample. Because of this, our spectroscopic method has internal controls on the quality of the sample-taking procedure. 3.4. Minimal detectable activity (MDA) calculations and experimental tests An important ﬁgure of merit is the sensitivity of our set-up or what is the minimal detectable activity (MDA). A usable approach [9,10] deﬁnes MDA as the maximum amount of activity for a particular nuclide that could be present and remain undetected in a real spectrum. The MDA expression takes into account detector parameters (resolution, photo peak efﬁciency and peak-to-compton ratio), crystal shielding (background), isotope properties and measuring time in the following combination [9]: Am MDA(E1 ) = 2bR(E1 ) B(E1 ) + A2m /4 + Am /2 ef (E1 ) · f · tm with the following notations: Am : the reciprocal of the fractional error (e.g. 20% means Am = 5); ef (E1 ): the absolute detector efﬁciency at E1 in the given source geometry; f : the branching ratio or relative intensity; tm : measuring time; b: shape factor; approx. 2 for symmetrical gaussian peaks; R(E1 ): the energy resolution (fwhm) at E1 , expressed in channels;

B(E1 ): the mean background in a channel in the E1 peak region; it is composed of the natural background contribution from the shielding and from the Compton contributions from higher-energy lines. In the case of the 1001 keV line from 234m Pa, the MDA is affected negatively by the very small intensity (0.00835); on the other side, the low background of our set-up in the 1000 keV region is favourable. The resulting calculated 20% MDA value is 20 Bq for an 11 h measuring time and 15 Bq for the longer 18 h measuring time. To conﬁrm our calculated MDA values, calibration samples containing a well-deﬁned amount of DU as a hot particle in the centre of the bottle have been prepared. The upper trace in Fig. 5 is part of the spectrum from a control soil sample doped with 2.4 (±0.2) milligramme DU. Three 228 Ac lines (from the natural 232 Th-series) are very similar, but the 1001 keV line from 234m Pa is exclusively and unambiguously present in the calibration spectrum. Using the uranium (and 234 Pa) conversion factor [10] of 12.27 Bq mg−1 , a value of 30 (±3) Bq is expected. The measured activity value after analysis with the Genie 2000 software is 33 (±7) Bq, in excellent agreement with the expected value. A second calibration source containing only 0.9 (±0.2) milligramme DU was used in a very long 100 000 s test measurement with satisfying results.

52

J. Uyttenhove

Summarising, we have proven, by MDA-calculations and experimentally, that we are able to detect DU particles in the milligramme region in an untreated soil sample by gamma spectrometry in a reasonable measuring time. 4. Organisation and results of the DU survey After the 1999 NATO intervention in Kosovo, a systematic DU survey was started in the summer of 2000 as a collaborative effort between Ghent University (Physics Laboratory) and the Belgian Army (MSW–Medical Research–Belgian Army) [11]. The 50 sampling sites were

Fig. 6. Map of the 52 collecting sites in Kosovo: regular sampling sites (circle), DU sites (black square). Sites 17 and 31 have been rejected because of incoherent 40 K data in the 3 samples, so that 50 sites remain. (“Incoherent” means that the values in the surface layer are far too low compared with the other layers; probably some recent modiﬁcations of the top layer occurred.)

Capabilities and limitations of high-resolution gamma spectrometry in environmental research

53

chosen allover Kosovo (Fig. 6); three samples on each point were taken at different depths. We included military positions, open ﬁelds, roadsides, and locations near places targeted with classical or DU-containing ordnances. The term “near” means that the samples are taken outside the impact area, but within a 25-meter radius around the impact point. Our study explicitly excludes the targeted places themselves for, besides DU-risks, such places can be very dangerous from the possible presence of unexploded munitions and toxic chemicals. The survey gives only information about possible widespread environmental DU contamination in Kosovo, not about localised contamination on impact sites. In none of the 150 measured samples was a signiﬁcant trace of the 1001 keV line found. Based on our MDA-considerations we can state with good conﬁdence that there is no DU present at our 50 sampling points in Kosovo, with MDA values as low as 15 Bq (corresponding to milligramme DU amounts in a 100–150 gram sample). Some samples, taken near 14 places where DU-ammunitions were used, have been re-examined very carefully with extralong measuring times (27.8 h), always with negative results. It was possible to perform all measurements within a 3 month period using gamma spectrometry; no special sample preparation or chemical process was needed. Our research does not address the situation in other regions of the FRY and our conclusions are based on 150 samples; it is possible that in other places traces of DU can be found. However, it is very unlikely that, if DU-particles should be widely spread over the region, not one single particle in the milligramme range is present in our samples.

5. Results for other nuclides The results for the natural nuclides in Kosovo are: U-series: mean value 25 Bq kg−1 (based on 214 Pb and 214 Bi (5 lines)) (22 Bq kg−1 in Belgium); Th-series: mean value 37 Bq kg−1 (based 228 Ac, 212 Pb and 208 Tl (5 lines)) (24 Bq kg−1 in Belgium); 40 K: mean value 370 Bq kg−1 (based on the 1460.8 keV line) (335 Bq kg−1 in the southern part of Belgium [Ardennes]). All data are for the surface layer A (0–10 cm); the corresponding mean values for Belgium are mentioned between parentheses. The artiﬁcial 137 Cs, resulting from atmospheric fall-out (nuclear bomb testing, Chernobyl accident) is easily measured via the intense 661.6 keV line; the mean value for the top layer is 53 Bq kg−1 . This ﬁgure is difﬁcult to compare with the values in Belgium, obtained by in-situ spectrometry and expressed in Bq m−2 , but in some places the concentrations in Kosovo are considerably higher than in Belgium (7 to 40 Bq kg−1 in 27 soil samples). It is worthwhile to note that we have found values for 137 Cs in Kosovo between practically zero and peaks in the range from 187 to 607 Bq kg−1 in some top layer samples. These data (Fig. 7) are the ﬁrst systematic post-Chernobyl results for that region; in the Atlas of Caesium Deposition on Europe After the Chernobyl Accident, published in 1998 by the European Commission [12], Kosovo and the surrounding regions (Yugoslavia, FYROM, Bosnia-Herzegovina) are white spots.

54

J. Uyttenhove

Fig. 7. Post map with the measured top-layer 137 Cs concentrations in Kosovo.

The higher concentrations in the central and northern part of Kosovo are probably correlated – like in Belgium – with local rainfall during the critical period between end-April and beginning-May 1986, but we have until now insufﬁcient information to conﬁrm this hypothesis.

6. General conclusions The post-Chernobyl survey in Belgium and the Kosovo DU study clearly illustrate the capabilities and limitations of high-resolution gamma spectrometry in environmental research. High-resolution spectrometry gives reliable quantitative results for all γ-emitting nuclides, natural or artiﬁcial, in one measurement. Most of the important isotopes in environmental research have suitable γ-lines with usable intensities. The in-situ technique provides a fast, reliable and sensitive means for the complete characterisation of the gamma radiation ﬁeld at a given location, with quantitative determination of the most important radionuclides in the soil. With laboratory measurements – even on an untreated soil sample like in the DU study – we are able to detect particles in the milligramme region by gamma spectrometry in a reasonable measuring time. A well-shielded and large (efﬁcient) detector even allows the use of lowintensity lines like the 1001 keV from 234m Pa decay.

Capabilities and limitations of high-resolution gamma spectrometry in environmental research

55

Other methods (mass-spectrometry, α-spectrometry) are sometimes more sensitive, but require special source preparation and have some other problems (U-isotopes ICPMS).

Acknowledgements The author is indebted to Dr. S. Pommé and the SCK-CEN (Mol, Belgium) for the collaboration in the post-Chernobyl 137 Cs survey. For the Kosovo survey, the excellent collaboration with the Belgian Army (General R. Van Hoof, commander in chief of Belgian Military Medical Service and MSW, Prof. M. Zizi) is strongly appreciated. The 52 places in Kosovo were visited by a military team under supervision in the ﬁeld by M. Lemmens.

References [1] J. Uyttenhove, et al., Health Phys. 73 (4) (1997) 644–646. [2] J. Uyttenhove, S. Pommé, B. Van Waeyenberge, in: IRPA 10 Proceedings, Hiroshima, Japan, 2000, CDRom, P-1a-14. [3] ICRU report 53, ICRU, Bethesda, MD, USA, 1994. [4] M. Balonov, P. Jacob, I. Likhtarev, V. Minenko, Pathways, levels and trends of population exposure after the Chernobyl accident, in: The Radiological Consequences of the Chernobyl Accident, Proceedings of the First International Conference, Minsk, Belarus, EUR 16544 EN, 1996, pp. 235–249, ISBN 92-827-5248-8. [5] Uranium, uranium appauvri, effets biologiques, CEA report (in French), http://www.cea.fr/actualite/ archives.asp?id=77. [6] R.L. Heath, Gamma Ray Spectrum Catalogue, 4th ed., 1998, pp. 832–835, http://id.inel.gov/gamma/data1. html or new data (2000) on 238 U: http://id.inel.gov/cgibin/byteserver.pl/gamma/pdf/u238.pdf. [7] A. Nzuruba, Nucl. Instrum. Methods A 424 (1999) 425–443. [8] M. Shoji, et al., Appl. Radiat Isot. 55 (2001) 221–227. [9] J. Cooper, Nucl. Instrum. Methods 82 (1970) 273–277. [10] J. Uyttenhove, A. Pofﬁjn, Radiat. Prot. Dosim. 34 (1990) 207–209. [11] UNEP/UNCHS Balkans Task Force (BTF), The potential effects on human health and environment arising from possible use of depleted uranium during the 1999 Kosovo conﬂict, October 1999. [12] European Commission, ISBN 92-828-3140-X, 1998.

56

Radon decay products in outdoor air J. Porstendörfer a,* , M. Gründel b a Am Hirtenberg 8, 37136 Waake, Germany b Isotope Laboratory of the Institute of Physical Chemistry, Georg-August-University, Tammannstr. 6,

37077 Göttingen, Germany

The values of the radon parameters, relevant for exposure and dose estimations, equilibrium factor (F ), unattached fraction (fp ), and activity size distribution of the radon progeny aerosol in outdoor air are reported. The results of these three parameters were obtained from investigations during the last ten years. The experimental results are described by model calculations.

1. Introduction The following is a summary of the results of our investigations to characterise the short-lived radon progeny in outdoor air, relevant for estimation of the exposure and the inhalation dose. These relevant radon parameters are: • equilibrium factor F = CEEC /C0 , which is the ratio of the equilibrium equivalent radon concentration CECC to the radon gas concentration C0 ; f f • unattached activities of the radon progeny expressed by fp = CEEC /CEEC , where CEEC is the equilibrium equivalent radon concentration of the unattached radon progeny; • relative activity size distribution of the aerosol attached radon decay products.

2. Measurement techniques In the following, we will brieﬂy mention the different measurement techniques used during our investigations. Details about these are available in special publications. For the measurement of these radon parameters special equipment was developed to carry out continuous measurements, also at low activity concentrations (1–10 Bq m−3 ) over longer time periods (up to three weeks). * E-mail address: [email protected] (J. Porstendörfer)

RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07006-8

© 2005 Elsevier Ltd. All rights reserved.

Radon decay products in outdoor air

57

F and fp were measured with our F –fp -monitor. This monitor made it possible to measure simultaneously and continuously the unattached and aerosol-attached progeny and the radon concentration. For the continuous measurement of the activity size distribution of the radon progeny aerosol a low-pressure Online-Alpha-Cascade Impactor (OACI) with nine stages 50% cutoff 60 nm and 15 500 nm was developed [1]. A Tube Diffusion Battery (TDB) at the entrance of the OACI was used to remove the unattached progeny. For the measurement of the number size distribution of the atmospheric aerosol in the diameter range 5–200 nm a Differential electrostatic aerosol Mobility Analyser (DMA) (TSI, Model 3071) in connection with a Condensation Nuclei Counter (CNC) (TSI, Model 3025) was used. The particle number size distribution in the diameter range 100–5000 nm was measured with a Laser Aerosol Spectrometer (LAS) (PMS, Model LAS-X). The particle number concentration of the atmospheric aerosol was measured with a Condensation Nuclei Counter (CNC) (TSI, Model 3025).

3. Equilibrium factor Radon formed in the soil is released into the atmosphere. After this exhalation radon and its decay products are distributed in the troposphere mainly by turbulent air mixing characterised by eddy diffusivity. The turbulent diffusion coefﬁcient varies with altitude according to the vertical variation of wind and atmospheric stability. Besides the eddy diffusivity, the concentration proﬁles are inﬂuenced by the decay constant of the radionuclides and by the wet and dry removal of the progeny from the atmosphere. The calculations of Jacobi and Andre [2] give basic information of the variation of radon and its progeny (characterised by the equilibrium factor F ) with height (Fig. 1). Four different turbulence conditions in the troposphere are assumed: the proﬁle SSN is strong, NNN a normal, WNN a rather weak, and IWN a very weak mixing throughout the troposphere. In the case of IWN there is a stable inversion in the atmosphere near the ground. The radon concentration can change up to factor 100 for different weather conditions (Fig. 2), which was conﬁrmed by several measurements (e.g. [3,4]). In addition, the prediction by model calculations shows, that the equilibrium factor in most cases varies between 0.54 and 0.76 at 1 m above the ground. Only for strong turbulent conditions the F -factor is greater than 0.8. Figure 3 shows a diurnal evaluation of a 3-week measurement period. The concentration of radon (C0 ) and of the decay products (CEEC ) show a maximum during the night, which is typical for the different turbulent mixing conditions during the day/night hours. Due to the higher turbulent mixing during the day hours, the radon concentration is lower and the F value is higher than during the night. This behaviour is in agreement with the theoretical prediction. The equilibrium factor varies between 0.57 and 0.76. Of course, the average daily variation of the radon concentration can vary depending on the place with its speciﬁc weather conditions. For Continental Europe with a predominantly continental climate (off shore areas), the average values in Fig. 3 are typical.

58

J. Porstendörfer, M. Gründel

Fig. 1. Distribution of radon concentration and F -factor above the ground (model calculations).

Fig. 2. Variation of the radon concentration and the F -factor (1 m above the ground) for different turbulence conditions of the troposphere.

Radon decay products in outdoor air

59

Fig. 3. Diurnal variation of the radon concentration (C0 ), equilibrium equivalent radon concentration (CEEC ), F -factor, wind speed and the temperature 1 m above the ground averaged over a measurement period of three weeks.

4. Unattached fraction In air, the radon decay products exist in two forms: (1) as unattached fraction, or (2) attached to the surface of aerosol particles. After their formation, the unattached decay products are predominantly (88%) positively charged. These ions react with trace gases and water vapour and become small particles called clusters. Simultaneously these clusters attached to aerosol particles and they can be neutralised. Lab-experiments show that the neutralisation rate strongly depends on the ionisation rate and the humidity in air [5]. The ionisation rate in air is mainly determined by the gamma and cosmic radiation and the alpha radiation of radon and thoron and their decay products. In outdoor air near ground level, the neutralisation rate is about 30 h−1 . Based on this investigation a fraction of about 30% of 218 Po clusters in outdoor air is positively charged.

60

J. Porstendörfer, M. Gründel

Theoretical considerations show that the dominant parameter which inﬂuences the fraction of the unattached radon progeny is the attachment rate to the atmospheric aerosol: ∞ ∂Z(d) ∂d X= β(d) ∂d 0 where • ∂Z(d) ∂d is the number size distribution, ∞ • Z = 0 ∂Z(d) ∂d ∂d is the particle concentration, and • β(d) is the attachment coefﬁcient describing the attachment probability to an aerosol particle with the diameter d. The unattached fraction of the short-lived radon decay product j with the decay constant λj is fj =

Cjf Cj

=

Cjf

(1)

Cjf + Cja

with j = 1: 218 Po; j = 2: 214 Pb; j = 3: 214 Bi/214 Po. Cjf and Cja are the activity concentrations of the unattached and aerosol-attached decay products. Neglecting the activity losses by plate-out on surfaces and because for almost all environmental air conditions is λj < X [6,7], the unattached fractions fj can be written approximately as f1 =

λ1 , X

f2 =

λ 2 R1 , X

f3 =

λ 2 λ 3 R1 . X2

With the equilibrium equivalent radon concentration CEEC = fp =

(2) 3

j =1 kj Cj ,

f CEEC 414 k 1 λ 1 + k 2 λ 2 R1 k 1 λ 1 + k 2 λ 2 R1 = = ≈ ¯ CEEC X Z[cm−3 ] βZ

one obtains (3)

with k1 = 0.105, k2 = 0.516, and the recoil factor R1 = 0.8, which is the fraction of the desorption from the particle surface after the alpha decay of 218 Po. β¯ = 1.4 × 10−6 cm3 s−1 is the average attachment coefﬁcient for the atmospheric aerosol obtained from measurements [6]. This prediction by model calculation is in agreement with results by measurements at different places [8]. The daily variation of the unattached fraction and the particle number concentration are illustrated in Fig. 4. The average diurnal variation of the number concentration of the aerosol in the atmosphere shows a signiﬁcant increase of the particle concentration during the day hours, dominantly caused by the photochemical reaction (gas to particle conversion by sun-light) and human activities (trafﬁc, combustion). The lower unattached fraction, measured during these hours with their higher aerosol particle concentration is in agreement with the theoretical prediction. The fp -values vary between 0.02 and 0.04 in agreement with the measurements of Hattori et al. [9].

Radon decay products in outdoor air

61

Fig. 4. Diurnal variation of the aerosol particle concentration (Z) and the unattached fraction (fp ) 1 m above the ground averaged over a measurement period of three weeks.

5. Activity size distribution of the aerosol The activity size distribution of the progeny aerosol Cj (d) and the number size distribution Z(d) of the aerosol are different, because the attachment probability β(d) is a function of particle diameter d [10]. The correlation between both size distributions can be expressed by [11] Cj β(d)Z(d), (4) X Cj is the activity concentration and X is the attachment rate, expressing the adsorption velocity of the decay product to the atmospheric aerosol. The activity size distribution Cj (d) can be calculated (equation (4)), if the number size distribution Z(d) of the atmospheric aerosol is known by measurement. This method of determining the activity size distribution includes the inaccuracy of the attachment coefﬁcient, derived and experimentally conﬁrmed only for spherical aerosol particles [10]. Figure 5 shows the daily variation of the number size distribution of the atmospheric aerosol average values of a continuous measurement period over 22 days [12]. Signiﬁcantly higher are the particle number concentrations in the size range smaller than 50 nm during the night hours than during the day hours. This diurnal variation is caused by the higher aerosol generation by the combustion processes and photochemical reactions during the daytime. Cj (d) =

62

J. Porstendörfer, M. Gründel

Fig. 5. The number size distribution in outdoor air at different times of the day/night.

Fig. 6. Calculated activity size distributions taking into account the measured number size distribution with the lowest (6.00 o’clock) and the highest (18.00 o’clock) particle concentrations during the day.

In Fig. 6 are two activity size distributions calculated with equation (4), taking into account the lowest (6.00 o’clock) and the highest (18.00 o’clock) particle concentrations with diameters < 50 nm [12,13]. For higher particle concentrations, an activity fraction in the nucleation size range exists. In the case of lower particle number concentrations as occurred during the nighttimes, no activity on nucleation particles was obtained. The direct measurement of the activity size distribution of the radon progeny aerosol gives typical values as illustrated in Fig. 7. These values were obtained with the Online Alpha Cascade Impactor averaged over a two-week measurement period. The experimental data can be approximated by a sum of two log-normal distributions. Most of the activity (about 80%) is adsorbed on particles of the accumulation mode with an Activity Median Diameter (AMD) 350 nm and a geometric standard deviation of σg = 2.2. The activity fraction on particles below 60 nm (collected on the back-up ﬁlter) was about 20%. This activity of the nucleation mode can be approximated by an AMD of about 40 nm. No coarse mode with particle sizes > 1000 nm was measured in contrast to previous measurements [12,13] where no tube diffusion battery in front of the impactor was used to remove the unattached radon progeny from the air stream. To see the variation of the activity size distribution during the measurement period, the AMDa value and the geometric standard deviation, σga , of the accumulation mode and the

Radon decay products in outdoor air

63

Fig. 7. Relative activity size distribution of 218 Po and the 214 Po aerosol in outdoor air averaged over a measurement period of 2 weeks.

Fig. 8. The variation of the activity median diameter (AMDa ) and the geometric standard deviation (σga ) of the accumulation size range and the activity on the back-up ﬁlter.

64

J. Porstendörfer, M. Gründel

Fig. 9. The diurnal variation of the activity fraction on back-up ﬁlter (< 60 nm) averaged over the measurement period May 31–June 21.

activity fraction on the back-up ﬁlter as function of time is presented in Fig. 8. The AMDa and the σga values vary between 250–500 nm and 1.8–3.0, respectively, but a diurnal variation and an inﬂuence of the weather parameters was not signiﬁcant. However, the variation of the activity on the back-up ﬁlter (sizes < 60 nm) showed a diurnal variation (Fig. 9). Its variation can be explained by the change of the number concentration of the atmospheric aerosol for particle diameters < 50 nm (see Fig. 5). A higher concentration of the nucleation particles (< 50 nm) during the day hours causes the higher activity fraction on the back-up ﬁlter.

References [1] J. Kesten, G. Butterweck, J. Porstendörfer, A. Reineking, H.J. Heymel, An online α-impactor for short-lived radon daughters, Aerosol Sci. Technol. 18 (1993) 156–164. [2] W. Jacobi, K. Andre, The vertical distribution of radon 222 and radon 220 and their decay products in the atmosphere, J. Geophysical Res. 68 (1963) 3799–3814. [3] G. Butterweck, Natürliche Radionuklide als Tracer zur Messung des turbulenten Austausches und der trockenen Deposition in der Umwelt, Dissertation, Georg-August-Universität, Göttingen, 1991. [4] J. Porstendörfer, G. Butterweck, A. Reineking, Daily variation of the radon concentration indoors and outdoors and the inﬂuence of meteorological parameters, Health Phys. 67 (1994) 283–287. [5] V. Dankelmann, A. Reineking, J. Porstendörfer, Determination of neutralisation rates of 218 Po ions in air, Radiat. Prot. Dosim. 94 (2001) 353–357. [6] J. Porstendörfer, T.T. Mercer, Inﬂuence of nuclei concentration and humidity upon the attachment rate of atoms in the atmosphere, Atmos. Environ. 12 (1978) 2223. [7] J. Porstendörfer, A. Reineking, K.H. Becker, Free fractions, attachment rates and plate-out rates of radon daughters in houses, in: P.H. Hopke (Ed.), Radon and its Decay Products – Occurrence, Properties and Health Effects, in: ACS Symp. Ser., vol. 331, 1987, pp. 285–300. [8] J. Porstendörfer, A. Reineking, Radon: characteristic in air and dose conversion factors, Health Phys. 76 (1999) 300–305. [9] T. Hattori, T. Schiji, K. Ishida, Equilibrium factor and unattached fraction of radon progeny in outdoor air, Radioisotopes 44 (1995) 710–714.

Radon decay products in outdoor air

65

[10] J. Porstendörfer, G. Röbig, A. Ahmed, Experimental determination of the attachment coefﬁcients of atoms and ions on monodisperse aerosols, J. Aerosol Sci. 10 (1979) 21–28. [11] J. Porstendörfer, Die Anlagerungsgeschwindigkeit der elektrisch geladenen und neutralen Emanationsfolgeprodukte an das atmosphärische Aerosol, Pure Appl. Geophys. 77 (1969) 175. [12] Ch. Zock, Die Messung der Aktivitätsgrößenverteilung des radioaktiven Aerosols der Radonfolgeprodukte und deren Einﬂuß auf die Strahlendosis beim Menschen, Dissertation, Georg-August-Universität, Göttingen, 1996. [13] J. Porstendörfer, Ch. Zock, A. Reineking, Aerosol size distribution of the radon progeny in outdoor air, J. Environ. Radioact. 51 (2000) 37–48.

66

Contribution of animal experimental data for the risk assessment of exposure to radon decay products G. Monchaux Institut de Radioprotection et de Sûreté Nucléaire (IRSN), BP 17, F-92262 Fontenay-aux-Roses, France

Experimental studies were used in addition to epidemiological studies to investigate the effects of exposure, exposure rates and other factors in predicting risks resulting from human exposures. The advantage of data from animals is that animals can be exposed under carefully controlled conditions and that exposure and exposure rate can be estimated more accurately. This review summarises data on lung cancer risk from radon experimental studies focusing on the exposure-rate effects. A dose–effect relationship was established in rats, which was very similar for medium and high cumulative exposures, to that observed in uranium miners. At high cumulative exposures up to 3000 WLM (10.8 J h m−3 ), an inverse dose-rate effect similar to that observed in uranium miners was also found in rats. In contrast, recent results from our group indicate that at relatively low cumulative exposures of 0.36 J h m−3 (100 WLM), comparable to lifetime exposures in high-radon houses or current underground mining exposures, the risk of lung cancer in rats decreases with PAEC, i.e., exposure rate. This conﬁrms the results obtained at lower cumulative exposure, showing that for a similar cumulative exposure of 25 WLM (0.09 J h m−3 ), the relative risk (RR) of lung cancer decreases with decreasing exposure rates. These data suggest that the induction of lung cancer results from a complex interplay between cumulative exposure and exposure rate, with an optimal combination of these two parameters, i.e., a combination of cumulative dose and dose rate that results in a maximum risk of lung cancer induction. They support the hypothesis that, at low doses, the risk of lung cancer is governed by the rate at which the dose is delivered, and not by the total cumulative dose alone. These data are also consistent with that of underground uranium miners showing an inverse dose-rate effect at high cumulative exposures, but a diminution of this effect at cumulative exposures lower than 50 WLM (0.18 J h m−3 ). They support both an inverse dose-rate effect at high cumulative exposures, as well as its diminution or disappearance at low cumulative exposures. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07007-X

© 2005 Elsevier Ltd. All rights reserved.

Contribution of animal experimental data for the risk assessment of exposure to radon decay products

67

1. Introduction An association between an excess risk of lung cancer and exposure to radon and its decay products has been demonstrated in uranium miners and other miners [1,2]. In various countries, measurements in dwellings have shown that in many homes, the radon concentrations are only one or two orders of magnitude lower than in typical underground mine situations [3,4]. Geographical epidemiological studies showed conﬂicting results and most of them did not show an excess lung cancer risk for people living in radon rich areas [5]. Some case– control studies, in particular Swedish studies, showed a signiﬁcant increased relative risk of lung cancer in relation with radon concentration in homes [6–8]. At the opposite, other studies, in particular the Canadian study of Létourneau et al. in the Winnipeg region [9], and two Finnish studies did not show any increase of lung cancer risk in relation with indoor radon exposure [10,11]. The risk related to indoor domestic exposure was estimated from the risk projections from underground miners data in association with measurements of indoor radon concentrations [12]. However, exposures and exposure rates were generally higher in mines compared with generally lower exposures and exposure rates in homes [13]. The pooled analysis of data from 11 cohorts of underground miners showed that the excess relative risk (ERR) of lung cancer increased as the exposure rate decreased. This pattern called protraction enhancement effect or inverse dose-rate effect was obvious at high cumulative exposures [14], but a diminution of this effect was observed for cumulative exposures below 50 WLM (0.18 J h m−3 ). Recently, the results of a meta-analysis including eight case–control studies [15] showed that the exposure–response trend was similar to model-based extrapolations from miners and to relative risk estimates computed directly from miners with low cumulative exposures. These results should be interpreted cautiously, until additional studies will be reported and the pooling of original data from different studies completed. The role of domestic radon exposure in the occurrence of lung cancer remains unclear. Experimental animal studies were used in addition to epidemiological studies to investigate the effects of exposure, exposure rate and other factors in predicting risks resulting from human radon exposures. This review summarises data on lung cancer risk from radon experimental studies performed by our group at CEA-COGEMA with emphasis on the most recent ﬁndings and analyses on the inﬂuence of dose rate.

2. Experimental data Experimental animal studies of radon-induced lung cancer are particularly helpful for understanding the carcinogenicity of human exposures both in the home and in the workplace. Animals can be exposed to a variety of agents under carefully controlled conditions and then sacriﬁced for the study of developing lesions or held for their lifespan for the development of tumours. Radon health effect data, developed primarily in adult male rats were provided by the Paciﬁc Northwest National Laboratory (PNNL) [16], formerly PNL, in USA, and in Europe by AEA-Technology in UK [17], and CEA-COGEMA in France [18]. Around 800 Syrian golden hamsters, 6000 speciﬁc pathogen-free Wistar rats, and 100 beagle dogs were exposed to mixtures of radon, radon progeny, diesel engine exhaust, uranium

68

G. Monchaux

ore dust, and cigarette smoke in PNNL studies. About 2000 SPF Sprague–Dawley rats were exposed to radon, radon progeny, at various cumulative exposures, and using Carnauba wax as a carrier aerosol, at AEA-Technology. More than 13 000 SPF Sprague–Dawley rats were exposed to mixtures of radon, radon progeny, ambient (natural) aerosols, cigarette smoke, mineral ﬁbres, minerals from metallic ores, diesel exhaust, ozone and chlorinated compounds in the CEA-COGEMA studies. Additional radon mechanistic carcinogenesis studies have been performed using intra-muscular injections of the promoter 5,6 benzo-ﬂavone (βNF) to clarify the role of cancer promoters in radon-induced lung cancers [19–21]. 2.1. Respiratory tract tumours Lesions of the extra-thoracic (ET) region, including tumours, were produced primarily in the nose and were associated with high-unattached fractions of radon decay products [22]. Major biological effects produced in the radon studies were respiratory tract tumours, pulmonary ﬁbrosis, pulmonary emphysema, and lifespan shortening. Appreciable ﬁbrosis, emphysema and lifespan shortening, although somewhat dependent of species, did not occur at exposure levels less than 1000 WLM (3.5 J h m−3 ). Lung tumours were produced in rats exposed to radon decay products alone, showing that associated exposures to other industrial or environmental airborne pollutants, such as uranium ore dust or cigarette smoke, are not required for lung cancer development. Respiratory tract tumours in rats were differentiated between malignant tumours (carcinomas and sarcomas) and preneoplastic lesions and pulmonary benign tumours. Preneoplastic lesions and pulmonary benign lesions were differentiated between non-neoplastic proliferative lesions – including alveolar epithelial hyperplasias and squamous cell metaplasias – and pulmonary benign tumours, adenomas. Among lung proliferative non-neoplastic lesions, bronchiolarisation or bronchiolo–alveolar hyperplasia resulted probably in the repopulation of denuded alveoli by bronchiolar epithelial cells. Highly ciliated cells could be observed in rats after inhalation of radon decay products, and most of the time they were originating from poorly differentiated bronchiolar cells. Simple bronchiolarisation might progress to focal alveolar hyperplasia with papillary structures also called adenomatosis, in which cells proliferate without evidence of destruction or invasion. Squamous metaplasia appeared somewhat like cysts with multiple layers of well-differentiated squamous cells, very often with a central core of keratin, and exhibiting later necrotic features. Adenomas, which are benign tumours, consisted in large adenomatoses with a more important papillary pattern and compressing the surrounding lung parenchyma. Lung carcinomas were differentiated into adenocarcinomas, squamous cell carcinomas and adenosquamous carcinomas. Adenocarcinomas could develop from any level of the respiratory tract and consisted of papillary neoplasms with well-differentiated cuboidal or columnar cells. They could be subdivided into mucus secreting (formerly classiﬁed as bronchogenic carcinoma), non-mucus secreting (formerly classiﬁed as non-bronchogenic carcinoma), anaplastic (undifferentiated) or large cell (poorly differentiated) types. Bronchiolo–alveolar carcinomas consisted of non-secreting papillary patterns of cuboidal cells from alveolar origin (type II alveolar cells) with dark round-shaped nuclei of different size. They frequently invaded the mediastinum. A more alveolar type without papillary features was also observed. Rat bronchiolo–alveolar carcinomas were considered and classiﬁed as a

Contribution of animal experimental data for the risk assessment of exposure to radon decay products

69

variant of adenocarcinomas [23]. Squamous cell (epidermoid) carcinomas consisted in irregular proliferation of stratiﬁed squamous epithelium that could be well or poorly differentiated. Adenosquamous carcinomas contained both glandular and squamous tissue. Although the oat-cell carcinomas that are common in humans were not found in rats, other histological types of lung carcinomas, especially squamous cell carcinomas and primitive lung adenocarcinomas were very similar to those observed in humans [24,25]. Unlike human lung tumours that are of bronchial origin, lung tumours observed in rats originate from terminal and respiratory bronchioles and the deep lung (alveoli), even they arise from the bronchial tree cells. It is noteworthy that squamous cell carcinomas observed in rats after radon exposure were very similar in the histology to those observed in humans [26]. Although the rat does not develop small cell carcinomas for itself in response to radon exposure, it is interesting to notice the similarity of positive bombesin staining reaction in rat squamous cell carcinomas and in human lung squamous cell carcinomas [25]. In the PNNL experiments, more than 70% of squamous cell (epidermoid) carcinomas, but only 20% of the adenocarcinomas were classiﬁed as fatal. Similar ﬁndings were also observed in the last series of experiments conducted at CEA-COGEMA. Finally, most of the radon-induced lung tumours are considered to originate peripherally and to occur near or at the bronchiolar–alveolar junction, in contrast to human lung tumours, which are generally more centrally located and bronchi associated [24]. 2.2. Dose–effect relationship It has been demonstrated that exposure to radon and its progeny induces lung cancer in rats [24]. An excess risk of lung cancers was observed in rats at cumulative exposure as low as 25 WLM (0.09 J h m−3 ) performed at relatively high potential alpha energy concentration (PAEC) of about 100 WL (2.1 mJ m−3 ) [25]. A dose–effect relationship was established showing that the incidence of lung cancers in rats exposed to radon/radon progeny increased with cumulative exposure [18]. The incidence of lung cancers in rats increased for cumulative exposures varying from 25 WLM (0.09 J h m−3 ) up to 3000 WLM (10.8 J h m−3 ), and decreased thereafter. For medium and high cumulative exposures, the pattern of this dose– effect relationship was very similar to that observed in underground uranium miners in whom a signiﬁcant excess of lung cancers was observed for cumulative exposure ranging from about 100 WLM (0.36 J h m−3 ) to 400 WLM (1.44 J h m−3 ) and above [1]. However, excess respiratory tract tumours were produced in rats at exposures considerably lower than 100 WLM (0.35 J h m−3 ), even at cumulative exposures (although not exposure rates) comparable to typical lifespan exposures in homes of about 20 WLM (0.07 J h m−3 ). A cumulative exposure of 25 WLM (0.09 J h m−3 ) at high exposure rates resulted in a tripled to four-fold increased lung cancer incidence in male Sprague–Dawley rats as compared to unexposed control rats [27]. 2.3. Inﬂuence of exposure rate In spite of the fact that signiﬁcant cumulative exposures at typical residential exposure rates of about 1.8 × 10−5 J h m−3 per week, approximately 0.005 WLM per week, cannot be tested in a short-lived species like the rat, the rat model is useful for reducing the uncertainties that exist in human data, particularly in regard to the exposure-rate effect. The “inverse dose-rate effect” observed in miners exposed at high cumulative exposure was also observed in rats. In the PNNL experiments, a trend towards an increasing tumour risk with

70

G. Monchaux

decreased exposure rate, has been reported in Wistar rats exposed at high PAEC of 100 WL (2.1 mJ m−3 ) and 1000 WL (21 mJ m−3 ), respectively, and at high cumulative exposures, varying from 640 WLM (2.3 J h m−3 ) up to 5120 WLM (18.4 J h m−3 ) [28]. In our studies in Sprague–Dawley rats, a similar trend was observed in rats exposed at high cumulative exposures ranging from 200 WLM (0.72 J h m−3 ) up to 3000 WLM (10.8 mJ h m−3 ) and at high exposure rates varying from 70 WLM per week (0.25 mJ h m−3 ) to 500 WLM per week (1.8 mJ m−3 ) [18,29]. In PNNL studies, protraction of exposures in rats produced a signiﬁcantly higher incidence of multiple primary lung tumours, more often of a different histological type than of the same types, and fatal primary lung tumours. Similar ﬁndings were observed in our studies in rats exposed at various PAEC. In the same way, it has been shown that at high cumulative exposure of 3000 WLM (10.8 J h m−3 ) and high PAEC of 1700 WL, protraction of exposure at 1 or 2 sessions per week protracted over 8 months resulted in a three-fold increase of lung cancer, as compared with rat exposed at daily exposure (Fig. 1). In contrast, the results obtained at low cumulative exposure, comparable to domestic indoor exposures showed no evidence of an inverse exposure-rate effect. Indeed, as shown on Fig. 2, chronic radon exposure at 25 WLM (0.09 J h m−3 ), protracted over an 18 months period at a PAEC of 2 WL (0.042 mJ m−3 ), resulted in less lung cancer frequency in rats than a similar cumulative exposure delivered over a 4 months period at a PAEC of 100 WL (2.1 mJ m−3 ) [30,31]. In this experiment, the lung cancer incidence was slightly lower in rats exposed at low exposure rate (0.60%) than in control animals (0.64%). In this series, the size of lung cancers increased with cumulative exposure in rats exposed at high exposure rates, as well as pleural extension. Moreover, the presence of multiple lung tumours or intrapulmonary metastases also increased in the group exposed at 50 WLM (0.18 J h m−3 ), and high PAEC of 100 WL (2.1 mJ m−3 , Group Rn50 HDR). Under the Fourth CEC Research and Development Framework Programme, a new series of experiments were carried out to investigate speciﬁcally the inﬂuence of exposure rate on lung cancer induction in rats [32]. These studies were conducted at relatively low cumulative ex-

Fig. 1. Inﬂuence of protraction of exposure in rats exposed at high cumulative exposure (3000 WLM) [3000 Pr: rats with protracted exposure].

Contribution of animal experimental data for the risk assessment of exposure to radon decay products

71

Fig. 2. Incidence of lung cancers in rats exposed at low cumulative exposure of 25 WLM and low PAEC of 2 WL (25 LDR) compared with controls and rats exposed at 25 and 50 WLM and high PAEC of 100 WL (25 HDR and 50 HDR).

posure comparable to lifetime exposures in high-radon houses or current underground mining exposures of about 100 WLM (0.36 J h m−3 ). The highest global incidence of lung cancers and the highest proportion of squamous cell carcinomas, were observed in rats exposed at high PAEC. In the other groups of rats, the global incidence of lung cancers and the relative frequency of squamous cell carcinomas decreased with decreasing PAEC, i.e., decreasing exposure rates. The larger and most invasive tumours were also observed in rats exposed at high PAEC and the size and invasiveness of tumours decreased with decreasing PAEC and/or protraction of exposure. Similarly, the number of non-neoplastic proliferative lesions and pulmonary benign tumours, in particular, the proportion of alveolar epithelial hyperplasia and adenomas decreased with decreasing PAEC. Figure 3 shows the pooled results of different experiments in rats exposed at low cumulative exposure of 100 WLM (0.36 J h m−3 ) and lower. It demonstrates obviously that the incidence of lung cancer depends not only on cumulative exposure (WLM) but also on PAEC (WL). These results indicate that at relatively low cumulative exposure of about 100 WLM (0.36 J h m−3 ) and lower, the risk of lung cancer in rats decreases with PAEC, i.e., exposure rate. They conﬁrm the results obtained at lower cumulative exposure [30,31], showing that for a similar cumulative exposure of 25 WLM (0.09 J h m−3 ), the relative risk (RR) of lung cancer decreases from 4.45 in rats exposed at 150 WL (3.15 mJ m−3 ) to 3.48 in rats exposed at 100 WL (2.1 mJ m−3 ) and to 0.9 in rats exposed at 2 WL (0.042 mJ m−3 ). Figures 4 and 5 pool the results of different experimental studies on the inﬂuence of cumulative exposure, PAEC and/or exposure rate on lung cancer risk. Figure 4 shows the combined inﬂuence of cumulative exposure expressed in terms of WLM and of PAEC, expressed in terms of WL, on the excess relative risk per unit exposure (ERR per WLM). For cumulative exposures higher than 100 WLM (0.36 J h m−3 ), and PAEC higher than 150 WL, the risk of lung cancer decreases with increasing cumulative exposure and increasing PAEC. In contrast, for cumulative exposures lower than 100 WLM (0.36 J h m−3 ), the risk of lung cancer decreases with decreasing PAEC. Moreover, the risk of lung tumour induction in rats appears to

72

G. Monchaux

Fig. 3. Incidence of lung cancers in rats exposed at low cumulative exposure of 100 WLM and lower and at various PAEC (rats exposed at 0.25 WLM are control rats).

Fig. 4. Excess relative risk (ERR) of lung cancer in rats exposed to radon decay products at various cumulative exposures (WLM) and PAEC (WL).

be maximal for cumulative exposures ranging from 25 WLM (0.09 J h m−3 ) up to 200 WLM (0.72 J h m−3 ), and PAEC ranging from 50 WL (1.05 mJ m−3 ) up to 150 WL (3.15 mJ m−3 ). Figure 5 shows the combined inﬂuence of cumulative exposure, expressed in terms of WLM and of exposure rate, expressed in terms of WLM per week, on the ERR per WLM. This ﬁgure exhibits a pattern that is very similar to that of Fig. 4. For cumulative exposures higher than 100 WLM (0.36 J h m−3 ), and exposure rates higher than 25 WLM per week (0.09 J h m−3 ),

Contribution of animal experimental data for the risk assessment of exposure to radon decay products

73

Fig. 5. Excess relative risk (ERR) of lung cancer in rats exposed to radon decay products at various cumulative exposures (WLM) and exposure rates (WLM per week).

the risk of lung cancer decreases with increasing cumulative exposure and increasing exposure rates. In contrast, for cumulative exposures lower than 100 WLM (0.36 J h m−3 ), the risk of lung cancer decreases with decreasing exposure rates. The highest risk of lung cancer induction occurs for cumulative exposures ranging from 25 WLM (0.09 J h m−3 ) up to 200 WLM (0.72 J h m−3 ) and exposure rates ranging from 5 WLM per week (18 mJ h m−3 per week) and 25 WLM per week (90 mJ h m−3 per week). These data suggest that the induction of lung cancer results from a complex interplay between cumulative dose and dose rate, with an optimal combination of these two parameters, i.e., a combination of cumulative dose and dose rate that results in a maximum risk of lung cancer induction.

3. Discussion The signiﬁcance of exposure rates in assessing the hazards of domestic radon exposure was addressed on biophysical grounds by Brenner [33], who concluded that, when cumulative exposures are sufﬁciently low that multiple traversals of target cells by alpha particles are rare (as it is the case of typical domestic radon exposures), all exposure rate enhancement effect disappear. The results of experiments conducted at Columbia University, using a micro-beam source [34], showed that traversal of cell nuclei by a single alpha particle induced signiﬁcantly lower oncogenic transformation in the C3H10T1/2 mouse ﬁbroblast system than in a Poisson distributed mean of one alpha particle. These ﬁndings suggest that cells traversed by multiple alpha particles contribute most to the risk. In this respect, based on dose-rate effect considerations, extrapolation of miner data with lower exposure rates to residential exposures where no target cells are traversed by more than a single alpha particle, may overestimate risks associated with typical residential exposures and exposure rates. Our recent data in rats support the

74

G. Monchaux

hypothesis that, at low doses, the risk of lung cancer is governed by the rate at which the dose is delivered, and not by the total cumulative dose alone. Recent data following chronic exposures to high concentrations of natural uranium ore dust alone also indicate that malignant tumour risks are not proportional to dose, but directly proportional to dose rate [35]. These data are also consistent with that of underground uranium miners showing an inverse dose-rate effect at high cumulative exposures, but a diminution of this effect at cumulative exposures lower than 50 WLM (0.18 J h m−3 ) [14]. These data support both an inverse dose-rate effect at high cumulative exposures, as well as its diminution or disappearance at low cumulative exposures. These experimental data suggest that at cumulative exposures comparable to those that may occur in underground mines and in homes, and at PAEC one order of magnitude larger than those currently observed in mines and about two orders of magnitude larger than those in dwellings, the existence of a practical threshold for lung cancer induction cannot be excluded, although statistical and theoretical modelling of the risk data may suggest different interpretations. Apparent thresholds for the induction of lung cancer have also been observed following low doses of alpha radiation, both in experimental animals – in rats exposed to PuO2 by inhalation [36] and in beagle dogs injected with 226 Ra [37] – and in humans – in radium dial painters [38] and Thorotrast-exposed patients [39]. Furthermore, the existence of a threshold related with dose rate is not restricted to alpha emitters since it has also been observed after external irradiation of mice with beta rays [40,41], after oral administration of tritiated water in mice [42], in beagle dogs fed 90 Sr [43] and after whole body irradiation of mice by gamma rays [44]. Moreover, quantitative modelling of data from animal studies provided risk coefﬁcients that can be compared with similarly derived coefﬁcients from epidemiological data. Different applications of the two-step mutation models were tested in the two larger data sets of rats exposed to radon from both PNNL [45] and CEA-COGEMA [46], as well as in various miner studies. Parameters derived in a recent study compared fairly well with those derived from uranium miners, suggesting that the effects of radon on rats may be used as an approximation for the effect on humans [47]. The objective of an European project in progress, is to perform a risk analysis of animal data, to compare the different sets of recent animal data on exposure rate with those of previously analysed data, to scale biological parameters and risks between rats an humans and to achieve statistical modelling of the underlying mechanisms.

Acknowledgements This work was supported in part by Grants FI4P-CT95-0025 (Commission of the European Communities) and COGEMA (PIC D11).

References [1] J.H. Lubin, J.D. Boice Jr., C. Edling, R.W. Hornung, G. Howe, E. Kunz, R.A. Kusiak, H.I. Morrison, E.P. Radford, J.M. Samet, M. Tirmarche, A. Woodward, Y.S. Xiang and D.A. Pierce, Radon and lung cancer risk: a joint analysis of 11 underground miners studies, US Department of Health and Human Services, Public Health Service, National Institutes of Health, NIH Publication No. 94-3644, Bethesda, MD, 1994.

Contribution of animal experimental data for the risk assessment of exposure to radon decay products

75

[2] J.H. Lubin, J.D. Boice Jr., C. Edling, R.W. Hornung, G. Howe, E. Kunz, R.A. Kusiak, H.I. Morrison, E.P. Radford, J.M. Samet, M. Tirmarche, A. Woodward, Y.S. Xiang, D.A. Pierce, J. Natl. Cancer Inst. 87 (1995) 817– 827. [3] National Research Council, Comparative Dosimetry of Radon in Mines and Homes, National Academy Press, Washington, DC, 1991. [4] The Householders’ Guide to Radon, 2nd ed., Department of the Environment, National Radiation Protection Board, Chilton, Didcot, UK, 1990. [5] B.L. Cohen, Health Phys. 68 (1995) 157–174. [6] S. Darby, E. Whitley, P. Silcocks, B. Thackrar, M. Green, P. Lomas, J. Miles, G. Reeves, T. Fearn, R. Doll, Br. J. Cancer 78 (1998) 394–408. [7] M.C.R. Alavanja, J.H. Lubin, J.A. Mahaffey, R.C. Brownson, Am. J. Public Health 89 (1999) 1042–1048. [8] G. Pershagen, G. Akerblom, O. Axelson, B. Clavensjo, L. Damber, G. Desai, A. Enﬂo, F. Lagarde, H. Mellander, M. Svartengren, G.A. Swedjemark, New Engl. J. Med. 330 (1994) 159–164. [9] E.G. Létourneau, D. Krewski, N.W. Choi, M.J. Goddard, R.G. McGregor, J.M. Zielinski, J. Du, Am. J. Epidemiol. 140 (1994) 310–322. [10] E. Ruosteenoja, Indoor radon and risk of lung cancer: an epidemiological study in Finland, STUK-A99, Finnish Center for Radiation and Nuclear Safety, Helsinki, 1991. [11] A. Auvinen, I. Mäkeläinen, M. Hakama, O. Castrén, E. Pukkala, H. Reisbacka, T. Rytömaa, J. Natl. Cancer Inst. 88 (1996) 966–972. [12] J.H. Lubin, J.D. Boice Jr., C. Edling, R.W. Hornung, G. Howe, E. Kunz, R.A. Kusiak, H.I. Morrison, E.P. Radford, J.M. Samet, M. Tirmarche, A. Woodward, Y.S. Xiang, D.A. Pierce, J. Natl. Cancer Inst. 87 (1995) 817– 827. [13] S. Darby, E. Whitley, G. Howe, S. Hutchings, R. A Kusiak, J.H. Lubin, H.I. Morrison, M. Tirmarche, L. Tomasek, E.D. Radford, R.J. Roscoe, J.H. Samet, Y.S. Xiang, J. Natl. Cancer Inst. 87 (1995) 374–384. [14] J.H. Lubin, J.D. Boice Jr., C. Edling, R.W. Hornung, G. Howe, E. Kunz, R.A. Kusiak, H.I. Morrison, E.P. Radford, J.M. Samet, M. Tirmarche, A. Woodward, Y.S. Xiang, Health Phys. 69 (1995) 494–500. [15] J.H. Lubin, J.D. Boice Jr., J. Natl. Cancer Inst. 89 (1997) 49–57. [16] F.T. Cross, Radon inhalation studies in animals, Report DOE/ER-0396, National Technical Information Service, Springﬁeld, VA, 1988. [17] C.G. Collier, J.C. Strong, S.T. Baker, T. Eldred, J.A.H. Humphreys, L.M. Cobb, Radiat. Res. 152 (1999) S141– S144. [18] G. Monchaux, J.P. Morlier, M. Morin, J. Chameaud, J. Lafuma, R. Masse, Environ. Health Perspect. 102 (1994) 64–73. [19] J.L. Poncy, P. Laroque, P. Fritsch, G. Monchaux, J. Chameaud, R. Masse, An experimental two-stage rat model of lung carcinoma initiated by radon exposure, in: F.T. Cross (Ed.), Indoor Radon and Lung Cancer: Reality or Myth? Part 2, Proceedings of the 29th Hanford Symposium on Health and the Environment, Battelle Press, Columbus, OH, 1992, pp. 803–819. [20] E. Douriez, P. Kermanac’h, P. Fritsch, M. Bisson, J.P. Morlier, G. Monchaux, M. Morin, P. Laurent, Radiat. Prot. Dosim. 56 (1994) 105–108. [21] G. Monchaux, J.P. Morlier, P. Rochefort, P. Fritsch, E. Douriez, M. Morin, R. Maximilien, Environ. Int. 22 (Suppl. 1) (1996) S917–S925. [22] F.T. Cross, R.F. Palmer, G.E. Dagle, R.H. Busch, R.L. Bushbom, Radiat. Prot. Dosim. 7 (1984) 381–384. [23] F.F. Hahn, G.A. Boorman, Neoplasia and preneoplasia of the lung, in: P. Bannash, W. Gössner (Eds.), Pathology of Neoplasia and Preneoplasia in Rodents, EULEP Color Atlas, vol. 2, Schattauer, Stuttgart, 1997, pp. 29–42. [24] R. Masse, Histogenesis of lung tumors induced in rats by inhalation of alpha emitters: An overview, in: Pulmonary Toxicology of Respirable Particles, Proceedings of the 19th Hanford Life Sciences Symposium, CONF791002, National Technical Information Service, Springﬁeld, VA, 1980, pp. 498–521. [25] G.E. Dagle, F.T. Cross, R.A. Gies, Morphology of respiratory tract lesions in rats exposed to radon/radon progeny, in: F.T. Cross (Ed.), Indoor Radon and Lung Cancer: Reality or Myth? Part 2, Proceedings of the 29th Hanford Symposium on Health and the Environment, Battelle Press, Columbus, OH, 1992, pp. 659–676. [26] F.T. Cross, A review of experimental radon health effects, in: W.C. Dewey, M. Edington, R.J.M. Fry, E.J. Hall, G.F. Whithmore (Eds.), Radiation Research: A Twentieth-Century Perspective, Proceedings of the 9th International Congress of Radiation Research, vol. 2, Academic Press, San Diego, 1992, pp. 476–781. [27] J. Chameaud, R. Masse, J. Lafuma, Radiat. Prot. Dosim. 7 (1984) 385–388.

76

G. Monchaux

[28] E.S. Gilbert, F.T. Cross, G.E. Dagle, Radiat Res. 145 (1996) 350–360. [29] R. Masse, J.P. Morlier, M. Morin, J. Chameaud, J. Lafuma, Radiat. Prot. Dosim. 45 (1992) 603–610. [30] J.P. Morlier, M. Morin, J. Chameaud, R. Masse, S. Bottard, J. Lafuma, C. R. Acad. Sci. Sér. III 315 (1992) 463–466. [31] J.P. Morlier, M. Morin, G. Monchaux, J.F. Pineau, J. Chameaud, J. Lafuma, R. Masse, Radiat. Prot. Dosim. 56 (1994) 93–97. [32] G. Monchaux, J.P. Morlier, J. Radiol. Protect. (2002), in press. [33] D.J. Brenner, Health Phys. 67 (1994) 76–79. [34] R.C. Miller, G. Randers-Pherson, C.R. Geard, E.J. Hall, D.J. Brenner, Proc. Natl. Acad. Sci. USA 96 (1999) 19–22. [35] R.E. Mitchell, J.S. Jackson, B. Heinmiller, Health Phys. 76 (1999) 145–155. [36] C.L. Sanders, K.E. Lauhala, K.E. McDonald, Int. J. Radiat. Biol. 64 (1993) 417–430. [37] O.G. Raabe, S.A. Book, N.J. Parks, C.E. Chrisp, M. Goldman, Radiat. Res. 86 (1989) 515–528. [38] R.E. Rowland, Radium in humans, in: A Review of US Studies, Argonne National Laboratory, 9700 South Cass Avenue, Argonne, IL 60439, September 1994, pp. 106–112. [39] M. Andersson, H.H. Storm, J. Natl. Cancer Inst. 84 (1992) 1318–1325. [40] A. Ootsuyama, H. Tanooka, Radiat. Res. 125 (1991) 98–101. [41] H. Tanooka, Int. J. Radiat. Biol. 77 (2001) 541–551. [42] O. Yamamoto, T. Seyama, H. Itoh, N. Fujimoto, Int. J. Radiat. Biol. 73 (1998) 535–541. [43] R.G. White, O.G. Raabe, M.R. Culbertson, N.J. Parks, S.J. Samuels, L.S. Rosenblatt, Radiat. Res. 136 (1993) 178–189. [44] J.R. Maisin, A. Wambersie, G.B. Gerber, G. Mattelin, M. Lambiet-Collier, J. Gueulette, Radiat. Res. 94 (1983) 359–373. [45] W.F. Heidenreich, P. Jacob, H.G. Paretzke, F.T. Cross, G.E. Dagle, Radiat Res. 151 (1999) 209–217. [46] W.F. Heidenreich, M.J.P. Brugmans, P.P. Little, H.P. Leenhouts, H.G. Paretzke, M. Morin, J. Lafuma, Radiat. Environ. Biophys. 39 (2000) 253–264. [47] H. Bijwaard, M.J.P. Brugmans, H.P. Leenhouts, Radiat. Environ. Biophys. 40 (2001) 269–277.

77

On the exposure circumstances and some further risk estimates regarding leukemia in ages 0–19 years and exposure to ionizing radiation in homes of uranium-containing alum shale-based concrete G. Åkerblom a , L. Hardell b , M. Fredrikson c , O. Axelson c a Swedish Radiation Protection Authority, SE 171 16 Stockholm, Sweden b Department of Oncology, Örebro Medical Centre, SE 701 85 Örebro, Sweden c Division of Occupational and Environmental Medicine, Department of Health and Environment,

Linköping University, SE 581 85 Linköping, Sweden

During the years 1929–1978, lightweight concrete made of uranium-rich black alum shale was used as a building material all over Sweden. About 400 000 dwellings (i.e., 10% of the dwellings) have walls and often also ﬂoors and ceilings of this material. To evaluate any possible risk of acute lymphocytic leukemia (ALL) from exposure to elevated gamma-radiation in homes built from uranium-containing alum shale concrete in Sweden, we identiﬁed the ALL cases aged less than 20 years during 1980–1989 and applied a case–control analysis that involved four out of eight originally drawn controls per case from the population registry, matching on age, gender and county. Measurements of gamma-radiation from the facades of houses were available as a result of the concerns about indoor radon in Sweden around 1980 as such houses might have had high radon concentrations. Using the existing measurements, exposure was assessable for 312 cases and 1418 controls from 151 municipalities covered by measurements. As reported elsewhere we found an elevated risk represented by an odds ratio of 1.4 (95% conﬁdence interval 1.0–1.9) for those always living in alum shale concrete houses, with the average exposure exceeding 0.10 μSv h−1 . Also a weak dose–response relation was noted. Further calculations presented here indicate a crude incidence rate of 4.43 cases per 100 000 person–years for the exposed versus 3.17 for unexposed subjects, i.e., the excess incidence was 1.26 cases per 100 000 person–years. Taken together with earlier Swedish studies on adults our ﬁndings suggest a likely effect regarding leukemia due to low-level exposure to indoor gamma-radiation in alum shale concrete houses. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07008-1

© 2005 Elsevier Ltd. All rights reserved.

78

G. Åkerblom et al.

1. Introduction During the years 1929–1978, lightweight concrete made of uranium-rich black alum shale was used as a building material in Sweden. About 400 000 dwellings have walls and often also ﬂoors and ceilings of this material. These dwellings, both ﬂats and detached houses, represent about 10% of the Swedish housing stock and are scattered all over Sweden. About 850 000 out of the 8.9 million inhabitants live in these dwellings, and all are exposed both to excess gamma radiation and higher than average radon concentrations. About 215 000 of the people living in alum shale-based concrete dwellings are in the ages of 0–19 years. The people living in the alum shale-based concrete buildings in Sweden probably constitute one of the largest populations that are exposed to enhanced natural radiation. We have investigated the possibility of an increased risk of acute lymphocytic leukemia (ALL) among the children living in these the alum shale-based concrete buildings [1]. Our ﬁndings corroborate earlier observations of an increased risk among adults as associated with living and working in houses of stony materials, i.e., not only those built from alum shale concrete but also from other concrete as also causing some degree of gamma radiation in excess of what is at hand in wooden houses [2–4]. Children and young adults were not included in these earlier studies and another main reason for now focusing on this cancer type among younger individuals was the indication in the earlier studies that the risk might be higher in the younger age groups. We therefore found it interesting to go further down in ages in order to obtain conﬁrmation or refutation for the possibility of a cancer risk from living in these houses. From a technical point of view, there was the likely beneﬁt of short latency time for leukemia among children and also that most children had lived only in one or two dwellings. Furthermore, acute lymphocytic leukemia (ALL) is the most common malignant disease in childhood. The only generally accepted etiologic factors are limited to ionizing radiation in utero and rare genetic disorders [5]. As an increased risk was indicated in our study of the children and young adults, we here elaborate some further aspects of the risk estimation. As an introductory note, we may also mention that in Italy, the risk of myeloid leukemia was increased in areas with volcanic versus sedimentary geology, implying a possible effect of indoor radon [6]. Such a relation was not conﬁrmed by later measurements of radon and gamma radiation, but the association with volcanic geology remained [7]. Studies in United States and United Kingdom also failed to ﬁnd any connection between indoor radon and childhood ALL or acute myeloid leukemia [8–10] However, geo-medical data from France indicate a possible role of background radiation for the development of leukemia [11] along with an effect from residence in granite-built houses [12]. The role of indoor gamma-radiation, rather than radon, as a risk factor clearly seems worth considering as both Swedish and Italian case–control studies on adult acute myeloid leukemia (AML) have indicated relationships to building materials with increased radioactivity [2–4,13]. 1.1. Alum shale-based lightweight concrete The alum shale-based concrete is used both as wall blocks and wall and ﬂoor elements [14] (Fig. 1). Alum shale is a uranium-rich black shale of late Cambrian age (500 million BC). The concentration of uranium in the shale is 50–400 ppm U (600–5000 Bq U-238/kg), thorium 8–15 ppm Th (30–60 Bq Th-232/kg) and potassium 3.5–6% (1100–1900 Bq K-40/kg). At

On the exposure circumstances in homes of uranium-containing alum shale-based concrete

79

Fig. 1. Blocks of alum shale-based lightweight concrete.

the production of the lightweight concrete shale, ash was mixed with slaked lime and cast to blocks. Because concerns about the radioactivity of this building material, the production of the alum shale-based lightweight concrete was stopped in 1975. The radium activity concentration of the lightweight concrete is 600–2600 Bq Ra-226/kg. The gamma radiation in rooms built of the lightweight concrete depends mainly on the radium concentration of the material and on how large parts of the walls that consist of the material. Usually the indoor gamma radiation is about 0.30–0.40 microsievert per hour (μSv h−1 ) in alum shale concrete houses, rarely lower than 0.20 μSv h−1 . Inside some 30 000 of these dwellings, the gamma-radiation is about 0.50 μSv h−1 or higher and may amount to even 1.20 μSv h−1 . At a constant exposure to a natural gamma radiation of 0.20 μSv h−1 the received dose to adults is 1 millisievert per year (mSv y−1 ). Infants and children are supposed to be 30%, respectively 10%, more sensitive to gamma radiation than adults [15]. Thus the excess doses to children living in alum shale-based dwellings are 1.5 to 6 mSv y−1 supposing 80% occupancy at home. The indoor gamma radiation in buildings that not contain alum shale-based lightweight concrete is normally 0.05–0.10 μSv h−1 (wooden houses) and 0.05–0.20 μSv h−1 (houses of bricks, concrete or stone); 0.10 μSv h−1 might be taken as a representative average radiation level for these constructions [16]. When the ﬁrst concerns about adverse health effects from indoor radon arose in the late 1970s and early 1980s, the Swedish Radon Commission initiated a comprehensive measurement program. At that time it was known that increased radon concentrations could occur in houses of alum shale-based concrete [17]. However, it was not known where these buildings were located. Therefore the Radon Commission 1979 initiated a nation-wide survey to map and register the locations of these houses. In the survey cars equipped by large NaI(Tl) scintillator crystals drove along the streets and roads and measured the gamma radiation from the facades (Fig. 2). In this way it was possible to detect a building with alum shale-based lightweight concrete at a distance of more than 25 m. To check that a detected house really was an alum shale-based lightweight concrete house, and to have information on the radioactivity of the material, measurements with a hand-hold gamma meter were carried out against the outer walls. The gamma radiation value for each house was recorded. Some houses already known to have been built from alum shale concrete were not measured, but were indicated in the map-based register. Car-borne mapping of the alum shale-based lightweight concrete buildings were made in the majority of the 286 Swedish municipalities. However, in many of

80

G. Åkerblom et al.

Fig. 2. Carborne survey for houses built of alum shale-based lightweight concrete.

the municipalities only parts of the municipality were surveyed. For 151 municipalities, the surveys covered the whole municipality. The precision in the surveys was better than 95%. At the time of the car-borne measurements, there was still little attention to the importance of radon leaking from the ground into the houses, which explains the speciﬁc interest in the alum shale concrete houses. Whereas indoor radon levels are variable with ventilation and can be reduced through various measures, the gamma-radiation from building materials remains unchanged over time.

2. Materials and methods 2.1. Subjects and their exposure In the study we identiﬁed cases of ALL aged less than 20 years and diagnosed in the period 1980–1989 through the nation-wide Swedish Cancer Registry, which is based on compulsory notiﬁcations of cancer diagnoses from all physicians. This registry includes a speciﬁc individual identiﬁcation number, which directly shows sex and date of birth and permits that the individual’s place of living can be found. In total there were 618 cases and as a precautionary measure we originally drew eight times as many controls from the registry of the Swedish population to have although only four of them were used by individual matching on year of birth, gender and county (usually with populations about 250 000–350 000). Due to incomplete measurements in several municipalities, only a limited number of our cases and their matched controls could be used, i.e., 312 cases and 1418 controls. The addresses and also the dates of moving could be ascertained for the cases and controls as records about earlier residency even after the time a person has moved. Hence, by checking against the map-based gamma measurement ﬁles, we could ﬁnd out whether the subjects had lived in an alum shale concrete house or not. Further details regarding exposure assessment are available in the original report of this study [1]. It may be added here that in the conditional logistic analysis another 86 subjects, representing 24 case–control sets, were lost. Considering exposure to gamma radiation, we ﬁrst noted whether subjects were ever registered on the address of a lightweight alum shale concrete building. Then we calculated an intensity-time product in millisievert (mSv) by multiplying the measured radiation (in μSv h−1 ) from the facade by the residency time until diagnosis for cases and the corresponding point in time for the matched controls. We assumed a somewhat arbitrary indoor occupancy

On the exposure circumstances in homes of uranium-containing alum shale-based concrete

81

Table 1 Background numbers and principle for calculation of risk estimates based on the formula I = RI0 F + I0 (1 − F ), where I is the overall incidence, R the relative risk associated with the exposure, F the fraction of persons at exposure, and I0 the incidence among the unexposed [20] Population in ages 0–19 years

Annual cases of leukemia cases

Derived incidence rate per 100 000

2 068 444

∼ = 70

3.38

F in formula obtained from the number of controls in Table 2 as 239/(239 + 1179) = 0.169; hence 3.38 = 1.4I0 × 0.169 + I0 (1 − 0.169); I0 = 3.17 and for the exposed RI0 = 4.44. The excess incidence is then 4.44 − 3.17 = 1.27.

time of 20 hours as probably slightly too high for teenagers and too low for the youngest children, but on the other hand, we applied no adjustment for shielding effects of the houses against outdoor radiation from the ground and of cosmic origin. As only addresses of houses with elevated gamma-radiation were registered and available, we used an exposure level of 0.10 μSv h−1 for unidentiﬁed houses, i.e., as not measured or otherwise were known not to be built from alum shale concrete. For those unmeasured houses that were known to be built of alum shale concrete, we used 0.20 μSv h−1 as representative but perhaps a somewhat low exposure estimate. The possibility of a dose–response relationship was considered by dividing those who always lived in alum shale concrete houses into two equal groups (differing in size by 2 subjects due to equal exposure) by their annual average exposure. Subjects who never lived in such houses constituted the reference category. 2.2. Statistical methods and analyses We applied conditional regression analysis of the matched sets, that is, those sets in which the case and at least one control remained. We also carried out stratiﬁed analyses, using EpiInfo [18], including the 86 subjects from the 24 lost sets but for whom we had exposure data. An etiologic fraction for the total population (or population attributable risk) was calculated to obtain some indication of the possible contribution of risk from the exposure studied. Here we now add some calculations concerning the number of cases that are likely to occur in the total of Sweden under the assumption that the relationship between alum shale houses and other houses is about the same also for those municipalities not included in the study due to incomplete measurements. As a base for these calculations, we also used the average age speciﬁc leukemia incidence in Sweden as obtainable from the annual publications of the national Cancer Registry. The methods of some calculations also appear in Table 1.

3. Results The 312 cases and 1418 controls included in the study are displayed in Table 2 regarding age and whether ever being registered at a lightweight alum shale concrete house address. Overall

82

G. Åkerblom et al.

Table 2 Odds ratios (OR) and 95% conﬁdence interval (CI) for acute lymphocytic leukemia by residence in an alum shalebased concrete (ASC) house Age group (years) 0–4.5 4.5–9.5 9.5–14.5 14.5–20 All ages

Cases Controls Cases Controls Cases Controls Cases Controls Cases Controls

Crude OR Mantel–Haenszel OR Conditional logistic OR∗ 95% conf. interval CI Crude ORSUBSET ∗ P -value for trend is 0.028

Never lived in an ASC house

Always lived in an ASC house Lower exposure Higher exposure

All exposed

119 566 77 379 34 168 13 66 243 1179

0 3 10 47 14 43 8 28 32 121

27 89 18 73 16 48 8 29 69 239

27 86 8 26 2 5 0 1 37 118

1.3

1.5

1.3 0.84–2.1 1.3

1.5 0.98–2.3 1.5

1.4 1.4 1.4 1.4

The table obtained by merging data from the original publication [1]. Those who always lived in an alum shale concrete house are subdivided in lower and higher degree of average annual exposure (cumulated exposure including background radiation divided by age). ∗ Leaving out 24 incomplete case–control sets – see text.

an estimated exposure > 0.10 μSv h−1 resulted in an odds ratio (OR) of 1.4. As can be seen by comparing the different risk estimates given in the table, it can be concluded that the inﬂuence from the matching factors in terms of confounding is negligible. The etiologic fraction for the population corresponding to the OR of 1.4 is 0.063, implying that 6.3% (erroneously reported as 6.8 in the original publication) of ALL in the ages considered would be attributable to living for some time in an alum shale concrete house if the association were causal. As there are some 70 cases of ALL each year in the ages considered here, the contribution of the exposure in the alum shale concrete houses would be about 4.4 cases (Table 1). The population in the ages considered in the study was 2 068 444 [19], an excess incidence of 1.26 cases per 100 000 person–years can be calculated. The total incidence for the exposed would be 4.43 cases per 100 000 person–years as compared to a background rate of 3.17 cases per 100 000 person–years for in these ages (i.e., the rate ratio is 1.4). Based on the distribution of controls with and without experience of living in an alum shale house, 16.8 and 83.1%, respectively, about 347,400 could be expected to have had experience of such living. The difference between the estimated fraction of alum shale concrete houses in Sweden, 10%, and the 16.8% calculated here indicates the mobility of the population with children and young adults.

On the exposure circumstances in homes of uranium-containing alum shale-based concrete

83

4. Discussion Our study suggests an effect of indoor ionizing radiation regarding ALL, thereby corroborating some of the earlier observations reported in the literature. There is always some possibility of bias in epidemiologic studies that could explain the result obtained, for example, selection phenomena, poor quality of the assessment of exposure, and confounding. As the data were derived from the national cancer registry covering the whole of Sweden, any serious selection bias appears as unlikely. The other important and reliable source of information was the somewhat unique population registries with information on moving. The facade measurements may not have discovered all houses built from alum shale concrete, however. If so, this would lead to some non-differential misclassiﬁcation of exposure, but the result would then be an underestimation rather than an exaggeration of the effect. A falsely increased risk could in principle result, however, as non-differential misclassiﬁcation is a chance phenomenon [20,21]. Population density and living in relation to proximity to main roads and petrol stations could perhaps cause some confounding [22–24], but these risk factors are relatively weak and should hardly correlate more strongly with living in alum shale concrete houses. Socio-economic factors are less likely to be seriously inﬂuencing the results as the alum shale concrete has been used to the same extent both for detached and apartment houses. Should radon play a role for the development of ALL, the correlation between indoor radon levels and gamma radiation makes confounding from radon exposure unavoidable in lack of data on radon in our study, but there is no clear indication of a role of radon regarding ALL [5–7,25]. Interestingly too, an Italian study showed a stronger correlation with urinary excretion of 8-hydroxydeoxyguanosine, a marker of oxidative DNA damage, for indoor gamma radiation than for radon [26]. This investigation taken together with the studies referred to in the introduction, suggests a likely effect regarding leukemia due to low exposure to indoor gamma radiation in alum shale concrete houses. References [1] O. Axelson, M. Fredrikson, G. Åkerblom, L. Hardell, Leukemia in childhood and adolescence and exposure to ionizing radiation in homes built from uranium-containing alum shale concrete, Epidemiology 13 (2002) 146–150. [2] U. Flodin, L. Andersson, C.-G. Anjou, U. Palm, O. Vikrot, O. Axelson, A case–referent study on acute myeloid leukemia, background radiation and exposure to solvents and other agents, Scand. J. Work Environ. Health 7 (1981) 169–178. [3] U. Flodin, M. Fredriksson, O. Axelson, B. Persson, L. Hardell, Background radiation, electrical work, and some other exposures associated with acute myeloid leukaemia in a case–referent study, Arch. Environ. Health 41 (1986) 77–83. [4] U. Flodin, M. Fredriksson, B. Persson, O. Axelson, Acute myeloid leukemia and background radiation in an expanded case–referent study, Arch. Environ. Health 45 (1990) 364–366. [5] D.G. Poplac, Acute lymphoblastic leukemia, in: P.A. Pizzo, D.G. Poplack (Eds.), Principles and Practice of Pediatric Oncology, J.B. Lippincott Company, Philadelphia, 1993, pp. 431–481. [6] F. Forastiere, A. Quiercia, F. Cavariani, M. Miceli, C.A. Perucci, O. Axelson, Cancer risk and radon exposure, Lancet 339 (1992) 1115.

84

G. Åkerblom et al.

[7] F. Forastiere, A. Sperati, G. Cherubini, M. Miceli, A. Biggeri, O. Axelson, Adult myeloid leukaemia, geology, and domestic exposure to radon and gamma-radiation: a case control study in central Italy, Occup. Environ. Med. 55 (1998) 106–110. [8] J.H. Lubin, M.S. Linet, J.D. Boice Jr., et al., Case-control study of childhood acute lymphoblastic leukemia and residential radon exposure, J. Natl. Cancer Inst. 90 (1998) 294–300. [9] M. Steinbuch, C.R. Weinberg, J.D. Buckley, L.L. Robison, D.P. Sandler, Indoor residential radon exposure and risk of childhood acute myeloid leukaemia, Br. J. Cancer 81 (1999) 900–906. [10] G.R. Law, E.V. Kane, E. Roman, A. Smith, R. Cartwright, Residential radon exposure and adult acute leukaemia, Lancet 355 (2000) 1888. [11] J.F. Viel, Radon exposure and leukaemia in adulthood, Int. J. Epidemiol. 22 (1993) 627–631. [12] D. Pobel, J.-F. Viel, Case-control study of leukaemia among young people near La Haugue nuclear reprocessing plant: the environmental hypothesis revisited, Br. Med. J. 314 (1997) 101–106. [13] A. Mele, M. Szklo, G. Visani, et al., Hair-dye use and other risk factor for leukemia and pre-leukemia: a case– control study, Am. J. Epidemiol. 139 (1994) 609–619. [14] G. Åkerblom, “Blue concrete” – a source of radon and gamma-radiation, in: SSI News, vol. 1, Swedish Radiation Protection Institute, Stockholm, 1995, pp. 4–5. [15] UNSCEAR (United Nations Scientiﬁc Committee on the Effects of Atomic Radiation), Sources and Effects of Ionising Radiation, UNSCEAR 2000 Report to the General Assembly, with Scientiﬁc Annexes, vol. I: Sources, United Nations, New York, 2000. [16] L. Mjönes, Gamma-radiation in Swedish dwellings, Radiat. Prot. Dosim. 15 (1986) 131–140. [17] B. Hultqvist, Studies on naturally occurring ionising radiation, with special reference to radiation doses in Swedish houses of various types, Kung. Svenska Vetenskap. Handl. (4) 6 (3) (1956). [18] A.G. Dean, J.A. Dean, D. Coulombier, et al., Epi-Info, Version 6: a Word Processing Data Base, and Statistics Program for Epidemiology on Microcomputers, Centers for Disease Control and Prevention, Atlanta, GA, 1994. [19] Statistical Abstract of Sweden 1990, in: Ofﬁcial Statistics of Sweden, vol. 76, 1990. [20] O. Axelson, Aspects on confounding in occupational health epidemiology, Scand. J. Work Environ. Health 4 (1978) 85–89. [21] T. Sohrahan, M.S. Gilthorpe, Non-differential misclassiﬁcation of exposure always leads to an underestimate of risk, an incorrect conclusion, Occup. Environ. Med. 51 (1994) 839–840. [22] F.E. Alexander, P. Boyle, P.M. Carli, et al., Population density and childhood leukaemia: results of the EUROCLUS Study, Eur. J. Cancer 35 (1999) 439–444. [23] R.M. Harrison, P.L. Leung, L. Somervaille, R. Smith, E. Gilman, Analysis of incidence of childhood cancer in the West Midlands of the United Kingdom in relation to proximity to main roads and petrol stations, Occup. Environ. Med. 56 (1999) 774–780. [24] U. Hjalmars, G. Gustafsson, Swedish Child Leukaemia Group, Higher risk for acute childhood lymphoblastic leukaemia in Swedish population centres 1973–1994, Br. J. Cancer 79 (1999) 30–33. [25] Committee on Health Risks of Exposure to Radon, Board on Radiation Effects Research, Commission on Life Sciences, National Research Council, Health Effects of Exposure to Indoor Radon, National Academy Press, Washington, DC, 1999. [26] A. Sperati, D.D. Abeni, C. Tagesson, F. Forastiere, M. Miceli, O. Axelson, Exposure to indoor background radiation and urinary concentrations of 8-hydroxydeoxyguanosine, a marker of oxidative DNA damage, Environ. Health Perspect. 107 (1999) 213–215.

85

The implementation in national legislation of Title VII of the Council Directive 96/29/Euratom: Some general remarks and the case of Italy S. Risica, F. Bochicchio, C. Nuccetelli Instituto Superiore di Sanità (National Institute of Health), Physics Laboratory, Viale Regina Elena 299, Rome, Italy

This paper analyses and discusses how the Italian legislation implementing Title VII of the Council Directive 96/29/Euratom deals with natural radioactivity of terrestrial origin in work environments, that is, both radon in workplaces and work activities with naturally occurring radioactive materials. Particular attention is given to the chosen values for Action Levels, the time frames for obligations – that is measurements, eventual application of corrective measures and, if necessary, application of the radiation protection system for practices – and their underlying rationale. Some qualifying choices of the legislation are highlighted and discussed. Possible difﬁculties in enforcing it are also analysed and some ﬁnal remarks are given to promote debate among experts.

1. Introduction The Council Directive 96/29/Euratom [1] – issued in May 1996 with the aim of renewing the basic safety standards of radiation protection in the EU – devoted its Title VII to work activities “. . . within which the presence of natural radiation sources leads to a signiﬁcant increase in the exposure of workers or of members of the public which cannot be disregarded from the radiation protection point of view.” In order to assist Member States in implementing this Title VII in their national legislation, the two technical guidelines Radiation Protection 88, Recommendations for the implementation of Title VII of the European Basic Safety Standards Directive (BSS) concerning signiﬁcant increase in exposure due to natural radiation sources [2] and Radiation Protection 95, Reference levels for workplaces processing materials with enhanced levels of naturally occurring radionuclides [3] were published by the EU Commission in 1997 and 1999, respectively. One of the qualifying points of these documents is the suggestion of ranges of Action Levels for the European Union, both for radon in the workplace and for effective doses in work RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07009-3

© 2005 Elsevier Ltd. All rights reserved.

86

S. Risica et al.

Table 1 Suggested Action Levels for workers in NORM industries [3] Effective dose (mSv y−1 )

Scenario

Normal Unlikely

<1

1–6

6–20

no regulation necessary

lower level of regulation

higher level of regulation

no regulation necessary

lower level of regulation

> 20

20–50 process not permitted∗

higher level of regulation

process not permitted∗

∗ This means that a thorough review of the working practices is necessary to make a detailed assessment of doses.

If these prove to exceed the worker dose limit as set out in the Directive 96/29/Euratom, then the process must cease.

activities with naturally occurring radioactive materials (NORM). In issuing its Directive, the EU Commission did not set limit values, most probably because it was felt that this initial attempt to limit natural radioactivity at work would come up against signiﬁcantly different levels of applicability and, therefore, feasibility, in different Member States. For this reason, Member States were left free to adopt different national approaches to the issue. In the ﬁrst technical guideline [2] for the working environment, the suggested Action Level of time averaged radon activity concentration is in the range of 500–1000 Bq m−3 : using ICRP 65 convention coefﬁcients [4] and an occupancy factor of 2000 h y−1 , this means 3–6 mSv y−1 . For work activities with NORM, the suggested Action Levels for workers are shown in Table 1. It is interesting to note that these Action Levels start out from the premise that 1 mSv y−1 is the dose criterion which, as regards practices, distinguishes the population from workers. Moreover, they are based on the belief that a worker exposed to natural radioactivity for his working conditions should not be exposed to a risk comparable to category A workers in practices, without any attempt to reduce this exposure. In Italy, the new Radiation Protection Act, implementing the 96/29/Euratom Directive, was issued in August 2000 [5] as a legislative decree. The decree was published as additions and amendments to the previous Radiation Protection Act [6], which had been issued when the 96/29/Euratom Directive was in an advanced form and therefore had already included many novelties (e.g. new dose limits). The draft of the 2000 decree, to be approved by the Council of Ministers, was prepared in a series of ofﬁcial meetings involving representatives of various ministries and radiation protection experts from institutes and agencies involved. In particular, the implementation of Title VII of the Directive required a long debate due to the novelty of the issue. For this reason, some informal meetings were organised among experts beforehand so that an agreement could be reached on the main decisions to be taken and then submitted to the ofﬁcial meetings. The authors of this paper attended, as experts, both types of meetings on natural radioactivity at work. 2. The decree implementing the Euratom directive in Italy In the new Radiation Protection Act, Title VII of the Directive was implemented dividing work activities in which the presence of natural radiation sources leads to a signiﬁcant increase

The implementation of Directive 96/29/Euratom in Italy

87

of the exposure of workers or members of the public which cannot be disregarded from a radiation protection point of view1 into six groups: (a) all work activities performed in underground workplaces (including tunnels, catacombs, caves, etc.) where workers and, if it is the case, members of the public are exposed to radon or thoron decay products or gamma radiation or any other exposure; (b) work activities in work environments not considered in (a) in well identiﬁed areas or in buildings with well deﬁned characteristics, where workers and, if it is the case, members of the public are exposed to radon or thoron decay products or gamma radiation or any other exposure; (c) work activities involving operations with, and storage of, materials not generally regarded as radioactive, but which contain natural radionuclides and cause a signiﬁcant increase in the exposure of workers and, if it is the case, members of the public; (d) work activities which produce residues not generally regarded as radioactive, but which contain natural radionuclides and cause a signiﬁcant increase in the exposure of members of the public and, if it is the case, workers; (e) work activities in thermal spas or mining activities not regulated in [omissis2 ]; (f) work activities in aircraft, as regards aircraft personnel. In this paper all work activities connected with terrestrial exposures are analysed and discussed, that is all activities listed, with the exception of (f). It can be noted that, in comparison with the EURATOM Directive [1], the Italian decree [5] makes some signiﬁcant distinctions concerning exposure to radon. Indeed, all work activities included in Article 40.2a of the Directive are divided into three different groups (a, b and e), which require quite different obligations. 2.1. Work activities in underground environments (a) For work activities in underground environments, the obligations for the employers are the following. (i) The mean annual value of radon activity concentration in air should be measured in all the workplaces, to verify if it exceeds the Action Level (see Section 2.5); measurements should be made following technical guidelines suggested by the National Technical Commission (see Section 2.6). (ii) If the radon activity concentration in air is lower than the Action Level, but higher than 80% of its value, the annual measurement should be repeated. (iii) If the radon concentration in air is higher than the Action Level, the employer should submit a report with the measurement results to the local authorities and take some remedial action to reduce the concentration signiﬁcantly below the Action Level. A new annual measurement is also required to verify the effectiveness of the remedial action undertaken. 1 The italics used in this paragraph indicate a non-ofﬁcial, literal translation of the law. 2 Mining activities recognised, according to certain criteria, as exposing workers to ionising radiation were already

regulated by the previous Radiation Protection Acts.

88

S. Risica et al.

(iv) If all the remedial action undertaken is not effective enough to reduce radon activity concentration to signiﬁcantly below the Action Level, the workplace should be subjected to the usual radiation protection control system for practices, where applicable. Enforcement of these obligations starts eighteen months after publication of the legislation in the Ofﬁcial Journal. Concerning the time scheduled for the obligations: the ﬁrst action should be completed within two years from the beginning of the work activity; the second the following year, whereas the third and, if necessary, the fourth, must be carried out within 3 years after the submission of the report, with an urgency correlated to the degree to which concentrations were in excess of the Action Level. The obligations for the work activities in underground environments are enforced over the whole Italian territory. This is due to the fact that soil is generally the main radon source and the number of underground workplaces was considered manageable, that is they are not too many. 2.2. Work activities in well identiﬁed areas or with well deﬁned characteristics (b) Nearly the same obligations provided for the work activities of (a) hold for the work activities of (b), but with a signiﬁcant difference: not all work activities are subjected to obligations, but only those located in selected areas (radon-prone areas) or in selected buildings whose characteristics could produce high indoor radon concentration (radon-prone buildings). Such areas and building characteristics are to be identiﬁed by the authorities of the 21 Italian regions on the basis of guidelines and general criteria suggested by the National Technical Commission (see Section 2.6). The ﬁrst identiﬁcation should be made within 5 years from the publication of the decree in the Ofﬁcial Journal (31 August 2000) that is within 1 September 2005. The list of identiﬁed areas should be published in the Italian Ofﬁcial Journal and should be updated as soon as new data are available. It is worth noting that the deﬁnition of a radon-prone building is not included in either the EURATOM Directive [1] or the EU technical guideline [2]. Thus, the decree introduces it to provide an additional tool for identifying those workplaces with high radon concentration. In fact, use of the radon-prone area concept alone would probably lead to a lack of identiﬁcation of workplaces with high radon concentration outside such areas. Workplaces in buildings made of materials with high activity concentration of natural radionuclides, e.g. tuff, are a possible example. On the other hand, most of the buildings in radon prone-areas have a relatively low radon concentration, and the use of the concept of a radon-prone building could reduce the number of workplaces to be monitored. Moreover, for the same reason, the decree states that, in radon-prone buildings or areas, workplaces located in basements and ground ﬂoors should be monitored ﬁrst. 2.3. Work activities with naturally occurring radioactive material (c and d) The Radiation Protection 95 guide [3] asserts correctly that “These processes [work activities with NORM] provide a major challenge for radiation protection. The existence of the radiation risk is incidental to the process and the concentration of natural radionuclides can occur at one or more stages during it.”

The implementation of Directive 96/29/Euratom in Italy

89

The EURATOM Directive calls on Member States to identify on their territory those work activities with NORM that may be of concern. In Italy, the work activities with NORM to be subjected to regulation were identiﬁed during the expert discussion cited earlier. The choice was made considering both the scientiﬁc knowledge about possible exposures of workers to natural radioactivity in Italy (see, e.g., [7–9]) and some radiation protection problems that had arisen in the past (e.g. about gas mantles), as well as the advice in the guide mentioned [3]. An initial list of work activities subjected to regulation was annexed to the legislative decree in which Action Levels were also established to facilitate their possible updating. The activities identiﬁed are: • phosphate industry and fertiliser warehouse for wholesale trade; • ore processing for extraction of tin, iron-niobium from pyrochlore and aluminium from bauxite; • processing of zircon sand and production of refractory material; • processing of rare earth; • processing and use of thorium compounds, as regards welding electrodes, lenses and optical glasses, and gas mantles; • production of titanium dioxide pigment; • oil extraction and reﬁning, and gas extraction, as regards presence and removal of scales in tubes and containers. Enforcement of the obligations for these work activities should start 3 years after publication of the law (that is 1 September 2003). This was done to allow a gradual application of this part of the decree concerning natural radioactivity at work, and to give priority to radon at work, considered the more urgent matter. The obligations for the employers of work with NORM are the following: (i) To carry out a preliminary evaluation of the radiological situation of the work environment on the basis of experimental measurements within two years of the beginning of the work activity. These measurements should be made according to the suggestions and technical guidance provided by the National Technical Commission (see Section 2.6). (ii) If the effective doses to workers and the population do not exceed the Action Levels (see Section 2.5), a new evaluation every three years or after signiﬁcant variations in the productive cycle is the only remaining obligation for the employer. (iii) If Action Levels are exceeded, the employer must have the working process analysed to assess the doses to workers and the population, submit a report with the dose assessment to local authorities and adopt corrective measures within three years. (iv) If, notwithstanding the corrective measures, doses continue to exceed the Action Level, the radiation protection control system for practices, for both workers and the population, should be applied. (v) If doses to workers turn out to be lower than the Action Level but higher than 80% of its value, measurements should be repeated every year. The radon activity concentration in indoor air due to the workplace characteristics (that is, from the underlying soil or building material) is excluded from these dose levels because it is independent of the work activity involved and is regulated by the Action Level for radon at work (see Sections 2.1 and 2.2).

90

S. Risica et al.

2.4. Work activities in thermal spas or mining activities not already regulated (e) The Italian decree, unlike the EURATOM Directive, regulates exposure to radon in spas as a particular case. In fact, the presence of radon in spas was considered a direct consequence of the work activity. On this basis, the Action Level was not chosen in terms of radon concentration, but of effective dose (1 mSv y−1 , see Section 2.5), which also includes the effective dose due to radon exposure. In this way, the Action Level, equal to the one applied for NORM industries, is more restrictive than that used for radon concentration for work activities of (a) and (c). In fact, with the conventional conversion dose factor of ICRP 65 [4] adopted by the decree, the Action Level of 1 mSv y−1 corresponds to an annual average radon activity concentration below 200 Bq m−3 , if radon is the only natural source of ionising radiation. As already speciﬁed in Section 2.3, this choice differs from that made for work activities (c) and (d), in which case the Action Level for workers excludes the contribution due to radon from soil and building materials, because it is independent of the work activity. The choice of this Action Level is supported by the 1987 ICRP Statement made in Como, (Italy) “. . .with respect to persons exposed while at work in such facilities, the entire system of dose limitation should apply. This would require, in addition to justiﬁcation and optimisation, the application of the dose limit, and all other relevant recommendations given for workers in ICRP Publications 26 and 32.” This means that any worker exposed to doses higher than 1 mSv y−1 should be considered an “exposed worker”. What differs in the case of spas is that corrective measures, aimed at reducing doses, should be adopted before deciding to classify the worker. The obligations (measurements, evaluation of the radiological situation, eventual adoption of corrective measures and time scheduled for all of them) for these work activities are the same as those for work activities with NORM (see Section 2.3). 2.5. Action levels As already stated, all Action Levels are provided in one of the annexes to facilitate updating. The Action Level for work activities (a) and (b) is 500 Bq m−3 of annual average radon activity concentration. If this value is exceeded, remedial actions should be adopted, unless the relevant effective dose is lower than 3 mSv y−1 for all workers. On the basis of the adopted conventional conversion dose factor of ICRP Publication 65 [4] this occurs when the occupancy time is lower than 2000 h y−1 . However, for nurseries, schools, kindergartens, primary and secondary schools, the Action Level is set as 500 Bq m−3 , regardless of occupancy time. Action Levels for workers and the population for work activities with NORM (c and d) are stated in 1 mSv y−1 and 0.3 mSv y−1 , respectively. For thermal spas (and mining activities not already regulated) the Action Level is set at 1 mSv y−1 for workers, as already mentioned. It was decided to balance these Action Levels, which are the lower values of the ranges suggested (500 to 1000 Bq m−3 and 1 to 6 mSv y−1 ), with delayed time schedules for the obligations.

The implementation of Directive 96/29/Euratom in Italy

91

2.6. The National Technical Commission on Exposure to Natural Radiation Sources A key role in the correct implementation of the decree was assigned to a National Technical Commission on Exposure to Natural Radiation Sources, intended to deal with the scientiﬁc and technical problems speciﬁc to natural radioactivity. The Commission is to be made up of 21 experts, coming from relevant ministries, national scientiﬁc institutions and agencies and regional authorities. As regards exposure to radioactivity of terrestrial origin, the main duties of this Commission are: (i) to work out guidelines for methods for measuring radon and thoron concentration in air in workplaces and for the assessment of relevant exposure; (ii) to set criteria for identifying radon-prone buildings and areas; (iii) to establish criteria for identifying the possible higher exposure for workers and the population in work activities with NORM, thermal spas and mining activities not already regulated, and work out guidelines for measurement methods suitable for performing appropriate assessments; (iv) to prepare proposals for updating the legislation; (v) to prepare proposals for standardised adoption of corrective measures to guarantee an optimum level of radiation protection; (vi) to prepare proposals for programmes for educational and refresher courses on radon and thoron measurements and remedial action application. The Commission should have been installed by 1 March 2001 (that is within six months of publication of the decree in the Ofﬁcial Journal). It should complete the ﬁrst and second duty within one year from its installation and the third duty within two years. 2.7. The National Archives for exposure to natural radiation sources at work As mentioned before, when an Action Level is exceeded, the employer is obliged to submit a report to the competent local authorities for the Environment, Health and Labour. The local authorities for Labour should transmit the data to the Ministry of Labour, which is responsible for setting up the National Archives. Data from these archives should be made available to competent authorities and the National Technical Commission. 3. Italian knowledge of exposure to natural radiation at work As regards exposures to radon activity concentration at work in Italy, very few measurements have been carried out in buildings. The only major experience concerns exposure to radon in schools [10]. In addition, some surveys have been carried out in caves and underground workplaces, like tunnels and Underground stations. However, a National Survey on Natural Radioactivity in dwellings was conducted in 1989– 1994 [11,12] which, together with successive local surveys and surveys in schools, could provide an initial base for identifying radon-prone areas and buildings. As regards work activities with NORM in Italy, some experience was gained in the past on radiation protection problems in the zircon sand industry, oil and gas extraction and the

92

S. Risica et al.

use of thoriated gas mantles, whereas other work activities are almost unexplored. Concerning the industrial use of zircon sand, a material largely used in Italian foundries and ceramic or refractory material industries, several radiological surveys and studies were carried out on worker exposures (see, e.g., [7,13]), whereas no external environmental survey was ever been made. As regards oil and gas extraction, an extensive investigation was carried out by ENI in 1992. Hundreds of wells and scores of plants, ﬁelds and platforms for both oil and gas extraction were monitored in Italy and Africa [14]. As for the phosphate industry, a joint investigation was carried out by the Italian Environmental Protection Agency with the Regional Environmental Protection Agency (Veneto Region) in 1998–1999 on the impact on the external environment in a zone of the Venice lagoon of a plant that for many years disposed of phosphogypsum. 210 Po and 210 Pb were measured in both sediments and molluscs [8]. It seems, however, that there are no active phosphate industries in Italy any more, so no exposed workers are expected to be identiﬁed for this type of process [14]. Nevertheless, land contamination could probably be identiﬁed in some areas, due to pre-existing industries and the release of relevant disposal material. No radiation protection investigation has ever been carried out in any of the other types of industries listed in the decree, such as ore processing for the extraction of tin or the processing of rare earth. As regards thermal spas, they could constitute a signiﬁcant problem in Italy as they are very numerous and spread throughout the whole territory. A 1997 estimate of the Ministry of Health, which is responsible for their authorisation, reported 104 spas for water drinking therapy, 339 for mud therapy and 369 for bathing therapy. In some of them, very high levels of radon activity concentration were measured in indoor air, with a consequent signiﬁcant dose to workers (see, e.g., [15]). Moreover, these spas often allow free access to the public even when they are not patients. In the authors’ opinion, an evaluation from the point of view of the justiﬁcation principle is needed for these environments. However, due to the complexity of the problem and its economical aspects a harmonisation among EU countries is desirable.

4. Discussion In the authors’ opinion, some open problems in the application of the Italian decree, could create serious difﬁculties. Several are worth quoting. • The National Technical Commission on Exposure to Natural Radiation Sources, under the Prime Minister’s Ofﬁce, has not been installed yet, causing a probable delay in fulﬁlling its ﬁrst two duties (see Section 2.6). • The measurement results of the radon activity concentration should be communicated to local authorities only if they exceed the Action Levels. This deprives us of a powerful tool for monitoring the present situation and the evolution of the Italian radon activity concentration in workplaces. • Both measurement of radon concentration and of some natural radionuclides (like 210 Po and 210 Pb), critical in NORM industries, and assessment of relevant doses are difﬁcult processes. The experimental and scientiﬁc capacity of the technicians involved should be improved.

The implementation of Directive 96/29/Euratom in Italy

93

• One of the last articles of the decree states that it should be implemented without any ﬁnancial support or even increase in expenses on the part of public institutions. Yet, it should be clear that ﬁnancial support is required to carry out reliable detailed surveys and to create and efﬁciently run the cited Commission and the National Archives.

5. Final remarks This presentation on the implementation of the Euratom Directive in Italy was also aimed at stimulating discussion among the experts, which is a fundamental step in pursuing a broader harmonisation of national regulations. For this purpose, the authors offer some general comments and personal views to further the debate. As regards work activities with NORM, application of Action Levels in terms of effective dose could come up against some difﬁculties. First of all, as is well known, the exposure of workers does not depend only on the activity concentration in materials, because materials are submitted to physical, thermal and chemical processes in which several radionuclides can be of different signiﬁcance. Therefore, each process should be speciﬁcally studied, whereas a limited operative and experimental experience is generally available. Moreover, working environments have mainly been studied, whereas only very few surveys have been carried out on the external environments around these types of plants. In the near future, some work activities that have never been investigated or investigated only in some countries could require broader and more detailed analysis to check if they could be of concern. In particular, geothermal energy plants, building material industries, waterworks, scrap metals arriving at foundries (see, e.g., [16]), etc., should be investigated. Moreover, it cannot be excluded that new materials could be identiﬁed as NORM in the future [17]. Finally, in the authors’ opinion, the new legislation will improve the radiation protection of workers, but will not help acquire more knowledge, because the employers’ duties will not encourage them to give experts access to their work environment or to provide them with radiological data relevant to their work activity.

References [1] [2] [3] [4] [5] [6] [7] [8] [9]

Council Directive 96/29/Euratom of 13 May 1996, Ofﬁcial J. Eur. Commun. Ser. L 159 (29.6.1996). Radiation Protection 88, European Commission, Luxembourg, 1997. Radiation Protection 95, European Commission, Luxembourg, 1999. ICRP Publication 65: Protection against radon-222 at home and at work, Ann. ICRP 23 (2) (1993). Decreto legislativo 26 maggio 2000, n. 241, Supplemento ordinario alla Gazzetta Ufﬁciale della Repubblica Italiana, August 31, 2000. Decreto Legislativo 17 marzo 1995, n. 230, Supplemento ordinario alla GURI 136 del 13 giugno 1995. National group for studying radiological implications in the use of zircon sand, Sci. Total Environ. 45 (1985) 135. ANPA (Agenzia Nazionale per la Protezione dell’Ambiente), Le discariche di fosfogessi nella laguna di Venezia: valutazioni preliminari dell’impatto radiologico, Marzo 2000, ISBN-88-448-0285-6. C. Testa, D. Desideri, M.A. Meli, C. Roselli, A. Bassignani, G. Colombo, R. Fresca Fantoni, Health Phys. 67 (1) (1994) 34.

94

S. Risica et al.

[10] L. Gaidolﬁ, M.R. Malisan, S. Bucci, M. Cappai, M. Bonomi, L. Verdi, F. Bochicchio, Radiat. Prot. Dosim. 78 (1) (1998) 73. [11] F. Bochicchio, G. Campos Venuti, C. Nuccetelli, S. Piermattei, S. Risica, L. Tommasino, G. Torri, Health Phys. 71 (5) (1996) 741. [12] F. Bochicchio, G. Campos Venuti, S. Piermattei, G. Torri, C. Nuccetelli, S. Risica, L. Tommasino, G. Torri, Proceedings of a Workshop on Radon in the Living Environment, Athens, Greece, April 19–23, 1999, Sci. Total Environ. 272 (1–3) (2001) 997. [13] M. Berico, C.M. Castellani, T. De Zaiacomo, M. Formignani, A. Ianni, C. Nobili, S. Sandri, R. Vasselli, Valutazione di dose per i lavoratori occupati in una industria che utilizza vernici a smalto contenenti silicato di zirconio, RT/AMB/99/2, 1999. [14] F. Trotti, S. Bucci, M. Belli, G. Colombo, B. Dalzocchio, G. Fusato, S. Maggiolo, S. Nava, G. Svegliado, C. Zampieri, in: Proceedings of the Conference NORM III, Bruxelles, 17–21 September 2001, in press. [15] A. Boschi, G. Gremigni, V. Sabbatini, in: Proceedings of the Conference Radon tra natura e ambiente Costruito. Radioprotezione Territorio Interventi Informazione Venezia, 24–26 november 1997, p. 287. [16] S. Risica, C. Bolzan, C. Nuccetelli, in: Proceedings of the Conference NORM III, Bruxelles, 17–21 September 2001. [17] M. Fabretto, in: Proceedings of the Conference Safety of Radiation Sources and Security of Radioactive Materials, Dijon, France, 14–18 September 1998, IAEA-TECDOC-1045, IAEA, Vienna, Austria, 1998.

95

Exposure potential of depleted uranium ammunition in the environment – a review J.P. Mc Laughlin Department of Experimental Physics, University College Dublin, Ireland

Depleted uranium (DU) is produced as a by-product or waste material in the production of enriched uranium for use in nuclear reactors or in weapons programmes. DU is thus a material which has been produced from natural uranium ore as a result of the application of technology. As such it falls into the category of a TENORM (technologically enhanced naturally occurring radioactive material). Large amounts of DU have been produced over past decades and it has found uses in such diverse areas as medicine, aviation, sport, various general industries and most prominently in the weapons industry. It is generally considered to have been ﬁrst used offensively as ammunition in the Gulf War in 1991 and since then in conﬂicts in the Balkans, most recently in different parts of Yugoslavia in 1999. From a health perspective, uranium in any isotopic form can be considered as primarily a chemically toxic heavy metal but most media attention has focused on the radio-toxic potential arising from the use of DU in these conﬂicts. Most concern initially was and continues to be expressed regarding the possible health effects on military personnel but its long-term effects on the general population living in areas where the environment has been contaminated by DU has so far received less attention. In this review, an account is given of some possible and likely scenarios where DU in the environment may have a potential for human exposure and health risks. With the exception of prolonged skin contact with DU metal and the rare practice of geophagia, it is very unlikely that in the long-term members of the general population in target areas will receive radiation doses of radiological signiﬁcance, but drinking water safety standards based on the chemical toxicity of uranium may be exceeded.

1. Introduction Natural uranium contains three uranium isotopes: 238 U (99.3% by mass), 235 U (0.72% by mass) and 234 U (0.006% by mass). In the process of enriching the 235 U content to 2% or greater, for use in nuclear reactors or in weapons programmes, depleted uranium (DU) metal is produced as a waste product. DU is essentially almost pure 238 U (99.8% by mass) with a 235 U RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07010-X

© 2005 Elsevier Ltd. All rights reserved.

96

J.P. Mc Laughlin

mass content of about 0.2% and its speciﬁc activity at about 15 Bq mg−1 is about 60% that of natural uranium (25.4 Bq mg−1 ). As products of the processing of uranium ore, both enriched uranium and DU are examples of TENORMs (technologically enhanced naturally occurring radioactive materials). While DU metal has found many commercial uses, ranging from radiation shielding in hospitals to pile driving in civil engineering, its most publicly prominent use has been in ammunition. In this latter application, its high density, self-sharpening and pyrophoric properties have made it an ideal material to be used as armour piercing incendiary penetrators. DU is also used as a protective armour on tanks and in some other military applications [1]. In the recent past it has been used extensively in the Gulf War (1991), in Bosnia and Herzogovina (mid-1990s) and in Yugoslavia (1999). It is not unreasonable to assume that DU of non-western origin may also have been used in some military applications in other recent conﬂicts. Present opinion is that DU will become increasingly used by the military in the future thus highlighting the need for studies of its long-term behaviour and potential for human exposure in different environments. The possible effects of DU on military personnel involved in the Gulf War is a matter of continuing interest in the media. In this paper, however, aspects of the use of DU in the 1999 conﬂict in Yugoslavia is the main focus. It is estimated that in 1999 approximately 30 000 DU penetrators, each of mass about 300 g, were used against targets in Kosovo and that a further 5000 were used on targets in the south of Serbia adjacent to Kosovo. The majority of these penetrators, which were ﬁred from A-10 Warthog aircraft, did not strike armour or other hard targets. Thus most of the penetrators remain intact and buried in the ground at depths which make them almost undetectable by surface radiation monitors. Starting already in 1999 UNEP (United Nations Environmental Programme) commenced a study and ﬁeld investigations into the potential effects of DU on human health and the environment in Kosovo [2,3]. This work has produced a number of recommendations including the need for a study of the rates of dissolution and migration of DU into water supplies and the food chain in Balkan target areas. To date, however, no systematic study of these topics has taken place. A limited amount of initial data on DU uptake by the general population in some Balkan target areas has become available but here again a systematic study has not yet taken place [4].

2. Health effects An adult human body typically contains about 100 μg of uranium which derives from natural uranium in our diets [1]. Depending on such factors as the natural uranium content of drinking water and food, there is however a wide variation worldwide in the natural uranium body burden. While the behaviour of uranium in the human body will be inﬂuenced by its chemical state and by individual metabolisms, its general behaviour in the body are well known. There is an extensive literature dealing with uranium exposure and toxicities, but here only the main features are summarised. 2.1. Uranium toxicity The majority of ingested uranium is excreted with only about 5% being absorbed. About two thirds of the uranium absorbed in the blood is rapidly eliminated through the kidney and

Exposure potential of depleted uranium ammunition in the environment – a review

97

within 24 hours up to nearly 90% will usually have been excreted from the body in urine. More than half of the uranium in the body is in the skeleton with most of the remainder to be found in the fat and muscle. Irrespective of isotopic composition uranium, both natural and depleted, can be considered as a mildly radioactive heavy metal thus presenting both radiological and chemical hazards. On the basis of animal studies, and on human studies to a lessor extent, the kidney is considered to be the organ of most susceptibility from the chemical toxicity perspective [1]. From a chemical exposure perspective, uranium is considered primarily as a nephrotoxic material. From a radiological perspective, the greatest doses arising from inhalation of uranium particles will be to the lung and associated lymph nodes. An increased risk of lung cancer is thus perceived to be the main radiation risk although some consider that the risk of lymphatic cancer has been underestimated [5]. It must, however, be emphasised that while dosimetry-based estimates of the cancer risk from inhaled uranium particles can be made, there has been no evidence of cancer induction by uranium in humans. Even in the case of animal studies, the induction of cancer could only be demonstrated at exceedingly high intakes of uranium. There is an extensive amount of published material on the toxic effects of uranium compounds on laboratory animals [6]. By means of ingestion, inhalation or injection these animals had incorporated large amounts of uranium compounds. The results of these animal experiments, due to the wide range of sensitivities of different animal species to uranium, are of limited applicability to the estimation of human health risks associated with uranium intakes. Human studies of individuals who received large acute intakes of uranium are very few and in these the main effects found were minor and reversible dysfunctions of the kidney [7]. In the case of workers in the uranium industry (processing ores or in uranium metal fabrication plants) there is little or no evidence for a general increase in clinically observable kidney disease [8]. Where exposures occurred resulting in kidney concentrations of about 1 μg uranium per gram of kidney, these gave rise to minor kidney dysfunction. While the total speciﬁc activities of the uranium isotopes of the DU penetrators was found to be about 13.8 × 106 Bq kg−1 is of interest to note that low activities of 239+240 Pu (< 50 Bq kg−1 ) in the DU penetrators have also been measured [9,10]. These activities are comparable to the trace amounts of 239 Pu that occur naturally in uranium ore due to capture of neutrons from spontaneous ﬁssion and alpha–neutron reactions [11]. Following the initial chemical processing of uranium ores, the naturally occurring plutonium enters the waste streams. Thus the trace amounts of plutonium and also of 236 U (∼ 7 × 104 Bq kg−1 ) measured in the DU is probably due to the presence of spent fuel or contamination in the processing plant. The presence of the traces of plutonium is of no radiological health signiﬁcance, both in absolute and relative terms, as the radiation dose from the plutonium in a 1 mg inhaled insoluble DU particles can be estimated to be < 1 nSv while that due to the DU itself is 0.15 mSv. 2.2. Limits of intake Because of its low radiotoxicity, annual ingestion of DU in amounts in excess of a limit would be required to give rise to radiation doses at a level considered to present a radiological health risk of signiﬁcance. For example, the dose factor for ingestion by the public of DU (0.2%) is estimated to be about 0.7 mSv g−1 thus giving rise to an ALI (Annual Limit on Intake) of about 1.4 g based on the ICRP radiological limit of 1 mSv y−1 for members of the public

98

J.P. Mc Laughlin

Table 1 Uranium and DU (0.2%) ALI (Annual Limit of Intake) and DDWC (Derived Drinking Water Concentration at 500 L y−1 ) for the public based on radiological limit (1 mSv y−1 ) and on chemical toxicity, respectively [12,14] ALI

Inhalation Ingestion

DDWC

Radiological

Chemical

Radiological

Chemical

134 mg (soluble) 8.3 mg (insoluble) 1410 mg

n/a

n/a

n/a

15.3 mg

2820 μg L−1

2 μg L−1

(see Table 1) [12,13]. On the other hand, an ingestion ALI of 15.3 mg was recommended in 1998 by the WHO for reasons of uranium chemical toxicity [14]. These large differences in recommended ALI values reﬂect the different paradigms of the signiﬁcance of health impacts used by bodies such as ICRP and WHO. The radiation limits of ICRP are based on what is termed “total detriment” which in turn is largely based on the risk of fatal cancer. The WHO assessment of the chemical toxicity of uranium is based on the TDI (tolerable daily intake), approach which evaluates the levels at which toxicological effects occur in various animal markers. These toxicological effects are generally non-cancerous and often are at a sub-clinical level. For example, in the case of uranium ingestion, the chemical toxicity limits have focused on its non-cancerous nephrotoxicity. These differences in approach and in the consequent values of recommended limits are a cause of considerable confusion for public health ofﬁcials in DU target areas. For many reasons, which are not always scientiﬁc, there are considerable quantitative differences between the limits for uranium intakes recommended by both international and various national agencies concerned with protecting human health. This creates difﬁculties for health ofﬁcials in target areas to decide upon an appropriate strategy in dealing with possible DU contamination. Table 1 gives a summary of some of the most relevant ingestion and inhalation recommended intake limits based on the recommendations by two international bodies: the International Commission on Radiological Protection (ICRP) and the World Health Organisation (WHO) [12,14]. It will be noted that the WHO chemical toxicity based DDWC is much more stringent than that of the ICRP which is radiologically based. Even where a limit may be exceedingly conservative or considered by some to be over-protective, health authorities generally consider that an intake or concentration of a substance that exceeds a current WHO health standard guideline is a matter of concern. It should be noted that the 2 μg L−1 provisional DDWC recommended by the World Health Organisation is considered to be protective for subclinical renal effects reported in epidemiological studies [15]. A recent study into the renal effects of uranium in drinking water in Finland suggests that the safe concentration of uranium in drinking water, based on changes in renal function, may be within the range 2–30 μg L−1 [16]. In this study, the drilled wells waters investigated had a median natural uranium concentration of 28 μg L−1 ranging to a maximum of 1920 μg L−1 , which are values well above the WHO drinking water standard of 2 μg L−1 .

Exposure potential of depleted uranium ammunition in the environment – a review

99

2.3. Risk perception The health situation in target areas is further complicated by the general belief in the Balkans, largely on the basis of unsubstantiated anecdotal and media accounts, that a major health hazard exists due to the presence of DU in the environment, and that many hundreds of deaths due to cancer have already occurred as a result of DU contamination. While there is no scientiﬁc evidence for this and while such perceptions are at variance with what is known concerning uranium toxicology and the latency period for cancer induction, these perceptions themselves may in fact be seen as a risk to health by causing unnecessary stress and anxiety in the general population. Although at a much lower level of intensity, this is not dissimilar to the psychosomatic distress evident in many individuals affected by the Chernobyl accident where it has been concluded that the accident has had a signiﬁcant long-term impact on psychological well-being and health related quality of life [17]. Notwithstanding the above considerations, it can, however, be argued that there is an ethical obligation on the scientiﬁc community to carry out a systematic study of exposure pathways in DU affected areas in conjunction with local health ofﬁcials to determine the actual situation thereby helping to create an informed perspective in the community which may help to allay unnecessary fears and reduce their perpetuation. As in all matters relating to public perception of radiation risks, this is perhaps an overly optimistic aspiration.

3. Behaviour of DU penetrators in soils In considering the long-term human exposure potential of the DU used in the Balkan conﬂicts in the 1990s, a clear distinction must be made between the small number of DU penetrators that were aerosolised by impact with heavy armour or other hard targets and the majority that were implanted almost intact in the ground. The former may have had some limited inhalation exposure signiﬁcance in the immediate and short-term periods following an attack. Here it is the essentially intact DU penetrators buried in the ground that are being considered as they may in the long term contribute to exposure of the general population principally by ingestion of water and food. 3.1. Dissolution and migration Most of the small number of buried penetrators which were recovered by UNEP were found to be intact or slightly damaged but already undergoing rapid corrosion [3]. In most soils, DU metal will rapidly oxidise to form uranium oxides. Penetrators recovered in Kosovo and Serbia within a year of their implantation in the ground were found to be heavily coated with black and yellow phases. Both of these alteration phases are uranium oxides with the yellow phase schoepite (UO3 2 H2 O) being the most abundant [3,18]. The behaviour of DU in target areas thus requires a detailed knowledge on the dissolution kinetics and solubility of schoepite in the environment, and of the transport of hexavalent uranium through the soils into the groundwater below. Although values of ∼ 100 μg kg−1 are quoted for the solubility of schoepite the actual solubility at a site is very speciﬁc to both the characteristics of the soil and the schoepite itself. Schoepite which is poorly crystalline or amorphous has been shown

100

J.P. Mc Laughlin

to have increased solubility up to a maximum of 10 000 μg kg−1 [19]. Its solubility can be increased also by the presence in the soil and groundwater of organic and inorganic acids such as humic and carbonate acids. Any study of human exposure to DU in the environment is complicated by the presence of natural uranium in the soils and ground water. While the average concentration of uranium in the Earth’s crust is 2–3 mg kg−1 , it can vary from as low as 0.3 mg kg−1 to 30 mg kg−1 in granites and as high as 300 mg kg−1 in shales [3,20]. The soils in Kosovo which were investigated by the UNEP team were found to have average uranium concentrations in the range ∼ 2–5 mg kg−1 [3]. The UNEP study, however, also established in some target sites that DU, in about 18 months after implantation, had in the immediately surrounding soil increased the uranium concentrations into the 102 –104 mg kg−1 range. This indicates that the DU schoepite is orders of magnitude more soluble than natural uranium contained in the mineral soil phases at the sites investigated. Many factors, such as soil Eh and pH values, play an important role in the processes of dissolution and migration and these limited studies by the UNEP team clearly show the need for detailed characterisation of target area soils. At DU munitions test sites in the US and UK, the corrosion rates in the ground were such that DU penetrators could reasonably be expected to be completely corroded in the ground at these sites from the metallic state in about 10 years. In order to predict the migration of uranium in a speciﬁc environment, it is necessary to understand the geology of the areas studied, as well as the hydrogeologic properties of the rocks that contain unconﬁned surface aquifers. 3.2. Contamination of drinking water In assessing the potential of DU to contaminate drinking water, a detailed knowledge is needed of the quantity of DU present, the geological and hydrogeological conditions to order to make a realistic prediction of the future contamination of groundwater. The areas studied by the UNEP team in Kosovo had a number of domestic wells where water could be collected for analysis. Also, the depth to the water table could be measured and it was generally at the depth of 4–10 m. Therefore drinking water, in these cases, is being used from unconﬁned aquifers and no conﬁning layer is present to shield the aquifers from DU contamination. A worst-case scenario with respect to the contamination of groundwater might be where DU penetrators lodge in a shallow groundwater alluvial aquifer where the corroding uranium might be released into a well on a time scale of a few years. On the other hand, best-case scenarios may be envisaged where penetrators lodge in absorbing soils of high organic carbon content and where the presence of clay layers form an effective impermeable barrier to the groundwater. It should be noted that in Kosovo and Serbia, many of the target locations are in rural areas where small villages and individual households take their drinking water from wells as distinct from run-off surface water reservoirs. At the date of the UNEP team’s visit in November 2000, no DU was detected in any of the ground waters analysed. The uranium concentrations varied from roughly 0.2 μg L−1 up to 2 μg L−1 and none of the uranium detected was DU. These concentrations are within the current WHO limit for drinking water of 2 μg L−1 [14]. It is, however, possible that the 18 months which had elapsed since implantation was not long enough for the DU to leach down to the water table. It was estimated that it might take a number of years for the DU to reach the wells. It is considered likely that many of the penetrators will have completely corroded

Exposure potential of depleted uranium ammunition in the environment – a review

101

in about 15 years. In place of the uranium metal, the surrounding soil will then contain high levels of schoepite which in time will dissolve and transport the uranium downward through the soil proﬁle. This downward transport may be limited by the presence of organic matter or minerals to which the uranium will be sorbed.

4. Exposure scenarios The three principal DU exposure pathways for the general population are by inhalation, ingestion and by skin contact. For military personnel, additional exposure pathways such as due to embedded DU fragments in the body and whole body exposure of crews of tanks and other military vehicles carrying DU ammunition or using it as shielding may occur. These military exposures are not discussed here. 4.1. External exposure External exposure doses from skin contact will occur at an equivalent dose rate of about 2.5 mSv h−1 from handling DU penetrators. This is mostly due to beta and gamma rays as the alpha particles are absorbed in the outer layer of skin. Deterministic effects, such as erythema due this dermal contact should not occur. While skin is relatively insensitive to radiation, a prolonged skin exposure of some weeks could however increase signiﬁcantly the risk of skin cancer [5]. In the present context, a young boy wearing a recovered DU penetrator as a pendant next to the skin for a number of weeks is a not unlikely scenario in which this might arise. To place this in some context, it should be noted that the recommended ICRP annual equivalent doses to the skin are 500 mSv (occupational) and 50 mSv (public) [12]. At a skin dose of 2.5 mSv h−1 from dermal contact these limits could be exceeded in about 200 and 20 h, respectively. Prolonged skin contact with solid pieces of DU should therefore be avoided. 4.2. Internal exposures The two principal exposure pathways are by inhalation and ingestion. Inhalation may be an important exposure pathway for the public in target areas in the immediate and short-term aftermath of an attack where DU was aerosolized by impact with a hard target. Resuspension of contaminated soil particles constitutes another inhalation pathway of less toxicological signiﬁcance as wind dispersion of the particles with time will tend to reduce the DU content of surface soil at a location. In this account, as the primary focus is on long-term exposure possibilities, the exposure of the public by means of inhalation is not considered. Ingestion of DU contaminated soil, water and food contaminated will give rise to internal exposure. In the case of DU contaminated soil, for most people it is unlikely that the occurrence of pica or the inadvertent ingestion of small quantities of soil or dust in target areas will exceed public radiation limits based on radiological hazard. Cases of occasional deliberate ingestion of soils, as happens with most young children, may give rise to enhanced radiation doses and chemical exposure to uranium. This could occur when playing in attacked vehicles where the ingested material might be very contaminated (at ∼ 10 g kg−1 ) with DU. It

102

J.P. Mc Laughlin

does, however, appear that most DU destroyed vehicles in Kosovo and Serbia have now been removed or made inaccessible to children which reduces this exposure possibility. Unfortunately, there is only a limited amounted of reliable quantitative information on soil ingestion by children and there appears to be none speciﬁc to uranium ingestion. Most recent data on soil ingestion are based on mass-balance studies of conservative tracer elements such as aluminium which are not absorbed in signiﬁcant amounts in the gut [21]. WHO suggests 20 mg as a typical level of daily inadvertent ingestion soil/dust but recent data suggests a central estimate closer to 80 mg. Of more potential concern than this is the practice of geophagia where many grams of soil are persistently and purposefully consumed per day. This is usually, but not exclusively, practised by pregnant females and children in economically underdeveloped societies and may in part be a response to some deﬁciencies in nutrition. If geophagia were to be practised in the local environment of a DU target site, where an uranium soil contamination of ∼ 1 g kg−1 is a reasonable estimate, taking a central estimate of geophagic ingestion of 26 g of soil per day the resulting radiation dose would be about 15 mSv y−1 [21]. This would also correspond to a uranium chemical exposure of about 9.5 g y−1 . From the perspective of both toxicologies, this scenario would be considered to carry unacceptable health risks. While there is no evidence to suggest that a “critical group” of geophagic individuals may be present in Balkan DU target areas, it would be remiss of any detailed exposure assessment of populations in all such present and future DU contaminated areas if the possibility of geophagia were not to be considered. Soil ingestion by cattle and other animals used for food is also a human exposure pathway that needs to be considered in exposure studies. Some preliminary investigations of DU body burdens in the Balkans have taken place. Using MC-ICPMS (multi-collector inductively coupled mass spectrometry) urine samples from a small number of residents of Bosnia-Herzogovina and Kosovo have been analysed for both natural and depleted uranium content [4]. The results obtained suggest DU body burdens from close to zero to about 46 μg. These estimated values are at most about half the natural uranium content of the body suggesting that health effects will not occur. Further more systematic studies of DU in urine in populations in target areas should take place.

5. Conclusions The use of DU munitions in the past decade has given rise to concern regarding its possible health effects. Excluding the immediate and short-term effects that DU ammunitions can have on persons in an attack, it is some of its potential long-term effects on the general population in target areas that has been the focus of this account. Of particular interest are the exposures that may take place by ingestion and skin contact. In the case of DU use in the Balkans, no systematic study has yet taken place to determine its rates of dissolution and migration in the soils of target areas. The ﬁndings of the UNEP investigations there indicate that these rates may be sufﬁciently large to allow DU to enter drinking waters on a time scale of a few years to some decades. From the human health perspective, DU can be considered as a mildly radioactive heavy metal with both radiological and chemical toxicity potential. At the levels of intake, both by inhalation and ingestion, which might reasonably occur in most scenarios considered, the radiological risks are considered as insigniﬁcant (i.e. < 1 mSv y−1 for the public). Notwithstanding the estimations of risk that can be made, even on the basis of low

Exposure potential of depleted uranium ammunition in the environment – a review

103

radiation doses, there is no evidence that uranium is carcinogenic to humans. It is possible, however, for those practising geophagia in target areas or for those having prolonged skin contact with DU metal, that radiological hazards of signiﬁcance could occur. In the former group, the chemical toxicity risk would be of more signiﬁcance. These two groups might in fact be reasonably considered as critical groups for DU exposure in the general population. In the case of its chemical toxicity, it is quite likely in target areas that the stringent health standards of the WHO may be exceeded for ingestion and for drinking water in particular. In the case of DU contaminated water supplies or soil, a number of preventative or remedial options exist which can be considered by local health authorities. Prohibiting or restricting access to or use of these resources is one option. Other options include the removal of DU from drinking water by, for example, precipitation with calcium carbonate or sorbent materials. In the case of DU in soils, the application of phosphates to the soil to form uranium compounds such as uranyl-phosphates and calcium–uranyl phosphates which are of very low solubility and bio-availability to crops is another option [22]. Apart from the socio-economic implications of these options, none of them can be assessed until systematic and thorough studies of DU contamination in the affected areas takes place.

References [1] N.H. Harley, E.C. Foulkes, L.H. Hilbourne, A. Hudson, C. Ross Anthony, A Review of the Scientiﬁc Literature as it Pertains to Gulf War Illnesses, Vol. 7: Depleted Uranium, National Defense Research Institute, USA, 1999. [2] The Potential Effects on Human Health and the Environment Arising from the Use of Depleted Uranium During the 1999 Kosovo Conﬂict, United Nations Environmental Programme, Geneva, October 1999. [3] Depleted Uranium in Kosovo – Post Conﬂict Environmental Assessment. Report of the United Nations Environmental Programme Mission to Kosovo, 5–19 November 2000, United Nations Environmental Programme, Geneva, April 2001. [4] N.D. Priest, M. Thirlwall, Early results of studies on the levels of depleted uranium excreted by Balkan residents, Arch. Oncol. 9 (4) (2001) 237–240. [5] The Health Hazards of Depleted Uranium Munitions, Part I, The Royal Society, London, May 2001. [6] M. Ellender, J.D. Harrison, H. Pottinger, J.M. Thomas, Induction of osteosarcoma and acute myeloid leukaemia in CBA/H mice by the alpha-emitting nuclides uranium-233, plutonium-239 and americium-241, Int. J. Radiat. Biol. 77 (2001) 41–52. [7] N.D. Priest, Toxicity of depleted uranium, Lancet 357 (January) (2001) 244–246. [8] D. Mc Geoghegan, K. Spinks, The mortality and cancer morbidity experience of workers at Springﬁelds uranium production facility, 1946-95, J. Radiol. Prot. 20 (2000) 111–137. [9] J.P. Mc Laughlin, L. Leon Vintro, K.J. Smith, P.I. Mitchell, Z.S. Zunic, Plutonium in depleted uranium penetrators, Arch. Oncol. 9 (4) (2001) 225–229. [10] J.P. Mc Laughlin, L. Leon Vintro, K.J. Smith, P.I. Mitchell, Z.S. Zunic, Actinide analysis of a depleted uranium penetrator, J. Environ. Radioact. (2002), in press. [11] D.M. Taylor, Environmental plutonium-creation of the universe to 21st century mankind, in: A. Kudo (Ed.), Environmental Plutonium, in: Radioactivity in the Environment Series, vol. 1, Elsevier, 2001, pp. 1–14. [12] ICRP Publication 60: 1990 Recommendations of the International Commission on Radiological Protection, Ann. ICRP 21 (1–3) (1991). [13] ICRP Publication 69: Age-dependent doses to members of the public from intake of radionuclides, Ann. ICRP 25 (3–4) (1995). [14] WHO, Health criteria and other supporting information, Addendum to vol. 2 in: Guidelines for Drinking-Water Quality, 2nd ed., World Health Organisation, Geneva, 1998. [15] World Health Organisation, Depleted Uranium: Sources, Exposures and Health Effects, in: WHO Monograph Series, vol. WHO/SDE/PHE/01.1, WHO, Geneva, April 2001.

104

J.P. Mc Laughlin

[16] P. Kurttio, et al., Renal effects of uranium in drinking water, Environ. Health Perspect. 110 (4) (2002) 337–342. [17] UNSCEAR, Sources and Effects of Ionizing Radiation, vol. II: Effects, United Nations, New York, 2000. [18] V. Ragnarsdottir, L. Charlet, Environmental behaviour of uranium, in: Environmental Mineralogy, Mineralogy Society of Great Britain and Ireland, London, 2000, pp. 333–377. [19] X.J. Bruno, I. Casas, B. Iagerman, M. Munoz, The determination of the solubility of amorphous UO2 (s) and the monomolecular hydrolysis constants of uranium (VI) at 25 ◦ C, in: Material Research Symposium Proceedings, vol. 84, 1987, pp. 153–160. [20] D. Langmuir, Uranium solution–mineral equilibria at low temperatures with applications to sedimentary ore deposits, Geochim. Cosmochim. Acta 42 (1978) 547–596. [21] The Health Hazards of Depleted Uranium Munitions, Part II. Annexe C (B. Smith–BGS): Estimate of infant doses from the direct ingestion of soils or dusts containing uranium and depleted uranium, The Royal Society, London, March 2002. [22] X.J. Bruno, et al., Estimation of the concentrations of trace metals in a natural system. The application of codissolution and co-precipitation approaches to El Berrocal (Spain) and Pocos de Caldas (Brazil), Chem. Geol. 151 (1998) 277–291.

3. Radioactivity measurements, releases and dosimetry

T . his Page Intentionally Left Blank

107

Changes in terrestrial natural radiation levels over the history of life P.A. Karam a , S.A. Leslie b a Department of Biological Sciences, Rochester Institute of Technology, 85 Lomb Drive, Rochester, NY 14623, USA b Department of Earth Sciences, University of Arkansas at Little Rock, 2801 South University,

Little Rock, AR 72204-1099, USA

Background radiation dose has changed throughout Earth’s history. Through geologic time, the uppermost layers of Earth became steadily enriched in radioactive 40 K, U, and Th. As this enrichment occurred, radioactive decay reduced the radioactivity concentrations in the crust. The result is that radiation levels from geologic and internal biologic emitters (40 K) likely peaked about 2 billion years ago (Ga) at a dose rate of about 7 mGy yr−1 , while radioactive decay of 40 K reduced radiation dose from internal emitters by a factor of 10 since life ﬁrst appeared [1]. Throughout Earth’s history radiation dose from galactic cosmic rays (GCRs) has likely increased by a factor of nearly 10 while dose from solar charged particles has dropped by a factor of about 8 through the last 4 Ga as a result of solar evolution. Solar UV emissions have increased steadily, but the formation of Earth’s ozone layer about 2 Ga resulted in the ﬂux of ionizing UV dropping by a factor of over 400 through time. In addition to these steady changes, episodic events such as very large solar ﬂares, nearby supernovae, and even gamma ray bursts have the ability to produce short-lived sea level radiation doses on the order of 1 Gy, and such events are expected to occur once or twice per species lifetime [2,3]. It is possible that these events exert a selection pressure in favor of resistance to radiation damage among organisms living at or near Earth’s surface. Changes in background radiation levels, atmospheric oxygen concentrations, and other environmental mutagens through geologic time may have inﬂuenced the manner in which modern organisms respond to DNA damage. In particular, changes in oxygen concentrations also affect rates of radiogenic DNA damage because of oxygen’s role in enhancing radiation damage [4]. Such changes may have inﬂuenced evolution for a wide variety of organisms. The speciﬁc environment in which ancient organisms lived (e.g., aerobic vs. anaerobic or photic zone vs. deep marine) likely affected their DNA damage rate and, hence, their individual exposures to DNA-damaging agents at times in the past [5]. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07011-1

© 2005 Elsevier Ltd. All rights reserved.

108

P.A. Karam, S.A. Leslie

1. Introduction Radiation is one of the many DNA-damaging agents to which life has been exposed since it ﬁrst appeared on Earth. The sources of natural background radiation include U, Th, and 40 K in the rocks and soil; the radioactive fraction of potassium that is part of our essential biochemistry, and radiation from cosmic sources. The damage done to the genome is related not only to the actual types and amounts of radiation to which life is exposed, but to other modifying factors, as well. The foremost of these is free oxygen, which strongly enhances radiogenic DNA damage. Earth’s ozone layer is made of oxygen and provides protection against ionizing wavelengths of UV radiation that would otherwise damage organisms at Earth’s surface. Therefore, the chemical evolution of the atmosphere (most signiﬁcantly, the addition of oxygen) is at least as important a factor in the history of radiation damage as is any other factor. This research touches on all of these sources of radiation damage. 1.1. Radiation dose from biological emitters Early life is thought to have arisen in the ancient oceans, probably on contact with the seaﬂoor. Although early organisms probably became free-ﬂoating early in the history of life, it is likely that the seaﬂoor remained a favored habitat and, indeed, many organisms continue to live in and on the ocean’s ﬂoor. In addition, fossil evidence strongly suggests that the earliest life quickly became colonial, forming relatively thick algal and bacterial mats and stromatolites. Even though the earliest organisms may have been sheltered from other sources of radiation (i.e., UV and cosmic radiation), they could not have escaped radiation from their internal biochemistry or the seaﬂoor on which they lived. Life, both at its inception and throughout most of its history, has been exposed to signiﬁcantly higher radiation levels than exist today from these inescapable sources of radiation exposure. The only variables in dose from biological 40 K are the time at which the dose rate is calculated and the organism’s K concentrations. If we assume a K concentration of 250 mmol L−1 ,1 we ﬁnd that the radiation dose rates from internal 40 K have decreased steadily since life evolved from about 5.5 to about 0.70 mGy yr−1 at present, as shown in Fig. 1. 1.2. Radiation dose from geologic emitters Though geologic time, Earth’s U, Th, and K were concentrated into Earth’s crust because large ions such as these are preferentially partitioned into the magma that erupts at Earth’s surface. As a result of this process, the concentrations of U, Th, and K in Earth’s crust have increased. At the same time, however, all three elements have undergone radioactive decay. The result of these competing processes is that radiation dose from geologic emitters remained nearly steady at 1.6 mGy yr−1 until about 2 Ga and has dropped steadily since that time [1]. This reﬂects the partition of radionuclides into the continental crust (via continental crust formation) at a rate similar to the decay rate during this period. The calculated radiation dose from all geologic materials was about twice that of current levels when the ﬁrst continental crust formed about 4 billion years ago (Fig. 1). This dose 1 A concentration of 250 mmol L−1 is used as an example of an average bacterium. See [1] for a more complete discussion of K-concentrations and resultant doses through geologic time.

Changes in terrestrial natural radiation levels over the history of life

(a)

109

(b)

Fig. 1. (a) Changes in background radiation levels through time. The x-axis starts at the present day and goes to 4 Ga. (b) Shows changes in cosmic radiation dose on an expanded scale.

was virtually identical to that from maﬁc2 rocks at that time because very little felsic crust existed. As the continental crust formation rate slowed, radiation levels began to fall, reﬂecting the decay of U, Th, and, most importantly, 40 K. The modeled radiation level at present (0.66 mGy yr−1 ) is higher than the 0.28 mGy yr−1 typically attributed to dose from geologic materials (NCRP, 1987) because soils (included in the 0.28 mGy yr−1 ﬁgure) are usually deﬁcient in U, Th, and 40 K compared to rocks, and the calculated doses reﬂect the chemical composition of rocks alone. It is likely that soils resembling those at present did not begin to form until the colonization of land by large plants about 380 million years ago. Before this time soils would be expected to closely resemble, both chemically and radiologically, the underlying rock. The calculated dose does compare favorably with doses derived from radionuclide composition information reported in Eisenbud and Gesell [6] and UNSCEAR [7] (0.71 and 0.77 mGy yr−1 , respectively). 1.3. Radiation dose from solar charged particles and galactic cosmic rays Changes in cosmic radiation exposure clearly fall into two major categories: long-term, gradual changes that stem from continuing physical and chemical processes and short-term events that occur infrequently and aperiodically. The most important factors controlling extraterrestrial radiation dose experienced on Earth include: • Solar evolution and its effects on UV emission, solar charged particle ﬂux, and galactic cosmic ray ﬂux. • Long-term changes in the chemical composition of Earth’s atmosphere and the effects on UV ﬂux and on radiogenic DNA damage. • Long-term changes in the terrestrial magnetic ﬁeld, including the impact of occasional magnetic ﬁeld reversals on the formation of cosmogenic radionuclides. • Short-term effects of large solar ﬂares on sea level radiation exposure. 2 Maﬁc rocks have high concentrations of magnesium and iron, and low concentrations of U, Th, and K. Basalt is an example of a maﬁc rock. Felsic rocks are rich in feldspars, quartz, muscovite, and feldspathoids and high concentrations of U, Th, and K. Granite is an example of a felsic rock.

110

P.A. Karam, S.A. Leslie

• Short-term effects of prompt gamma rays from supernovae and gamma ray bursts and the radiation dose generated at sea level from these events. • Intermediate-term (i.e., over several months) radiation exposure from the decay of supernova-produced radioactivities. The Sun contributes to terrestrial radiation dose by emission of charged particles (the solar wind) and by the action of its magnetic ﬁeld, which helps to exclude high-energy galactic cosmic rays (GCR) from the inner Solar System. Galactic cosmic rays are more effective than solar charged particles at producing sea level radiation dose because of their higher energies, so the effect of changes in solar activity (which affects solar magnetic ﬁeld strength) on sea level radiation dose is not simple. This is shown in Fig. 1b. Over the Sun’s history, solar activity has dropped steadily [8] as its rotation has slowed. This, in turn, means that the solar wind has weakened and solar charged particles have had lesser importance in sea level radiation dose while GCRs have become steadily more important [9,10]. Calculations show that the relative contributions of solar and galactic cosmic rays to sea level radiation exposure have changed continuously through time and that, on the early Earth, solar cosmic rays were a more important source of radiation than were GCRs [9,10]. The lowest calculated cosmic radiation dose from these combined sources occurred about 1 billion years ago at a value of about 0.2 mGy yr−1 (from a peak value of almost 1.2 mGy yr−1 in the earliest Solar System) before gradually increasing to modern values. However, we note that, even at the highest levels, cosmic radiation played only a minor role in the overall background radiation “picture” at any time in the past. Changes in radiation dose from cosmic, geologic, and biological sources are summarized in Fig. 1a and 1b. 1.4. Ultraviolet radiation dose The surface of the Sun behaves, to a reasonably good ﬁrst-order approximation like a black body in that its spectrum depends primarily upon its temperature. The presence of absorption lines for some elements and emission lines for others causes the Sun to depart from the ideal, but these effects are not considered in this paper. Over its lifetime, the Sun has become gradually hotter with a corresponding spectral shift to shorter wavelengths, so UV emissions today are much higher than in the past. At the time when life appeared at 4.0 Ga the luminosity of the Sun was approximately 70% of its present luminosity [11,12]. Plank’s law states that the energy emitted in any wavelength is related to the black body curve. According to Planck’s Law, the energy emitted in a speciﬁc wavelength (for example, the UV emission) of a black body is proportional to 1/(e1/T − 1). As the Sun has become hotter, its UV emissions have increased as well. In fact, UV emissions in the early Sun, when the Sun’s surface temperature was 90% the value today, were about 50% current levels. Current solar evolution models suggest that the Sun’s surface temperature has risen nearly linearly with time [8]. Therefore, UVB and UVC emissions have changed as well, increasing to modern levels as the Sun has evolved. Increases in emissions of ionizing UVB and UVC, however, only address the source term. Of equal importance is to determine the amount of radiation reaching Earth’s surface. The primary UV shield is ozone. Before the formation of Earth’s ozone layer, UVB and UVC

Changes in terrestrial natural radiation levels over the history of life

111

reached Earth’s surface almost unattenuated. Thus, surface UV ﬂux was signiﬁcantly higher in the distant past than it is today even with lower source emissions. A series of rocket-borne experiments have provided information on the solar UV spectrum in space and the absorption cross-section for ozone across this spectrum. Others have reported on the average column density of ozone molecules in the atmosphere [13]. These data can be combined to arrive at the fraction of incident UV radiation that reaches sea level, assuming that scattering and absorption by other atmospheric constituents is negligible. The resulting sea level photon ﬂux for each wavelength is then in units of photons cm−2 s−1 . The photon ﬂux is converted to an energy ﬂux by calculating the energy of individual photons. Multiplying the photon energy by the photon ﬂux yields an energy ﬂux at sea level. Because the UV spectrum is reported in wavelength bands, the central wavelength of each 5 nm band was used for these calculations. Summing the UV ﬂux shows that, today, the solar UV ﬂux is about 14 000 ergs cm−2 s−1 in space and is attenuated to about 29 ergs cm−2 s−1 at sea level [9]. In the past, Earth had noticeably less oxygen in the atmosphere and correspondingly less ozone. This allowed high levels of UV to penetrate to sea level. To calculate sea level UV ﬂux in the past, ozone concentrations were assumed to be directly proportional to atmospheric oxygen levels at that time (Cockell, personal communication). For example, today’s atmosphere contains about 21% oxygen and has a column density of about 8 × 1018 molecules of ozone cm−2 . However, a billion years ago, the atmosphere had only about 15% today’s oxygen concentrations (or about 3% oxygen) [12], giving it an ozone column density of about 1.2 × 1018 molecules cm−2 . Summing the photon and energy ﬂux at sea level a billion years ago, we ﬁnd that it was about 2200 ergs cm−2 s−1 and the incident photons were primarily of wavelengths between 280 and 300 nm [9]. Another factor that must be taken into account is the changing solar spectrum over time. As noted above, the Sun has become steadily hotter as it has evolved, emitting correspondingly higher levels of UV radiation. We assume that the early (4 Ga) Sun was 70% as luminous as today’s Sun and that solar luminosity and temperature increased in a linear fashion in the intervening time [8]. The changes in solar luminosity were used to derive the Sun’s surface temperature that was, in turn, used to determine the ﬂux of UV photons in each wavelength band at that time in the past. These calculations show that the early earth was subjected to signiﬁcantly higher levels of ionizing UV radiation prior to the formation of the ozone layer. This conclusion is hardly a surprise and has, in fact, been reached by many in the past [14,15]. These calculations also show that UV ﬂux on the earliest earth was approximately 430 times as intense as on Earth today. This value compares favorably with Cockell’s ratio of 470 times the relative DNA damage rates to early life from UV radiation [14]. Cockell also noted that other factors, including other atmospheric gases, particulates, or sulfur compounds may have helped screen early life from UV irradiation, but their presence cannot be known with any degree of certainty. Changes in sea level UV ﬂux are shown in Fig. 2. The DNA-weighted UV irradiance at sea level today is about 1 erg cm−2 s−1 . Therefore, earliest life (if exposed at Earth’s surface) would have received an irradiation of over 400 ergs cm−2 s−1 . This is equivalent to an E. coli mutation doubling dose every quarter second [16], which is much too high for a living organism to survive. Since life exists today, it is evident that early life found a suitably shielded habitat, either in the deep sea, beneath a surface layer of sediments, or some other UV-opaque material. However, it is likely that many

112

P.A. Karam, S.A. Leslie

Fig. 2. Changes in sea level ﬂux of UVB and UVC through time. The sharp drops are due to oxygenation of the atmosphere and the slight rises between these drops in ﬂux are due to the effects of solar evolution.

dominant life forms on the early Earth were photosynthetic, especially after about 3 billion years ago. Accordingly, it is highly probable that these organisms were still exposed to levels of UV radiation signiﬁcantly higher than those experienced by modern organisms [17]. In fact, if we can draw corollaries from similar organisms that are extant today, it is likely that many early photosynthetic organisms lived in waters less than a few meters in depth, placing them well within the zone into which UVB and UVC can penetrate [12,14,18]. Other organisms, such as stromatolites may have evolved the ability to survive in the photic zone by developing UV-resistant outer (and upper) layers that protected the more sensitive cells beneath [12,19]. Survival strategies used by such organisms have been previously discussed in the scientiﬁc literature (see, for example, [15]) and are not discussed further here. Cockell also pointed out that the DNA-weighted UV ﬂux in the ancient oceans, prior to the formation of an ozone layer, would have decreased to roughly present-day surface levels at a depth of about 30 m, while he and others suggest the early oceans may have been covered in part by UV-opaque molecules that would have served to help protect the life beneath [20,21].

2. Radiation dose from episodic cosmic events Some have speculated that occasional events such as supernovae, gamma ray bursts, or solar “super-ﬂares” may have contributed sufﬁcient radiation dose to organisms at Earth’s surface to have caused mass extinctions [22–25]. Calculations suggest otherwise because of the rarity of catastrophic cosmic events near enough to earth to deliver lethal radiation dose to our planet’s surface, but it is still possible for cosmic events to affect terrestrial life [26,27]. Radiation exposure at Earth’s surface from gamma ray bursts is likely to reach levels of about 1 Gy approximately once per species lifetime [26], but supernova emit too little prompt gamma radiation to have a signiﬁcant effect on sea level radiation levels. On the other hand, supernovae produce approximately 1028 kg of radioactive 56 Ni when they explode. The decay of this radioactive nickel to stability (via 56 Co) releases nearly 1043 J of gamma energy over the course of a few years. Although over 90% of this energy is absorbed by the supernova remnant, the remainder escapes into space and can produce a radiation dose of up to 1 Gy at Earth’s surface with a mean interval of about 5 million years [27]. Two lines of evidence point to periodic solar “superﬂares” occurring on similar time scales [4,28]. Superﬂares emit substantially more energy than normal solar ﬂares and are also capable of producing very high (about 1 Gy) radiation levels at sea level approximately

Changes in terrestrial natural radiation levels over the history of life

113

once per species lifetime. Accordingly, we feel is safe to say that the vast majority of organisms will never experience such high radiation levels, but each species will. Because of this, it is possible that even such infrequent events can contribute towards species maintaining the ability to repair radiation damage in spite of steadily declining terrestrial levels through time. We also note that radiation dose in space from these events is likely to be sufﬁciently high as to place some constraints on the ability of living organisms to travel between planets or between planetary systems. In order to survive a journey of several tens of millions of years, organisms would need to reside in bodies of rock or ice at least several tens of cm in diameter [26,27]. Bodies of this size are relatively rare, suggesting that interplanetary “seeding” may be similarly rare.3 It is worth noting that there is a high probability that nearby supernovae have frequently produced sufﬁcient UV ﬂux to have had a major impact on terrestrial life. Due to intense UV emissions from supernovae, it is probable that Earth experiences a sea level UV dose sufﬁcient to cause a doubling of the background mutation rate with a mean interval of about 200 000 years (Scalo, personal communication). The mean interval between events that would contribute a UV ﬂux that is greater than that of the Sun is only a few thousand years.

3. Modifying effects of changes in the terrestrial atmosphere The terrestrial atmosphere has changed in composition and quantity since Earth ﬁrst formed. Of all the atmosphere’s constituents, the most important from the standpoint of this research is the presence of free oxygen in the atmosphere because of oxygen’s role in forming the ozone layer and its role as a powerful modiﬁer of radiogenic DNA damage. The following discussion summarizes more detailed information found in [3]. The earliest atmosphere was likely anoxic, and oxygen levels increased in steps at distinct times in the past. The ﬁrst step occurred about 3 billion years ago, when oxygenic photosynthesis was “invented” by early cyanobacteria [29]. Although the great majority of this oxygen reacted with reduced ions in Earth’s oceans, atmosphere, and crust, it is likely that trace amounts remained in the atmosphere and that larger amounts were dissolved in seawater in “oxygen oases”; isolated or restricted basins that were not in equilibrium with the bulk ocean. Accordingly, although background radiation levels were nearly ten times those found today, radiogenic DNA damage rates were only about 2 to 2.5 times present levels. About 2 billion years ago, when the reduced ions had been oxidized, oxygen began to accumulate in the atmosphere and oceans, leading to a sharp increase in radiogenic DNA damage. This damage decreased with decaying radiation levels until the next sharp increase in atmospheric oxygen levels about 440 million years ago, when plants began colonization of land. From that time until the present, radiogenic DNA damage rates have dropped steadily along with background radiation levels, as shown in Fig. 3. 3 Consider, for example, that the putative martian meteorite ALH84001 was in space for about 16 million years to

travel the 75 million km from Mars’ orbit to that of the Earth.

114

P.A. Karam, S.A. Leslie

Fig. 3. Changes in atmospheric oxygen concentrations and radiogenic DNA damage rates through time.

The effects of increasing atmospheric oxygen levels on sea level UV ﬂux, discussed earlier, was even more profound.

4. Other environmental mutagens There are, of course, many environmental mutagens, all of which have changed through time. These include the evolution of eukaryotic life, which exposed cells to the leakage of free radicals from mitochondria, changes in dissolved oxygen levels, which led to oxidative DNA damage even in the absence of radiation, and the likely presence of other, as yet unknown, DNA damaging agents in the ancient oceans. To fully understand the response of modern organisms to DNA insult, we must better understand how all of these factors have changed with time. We have made preliminary efforts to do so [5], and are now undertaking more detailed calculations. We also realize that not all organisms will be exposed to all mutagens, so it may also be necessary to determine the “lifestyle” of a particular organism and its evolutionary antecedents in order to better appreciate the manner in which it will respond to DNA damage. This is obviously far beyond the scope of this paper, and will make for many years of fascinating research. The results of our initial work are presented in Fig. 4 and in previous work [6].

Fig. 4. Changes in environmental mutagens and mutation rates through time [5].

Changes in terrestrial natural radiation levels over the history of life

115

5. Discussion and conclusions Many of the DNA damage repair mechanisms in modern organisms are conservative [30]; they are very similar in widely disparate kingdoms such as eubacteria, archaebacteria, and animalia (the animal kingdom). This observation supports the idea that such repair mechanisms evolved only once in the common ancestor to all modern life forms, before life diverged to form the modern kingdoms. If this is the case, then the mutation repair mechanisms in humans likely has its evolutionary roots in an environment that was far more mutagenic than today’s, and may have retained the ability to successfully repair levels and rates of DNA damage in excess of those found today. This possibility gives a geological and historical context to the idea of a threshold level, below which life’s cellular repair mechanisms can adequately repair radiation damage with little or no expected harmful effects to the organism. The discovery that adaptive response to radiation can be induced by exposure to elevated levels of background radiation, such as those found in Ramsar, Iran [31], lends some credence to this speculation. The earliest life consisted of prokaryotes, small cells with few organelles and no nucleus. Some prokaryotes (archaebacteria) can live in extreme environments not inhabited by other kingdoms of life, although organisms from these domains of life are common in virtually all environments on Earth. Eukaryotes are thought to have evolved about 2 billion years ago [32] as the result of symbiosis between prokaryotes [33]. Eukaryotes are larger and more complex than prokaryotes, containing numerous organelles and a nucleus containing genetic information. Fungi, protists, plants, and animals are composed of eukaryotic cells. Eukaryotic cells are generally more sensitive to the effects of radiation than are prokaryotes. This sensitivity could reﬂect the greater complexity of eukaryotic cells, making their functioning easier to disrupt; the relatively greater chromosome volume, giving more “targets” for mutating events; their evolution in a background radiation ﬁeld lower than that in which prokaryotes evolved; some combination of these; or another cause entirely. In any event, it is interesting to note that many mutation repair mechanisms are shared by both prokaryotic and eukaryotic cells, while others are unique to each kingdom of life [34]. The United Nations Science Committee on the Effects of Atomic Radiation published a table showing lethal radiation dosage ranges for a number of phyla [35]. In general, more recently evolved organisms show higher sensitivity to the effects of radiation as evidenced by lower lethal dose ranges. It is tempting to claim that this demonstrates the evolution of these organisms in successively lower radiation ﬁelds over time. However, with the possible exception of eukaryotes noted above, this is likely an oversimpliﬁcation because multi-cellular life forms evolved quickly with respect to the half-life of 40 K (the primary source of radiation exposure), so dose rates did not change signiﬁcantly between the evolution of molluscs (about 550 million years ago) and the ﬁrst mammals (about 200 million years ago). Halliwell and Aruoma [36] similarly note that in vivo • OH formation resulting from exposure to background radiation and reaction of H2 O2 with metal ions may have helped stimulate the development of some DNA repair systems. If this is the case, it may be possible for these repair mechanisms to have retained the ability to efﬁciently and accurately repair higher rates of DNA damage than exist at present. In addition, there may be a qualitative difference between DNA damage from even low-LET radiations and from other mutagens [37,38]. This suggests that it is possible that the response

116

P.A. Karam, S.A. Leslie

of modern organisms to radiation insult may, in part, reﬂect the mutagenic environment in which these repair mechanisms evolved. However, in order to fully assess what capabilities current repair mechanisms possess, it is necessary to develop a more complete understanding of the environment in which they have evolved since ﬁrst appearing in our most ancient ancestors. For that reason, it is possible that the changes in natural radiation levels through the history of life on Earth may help in better understanding the ability of modern organisms to repair mutations due to exposure to background or anthropogenic radiation.

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35]

P.A. Karam, S.A. Leslie, Health Phys. 77 (6) (1999) 662–667. O’Brien, Health Phys. 36 (1) (1978) 63–65. P.A. Karam, Radiat. Phys. Chem. 64 (2002) 77–87. P.A. Karam, et al., Health Phys. 81 (5) (2001) 541–552. P.A. Karam, S.A. Leslie, Paleobios 21 (2) (2001) S77. M. Eisenbud, T. Gesell, Environmental Radioactivity, Academic Press, 1997. UNSCEAR, Sources and Effects of Ionizing Radiation, United Nations, New York, 1958. Bahcall, et al., Astrophys. J. 555 (2001) 990–1012. P.A. Karam, Health Phys. 84 (3) (2003) 322–333. P.A. Karam, Unpublished PhD dissertation, 2001. Sackman, et al., Astrophys. J. 418 (1993) 457–468. G.B. Kasting, C. Chang, in: J.W. Schopf, C. Klein (Eds.), The Proterozoic Biosphere: A Multidisciplinary Study, Cambridge Univ. Press, New York, 1992, pp. 9–12. A.N. Cox, Allen’s Astrophysical Quantities, Springer-Verlag/AIP Press, New York, 2000. Cockell, J. Theor. Biol. 193 (1998) 717–729. Cockell, Planet. Space Sci. 48 (2000) 203–214. Scalo, Wheeler, in: L.M. Celnekier (Ed.), Frontiers of Life, 12th Recontres de Blois, 2001. G.B. Kasting, et al., in: J.W. Schopf, C. Klein (Eds.), The Proterozoic Biosphere: An Interdisciplinary Study, Cambridge Univ. Press, New York, 1992, 1992, pp. 159–163. Cockell, Evolution, and UV Radiation, Springer-Verlag, New York, 2001. Rambler, Margulis, Science 210 (1980) 638–640. Cleaves, Miller, Proc. Natl. Acad. Sci. 95 (1998) 7260–7263. Cockell, Knowland, Biol. Rev. 74 (1999) 311–345. Terry, Tucker, Science 159 (1968) 421–423. Ellis, et al., Astrophys. J. 470 (1996) 1227. Annis, J. British Interplanet. Soc. 52 (1999) 19–22. Scalo, Wheeler, Astrophys. J. 576 (2001) 723–737. P.A. Karam, Health Phys. 82 (4) (2002) 491–499. P.A. Karam, Health Phys. Radiat. Phys. Chem. 64 (2002) 77–87. Schaeffer, Astrophys. J. 529 (1999) 1026–1030. Dismukes, et al., Proc. Natl. Acad. Sci. 98 (5) (2001) 2170–2175. Mackindon, James, Health Phys. 59 (1) (1990) 29–34. O. Ghiassi-nejad, et al., Health Phys. 82 (1) (2002) 87–93. Runnegar, Early life on Earth, in: Nobel Symposium No. 84, Columbia Univ. Press, New York, 1994. Margulis, Sagan, Microcosmos: Four Billion Years of Microbial Evolution, Simon and Schuster, New York, 1986. N. Prasad, Handbook of Radiobiology, second ed., CRC Press, Boca Raton, FL, 1995. UNSCEAR, Sources and Effects of Ionizing Radiation, Report to the General Assembly, United Nations, New York, 1996.

Changes in terrestrial natural radiation levels over the history of life

117

[36] Halliwell, Aruma, Federat. Eur. Biochem. Soc. 281 (1,2) (1991) 9–19. [37] Milligan, et al., Mechanisms of DNA Strand Breakage by Ionizing Radiation, Battelle Press, Columbus, OH, 1995. [38] Ward, et al., in: J.D. Zimbrick, O. Fuciarelli (Eds.), DNA: Structure/Function Relationships at Early Times, Battelle Press, 1995.

118

Gamma ray direction ﬁnder K. Fujimoto, Y. Noda National Institute of Radiological Sciences 9-1, Anagawa-4, Inage-ku, Chiba 263-8555, Japan

New types of detector were designed in the present paper to acquire directional information of incident photons. A wind rose-type detector with dozens of scintillators and a position sensitive photo multiplier tube (PSPMT) is in prospect although the detector could identify only lateral directions. One of the simplest and cheapest detector systems is designed to be one pair of NaI(Tl) scintillators with a triangle shape lead shield. This type of detector could be developed using currently available equipment. The disadvantage of this detector is that it could identify only one dominant incident angle. Among the several detectors designed, a spherical honeycomb-type detector made of many pieces of scintillator, light guides and a PSPMT could be ideal for this purpose. The gamma ray direction ﬁnder would be applicable in various ﬁelds where directional information of incident photons is needed.

1. Introduction Conventional gamma ray detectors are designed not to have angular dependence in their response. Especially for environmental gamma ray monitoring an isotropic response is desirable. When the detector is coupled with a multi-channel analyzer the energy spectrum can then be obtained. However, at present no detectors are capable of identifying the direction of incident gamma rays. Only one type of detector that is recently commercially available has a mono-directional response to gamma rays and can display the radioactive source position on a video image. This detector, however, is a single channel detector in the sense of directional measurements. If directional information of incident gamma rays is sought with this system, the detector has to be rotated to other angles for additional measurements. If one measurement could provide directional information of incident gamma rays, it would be very useful for the location of contamination, to search for missing radioactive sources, in identiﬁcation of highly radioactive areas, in explaining rate differences often found by car-borne measurements, etc. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07012-3

© 2005 Elsevier Ltd. All rights reserved.

Gamma ray direction ﬁnder

119

2. Approach Several detectors designed to acquire directional information on incident gamma rays are presented in this paper. The new type of detectors is called here the “Gamma Ray Direction Finder” named after the detector system built to reveal the direction of radio waves. There are two ways to obtain directional information. One is the measurement with coincidence method, and the other is with shielded detectors. 2.1. Coincidence method Small detectors are placed around a large detector and are used to identify the incidence of gamma rays. The small detectors provide the directional information of incident gamma rays and the larger detector provides the energy spectrum of the incident gamma rays. A glass ﬁber looped around the surface of a spherical scintillation detector such as NaI(Tl) could tell the location where the interaction of an incident gamma ray occurred by the time difference of the light signals reaching the two ends of the glass ﬁber. When the glass ﬁber illuminates the detector system opens the gate for the inner detector to accumulate the gamma ray energy. One of the examples of this detector is shown in Fig. 1 and labeled “C1.” Photodiodes placed around the surface of a spherical detector would also work like the glass ﬁber detector as mentioned above. An image of this type of detector is shown as C2 in Fig. 1. The energy spectra are summed as a function of direction of incident gamma ray. After a certain time period of measurement an incident gamma ray spectrum in each direction would be obtained. The drawbacks of these systems are the small gamma interaction probability and the energy loss in the glass ﬁber or photodiodes. 2.2. Shielded detector When a detector is covered with shielding material and has an open end only in one direction, the detector can accumulate energy information of the gamma rays coming from that direction. S1 in Fig. 1 shows one design of this type of detector system. The shielding material with a hole rotates around a cylindrical detector with a constant speed. The inner detector ac-

Fig. 1. Several designs of gamma ray direction ﬁnder.

120

K. Fujimoto, Y. Noda

cumulates the energy information of the incident gamma rays by the mode synchronized with the rotation of the shielding material. The drawback of this system is the necessity of rotation of the shielding material and limited acquisition of information in lateral directions. Dozens of cuboid scintillators lined around an axis like a wind rose are mounted in shielding material as one of the direction ﬁnder (see S2 in Fig. 1). The top and bottom of the assembly are also shielded. Each detector of the system would have a response only to the gamma ray coming from the open face although it could identify only the lateral direction of incident gamma rays. S3 in Fig. 1 shows an ideal detector system. Dozens of scintillation detectors are assembled in the shape of a spherical honeycomb. Light pipes and a position sensitive photomultiplier tube (PSPMT) are installed inside the assembly. The scintillators are placed one by one and covered with light-tight material to shut out optical signals arising from the adjacent detectors. They could detect all gamma rays coming from any direction. The shortcoming of this system is the difﬁculty of manufacturing the complicatedly shaped scintillators. The cross talk among scintillators may also cause problems. S4 in Fig. 1 shows a more realistic design of gamma ray direction ﬁnder. Limited numbers of cylindrical type scintillation detectors are installed in a spherical shaped shielding material. Each detector has only one open face. The light pipes and PSPMT are connected to each detector in the center of the assembly. More detailed design is discussed below.

3. Design/calculation Only design S4 is considered in detail in the present paper although several other detector systems may have good possibilities. The response of each detector and the total response of the system are calculated as a function of incident gamma ray angle. The response of the ideal detector system should be constant for any incident gamma ray coming from all directions. A relatively large number of detectors should be placed on the horizontal plane since the dominant number of gamma rays comes from the horizon in the normal radiation ﬁeld [1]. Eight might be an appropriate number of detectors installed on the horizontal plane taking into account the angular resolution of each detector (they are named in the present paper as H1 to H8 as shown in Fig. 2). On the top and bottom of the sphere of the system one detector should

Fig. 2. Positioning of detectors in design S4.

Gamma ray direction ﬁnder

121

be installed (T1 and T2). Between the top or bottom and the horizon another set of detectors should be installed. If the detectors are located in the middle longitude between the two detectors that are placed on the horizontal plane, the total number of detectors on the plus and minus 45-degree latitudes is 8 for each (S1 to S8 and S9 to S16). Therefore, the total number of detectors in this system is 26. The 26 detectors are placed as shown in Fig. 2. The spherical coordinates of an incident gamma ray are expressed as (α, β) and corresponding rectangular coordinates are (b1 , b2 , b3 ). The response of each detector is assumed to be proportional to the cross section of the detector to an incident gamma ray since only one face of the cylindrical detector emerges from the shielding material. The cross section can be obtained as a cosine of the angle of two vectors, i.e., vectors representing detector axis and incident gamma ray. The cosine is derived from the scalar product of two vectors as shown scalar product: (a, b) = a1 xb1 + a2 xb2 + a3 xb3 = |a||b| cos ψ, therefore,

cos ψ = a1 xb1 + a2 xb2 + a3 xb3 ,

(1)

where a = (a1 , a2 , a3 ) is a unit vector representing the detector axis, b = (b1 , b2 , b3 ) is a unit vector representing the incident gamma ray, and ψ is angle of detector axis and incident gamma ray. 4. Results and discussion Based on the calculation of each cos ψ mentioned above the response of each detector is obtained for incident gamma rays with several angles. The responses to gamma rays coming from the horizon and at −45◦ of the dip of the horizon (3π/4) are shown in Figs. 3 and 4, respectively, as examples. Negative values shown in Figs. 3 and 4 demonstrate that those detectors are located on the backside of the detector system from the view of the incident gamma ray. A total response of the detector system is shown in Figs. 5 and 6, that is the sum of each detector’s response. Negative values as shown in Figs. 3 and 4 are not included. Each line shows the response to the gamma ray coming from the direction (α, β). The ﬂat line and overlapping of all lines are an ideal response. In order to have ﬂat responses in Fig. 5 and to have smallest differences among the lines for each incident gamma ray angle (β in Fig. 6), the weighting factors√ of the Si detectors are adjusted. The most appropriate weighting factor is estimated to be 1/ 2 for Si. The weighting factor of the other detectors (T1, T2 and Hi) is 1.0. Then the directional dependence has been reduced to be within 10% for any incident gamma ray as shown in Figs. 7 and 8. About 8–11 detectors can be seen from any direction when the 26 detectors are installed as shown in Fig. 2. However, the effective number of detectors that can be used in a measurement for gamma √ rays coming from one direction is 6 to 7 as shown in Fig. 5. When the weighting factor of 1/ 2 is applied to Si, the effective number of detectors is reduced to 5–5.5. After the completion of measurement all information obtained by the 26 detectors is deconvoluted to estimate the concentration distributions around the detector system. Three-dimensional concentration maps for several identiﬁed radionuclides, such as 40 K, U, and Th series, and Cs, could be obtained using the same kind of methodology applied in computer tomography (CT) that is commonly used in the medical ﬁeld.

122

K. Fujimoto, Y. Noda

Fig. 3. Response of each detector to gamma rays of β = π/2.

Fig. 4. Response of each detector to gamma rays of β = 3π/4.

Gamma ray direction ﬁnder

123

Fig. 5. Detector system response to gamma ray (α, β) as a function of β.

Fig. 6. Detector system response (Y axis) to gamma ray (α, β) as a function of α.

5. Conclusion The design S4 shown in Figs. 1 and 2 having 26 detectors mounted in a spherical shielding material would be a promising detector for a gamma ray direction ﬁnder. Directional dependence could be adjusted by the weighting factors of Ti, Hi, and Si. The most appropriate

124

K. Fujimoto, Y. Noda

Fig. 7. Detector system response to gamma ray (α, β) as a function of β after weighting adjustment.

Fig. 8. Detector system response to gamma ray (α, β) as a function of α after weighting adjustment.

√ weighting factors are 1.0, 1.0 and 1/ 2 for Ti, Hi, and Si, respectively. The directional dependence could be conﬁned within 10% as a total system. When an actual detector system is manufactured the response of the system should be checked by calibration sources with several energies. Then the weighting factors could be further adjusted to have a ﬂat response. The gamma ray direction ﬁnder would be a useful and innovative tool for the location of contam-

Gamma ray direction ﬁnder

125

ination, searching for missing radioactive sources, identiﬁcation of highly radioactive areas, investigating of dose rate differences often found by car-borne measurement, and so on.

Reference [1] K. Fujimoto, S. Kobayashi, Shielding effect of snow cover on indoor exposure due to terrestrial gamma radiation, in: Proc. of the 7th International Conf. of the International Radiation Protection Association, 1988, pp. 910–913.

126

Continuous measurement of environmental gamma radiation using a Ge semiconductor detector and 222Rn concentration in air T. Ichiji, T. Hattori Nuclear Energy Systems Department, Komae Research Laboratory, Central Research Institute of Electric Power Industry, 2-11-1 Iwado Kita, Komae-shi, Tokyo 201-8511, Japan

Environmental gamma radiation was measured using a Ge semiconductor detector over a period of about 1 year in Tokyo. Vertical distributions of 222 Rn in air were estimated qualitatively using the ratios of the count rates of 1120 and 1765 keV gamma radiations to that of 609 keV emitted from 214 Bi. Diurnal and seasonal variations of the ratios were investigated. In the case of diurnal variations, the ratios at the break of dawn were small. This indicates that the difference in 222 Rn concentration in air in the vertical direction is relatively large at the break of dawn. As for seasonal variations, the ratios in summer were larger than those in winter. This indicates that the difference in 222 Rn concentration in air in the vertical direction is relatively small in summer.

1. Introduction Variations of 222 Rn concentrations in air are caused by meteorological conditions, for example, air temperature, atmospheric pressure, humidity, insulation, wind direction and wind speed. In the global meteorological condition, 222 Rn concentrations in air increase because 222 Rn originating from the Asia continent is transported by wind and continental air masses. Hattori and Ichiji [1] reported a method of estimating seasonal variations of the 222 Rn originating from Asia, using the correlation between 222 Rn and 212 Pb concentrations in air. One of the most important parameters of local meteorological conditions is the air mixture condition in the vertical direction. The vertical distribution of 222 Rn inﬂuences diurnal variations of 222 Rn in air. Nishikawa et al. [2] measured the height of the temperature inversion layer using acoustic radar and showed the relationship between 222 Rn in air and the height of the inversion layer. Kataoka et al. [3] observed 222 Rn concentrations and meteorological elements at different heights and compared them. Kataoka and co-workers [4,5] measured vertical distributions of 222 Rn progeny concentrations in air on a meteorology tower at heights of 1, 10 RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07013-5

© 2005 Elsevier Ltd. All rights reserved.

Continuous measurement of environmental gamma radiation

127

and 100 m above ground level. Nishikawa et al. [6,7] estimated the gamma radiation dose rate due to atmospheric 222 Rn progeny, using the Monte Carlo calculation. In order to estimate vertical distributions of 222 Rn in air, large-scale measurements requiring the use of, for example, a high building or an airplane have been needed. Continuous measurement is difﬁcult when vertical distributions of 222 Rn in air are measured using an airplane. On the other hand, the coverage area is limited when vertical distributions of 222 Rn in air are measured using a high building. In this paper, an easy method of estimating the vertical distribution of 222 Rn in air using count rates of various gamma radiations emitted from 214 Bi is discussed.

2. Materials and methods Environmental gamma radiation was measured using a Ge semiconductor detector, during the period from July 2000 to August 2001 at our laboratory in the suburbs of Tokyo. 222 Rn concentrations in air were simultaneously measured every hour. Rainfall intensity was also measured every 10 min. In the following sections, the methods of measuring environmental gamma radiation and 222 Rn concentrations in air are described in full detail. 2.1. Measurements of environmental gamma radiation The Ge semiconductor detector was installed outdoors at a height of 1 m above the ground in a double-shielded tent. A schematic diagram of the gamma radiation measurement system is shown in Fig. 1. The relative efﬁciency of the Ge semiconductor detector is 65%. In order to detect gamma radiation from the sky, a 1-m-diameter, 10-cm-thick Pb plate shield was placed under the gamma radiation measurement system and a 5-cm-thick Pb shield was placed around the detector. The atmosphere in the tent was kept at a temperature of about 20 ◦ C for stable operation of the detector, using an air conditioner. The gamma radiation energy spectra were automatically recorded on a hard disk of a personal computer every 10 min. The regions of interest (ROI) were set at 352 keV gamma radiation emitted from 214 Pb and 609, 768, 1120, 1238, 1764 and 2204 keV gamma radiation emitted from 214 Bi. The count rates of gamma radiation, the energy resolutions and the peak energy channels were simultaneously recorded on the hard disk. If problems arise, it is possible to check the energy resolutions and the drift of the peak energy channels. An electric refrigerator was used to cool the detector. It was not necessary to supply liquid nitrogen manually. 2.2. Measurement of 222 Rn concentrations in air Hourly data of the 222 Rn concentrations were measured with an electrostatic monitor using the method reported by Iida et al. [8]. The 222 Rn concentration monitor is shown in Fig. 2. The 222 Rn concentrations in air were measured at a height of 1.5 m. The air sample was pumped into the 15 L vessel through a membrane ﬁlter to remove the 222 Rn progeny at a ﬂow rate of 0.3 L min−1 . The 222 Rn in the air sample decays to 218 Po by emitting an alpha particle. The positive 218 Po ions are collected electrostatically on an electrode of aluminized Mylar covering a ZnS(Ag) scintillator. A photomultiplier tube detects the scintillations due

128

T. Ichiji, T. Hattori

Fig. 1. Schematic diagram of gamma radiation measurement system.

Fig. 2. Monitor of 222 Rn concentration.

Continuous measurement of environmental gamma radiation

129

to alpha particles emitted from the 222 Rn progeny. The scintillation pulse is ampliﬁed and fed into a personal computer. The air sample is dehumidiﬁed to decrease the relative humidity to lower than 1% before passing it through the ﬁlter, since the collection efﬁciency of the 218 Po ions depends on the humidity in the vessel. The mean concentration of 222 Rn for each hour is calculated, using alpha counts at one-hour intervals, by compensating the counts due to the 222 Rn progeny accumulated during the previous 5 h. The sensitivity of the monitor is about 1 Bq m−3 which is the concentration level at which the relative statistical error is 50%. The ﬂow rate of the sampled air is sufﬁciently low to eliminate the effect of 220 Rn. The 222 Rn concentration in the vessel follows the outside one with a delay of 30 min if the outdoor concentration changes markedly. In the case of a decrease of the ﬂow rate or an increase of the humidity, the monitor stops the measurement and displays an alarm message.

3. Results and discussion Rn-222 concentrations in air and environmental gamma radiation were measured continuously for about 1 year during the period from July 2000 to August 2001. Environmental gamma radiation increases if it rains, since 222 Rn progeny, which come from the air and clouds and are contained in rainwater, accumulate on the ground. In order to eliminate the increase of gamma radiation due to rain, data from the beginning of rainfall until three hours after the end of rainfall were removed. Relationships between 222 Rn concentrations in air and gamma radiation count rates of 222 Rn progeny are shown in Fig. 3. Correlation coefﬁcients were large between 222 Rn concentrations in air and count rates of gamma radiation whose transition probabilities (probability of photon emission per disintegration) are more than 10% for emission from 214 Bi (609, 1120, 1765 keV). Correlation coefﬁcients and transition probabilities of each occurrence of gamma radiation emitted from 214 Bi are shown in Table 1. Diurnal variations of 222 Rn concentrations in air and count rates of gamma radiation emitted from 214 Bi were averaged for each season. For example, diurnal variations in winter are shown in Fig. 4. Diurnal variations of count rates of gamma radiation whose transition probabilities are more than 10% for emission from 214 Bi showed a similar tendency those of 222 Rn concentrations in air. Rn-222 concentrations in Table 1 Gamma radiation energy, transition probability and the correlation coefﬁcients between 222 Rn concentrations in air and gamma radiation count rates of 222 Rn progeny Nuclide

Energy (keV)

Transition probability (%)

Correlation coefﬁcient

214 Pb

352 609 768

35.8 44.8 4.8

0.67 0.74 0.20

214 Bi

1120 1238 1764 2204

14.8 5.9 15.4 4.9

0.46 0.22 0.59 0.28

130

T. Ichiji, T. Hattori

Fig. 3. Relationships between 222 Rn concentrations in air and count rates of gamma radiation of 222 Rn progeny.

air and the count rates of gamma radiation emitted from 214 Bi were high at the break of dawn and low during the daytime. Typical vertical distributions of 222 Rn in air are shown in Fig. 5. Rn-222 is stored at low altitude at the break of dawn because a temperature inversion layer is formed. Rn-222 diffuses to high altitude during the daytime as the temperature inversion layer disappears after dawn. The difference in 222 Rn concentration in air in the vertical direction is larger at the break of dawn than during the daytime. Ratios of the count rates of gamma radiation of 1120 and 1765 keV to that of 609 keV were calculated. Vertical distributions of 222 Rn in air were estimated qualitatively using the ratios. Ratios of the count rates of high-energy gamma radiation to that of 609 keV are relatively large, if the difference in 222 Rn concentration in air in the vertical direction is relatively small, because lower energy gamma radiation is more easily absorbed by air than higher energy gamma radiation. The ratios at 0, 6, 12 and 18 h are shown in Tables 2 and 3 for each season. In the case of diurnal variations, the ratios at 6 h (at the break of dawn) were small in each season. This indicates that the difference in 222 Rn concentration in air in the vertical direction is relatively large at the break of dawn. As for seasonal variations, the ratios in summer were larger than

Continuous measurement of environmental gamma radiation

131

Fig. 4. Diurnal variations of 222 Rn concentrations in air and count rates of gamma radiation of 214 Bi in winter.

those in winter. This indicates that the difference in 222 Rn concentration in air in the vertical direction is relatively small in summer. The inﬂuence of height on the count rates measured with the Ge semiconductor detector was calculated taking into account the efﬁciency of the detector and the transition probability of each gamma radiation from 214 Bi. Figure 6 shows the calculated arbitrary count rate of the gamma radiation from 214 Bi existing at each altitude measured with the Ge semiconductor detector. The gamma radiation from 214 Bi existing above the height of several hundred meters is negligible. The inﬂuence of the plate-out activity of the Rn progeny on the surface of the tent was calculated using the plate-out velocity of 2 m h−1 reported [9]. The count rate of γ-radiation from the plate-out 222 Rn progeny is about one percent of that of the total γ-radiation.

132

T. Ichiji, T. Hattori

Fig. 5. Vertical distribution of 222 Rn in air at break of dawn and during the daytime. Table 2 Ratios of the count rates of 1120 keV gamma radiation to that of 609 keV

0h 6h 12 h 18 h

Spring

Summer

Autumn

Winter

0.44 ± 0.018 0.39 ± 0.015 0.46 ± 0.021 0.49 ± 0.024

0.58 ± 0.029 0.47 ± 0.021 0.60 ± 0.029 0.49 ± 0.026

0.49 ± 0.024 0.40 ± 0.016 0.54 ± 0.026 0.56 ± 0.031

0.41 ± 0.013 0.37 ± 0.011 0.41 ± 0.016 0.42 ± 0.018

Table 3 Ratios of the count rates of 1765 keV gamma radiation to that of 609 keV

0h 6h 12 h 18 h

Spring

Summer

Autumn

Winter

0.42 ± 0.013 0.38 ± 0.011 0.40 ± 0.015 0.48 ± 0.018

0.49 ± 0.021 0.43 ± 0.015 0.53 ± 0.022 0.49 ± 0.020

0.49 ± 0.019 0.38 ± 0.012 0.49 ± 0.019 0.48 ± 0.022

0.41 ± 0.010 0.36 ± 0.0086 0.41 ± 0.012 0.44 ± 0.014

Continuous measurement of environmental gamma radiation

133

Fig. 6. Calculated arbitrary count rate of gamma radiation from 214 Bi existing at each altitude measured with the Ge semiconductor detector.

4. Conclusion Environmental gamma radiation was measured using a Ge semiconductor detector over a period of about 1 year in Tokyo. Vertical distributions of 222 Rn in air were estimated qualitatively using the ratios of the count rates of 1120 and 1765 keV gamma radiations to that of 609 keV emitted from 214 Bi. Diurnal and seasonal variations of the ratios were investigated. In the case of diurnal variations, the ratios at the break of dawn were small. This indicates that the difference in 222 Rn concentration in air in the vertical direction is relatively large at the break of dawn. As for seasonal variations, the ratios in summer were larger than those in winter. This indicates that the difference in 222 Rn concentration in air in the vertical direction is relatively small in summer. The vertical distributions of 222 Rn in air could be estimated qualitatively using gamma radiation measured with a Ge semiconductor detector at the ground level. Quantitative estimation could be accomplished in the further study.

References [1] T. Hattori, T. Ichiji, Estimates of seasonal variations of 222 Rn from different origins by using the correlation between 222 Rn and 212 Pb concentrations in air, in: Radon and Thoron in the Human Environment, 1997, pp. 259– 264. [2] T. Nishikawa, M. Aoki, S. Okabe, Relation between diurnal variation of atmospheric radon daughter concentration and height of inversion layer, in: Atmospheric Radon Families and Environmental Radioactivity, vol. II, 1990, pp. 135–142. [3] T. Kataoka, Y. Ikebe, M. Shimo, T. Iida, K. Ishida, S. Minato, Inﬂuence of short-lived radon-222 daughters present in atmosphere on natural environmental gamma-radiation ﬁeld, J. Nucl. Sci. Technol. 19 (1982) 831– 836. [4] T. Kataoka, E. Yunoki, K. Michihiro, H. Sugiyama, K. Matsunaga, H. Tanimoto, T. Ishida, T. Mori, 222 Rn concentration in outdoor air and some meteorological elements at different heights of the ground level above sea level, in: Atmospheric Radon Families and Environmental Radioactivity, vol. II, 1990, pp. 143–149. [5] T. Kataoka, E. Tsukamoto, E. Yunoki, K. Michihiro, H. Sugiyama, H. Shimizu, T. Mori, K. Sahashi, S. Fujii, Variation of 222 Rn concentration in outdoor air due to variation of the atmospheric boundary layer, Radiat. Prot. Dosim. 45 (1/4) (1992) 403–406. [6] T. Nishikawa, M. Aoki, S. Okabe, Monte Carlo calculation of gamma-ray ﬂux density due to atmospheric radon daughters, J. Nucl. Sci. Technol. 26 (1989) 525.

134

T. Ichiji, T. Hattori

[7] T. Nishikawa, M. Aoki, S. Okabe, Calculation of gamma-ray dose rate due to atmospheric radon daughters, in: Atmospheric Radon Families and Environmental Radioactivity, vol. III, 1995, pp. 69–71. [8] T. Iida, Y. Ikebe, K. Tojo, An electrostatic radon monitor for measurements of environmental radon, Res. Lett. Atmos. Electr. 11 (1991) 55–59. [9] L. Morawska, M. Jamriska, Deposition of radon progeny on indoor surfaces, J. Aerosol Sci. 27 (2) (1996) 305– 312.

135

Uranium concentrations and isotopic ratios in Austrian drinking waters M. Gegner, K. Irlweck Institute for Inorganic Chemistry, University of Vienna, Währingerstr. 42, A-1090 Vienna, Austria

With regard to the discussion about limits for the natural uranium concentration in public drinking water, our investigations present a ﬁrst survey of the situation in Austria. From different locations ∼ 1.5 L samples of tap water were collected. Adding 232 U as a spike uranium was separated by anionic exchange and measured α-spectrometrically after microprecipitation with neodymium ﬂuoride. Measurements were carried out with a Silicon surface barrier detector, combined with a multi channel analyser system (Canberra, MCA S 100) using counting times of 2000 min. With spike recoveries of 71–100% and a counting efﬁciency of 28.7% a minimum detectable activity of 0.2 mBq 238 U L−1 corresponding to 0.015 μg U(nat) L−1 could be achieved. Our results showed uranium concentrations ranging between less than 0.1 and 17 – in one special case up to 80 – μg U(nat) L−1 depending on the geological background of the water table used for the water supply. Isotopic ratios of 234 U/238 U were generally above 1.0 in some cases, even much higher, viz. up to 3.0. Only for one sample this ratio was 0.8. In conclusion, we can say that the World Health Organization (WHO) demand to set a limit of 2.0 μg U(nat) L−1 would be failed in a lot of regions and could cause serious problems with the public water supply.

1. Introduction Recently different limits for the natural uranium concentration in public drinking water were discussed. Due to the chemical toxicity of uranium the kidneys become the critical organ [1]. The most stringent limitation was derived from the WHO [2], which demanded 2.0 μg U(nat) L−1 . In different countries [3–5] such limits ranged from 20 μg L−1 in Canada up to 100 μg U(nat) L−1 in Finland, where additionally an activity concentration limit of 4 Bq L−1 was set, taking into account enhanced 234 U concentrations, which need not necessarily be in secular equilibrium with 238 U. Very often isotopic ratios of 234 U/238 U much higher than 1.0 were found. In the USA [3] the US Environmental Protection Agency (2000) recommended a maximum concentration level of 30 μg L−1 for naturally occurring uranium. This limit was RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07014-7

© 2005 Elsevier Ltd. All rights reserved.

136

M. Gegner, K. Irlweck

set explicitly due to economic reasons after considering the costs of mitigating wide spread natural concentration levels instead of the 20 μg L−1 proposed previously. The scope of our work was to get a ﬁrst survey of the situation in Austria. Therefore, we collected drinking water samples from different regions, especially from those where higher uranium concentration levels could be expected because of the geological background.

2. Experimental methods About 1–2 L of drinking water samples were taken mainly from public water supplies and in some cases also from private wells. These locations are shown on a geological map in Fig. 1. For our samplings polyethylene bottles were used without adding any acid. Some tests on absorption effects on this material showed that sampling can be carried out in this way without any losses of uranium, if these bottles are rinsed in the laboratory later on with half-diluted concentrated nitric acid. These washings were combined with the bulk of the sample in a 3 L beaker and evaporated to dryness after adding 232 U as a spike. Any organic material was destroyed by dissolution of this residue and heating with HNO3 and H2 O2 . After evaporation to dryness the residue was transformed into the chlorides with concentrated hydrochloric acid and dissolved in 8 M HCl. Uranium was separated by anionic exchange on Dowex 1 × 2 (100–200 mesh) from 8 M HCl solution [6] and eluted with 0.1 M HCl. Sources for α-spectroscopic measurements were prepared by microprecipitation of the uranium with neodymium ﬂuoride [7]. With counting times of 2000 min using a Si surface barrier detector combined with a multi channel analyser system (MCA S 100, Can-

Fig. 1. Geological map of Austria with sampling stations.

Uranium concentrations and isotopic ratios in Austrian drinking waters

137

berra) a detection limit of 0.2 mBq 238 U L−1 , i.e., 0.015 μg U(nat) L−1 , could be achieved. Radiochemical yields ranged between 71 and 100% with an arithmetic mean of 88.1 ± 8.7%.

3. Results and discussion Uranium concentrations in drinking water samples from 41 locations are presented in Table 1 together with the measured activity concentrations of 238 U and 234 U as well as the activity ratios of 234 U/238 U. The content of U(nat) was calculated via the factor 12.3 mBq 238 U μg−1 U for the speciﬁc activity of this uranium isotope. Table 1 Uranium concentrations and 234 U/238 U activity ratios of Austrian drinking waters Code No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

Location Vorarlberg Dornbirn Tyrol Fieberbrunn St. Johann Lienz Salzburg Böckstein/Kurhaus Evianquelle Dienten Saalfelden Salzburg Stadt Upper Austria Ansfelden Burgkirchen Frankenmarkt Lindach Munderﬁng Ranshofen Spital/Pyhrn Carinthia Hermagor Mallnitz Oberdrauburg Obervellach Villach Styria Judenburg Leoben/Goess Mönichwald Spital/Semm. Unterreith

Date of Volume % spike sampling (L) recovery

238 U

(mBq L−1 )

234 U

(mBq L−1 )

U(nat) (μg L−1 )

Ratio (234 U/238 U)

2001 04 14

2.0

78.3

27.2 ± 0.7

27.5 ± 0.7

2.21±0.06

1.01±0.03

2001 07 18 2001 07 18 2001 07 20

2.11 2.11 1.03

88.4 102.4 71.3

2.1 ± 0.3 3.5 ± 0.3 6.9 ± 0.8

3.9 ± 0.4 9.7 ± 0.6 9.5 ± 0.9

0.17±0.02 0.28±0.03 0.56±0.06

1.85±0.28 2.76±0.29 1.38±0.19

2001 07 19 1998 09 25 2001 07 18 2001 07 18 2001 09 15

1.58 1.4 1.96 2.03 1.4

82.6 58.2 98.1 92.2 86.5

975 ± 24 1251 ± 31 199.4 ± 8.0 197.2 ± 8.0 1.7 ± 0.1 2.3 ± 0.1 1.7 ± 0.2 3.1 ± 0.3 2.6 ± 0.3 3.2 ± 0.3

79.2 ± 1.9 16.4 ± 0.7 0.14±0.01 0.13±0.02 0.21±0.02

1.28±0.01 0.98±0.02 1.38±0.12 1.87±0.32 1.26±0.19

2001 06 24 2001 06 24 2001 06 24 2001 06 24 2001 06 24 2001 06 24 2001 07 16

1.48 1.62 1.88 1.62 1.54 1.52 2.0

85.1 70.1 96.6 94.5 92.4 89.5 94.6

23.7 ± 1.2 5.2 ± 0.5 3.5 ± 0.3 8.5 ± 0.6 1.2 ± 0.2 5.3 ± 0.5 1.7 ± 0.2

24.6 ± 1.2 9.3 ± 0.8 4.2 ± 0.4 9.5 ± 0.6 2.1 ± 0.3 8.2 ± 0.6 2.1 ± 0.2

1.92±0.10 0.42±0.04 0.28±0.03 0.69±0.05 0.10±0.02 0.43±0.04 0.14±0.02

1.04±0.06 1.77±0.22 1.21±0.16 1.12±0.10 1.73±0.40 1.55±0.18 1.24±0.19

2001 07 22 2001 07 20 2001 07 20 2001 07 20 2001 07 23

1.58 2.08 1.6 2.0 1.48

94.2 77.4 83.2 78.8 93.4

20.0 ± 0.9 148 ± 4 133 ± 4 20.5 ± 0.5 5.2 ± 0.5

24.5 ± 1.1 161 ± 5 108 ± 3 21.5 ± 0.5 7.4 ± 0.6

1.63±0.08 12.0 ± 0.4 10.8 ± 0.3 1.66±0.04 0.42±0.04

1.22±0.07 1.09±0.02 0.81±0.02 1.05±0.03 1.42±0.17

2001 06 10 2001 07 23 2001 05 23 2001 05 19 2001 07 16

1.53 1.59 2.0 2.0 2.1

88.6 72.6 95.0 93.5 100.8

1.06±0.07 1.70±0.11 0.08±0.01 0.91±0.05 3.51±0.12

0.94±0.08 1.04±0.09 1.66±0.37 1.38±0.09 1.05±0.04

13.1 ± 0.8 20.9 ± 1.4 0.98±0.17 11.2 ± 0.6 43.1 ± 1.4

12.3 ± 0.8 21.7 ± 1.4 1.62±0.22 15.4 ± 0.8 45.1 ± 1.5

(continued on next page)

138

M. Gegner, K. Irlweck

Table 1 (continued) Code No. 27 28 29 30 31 32 33 34 35 36 40 37 38 39 41

Location Lower Austria Amstetten Eisgarn Grein Hirschenwies Hollenstein Kamegg Münchendorf Schottwien Semmering Zellerndorf Waidhofen Burgenland Eisenstadt Rust St. Margareten Halbturn

Date of sampling

Volume % spike (L) recovery

238 U (mBq L−1 )

234 U (mBq L−1 )

U(nat) (μg L−1 )

Ratio (234 U/238 U)

2001 07 07 2001 04 21 2001 07 07 2001 07 22 2001 04 16 1999 06 06 2001 05 25 2001 05 19 2001 05 19 1999 05 23 2002 01 16

1.8 1.56 2.07 2.04 2.0 4.94 2.01 2.0 2.0 5.18 1.60

80.0 75.4 95.0 94.9 80.3 38.8 97.4 88.6 79.7 61.7 86.8

10.9 ± 0.7 13.6 ± 0.8 8.3 ± 0.4 12.1 ± 0.5 46.5 ± 1.6 45.4 ± 1.5 0.72±0.15 1.02±0.18 4.5 ± 0.2 6.0 ± 0.3 51.6 ± 2.6 162.0 ± 7.6 15.7 ± 0.4 24.6 ± 0.5 3.1 ± 0.3 5.1 ± 0.4 12.4 ± 0.7 18.9 ± 1.0 88.0 ± 3.4 96.2 ± 3.7 38.6 ± 1.3 70.2 ± 2.3

0.89±0.06 0.67±0.03 3.78±0.13 0.06±0.01 0.37±0.02 4.19±0.21 1.28±0.03 0.25±0.03 1.00±0.06 7.15±0.28 3.14±0.09

1.25±0.10 1.46±0.09 0.98±0.04 1.42±0.38 1.33±0.08 3.14±0.07 1.57±0.05 1.64±0.21 1.53±0.11 1.07±0.02 1.82±0.04

2001 05 25 2001 05 25 2001 05 25 2002 02 13

2.08 2.09 2.05 1.20

94.6 93.6 92.1 94.9

11.5 ± 0.6 12.0 ± 0.7 8.9 ± 0.3 24.4 ± 0.7

0.94±0.05 0.98±0.05 0.73±0.02 1.96±0.06

1.46±0.07 1.17±0.08 1.47±0.06 1.23±0.04

16.8 ± 0.8 14.1 ± 0.7 13.1 ± 0.4 30.1 ± 1.1

All errors are given as ±1σ standard deviation due to statistical uncertainty of α-counting.

These results show that in general the uranium concentration in Austrian drinking water is signiﬁcantly lower than 20 μg L−1 . Our highest value with 80 μg L−1 for drinking water used in the spa of Böckstein remained as an exception. This water is taken directly from a spring collected inside the “healing gallery”. The Evianquelle, another spring in Böckstein situated outside the gallery somewhat below it, which is used for public water supply, has only a uranium concentration of 16.4 μg L−1 . Whereas drinking water in the south of the Central Alps (Hohen Tauern) at the same altitude as Gastein/Böckstein contains 12 μg L−1 (Mallnitz), while in the north of the river Salzach where a high limestone massif exists, only low uranium concentrations have been found, e.g., Dienten (Hochkönig) 0.14 μg L−1 . In some parts of Lower Austria (Waldviertel), higher concentrations could be expected because of the geological background (Bohemian Massif), and strong local variations have been observed. So a well water in Eisgarn (in the north of Gmünd) shows only 0.7 μg L−1 and a tap water from Hirschenwies (close to Weitra) even less, viz. 0.06 μg L−1 . On the other side in Waidhofen/Thaya 3.1 and samples from Kamegg (well water), Grein and Zellerndorf (in the northern Weinviertel) with 4.5, 3.8 and 7.5 μg L−1 , respectively, are all above the WHO limit, but far below the US EPA recommendation. In special regions, however, like Fieberbrunn in Tyrol and Forstau (Unterreith) in Styria, where uranium deposits are known, uranium concentrations in drinking water are low or only slightly enhanced. Results of water samples from the Vienna Basin can be compared generally with those from the Molasse Zone ranging between 0.1 and 2.2 μg L−1 . Isotopic ratios of 234 U/238 U in the most cases were above 1.0, with the exception of a sample from Oberdrauburg (Carinthia). There, a ratio of 0.81 ± 0.02 was found. In well water

Uranium concentrations and isotopic ratios in Austrian drinking waters

139

from Kamegg we observed the highest value of our measurements, namely 3.03 ± 0.13. Such high isotopic ratios are not unusual for ground waters in certain geological formations, e.g. [8]. In summary, we can say, that the USEPA limit of 30 μg L−1 is hardly ever seen in Austrian water. Following, however, the recommendations of the WHO a lot of implications would occur. In some regions public drinking water supply would be in question. More detailed investigations must clarify the situation for each of the public water resources in some regions. In principle mitigation of the higher uranium concentrations is possible, e.g., by ionic exchange devices, but then new problems would arise. So the waste disposal of the accumulated uranium has to be solved. With regard to all these additional costs for puriﬁcation of drinking water from the involved water supplies in our opinion the implementation of such an extremely low limit must be really justiﬁed.

Acknowledgements This work has been performed under contract No. 353.019/1-X/9/00 with the Federal Ministry of Social Security and Generations. We thank Sekt. Chef Mag. Dr. E. Bobek and Min. Rat Dr. J. Zechner for their interests.

References [1] E.A. Maynard, W.L. Down, H.C. Hodge, Oral toxicity of uranium compounds, in: C. Voegtlin, H.C. Hodge (Eds.), Pharmacology and Toxicology of Uranium Compounds, McGraw–Hill, New York, 1953. [2] WHO, Health criteria and other supporting information, Addendum to vol. 2 in: Guidelines for Drinking-Water Quality, 2nd ed., World Health Organisation, Geneva, 1998. [3] US EPA, Proposed drinking water standards, US EPA 65 FR 76707, US Environmental Protection Agency, December 7, 2000. [4] Fisenne, Environ. Int. 22 (1996) 243. [5] L. Salonen, Finnish Centre for Radiation and Nuclear Safety, personal communication. [6] J. Kritil, J. Mencel, A. Moravec, J. Radioanal. Nucl. Chem. Lett. 21 (1975) 115. [7] D.F. Hindman, Anal. Chem. 55 (1983) 2460. [8] J. Rosholt, J. Geophys. Res. 88 (1983) 7315.

140

Natural radionuclides in radium-rich soils in North-East Estonia E. Realo, K. Realo Institute of Physics, University of Tartu, 142 Riia St., 51014 Tartu, Estonia

Natural radionuclide contents of surface soils were studied in areas near the coast of the Gulf of Finland in NE Estonia, 1998–2001.1 The activity concentrations of radionuclides in soil cores were analyzed by applying laboratory HPGe γ-spectrometry to ﬁnd their depthdependent and area-dependent features in soil. In the region, the average activity concentrations of 40 K, 210 Pb, 226 Ra, 232 Th and 238 U were 564, 95, 90, 32 and 65 Bq kg−1 , respectively. Radon emanation coefﬁcients of soils determined γ-spectrometrically vary in the range from 0.1 to 0.55. Limited areas in the region are characterized by relatively high 226 Ra concentrations up to 325 Bq kg−1 . A notable deﬁciency of 238 U in comparison with 226 Ra and a relatively low ratio of 232 Th/226 Ra in comparison with normal soils were identiﬁed in sites with high 226 Ra concentration. In the region, the mean external γ-dose rate, 138 nSv h−1 , is about 40% higher than the corresponding countrywide mean. Surface soils characterized by a signiﬁcant non-equilibrium of 210 Pb in comparison with 226 Ra were found. The inventories of the unsupported 210 Pb fraction were determined by ﬁtting the activity concentration–depth curves by means of a one-dimensional diffusion model accounting for its unsupported and supported fractions in soil. The average deposition ﬂuxes of 210 Pb ranged from 69 to 177 Bq m−2 a−1 with a mean value of 116 Bq m−2 a−1 . The corresponding 210 Pb concentrations in precipitation range between 102 and 260 Bq m−3 (the mean value is 171 Bq m−3 ). The deposition ﬂuxes were weakly correlated with the 226 Ra activity concentrations in soil. The areas of high soil radium concentration and the alum shale layers rich in uranium are assumed to be the most probable sources of enhanced radon (and 210 Pb) in air of the region.

1. Introduction In the environment radionuclides generate exposure to humans via different pathways, the estimation of which needs reliable data on their contents in soils. In Estonia systematic studies 1 Partially supported by the Estonian Science Foundation Grants 2770 and 4691.

RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07015-9

© 2005 Elsevier Ltd. All rights reserved.

Natural radionuclides in radium-rich soils in North-East Estonia

141

in the ﬁeld were launched, for historical reasons, only in the early 1990s. The territory of Estonia (∼ 45 200 km2 ) is characterized by a complex geological structure and a signiﬁcant variety of soil types with widely varying radionuclide concentrations. Our previous countrywide study from 1991 to 1993 [1–3] identiﬁed a few areas with markedly high 226 Ra activity concentrations in surface soils in NE Estonia. Independent air-borne [4] and car-borne surveys [5] in 1991–1995 helped to specify the areal distribution of high radium regions. Much less is known about the behavior of 210 Pb in these soils. In addition to the natural radiological situation, a relatively heavy atmospheric ﬂy ash deposition load caused by two large oil-shale-ﬁred power plants has an impact on the region [2,6]. A few features of 210 Pb atmospheric deposition near these power plants have been considered in our previous study [7]. We have found sites where the supported 210 Pb fraction dominates over the unsupported fraction in soil and where its signiﬁcant deﬁciencies or above-average depositions occur. Although the highest deposition ﬂuxes were found near the power plants, there was no clear evidence to support the major contribution from 210 Pb deposition emitted with ﬂue gases and ﬂy ash. These results stimulated us to launch additional 210 Pb studies involving natural, including radium-rich soils far from the power plants. Lead-210 (t1/2 = 22.2 a) is formed from the α-decay of its parent 222 Rn (t1/2 = 3.82 d), a decay product of 226 Ra. Because of the 226 Ra recoil, some noble gas radon atoms are released from mineral grains into the pore space of soil, while others remain ﬁxed in soil particles. The released radon atoms are transported by advection and diffusion through the pores, a fraction of them being released into the atmosphere. Both released and ﬁxed radon nuclei undergo a number of decays to produce 210 Pb. Air concentrations of 210 Pb determined by different authors vary over a wide range from 0.1 to 2.5 mBq m−3 with the worldwide average of 0.5 mBq m−3 [8]. Its residence time in the atmosphere is considered to be relatively short, ca. 5 days [9]. Precipitation in the form of rain or snow predominantly governs a reasonably constant annual depositional ﬂux from the atmosphere to the soil. The long-term accumulation compensates losses caused by radioactive decay and by migration to deeper soil layers and generally gives rise to an enhanced (in comparison with 226 Ra) concentration of 210 Pb in the surface horizons of soil. Therefore, in these soil layers the total 210 Pb concentration consists of two components: the atmospherically derived unsupported fraction and the in situ produced supported fraction. Inventories of the unsupported fraction in soil can be directly related to the average annual atmospheric 210 Pb deposition ﬂux. The present paper reports our recent results on the geographical and depth distributions of natural radionuclides analyzed in soil proﬁles collected in the Lääne-Virumaa and IdaVirumaa Counties, NE Estonia, in 1998–2001.

2. Measurements and methods Soil proﬁles from natural uncultivated areas were sampled down to the depth of about 20 cm by means of a special soil corer. The corer design minimized compaction of soil. As a rule, two cores within a distance of 1 m were sampled from each site. Geographical coordinates of the sampling sites were determined using a MAGELLAN GPS receiver. In the laboratory, the cores were sliced into 2 or 3 cm sections. The corresponding sections of two cores were bulked and thoroughly mixed. All samples were dried at 105 ◦ C, homogenized, sieved through

142

E. Realo, K. Realo

2-mm-mesh sieve and stored in sealed 57 cm3 metal containers. Two low-background HPGe γ-spectrometers with a 42% coaxial detector and an 800 mm2 planar detector, respectively, were used for the analyses of 40 K, 226 Ra, 232 Th, 235 U, 234m Pa and 210 Pb, 234 Th. The activity concentration of 238 U was determined as the measured concentrations of both 234 Th and 234m Pa, assuming equilibrium in the 238 U subseries. Lead-210 was analyzed by means of its 46.5 keV line. In low-energy spectrum analysis, a self-attenuation correction method based on a direct 46 or 63 keV transmission measurement for each sample was applied [7]. At the sampling sites, external γ-dose rate measurements were performed using portable dose rate meters. For comparison, dose rates were also evaluated using the concentrations of radionuclides in soil and the corresponding dose conversion factors [8]. Radon emanation coefﬁcients of a few samples from each proﬁle were calculated using time-dependent intensities of the 214 Pb/Bi γ-lines, as described in [7]. The calibration of spectrometers was carried out using IAEA-RG-SET certiﬁed reference materials. As a result of analysis, depth-dependent activity concentrations of 210 Pb and other nuclides i, CPb (x) and Ci (x), respectively, have been determined. Reasonably constant activity concentration values versus depth have been found for the following radionuclides, i = 40 K, 226 Ra, 232 Th and 235 U. Their distributions of activity concentrations, Ci (x) at depth, x, in soil proﬁles were least-square ﬁtted using the expression x Ci (x) = C0,i + C1,i 1 − exp − (1) , l0,i where C0,i = Ci (0), C1,i is the amplitude of the depth-dependent concentration change. The total 210 Pb inventory in a soil proﬁle, CPb (x), to depth, x, has been described as a sum of two components x x + CRa 1 − κ exp − . CPb (x) = Cuns (x) + Csup (x) = Cuns,0 exp − (2) luns lsup The ﬁrst term, an exponential distribution model with relaxation length, luns , for the depthdependent concentration of unsupported fraction, Cuns (x), can be derived assuming a longterm steady-state deposition and 1D migration of 210 Pb from surface to deeper soil horizons. Cuns,0 is the surface concentration of the fraction. We have successfully applied the model in our previous work, when activity concentrations of 226 Ra, CRa , are smaller than 30 Bq kg−1 [7]. In this case the assumption of the approximate equality of Csup (x) ≈ CRa is justiﬁed. The second term for the depth-dependent concentration of supported 210 Pb, Csup (x), follows from a simple stationary radon diffusion and exhalation model in the long-term limit (see, e.g., [10]). This term introduces a depth-dependent decrease of the supported 210 Pb (relative to 226 Ra) in soils with high 226 Ra content. κ represents the radon emanation coefﬁcient and lsup is the diffusive relaxation depth of 222 Rn or the relaxation length of supported 210 Pb. The least-squares ﬁtting procedure for the determined depth distributions of total 210 Pb activity concentrations in soil proﬁles was performed using the ORIGIN5 software. The best-ﬁt parameters were applied to calculate the annual atmospheric deposition ﬂux, F (Bq m−2 a−1 ), of 210 Pb, F = λCuns,0 ρluns ,

(3)

Natural radionuclides in radium-rich soils in North-East Estonia

143

where λ = 0.03122 a−1 is the radioactive decay constant of 210 Pb and ρ is the density of soil, which was found to be constant (within ±7%) along the studied soil proﬁles. Furthermore, considering a relatively small sampling area and the predominant role of wet deposition (rain and especially snow), deposition ﬂuxes per unit precipitation p (m a−1 ) were calculated: P = F /p. P (Bq m−3 ) is actually an average 210 Pb activity concentration in precipitation. A long-term annual mean precipitation rate, p = 0.68 m a−1 , for the eastern Virumaa County has been established [11].

3. Results and discussion 3.1. Potassium-40, uranium and thorium series Fitting equation (1) to i = 40 K, 226 Ra and 232 Th concentration data, the ratio of the two parameters has been found approximately identical: C0 /C1 = (0.94 ± 0.02)/(0.06 ± 0.02), while the relaxation length, l0,i , was 2.6, 2.2 and 5.4 cm, respectively. This result conﬁrms the almost identical concentration distribution with depth of these radionuclides with a relatively small decrease of 6% in the topmost soil layer. At the same time, on average, 238 U shows a homogeneous distribution with depth. Compared to the average distribution the individual depth dependencies, nevertheless, show site-speciﬁc variations up to ±15–20%. The mean activity concentration values with their statistical parameters for 30 surface soil samples of the region are presented in Table 1. In this region, mean concentrations of 226 Ra and 40 K are higher than the worldwide averages of 37 and 520 Bq kg−1 , respectively. The latter practically coincide with the corresponding Estonian averages of 35 and 513 Bq kg−1 . Our results conﬁrm that soils in Estonia are relatively deﬁcient in thorium. The mean value for 232 Th is about 25% lower than its worldwide average, 44 Bq kg−1 [8], but larger than the countrywide mean of 27 Bq kg−1 [3]. The maximum 226 Ra concentrations, reaching up to 325 Bq kg−1 near Sillamäe, are about threefold the regional and up to tenfold the countrywide mean. Relatively high skewness and kurtosis values of all distributions, except 210 Pb, demonstrate the complexity of the distributions. Lead-210 concentration data can be approximated by a normal distribution. A statistical analysis of the depth-averaged activity concentrations of radionuclides from all soil locations indicated a reasonable correlation between 40 K and 232 Th with a coefﬁcient Table 1 Statistical parameters for activity concentrations C (Bq kg−1 dry wt) of radionuclides in surface soils collected in NE Estonia, 1998–2001 Activity concentration C (Bq kg−1 dry wt)

Statistical parameters

Mean Standard deviation Range Kurtosis Skewness

40 K

210 Pb

226 Ra

232 Th

564 275 138–1152 −1.0 0.16

95.1 46.1 37–209 0.1 1.0

90.3 69.0 15–325 2.9 1.5

32.5 15.6 5–62 −1.1 0.29

144

E. Realo, K. Realo

of 0.71 and also between 226 Ra and 232 Th with a coefﬁcient of 0.67. Activity concentrations of 238 U calculated from the 235 U data, assuming the activity ratio of 21.7, almost coincided with those determined from 234 Th and 234m Pa, e.g., for 30 sites, the average ratio of both determinations was 0.98 with a correlation coefﬁcient of 0.95. A few other important features followed from the data analysis: (1) Areas of high radium content are rather limited and are adjacent to “normal radium” areas in the region. Maximum 226 Ra concentrations increase from west to east in these distinct areas. (2) We can distinguish between two different soil “groups” according to their 232 Th/226 Ra activity concentration ratio. With a few exceptions, the radium-rich areas (> 100 Bq kg−1 ) were characterized by ratios in the range 0.18–0.5, while the ratio for “normal radium” soils varies from 0.5–1 [3] (see Fig. 1b). (3) Preliminary estimations show a notable non-equilibrium between the concentrations of 210 Pb, 238 U and 226 Ra. While the 238 U activity concentrations in normal soils are extremely well correlated with those of 226 Ra with a coefﬁcient of 0.95 ± 0.02, radium-rich soils, on average, are depleted in 238 U by 27 ± 4%. It follows from our statistical analysis (see also in Fig. 1a) that in low radium soils the activity concentrations of 238 U and 226 Ra are practically equal. At present, there is no explanation to the observed uranium depletion and further studies are needed. The thorium series is practically in equilibrium. (4) Only limited data are available on the outdoor and indoor [12] radon exposures in these areas. In many countries, the areas with radium-rich soils are considered as radon-prone areas, where speciﬁc additional monitoring and survey are needed.

Fig. 1. Activity concentration (Bq kg−1 ) of 238 U (a) and 232 Th (b) vs. 226 Ra activity concentration for sites in NE Estonia. Equal concentration lines, y = x, are shown as dotted lines in (a) and (b). Solid lines are shown as: (a) a ﬁtted dependence C(238 U) = 0.69C(226 Ra) and (b) a calculated dependence C(232 Th) = 0.18C(226 Ra).

Natural radionuclides in radium-rich soils in North-East Estonia

145

Fig. 2. Map of NE Estonia with sampling sites. Symbol size is proportional to CRa (Bq kg−1 ).

We have calculated the mean external γ-dose rate value of 138 nSv h−1 for NE Estonia, while the dose rate values of individual sites range from 76 to 263 nSv h−1 [3]. For radium-rich areas, the measured values are typically about 10% lower than the calculated ones. In some locations values of 200 nSv h−1 and higher have been measured. For comparison, the countrywide mean value is 94 nSv h−1 [3] and the corresponding worldwide average is 103 nSv h−1 [8]. Approximately 40% of the external exposure of the region is generated by 226 Ra, followed by the cosmic radiation contribution of 25%, whereas, in a countrywide average the contributions from cosmic radiation (37%) and 40 K (27%) are predominant [3]. Figure 2 presents the geographical distribution of the sampling sites on the contour map of NE Estonia, where symbol size is proportional to the site-speciﬁc 226 Ra activity concentration. Because of the coarse sampling grid used, the given 226 Ra distribution is only a rough estimate of the real complex situation. 3.2. Lead-210 and radium-226 A few typical depth-dependent 210 Pb and 226 Ra activity concentrations are plotted in Fig. 3. The 226 Ra concentration, as shown above, is fairly constant along the proﬁle with a small decrease in the topmost layer. An entirely different distribution is characteristic of 210 Pb and signiﬁcant site-speciﬁc differences can be found between the depth-dependent concentration behavior of different sites. A considerable 210 Pb deﬁciency (relative to 226 Ra) is observed in most of the radium-rich proﬁles and its surplus in soils with CRa < 50 Bq kg−1 . All the determined depth-dependent 210 Pb activity concentration curves have been least-square ﬁtted to ﬁnd the unsupported fraction inventories, deposition ﬂuxes and concentrations in precipitation. In Table 2 the average ﬁtting parameters and their ranges for ﬁnding the unsupported 210 Pb fraction are summarized, displaying the occurrence of signiﬁcant site-speciﬁc variations of parameters. From site to site, the relaxation length of supported fraction, lsup , varies in the range of 0.01 to 4 m (the mean is 0.43 m). While the unsupported fraction ﬁtting parameters, Cuns and luns , follow closely the normal distribution, the parameters, lsup , have a notably asymmetric distribution with a kurtosis of 14.8 and a skewness of 3.7. At present, the analysis of uncertainties of the ﬁtting process is in progress. For different sampling sites, 226 Ra activity concentrations vary in a very broad range from 15 to 325 Bq kg−1 . The measured radon emanation coefﬁcients of soils, κ, vary from 0.1 to

146

E. Realo, K. Realo

Fig. 3. The 210 Pb and 226 Ra activity concentrations (Bq kg−1 ) as a function of depth in typical soil proﬁles. Solid lines – total 210 Pb ﬁts using equation (2) (see text).

Table 2 Average ﬁtting parameters for 210 Pb unsupported fraction, assuming equation (2)

Mean Median Standard deviation Range Kurtosis Skewness

Cuns (Bq kg−1 )

luns (m)

lsup (m)

80.0 83.8 25.7 41.7–131.6 −0.64 0.34

0.046 0.045 0.015 0.026–0.08 −0.10 0.81

0.43 0.13 0.89 0.01–4.0 14.8 3.7

0.55 with a mean of 0.29, while the values demonstrate poor correlation with both the density of soil and the CRa values (correlation coefﬁcients of −0.10 and −0.19, respectively). The relaxation lengths of the unsupported fraction, luns , are limited to a relatively narrow region from 2.6 to 8 cm. Considering equation (2), it follows that about 63% of the atmospherically deposited (unsupported) 210 Pb is bound in the upper 3 to 8 cm soil layer. This ﬁnding seems to conﬁrm the attachment of unsupported lead to organic matter in the surface soil [13]. The diffusive relaxation lengths of radon, lsup , which characterize the depth distribution of the supported fraction, vary in a comparatively wide range from 0.01 to 4 m with a mean of 0.43 m. Considering the observed wide variability of soil properties and possible differences in hydrological conditions in different sites and also ﬁtting uncertainties, the result is logical. The value of lsup is found to be strongly dependent on the water content in soil and may range from less than 0.02 m in saturated soils to more than 2 m in well aerated soils (see, e.g., [13]). The results of the calculations using equation (3) are the following. Atmospheric deposition ﬂux values, F , in the range from 69 to 177 Bq m−2 a−1 have been obtained. Their values approximately follow a normal distribution with a mean of Fmean = 116 Bq m−2 a−1 . Using a long-term annual mean precipitation rate for NE Estonia, p = 0.68 m a−1 , the corresponding

Natural radionuclides in radium-rich soils in North-East Estonia

147

Fig. 4. 210 Pb deposition ﬂux F (Bq m−2 a−1 ) vs 226 Ra activity concentration (Bq kg−1 ) in soil for 19 sites, NE Estonia, 1999–2001. Solid line is a ﬁt to the data points, F = 97 + 0.15CRa (the correlation coefﬁcient is 0.39).

Table 3 Comparison of 210 Pb deposition ﬂux and concentration in precipitation data Location Norfolk, USA SE Virginia, USA Devoke Water, UK Harwell, UK Geneva, Switzerland NE Estonia, mean (min/max)

Precipitation (m a−1 ) 1.38 1.39 1.84 0.65 0.96 0.68

210 Pb ﬂux, F

210 Pb concentration, P

(Bq m−2 a−1 )

(Bq m−3 )

152 142 179 54 150 116 (69/177)

110 102 97 83 156 171 (102/260)

Reference [10] [11] [9] [9] [12] This work

concentrations in precipitation, P = F /p, vary in the range of 102 to 260 Bq m−3 and have a mean value of 171 Bq m−3 . Although all the sampling sites are located in the coastal region of the Gulf of Finland (< 10 km from the sea), the 210 Pb deposition ﬂuxes and the distance of the sampling sites from the sea show a rather moderate inverse correlation with a coefﬁcient of −0.28. The ﬂux averaged over the 5 closest locations (< 100 m to the seashore), 92 Bq m−2 a−1 , is approximately 75% of the corresponding mean value. There is also a tendency that higher deposition ﬂuxes are found at sites characterized with higher 226 Ra concentrations in soil (the correlation coefﬁcient is 0.39). The corresponding dependence, F vs CRa , and a linear ﬁt are shown in Fig. 4. In Table 3, the calculated F -values are compared with those found by other authors [13–16]. The comparative material reveals that the mean 210 Pb deposition ﬂux as well as its range in NE Estonia are comparable to or less than the values reported for other countries. At the same time, the 210 Pb concentrations in precipitation are about 1.5 to 2.5 times higher than most of the referenced values. This conclusion seems to support the assumption about locally enhanced 210 Pb (and hence 222 Rn) concentration in air in NE Estonia. Indirect evidence of this assumption can also be found in the results of an atmospheric transport modeling study 210 Pb

148

E. Realo, K. Realo

for air-borne 210 Pb [17], revealing that areas in E Estonia are identiﬁed amongst the relatively intense 210 Pb source regions. The speciﬁc rendsina-type soils are found in the radium-rich areas located near the coast of the Gulf of Finland. These comparatively thin soils overlay the Ordovician carbonate rock layers, which cover the alum shale sedimentary rock, dictyonema argillite, rich in uranium in the background of this region. The most probable intensive radon and 210 Pb sources are the soil areas with a high 226 Ra (238 U) concentration and the uranium-rich alum shale layers. A detailed geographical distribution of these areas has not yet been established. Further analysis based on the present as well as the earlier data presented by the authors should clarify the contributions to the 210 Pb inventory from emissions of ﬂue gases and ﬂy ash by two large oil-shale-ﬁred power plants located in the region.

4. Conclusions In the coastal regions of NE Estonia, the areas of high radium soil content are rather limited and neighbor “normal radium” areas. Maximum 226 Ra concentrations in these distinct areas show the tendency to increase towards the east. While sampling sites cover a wide range of soil types, two different soil “groups” according to their 232 Th/226 Ra activity concentration ratio have been identiﬁed: in radium-rich soils the ratio varies in the range of 0.18 to 0.5 and in “normal” soils from 0.5 to 1. Preliminary estimations have shown a notable nonequilibrium between the concentrations of 238 U and 226 Ra in radium-rich soils. The reasons for the observed depletion in 238 U are not clear and further detailed radioecological research is necessary. We have found surface soils characterized by a signiﬁcant deﬁciency or by a surplus of 210 Pb. Depth-dependent activity concentrations have been successfully least-square ﬁtted using a 1D diffusion model describing both unsupported and supported 210 Pb fractions in soil. About 2/3 of the unsupported 210 Pb fraction inventory was found bounded in the upper 2 to 8 cm soil horizon. As a rule, the supported fraction inventories of 210 Pb are considerably smaller than those of 226 Ra. The ﬁrst assessment of the 210 Pb atmospheric deposition ﬂuxes and its concentrations in precipitation based on analysis and modeling in areas of NE Estonia, involving those of high 226 Ra soil content, has been made. The ﬂuxes show reasonable correlation with the 226 Ra activity concentrations in soil cores, but practically no correlation with the soil density values or with the determined radon emanation coefﬁcients. The areas with high 226 Ra (238 U) concentration in soil and the low-lying uranium-rich alum shale layers in the background are assumed to be the most probable sources of 222 Rn in air of the region. The role of two large oil-shale-ﬁred power plants in the inventory of 210 Pb in soil is still uncertain, and additional analyses are needed. The analysis may also be useful to clarify the possible use of the 210 Pb inventory in soil proﬁles as an indicator of high 226 Ra concentrations in deeper soil layers, for radon migration characterization, etc.

Acknowledgement The authors gratefully acknowledge M. Lust for ﬁeld and laboratory assistance.

Natural radionuclides in radium-rich soils in North-East Estonia

149

References [1] J. Jõgi, R. Koch, E. Realo, K. Realo, Tehnika ja Tootmine (Technology and Production) 8 (1994) 11; Tehnika ja Tootmine (Technology and Production) 9 (1994) 14 (in Estonian). [2] E. Realo, K. Realo, J. Jõgi, J. Environ. Radioact. 33 (1996) 77. [3] E. Realo, in: Proc. of the Regional IRPA Congress, Stockholm, 1998, p. 193. [4] L. Saare (Ed.), Estonian Environment 1994, Environm. Information Centre, Tallinn, 1994. [5] S. Ylatalo, J. Karvonen, T. Ilander, T. Honkamaa, H. Toivonen, STUK Report No. A 134, STUK, Helsinki, 1996. [6] K. Realo, E. Realo, in: J. Søgaard-Hansen, A. Damkjær (Eds.), Proc. of 12th NSFS/IRPA Ordinary Meeting, Skagen, Denmark, Risø National Laboratory, Roskilde, 1999, p. 229. [7] K. Realo, E. Realo, in: B. Obeli´c, et al. (Eds.), Proc. of the IRPA Regional Congress on Radiation Protection in Central Europe, Dubrovnik, Croatia, IRPA/Croatian RPA, Dubrovnik, 2001, p. 127. [8] UNSCEAR, Sources and Effects of Ionizing Radiation, vol. I: Sources, United Nations, New York, 2000. [9] T. Tokieda, K. Yamanaka, K. Harada, S. Tsunogai, Tellus 48B (1996) 690. [10] H.E. Moore, S.E. Poet, J. Geophys. Res. 81 (1976) 1056. [11] Statistical Ofﬁce of Estonia, State of Environment, http://gatekeeper.stat.ee:8000/px-web.2001/dialog/statﬁleri. asp. [12] L. Pahapill, R. Rajamäe, Radon in houses, http://www.envir.ee/kiirgus/image/radoon.gif, 2002. [13] J.T. Smith, P.G. Appleby, J. Hilton, N. Richardson, J. Environ. Radioact. 37 (1997) 127. [14] C.R. Olsen, I.L. Larsen, P.D. Lowry, N.H. Cutshall, J.F. Todd, G.T.F. Wong, W.H. Casey, J. Geophys. Res. 90 (1985) 10487. [15] J.F. Todd, G.F. Wong, C.R. Olsen, I.L. Larsen, J. Geophys. Res. 94 (1989) 11106. [16] S. Caillet, P. Arpagaus, F. Monna, J. Domonik, J. Environ. Radioact. 53 (2001) 241. [17] J. Paatero, J. Hatakka, Health Phys. 79 (2000) 691.

150

Depleted uranium determination in soil samples ˇ I. Bikit, J. Slivka, M. Krmar, M. Veskovi´c, L. Conki´ c, ´ ci´c, D. Mrdja E. Varga, S. Curˇ Institute of Physics, Faculty of Sciences, University of Novi Sad, Trg D. Obradovi´ca 4, 21000 Novi Sad, Yugoslavia

When the issue of depleted uranium (DU) presence in the environment emerged, methods for the analytical discrimination of DU against natural uranium had to be developed. We present here a simple gamma-spectrometric method, based on the 238 U–226 Ra activity (non) equilibrium. Preliminary calculations that are still under way lead to the result that the lower limit of detection of DU is about 10 Bq kg−1 for a 50 ks measurement and thus the method is appropriate for the determination of small amounts (≈ 100 Bq kg−1 ) of DU in environmental samples. The method is tested on about 90 soil samples. 1. Introduction Depleted uranium is a by-product of the uranium enrichment process. During this process, the ﬁssile isotope uranium-235 is separated from uranium. The remaining uranium, which is usually 99.8% uranium-238, is called “depleted uranium”. While the term “depleted” implies that it is not particularly dangerous, in fact, this waste product of the nuclear industry is “conveniently” disposed-of by producing efﬁcient anti-armour weapons [1]. In the 1950s, the United States Department of Defense became interested in using depleted uranium metal in weapons because of its extremely dense, pyrophoric qualities and because it was cheap and available in huge quantities. It is now given practically free of charge to the military and arms manufacturers and is used both as tank armour and in armour-piercing shells known as depleted uranium penetrators. After the impact most of the DU is converted at high temperature into an aerosol, that is, minute insoluble particles of uranium oxide, UO2 or UO3 , in a mist or fog [2,3]. The aerosol resists gravity and is able to travel kilometers in air. Once on the ground, it can be resuspended when the sand is disturbed by motion or wind. Penetrators which miss their targets and hit soft soil remain compact and may be slowly dissolved and enter the deeper soil layers or underground waters. In order to identify the exposure to DU, specialized equipment for detection of low-level concentrations of DU in soil is required. This is a complex problem because the soil already has a certain concentration of natural uranium [4]. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07016-0

© 2005 Elsevier Ltd. All rights reserved.

Depleted uranium determination in soil samples

151

2. The experimental method For the quantitative measurement of natural U, several gamma-lines are routinely used [1–3]. The most prominent of those are presented in Table 1. The limited accuracy deconvolution of the 186 keV doublet can be avoided when good low energy efﬁciency detectors in iron shields (not generating X-rays of Pb) are used [5]. By deﬁnition DU is recognized by a lower result for the 235 U/238 U isotopic ratio than its natural value of 7.25 × 10−3 . However, the 235 U content in natural soil samples can be determined by routine measurements (Tm ≈ 1 d) but with signiﬁcant statistical uncertainties, thus making low DU contamination invisible. On the other hand, for DU determination we can use the fact that during the chemical treatment of uranium the 238 U–226 Ra equilibrium is completely destroyed. Freshly processed U does not have 226 Ra, and owing to its half-life of T = 1620 y, the 226 Ra content of processed U is negligible. This conclusion was veriﬁed by direct measurements on DU penetrators. The A(226 Ra)/A(238 U) activity ratio is derived from the energetically close lines of 226 Ra and 238 U, the measurement of which is not affected by self-absorption in the dense penetrator [6,7]. Taking these considerations into account the DU concentration in the sample can be evaluated by the simple formula A(DU) = A 238 U − KA 226 Ra . (1) The factor K which is the activity ratio in the noncontaminated soil sample K=

A(238 U) A(226 Ra)

(2)

limits a bit the generality and simplicity of this method. The A(238 U)/A(226 Ra) equilibrium in soil may be inﬂuenced by the difference in the solubility of U and Ra in various soil types. From the measurement of more than 200 soil samples from Serbia we derived the mean value K = 1.3 ± 0.3,

(3)

which is used in the subsequent analysis. Table 1 The most prominent gamma-rays of 238 U and 226 Ra daughters Parent

Nuclei

Eγ (keV)

Iγ (%)

238 U

234 Th

63.3 92.3 92.8

3.8 2.7 2.7

226 Ra

214 Pb

241.9 295.1 351.9

7.46 19.2 37.1

214 Bi

609.3 1120.3 1764.5

46.1 15.0 15.9

152

I. Bikit et al.

It can be noted that the K value may be affected by the errors in the detection efﬁciency curve for voluminous sources. From the chemistry of U in soil it is hard to understand some reported [5] values of K 1. The 1620 y half/life daughter (226 Ra) activity can dominate the parent (238 U) activity only by unusually fast geochemical processes. Anyway, if the K value is not known from previous measurements, it can be determined from measurements on safely noncontaminated soil samples of the same type as the investigated one. Low-level high-resolution gamma spectroscopy is a very convenient technique for uranium determination in environmental samples with 238 U activity concentration above 1 Bq kg−1 . The radionuclide content of the samples was measured by means of the reversed electrode “GMX”-type HPGe spectrometer made by ORTEC. Semiconductor detectors of type “GMX” are the most suitable for the measurement of the intensity of gamma rays below 100 keV due to their extended efﬁciency in the low-energy region. The detector was operated inside a pre-World War II cast iron shield. The inner volume of the measuring chamber is 1 m3 with wall thickness 25 cm. In this shielding conﬁguration the background is free of any lead X-rays. Thus detection of the low-energy gamma rays following the decay of the ﬁrst 238 U daughters (234 Th) becomes very efﬁcient with this measuring equipment. The nominal efﬁciency of the detector is 32% and the resolution is 1.9 keV at 1332 keV. The detector was calibrated by means of reference radioactive materials in cylindrical geometry (NBS Standard Reference Material 4350B). The matrix effects were taken Table 2 Activity concentration of characteristic radionuclides in soil samples in Novi Sad Sample

Soil, TV Novi Sad I (06. ’99) Soil, TV Novi Sad II (06. ’99) Soil, oil reﬁnery (reservoir) (04. ’99) Soil, oil reﬁnery (gate 4) (04. ’99) Soil, oil reﬁnery (Uljara 1) (05. ’99) Soil, oil reﬁnery (workshop) (05. ’99) Soil, Ribnjak (07. ’99) Soil, TV-tower, Venac (06. ’99) Soil, EPS, Venac (09. ’99) Concrete, Žeželj bridge (04. ’99) Sand, Naftagas (09. ’99) Soil, Iriški Venac 1a (09. ’99) Soil, Iriški Venac 1b (09. ’99) Soil, Iriški Venac 2 (09. ’99) Soil, Iriški Venac 7 (09. ’99) Soil, Iriški Venac 8 (09. ’99) Soil, Iriški Venac 10 (09. ’99) Soil, Iriški Venac 12 (09. ’99) Sediment, canalDTD-mouth of Danube (09. ’99) Sediment, Danube-Subi´c (09. ’99) ∗U

Activity concentration (Bq kg−1 ) UD ∗

238 U

226 Ra

232 Th

40 K

< 26 25 ± 9 11 ± 6 20 ± 6 20 ± 8 20 ± 11 24 ± 1 12 ± 5 27 ± 20 14 ± 5 19 ± 11 80 ± 30 56 ± 23 80 ± 30 64 ± 20 80 ± 30 35 ± 27 54 ± 17 33 ± 10

23.4 ± 1.2 27.9 ± 1.2 10.6 ± 0.4 18.7 ± 0.5 16.1 ± .0.7 12.5 ± 0.6 23.2 ± 2.0 19.5 ± 0.4 30.3 ± 0.9 10.4 ± 0.3 17 ± 1 46 ± 2 54 ± 3 46 ± 2 41 ± 2 46 ± 2 138 ± 7 39 ± 2 31.3 ± 1.4

31 ± 10 37.0±1.5 11.0±0.5 20.3±1.1 17.5±1.1 12.3±1.1 39.3±2.4 24.1±0.7 41.8±2.2 10.9±0.8 3±2 62 ± 4 49 ± 3 62 ± 4 60 ± 3 62 ± 4 44 ± 4 55 ± 4 30 ± 2

382 ± 13 479 ± 21 219 ± 11 281 ± 17 312 ± 21 273 ± 14 380 ± 60 325 ± 13 434 ± 23 151 ± 10 290 ± 20 560 ± 50 450 ± 40 580 ± 40 730 ± 40 580 ± 40 440 ± 70 680 ± 40 459 ± 27

32 ± 2

408 ± 29 < 10

55 ± 10

D – depleted uranium activity concentration.

49 ± 2

< 26 <9 <9 <6 <8 < 11 < 2.2 <5 < 20 <5 < 11 < 30 < 23 < 30 < 20 < 30 < 28 < 17 < 10

137 Cs

< 0.27 0.6±0.4 0.8±0.2 1.0±0.4 0.5±0.3 3.2±0.6 < 0.4 0.6±0.1 < 0.4 0.7±0.2 0.8±0.4 27 ± 2 86 ± 5 27 ± 2 29 ± 2 27 ± 2 188 ± 13 144 ± 8 25.1±1.6 46 ± 2

Depleted uranium determination in soil samples

153

into account by means of a computer code. The typical time of measurement of the samples was 50 ks. Soil samples of mass of about 400 g were dried at 105 ◦ C to constant mass. Then the soil samples were transferred to cylindrical containers and sealed. The measurements were performed after more than 30 days following the sealing procedure, in order to bring the 222 Rn daughters in equilibrium. In this way the 226 Ra activity could be determined from the 214 Bi and 214 Pb activity. The most prominent gamma rays of 238 U and 226 Ra daughters are listed in Table 1. 3. Experimental results and conclusions The experimental results are presented in Tables 2 and 3. For a moderately contaminated sample (samples 1–7 in Table 3) the experimental uncertainty for depleted uranium is 27 Bq kg−1 Table 3 Activity concentration of characteristic radionuclides in soil samples from Southern Serbia Region

Sample

Code

238 U

226 Ra

(Bq kg−1 )

(Bq kg−1 )

UD (Bq kg−1 )

40 K

232 Th

(kBq kg−1 )

(Bq kg−1 )

0.450 ± 0.022 0.480 ± 0.023 0.53 ± 0.04 0.587 ± 0.028 0.554 ± 0.027 0.73 ± 0.04 0.689 ± 0.297 0.522 ± 0.026 0.187 ± 0.022 0.128 ± 0.011 0.116 ± 0.010 0.137 ± 0.010 0.145 ± 0.010 0.141 ± 0.010 1.28 ± 0.09 0.76 ± 0.03 0.695 ± 0.028 0.81 ± 0.04 0.81 ± 0.03 0.76 ± 0.04 0.51 ± 0.027 1.14 ± 0.08 1.15 ± 0.07 1.18 ± 0.08 1.02 ± 0.07 1.21 ± 0.07 1.19 ± 0.08 1.29 ± 0.08 0.93 ± 0.06 0.85 ± 0.04 0.98 ± 0.05

34.4±1.9 34.5±2.3 30 ± 3 46.5±2.5 42.2±2.4 52.0±2.9 51.2±2.7 41.8±2.5 13.5±1.1 11.4±0.9 13.1±1.0 12.4±0.9 14.0±1.0 11.4±0.7 72 ± 7 55 ± 3 52 ± 3 46.9±2.6 49.7±2.6 47.7±2.6 30.9±2.0 52 ± 3 62 ± 4 48.6±2.8 67 ± 4 45 ± 3 51 ± 4 31.7±2.4 41 ± 3 61 ± 3 64 ± 4

post Rn 1 1 1 1 1 1

1 2 Average 3&4 5 6 7

1 2 2 2 2 2 2 3 3

8 Average 1 2 3 4 5 Average 1

3 3

2 3

3 4 4 4 4 4 4 4 4 4 4

4 Average A Average B 1 2 3 4 5 6 7 8

LVJ18 LVJ19 LVJ4 LVJ20 LVJ21 LVJ22 LVJ22-1 LVJ23 LVJ3 LVJ24 LVJ25 LVJ26 LVJ27 LVJ28 LVJ5 LVJ14 LVJ14-1 LVJ15 LVJ16 LVJ16-1 LVJ17 LVJ2 LVJ1 LVJ6 LVJ8 LVJ7 LVJ9 LVJ10 LVJ11 LVJ12 LVJ13

29 ± 7 25 ± 6 5000 ± 1300 70 ± 12 42 ± 8 191 ± 24 201 ± 25 32 ± 11 115 ± 15 17 ± 10 14 ± 5 75 ± 10 19 ± 5 16 ± 6 2000 ± 600 133 ± 19 113 ± 18 107 ± 14 115 ± 14 108 ± 14 56 ± 9 42 ± 9 74 ± 10 38 ± 8 65 ± 16 33 ± 8 41 ± 8 44 ± 7 40 ± 11 51 ± 9 47 ± 9

15.7±2.9 15 ± 4 19.8±2.1 29 ± 5 30 ± 4 35 ± 4 52 ± 3 24 ± 4 13.2±2.3 8.3±2.2 9.6±1.4 10.8±1.9 11.3±1.8 7.5±1.8 116 ± 11 82 ± 7 76 ± 6 67 ± 7 68 ± 6 72 ± 7 30 ± 4 22 ± 3 29.7±2.8 19 ± 3 26 ± 4 22.4±2.2 26.4±2.6 24 ± 3 20 ± 3 25 ± 4 27.2±2.1

< 18 < 15 5000 ± 1300 32 ± 16 < 16 146 ± 27 133 ± 30 < 15 98 ± 16 < 17 <8 61 ± 11 < 10 < 13 1850 ± 600 < 70 < 44 < 57 < 63 < 53 17 ± 14 13 ± 12 35 ± 14 < 27 31 ± 19 < 15 < 19 13 ± 11 14 ± 13 < 36 < 24

154

I. Bikit et al.

or 18%. If we accept the most simple deﬁnition of the detection limit aL (UD ) = (UD ) this result together with the data in Table 2 leads to a preliminary general order of magnitude estimate of aL (UD ) ≈ 10 Bq kg−1 . From Tables 2 and 3 it can be concluded that the measurements on samples collected from Novi Sad prove the absence of detectable amounts of depleted uranium. However, in the southern part of Serbia samples contaminated with depleted uranium were found. Our method for depleted uranium detection can be employed for further surveying the radioactivity of the region and especially for the investigation of depleted uranium migration from the contaminated areas.

Acknowledgement The authors acknowledge the ﬁnancial support of the Ministry of Sciences and Technology of Serbia in the frame of the project Nuclear Spectroscopy and Rare Processes (No. 1859).

References [1] P. Loewenstein, Industrial Uses of Depleted Uranium, vol. I, American Society for Metals, 1989. [2] Health and Environmental Consequences of Depleted Uranium Use in the US Army, US Army Environmental Policy Institute (AEPI), 1995. [3] H. Livingstone, Depleted Uranium Weapons, The Edge Gallery, London, 1995. ˇ [4] I. Bikit, J. Slivka, M. Krmar, Lj. Conki´ c, M. Veskovi´c, Ž. Ðurˇci´c, N. Žiki´c, in: 3rd Int. Symp., Federal Ministry for Development, Science and Environmental, University of Novi Sad, Yugoslavia, 1998, p. 351. [5] Z.S. Žuni´c, D.J. Karangelos, M.J. Anagnostakis, E.P. Hinis, S.E. Simopoulos, in: Book of Abstracts of the 7th International Symposium on Natural Radiation Environment (NRE-VII), Rhodes, Greece, May 2002, p. 202. ˇ [6] S. Manojlovi´c, I. Bikit, J. Slivka, M. Veskovi´c, D. Dozet, Lj. Conki´ c, in: 10th World Fertilizer Congress of CIEC, Contributed Papers, International Scientiﬁc Centre of Fertilizers, Nicosia, Cyprus, 1990. ˇ ´ ci´c, D. Mrda, [7] I. Bikit, J. Slivka, M. Krmar, Lj. Conki´ c, E. Varga, S. Curˇ ¯ Arch. Oncol. (2001) 241.

155

Measurement of snow cover based on external radiation J. Paatero a , E. Kyrö b , J. Hatakka a , V. Aaltonen a , Y. Viisanen a a Finnish Meteorological Institute, Air Quality Research, Sahaajankatu 20E, FIN-00880 Helsinki, Finland b Finnish Meteorological Institute, Arctic Research Centre, Tähteläntie 62, FIN-99600 Sodankylä, Finland

Data on the amount of snow (snow water equivalent, SWE (kg m−2 )) is used in Finland in the planning of ﬂood control actions and hydroelectric power production. In this work we have measured external dose rate and compared the results with simultaneous SWE values in order to study the applicability of attenuation of soil-originated gamma radiation in the snow cover for the SWE measurements. 1. Introduction The external dose rate in Finland varies between 0.04 and 0.30 μGy h−1 depending mainly on the amount of radionuclides in the ground. These radionuclides comprise predominantly of the natural decay series of uranium-238 and thorium-232 and the primordial nuclide potassium40. Other factors affecting the external radiation include: • time of year, because the gamma radiation from the Earth’s crust is attenuated in the snow cover in winter, • the latitude, because the Earth’s magnetic ﬁeld directs the charged particles of the cosmic radiation towards the magnetic poles, • the amount of radionuclides suspended in the air, • the amount of radionuclides deposited to the soil surface, and • the altitude, because the dose rate due to the cosmic radiation increases as a function of elevation. The four last factors are of minor importance in Finland [1,2]. Finland is covered with snow during a time period ranging from a few weeks on the southwestern archipelago to over half a year in the north. Data on the amount of snow (snow water equivalent, SWE (kg m−2 )) is needed in the planning of ﬂood control actions, hydroelectric power production and agriculture. The data is also used in monitoring the weight on building roofs and in the forecasts of the ground water level. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07017-2

© 2005 Elsevier Ltd. All rights reserved.

156

J. Paatero et al.

In this work we have measured external dose rate with an ionisation chamber and compared the results with simultaneous SWE values in order to study the applicability of attenuation of soil-originated gamma radiation in the snow cover for the SWE measurements. We have also measured concurrently the external radiation with scintillation detectors to ﬁnd out their count rate response as a function of snow cover which modiﬁes the gamma-ray energy spectrum. 2. Experimental The measurement site was on a ﬂat ﬁeld at the Arctic Research Centre of the Finnish Meteorological Institute (FMI), Sodankylä, Finland (67◦ 22 N, 26◦ 39 E, 178 m above the sea level) about 120 km north of the Arctic circle. The terrain in the area consists mainly of sand and gravel. Together with the air quality monitoring station at Pallas the Arctic Research Centre form the Pallas-Sodankylä Global Atmosphere Watch (GAW) station. The GAW programme is managed by the World Meteorological Organisation (WMO) [3–5]. The measurement of external dose rate was made with a pressurised ionisation chamber (Eberline FHT191N) two metres above the ground from June 1999 to July 2000. The obtained 10-minute values were processed to daily mean values. The scintillation detector system employed in this study consists of two 76 mm × 76 mm NaI(Tl) detectors connected to a HV supply/ampliﬁer/single channel analyser unit (Alnor RD-1600). The detectors are separated by a lead shield 25 mm thick situated 1.5 m above the ground (Fig. 1). This counting geometry causes the lower detector to be more sensitive to gamma radiation from the ground and deposited gamma emitters. The upper detector is more sensitive to the gamma-emitting radionuclides suspended in the air. The coincidence counts, which are also recorded, are mainly caused by cosmic ray muons travelling through both detectors and partly by chance coincidences. The detector system is placed in a temperature-regulated shelter made of plywood and polystyrene insulation. All pulses exceeding the lower level discriminator (corresponding to a photon energy of 50 keV) were recorded in 10-minute intervals and processed to daily values [6,7].

Fig. 1. NaI(Tl) detector system.

Measurement of snow cover based on external radiation

157

The SWE measurements (every ﬁfth day and twice a month from a larger area) were made by sticking a corer through the snow cover and weighing the mass of the collected snow core with a balance. The obtained SWE results were interpolated both in time and space to get a daily estimate of a regional SWE value. These calculations were done at the Finnish Environmental Institute.

3. Results and discussion The snow cover exponentially reduces the gamma radiation coming from the Earth’s crust to the air (Fig. 2). A curve ﬁtted to the observed data corresponds to the function DR = 0.0317e−0.00954∗SWE + 0.0548, μGy h−1 .

(1)

The constant term can be attributed to the where DR is the external dose rate in cosmic radiation. The data points clearly below the ﬁtted curve are recorded during a couple of weeks in late spring when the snow melts very rapidly (Fig. 3). In this case the melted water

Fig. 2. Dose rate vs. SWE at Sodankylä, Finland, winter 1999/2000.

Fig. 3. Dose rate at Sodankylä, Finland, June 1999–July 2000.

158

J. Paatero et al.

Fig. 4. NaI(Tl) detector system count rate vs. dose rate at Sodankylä, Finland. The R 2 values are 0.958, 0.971, and 0.969 for the upper detector, lower detector and sum count rates, respectively.

attenuates gamma radiation but is not trapped into the corer leading to underestimated SWE values. The sudden peaks in the dose rate curve are due to the precipitation, which scavenges airborne short-lived radon-222 progeny to the ground. The NaI(Tl) detector system response is very linear in the dose rate range found in Sodankylä (Fig. 4). Any higher order polynomial or logarithmic function would improve the least squares ﬁt only marginally and the count rate to dose rate conversion can be achieved with a simple linear conversion.

4. Conclusions The measurement of SWE based on the attenuation of gamma radiation offers several advantages compared to snow cover coring and subsequent weighing. The present method takes into consideration the melted water, which is not trapped into the corer. It also integrates SWE from a larger area compared to point data obtained with a corer. The method also provides continuous data. All these factors lead to improved prognoses of melt water amount, which helps the planning of ﬂood protection measures and facilitates optimised hydroelectric power production. The accuracy of the method could be further improved by replacing the ionisation chamber with a NaI(Tl) scintillation detector and by recording the count rate of a single soil-originated gamma peak, e.g., that of 40 K. This would signiﬁcantly lower the background caused by cosmic radiation. In addition, during summer season the measurement of external dose rate can provide data on the evapotranspiration of water from the soil [2]. In this case, too, the method is continuous and automatic unlike the traditional method of measuring the water level in an open vessel. The traditional method also ignores the effect of plant transpiration. The present method thus

Measurement of snow cover based on external radiation

159

gives a better view of the soil moisture content, which can improve the accuracy of forest ﬁre risk assessments.

References [1] S. Ristonmaa (Ed.), Preparedness events and national radiation surveillance, Annual report 1997, STUK – Radiation and Nuclear Safety Authority, Helsinki, 1998. [2] J. Hatakka, J. Paatero, Y. Viisanen, R. Mattsson, Radiochemistry 40 (1998) 534. [3] J. Paatero, J. Hatakka, Y. Viisanen, Concurrent measurements of airborne radon-222, lead-210 and beryllium-7 at the Pallas-Sodankylä GAW station, Northern Finland, Report No. 1998:1, Finnish Meteorological Institute, Helsinki, 1998. [4] http://fmiarc.fmi.ﬁ. [5] http://fmigaw.fmi.ﬁ. [6] J. Paatero, J. Hatakka, R. Mattsson, I. Lehtinen, Radiat. Prot. Dosim. 54 (1994) 33. [7] J. Paatero, J. Hatakka, Boreal Environ. Res. 4 (1999) 285.

160

Ra and U isotopes determination in phosphogypsum leachates by alpha-particle spectrometry J.L. Aguado a,* , J.P. Bolívar a , E.G. San Miguel a , R. García-Tenorio b a Departamento de Física Aplicada, Universidad de Huelva, Facultad de Ciencias Experimentales,

21007 Huelva, Spain b Departamento de Física Aplicada, Universidad de Sevilla, E.T.S. Arquitectura, Avda Reina Mercedes 2,

41012 Sevilla, Spain

On-site studies have shown the radiological impact on the Tinto and Odiel rivers (SW of Spain) by the phosphogypsum produced as wastes in several phosphoric-acid industries located in this area. Leaching of natural radionuclides (226 Ra, 210 Po, 234 U, 238 U) from phosphogypsum piles formed on the bank of the Tinto river has been identiﬁed as one of the main routes of radiological impact. The aim of this work was to study the 226 Ra and U-isotope leaching behaviour from phosphogypsum for a better understanding of the radiological impact caused in these aquatic systems.

1. Introduction Several factories located in Huelva (SW of Spain) are devoted to the production of phosphoric acid from sedimentary phosphate rocks by the so-called “wet process”: Ca10 (PO4 )6 F2 + 10H2 SO4 + 20H2 O = 6H3 PO4 + 10CaSO4 2H2 O + 2HF.

(1)

As a result of this process, the signiﬁcant natural radioactivity associated with the phosphate rock (∼ 1000 Bq kg−1 of 238 U with its decay products) is fractionated between the phosphoric acid and the waste of the process, phosphogypsum. In fact, it has been estimated that 90% of the radium and polonium is fractionated to the phosphogypsum, while the majority of the uranium (> 60%) could be found in the phosphoric acid [1]. Since the middle of the 1960s, the phosphogypsum produced in these factories was managed as follows: the major part of the gypsum (80%) was transported from the factories with * E-mail address: [email protected] (J.L. Aguado).

RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07018-4

© 2005 Elsevier Ltd. All rights reserved.

Ra and U isotopes determination in phosphogypsum leachates by alpha-particle spectrometry

161

water and stored in big piles. These gyp-stacks are located next to the factories in a salt-marsh area on the bank of the Tinto river. The drainage waters and the waters used for the transportation of phosphogypsum were released to this river. The rest of the phosphogypsum (20%) was released directly to the other neighbouring river the Odiel. The new management policy of the factories has determined that the 3 × 106 metric tonnes of phosphogypsum produced annually in Huelva must be stored in the piles, while the associated waters are stored, chemically treated and reused for transportation in a closed circuit. Previous works [2,3] have shown an unambiguous radiological impact on the environment near to the factories: river waters, soils and sediments reveal enhanced concentrations of the 238 U-series radionuclides. These on-site studies have shown that leaching of radionuclides from phosphogypsum piles is one of the main pathways of impact. In this work, the leachability of 226 Ra and U-isotopes in water from the phosphogypsum was analysed by means of “forced leaching” experiments. Based on the results obtained we tried to understand the radiological impact caused by this process in the neighbouring aquatic systems. 2. Material and methods Two different samples of phosphogypsum have been considered for the leaching experiments: fresh phosphogypsum, collected from one of the factories before its transport to the piles, and “old” phosphogypsum, collected from the surface of the piles. The radioactive concentrations of these samples are shown in Table 1. These results are in good agreement with Ra and U-isotope concentrations previously found in these piles [4] and with typical values reported in the literature [5]. It is remarkable that 226 Ra concentrations are 3 or 4 times higher than U-isotope speciﬁc activities. Otherwise, 238 U and 234 U are almost in secular equilibrium. Aliquots containing three grams of phosphogypsum were mixed with 60 ml of ﬁltered fresh or marine water (U and Ra concentrations in waters were negligible). The phosphogypsum/water ratio was chosen according to the usual mixtures used by the factories to transport phosphogypsum to the piles, because it has been found that the liquid : solid ratio plays an important role in the leachability of radium from phosphogypsum [6]. Each aliquot was stirred with water at 800 rpm under laboratory conditions of pressure and temperature (1 atm, 20 ◦ C) for a certain interval of time (5, 10, 15, . . . , 2880 min). After stirring, the solution was immediately ﬁltered through 0.8 μm Millipore paper and the pH of the liquid fraction was measured. Then, the leachates were subjected to 226 Ra and U-isotope analysis. When the speciﬁc activities of the leachates were known, the 226 Ra and U percentages of activity transferred from the phosphogypsum to the water (ACR) were calculated. Table 1 Activity concentrations (Bq kg−1 ) of 226 Ra and U-isotopes in phosphogypsum samples Isotope

Fresh phosphogypsum

“Old” phosphogypsum

238 U

233 ± 17 241 ± 17 818 ± 40

175 ± 12 185 ± 12 825 ± 34

234 U 226 Ra

162

J.L. Aguado et al.

U and Ra-isotopes were isolated from the leachates using a new sequential extraction procedure [7]. First, the pH of the liquid fraction was adjusted to 1. Then Ra was co-precipitated with Pb, while U remained in the supernatant. The precipitate with Ra was redissolved with EDTA and interfering elements were eliminated with the proper use of two chromatographic columns. Finally, Ra was electrodepositated in organic media onto a stainless steel disc. On the other hand, the supernatant fraction with U was evaporated to dryness and redissolved with 8 M nitric acid. After that, a solvent extraction procedure was applied to obtain a puriﬁed solution with uranium [8]. Finally, a thin U-radioactive source was obtained by electrodeposition in aqueous media. Typical radiochemical recoveries for U and Ra isolation from leachates were greater than 70%. Radioactive sources were counted by alpha-particle spectrometry using low background ion-implanted silicon detectors with a 0.25 absolute efﬁciency. Alpha emissions were directly identiﬁed in the spectra, the FWHM of the peaks obtained being in the 30–40 keV range. The counting periods were 2–3 days, so the uncertainties of the measurements were below 10%. Finally, the minimum detectable activities for 226 Ra and U isotope determinations were less than 1 mBq.

3. Results and discussion Figure 1 shows the 226 Ra activity ratio (Ra-ACR) obtained from the forced leaching experiments with fresh water and fresh phosphogypsum and for different stirring times. This coefﬁcient was quite variable at short times of stirring but tends to a steady value for longer contact times (24–48 h). The corresponding 226 Ra concentrations in the acidic liquid fraction obtained (pHs of the leachates were closed to 3.0) ranged from 3 to 0.5 Bq L−1 . If the leaching was performed with marine water the initial Ra-ACRs were higher (∼ 8%) than ratios found with fresh water, this difference being related to the presence of ions in the marine water that could be initially adsorbed to the phosphogypsum instead of radium.

Fig. 1. 226 Ra transferred from fresh phosphogypsum to ﬁltered fresh water versus time of stirring.

Ra and U isotopes determination in phosphogypsum leachates by alpha-particle spectrometry

163

Fig. 2. 238 U transferred from fresh phosphogypsum to ﬁltered marine water versus time of stirring.

Nevertheless, the concentration of 226 Ra in the leachates for higher times of stirring were similar for fresh and marine water (0.5 Bq L−1 ). Moreover, it is important to mention that the time evolution of the solubility of phosphogypsum in the liquid fraction shows a tendency similar to that in Fig. 1. It varies from 7 (short times) to 3 mg ml−1 (long times). So, the leachability of 226 Ra to water could be related to the solubility of the phosphogypsum itself. However, the 226 Ra concentration in the leachates was disproportionate to the dissolution of the phosphogypsum itself, in agreement with the results described in previous works [5,6,9]. When similar leaching experiments were performed with “old” phosphogypsum, the RaACRs found at different times of stirring were very uniform (0.2–1.0%). Speciﬁc activities varied in a quite narrow range: 0.1–0.5 Bq L−1 . These results were in agreement with a previous study about the fractionation of 226 Ra in phosphogypsum [10], and shows that a major proportion of 226 Ra could be extracted by dissolution with water from fresh phosphogypsum, whereas 226 Ra tends to be associated with refractory forms of the gypsum as the age of phosphogypsum increases. Figure 2 shows the 238 U activity ratio (U-ACR) obtained between ﬁltered marine water and fresh phosphogypsum. It is evident that the leachability of U was different from 226 Ra. The ratios found with fresh and marine water were in the 40–60% range and concentrations of 238 U in the leachates were close to 6 Bq L−1 for longer times of stirring. Additionally, 238 U and 234 U were in secular equilibrium in the leachates (average activity ratio between 234 U and 238 U was 0.99 ± 0.02, error = 1 standard deviation). Consequently, the leachability of the isotopes studied from fresh phosphogypsum clearly follows the pattern U > Ra. The activity ratios 226 Ra/238 U found on the leachates were less than one (0.21 ± 0.04 with fresh water and 0.61 ± 0.19 with marine water; errors = 1 standard deviation), while this activity ratio in bulk phosphogypsum is clearly higher than one. If we take into consideration that the sediments collected from the Tinto river are enriched in U in relation to Ra, we can ratify the leaching as the main pathway towards radiological impact of the phosphogypsum piles on the neighbouring area.

164

J.L. Aguado et al.

The experiments with “old” phosphogypsum gave us U-ACRs next to 2% and 238 U concentrations in the leachates lower than 0.5 Bq L−1 . These results are clearly related to the acidic content of the liquid fraction. If phosphoric acid was present in the phosphogypsum sample, the pH of the leachates was less than 3 and U-ACR tends to higher values. On the contrary, the absence of acid in old phosphogypsum (pH of the leachates was higher than 7) leads to a reduction of transfer. This was another difference of the U behaviour cf. Ra and it has been ratiﬁed with the new storage policy now followed in the factories. In fact, acidic waters from a reservoir located at the new piles showed U-concentrations of 200 Bq L−1 , whereas 226 Ra concentrations were clearly lower and very similar to our results for long contact times (0.8 Bq L−1 ). Finally, we have leached phosphogypsum samples collected at a depth of 60–70 cm inside the piles. The corresponding leachates were more acidic (pH varies in the 5–6 range) than liquid fractions obtained with superﬁcial “old” phosphogypsum. Consequently, the U-concentrations found in the leachates of this sample with fresh water were one order of magnitude higher than those found in superﬁcial samples (∼ 2 Bq L−1 ) and the U-ACRs were 25%. This fact indicates that, probably, the acidic waters used to transport the gypsum to the piles ﬁltered through the stacks and, consequently, enrich with acid groups the deeper gypsum layers. On the other hand, 226 Ra concentrations in these leachates ranged between 0.1 and 0.4 Bq L−1 , which are similar to the concentrations found in the leachates from superﬁcial samples.

4. Conclusions Forced leaching experiments on phosphogypsum reveals that the main transfer of Ra and U from the phosphogypsum wastes to the water was performed with fresh phosphogypsum. Consequently, waters used for phosphogypsum transportation to the piles for its storage could be enriched in U-isotopes and 226 Ra. Additionally, U concentrations in these leachates were higher than for Ra, in agreement with the results found in sediments collected in the Tinto river, just in the zone where these waters drained after phosphogypsum storage. These sediments show U/Ra activity ratios higher than one, in contrast to the U/Ra activity ratios lower than one found in bulk phosphogypsum. On the other hand, leachates obtained from forced leaching experiments with phosphogypsum previously washed and stored in the piles (“old” phosphogypsum) showed for both U and Ra very low transfer to the waters. However, it was found that U mobilisation from these wastes depends on the presence of traces of phosphoric acid in the phosphogypsum (i.e., on the acidity of the leachates).

References [1] [2] [3] [4]

J.P. Bolívar, R. García-Tenorio, M. García León, J. Radioanal. Nucl. Chem. Lett. 214 (2) (1996) 77. J.P. Bolívar, R. García-Tenorio, M. García León, Appl. Radiat. Isot. 47 (9/10) (1996) 1069. J.L. Aguado, J.P. Bolívar, R. García-Tenorio, M. García León, Verh. Internat. Vrein. Limnol. 26 (1998) 893. J.L. Mas, J.P. Bolívar, R. García-Tenorio, J.L. Aguado, E.G. San Miguel, González-Labajo, J. Health Phys. 80 (1) (2001) 34.

Ra and U isotopes determination in phosphogypsum leachates by alpha-particle spectrometry

165

[5] P.M. Rutheford, M.J. Dudas, R.A. Samek, Sci. Total Environ. 149 (1994) 1. [6] P.P. Haridasan, C.G. Maniyan, P.M.B. Pillai, A.H. Khan, J. Environ. Radioact. 62 (1) (2002) 287. [7] J.L. Aguado, J.P. Bolívar, E. Gutiérrez San-Miguel, R. García-Tenorio, in: Proceedings of the XIV Radiochemical Conference, Marianske Lazne, Czech Republic, 2002. [8] E. Holm, R. Fukai, Talanta 24 (1977) 659. [9] P.M. Rutheford, M.J. Dudas, J.M. Arocena, J. Environ. Qual. 24 (1995) 307. [10] W.C. Burnett, G. Schaefer, M.K. Schultz, in: Spec. Publ. (Environmental Radiochemical Analysis), vol. 234, Royal Society of Chemistry, 1999, p. 1.

166

Efﬁciency calibration for 210Pb gamma-spectrometric determinations in sediment samples E.G. San Miguel a , J.P. Pérez-Moreno a , J.P. Bolívar a , J.L. Aguado a , R. García-Tenorio b a Departamento Física Aplicada, EPS La Rábida, Universidad de Huelva, 21819-Palos, Huelva, Spain b Departamento Física Aplicada II, ETS Arquitectura, Universidad de Sevilla, Avda. Reina Mercedes 2,

41012 Sevilla, Spain

A method of efﬁciency calibration for 210 Pb determinations by gamma-ray spectrometry in sediment samples is outlined. This method, developed for a cylindrical sample geometry with coaxial Ge detectors, is based on the gamma transmission method. It accounts for variable sample height and supplies a fundamental advantage: individual self-absorption corrections can be easily determined knowing the apparent densities and the elemental composition of the samples. This calibration has been used for dating sediments of the Huelva estuary (southwest of Spain).

1. Introduction Historically, 210 Pb has been determined in environmental samples mainly through its granddaughter 210 Po by alpha-particle spectrometry [1] or, to a less extent, through its beta daughter 210 Bi [2]. Nevertheless, since Gäggeler et al. [3] suggested the determination of 210 Pb through gamma spectrometry via its 46.5 keV gamma ray, the application of this technique for 210 Pb determination has increased drastically. The main advantage of the 210 Pb determination by gamma-ray spectrometry in environmental samples is that no radiochemical methods are required for its isolation as a previous step to the measurement, making it a direct and nondestructive technique. However, an important drawback needs to be considered: the high self-absorption of the soft gamma particle emitted, that strictly depends on the composition and apparent density of the analysed sample. Cutshall et al. [4] devised a technique for 210 Pb self-absorption corrections which needs the use of a strong 210 Pb point source and to measure the samples twice: with and without the point source on top of the sample holder. Since the work of Cutshall et al., many other authors RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07019-6

© 2005 Elsevier Ltd. All rights reserved.

Efﬁciency calibration for 210 Pb gamma-spectrometric determinations in sediment samples

167

have treated the photon self-absorption problem in environmental samples for different energy ranges and geometries. In this work, we propose an alternative method for determining the 210 Pb efﬁciency calibration in sediments samples aimed at a standard coaxial gammaspectrometric system. This method takes into account the correction for the self-absorption of 46.5 keV γ-ray, considering that the full energy peak efﬁciency depends on composition and apparent density of the sediment. It uses a solid environmental sample highly enriched in 210 Pb as a calibration matrix. In this way, we avoid the process of homogenisation needed when a 210 Pb spike is added to a solid blank for manufacturing the calibration sample.

2. Materials and methods We have determined the concentrations of 210 Pb in sediments and phosphogypsum (PG) samples by gamma spectrometry through the method proposed. In these samples, as well as in the calibration matrix (phosphate rock, PR) we have also determined 210 Pb through 210 Po by alpha spectrometry for comparison and validation. 2.1. Gamma-ray spectrometry All the 210 Pb determinations by gamma-ray spectrometry were performed with the same system: a Canberra Extended Range Ge detector (XtRa) model GX3519, with 38% of relative efﬁciency and FWHM (full width half maximum) of 0.95 keV at 122 keV and 1.9 keV at 1330 keV. The detector works coupled to a conventional electronic chain, including a multichannel analyser, and is shielded with 15 cm thick Fe. Prior to doing the measurements, each sample was dried at 60 ◦ C, homogenised, ground, introduced in polyethylene containers of cylindrical geometry (diameter 6.5 cm) and sealed. Regarding the gamma measurements of the sediment samples, the statistical uncertainties from the net counts under the 210 Pb photopeak were kept below 5% by maintaining a minimum of 24 h counting time, while in the calibration matrix (PR containing about 1 Bq g−1 of 210 Pb) the same counting time led to uncertainties of 1%. The minimum detectable activity (MDA) is lower than 0.5 Bq for 210 Pb. 2.2. Alpha-particle spectrometry In aliquots of the samples analysed by gamma spectrometry, the concentrations of 210 Po (210 Pb) have also been determined by alpha spectrometry. These samples were in all cases older than two years to assure secular equilibrium between the two nuclides. For the 210 Po determination, we have applied a sequential solvent extraction method with tributylphosphate (TBP) [5] that allows Po-isotope isolation in the ﬁrst stage. Prior to the treatment, the samples were traced with known amounts of 209 Po to evaluate the recovery yields of the radiochemical method. The Po-isotopes were isolated in an aqueous phase after a solvent extraction, to be ﬁnally selectively self-deposited onto silver disks [1]. Recovery yields for Po of 60–70% were obtained. The Po planchets were measured using an EG&Ortec alpha spectrometry system with ionimplanted silicon detectors, and the activities were quantiﬁed by the application of the isotope

168

E.G. San Miguel et al.

dilution technique. Counting times ranged from two days to one week, depending on the activity concentrations and the recovery obtained in the chemical separation. The minimum detectable activity (MDA) is lower than 1 mBq for 210 Po. It is worth noting that the MDA obtained for 210 Pb is about 500 times higher than for 210 Po. 2.3. Composition analysis The elemental composition of different samples was determined by TTPIXE in the 3 MV Van de Graaff accelerator of the “Instituto Tecnológico e Nuclear” (ITN), Sacavém (Portugal). A complete description of the technique can be found elsewhere [6].

3. Full energy peak determination 3.1. The method In this work, we have used as a solid calibration material, a phosphate rock (PR), in which activities were determined in several aliquots through its daughter 210 Po by alphaparticle spectrometry [1]. As it is known, phosphate ores contain high activities of uranium and, as a consequence, also of 210 Pb [7], which is a member of its radioactive decay series and stays in secular equilibrium with 210 Po. Once the 210 Pb activity in the calibration matrix (PR) has been determined with good accuracy, we can obtain the full energy peak efﬁciency in the calibration matrix, named hereafter εc , for a ﬁxed geometry. In this work, we have used a cylindrical sample geometry (6.5 cm diameter) allowing a variable sample height (h). In fact, εc for different heights (from 1 to 5 cm with intervals of 0.5 cm) have been determined. We have calibrated at different heights, because the amount of dry material in the sediments to be analysed can be variable from one sample to another, with the result that it is not always possible to ﬁx a constant height for gamma measurements. However, calibration and real samples (sediments in our case) are likely to have different densities and compositions. As a consequence, the 210 Pb self-absorption in both samples may be quite different. Then, to obtain the full energy peak efﬁciency for the real sample (εp ), we have to correct εc by applying the equation 210 Pb

εp = f · εc ,

(1)

where the correction factor f depends on densities and the mass attenuation coefﬁcients (mac) of both samples, as well as on the sample height used in the gamma measurements. This correction factor takes into account the difference in 210 Pb self-absorption between the real and the calibration samples. This correction factor can be determined without any additional gamma measurement in the following way: if we assume a normal incidence of the 46.5 keV gamma-rays emitted by the sample on the detector, the number of photons transmitted without suffering any interactions in the sample is given by the well-known self-absorption equation 1 − e−μρh , ns = n0 (2) μρh

Efﬁciency calibration for 210 Pb gamma-spectrometric determinations in sediment samples

169

where μ is the mac (cm2 g−1 ), ρ is the apparent density (g cm−3 ), and h is the selected height (cm) of the sample, while n0 is the number of photons that originate within the sample and that directly travel towards the detector (number of photons that theoretically would reach the detector in the absence of self-absorption). In this way, theoretically, the full energy peak efﬁciencies εp and εc (for real and calibration samples, respectively) will be given by 1 − e−μp ρp h 1 − e−μc ρc h εp = ε0 (3) and εc = ε0 , μp ρp h μc ρc h where ε0 is the full energy peak efﬁciency in the absence of self-absorption and only depends on the photon energy and geometry. That is, ε → ε0 if μ → 0, for both cases. The last statement can obviously be assumed only if there is not a signiﬁcant difference between the attenuations of the 46.5 keV gamma-rays in the real and calibration samples. In this case, the effective solid angle subtended between the samples and the detector can be considered independent from the type of sample for all our practical purposes. As a result, if we take equations (3) into account, the correction factor f = εp /εc can be written as μc ρc h 1 − e−μp ρp h , f= (4) μp ρp h 1 − e−μc ρc h where the apparent densities of the real (ρp ) and calibration samples (ρc ) are known, h is the height of the sample and μ is obtained through the Bragg’s formula μ= ωi μi . (5) The mass fraction of each component i in the calibration sample, ωi , can be obtained easily since its composition is known (mainly Ca10 (PO4 )6 F2 ), while the mass attenuation coefﬁcients, μi , of each component i for 46.5 keV, can be deduced by interpolation from tabulated data [8]. Thus, the unique quantity which remains initially unknown in equation (4) is the mass attenuation coefﬁcients (mac) for real samples, μp . However, if we determine the elemental composition of the sediment samples whose 210 Pb activity is being measured, the mac (μ) for each element at 46.5 keV can be estimated using Newton’s interpolation in the 20–80 keV energy range. In this way, the mac for sediments can also be obtained using Bragg’s formula. Therefore, the correction factor can be calculated and used to determine εp . We use the 210 Pb calibration proposed for the determination of this nuclide in the different layers of sediment cores in order to apply the 210 Pb dating method. The chronology established in these cores will allow us to study the historical evolution of different heavy metal and other contaminant concentrations in the aquatic system studied. For this reason, the determination of the elemental composition of the sediment samples analysed is essential for our environmental studies, and it will be simultaneously available for the estimation of the correction factors through the determination of the mac for sediments. No additional work is then needed for the proper 210 Pb determinations. 3.2. Full energy peak efﬁciency in the calibration sample The average of 210 Po speciﬁc activities determined in eight different aliquots of the calibration matrix (PR) is 1103 ± 16 Bq kg−1 (1σ ). Once the 210 Pb speciﬁc activity was known,

170

E.G. San Miguel et al.

Fig. 1. Full energy peak efﬁciency (εc ) in phosphate rock for different sample heights.

we determined εc . The calibrating matrix was introduced in the polyethylene containers and placed in front of the detector at a distance of a few millimetres from the window. We have generalised the full energy peak efﬁciency in the calibration sample by its determination at different sample heights. In our case, we have determined εc for heights ranging from 1 to 5 cm, with intervals of 0.5 cm. In Fig. 1, we have plotted these values of εc , with the function obtained by least squares ﬁtting of εc (h) data. We have selected a polynomial function of order two due to its simplicity and the good results supplied. The parameters found in this ﬁtting were εc (h) = (0.114 ± 0.006) − (36 ± 4) × 10−3 h + (35 ± 5) × 10−4 h2 , r 2 = 0.985.

(6)

This function may be useful in some cases when the height of a real sample needs to be interpolated between two exact heights of the calibration sample, considering that the uncertainties predicted by equation (6) are in the 2–4% range.

4. Validation of the model In order to validate the approach proposed for 210 Pb determination in solid samples by gamma-ray spectrometry, we have determined the full energy efﬁciency at 46.5 keV using phosphogypsum (PG) as a solid matrix with different sample heights. PG is the main byproduct formed in the treatment of the phosphate rocks with sulphuric acid by the extraction and isolation of phosphoric acid. This material, whose composition is mainly CaSO4 ·2H2 O, is also enriched in 210 Pb [7]. As the PG matrix used is older than two years, 210 Pb and 210 Po are found to be in secular equilibrium. The average 210 Pb activity determined in this matrix by alpha spectrometry (210 Po) was 1079 ± 23 Bq kg−1 (1σ ). The 210 Pb full efﬁciency energy in gamma-ray spectrometry could then be determined for ﬁve different heights (from 1 to 5 cm, with 1 cm interval) using the same counting geometry as for the calibration sample. As was the case for PR, a least squares weighted ﬁtting of the

Efﬁciency calibration for 210 Pb gamma-spectrometric determinations in sediment samples

171

Fig. 2. Experimental and model correction factor functions (relating phosphate rock and phosphogypsum 210 Pb efﬁciencies).

full energy peak efﬁciency versus sample height (h) was obtained, with the following results: εPG (h) = (0.147 ± 0.010) − (41 ± 6) × 10−3 h + (37 ± 9) × 10−4 h2 , r 2 = 0.992.

(7)

The uncertainties predicted by this equation are in the 1–2% range. Then, if we take into account equations (6) and (7), we can thus obtain an experimental expression for the correction factor (fEXP ) (PR taken as the calibration sample) as follows: fEXP (h) =

εPG (h) . εPR (h)

(8)

On the other hand, considering the mac and apparent densities for PG (0.569 cm2 g−1 ; 1.24 g cm−3 ) and PR (0.717 cm2 g−1 ; 1.60 g cm−3 ), we can derive an expression for determining the correction factor from the model, fMOD (h), by applying equation (4). The values deduced for both expressions are reported in Fig. 2. As can be seen from Fig. 2, there is good agreement between fEXP and fMOD considering the experimental uncertainties (1σ ). The small discrepancies found in the extreme sample heights are mostly derived from the ratio of two functions obtained by least squares weighted ﬁtting in the determination of fEXP . Deviation in the extreme values will be more effective for fEXP (h). Nevertheless, if we consider uncertainties at the 2σ level, we do not ﬁnd signiﬁcant differences between fEXP and fMOD . 5. Validation of the calibration In order to validate the efﬁciency calibration method proposed, we have determined 210 Pb activities through gamma spectrometry by the proposed method and also by alpha spectrometry in sediment samples of the Huelva estuary (SW of Spain). These sediments were old enough to assure secular equilibrium between the two radionuclides. The 210 Pb speciﬁc activities ranged from 20 to 600 Bq kg−1 (the estuary is contaminated by wastes enriched in radionuclides of

172

E.G. San Miguel et al.

Fig. 3. 210 Pb speciﬁc activities determined by gamma-ray spectrometry against 210 Po speciﬁc activities determined by alpha-particle spectrometry.

the uranium series). 210 Pb activities measured by gamma and alpha spectrometry are plotted in Fig. 3. The best linear weighted ﬁtting is also reported, which gave the following results: 210

Pbγ = (5 ± 12) + (0.96 ± 0.04) · 210 Pbα ,

r 2 = 0.995.

(9)

As can be seen, the intercept and slope in equation (9) are not statistically different from zero and one, respectively, at 95% conﬁdence level. This result indicates that there is no difference between the two methods. The good agreement between 210 Pb and 210 Po is a ﬁrst validation of our proposed calibration method for 210 Pb determination in sediment samples.

6. Application of the method to sediment dating As an example of the use of this calibration we have dated an estuarine sediment core by the application of the 210 Pb dating method. With this aim, 210 Pb and 226 Ra activities were determined in different layers of the sediment core. 226 Ra activities were obtained by applying an efﬁciency calibration developed by us for the energy of gamma rays above 150 keV [9], while 210 Pb activities were obtained by the method proposed in this paper. From the values obtained, it was possible to calculate for each layer the unsupported 210 Pb concentrations, and to determine the logarithmic proﬁle of unsupported 210 Pb residual inventory against mass depth (based on the knowledge of the material accumulated per unit area for each layer). This proﬁle clearly follows a linear trend (Fig. 4). In fact, the linear weighted ﬁtting of this proﬁle gave the following results: Ln(AΣ − An ) = (4.70 ± 0.16) − (0.27 ± 0.05) · m,

r 2 = 0.940.

(10)

This relationship between the 210 Pb excess residual inventory and mass depth allows us to obtain a mean sedimentation rate of 0.12(2) g cm−2 yr−1 for this sediment core through the application of the CRS-MV 210 Pb dating model [10]. Although obtaining and validating a value of the sedimentation rate were not objectives of this research, it is worth noting that there is good agreement between the sedimentation rate

Efﬁciency calibration for 210 Pb gamma-spectrometric determinations in sediment samples

173

Fig. 4. Residual inventory of 210 Pb excess (logarithmic proﬁle) against mass depth in a sediment core of the Huelva estuary.

obtained through the 210 Pb dating method and the information provided by the analysis of the vertical proﬁle of 137 Cs in the sediment core. This fact not only validates the applicability of the 210 Pb model elected, but also the method of 210 Pb determination which is the main objective of this paper.

7. Conclusions A method for the analysis of 210 Pb by gamma-ray spectrometry in sediment samples using coaxial Ge detectors has been outlined. Special attention was paid to self-absorption corrections of its soft gamma-ray. The differences in self-absorption between the calibration and the real sample are estimated through a correction factor, which depends on the composition (through the mass attenuation coefﬁcient) and the apparent density of each sample. The results obtained by the application of the proposed calibration method to superﬁcial sediments shows good agreement with those obtained by alpha-particle spectrometry. This fact validates it. Additionally, this calibration has been used successfully for dating sediment cores through the application of the 210 Pb method.

References [1] F. El-Daoushy, K. Olsson, R. García-Tenorio, Hydrobiology 214 (1991) 43. [2] Y.A. Sapozhnikov, O.B. Egorov, I.P. Eﬁmov, S.V. Pirogova, N.K. Kutseva, J. Radioanal. Nucl. Chem. Lett. 176 (1993) 353. [3] H. Gäggeler, H.R. von Gunten, U. Nyffeler, Earth Planet. Sci. Lett. 33 (1976) 119. [4] N.H. Cutshall, I.L. Larsen, C.R. Olsen, Nucl. Instrum. Methods 206 (1983) 309. [5] E. Holm, R. Fukai, Talanta 24 (1977) 659. [6] J.E. Martín, R. García-Tenorio, M.A. Respaldiza, J.P. Bolívar, M.F. da Silva, Nucl. Instrum. Methods B 136/138 (1998) 1000. [7] E.M. Hussein, Health Phys. 67 (1994) 280.

174

E.G. San Miguel et al.

[8] J.H. Hubbell, S.M. Seltzer, NISTIR-5632, 1995. [9] J.P. Pérez-Moreno, J.P. Bolívar, R. García-Tenorio, E.G. San Miguel, J.L. Aguado, J.L. Más, F. Vaca, Radiat. Phys. Chem. 61 (2001) 437. [10] W.R. Schell, M.J. Tobin, in: M. García-León, R. García-Tenorio (Eds.), Proc. 3rd International Summer School: Low-Level Measurements of Radioactivity in the Environment, Techniques and Applications, 1994, p. 355.

175

238 U

and its daughter products in Greek surface soils

M.J. Anagnostakis, E.P. Hinis, S.E. Simopoulos Nuclear Engineering Section, Mechanical Engineering Department, National Technical University of Athens, 15780 Athens, Greece

High concentrations of natural radionuclides in the soil can result in high dose rates outdoors mainly due to external exposure. In dose calculations, the main radionuclides of interest are 226 Ra, 232 Th and 40 K. Research till recent years has focused mainly on the evaluation of 226 Ra concentration, due to radon exhalation from the ground, and due to its relatively simple determination through its gamma emitting decay products. Other radionuclides of the uranium series such as 238 U and 210 Pb, emitting low energy photons, are not usually determined due to γ-spectroscopic difﬁculties, and for dosimetric calculations an assumption of radioactive equilibrium among the nuclides of the uranium (238 U) series as well as the nuclides of the thorium (232 Th) series is usually adopted. Radioactive disequilibrium may exist among the nuclides of the uranium series, and this is the case for the relatively long-lived nuclides 238 U (T1/2 = 4.47 × 109 y), 226 Ra (T1/2 = 1600 y) and 210 Pb (T1/2 = 22.2 y) for several reasons, such as leaching, rock weathering, radon exhalation from the ground, etc. Radioactive equilibrium may also be disturbed because of human activities, such as the operation of coal-ﬁred power plants and the resulting ﬂy-ash dispersion and fallout. An extensive research program has been undertaken by the Nuclear Engineering Section of the National Technical University of Athens (NES-NTUA) for the investigation of the natural and artiﬁcial radioactivity in Greek surface soils. This research has resulted in the mapping of 226 Ra, 232 Th and 40 K in Greek surface soils, as well as the mapping of radionuclides of the Chernobyl fallout in Greece. The gamma spectroscopic determination of 238 U and 210 Pb, which emit low energy photons at 63.29 and 46.5 keV, respectively, with very low yields, requires the use of detectors with high efﬁciency in the low region below 200 keV, such as LEGe or XtRa detectors, and special techniques for correction due to self-absorption inside the samples. Due to the time consuming and difﬁcult determination of the low activities of 238 U and 210 Pb in environmental samples, such as soil, it is only recently that such results are reported in the literature. In the framework of this research soil samples from all over Greece are analyzed for 238 U, 226 Ra and 210 Pb. From the samples analyzed so far, the conclusion may be drawn that, when 226 Ra concentration is much lower than about 25 Bq kg−1 , which RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07020-2

© 2005 Elsevier Ltd. All rights reserved.

176

M.J. Anagnostakis et al.

is incidentally the mean value for 226 Ra in Greek surface soils, radioactive equilibrium is signiﬁcantly disturbed, and the 226 Ra/238 U ratio may be as low as 0.3. In the case where 226 Ra activity is much higher than 25 Bq kg−1 , radioactive equilibrium is again signiﬁcantly disturbed with the 226 Ra/238 U ratio reaching values as high as 4. The above ﬁndings indicate that very high values of 226 Ra in surface soils may be the result of 226 Ra transport to that soil. Concerning the concentration of 210 Pb in surface soils, the analyses performed show that it is usually higher than that of 226 Ra, probably due to radon exhalation from the ground. Analyses of soil samples from deeper soil layers are also performed, in order to investigate the vertical proﬁle of the radionuclides under study.

1. Introduction The assessment of the radiation doses to humans due to natural sources is of particular importance because natural sources contribute signiﬁcantly to the collective dose of the world population. Amongst the natural radiation sources, those of terrestrial origin, i.e., the radionuclides of the 238 U and 232 Th series and 40 K, are important to the dose assessment because of the direct exposure and the radon and thoron exhalation from the substrate soil and building materials. Dosimetric calculations due to natural gamma emitting radionuclides of terrestrial origin are almost always based upon the measurement of relatively easily measured radionuclides, such as 226 Ra, 228 Ra and 40 K, and the assumption of radioactive equilibrium within the uranium (238 U) and thorium (232 Th) series [1]. However, radioactive equilibrium within the uranium (238 U) and the thorium (232 Th) series may be disturbed in the soil, because of the removal of radionuclides, as a result of physical or chemical processes. Among the reasons for radioactive equilibrium disturbance are: the agency of water on soil and rocks (leaching), the recoil of atoms as a result of radioactive disintegration, radon exhalation from the soil and deposition of radon daughters on the soil surface, and deposition of natural radionuclides on the ground surface due to human activities (e.g., ﬂy-ash escaping from coal ﬁred Power Plants). Among the nuclides of the uranium series, radioactive equilibrium disturbance is most probable among the relatively long-lived nuclides of the series: 238 U (T1/2 = 4.47 × 109 y), 234 U (T1/2 = 2.45 × 106 y), 226 Ra (T1/2 = 1600 y) and 210 Pb (T1/2 = 22.2 y). Laboratory measurements indicate that in many cases signiﬁcant disturbance of radioactive equilibrium may exist in the soil, between 238 U, 226 Ra and 210 Pb. According to [2] a weathered rock or soil sample with the 238 U decay series truly in secular equilibrium would be unusual, considering the opportunities for non-equilibrium to occur. From the data presented in the same report, values of the 226 Ra/238 U up to 0.62 for natural soil, and 0.073 for granitic soil are given. In [3] values of ratio 238 U/226 Ra between 0.27 and 1.81 are reported, for quaternary deposits of limestone and tufa, while in [4] values of 226 Ra/238 U ranging between 0.78 and 9.1 are reported, for soil collected at various depths in localized areas, lying at the bottom of a valley, which is characterized as an area of anomalous 226 Ra/238 U disequilibrium. Previous research conducted at NES-NTUA [5] has resulted in the mapping of 226 Ra, 232 Th and 40 K in Greek surface soils. The aim of the present research is the mapping of 238 U and 210 Pb and investigation of the radioactive equilibrium disturbance among 238 U, 226 Ra and 210 Pb, as well as the ratio of 238 U/232 Th and 226 Ra/232 Th in Greek surface soils. These mappings will allow for the localization of areas with signiﬁcant disturbance of radioactive

238 U and its daughter products in Greek surface soils

177

equilibrium. Moreover, mapping of 210 Pb could lead to very useful conclusions regarding the radon exhalation rate from the ground, and the estimation of radon potential, in locations where there is no deposition of natural radionuclides on the ground due to human activities. According to [6] 210 Pb could be an appropriate isotope to characterize the radon potential in a territory, since, due to its relatively long half-life, its concentration in the investigated medium should be on average proportional to the amount of radon transported through it. Depth proﬁles of 210 Po in the soil are reported in [7], with activity ratios from surface to 20 cm depth higher than 10, indicating that the mean source of 210 Po in surface soil is the deposition of radon progeny.

2. Soil sampling and sample preparation An extensive soil sampling campaign has been conducted by NES-NTUA during the period 1986–1999. The sampling covered all geographical subdivisions of the mainland. A total of 1648 1-cm thick surface soil samples of about 1000 cm3 each have been collected from apparently undisturbed sites in open areas. Particular care was taken to avoid sampling at sites where undulations of soil might have made control of 1 cm sampling depth difﬁcult. The soil at the sampling sites was not covered with grass or other vegetation. In several locations repetitive samplings were conducted simultaneously or at a later time, in order to ensure the reproducibility of the results obtained. The collected samples were brought to the laboratory, air-dried under ambient temperature and after all foreign particles, such as glass pebbles, rubbish or grass were removed, their water content was determined. An ample quantity of carefully cleaned soil was used to ﬁll a 0.282 cc plastic cylindrical box, of 70 mm height. The boxes were weighed, hermetically sealed and covered with a ﬁlm of epoxy resin to ensure that no gas would escape [5].

3. Experimental procedure All the collected soil samples (1674) have been analyzed at NES-NTUA within the period 1986–1999 using high-resolution high efﬁciency Ge detector set-ups, for the determination of 226 Ra, 232 Th, 40 K. Two of these nuclides, 226 Ra and 232 Th, were determined indirectly, through the radon and thoron progeny in equilibrium, respectively. Details about the techniques used may be found in [8]. According to the results obtained from this analysis, the natural radioactivity content of Greek surface soils is rather low, namely: 226 Ra: 25 ± 19 Bq kg−1 , 228 Ra: 21 ± 16 Bq kg−1 and 40 K: 355 ± 220 Bq kg−1 [5]. In the framework of this research, a subset of 286 soil samples was further analyzed using high resolution Ge detectors, with high efﬁciency in the low energy region, such as a LEGe detector (530 eV fwhm @ 122 keV of 57 Co) and an XtRa detector (1003 eV fwhm @ 122 keV), aiming at the determination of 238 U and 210 Pb. For the γ-spectroscopic analysis the in-house developed UNIX software SPUNAL was used. SPUNAL analyses complex γ-spectra, with up to 10 component multiplets, using a modiﬁed Marquardt algorithm, and it is suitable for the detection and the analysis of close-lying and/or overlapping peaks which may constitute the components of a multiplet [9]. Characteristic cases are the photopeak at ∼186 keV from

178

M.J. Anagnostakis et al.

the adjacent 185.72 (235 U) and 186.25 keV (226 Ra) photons, and the photopeak at ∼ 92.5 keV from the adjacent 92.38 and 92.80 keV photons of 234 Th. In cases where the LEGe detector was used, it was found that the analysis of the multiplet photopeak at ∼186 into its two components, at 185.75 and 186.25 keV, leads directly to the accurate determination of 235 U and 226 Ra. However, for the low speciﬁc activity soil samples, this analysis requires long analysis times, up to 5 × 105 s, to obtain good enough statistics to permit the analysis of the multiplet photopeak. In the case of the analysis of a spectrum collected in the energy region 40–200 keV, efﬁciency corrections due to self-absorption are automatically performed by the software used, based on the technique reported in [10]. According to this technique, the detector is calibrated for efﬁciency using a standard calibration source. The resulting efﬁciency is then corrected to take into account the difference in the self-absorbing properties between the materials of the sample to be analyzed and the calibration source at the energy of interest, using an efﬁciency correction factor η. The values of the linear attenuation coefﬁcient μ for soil, which, among others, are needed for the calculation of η, are provided by the correlation μ(E, ρ) introduced in [9]. By applying the above efﬁciency correction technique, it was found that for the energy of 46.54 keV (210 Pb) and the sample geometry used, the efﬁciency for the lighter soil analyzed (0.53 g cm−3 ) was as much as 2.2 times higher than the efﬁciency for the heavier soil (1.85 g cm−3 ), which shows the need for self-absorption corrections in this kind of measurements. For every sample analyzed in the framework of this research: • The 238 U activity was determined indirectly from the 63.29 keV photons emitted by its decay product 234 Th. When the multiplet photopeak at ∼ 186 keV was analyzed, 238 U was also determined indirectly from the 185.72 keV photons of 235 U, under the assumption of a natural isotopic abundance of uranium in the soil. • The 210 Pb activity was determined using its 46.54 keV photons. Since many of the samples were analyzed for 210 Pb a long time after the sampling, decay corrections were performed for the determination of the initial 210 Pb activity of the sample at the sampling date. For those samples that were also analyzed with the LEGe detector calibrated at the low energy region, and in the cases where the multiplet photopeak at ∼ 186 keV was analyzed into its two components: •

238 U was determined indirectly from the 185.72 keV photons of 235 U, under the assumption

•

226 Ra

of a natural isotopic abundance of uranium in the soil. was determined directly from its 186.25 keV photons.

4. Experimental results The results of the determination of 238 U and 226 Ra in some samples which were analyzed using both the XtRa detector and the LEGe detector are presented in Table 1. From these results, it may be concluded that the analysis of the multiplet photopeak at 186 keV into its two components at 185.72 (235 U) and 186.25 keV (226 Ra), leads to the accurate determination of 238 U and 226 Ra, respectively.

238 U and its daughter products in Greek surface soils

179

Table 1 Determination of 238 U and 226 Ra, for some samples Sample code

X0237 X0292 X0338 X0554 X0640 X1005 X1024 X1664

226 Ra activity (Bq kg−1 ) ± total error,

238 U activity (Bq kg−1 ) ± total error,

determined from γ-emitting radon progeny

186.26 keV

determined from 63.29 keV (234 Th)

185.72 keV (235 U)

145 ± 6 75 ± 4 84 ± 5 104 ± 6 79 ± 4 133 ± 6 147 ± 6 280 ± 12

152 ± 17 75 ± 18 96 ± 21 127 ± 26 95 ± 21 139 ± 26 158 ± 35 269 ± 41

166 ± 17 86 ± 13 15 ± 4 82 ± 14 105 ± 16 67 ± 12 40 ± 11 58 ± 19

145 ± 17 86 ± 19 14 ± 4 79 ± 19 88 ± 20 66 ± 17 43 ± 19 60 ± 29

Table 2 238 U, 226 Ra and 210 Pb radioactivity of Greek surface soils (in Bq kg−1 ) Radionuclide

Sample size

238 U

286 286 1674 285

226 Ra 226 Ra 210 Pb

Arithmetic mean ± standard deviation 40 ± 37 37 ± 39 28 ± 13 83 ± 60

Range 4–480 0.5–372 0.5–372 19–434

The results of the analysis of the 286 samples of the sub-set analyzed during the present work for 238 U, 226 Ra and 210 Pb, as well as the results of the analysis of the whole data set (1674 samples) for 226 Ra compiled from [5], are presented in Table 2. It is interesting to notice in this table that the mean value for 226 Ra calculated from the data sub-set is higher than that calculated from the whole set of the 1674 samples. This clearly shows the effect of the sample size, and indicates that a larger sub-set out of the whole data set should be analyzed for 238 U and 210 Pb, in order to obtain representative results for the whole country. Figure 1 presents the correlation between the speciﬁc activities of 238 U and 226 Ra, and Fig. 2 that between 210 Pb and 226 Ra, with correlation coefﬁcients of 0.7829 and 0.6663, respectively. The correlation coefﬁcient between 210 Pb and 226 Ra reported in this work is lower than the value 0.8745 reported in [6] for 64 surface soil samples, collected up to the thickness of 5–10 cm. The amount of data for 238 U and 210 Pb that were distributed over the whole country allowed for the mapping of 238 U and 210 Pb. The mapping was performed using the in-house built Data Base/Geographical Information System (DBGIS). This system contains detailed information regarding each soil sample collected, including the geographical coordinates of the sampling location, and in-house developed statistical and plotting routines, which allow the mapping of the data in various formats and at any desirable scale, including the drawing of isolines. This DBGIS has been used for mapping 226 Ra, 232 Th and 40 K [5], using the 1674 sample data set. The mappings of 238 U and 210 Pb, respectively, produced during this work from the

180

M.J. Anagnostakis et al.

Fig. 1. Correlation between 238 U and 226 Ra in surface soil.

Fig. 2. Correlation between 210 Pb and 226 Ra in surface soil.

sub-set, are presented in Figs. 3 and 4, respectively. Both mappings are presented in the form of a four-class histogram, obviously different for each one of the radionuclides investigated. By comparing these mappings and the 226 Ra mapping from [5] it was concluded that: (1) the 226 Ra and 238 U mappings are quite similar, though in the case of 226 Ra the mapping is more detailed due to the larger sample size; (2) the 210 Pb mapping presents signiﬁcant differences from the other two, indicating in many cases disturbance of radioactive equilibrium within the uranium series.

238 U and its daughter products in Greek surface soils

Fig. 3. 238 U mapping of Greek surface soils.

Fig. 4. 210 Pb mapping of Greek surface soils.

181

182

M.J. Anagnostakis et al.

Fig. 5. 226 Ra/238 U mapping of Greek surface soils.

Fig. 6. 210 Pb/226 Ra mapping of Greek surface soils.

238 U and its daughter products in Greek surface soils

183

5. Radioactive equilibrium disturbance within the uranium series From the speciﬁc activities of 238 U, 226 Ra and 210 Pb, the ratios 210 Pb/226 Ra and 226 Ra/238 U were calculated at each sampling point. The 226 Ra/238 U ratio proved to follow a normal distribution while that of 210 Pb/226 Ra was almost lognormal. Figures 5 and 6 present the mappings of 226 Ra/238 U and 210 Pb/226 Ra, respectively. Table 3 presents the results of these ratios for the whole country and Fig. 7 presents both ratios for each sampling point. The 226 Ra/238 U ratio ranges from the unique very low value of 0.06 to the value of 4.8, having a mean value not signiﬁcantly different from 1. The lowest value observed was in a sample collected at the Katara col, at a height of 1690 m, which is quite often covered by snow during wintertime. In this sample the lowest 226 Ra content (0.5 Bq kg−1 ) was detected, while the 238 U activity was also relatively low (8 Bq kg−1 ). Regarding the 210 Pb/226 Ra ratio of this sample, an extreme value (71) was observed. This is an indication that 226 Ra has been removed from the surface soil presumably due to the agency of water. The highest value for 226 Ra/238 U (4.8) was observed in a sample that was collected from a plateau on Olympus mountain, at a height of 2700 m, that is covered with snow for most of the winter and spring. A total of 16 samTable 3 Radioactive equilibrium disturbance in Greek surface soil Activity ratio

Sample size

Mean value ± standard deviation (σ )

Geometric mean

Range

Upper 5%

226 Ra/238 U

286 285

0.94 ± 0.61 3.5 ± 5.0

0.95 ± 0.38 4.2 ± 4.3

0.06–4.8 0.8–71

16 samples, ratio > 2 16 samples ratio > 8

210 Pb/226 Ra

Fig. 7. The 226 Ra/238 U and 210 Pb/226 Ra ratios for the 286 sampling points (points with 210 Pb/226 Ra ratio higher than 10 are not presented).

184

M.J. Anagnostakis et al.

ples (the upper ∼ 5%) have 226 Ra/238 U ratio higher than 2. A characteristic case with high 226 Ra/238 U values is that of the Southeastern part of Peloponnesus. The mean value of 210 Pb/226 Ra ratio is much higher than 1 (3.5) indicating signiﬁcant disturbance of equilibrium in surface soil. A total of 16 samples (about 5%) have values of the ratio 210 Pb/226 Ra higher than 8, while there was no sample with ratio signiﬁcantly lower than 1. In areas where lignite-burning power plants operate, radioactive equilibrium disturbance is relatively low, compared to the rest of the country, indicating that the human activity is not the major contributor to this disturbance. Furthermore, there exist some rather wide areas with signiﬁcant disturbance of radioactive equilibrium between 210 Pb and 226 Ra, such as: (1) a large area about Lamia–Aidipsos–Atalanti, where a lot of tectonic faults and spas are located, and (2) a large area in the central part of Greece about Ioannina.

6. The activity ratios of 226 Ra/232 Th and 238 U/232 Th For every sampling point, the activity ratios 226 Ra/232 Th and 238 U/232 Th were also calculated. From Table 4, where the results are presented, it is concluded that these ratios may have very high values, up to 110, which in most cases should be attributed to the very low 232 Th and not to high 226 Ra concentrations, of the respective samples, as is clear from Fig. 8. Table 4 226 Ra/232 Th and 238 U/232 Th activity ratios of Greek surface soils Activity ratio

Sample size

226 Ra/232 Th 1578 238 U/232 Th 285

Arithmetic mean value ± standard deviation (σ )

Geometric mean ± standard deviation (σ )

Range

Upper 5%

1.82 ± 4.18 2.93 ± 8.20

1.52 ± 1.57 2.31 ± 3.74

0.26–110 0.3–127

90 samples ratio > 4.5 14 samples ratio > 8

Fig. 8. Dependence of the 226 Ra/232 Th ratio on the 226 Ra, 232 Th activity of the samples.

238 U and its daughter products in Greek surface soils

185

Fig. 9. Mapping of the 226 Ra/232 Th activity ratio.

High values of 226 Ra/232 Th, up to 8, are reported in [11]; according to this report, Ra-rich areas in Estonia (> 100 Bq kg−1 ) were characterized by a 226 Ra/232 Th ratio higher than 2. The mapping of 226 Ra/232 Th over Greece is presented in Fig. 9. In 90 samples (about 5%) the value of 226 Ra/232 Th is higher than 4.5. Surprisingly, most of the samples with values of the ratio higher than 5 are very well localized, for example around the lignite ﬁeld basin of Megalopolis and the island of Corfu in the North-West part of Greece.

7. Conclusions The results of this work lead to the following conclusions: (1) There exist locations within Greece with signiﬁcant disturbance of radioactive equilibrium within the uranium series in surface soil, and very high 226 Ra/238 U and 210 Pb/226 Ra values. (2) In most cases, this disturbance cannot be attributed to human activities such as fossil fuel burning, and the reasons for this should be further investigated. (3) Some of the highest values of 210 Pb/226 Ra were found in areas with spas and tectonic faults and it should be further investigated whether the existence of spas and faults are the cause of such disturbances. In these areas the vertical proﬁle of the disturbance should be investigated. Furthermore, the dependence of the 210 Pb/226 Ra ratio with the radon

186

M.J. Anagnostakis et al.

potential of soil gas should be investigated in locations with high values of 210 Pb/226 Ra ratio. (4) The most signiﬁcant disturbance of the 226 Ra/238 U ratio was observed at very high altitudes, where the agency of water is more pronounced. (5) According to the results obtained, very high 226 Ra/232 Th ratios are due to very low 232 Th activity. From the present work it is concluded that radioactive equilibrium disturbance – sometimes signiﬁcant – within the uranium series is a very common phenomenon in surface soils, and there are indications that the natural radioactivity of surface soils is in some cases different from that of deeper soil layers. This conclusion should be taken into consideration when external exposure of humans due to natural radionuclides of terrestrial origin is calculated by simple models, which are based on the measured activity of speciﬁc natural radionuclides in surface soil, and the assumption of radioactive equilibrium.

References [1] UNSCEAR, United Nations Scientiﬁc Committee on the Effects of Atomic Radiation Report, New York, 1988. [2] D.G. Coles, J.W.T. Meadows, The direct measurement of ppm levels of uranium in soils using high-resolution Ge(Li) gamma-ray spectroscopy, Sci. Total Environ. 5 (1976) 171–179. [3] F.B. Ribeiro, A. Roque, P.C. Boggiani, J.-M. Flexor, Uranium and thorium disequilibrium in quaternary carbonate deposits from the Serra da Bodoquena and Pantanal do Miranda, Mato Grosso do Sul State, central Brazil, Appl. Radiat. Isot. 54 (2001) 153–173. [4] M. Dowdall, J. O’Dea, 226 Ra/238 U disequilibrium in an upland organic soil exhibiting elevated natural radioactivity, J. Environ. Radioact. 59 (2002) 91–104. [5] M.J. Anagnostakis, E.P. Hinis, S.E. Simopoulos, M.G. Angelopoulos, Natural radioactivity mapping of Greek surface soils, Environ. Int. 22 (S1) (1996) S3–S8. [6] B. Myslek-Laurikainen, S. Wolkowicz, R. Strzelecki, M. Biernacka, M. Matul, 210 Pb concentration in soils in Poland and its behavior in radon rich regions, Proceedings of a Workshop on Radon in the Living Environment, Athens, Greece, April 19–23, 1999, Sci. Total Environ. 272 (1–3) (2001) 879–888. [7] N. Kaunakara, D.N. Avadhani, H.M. Mahesh, H.M. Somashekarappa, Y. Narayana, K. Siddappa, Distribution and enrichement of 210 Po in the environment of Kaiga in South India, J. Environ. Radioact. 51 (2000) 349–362. [8] S.E. Simopoulos, M.G. Angelopoulos, Natural radioactivity releases from lignite power plants in Greece, J. Environ. Radioact. 5 (1987) 379–389. [9] N.P. Petropoulos, M.J. Anagnostakis, S.E. Simopoulos, Photon attenuation, natural radioactivity content and radon exhalation rate of building materials, J. Environ. Radioact. 61 (2002) 257–269. [10] M.J. Anagnostakis, S.E. Simopoulos, An experimental and numerical method for the efﬁciency calibration of low-energy germanium detectors, Environ. Int. 22 (S1) (1996) S93–S99. [11] E. Realo, K. Realo, Natural radionuclides in radium-rich soils in North-East Estonia, in: J.P. McLaughlin, S.E. Simopoulos, F. Steinhäusler (Eds.), The Natural Radiation Environment VII, Proc. VIIth International Symposium on the NRE, Rhodes, Greece, 20–24 May 2002, in: Radioactivity in the Environment, vol. 7, Elsevier, Amsterdam, 2004, this volume.

187

Data leading to the investigation of a relation between seismic activity and airborne radon decay product concentrations outdoors D.J. Karangelos, N.P. Petropoulos, M.J. Anagnostakis, E.P. Hinis, S.E. Simopoulos Nuclear Engineering Section, Mechanical Engineering Department, National Technical University of Athens, 15780 Athens, Greece

Systematic monitoring of airborne radon decay product concentrations outdoors has been a part of the Athens area environmental radioactivity monitoring program of the Nuclear Engineering Section of the National Technical University of Athens (NES–NTUA) since the mid-1980s. One of the monitoring systems consists of a sodium-iodide (NaI) detector and appropriate electronics including a spectrum stabilizer suitable to compensate for spectrum shifts mainly due to large temperature variations. The photons to be monitored are mainly those of 214 Bi (609 keV); single channel analysis in regular time intervals is being employed. Data logging is fully automated. Through this 15-year long operation of the system, it was observed that there have been cases when the radon decay product concentration outdoors was signiﬁcantly and unreasonably increased during periods lasting from 1 hour to 2 days. A-posteriori analysis showed that 6 to 10 days later a pronounced seismic activity was observed within a radius of about 300 km. This ﬁnding suggested that an increase of radon decay product concentration outdoors might be an earthquake precursory signal for the area under study. This paper presents a ﬁrst attempt at the analysis of such data. The analysis results indicate that the monitoring method gives a reliable indication of an earthquake within a reasonable time window. The method may prove useful if combined with other similar earthquake forerunner tools. 1. Introduction The continuous monitoring of airborne radon decay product concentrations outdoors has been a part of the Athens area environmental radioactivity monitoring programme of the Nuclear Engineering Section of the National Technical University of Athens (NES–NTUA), since approximately 1986. The main monitoring system consists of a sodium-iodide (NaI) detector RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07021-4

© 2005 Elsevier Ltd. All rights reserved.

188

D.J. Karangelos et al.

and appropriate electronics including a spectrum stabilizer suitable to compensate for spectrum shifts mainly due to large temperature variations. The radon progeny photons monitored are mainly those of 214 Bi (609 keV) employing single channel analysis. The measurement rate is automatically adjusted if an increase in the count rate is observed. Each single channel measurement lasts 100 s. Data acquisition is computer driven and fully automated. Through this 15-year long operation of the system, it was observed that there have been cases when the radon decay product concentrations outdoors were signiﬁcantly and unreasonably increased during periods lasting from 1 hour to 2 days. A-posteriori analysis showed that 6 to 10 days later a pronounced seismic activity was observed within a radius of about 300 km. Furthermore, it was heuristically concluded that: (a) a radon decay product concentration increase has been detected prior to all earthquakes of magnitude > 4.5 R within a radius of 200 km from Athens, as reported by the Institute of Geodynamics of the National Observatory of Athens (GI–NOA) [1]; (b) a radon decay product concentration increase has been detected prior to all reported earthquakes of magnitude > 3.5 R within a radius of 50 km from Athens; and (c) a small number of false positive increases have been detected (about 5–10%). These conclusions suggested that an increase of radon decay product concentration outdoors might be an earthquake precursory signal for the area under study. Yet, no wellestablished justiﬁcation of the phenomenon was pursued. The September 1999 Athens earthquake of magnitude 5.9 R (115 deaths, multimillion EURO damage), preceded by a distinct signal, initiated a systematic study of radon decay products concentration increases outdoors as earthquake precursory signals, in the Athens wider area. Figure 1 is a graph of both the recorded data for the 214 Bi photons in counts per 100 s and the earthquakes in magnitude

Fig. 1. The September 1999 Athens earthquake and its respective precursory signal: +: earthquake events; |: 214 Bi counts per 100 s.

Seismic activity and airborne radon decay product concentrations outdoors

189

measured on the Richter scale reported by the GI–NOA in 1999 for Greece. It is evident that there exists a pronounced signal nine days before the seismic event.

2. Methods Following the September 1999 Athens earthquake, adaptive control of sampling intervals was introduced so that every case of signiﬁcant radon decay product concentration increase (> 200 counts per 100 s at the 609 keV peak of 214 Bi) is represented with adequate statistics. Furthermore, data collection was enhanced to record humidity and a second single channel analyzer was added to record the total count rate in the energy region above the electronic noise threshold up to approximately 2000 keV; both these parameters might indicate a probable false positive signal, which has to be ﬁltered out. The sampling location was moved from the ground level to the roof of a seven-meter high building. In addition, spectrum collection over a 24 h period was also employed. Data acquisition was fully computerized and feeding a database. Close analysis of the recorded NaI spectra collected over 2 years (2000–2001) proved that they contain integrated information within which signals from radon decay product concentration increases lasting up to a few hours might be averaged and lost. Therefore, it was decided that only the recorded data for the 214 Bi photons in counts per 100 s should be further evaluated and treated. Figure 2 is a graph of both the recorded data for the 214 Bi photons in counts per 100 s and the recorded earthquakes in magnitude measured on the Richter scale in 2001. A simple comparison with the respective 1999 data indicates that there is evident improvement in the precursory signal recordings in terms of signal magnitude and signal discrimination from background. The improvement may be attributed to the sampling location change. The humidity and the total γ-background recordings showed that comparatively high

Fig. 2. The precursory signals received and the earthquakes reported for 2001 within Greece.

190

D.J. Karangelos et al.

signal values could not be considered as precursory if humidity and/or total γ-background is increased as well.

3. Raw data analysis The total number of 214 Bi data points for year 2001 is approximately 2900 (about eight sampling intervals per day). Figure 3 is a graphic representation of the inﬂuence of the total count rate and humidity on the recorded data. The total count rate may be increased due to several reasons, the most important of which are: (a) existence of artiﬁcial radioactivity in the ambient air; (b) washout and rainout processes, which result in increased deposition of radon decay products near and on the surface of the sodium iodide detector; (c) increased radon concentration outdoors, due to phenomena such as diurnal variation, which consequently result in increased progeny concentration recordings. Reason (a) can be safely excluded for the duration of 2001, when no nuclear weapon testing and/or major accidental releases of radionuclides from nuclear power plants were reported. Reasons (b) and (c) are certainly valid throughout the examined period. Their effect on the data may be summarized as follows: 1. Recorded data, when relative humidity is higher than 85%, are a consequence of washout and/or rainout deposition processes; thus they may not be considered as earthquake precursory signals. 2. Recorded data that satisfy the condition: Data value > 0.03x + 7

Fig. 3. Inﬂuence of total count rate and humidity on the recorded data.

(1)

Seismic activity and airborne radon decay product concentrations outdoors

191

Fig. 4. The distribution of the recorded data in year 2001.

where x = the respective value of the total count rate, that is the boundary line set by ﬁtted function F 1, should be further examined for consideration as earthquake precursory signals. Figure 4 presents the distribution of the recorded data for the 214 Bi photons in counts per 100 s (cps100 ) in the year 2001. The mean value of the 2900 recordings is calculated equal to 195 ± 28 cps100 . It is anticipated that these data, being in their essence radon progeny air concentration data, should follow a normal or lognormal distribution – this has been concluded in many, mainly indoor air, studies, using integrating instrumentation [2]. However, in this case only data with values up to 193 cps100 follow the normal distribution at the 95% conﬁdence level, with mean μ = 183 cps100 and standard deviation σ = 4 cps100 . Furthermore, the same data with values up to 193 cps100 also follow the lognormal distribution at the 95% conﬁdence level, with geometric mean μg = 183 cps100 and geometric standard deviation σg = 4 cps100 . Such data represent about 70% of the total recordings. The remaining 30% of the data are considered as outliers. This high percentage of outliers may be attributed to: (a) the adequately chosen sampling time (100 s), which prevents integration or averaging of short lasting radon concentration increases; (b) the biasing of the frequency histogram due to the adaptive control of sampling intervals. This means that the number of sampling intervals within a day is increased when a value greater than 200 counts per 100 s (≈ μ+3σ ) is recorded; a fact that signiﬁcantly increases the possibilities of more outlier recordings. Following this preview, the recorded data frequency histogram was successfully deconvoluted at the 95% conﬁdence level to two separate lognormal distribution peaks: (i) one with geometric mean value μg = 183 cps100 and geometric standard deviation σg = 4 cps100 ;

192

D.J. Karangelos et al.

(ii) one with geometric mean value μg = 201 cps100 and geometric standard deviation σg = 50 cps100 . Distribution (i) represents the anticipated lognormal distribution for the concentration of radon progeny in air. Distribution (ii) represents excess data recordings which were conducted due to the automatic increase of sampling intervals if a value greater than 200 is measured. It is reasonable to assume that distribution (ii) contains data which have to be considered as precursory earthquake signals. To facilitate an initial and possibly straightforward association of signals with potentially detectable earthquakes, the nominal threshold in order to accept data as precursory earthquake signals is set at 220 cps100 (≈ μg + 0.4σg ). Considering the characteristics of distribution (ii) the possibility of recording such signal values is approximately 35%. Last but not least Fast Fourier Transform analysis was performed on all collected data for year 2001; it was concluded that there is no apparent signal periodicity. In synopsis, according to the above analysis, recorded data to be considered as precursory earthquake signals are: (a) those recorded when relative humidity is lower than 85%; (b) those that satisfy inequality (1); (c) those that are higher than 220 cps100 at the 609 keV of 214 Bi. This means that within year 2001, about 300 such signals satisfy these three conditions (approximately 10% of the total). The same analysis has been performed on the recorded data of year 2000 with similar results.

4. Existing information on precursory signals of atmospheric origin At present it seems difﬁcult to suggest a sound plausible explanation for the received precursory signals. The simplest idea for these radon progeny concentration increases is that they might be connected to intensiﬁed radon exhalation around the focal point before an earthquake, e.g., [3]. Nevertheless, there is no experimental evidence that radon concentration in ambient air increases around the monitoring station within the expected precursor time. Yet, the phenomenon might be a side effect of such intensive radon exhalation occurring well before fault rupture, inasmuch as excess radon might result in changes of atmospheric conductivity or changes of the electric ﬁeld, as suggested in [4] and discussed for example in [5], and [6]. Scientists of [7], however, though they agree that there exists radon exhalation before strong earthquakes, argue that it is the metallic aerosols emanated along with radon from the focal point area that produce atmospheric electric ﬁeld modiﬁcations. Despite the fact that the respected review of the behaviour of radon progeny in air compiled by [8] neglects the electric mechanisms as part of the physical processes governing radon progeny atmospheric deposition, convincing examples show that electric ﬁelds might have an essential inﬂuence [9–14]. According to [12], the electric mechanism of deposition appears important in conditions of low wind speed and is further enhanced under thunderclouds and high voltage power lines. Furthermore, as experimentally tested by [15] the electrical component of electromagnetic ﬁelds could signiﬁcantly affect the movement of charged decay products; in addition induced

Seismic activity and airborne radon decay product concentrations outdoors

193

electric ﬁelds should be taken into account especially in strong magnetic ﬁelds. Such changes might affect radon progeny deposition rates for short time periods. Further to these ideas, [5] reviews some of the accumulated knowledge regarding the seismo-electromagnetic effects in the atmosphere occurring a few hours to a few days prior to earthquakes, due to other kinds of lithospheric processes. The resulting seismic electric ﬁelds penetrate up to the ionosphere with the electric ﬁeld at the Earth’s surface estimated in the order of 103 –104 V m−1 , within a frequency range between 1 and 103 Hz. Such electromagnetic conditions are quite similar to those observed in the vicinity of high voltage power lines, which highly affect radon progeny deposition rate, as reported by [11]. Therefore, it seems again reasonable to hypothesize that the radon progeny variations recorded in the monitoring station might be in fact radon progeny deposition rate increases due to certain seismo-electromagnetic effect propagation through the atmosphere. It is worth mentioning that [16] notes that atmosphere and ionosphere modiﬁcations are inherent up to several days before earthquakes of magnitude M 4 with focus depths less than 200 km. These modiﬁcations are observable in zones known as the earthquake preparation zones. The radius logarithm of such areas around the focal point is proportional to the earthquake magnitude and ranges up to 250 km for earthquakes up to magnitude 5. These observations justify recordings of precursory signals at the monitoring station, associated with earthquakes occurring at considerable distances. The discussion presented approximates an initial cause–effect relationship between earthquakes and received precursory signals.

5. Signal association with earthquakes Despite the fact that there exists no plausible explanation of a relation between seismic activity and airborne radon decay product concentration precursory signals, we proceeded with a technical analysis of our data, to obtain a ﬁrst step towards signal association with earthquakes. The experience accumulated by NES–NTUA in the past 15 years with the described monitoring method leads to the result that the signals satisfy the ﬁltering conditions (a), (b) and (c) presented at the end of Section 3 and that the signal to earthquake association analysis should be based on the experimental evidence that: (1) the precursory signals are received within the earthquake dilatancy generation stage; thus the signals are considered as short-term precursors with precursor time in the range up to 14 days; (2) the precursory signals received within a time span of two days may be associated with one or more earthquakes, which happen within the precursor time range; (3) precursory signals are generally received prior to earthquakes of magnitude M 4, within a distance of about 300 km. The earthquake data used for the analysis are those provided by the Institute of Geodynamics of the National Observatory of Athens, as published in its web site for year 2001 (a total of 2671 events) [1]. In order to take into account notes (1) and (2) above, the precursory signals data set is divided into 180 successive segments, each lasting two days. The resulting

194

D.J. Karangelos et al.

Table 1 Precursory signals to earthquake association analysis results for year 2001 Precursory signal (day-of-the-year)

Earthquake (day-of-the-year)

Magnitude of earthquake (R)

Epicentre distance (km)

Depth (km)

Precursor time (days)

105 109 139 159 165 187 205 213 215 231 235 241 247 269 275 287 299 315 317 327 341

117 120 148 167 174 NA-1 NA-2 221 222 242 246 249 258 278 287 297 305 322 326 341 NA-1

4.6 4.9 5.1 4.4 4.7

316 428 444 296 427

37 29 5 5 5

12 11 11 8 9

4.1 3.6 4.4 3.5 3.8 5.2 4.4 4.6 3.5 4.0 3.5 4.0 5.0

122 127 331 122 94 188 379 278 110 241 19 124 156

9 5 5 5 10 5 5 5 6 5 35 16 24

8 7 11 11 8 11 9 12 10 6 7 9 14

Mean Precursor Time:

9.6

NA ≡ Not Associated.

signal distribution out of this process revealed a number of cases when one or more precursory signals have been recorded within this time span of two days. These signals have been received on days 105, 109, 139, 151, 159, 165, 187, 205, 213, 215, 231, 235, 241, 247, 269, 275, 287, 299, 315, 317, 327, 341, 355, 357, 359, 361, 363 and 365 of year 2001. Signals received between days 355 and 365 are not being considered in this association analysis for year 2001, since their respective earthquake might have happened in year 2002. The method further used to associate a precursory signal with an earthquake is based on note (3) above. The earthquakes of local magnitude ML 3.5 recorded within 14 days after a signal are sorted by magnitude. The earthquake with the maximum magnitude reported within this 14 day span is considered to be the earthquake most probably associated with the signal. In case more than one earthquake has approximately the same maximum magnitude in this time span, then the earthquake closest to the monitoring station is considered as the one detected. Moreover, if this distance is also approximately equal, then the earthquake which occurred closest to the earth’s surface is qualiﬁed as detected. Table 1 presents the results of such an association of the precursory signals with earthquakes for 2001. Signals characterized as NA-1 could not be associated with any earthquake with magnitude ML > 3.5. The signal characterized as NA-2 may be associated with more than one earthquake with magnitude ML > 3.5. Following these results, the precursor time is calculated at 9 ± 2 days (range: 7 to 14 days). The precursor

Seismic activity and airborne radon decay product concentrations outdoors

195

Fig. 5. Epicentral distance of earthquakes from the monitoring station (R in km) vs earthquake magnitude.

time results seem to agree well with the 9 ± 1 day reported in the work of [17] concerning γ-background variations before large-scale regional earthquakes. All the earthquakes of year 2001 with magnitude ML 3.5 are graphically represented as dots in Fig. 5. Furthermore, the detected earthquakes are noted in the same graph using hollow circles. This plot does not imply any relation between earthquake magnitudes and epicentral distance from the monitoring station, but it was adopted to give a visual impression of the method proposed. It is obvious from Fig. 5 that the detected earthquakes can be identiﬁed provided that they fall underneath the regression line: Log10 (R) = 0.41ML + 0.53 (correlation coefﬁcient: 0.68)

(2)

where: R is the epicenter distance from the monitoring station in km, and ML is the earthquake local magnitude. It is believed that this linear relation between magnitude ML and distance R of detected earthquakes presented in Fig. 5 successfully averages in a qualitative way the precursory signal detection capability of the monitoring station examined. Undetected earthquakes lying below the line in Fig. 5 are those which could not be associated with any signal mainly due to the ﬁltering constraints (a), (b) and (c) presented at the end of Section 4. Data collected in years prior to 2001, as well as data collected in early 2002, give substantial evidence that earthquakes lying below this line have been detected by the system presented. The observations of [16] regarding the proportional relationship between the detected earthquake magnitudes and the logarithm of the radius around the area where the precursory signal may be received are in fairly good agreement with the results contained in Table 1 and presented in Fig. 5.

196

D.J. Karangelos et al.

Fig. 6. Detected earthquake precursor time (in days) vs from earthquake magnitude.

Figure 6 is in agreement with the ﬁndings reported by [18] of several short-term precursors, for which the smaller the magnitude, the shorter the precursor time.

6. Future research and conclusions The ultimate scope of full exploitation of every earthquake precursory signal is to successfully correlate it with earthquake magnitude, depth, epicentre coordinates and rupture time. At present, and given the fact that there is no clear cause–effect relationship between received signals and associated earthquakes, this task cannot be pursued using this method. Future research is planned to focus on widening as far as possible the experimental information collected at the monitoring station. To this end: (a) the daily number of sampling intervals will be further increased. It is planned to quasicontinuously cover 24 h with 100 s measurements of the counts at the 609 keV peak of 214 Bi; (b) the system will be complementarily enhanced to monitor photons of 214 Bi at the 1764 keV peak; and (c) monitoring of 212 Pb photons collectively at the 295 and 352 keV peaks will be considered in combination with Compton suppression techniques. In addition, a duplicate monitoring station set-up will be operated from time to time at the same location to exclude all possibilities that recordings of radon progeny concentration increases are not due to any malfunction. Continuous radon concentration monitoring of the ambient air will also be employed to better establish the collected evidence that the radon progeny concentration increases are independent of radon concentration changes in ambient air. An electronic rain detector will also be employed for distinguishing rainout deposition from other recordings. Setting up of two more monitoring stations at a distance of about 100

Seismic activity and airborne radon decay product concentrations outdoors

197

to 150 km from Athens, and the appropriate telemetry, has already been designed. The stations will be placed at preselected locations of seismological interest. Their operation will help to further explain the observed phenomena. Furthermore, it is expected that this small network might be proven useful in estimating at least some of the forthcoming earthquake magnitudes. On the other hand, it seems very unlikely that the earthquake epicentre may be predicted. Anyhow, further experimenting is necessary, but, of course, the experimental conditions are highly unpredictable and undesirable. The method may prove useful if combined with other earthquake precursory tools.

References [1] Institute of Geodynamics–National Observatory of Athens, Earthquake Catalogue Archive, http://www. gein.noa.gr/services/cat.html. [2] UNSCEAR, United Nations Scientiﬁc Committee on the Effects of Atomic Radiation Report, New York, 1988. [3] C.Y. King, Gas geochemistry applied to earthquake prediction: an overview, J. Geophys. Res. 91 (12) (1986) 269–281. [4] E.T. Pierce, Atmospheric electricity and earthquake prediction, Geophys. Res. Lett. 3 (1976) 185–188. [5] M. Parrot, Use of satellites to detect seismo-electromagnetic effects, Adv. Space Res. 15 (11) (1995) 27–35. [6] O.A. Molchanov, M. Hayakawa, T. Oudoh, E. Kawai, Precursory effects in the subionospheric VLF signals for the Kobe earthquake, Phys. Earth Planet. Interiors 105 (1998) 239–248. [7] S.A. Pulinets, V.A. Alekseev, A.D. Legen’ka, V.V. Khegai, Radon and metallic aerosols emanation before strong earthquakes and their role in atmosphere and ionosphere modiﬁcation, Adv. Space Res. 11 (1997) 2173–2176. [8] J. Porstendörfer, Properties and behavior of radon and thoron and their decay products in the air, J. Aerosol Sci. 25 (1994) 219–263. [9] M.H. Wilkening, Inﬂuence of the electric ﬁelds of thunderstorms on radon-222 daughter ion concentrations, in: H. Dolezalek, R. Reiter (Eds.), Electric Processes in Atmospheres, Steinkopf, Darmstadt, 1977. [10] T. Schneider, M. Bohgard, A. Gudmundsson, A semiempirical model for particle deposition on to facial skin and eyes. Role of air currents and electric ﬁelds, J. Aerosol Sci. 25 (1994) 583–593. [11] D.L. Henshaw, A.N. Ross, A.P. Fews, A.W. Preece, Enhanced deposition of radon daughter nuclei in the vicinity of power frequency electromagnetic ﬁelds, Int. J. Radiat. Biol. 69 (1996) 25–38. [12] H. Tammet, V. Kimmel, Electrostatic deposition of radon daughter clusters on the trees, J. Aerosol Sci. 29 (1996) S473–S474. [13] A.P. Fews, D.L. Henshaw, P.A. Keitch, J.J. Close, R.J. Wilding, Increased exposure to pollutant aerosols under high voltage power lines, Int. J. Radiat. Biol. 75 (1999) 1505–1521. [14] A.P. Fews, D.L. Henshaw, R.J. Wilding, P.A. Keitch, Corona ions from powerlines and increased exposure to pollutant aerosols, Int. J. Radiat. Biol. 75 (1999) 1523–1531. [15] K. Oda, J. Saegusa, T. Yamamoto, Behavior of radon progeny in low frequency electromagnetic ﬁelds, Radiat. Measur. 31 (1999) 331–336. [16] V.N. Oraevsky, Y.Y. Ruzhhin, A.K. Depueva, Anomalous global plasma structures as seismoionospheric precursors, Adv. Space Res. 15 (11) (1995) 127–130. [17] I.O. Nevinsky, T.V. Tsvetkova, Gamma-background variations in underground low background setups, Atomnaya Energia 72 (6) (1992) 622. [18] T. Rikitake, Earthquake Prediction, Elsevier, Amsterdam, 1976.

198

Calibration of a HPGe detector for in-situ gamma spectrometry: a comparison between a Monte Carlo based code and an experimental method F. D’Alberti, M. Forte Safety, Security and Radiological Protection Unit, TP 510, Joint Research Centre, I-21020 Ispra (VA), Italy

This paper presents an analysis of data from in-situ gamma spectrometry measurements, carried out by the JRC–SSPRP Unit on uncontaminated soils using efﬁciency calibration factors determined both experimentally and mathematically. The experimental calibration was performed following the traditional procedure, which consists in combining general radiation ﬂux equations with point source measurements at various angles from the detector axis. The mathematical calibration method is based on a computer simulation of the counting conditions (measurement geometry conﬁguration, sample nature, etc.). For this purpose a commercial Monte Carlo based code has been used. In-situ spectra were collected in some undisturbed soils in Ispra region. The two calibration methods were applied for the activity determination of some selected natural radionuclides (U and Th natural series, K-40) and of Cs-137. A necessary input for a correct calibration of in-situ measurements is the relaxation length of the radionuclides in soils, which was assessed by laboratory gamma analysis of core samples, collected at each site investigated. For these selected radionuclides, the in-situ activity calculated with the two calibration methods was then compared with the activity of the core samples. The experimental and the mathematical methods give congruent results for in-situ measurements, both for uniform and non-uniform radionuclide distributions. The comparison between these results and those obtained with laboratory analyses of soil samples can be considered as satisfactory for our environmental applications. The chosen Monte Carlo based code can be used instead of other techniques mainly for its time saving properties in the determination of efﬁciency calibration curves. 1. Introduction The Laboratory for Radioactivity Measurements of the Safety, Security and Radiological Protection Unit, Joint Research Centre of Ispra (Italy), carries out on a regular basis in-situ highresolution gamma spectrometry on soils with a portable HPGe. This work consists of the RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07022-6

© 2005 Elsevier Ltd. All rights reserved.

Calibration of a HPGe detector for in-situ gamma spectrometry

199

comparison of activity values for selected radionuclides (some natural isotopes and Cs-137) using efﬁciency calibration curves obtained with two methods. A “mathematical method”, developed with the Monte Carlo based commercial software ISOCS (Canberra), and an “experimental method”, following the traditional calibration procedure, with point sources at various angles from the detector axis and theoretical considerations on photons transport. The reliability of the two methods is evaluated, for the Laboratory purposes applications, comparing the results obtained on soils with direct in-situ measurements and in laboratory with measurements on core samples taken from the same soils.

2. Materials and methods The main detector speciﬁcations and performance data are summarised in Table 1. The detector is connected to a compact modular data acquisition system (InSpector, Canberra). A software package (Genie-2000, Canberra) manages acquisition and analysis of the spectra. 2.1. Theoretical introduction The following hypotheses are formulated for the vertical concentration proﬁle of radionuclides in soil: uniform distribution for natural radionuclides, exponential decrease with depth for fallout products, where the activity per mass of soil at depth z (expressed in g cm−2 ) is given by S(z) = S0 exp(−z/λ)

(1) (Bq g−1 )

(g cm−2 )

where S0 is the surface activity and λ is the relaxation length of the radionuclide distribution in soil (deﬁned as the depth at which the radionuclide activity is 1/e times the surface activity). The typical in-situ measurements scheme is reproduced in Fig. 1, with the detector facing downwards placed in O at height h (expressed in g cm−2 ) from the soil–air interface. Referring to Fig. 1, let μa and μs (cm2 g−1 ) be the gamma-ray attenuation coefﬁcients in air and soil, A = S0 · λ (Bq cm−2 ), μa h = k and q = 1/μs · 1/λ. Table 1 Detector speciﬁcations and performance data Model Crystal diameter Crystal length Window Endcup Energy range Relative efﬁciency FWHM at 122 keV at 1332 keV Peak/Compton

Canberra GX3018-coaxial HPGe p-type 56 mm 55 mm 0.5 mm Carbon-epoxy 1.5 mm Al 5 keV–10 MeV 32.1% 0.835 keV 1.72 keV 64.8:1

200

F. D’Alberti, M. Forte

Fig. 1. Schematic model for the calculation of the total photon ﬂux seen by the detector.

The equation for the total uncollided gamma-ray ﬂux in air at height k, Φ, are

A E1 (k) − ekq E1 (k + kq) for an exponential distribution, 2

S0 −k φ= e − kE1 (k) for a uniform distribution 2μs ∞ where E1 (ξ ) = ξ e−x /x dx is the exponential integral function. φ=

(2) (3)

2.2. Experimental method Efﬁciency calibrations have been experimentally determined following the standard procedure with point-source measurements at various angles from the detector axis [1,2]. The calibration factors for in-situ measurements can be expressed in terms of full energy peak count rate (N ), speciﬁc activity in the soil (A) and uncollimated gamma-ray ﬂux (φ) by the following expression: Nf /A = (N0 /φ)(Nf /N0 )(φ/A)

(4)

where: • Nf /A is the count rate (counts per second) in the full-energy peak at the gamma energy of the considered nuclide, per unit of speciﬁc activity (Bq cm−2 or Bq g−1 ) of the nuclide in the soil. • N0 /φ is the count rate in the full-energy peak for unit ﬂux (γ cm−2 s−1 ) for gamma rays of the same energy incident perpendicularly on the detector window. • Nf /N0 is the angular correction factor, speciﬁc to the detector at the considered energy, for a selected distribution of the nuclide in the soil. This factor takes into account the nonuniform response of a cylindrical Ge detector to the gamma rays coming from different directions. • φ/A is the total uncollimated ﬂux at the considered energy reaching the detector for unit speciﬁc activity of the nuclide in the soil. These factors were determined with the following procedures.

Calibration of a HPGe detector for in-situ gamma spectrometry

201

N0 /φ was determined by counting a multi-gamma point source placed at 1 meter from the detector window along the detector axis. The source contains radionuclides of known activity covering the energy range 60–1840 keV (241 Am, 109 Cd, 57 Co, 123 Te-m, 51 Cr, 113 Sn, 85 Sr, 137 Cs, 88 Y, 60 Co). In all the spectra the counting uncertainties are less than 2% (in this work all the uncertainties are always referred to 1 sigma, i.e., with a 68% conﬁdence level). The ﬂux at each energy E was corrected for radioactive decay and attenuation. The uncertainties on the calculated N0 /φ values are of the order of 4%, including counting statistics, the certiﬁed uncertainty on the nuclide activity value and the possible imprecision in the distance sourcedetector window (0.5 cm). Eventually, an interpolation curve has been obtained. Nf /N0 can in principle be determined by measuring point sources positioned at a ﬁxed distance at various angles from the detector axis. Let Nϑ /N0 = Rϑ be the ratio between the full energy peak area at E energy in the spectrum of the source positioned at angle ϑ(Nϑ ) and the area in the spectrum with the source on axis (N0 ). The expression for Nf /N0 is: Nf 1 = N0 φ

π/2

φ(ϑ)Rϑ dϑ.

(5)

0

To determine Nf /N0 , we have used point sources of 241 Am, 152 Eu, 137 Cs, 133 Ba, 60 Co and 22 Na measured at the ﬁxed distance of 1 meter from the crystal geometric centre at step increments of cos ϑ = 0.2 with ϑ ranging between 0 and 90 degrees (where ϑ is as in Fig. 1). The net area in each spectrum has been calculated for the photopeak at E energy. The Nf /N0 factors have been calculated with numerical integration of equation (5) for different proﬁles of activity distribution in the soil. Nf /N0 factors are slightly different from 1 only for energies lower than 100 keV, especially for low λ (Nf /N0 = 0.6 for 60 keV and λ = 2 g cm−2 ), while they get a maximum (Nf /N0 = 1.02 for a uniform proﬁle) for energy about 250 keV. φ/A has been calculated, for the vertical proﬁle of interest, from equations (2) and (3) using an analytical approximation of the integral exponential function [3] and using XCOM [4] to calculate the values of μa and μs for the energies of interest. Nf /A has been thus determined from equation (4) where N0 /φ at E energy is obtained from the interpolation curve and Nf /N0 from an interpolation of the values calculated for the speciﬁc λ. The uncertainty on the ﬁnal efﬁciency values is of the order of 4–5%. 2.3. Mathematical method The software used for the mathematical calculations is the “ISOCS” (In-Situ Object Counting System) by Canberra. ISOCS is a Monte Carlo based code [5,6]. This software provides efﬁciency calibration values requiring the user only to input data about the characteristics of the presumed distribution of the sources in the soil and the measurement conﬁguration (sourcedetector displacement and soil physical characteristics). All the other information and data necessary for the calculation of the calibration values are contained in a “characterisation ﬁle”, typical of each detector. Such a ﬁle is properly created at Canberra premises from the geometrical response of the detector to sources located inside a 500 meters radius sphere, centred in the detector, and over photon energies ranging from 45 to 7 MeV. A user friendly interface allows a choice from a wide variety of source characteristics and measurement conﬁgurations.

202

F. D’Alberti, M. Forte

For our purposes, we have chosen a measurement conﬁguration like the one in Fig. 1 where the soil is made of a cylinder, centred on the detector axis. The cylinder can be made of a unique layer with homogeneous radionuclide concentration (uniform proﬁle) or made of n layers with a different concentration of the same radionuclides (exponential proﬁle). Alternatively, to reproduce an exponential proﬁle, the software offers a cylinder template allowing the introduction of the relaxation length value L (cm), speciﬁc for the considered distribution. Eventually, we have calculated the efﬁciency calibration curves referred to the soil mass (“massimetric efﬁciency” expressed in cps Bq−1 g−1 ) or to the soil area (cps Bq−1 m−2 )). In order to determine the accuracy required for the input data, we have evaluated the effect on the calculated efﬁciency of arbitrary variations in all the parameters introduced in the selected model. For this purpose, we have calculated the efﬁciency values for energies in the range 60–2650 keV, varying one parameter at time and maintaining the others unchanged, as described in the following two paragraphs. 2.3.1. Uniform distribution of radionuclides in soil We have assumed the following base conﬁguration: Diameter: 80 m; Depth: 50 cm; Soil composition “Stdsoil ”: H2 O (10%) + SiO2 (67.5%) + Al2 O3 (13.5%) + Fe2 O3 (4.5%) + CO2 (4.5%); Soil density: 1.6 g cm−3 ; Standard atmospheric parameters: T = 20 ◦ C, P = 760 mm Hg, R.H. = 50%. As for geometric parameters, we have observed that ISOCS massimetric efﬁciency gets a maximum asymptotic value, for all the selected energies, for diameters higher than 80 meters and soil depths higher than 50 cm, while efﬁciency is quite the same for soil density ranging from 1.2 to 2 g cm−3 . In this work we deﬁne as “inﬁnite” those parameters that cause efﬁciency values differing by less than 1% from the asymptotic values. As for soil composition inﬂuence, we have evaluated the variations in the calculated efﬁciency with the following compositions: Hsoil : H2 O (20%) + SiO2 (60%) + Al2 O3 (12%) + Fe2 O3 (4%) + CO2 (4%); Fesoil : H2 O (10%) + SiO2 (62%) + Al2 O3 (12%) + Fe2 O3 (12%) + CO2 (4%). The obtained results are reported in Table 2 in terms of ratios between the efﬁciency values calculated with the modiﬁed conﬁguration and with the base conﬁguration. The dependence on the soil composition is critical only for energies < 100 keV. In our case, in which we intend to calculate the activity of natural radionuclides in soils, the lower energy line of interest for in-situ measurements is 295 keV (214 Pb), so we have reasonably assumed the soil composition Stdsoil for all the efﬁciency calculation. The atmospheric conditions during the in-situ measurements were similar to the variations considered for the standard atmospheric parameters. The consequent variations in efﬁciency are not so critical; however the real values of the atmospheric parameters at the measurement time were used in our calculations with ISOCS. In conclusion, for the uniform proﬁle, we have used a cylindrical template with diameter 80 m, depth 50 cm and soil physical parameters of the base geometry.

Calibration of a HPGe detector for in-situ gamma spectrometry

203

Table 2 Ratios between the efﬁciency values calculated with the modiﬁed conﬁguration and with the base conﬁguration for uniform distribution E (keV) 60 88 122 159 320 662 1173 1836 2650

Diameter (m) 300/80

Depth (cm) 80/50

Density (g cm−3 ) 2/1.6

Soil

Soil Fesoil/Stdsoil

T (◦ C) 30/20

P (mm Hg) 720/740

R.H. (%) 30/50

Hsoil/Stdsoil

1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00

1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00

1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00

1.02 1.00 0.99 0.99 0.99 0.99 0.99 0.99 0.99

0.84 0.93 0.97 0.99 1.00 1.00 1.00 1.00 1.00

0.97 0.97 0.97 0.97 0.98 0.98 0.99 0.99 0.99

1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00

1.02 1.02 1.02 1.02 1.01 1.01 1.01 1.01 1.01

2.3.2. Exponential distribution of radionuclides in soil In this case the efﬁciency (expressed in cps Bq−1 m−2 ) increases with increase in diameter, while it decreases with increase in depth, giving the asymptotic values in both cases. As is obvious, the “inﬁnite” diameter and depth are a function of the relaxation length: increasing λ, the “inﬁnite” depth increases and the diameter decreases (e.g., with soil density 1.6 g cm−3 and L = 5 cm, we have inﬁnite depth at 30 cm and inﬁnite diameter at 120 m; with soil density 1.6 g cm−3 for L = 10 cm, we have inﬁnite depth at 50 cm and inﬁnite diameter at 80 m). Concerning the dependence on the soil composition and atmospheric parameters, the considerations are analogous to the uniform distribution case, while, for a ﬁxed L, the efﬁciency decreases signiﬁcantly with increase in soil density. In conclusion, for a speciﬁc exponential distribution, an evaluation of the appropriated geometric parameters is required as well as the knowledge of the real soil density. It is recommended by ISOCS speciﬁcations to associate the following uncertainties to the calculated efﬁciency values: 10.5% for energies < 100 keV, 7.5% for energies up to 400 keV, 4.5% for higher energies. We have used these uncertainties for the error propagation in activity calculations.

3. Results and discussion The soils under study were plain and apparently undisturbed, near Ispra site (northern Italy). Four sites were considered, named: S1, S2, S3, S4. The in-situ spectra were collected in spring–summer 2000. Acquisition time was set to two hours. The detector was placed on a tripod, facing downward at 1 meter height above the soil–air interface. After each measurement soil samples were collected by coring at step increments of 5 cm to a 30 cm depth. The core samples corresponding to the same soil layers were mixed together, placed directly in a 1 litre Marinelli beaker, counted for 1000 minutes with a shielded coaxial HPGe (42% efﬁciency) and analysed with the appropriate laboratory calibration. The in-situ spectra were

204

F. D’Alberti, M. Forte

Table 3 Nuclides and gamma lines used for in-situ-laboratory comparison Nuclide

40 K

137 Cs

208 Tl

214 Bi

214 Pb

228 Ac

Energy (keV)

1460.86

661.66

583.19 860.56

609.32 1120.29 1764.52

295.21 351.92

338.40 911.07

analysed using both experimental and ISOCS efﬁciency curves, with the appropriate “inﬁnite” parameters. The in-situ counting uncertainties for the selected lines are between 3% and 8%. In Table 3 we report the nuclides used for the comparison, with the related gamma emissions. The core sample analyses show a near uniform distribution of natural radionuclides in the soil, while the activity of 137 Cs decreases with depth in soils S2 and S3. In soil S1, the maximum 137 Cs concentration is in the layer 5–10 cm, while in soil S4 the 137 Cs concentration is quite uniform (probably due to soil mixing after Chernobyl fallout). For in-situ measurements of natural radionuclides we have used the efﬁciency calibration curve for a uniform distribution. For 137 Cs we have determined the L value for each site with a least squares ﬁt of the concentrations in the different layers. In this way we have calculated L between 7 and 10 cm. These values, together with the appropriate average soil density, were used to calculate the theoretical efﬁciency with ISOCS and the experimental efﬁciency with equation (4). The LAB activity of natural radionuclides is the average of the activities calculated in the soil layers. The LAB activity of 137 Cs is calculated as follows: for soils S1, S2, S3, in which there is an exponential distribution, the LAB activity is the sum of the activities per unit area in each layer; for soil S4, in which there is a uniform distribution, the LAB activity is the average activity per unit area in the ﬁrst 30 cm. Table 4 reports, for each radionuclide, the activity calculated from laboratory measurements on soil samples (“LAB”) and the in-situ activity, calculated both with experimental (“in-situ EXP”) and ISOCS (“in-situ ISOCS”) calibration curves. The deviations between the in-situ activity calculated with the experimental calibration procedure and with ISOCS are less than 8%. It can be noted that activity values calculated with experimental calibration curves are always lower than those calculated with ISOCS calibration curves. This could be due to a systematic error introduced either by the experimental calibration procedure or by ISOCS calculation procedures, which we have not investigated yet. As for natural radionuclides, the agreement between laboratory and in-situ results can be considered satisfactory and congruent with the results obtained in analogous studies by other authors [7–9]. The maximum deviations appear in the case of 214 Bi and 214 Pb, especially in soil S4, where 214 Bi in-situ activity is about 30% higher than the laboratory one. This effect could be related to the different 222 Rn emanation from in-situ soil and from laboratory samples. For 214 Bi we have additionally to take into consideration the coincidence-summing effect that leads to a laboratory activity value lower than the real one. These effects have not been studied in depth yet, being beyond the purpose of this work.

Calibration of a HPGe detector for in-situ gamma spectrometry

205

Table 4 Activity of some radionuclides calculated from laboratory measurements on soil samples (“LAB”) and the in-situ activity, calculated both with experimental (“in-situ EXP”) and ISOCS (“in-situ ISOCS”) calibration curves Site

in-situ EXP

in-situ ISOCS

LAB

Ac-228 (Bq kg−1 )

S1 S2 S3 S4

31.30 ± 0.87 27.58 ± 1.72 28.96 ± 1.80 23.28 ± 0.95

33.05 ± 1.05 28.43 ± 1.82 30.46 ± 1.95 24.38 ± 1.15

35.67 ± 0.52 27.80 ± 1.00 33.96 ± 1.22 24.59 ± 0.89

TI-208 (Bq kg−1 )

S1 S2 S3 S4

12.44 ± 0.35 10.45 ± 0.67 11.59 ± 0.72 8.07 ± 0.70

13.12 ± 0.41 10.82 ± 0.71 12.28 ± 0.78 8.54 ± 0.75

13.05 ± 0.21 9.93 ± 0.39 12.50 ± 0.48 8.84 ± 0.30

Pb-214 (Bq kg−1 )

S1 S2 S3 S4

31.52 ± 0.91 27.10 ± 1.57 28.10 ± 1.69 25.68 ± 1.44

33.11 ± 1.29 27.70 ± 1.77 29.66 ± 1.97 27.10 ± 1.70

30.50 ± 1.22 26.43 ± 1.15 30.52 ± 1.61 21.74 ± 1.10

Bi-214 (Bq kg−1 )

S1 S2 S3 S4

31.04 ± 0.65 26.52 ± 0.70 26.87 ± 1.18 24.71 ± 0.77

33.15 ± 0.78 27.58 ± 0.78 28.71 ± 1.30 26.26 ± 0.86

27.97 ± 0.34 24.00 ± 0.68 27.25 ± 1.03 19.56 ± 0.68

K-40 (Bq kg−1 )

S1 S2 S3 S4

378.74 ± 9.19 437.58 ± 12.26 380.61 ± 11.33 321.87 ± 9.68

406.14 ± 12.11 462.11 ± 15.22 408.11 ± 14.06 345.16 ± 11.99

406.12 ± 5.34 430.96 ± 17.98 433.74 ± 32.63 356.15 ± 13.98

Cs-137 (Bq m−2 )

S1 S2 S3 S4

8526.98 ± 210.31 11 233.19 ± 337.36 13 005.90 ± 372.66 13 099.20 ± 465.60

9019.50 ± 249.34 11 523.35 ± 370.80 13 650.33 ± 400.24 13 828.80 ± 523.20

14 606.12 ± 267.97 13 050.49 ± 340.27 14 140.05 ± 304.10 15 596.00 ± 1218.90

As for 137 Cs, the results are acceptable, except in soil S1, where the deviation between laboratory and in-situ activity is around 40%. In this case the ﬁrst approximation of an exponential proﬁle used to determine the λ value is not correct. The correct efﬁciency value can instead be calculated assuming, in the ISOCS cylinder template, 6 soil layers, each with the activity calculated in laboratory on the core samples. The result thus obtained for 137 Cs deviates 10% from the laboratory activity. 4. Conclusions The agreement between the activity concentrations calculated in-situ with ISOCS and with experimental procedures can be considered satisfactory and these values are comparable to those calculated in the laboratory on samples from the same site. In conclusion the in-situ gamma spectrometry technique with ISOCS is convenient for our purposes aimed at the evaluation of the concentration of natural radionuclides in soils, mainly because it does not require time consuming soil core sampling and measurement.

206

F. D’Alberti, M. Forte

References [1] H.L. Beck, J. De Campo, C.V. Gogolak, In situ Ge(Li) and NaI(Tl) gamma-ray spectrometry, USAEC Report HASL-258, 1972. [2] K.M. Miller, Field Gamma-Ray Spectrometry, USDOE EML Procedures Manual HASL-300, 28th ed., 1997, vol. I, Section 3.3. [3] M. Abramowitz, I.A. Stegun, Handbook of Mathematical Functions, Dover, New York, 1965. [4] M.J. Berger, J.H. Hubbell, XCOM: photon cross sections on personal computers, NBSIR 87-3597, National Bureau of Standards, 1987. [5] F.L. Bronson, L. Wang, Validation of the MCNP Monte Carlo Code for germanium detector gamma efﬁciency calibrations, in: Waste Management ’96, Tucson, AZ, USA, 1996. [6] R. Venkataraman, F. Bronson, V. Atrashkevich, B.M. Young, M. Field, Validation of In Situ Object Counting System (ISOCS) mathematical efﬁciency calibration software, Nucl. Instrum. Methods (A) 422 (1999) 450–454. [7] L. Daling, Z. Chunxiang, G. Zujie, L. Xian, H. Guorong, Gamma-spectrometric measurements of natural radionuclide contents in soil and gamma dose rates in Yangjiang, PR China, Nucl. Instrum. Methods (A) 299 (1990) 687–689. [8] S. Bortoluzzi, M. Montalto, M. Nocente, R. Giacomelli, P. Spezzano, Misura “in campo” della concentrazione di radioemettitori gamma nel terreno, Rapporto tecnico ENEA, Centro Ricerche Energia Saluggia, Vercelli, 1990. [9] R.R. Benke, K.J. Kearfott, Comparison of in situ and laboratory gamma spectroscopy of natural radionuclides in desert soil, Health Phys. 73 (2) (1997) 350–361.

207

A new technique for accurate measurements of Ra-226 with γ-spectroscopy in voluminous samples M. Manolopoulou, S. Stoulos, D. Mironaki, C. Papastefanou Nuclear & Elementary Particle Physics Division, Physics Department, Aristotle University of Thessaloniki, Thessaloniki 54124, Greece

Accuracy in the measurements of Ra-226 concentrations by γ-spectroscopy depends on an appropriate estimation of Rn-222 decay product concentrations (Pb-214 and Bi-214) in the sample. As an inert gas, radon can leak from the sample vessel and/or can be accumulated in the upper, void part of it. In the latter case radon decay products can be attached on the inner surface of the sample container, and therefore, the produced γ-rays have a different geometry and a self-absorption factor from the one assumed (similar to the calibration sample). The inﬂuence of the above procedures on the accuracy of the measurements depends mainly on the emanation factor of radon in the sample, the geometry used in the measurement and the structure of the sample container. In this paper, results of the tests performed with a new technique developed in order to eliminate radon diffusion through the sample, are presented. Two materials with different effective Ra-226 concentrations were used for the tests: soil with 5.2 Bq kg−1 effective Ra-226 (total Ra-226 concentration 20.3 Bq kg−1 and emanation factor 25.3%) and phosphate fertilizer with 21.5 Bq kg−1 effective Ra-226 (total Ra-226 concentration 305.1 Bq kg−1 and emanation factor 7.1%). From γ-spectroscopy measurements it was found out that the increase in determined Ra-226 concentration, when charcoal was added to the samples, was about 35% of the effective Ra-226 of the sample. This could lead to an underestimation of about 10% in Ra-226 concentrations in voluminous samples with high emanation factor (25–30%).

1. Introduction Determination of Ra-226 concentration by γ-spectroscopy is based mainly on measurements of Rn-222 decay product concentrations. In solid samples a part of the Rn-222 produced by Ra-226 decay is captured in the grains, while another part of the Rn-222 emanates into the interstitial space of the sample, either due to diffusion in the grain or to recoil processes. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07023-8

© 2005 Elsevier Ltd. All rights reserved.

208

M. Manolopoulou et al.

The fraction of Rn-222 that emanates from the grains is usually called effective radium and depends on the physicochemical properties and the size of the sample grains [1–3]. Radon as an inert gas can leak from the sample vessel and/or can be accumulated in the upper, void part of it. Radon progeny gamma-rays originating from the upper part of the vessel present a smaller geometry and a larger self-absorption factor than those that are generated in the sample. For the estimation of Ra-226 concentration one assumes that the radioactivity is homogeneously distributed in the sample, as it is in the calibration sample. As a result, the accuracy in the determination of Ra-226 concentration in the sample depends on the amount of the effective radium in the sample as well as the geometry characteristics of the sample vessel. In this paper, the results of the tests performed with a new technique developed in order to eliminate radon diffusion through the sample, are presented.

2. Experimental methods and instrumentation The new technique consists of mixing regenerated charcoal in powder form (Ø less than 400 μm) with the sample before sealing it hermetically and storing it in a freezer during Rn-222 in-growth. Charcoal is a well-known effective adsorber of radon [4–6]. Gamma spectroscopy measurements were performed using a high efﬁciency HPGe (42%) with resolution 2.0 keV at 1.33 MeV of Co-60, linked to an appropriate data acquisition system. The calibration of the spectrometer was realized with IAEA gamma spectrometry reference materials (RG-set). Uranium concentration in the calibration sample (volume 1 L) is certiﬁed to have 400 ppm with conﬁdence interval ±2 ppm at a signiﬁcance level of 0.05. U-238, Ra-226 and Pb-210 were conﬁrmed to be in radioactive equilibrium. The efﬁciency calibration was done with the radionuclide speciﬁc efﬁciency method in order to avoid uncertainty in the resulting gamma-ray intensities as well as the inﬂuence of the summation effect [7]. Radon decay product concentration in the samples was determined as the weighted values from 295 keV and 352 keV for Pb-214 and 609 keV, 1120 keV and 1764 keV for Bi-214. Two sample materials were used for the tests of the technique with different radium concentration and emanation factor: phosphate fertilizer and soil. Samples were pulverized to grains of 400 μm in size and homogenized. The characteristics of the materials are presented in Table 1. Solid density is the density of the grains while bulk density is the density of the sample. The errors presented in Table 1 were calculated by propagation of systematic and random errors of the measurements. Marinelli beaker geometry was used during these tests, being a usual geometry for voluminous samples. Regenerated charcoal in powder form (Ø < 400 μm) was mixed with the Table 1 Speciﬁc characteristics of phosphate fertilizer and soil samples

Phosphate fertilizer Soil

Bulk density (g cm−3 )

Solid density (g cm−3 )

Radon exhalation rate from grain size Ø < 400 μm (Bq m−2 h−1 )

Emanation factor (%)

1.24 ± 0.02 1.27 ± 0.04

2.25 ± 0.09 2.23 ± 0.03

0.162 ± 0.001 0.035 ± 0.004

7.1 ± 0.2 25 ± 3

A new technique for accurate measurements of Ra-226 with γ-spectroscopy in voluminous samples

209

Fig. 1. Schematic diagram of a Marinelli beaker. The dimensions of the sample and the position of the LR-115 detector are presented.

sample from 0.1% up to 2% per weight. The beaker was ﬁlled with 1 L sample leaving an empty space of 0.4 L above the sample (Fig. 1). The sample was sealed hermetically and stored in a freezer during the Rn-222 build up period. In order to monitor Rn-222 concentration in the upper space of the Marinelli beaker a solidstate nuclear track-etch detector (LR-115) was placed on the cover of the Marinelli beaker (Fig. 1). A small (Ø 5 cm) and thin plastic barrier above the sample ensured that there was no direct registration of α-particles from the sample. Radon leakage rates from the Marinelli beaker as well as radon exhalation rates from the samples (Table 1) were determined by enclosing the samples in a small radon chamber (about 0.1 m3 ), where Rn-222 concentration was monitored during the build up period using standard Lucas α-scintillation cells [8,9]. 3. Results and discussion Radon leakage rates from the Marinelli beaker as a function of the weight percentage of charcoal addition in phosphate fertilizer and soil samples are presented in Fig. 2. The critical limit of the detection at 90% conﬁdence is 3.1·10−6 Bq kg−1 s−1 . From the data presented in Fig. 2, it turns out that only a very small fraction of emanating Rn-222 escaped out of the Marinelli beaker when the samples were mixed with charcoal. The leakage rate of the samples without any charcoal is (20.9 ± 2.4) · 10−6 Bq kg−1 s−1 and (43.7 ± 8.5) · 10−7 Bq kg−1 s−1 for phosphate fertilizer and soil sample, respectively. The track registration rate on LR-115, which was placed at the top of the Marinelli beakers (Fig. 1), as a function of the weight percentage of charcoal addition in phosphate fertilizer and soil samples is presented in Fig. 3. Radon and its decay product concentration in the upper space of the Marinelli beaker decreased as the weight percentage of charcoal in the samples increased indicating that charcoal did not allow radon diffusion out of the sample volume.

210

M. Manolopoulou et al.

Fig. 2. Radon leakage rates from the Marinelli beaker as a function of the weight percentage of charcoal addition in phosphate fertilizer and soil samples.

Fig. 3. Track registration rates on LR-115 placed at the top of the Marinelli beaker as a function of the weight percentage of charcoal addition in phosphate fertilizer and soil samples.

The determined Ra-226 concentrations in the samples by gamma-ray spectroscopy rise with the increment of charcoal addition (Table 2). The increase of Ra-226 concentration should be attributed to the radon fraction the diffusion of which was prevented by charcoal and conse-

A new technique for accurate measurements of Ra-226 with γ-spectroscopy in voluminous samples

211

Table 2 Ra-226 concentration (CRa ) as a function of the weight percentage of charcoal addition in phosphate fertilizer and soil samples Phosphate fertilizer

Soil ± (Bq kg−1 )

Charcoal percentage (p.w.)

CRa

0.0% 0.1% 0.2% 0.4% 0.6% 0.8% 1.0% 2.0%

299.2 299.4 305.2 303.6 306.8 304.8 305.3 304.9

σ∗ 0.89 0.93 0.98 0.95 0.94 0.95 1.11 0.98

± (Bq kg−1 )

Charcoal percentage (p.w.)

CRa

0.0% 0.2% 0.5% 0.8% 1.0% 2.0%

18.00 19.10 20.57 20.38 20.20 20.45

σ∗ 0.11 0.13 0.12 0.13 0.13 0.13

∗ σ is the total uncertainty of the measurement.

quently was kept in the sample volume, since radon decay product gamma-rays originating from the upper part of the Marinelli beaker have a smaller geometry factor and undergo larger self-absorption than those that are generated in the sample. So, the contribution of this part of the radon to the peak areas of the gamma-ray spectrum was smaller when it occupied the upper space of the Marinelli beaker when compared with the contribution it had when it was distributed in the sample. The effective radium of the samples as calculated from the exhalation rates (Table 1) was 21.5 ± 0.6 Bq kg−1 and 5.2 ± 0.6 Bq kg−1 for phosphate fertilizer and soil sample, respectively. Considered as a better estimate of Ra-226 concentration, the mean value of the samples measured with 1% and 2% of charcoal addition were calculated to be 305.1 ± 0.7 Bq kg−1 and 20.3 ± 0.1 Bq kg−1 for phosphate fertilizer and soil, respectively. The difference between the above concentrations and those measured without any charcoal addition was 5.9 ± 1.2 Bq kg−1 for phosphate fertilizer and 2.3 ± 0.2 Bq kg−1 for the soil sample. These differences represent (27 ± 5)% and (44 ± 6)% of effective radium of the corresponding samples. Assuming that in the deﬁned geometry the effective Ra-226 concentration with no charcoal addition is underestimated by a factor of about 35% (mean value of 27% and 44%), this could lead to an underestimation of Ra-226 concentration of almost 0% for samples with low emanation factor (less than 1%) to about 10% for samples with high emanation factor (25– 30%), thus resulting in a systematic error which depends on the physicochemical properties of the sample.

References [1] [2] [3] [4]

D.A.W. Bossus, Radiat. Prot. Dosim. 7 (1984) 73. S. Simopoulos, D. Leonidou, Atomkernenergie-Kerntechnick 49 (1986) 105. L. Morawska, Health Phys. 57 (1989) 23. K.P. Strong, D.M. Levins, US DOE Report CONF 7870819, 1979.

212 [5] [6] [7] [8] [9]

M. Manolopoulou et al. A.C. George, Health Phys. 46 (1984) 867. L. Zikovsky, Health Phys. 75 (1998) 313. G. Gilmore, J.D. Hemingway (Eds.), Practical Gamma-Ray Spectrometry, Wiley, 1995. N. Jonassen, J.P. McLaughlin, in: Proc. Natural Radiation Environment III, vol. 2, CONF-780422, 1980, p. 1211. W.H. van der Spoel, et al., Health Phys. 72 (1997) 766.

4. Radon and thoron

T . his Page Intentionally Left Blank

215

Radon concentration in the tunnels of a hydroelectric power station under construction in Italy, a case study S. Verdelocco, D. Walker, P. Turkowsky Joint Research Centre, Via E. Fermi 1, 21020 Ispra (Va), Italy

The purpose of this study was to measure the radon concentration in air, in the tunnels of a hydroelectric power station in Piedmont, Italy. Both active and passive methods were used for radon measurements. In addition gamma spectrometry was performed on various rock samples. The results showed how radon concentration in tunnels is inﬂuenced by the ventilation rate both for active and for passive measurements. It was clear that in this type of environment the only solution to the problem of reducing the radon concentration in air was to use a good ventilation system that should always be in service during periods of work. The source of radon in the tunnels was not attributed to the rock type (metamorphic rocks), as was veriﬁed with spectrometry analysis.

1. Introduction The purpose of this study was to measure the radon concentration in air, in the tunnels of a hydroelectric power station in Piedmont, Italy. This power station is the only hydroelectric installation under construction today in Italy and is predominantly underground. The necessity for monitoring arose from the presence in this area of various uraniferous outcrops identiﬁed by drilling and mineral research performed in the past. For this reason it could not be excluded that some uraniferous deposits might be encountered while boring the tunnels. The geology of the area is predominantly characterised by metamorphic rocks; these kinds of rocks usually rich in uranium minerals could also be a radon source. The study commenced in 1998 and was completed in 2000. Since the site where the monitoring was performed presented anomalous environmental characteristics (water, dust, humidity, vibrations, etc.), it was necessary to adapt the methodology used, choosing the equipment that could give the best results considering the type of environment. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07024-X

© 2005 Elsevier Ltd. All rights reserved.

216

S. Verdelocco et al.

2. Geology of the area The geology of the area is characterised by outcrops belonging to the Ambin Massif (northwest zone) and to the Piemontese Area (central-east zone). The sequences, both of the Ambin Massif and of the Piemontese Area, are comprised of metamorphic rocks such as: quartzite, micaceous quartzite, dolomitic and calcitic marble, gneiss, mica schist, carbonatic schist, and calcareous schist. The superﬁcial coverings in the area are drift deposits, ﬂuvial deposits, recent alluvial deposits, colluvium, and detrital-colluvial covering. The structural setting is characterised by folds (Alpine compressional phase) normal faults and discontinuities (regional stretching stress).

3. Materials and methods Both active and passive methods were used for radon measurements. The Genitron Alphaguard monitor was used for active measurements. Regular checks on the air radon concentration, at the tunnel work-faces, were performed during working periods. The passive method, Karlsruhe diffusion chambers sealed in polyethylene, was used to measure the longterm radon concentration including non-working periods. In addition gamma spectrometry was performed on various rock samples. 3.1. Working methods and problems related to the conditions in the tunnels The Alphaguard instrument was chosen for real time radon measurements as its characteristics best met the frequently severe conditions in the tunnels. Notwithstanding the fact that it uses an ionisation chamber, hence normally considered delicate, it remains unaffected by high humidity. It is also capable of measuring the very high concentrations that may be found in these areas. In all tunnels the measurements were performed as close as possible to the work face (being the area where the workers pass most of there time) depending on the state of work. For the tunnels excavated using the TBM (tunnel boring machine) the monitor was placed on the machine head as close as possible to the cutting tool. On a number of occasions the measurements were performed with the machine boring with subsequent very high vibrations. On other occasions they were performed in the presence of gas and dust due to drilling and coating operations. The measurements were always performed for at least half an hour for each measurement site. The monitor was always covered by its protective bag and frequently with an umbrella. The instrument case was always used as a stand. The use of dosimeters for long term radon and ambient measurements presented problems other than those linked to humidity and dust but rather problems related to the continuous advancement of the work face. Initially the dosimeters were attached to the walls or placed in the wall supports. The area was highlighted with red paint or brought to the attention of the workers. Frequently after one or two months the dosimeters were lost, either cemented into the wall covering or buried under the ﬂoor covering, or simply damaged. Because of this highly visible steel plates were manufactured with supports for the dosimeters. These plates were ﬁxed to the tunnel walls. In this way the visibility of the plates was guaranteed and they were made easy for the workers to move and reposition safely. For the tunnels excavated using a TBM the situation was made

Radon concentration in the tunnels of a hydroelectric power station under construction in Italy

217

easier by positioning the dosimeters in the TBM control room. In this way the measurements could continue while the work face advanced. A further problem, related to the presence of water, mud and dust, was resolved by placing the dosimeters in polythene bags impermeable to water but permeable to radon. These dosimeters were changed at each visit and therefore remained exposed for a period of at least one month.

4. Results Measurements were made with the active monitor, with the ventilation in and out of service. In general with the ventilation in service the radon concentration in air was below 100 Bq m−3 , whilst with the ventilation out of service radon concentrations of over 10 000 Bq m−3 were measured. Figure 1 shows the instantaneous measurement results of all the tunnels monitored. Measurements performed with passive detectors showed the same results: radon concentration below 1000 Bq m−3 during normal periods and over 10 000 Bq m−3 during periods in which the construction site was closed. Figure 2 shows some of the long-term measurement results. The radon concentration measured with dosimeters reﬂects the mean concentration for the complete exposure period. These measurements should be considered excessive from a precautionary point of view because they include periods when work was suspended and the site was closed (i.e., during Christmas holidays). In this case the absence of recirculation of the air present in the tunnels aggravates the accumulation of radon. It should be noticed that the high concentration measured in tunnel 5 were due to the fact that the drilling (and the ventilation) was stopped for a period because of technical problems. Figure 3 shows the difference on radon concentrations for two tunnels measured in the same day with working and non-working ventilation.

Fig. 1. Instantaneous measurement results.

218

S. Verdelocco et al.

Fig. 2. Long term radon results.

Fig. 3. Radon concentrations in two different tunnels with working and non-working ventilation.

It was veriﬁed that radionuclides of the uranium and thorium series contained in the rock samples analysed were in equilibrium. Measurements on rock samples showed that the concentrations of uranium-238 and thorium-232 were lower than 1 Bq g−1 . According to the legislative decree which was in force at the time the measurements were performed (d.lgs. 230/1995 point 1.3 attachment I) these are the Italian limits above which the materials must

Radon concentration in the tunnels of a hydroelectric power station under construction in Italy

219

Fig. 4. Results of gamma spectrometry measurements performed on rock samples taken in 1998.

Fig. 5. Results of gamma spectrometry measurements performed on rock samples taken in 1999.

be considered radioactive [1]. Figures 4 and 5 show results of gamma spectrometry measurements performed on rock samples taken respectively in 1998 and 1999. 5. Conclusion The results showed how radon concentration in tunnels is inﬂuenced by the ventilation rate both for active and for passive measurements. It was clear that in this type of environment

220

S. Verdelocco et al.

the only solution to the problem of reducing the radon concentration in air was to use a good ventilation system that should always be in service during periods of work. The radon source in the tunnels was not attributed to the rock type (metamorphic rocks), as was veriﬁed with the spectrometry analysis. It is probable that the large volumes of water in the tunnels acted as a conduit for the radon, since the highest radon concentrations were measured in tunnels with a large volume of water ﬂowing through faults and fractures. With regard to the technical aspects of the monitoring campaign, it is difﬁcult to deﬁne a standard measurement routine, because it will depend on the work programme for the site and on the decisions of the local authorities. With regard to the progression of the measurements, the number performed depends in ﬁrst instance on the length of the tunnels and the excavation rate. The tunnels in which the most measurements were performed were those with extended excavation times due to problems encountered during excavation (meeting a fault plane, water sources, collapse of rock wedges, etc.), or excessive length. Meanwhile those tunnels with the smallest number of measurements were the shortest or those that encountered the least problems. The occasional gaps in the results obtained may be alleviated by the following conditions: information provided to the workers on measurements in progress, positioning of dosimeters in an area that guarantees their safety (often difﬁcult to identify). Dosimeters may be damaged by the progression of the excavation and the difﬁcult conditions such as dust, mud, water, cement and excavation using explosives. With regard to the measurement methods it was found to be essential for the accuracy of the measurements to use both active and passive methods. With the active method it is possible to have a real time measurement of the radon concentration and evaluate the efﬁciency of protective measures during actual periods of work (ventilation strength). Alternatively the passive method being an integrated measurement provides the mean radon concentration during a given time period, including periods when work was suspended and radon was allowed to accumulate.

References [1] Decreto legislativo 17 marzo 1995, n. 230, supplemento ordinario alla GURI 136 del 13 giugno 1995.

Further reading [1] T. Domanski, W. Chruscielewski, M. Hofman, Monitoring the exposure to radon decay products in mine air using passive track detectors, Health Phys. 40 (1980) 211. [2] Recommendations for the implementation of Title VII of the European Basic Safety Standards Directive (BSS) concerning signiﬁcant increase in exposure to natural radiation sources, Radiation Protection 88, European Commission, Luxembourg, 1997. [3] Genitron Instruments, Report of the radon monitor AlphaGUARD in mines. [4] S.E.A. Impianto Idroelettrico Pont Ventoux-Susa, Nota Geologica Informativa, Commento Alle Tavole Geologiche Doc. SV9714RGD00. [5] http://members.tripod.it/PontVentoux/. [6] S. Verdelocco, D. Walker, P. Turkowsky, C. Osimani, Misure di radon-222 e radioattività ambientale nell’impianto idroelettrico di Pont Ventoux-Susa, Piemonte, Rapporto della Commissione Europea EUR 19656 IT, 2000.

221

Determination of the radon potential of a building by a controlled depressurisation technique (RACODE) W. Ringer a , H. Kaineder b , F.-J. Maringer c , P. Kindl d a Department of Radiation Protection, Federal Ofﬁce of Agrobiology, Derfﬂingerstraße 2, 4020 Linz, Austria b Unterabt. Lärm- und Strahlenschutz, Amt der Oö. Landesregierung, Stockhofstraße 40, 4020 Linz, Austria c Low-Level-Radiometry Laboratory Arsenal, Austrian Research Centers Seibersdorf,

Faradaygasse 3, 1030 Wien, Austria d Institute of Technical Physics, Technical University of Graz, Petersgasse 16, 8010 Graz, Austria

Action levels and limits for radon in homes apply to the annual mean radon concentration. Because the indoor radon concentration varies strongly with time short term measurements are often not accurate; on the other hand, long-term measurements do not allow rapid assessment of the exposure to radon. This paper presents methodology and results of a new method for the rapid determination of the building radon potential (RACODE (radon potential determination by controlled building depressurisation)). A fan is used to produce a small pressure differential (10–50 Pa) between building and outdoors and the measurement of the ﬂow rate and the radon concentration of the fan exhaust air at steady state yields the convective radon entry rate. Furthermore building characteristics like air exchange rate, equivalent leakage areas, and leakage distribution are determined. Sometimes the use of a second fan is necessary to produce the pressure differential in the lower part of the building only and to determine the leakage distribution of the building. With appropriate modelling the mean radon concentration is deduced from these data. RACODE was applied to eight buildings where the mean radon concentration was known from long-term passive measurements (2 × 3 months). The radon concentrations obtained agree well in most cases with those from the long-term measurements. The uncertainty depends strongly on the type of building, i.e. whether it is possible to simulate stack effect conditions well enough with the fan(s) and whether the leakage distribution can be determined accurately. Besides the determination of the mean radon concentration RACODE should be useful for the rapid assessment of the effectiveness of mitigation measures if the same kind of measurements at deﬁned pressure conditions are performed before and after mitigation. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07025-1

© 2005 Elsevier Ltd. All rights reserved.

222

W. Ringer et al.

1. Introduction Long-term exposure to radon may constitute the highest health risk to individuals from indoor air pollution in residences. Prolonged inhalation of air with elevated radon levels increases signiﬁcantly the risk of lung cancer. Therefore the need to identify homes that expose their occupants to high long-term radon concentrations has gained increased attention in recent years. The standard means of such identiﬁcations are based on either a short-term (2–7 days) sampling of indoor air or a long-term (30 days or longer) measurement of indoor radon concentrations. The indoor radon concentration in a building shows strong daily and seasonal variations due to the complex relationship of radon entry rate and ventilation rate with meteorological and house occupant related factors. Therefore short-term samples have not been shown to be good indicators of long-term average radon concentrations. Long-term tests are more accurate in determining annual average radon exposures but do not allow rapid assessment. Even the long-term test is subject to yearly variations. The indoor concentration of radon and its decay products, or of any other airborne pollutant, depends on three factors: the entry or production rate from various sources, the ventilation rate, and the rates of chemical or physical transformation or removal. Because of its relatively long half-life and lack of chemical activity, radon itself acts much like a stable pollutant whose indoor concentration is determined by only two factors: the entry rate and the ventilation rate [1]. Sources of indoor radon are convective and diffusive entry of radon bearing soil gas, building materials, and domestic water supplies. However, a number of studies conﬁrm that pressure-driven (convective) ﬂow is the dominant source of radon in most houses with elevated concentrations [2–5]. Convective ﬂow of radon-bearing soil gas into a house results from a pressure deﬁcit at the base of the house relative to the surrounding soil. Causal mechanisms include: temperature differences between indoors and out, wind on the house superstructure, temporal variations in barometric pressure, and unbalanced building ventilation. Moreover, inﬁltration of outdoor air into the building – thus the ventilation rate – is governed by pressure differences across the building shell, too. Over the past two decades, substantial efforts have been made to develop both experimental and modelling techniques to predict long term mean indoor radon concentrations [6–15]. Many of these short-term diagnostic tests rely on the measurement of soil properties (soil permeability, soil gas radon concentration) and use various modelling techniques for the calculation of indoor radon levels. These methods are limited in their reliability by the fact that soil properties can be highly inhomogeneous and anisotropic over space and time. Extensive sampling is required to determine the variability at a speciﬁc location (house). Furthermore, data on building characteristics have to be collected. Therefore, a technique which measures directly the radon entry rate of a speciﬁc building is required. If this technique simultaneously allows one to determine the ventilation characteristics of the building, then both radon entry and ventilation rates can be estimated and simpliﬁed modelling should give satisfactory estimates of mean indoor radon levels. Rapid evaluation of the effectiveness of remedial actions (by applying the same technique before and after mitigation) would be possible.

Determination of the radon potential of a building by a controlled depressurisation technique (RACODE) 223

2. Method In this study, a Blower Door (BD) was used to mechanically depressurise the building and to determine the convective radon entry rate (Sr) and the ventilation rate (Fig. 1). Blower Doors are conventionally employed for air tightness tests. A fan is mounted in a doorframe. Pressure differentials between indoor and outdoor of up to ±60 Pa can be established. Sr is determined by measuring the ﬂow rate through the BD fan and the radon concentration of the BD exhaust air. The ventilation rate is obtained by the standard air tightness test. With further series leakage tests (Guard Zone, Opening A Door, Adding A Hole, . . .) the leakage distribution of the building shell and within the building is deduced. With the data obtained from the measurements described above the mean indoor radon concentration is derived using two models. Model A is based on considerations by Sherman [11] to calculate the indoor radon concentration Cr taking into account building characteristics:

2 1+n Hb nr Sr−4 Pa × 2 × Z × βs + × × Cr = ELA4 × v0 fXs 1+n×R H nr −n

ρ × g × H (1) × 2 × p0 where Sr−4 Pa is advective radon entry rate at −4 Pa [Bq s−1 ], Sr−4 Pa = Cf−4 Pa × Qf,−4 Pa ; Cf−4 Pa is radon concentration at BD exhaust at −4 Pa [Bq m−3 ];

Fig. 1. Blower Door and radon monitor.

224

W. Ringer et al.

Qf,−4 Pa is ﬂow rate through BD at −4 Pa; ELA4 is the effective leakage area of whole building at 4 Pa [m2 ]; v0 is the reference velocity [2.58 m s−1 ]; fXs is the leakage asymmetry factor, n 2 fXs ≡ 1 − Xs2 × (1 − Xs )1/n + (1 + Xs )1/n

(2)

with Xs is leakage asymmetry, ELA4H − ELA40 ; ELA4 ELA4H is ELA4 at top of building; ELA40 is ELA4 at bottom of building; n is the ﬂow exponent for ventilation; R is the box parameter, Xs ≡

(3)

ELA4H + ELA40 ; (4) ELA4 Z is the zone factor, Z = 1 for single-zone buildings, Z = 0.6 for two-zone buildings (i.e., when pzb > 0.2pbo ); βs is the neutral pressure level factor, R≡

βs ≡

1+

1 1−Xs 1/n ;

(5)

1+Xs

H is height of building above grade level [m]; Hb is height of building below grade level [m]; nr is ﬂow exponent for radon entry; ρ is the difference in air density due to temperature difference T [kg m−3 ]; g is acceleration of gravity [m s−2 ]; p0 is the reference pressure [4 Pa]. Model B uses a powerful computer program (CONTAM96) [16] to simulate airﬂows and radon concentrations. Leakage characteristics of the building can be included in the simulation as well as meteorological parameters or airing.

3. Results and discussion 3.1. Preliminary tests A BD was applied for radon measurements within the Austrian radon mitigation research project SARAH for the ﬁrst time. The BD was mounted in a door which leads to the building or the group of rooms under investigation. All exterior openings (doors, windows, exhausts, chimneys) were closed (sealed) while all interior doors are opened. Then a pressure difference of 50 Pa across the building shell (depressurisation) was applied. The radon concentration of the BD exhaust fan

Determination of the radon potential of a building by a controlled depressurisation technique (RACODE) 225

Fig. 2. Radon concentration of the BD exhaust air at location Haid at various weather conditions.

air was measured with a continuous radon monitor. Measurements were made until the radon concentration of the exhaust fan air reaches equilibrium (about three hours). In order to verify that building depressurisation overrides meteorological effects on the radon entry rate, the extended BD method was applied ﬁve times at a house in Haid (no basement, high soil permeability, Fig. 2) and twice at an old farmhouse in Königswiesen (no basement, high soil permeability, high soil gas radon concentration, Fig. 5) at quite different weather conditions. The results suggest that there is virtually no weather impact on the BD results. Therefore, the extended BD method is a building-speciﬁc but weather-independent method for characterising the convective radon entry rate. 3.2. Systematic measurements with the aim to quantify building radon potentials Two sets of BD experiments were conducted at 8 houses with known mean annual radon concentrations from long-term passive radon measurements (TrackEtch- or EPERM-detectors). The detectors were exposed in the two most occupied rooms of the home for three months both during winter and summer time. Calculated annual averages range from 35 to 525 Bq m−3 with an uncertainty of about ±30%. All houses were located in Upper Austria. Most houses were situated on gravel terrace, some on loamy soil and one house on granite. At each location soil moisture in the upper soil layer (0–30 cm) and both soil gas radon concentration and resistance to soil gas ﬂow at a depth of 1 m were determined (Table 1). The extended BD method was applied to the homes twice: the ﬁrst set of experiments was conducted in July, the second one the following year in March. The results of the ﬁrst set of

226

W. Ringer et al. Table 1 Range of values of the most important parameters Parameter

Range

Soil moisture Soil gas radon concentration Resistance to soil gas ﬂow Effective Leakage Area at 4 Pa Flow exponent n Air exchange rate at normal conditions Radon entry rate per Pa

[vol%] [Bq m−3 ] [Pa ml−1 min−1 ] [m2 ] [L h−1 ] [Bq s−1 Pa−1 ]

27–70 9500–68 000 1.5–22.4 0.015–0.196 0.55–0.83 0.18–1.61 0.15–5.74

experiments showed that the radon concentration at the BD exhaust could not be compared directly to the average radon concentrations obtained from the long-term passive radon measurements. The speciﬁc ventilation characteristics of a building have to be taken into account. Furthermore, if the air tightness of the building shell is low the radon concentration at the BD exhaust is low, too, and the contribution of the ambient air radon to the radon entry rate becomes signiﬁcant. Consequently, to make sure that the BD exhaust radon concentration is signiﬁcantly higher than the outdoor level at the second set of measurements attempts were made to depressurise only the lower part of the building and to avoid intra-building air ﬂows using the Guard Zone method (Fig. 3).

Fig. 3. Applying the Guard Zone method.

Determination of the radon potential of a building by a controlled depressurisation technique (RACODE) 227

Further BD tests were conducted to determine the leakage of exterior walls, leakage between ﬂoors or leakage to the loft by using diagnostic BD methods. 3.3. Derivation of the mean indoor radon concentration by modelling The measured radon entry rates, leakage characteristics (in terms of ELA4 (Effective Leakage Area at 4 Pa) and ﬂow exponent n), and building dimensions are used as input data for two models. Model A is given in Section 2; it assumes convective transport only, steady state conditions, no stack or wind effects in the soil due to the presence of the house, and radon entry at basement ﬂoor level. Model B uses an indoor air quality analysis computer program (CONTAM96) which is designed to help the user determine airﬂows (inﬁltration, exﬁltration, room-to-room, etc.), contaminant concentrations, and personal exposure. An important feature of CONTAM96 is the possibility to include weather parameters in the simulation. Depending on the type of simulation selected, steady state or transient weather data may be needed. Furthermore, time schedules for leakage characteristics can be generated. This is useful for modelling e.g. the opening/closing of doors or windows or the operation of a kitchen exhaust fan. Once the building is idealised by the tools provided by CONTAM96 it is useful to simulate the BD measurements performed at the real building to check for the correct setup of the model. Figure 4 shows the measured and modelled radon concentrations at the BD exhaust at building RiWo. The mean radon concentrations Cr for the winter period were deduced with Models A and B and compared to the long term passive radon measurements over the same period of time (Table 2).

Fig. 4. Comparison of measured and calculated BD radon concentrations.

228

W. Ringer et al.

Table 2 Measured and derived indoor radon concentrations at 8 buildings (in Bq m−3 )

CrModel A CrModel B Crmeasured

Kain

Klem

Kony

Rich

RiEl

RiSt

RiWo

SlaM

106 90 211

75 – 85

811 603∗ /166† 125

172 153 121

405 – 127

18 13 174

651 689 814

126 – 140

∗ Basement window closed. † Basement window open.

For single zone and two-zone buildings (Rich, RiSt, Kony, RiWo) the results agree quite well (although the measured radon entry rate at RiSt is totally wrong). Due to the complex building design the determination of the building leakage characteristics was not sufﬁcient at three houses for using CONTAM96 (Klem, RiEl, SlaM). At Kain the radon entry rate could not be measured appropriately since only half of the radon entry area was depressurised and due to the different set-up of the ﬂoor (foundation) in the two halves the radon entry rate might not be uniform over the whole radon entry area. At Kony a basement window is usually open but was closed when conducting the BD tests. Adding an open window to the CONTAM96 simulation yields a mean radon concentration close to the measured one. The uncertainty of the derived results is estimated as ±30%.

Fig. 5. Radon concentration of the BD exhaust air before and after mitigation in Königswiesen.

Determination of the radon potential of a building by a controlled depressurisation technique (RACODE) 229 Table 3 Comparison of radon reduction efﬁciencies Extended BD method Passive long-term measurements Mean Reduction Rn concentration Rn concentration Reduction Mean before mitigation after mitigation factor Rn concentration Rn concentration factor before mitigation after mitigation Bq m−3 Gutau 457 Königswiesen 1280 Traun 496

Bq m−3 48 250 290

9.5 5.1 1.7

Bq m−3

Bq m−3

540 870 490

101 294 247

5.3 3.0 2.0

3.4. Veriﬁcation of mitigation works Blower Doors may present an efﬁcient tool to rapidly assess the effect of mitigation works. If the measurements are carried out before and after mitigation in the same manner then neither leakage measurements nor modelling is required but the radon concentration at the BD exhaust can be compared directly. BD measurements before and after mitigation where conducted at three houses where the remedial measures were as follows: Gutau: active sub-ﬂoor depressurisation; Königswiesen: active sub-house depressurisation; Traun: passive sub-ﬂoor (crawl space) ventilation. Figure 5 shows as an example the results at Königswiesen. Follow-up long-term measurements after mitigation (three years, using track-etch detectors) were performed. In Table 3 the radon reduction factors calculated from the extended BD method and from long term measurements are compared. Good agreement is found at all three buildings.

4. Conclusions In this study a Blower Door (BD) was used to mechanically depressurise houses in a controlled and reproducible manner to determine the radon entry rate. Furthermore, diagnostic tests like Opening A Door or Guard Zone were carried out to characterise the leakage pattern (i.e., ELA and n) of the house. With these data and (simpliﬁed) modelling mean indoor radon concentrations are deduced and compared with the results of long-term passive radon measurements. One main advantage of using a BD to depressurise the building is that it simulates the real convective radon entry situation due to stack effect and thus averages over all soil gas radon concentrations, soil permeabilities, and radon entry paths. To determine the radon potential of a house, the results of this study suggest three distinct steps:

230

W. Ringer et al.

(i) Determination of the radon entry rate Sr • Conduct a depressurisation test taking into account the following requirements: – The depressurised area should enclose most of the radon entry area, ideally all of it. – The steady state BD exhaust air radon concentration has to be signiﬁcantly above the outdoor level (i.e., greater than 100 Bq m−3 ); this may require to reduce inﬁltration of ambient air, e.g., by depressurisation of the lower part of the building only (e.g., basement), using the Guard Zone method, extensive sealing, closing the interior doors of the upper level(s). – The building or zone, respectively, should be depressurised to about 10–15 Pa only to avoid soil depletion effects; record Cf and the airﬂow rate of the fan Qf after Cf has reached a steady state. • Calculate Sr = Cf × Qf (for above experimental set-up the contribution of ambient air radon to Sr can be neglected). (ii) Determination of the leakage pattern A standard BD test and further diagnostic BD tests should be conducted to characterise the leakage pattern of the building, i.e., ELA4inﬁltration , ninﬁltration , ELA4exﬁltration , and nexﬁltration for each zone of the building; for these measurements alterations to the building shell (e.g., sealing) should be kept to a minimum. (iii) Modelling In this study a modiﬁed model by M. Sherman (Model A) and an indoor air quality analysis computer program (CONTAM96, Model B) were tested successfully. Furthermore, CONTAM96 is a very versatile model and very useful to simulate the impact of various modiﬁcations to the building shell or different temperature settings inside and outside the building. Weather ﬁles can be generated to simulate real meteorological conditions and time schedules for leakage paths can be created to simulate airing, for example. Openings which had to be sealed for the BD tests can be introduced in the model. Therefore not only mean radon concentrations can be calculated but the time course of the radon concentration in each zone of the building can be determined. Above considerations are put together into a test procedure named RACODE (radon potential determination by controlled building depressurisation). The test procedure is expected to give good results for single-zone and two-zone buildings with basement. Multi-zone buildings can be very complex; characterisation of the leakage pattern may require a great deal of work with the uncertainties of the measured leakage rates between zones being rather high. Even when diagnostic pressure tests yield a good description of the leakage pattern, the use of these data for the models requires a degree of engineering. Either both the leakage asymmetry and the box parameter have to be estimated (Model A) or the exact location, the ELA, and n of the ﬂow paths have to be entered into the model (Model B). The next step is to apply RACODE to houses with known mean radon concentrations and to further improve the method. Some research will be needed to better quantify the temporal variability of the radon entry rate as a function of geology (soil permeability) and meteorology. Measurements at three mitigated houses indicate that BD tests can be very helpful for rapid veriﬁcation of the mitigation success if identical measurements are conducted before and after

Determination of the radon potential of a building by a controlled depressurisation technique (RACODE) 231

mitigation. This is true for most mitigation measures (e.g., sub-ﬂoor ventilation, sub-ﬂoor or sub-house depressurisation, sealing) but not all (e.g., if the basement is slightly depressurised in winter by an open chimney). The cost of performing RACODE is about € 80–100, i.e., about ten times higher than for traditional short-term or long-term measurements. Therefore it will be mainly applied in situations where conventional measurements are not suitable, for example if the radon levels need to be known quickly (e.g., after remediation), for research purposes (inﬂuence of leakage pattern on radon level, localisation and quantiﬁcation of radon entry paths, etc.), or if air tightness tests are performed anyway.

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16]

A.V. Nero, in: W.W. Nazaroff, A.V. Nero (Eds.), Radon and Its Decay Products in Indoor Air, Wiley, 1988. W.W. Nazaroff, S.R. Lewis, S.M. Doyle, B.A. Moed, A.V. Nero, Environ. Sci. Technol. 21 (1987) 459–466. W.W. Nazaroff, Health Phys. 55 (1988) 1005–1009. G. Åkerblom, P. Anderson, B. Clavensjo, Radiat. Prot. Dosim. 7 (1984) 49–54. R.G. Sextro, et al., in: P.K. Hopke (Ed.), Radon and Its Decay Products: Occurrences, Properties, and Health Effects, American Chemical Society, Washington, DC, 1987, pp. 10–29. W.W. Nazaroff, R.G. Sextro, LBL-25886, LBL, University of California, Berkeley, USA, 1988. B.H. Turk, J. Harrison, R.J. Prill, R.G. Sextro, Health Phys. 59 (4) (1990) 405–419. A.J. Gadgil, Radiat. Prot. Dosim. 45 (1) (1992) 373–380. A. Cripps, Time-Dependent Modelling of Soil Gas Movement: A Literature Review, BRE Report 298, Building Research Establishment, England, 1995, ISBN 1 86081 0578. D. Saum, M. Modera, Inﬁltec, Falls Church, VA, USA, 1991. M.H. Sherman, LBL-31305, LBL, University of California, Berkeley, USA, 1992. R.G. Sextro, et al., LBL-33682, LBL, University of California, Berkeley, USA, 1993. V. Rogers, K.K. Nielson, V.C. Rogers, R.B. Holt, EPA/600/SR-96/126, US EPA, 1996. K. Garbesi, et al., Environ. Sci. Technol. 27 (3) (1993) 466–473. K. Garbesi, PhD thesis, LBL, University of California, Berkeley, USA, 1994. G.N. Walton, NISTIR 6056, US Department of Commerce, NIST, Building and Fire Research Laboratory, Gaithersburg, MD 20899, USA, 1997.

232

Population exposure to inhaled radon and thoron progeny O. Iacob, C. Grecea, E. Botezatu Radiation Protection Department, Institute of Public Health, 14, Victor Babes Street, Iasi 6600, Romania

The radiation dose from inhaled radon and thoron progeny indoors is the dominant component of population exposure to natural radiation sources accounting for about 60 per cent in Romania. The radon and thoron short-lived decay product concentrations were measured in 760 typical urban and rural Romanian houses, randomly selected throughout the country.The method used to determine the volumetric activity of 218 Po, 214 Pb, 214 Bi and 212 Pb is the active one of sucking a known volume of air through an open-faced high-efﬁciency ﬁlter paper and counting the deposited activity with a ZnS(Ag) alpha scintillation counter. The detection limits were of 1 Bq m−3 for each of 218 Po, 214 Pb, 214 Bi and 0.1 Bq m−3 for 212 Pb. Internal exposures due to inhalation of radon and thoron progeny indoors and outdoors were expressed in terms of effective dose, individual and collective. The population-weighted averages of equilibrium equivalent concentration (EEC) of radon and thoron indoors were 25 Bq m−3 and 1.1 Bq m−3 respectively, with the corresponding average annual per capita effective dose of 1.88 mSv and 0.37–2.25 mSv in total. Outdoors values of EEC, were 5.7 Bq m−3 for radon progeny and 0.3 Bq m−3 for thoron progeny and the corresponding annual per capita effective doses were 0.14 mSv and 0.04 mSv, 0.18 mSv in total. The overall annual per capita effective dose arising from internal exposure to inhaled radon isotopes was estimated at 2.43 mSv with an associated annual collective effective dose of 55 112 man Sv.

1. Introduction The radiation dose from inhaled radon and thoron progeny indoors is the dominant component of population exposure to natural radiation sources accounting for about 60 per cent in Romania [1]. The objectives of our work were to determine the general distributions of radon and thoron progeny concentrations in dwellings, the magnitude of individual and collective exposures and to assess the potential lung cancer risk. Since our last study [2], we have extended the indoor radon and thoron progeny measurements in rural areas. The rural houses differ from those built in urban areas in structure, construction technology, design, building RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07026-3

© 2005 Elsevier Ltd. All rights reserved.

Population exposure to inhaled radon and thoron progeny

233

materials, and heating. They are usually single-family dwellings where raised levels of radon and thoron progeny are expected to occur because of their activity concentrations in subjacent ground and surrounding soil [3].

2. Materials and methods The radon and thoron short-lived decay product concentrations have been measured in 760 typical urban and rural Romanian houses, randomly selected throughout the country. The standardized method used to determine the volumetric activity of 218 Po, 214 Pb, 214 Bi and 212 Pb is the active one of grabbing a known volume of air (about 0.6 m3 ) through an open-faced highefﬁciency ﬁlter paper (98%) at known ﬂow rate (0.01–0.08 m3 min−1 ) for a certain collection time (usually 10 minutes) and counting the deposited activity with a ZnS alpha scintillation counter (40% efﬁciency) during four counting intervals. A computer program, very versatile in handling multiple input parameters, has been developed to solve the decay equations for obtaining the activity concentrations of daughters in air, to calculate the equilibrium equivalent concentration (EEC), the potential alpha energy concentrations and the equilibrium factor F , for radon daughters. The detection limits of our method are, in optimum sampling and counting conditions, 1 Bq m−3 for each of 218 Po, 214 Pb, 214 Bi and 0.1 Bq m−3 for 212 Pb [4]. Internal exposure due to inhalation of radon and thoron progeny indoors and outdoors has been expressed in terms of effective dose [5]. In dose estimates, the dose conversion coefﬁcients adopted by the UNSCEAR-2000 Report of 9 nSv (Bq h m−3 )−1 has been used for radon daughters, indoors and outdoors, and, for thoron progeny, of 40 nSv (Bq h m−3 )−1 indoors as well as outdoors [6]. An indoor occupancy factor of 0.75 and a population size of 22.68 million inhabitants have been considered in all calculations.

3. Results and discussion The results of all measurements show a log-normal distribution, as is illustrated in Fig. 1 for 222 Rn progeny and in Fig. 2 for 212 Pb, the most important short lived decay product of thoron. The average value of equilibrium factor F for indoor radon short-lived decay products was 0.51 ± 0.19, ranging from 0.1 to 0.9. The value of F varies in dwellings with ventilation rate and presence of aerosol sources such as smoking or cooking and is affected to some extent by the incoming air. High ventilation rate and low aerosol concentration give values of F signiﬁcantly lower than 0.5. When ventilation is low and aerosol concentration is high the value of F is higher and so is also the dose. Figure 3 shows the results in histogram form. The results of our study, as equilibrium equivalent concentrations of radon and thoron and the resulting annual effective dose for an adult, are presented in Table 1. In detached houses, the average values of EEC of radon and thoron were 36.3 Bq m−3 and 1.4 Bq m−3 , respectively. In blocks of ﬂats, these average values were approximately two times lower, 11.7 Bq m−3 for radon progeny and 0.8 Bq m−3 for thoron progeny. The population-weighted averages of the EEC of radon and thoron indoors have been estimated

234

O. Iacob et al.

Fig. 1. Log-normal cumulative frequency plot of radon daughter concentrations.

Fig. 2. Log-normal cumulative frequency plot of 212 Pb concentrations.

Fig. 3. Distribution of equilibrium factor F for indoor radon daughters.

Population exposure to inhaled radon and thoron progeny

235

Table 1 EECs of radon and thoron and the resulting annual individual exposures (adult) Source of exposure

Location

Equilibrium equivalent concentration (Bq m−3 ) average

222 Rn daughters

220 Rn daughters

Indoors Detached house Block of ﬂats Outdoors Indoors Detached house Block of ﬂats Outdoors

25.0∗ 36.3 11.7 5.7 1.1∗ 1.4 0.8 0.3

range 10–564 3.8–21.6 2.0–8.3 0.2–6.4 0.2–2.8 0.1–0.6

Annual individual effective dose (mSv) average 1.48∗ 2.15 0.69 0.11 0.29∗ 0.37 0.21 0.03

range 0.59–33.3 0.22–1.28 0.04–0.17 0.05–1.68 0.05–0.59 0.01–0.05

∗ Population-weighted average.

Table 2 EECs of radon and thoron with respect to building materials Building material

Equilibrium equivalent concentration (Bq m−3 ) Radon daughters Thoron daughters average range average range

Brick Wood Adobe Raw clay Hollow masonry blocks Autoclaved concrete

19.0 ± 14.4 8.3 ± 5.8 18.5 ± 10.2 26.4 ± 16.2 13.7 ± 11.8 7.0 ± 3.5

2.3–564 1.6–37.8 2.5–101 5.5–375 2.7–39.0 3.4–13.3

0.68 ± 0.52 1.11 ± 0.79 1.46 ± 1.04 1.66 ± 1.27 0.77 ± 0.48 0.46 ± 0.27

0.1–5.9 0.2–6.4 0.1–12.8 0.1–6.1 0.3–1.9 0.2–0.9

at 25 Bq m−3 and 1.1 Bq m−3 , respectively. The corresponding values for outdoors were 5.7 Bq m−3 and 0.3 Bq m−3 for radon and thoron progeny, about 5 times smaller than indoors. Table 2 presents the average equilibrium equivalent concentrations of radon and thoron determined with respect to building materials. The average values of EEC of radon in rural dwellings varied from 7.0 to 26.4 Bq m−3 as a function of building materials used, with individual values ranging from 1.6 Bq m−3 to 564 Bq m−3 . The average values of EEC of thoron were between 0.46 Bq m−3 and 1.66 Bq m−3 taking into account the building materials. The individual values ranged from 0.1 Bq m−3 to 12.8 Bq m−3 . Data in Table 2 clearly indicate that rural houses built of raw clay in a wooden framework presented substantially higher radon and thoron levels than houses constructed of wood, brick or concrete blocks. The corresponding average individual annual effective doses are listed in Table 3. The average annual effective doses received by people living in houses constructed of different types of building materials had values between 0.53 mSv in cellular autoclaved concrete houses and 2.0 mSv in houses built of raw clay in a wooden framework from both radon and

236

O. Iacob et al.

Table 3 Annual effective doses from 222 Rn and 220 Rn progeny with respect to building materials Building material Radon daughters average limits Brick Wood Adobe Raw clay Hollow masonry blocks Autoclaved concrete

1.12 ± 0.85 0.49 ± 0.34 1.09 ± 0.60 1.56 ± 0.96 0.81 ± 0.70 0.41 ± 0.21

Annual effective doses (mSv) Thoron daughters average limits

0.14–33.3 0.09–2.24 0.15–5.97 0.33–22.2 0.16–2.21 0.20–0.70

0.18 ± 0.14 0.29 ± 0.21 0.38 ± 0.27 0.44 ± 0.33 0.20 ± 0.13 0.12 ± 0.07

0.03–1.55 0.05–1.68 0.03–3.36 0.03–1.60 0.08–0.50 0.05–0.24

Total 1.30 ± 0.89 0.78 ± 0.39 1.47 ± 0.65 2.0 ± 1.37 1.01 ± 0.71 0.53 ± 0.22

Table 4 Annual effective doses and the estimated health effects Location

Exposure source

Annual per capita (mSv)

Effective dose collective (man Sv)

Potential number of radioinduced lung cancers

Indoors

222 Rn progeny

1.88 0.37 2.25

42 638 8392 51 030

3725

Total outdoors

0.14 0.04 0.18

3175 907 4082

298

Total

2.43

55 112

4023

220 Rn progeny

Total indoors Outdoors

222 Rn progeny 220 Rn progeny

thoron progeny inhalation. Table 3 also conﬁrms that radon and thoron progeny inhaled in detached houses is the most variable source of human exposure, with annual exposures ranging over three orders of magnitude: from 0.09 mSv to 33.3 mSv for radon daughters and from 0.03 mSv to 3.36 mSv for thoron daughters. The estimated overall effective dose arising from internal exposure to inhaled radon isotopes in the course of a year is summarized in Table 4. The average annual per capita effective doses received by the population indoors were estimated at 1.88 mSv and 0.37 mSv from radon and thoron progeny respectively, 2.25 mSv in total. Exposures for children have been calculated on the assumption that effective doses for the age group of up to ten years might on average be a factor of 2 higher than for adults. The corresponding values outdoors (0.14 mSv and 0.04 mSv) are much smaller. The overall annual per capita effective dose arising from internal exposure to inhaled radon isotopes was estimated at 2.43 mSv with an associated annual collective effective dose of 55 112 man Sv. Implications for public health of radon and thoron progeny exposure indoors can be estimated from this collective dose by applying the nominal fatality and detriment coefﬁcient of 7.3 × 10−2 Sv−1 adopted by the ICRP in its Publication 65 for risk assessment [7]. As an estimate, 4023 lifetime radio-induced lung cancers each year might be attributed to inhalation of radon and thoron short-lived decay products.

Population exposure to inhaled radon and thoron progeny

237

4. Conclusions This study on the population exposure to inhaled radon and thoron progeny in Romania led to the following results: – The values of indoor equilibrium equivalent concentrations of radon and thoron presented a log-normal distribution. – The average values of EEC of radon were 36.3 Bq m−3 in detached houses and 11.7 Bq m−3 in a block of ﬂats with individual values ranging from 3.8 to 564 Bq m−3 . – The average values of EEC of thoron were 1.4 Bq m−3 in detached houses and 0.8 Bq m−3 in a block of ﬂats, with individual values ranging from 0.2 to 6.4 Bq m−3 . – The indoor population-weighted average of the EECs of radon and thoron were estimated at 25.0 and 1.1 Bq m−3 , respectively. – The average annual effective dose for an adult was 1.48 mSv from indoor inhaled radon progeny, individual values ranging between 0.22 and 33.3 mSv. – The average annual effective dose for an adult from thoron progeny inhalation indoors was 0.29 mSv, individual values ranging between 0.05 and 1.68 mSv. – The corresponding average annual per capita effective doses were 1.88 and 0.37 mSv, 2.25 mSv in total. – Outdoors values of EEC of 5.7 for radon and 0.3 Bq m−3 for thoron and, consequently, the corresponding annual per capita effective doses (0.14 and 0.04 mSv, 0.18 mSv in total) are much smaller. – The overall annual per capita effective dose arising from internal exposure to inhaled radon isotopes was estimated at 2.43 mSv with an associated annual collective effective dose of 55 112 man Sv. As an estimate, 4023 lifetime radio-induced lung cancers each year might be attributed to inhalation of radon and thoron short-lived decay products by the Romanian population.

References [1] O. Iacob, J. Prev. Med. 4 (2) (1996) 73–82. [2] O. Iacob, C. Grecea, O. Capitanu, V. Rascanu, J. Prev. Med. 9 (2) (2001) 5–12. [3] O. Iacob, C. Grecea, in: International Conference of High Levels of Natural Radiation and Radon Areas, vol. 2, Munich, 2002, pp. 156–158. [4] ICRP Publication 60: 1990 Recommendations of the International Commission on Radiological Protection, Ann. ICRP 21 (1–3) (1991). [5] IRS, Romanian Standard SR-13397, 1997. [6] UNSCEAR, Exposures from natural radiation sources, Annex B in: Sources and Effects of Ionizing Radiation, UNSCEAR 2000 Report to the General Assembly, with Scientiﬁc Annexes, United Nations, New York, 2000, pp. 50–123. [7] ICRP Publication 65: Protection against radon-222 at home and at work, Ann. ICRP 23 (2) (1993).

238

Preferential radon transport through highly permeable channels in soils R.B. Mosley US Environmental Protection Agency, Ofﬁce of Research and Development, National Risk Management Research Laboratory, Research Triangle Park, NC 27711, USA

Indoor radon levels (that can pose a serious health risk) can be dramatically increased by air that is drawn into buildings through pipe penetrations that connect to permeable channels in soils. The channels, commonly containing gravel bedding around utility pipes, act as a collection plenum for soil radon and can draw air from distances approaching 100 m. Equations characterizing air and radon ﬂow in such channels are developed and compared with ﬁeld data in this paper. This pollutant entry mechanism has recently attracted new attention because of its relevance to entry of volatile organic compounds from contaminated groundwater, leaking storage tanks, landﬁlls, and other sources of soil vapor contamination. Three test channels were constructed to simulate conditions associated with utility line installations. The channels were constructed in 0.3 m wide trenches at depths between 0.9 and 1.2 m. A 63 m long channel was ﬁlled with clean gravel, and two 30 m long channels were ﬁlled with native soil and sand, respectively. The trench volumes above the channels were backﬁlled with native soil. A suction tube was installed at one end of each channel for pumping air from it, and air-sampling tubes were installed at regular intervals along the channel to monitor air pressure distributions and radon concentrations. Site sampling characterized soil radium, emanation, moisture, particle size, density, speciﬁc gravity, permeability, and diffusion coefﬁcient properties. Soils were relatively homogeneous in all respects except for reduced density in the recompacted soil above the channels and slightly reduced moisture in near-surface soils. The soil radon generation rate was 99.9 Bq m−3 s−1 , compared to 7.4 Bq m−3 s−1 for the gravel and 29.6 Bq m−3 s−1 for the sand. The effective permeability of site soils was observed to increase over a 3-month period during the spring and early summer. Experiments, in which air was extracted through the suction tube, showed that pressures and air ﬂow rates decreased exponentially with distance along each channel, as predicted. Pressure inﬂuences in the gravel channel propagated more than 30 m, while those in the soil and sand channels were limited to approximately 5 m. Pressures calculated from independent permeability measurements agreed with measured pressure proﬁles in the gravel channel within an average of 12%. Radon concentrations in the channels were lower than in surrounding soil because of RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07027-5

Published by Elsevier Ltd.

Preferential radon transport through highly permeable channels in soils

239

their greater porosity and reduced radon source strengths. With constant suction from the end of the gravel channel, radon concentrations within 10 to 20 m of the suction point were diluted by inﬁltrating surface air despite the increased advective transport of radon from surrounding soils into the channel. The resulting radon proﬁle had concentrations near the outlet that were about 75% of the concentrations at 60 m distance. Calculated concentrations near the outlet averaged within 14% of measured concentrations, and calculated concentrations at the opposite end were within 2% of the measured concentrations. Despite dilution by inﬁltrating surface air, radon produced by the gravel channel was approximately 12.7 Bq s−1 , which is sufﬁcient to produce indoor radon concentrations above 222 Bq m−3 in a typical single-story house with a ventilation rate of 0.5 air changes per hour.

1. Background Radon (222 Rn) gas enters buildings primarily from radium (226 Ra) in foundation soils. If the radon entry rate is elevated and the building is not well ventilated, the radon can accumulate to levels that signiﬁcantly increase the risk of lung cancer in chronically exposed occupants. Their degree of health risk is proportional to their long-term average level of radon exposure. The US Environmental Protection Agency (EPA) [1,2] recommends remedial action if indoor radon levels average 148 Bq m−3 (4 pCi L−1 ) or higher. Indoor radon concentrations are expected to be proportional to soil radium concentrations. However, the dependence of indoor radon on soil properties is sometimes obscured by factors such as: ﬂuctuating building ventilation rates and air pressures, heterogeneity of soil radon sources and transport rates, and poorly characterized cracks and openings in the building foundation. These complicating effects are sufﬁciently inﬂuential that some empirical studies, when not properly designed, have even failed to show a correlation between indoor radon and soil properties [3]. However, most studies show clear correlation of radon levels with soil properties. Mathematical models [4–8] have helped quantify the amount of radon produced by soils and how it enters houses and accumulates indoors. Model calculations of soil radon entry have shown excellent agreement [7] with measured data for carefully constructed test structures and for many houses. However, serious discrepancies are also observed in many cases. For example, radon levels in the Lawrence Berkeley Laboratory (LBL) test structures exceeded model calculations by as much as a factor of 8 when soil gas ﬂow was modeled as the only radon entry mechanism [9]. Although radon diffusion can also cause signiﬁcant radon entry, the LBL data suggest that pressure-driven soil gas ﬂow caused the anomalous radon levels. Several explanations have been offered for excessive soil gas ﬂow into structures. These include enhanced permeability of backﬁll soils or heterogeneous layers, anisotropic soil permeability due to sedimentary deposition, and permeable soil channels associated with animal burrows or buried utility lines. While backﬁll zones and soil layering have been the subject of previous ﬁeld studies, permeable soil channels are more difﬁcult to ﬁnd and have generally been ignored. Mosley [10,11] has developed a mathematical model indicating important contributions from the permeable channels commonly associated with utility pipes. Since the channels connect to houses at pipe penetration points, the houses can potentially draw soil gas from the channels through leaks in the pipe-concrete joint. Soil gas entry from the channel is enhanced

240

R.B. Mosley

by the common use of permeable gravel bedding in pipe trenches. Even when native soil is used to backﬁll around pipes, it is not ordinarily compacted to achieve a permeability as low as that of the surrounding soil. Mechanical vibrations and temperature changes in the pipe may also create a concentric permeable zone. The permeable channel model [10] indicates that air movement along pipe channels could account for 50 to 75% of indoor radon concentrations if the channel permeability is approximately 10 000 times that of the surrounding soil. While gravel in pipe channels could readily provide such a permeability difference, the rate of air movement in the channels needs empirical conﬁrmation. This paper describes a ﬁeld study aimed at testing the permeable channel model equations for preferential air ﬂow. The study involved the construction of test channels in a homogeneous, low-permeability soil and measurement of air pressure and ﬂow distributions in the channels. Radon concentrations were also measured along the channels and in air drawn from the channels for comparison with radon source strengths computed from radium and radon emanation rates in the surrounding soils. The measurements were compared with trends predicted by the permeable channel model to estimate the potential signiﬁcance of indoor radon entry from air ﬂow along buried utility lines. The equations characterizing preferential air ﬂow along permeable soil channels are presented in Section 2, along with their implications for indoor radon contributions. The experimental measurements in the permeable channels are described in Section 3, including site characterization, channel construction, and measurement methods. The results of the site characterization and channel air ﬂow measurements are presented in Section 4, followed by model analyses and comparisons with the experimental data in Section 5. Conclusions are summarized in Section 6.

2. Pertinent equations 2.1. Air ﬂow and pressure equations The rate of air ﬂow and the distributions of air pressures along a permeable channel connected to a house are deﬁned by Mosley [10] for the simpliﬁed geometry illustrated in Fig. 1. The deﬁnitions assume that homogeneous soil extends inﬁnitely on either side of the channel and from the surface to a depth much greater than the channel depth. The channel is deﬁned as a circular cylinder of inﬁnite length extending from the house with a deﬁned pressure difference relative to the outdoor pressure at the soil surface. The air ﬂow and pressure distributions are derived [10] using Forchheimer’s extension of Darcy’s law to obtain the following expressions for air ﬂow and pressure as a function of distance from the house-end of the channel: q(z) = − 3πb2 /2f sech2 (αz + δ), (1) P (z) = (3μ/4f k2 α) tanh(αz + δ) sech2 (αz + δ) where q = entry ﬂow rate into a short segment of the channel (m3 s−1 ), z = position along the channel from the house (m), b = radius of the channel (m),

(2)

Preferential radon transport through highly permeable channels in soils

241

f = Forchheimer constant for the channel ﬁll material (s m−1 ), sech = hyperbolic secant function, α = {k1 /[b2 k2 ln([h + (h2 − b2 )1/2 ]/[h − (h2 − b2 )1/2 ])]}1/2 , k1 = air permeability of the soil (m2 ), k2 = air permeability of the material in the channel (m2 ), h = depth of the channel from the soil surface (m), δ = tanh−1 ({1 − [2f/3πb2 ]QT }1/2 ), tanh = hyperbolic tangent function, QT = total ﬂow rate through the channel (m3 s−1 ), P (z) = air pressure in the channel at z relative to ambient air pressure (Pa), and μ = air viscosity (1.85 × 10−5 Pa s−1 ). The idealized geometry shown in Fig. 1 is used to approximate an experimental conﬁguration as shown in Fig. 2. Experimentally, the cylindrical channel is approximated by a square geometry, and the uniform soil surrounding the channel has been disturbed and re-compacted for construction of the channel. Although equations (1) and (2) assume the properties of the undisturbed (region 1) and recompacted (region 1’) soils are identical, measurements show these assumptions to be incorrect. An attempt is made to account for such differences by using an effective permeability.

Fig. 1. (a) Schematic diagram of a house with a permeable channel to a basement entry point, and (b) a cross section of the permeable channel.

242

R.B. Mosley

Fig. 2. Schematic diagram of experimental channel cross section.

2.2. Equations for indoor radon entry The amount of radon entry from the permeable channel to the indoor environment is the total amount of radon that gets swept past the z = 0 point in the channel. This quantity has been approximated [10] by ﬁrst computing the radon entry into incremental sections along the length of the channel, and then integrating the amount of radon along the total channel length. The radon activity at the surface of the channel is obtained by solving the steady-state radon transport equation: D∇ 2 C − v/ε · ∇C + G − λC = 0

(3)

where D = diffusion coefﬁcient of radon in the pore space (cm2 s−1 ), ∇ = gradient operator, C = radon concentration (Bq m−3 ), v = Darcy velocity of gas ﬂow (m s−1 ), ε = soil porosity (dimensionless), G = radon generation rate (Bq m−3 s−1 ), and λ = radon decay constant (2.1 × 10−6 s−1 ). The components of air velocity are given by [10]: √ 4k1 P (z) h2 − b2 vx = √ √ μ ln[(h + h2 − b2 )/(h − h2 − b2 )] 2xy × (h2 − b2 + x 2 + y 2 )2 − 4y 2 (h2 − b2 )

and

(4)

Preferential radon transport through highly permeable channels in soils

√ −4k1 P (z) h2 − b2 vy = √ √ μ ln[(h + h2 − b2 )/(h − h2 − b2 )] h2 − b2 + x 2 − y 2 × (h2 − b2 + x 2 + y 2 )2 − 4y 2 (h2 − b2 )

243

(5)

where vx = horizontal air velocity perpendicular to the channel (m s−1 ), x = horizontal distance from the channel (m), y = vertical distance from the soil surface (m), and vy = vertical air velocity perpendicular to the channel (m s−1 ). Equations (3) to (5) are solved numerically to obtain the radon concentrations in the x–y plane at various values of z. The resulting concentrations at the surfaces of the cylindrical channel are then averaged over its circumference. The rate of radon entry into the building, approximated by the radon entry into the channel, is then obtained by numerically integrating the products of the radon concentrations and air ﬂow velocities into the channel as: 0 Cs (z)v(z) dz Entry = 2πb (6) ∞

where Entry = radon entry rate (Bq s−1 ), Cs (z) = average radon concentration at the channel surface (Bq m−3 ), and v(z) = average velocity of soil gas entering the channel at z (m s−1 ). As an alternative to the numerical solution of this set of equations, the entry rate is approximated using advective ﬂow through the recompacted soil and diffusion through the undisturbed soil. A second alternative approach uses similar assumptions to compute the radon concentration in the channel and the entry rate as the product of concentration and ﬂow rate at the building-end of the channel. 2.3. Approximations that apply for the experimental conditions The experimental conditions, described in Section 3, allow several approximations to be made to equations (1) through (6). For example, δ is always greater than approximately 1.75, so that tanh(αz + δ) is approximately unity, and sech(αz + δ) can be approximated by 2 exp[−(αz + δ)]. Making these substitutions into equations (1) and (2) gives: q(z) = −QT exp(−2αz)

(7)

P (z) = μQT /2πb2 αk2 exp(−2αz).

(8)

and

A comparison of measured radon concentrations and entry rates with model predictions requires the solution of the multidimensional equations (3) through (5), and subsequent integration of equation (6). In most instances, these solutions would require numerical methods of

244

R.B. Mosley

evaluation. Some analytical approximations have been applied to equations (3) through (5) to obtain approximate closed-form solutions. For the experimental values of h and b, the soil-gas velocity into the channel varies only slightly from the top of the channel to the bottom. Therefore, it is a good approximation to represent the x and y components of the velocity vector (equations (4) and (5)) by a radial velocity scaler that is independent of angle in cylindrical coordinates. Therefore, the velocity at the mid-plane of the channel is given by P (z) = μQT /2πb2 αk2 exp(−2αz) (9) where vr (z) = radial air velocity perpendicular to the channel at z (m s−1 ). The vr (z) is obtained from the sum of the squares of equations (4) and (5). The velocity given by equation (9) is the linear velocity, not the Darcy velocity. The average velocity, vz (z), inside the channel is readily obtained from equation (7) after dividing by the cross sectional area to give: vz (z) = vz (0) exp(−2αz)

(10)

where vz (0) = −QT An approximate solution to equations (3) through (5) is obtained by using equation (9) in an approximate version of equation (3). As will be shown in Section 4, the permeability of the backﬁll soil over the channels is greater than the permeability of the undisturbed native soil. This causes the advective ﬂow into the channel to be mainly from ﬂow through the backﬁll soil, so that this advective transport can be approximated over a small segment of the top surface. Furthermore, the air velocity in the backﬁll soil near the suction point is sufﬁciently high that the diffusive term in this region can be neglected. Transport through the rest of the soil around the channel is mainly by diffusion, so the transport in this undisturbed-soil region can be approximated by the analytical solution to the two-region, inﬁnite-medium cylindrical diffusion equation. This solution contains the modiﬁed Bessel functions I0 , I1 , K0 , and K1 . For the region of interest around the channel, these Bessel functions can be represented by their asymptotic approximations. The resulting advective and diffusive expressions can be combined to give: √ G2 /λ + (2/b) D1 /λG1 /λ 1 3 2αz Cb (z) = (11) 1 − exp −λεπh e /QT √ 3 1 + (2/b) D1 /λ /(πb2 ).

where Cb (z) = average radon concentration in the channel (Bq m−3 ), G1 = radon generation rate in region 1 (Bq m−3 s−1 ), and G2 = radon generation rate in region 2 (Bq m−3 s−1 ). The 1/3 factor in the last term in brackets exceeds 1/4 of the channel circumference because the depleted concentration from advection is assumed to decrease the channel boundary radon concentration near the advective zone. This depletion zone in the diffusion component of the concentration was assumed to increase linearly for two diffusion lengths on each side of the advective zone. Equation (6) assumes that the radon entry rate into a building equals the advective radon entry rate through the walls of the channel. This assumption neglects diffusive entry through the walls of the channel as well as decay while radon is in transit in the

Preferential radon transport through highly permeable channels in soils

245

channel. An analytical solution to equation (6) is obtained by using equations (9) and (11). The resulting expression is: √ G2 /λ + (2/b) D1 /λG1 /λ 1 3 2αz Cb (z) = (12) 1 − exp −λεπh e /QT . √ 3 1 + (2/b) D1 /λ A simpler and more direct expression for the entry rate is the product of the average channel radon concentration and the axial velocity at z = 0. Using equations (10) and (11), this product gives: √ 2bQT G2 + (2/b) D1 /λG1 1 3 2αz Entrya = (13) 1 − exp −λεπh e /QT . √ εhλ 6 1 + (2/b) D1 /λ Both equations (12) and (13) were derived for comparison to the experimental data. The radon generation rate, G, in a given medium is deﬁned as: G = RρEλ/ε

(14)

where R = radium concentration (Bq kg−1 ), ρ = dry density (g cm−3 ), E = radon emanation coefﬁcient (the fraction of decaying radium atoms that result in radon atoms suspended in the gas phase), ε = material porosity (fraction) = 1 − ρ/ρg , and ρg = speciﬁc gravity (g cm−3 ).

3. Experimental design and measurements A set of permeable test channels was constructed for measuring air ﬂow and pressure distributions to compare with the predictions from equations (1) and (2). This section describes the characterization of the site, the design and construction of the channels, and the measurements that were conducted on the completed channels. The selected site had favorable permeability properties, and was also observed to have adequate size, level surface topography, uniform clayey textures, favorable location, and possibly adequate soil thickness. In further investigations, soil samples collected at the east, center, and west areas of the site were tested for moisture content and radium concentration. The moisture averaged 25.8 ± 3.7%, and the radium concentrations averaged 80.1 ± 18.5 Bq kg−1 . The high moisture was typical of the clayey soil texture, and the radium concentrations were sufﬁciently high to generate measurable radon concentrations even during channel air ﬂow experiments. 3.1. Channel construction Three permeable channels were constructed at the site spanning a permeability range from 10−14 to 10−8 m2 . Figure 3 illustrates the cross section of each channel. The channels were constructed by excavating 0.3-m wide trenches to depths of 1.2 m. The lengths and layouts of the trenches at the site are illustrated in Fig. 4. Soils removed from each trench were placed

246

R.B. Mosley

Fig. 3. Cross section of the three permeable channels.

Fig. 4. Relative positions of the sampling locations and the air access tubes along the channels.

at its side for later use as backﬁll. After trench excavation, a 0.3-m layer of gravel, native soil, or sand was installed in each trench, as illustrated in Fig. 3. A sheet of permeable, woven geomembrane was then installed over the gravel and sand layers to protect against inﬁltration of the clayey ﬁll soil. Air suction and sampling tubes were next installed in the trenches. At the east end of each trench, a metal duct ﬁtting was installed to connect a ﬂexible plastic tube to the 0.3-m square cross section of the ﬁll material, as shown in Fig. 5. The duct ﬁtting was ﬁlled with gravel, sand, or soil from the adjoining channel. Air sampling tubes (1.2-m long, 6-mm diameter polyethylene) were then installed into the center of each section of channel ﬁll as shown in Figs. 3 and 5 at the locations indicated in Fig. 4.

Preferential radon transport through highly permeable channels in soils

247

Fig. 5. Construction diagram of the channel and air sampling tubes.

Native soil was ﬁnally placed back into the trenches and compacted with a 0.3 × 0.3-m power compactor. Final compaction of the soil was augmented by driving over the surface with the excavation backhoe. The ends of all tubes were kept sealed except when connected to pump, sampling, or pressure measurement ﬁttings. The ﬁnished channels were allowed to equilibrate and settle for approximately 90 days after construction, during which time the site was generally covered with snow. 3.2. Site description and characterization A single boring was completed to a depth of 2.7 m to investigate the extent of the clayey surface soils. The boring utilized a 5-cm diameter soil auger (Model 405.23, Arts Manufacturing & Supply, American Falls, ID). Visual observations of soils from the auger cuttings indicated moist clayey soils with no prominent layering throughout the entire proﬁle. The borehole was logged with a gamma scintillation detector (Model 44-3, Ludlum Measurements, Inc., Sweetwater, TX) to estimate the relative distribution of radium and thorium activities. The gamma logging was limited to a depth of 2.2 m by the detector cable length. Extensive soil sampling utilized soils excavated from trenches during construction of the permeable channels. As illustrated in Fig. 4, samples were collected at nine locations in the ﬁrst trench and at seven locations in the other two trenches. The soil sampling locations were spaced closer together at the east end, where air pumping and pressure measurements were planned to emphasize the material properties more than those at the more distant locations. At each sampling location, soils were collected to represent three different depth intervals: 0 to 0.4 m, 0.5 to 0.8 m, and 0.9 to 1.2 m. Therefore, a total of 69 samples were collected to characterize the soils around the test channels. Each sample was immediately sealed in a polyethylene bag for storage pending laboratory compositing and analysis. Triplicate samples were also collected from the gravel and sand materials hauled to the site for ﬁlling the permeable channel sections of the trenches. Laboratory analyses included measurements of soil radium concentration, radon emanation coefﬁcient, moisture, density, speciﬁc gravity, and particle size distribution. Soil radium, radon emanation, and moisture measurements were performed on 35 individual soil samples,

248

R.B. Mosley

four composite soil samples, and the triplicate sand-and-gravel-ﬁll samples using a previously validated protocol [12]. Triplicate density measurements for the sand and gravel utilized standard laboratory Proctor compaction equipment. Soil density was determined from two in-situ samples of undisturbed soils and from three in-situ samples of recompacted soils using the drive-cylinder method [13]. Speciﬁc gravity was measured on seven soil samples and single sand-and-gravel samples by displacement techniques [14]. Particle size distributions were measured with sieve and hydrometer methods [15] on six soil samples and single sand-andgravel samples. The radon diffusion and air permeability properties of the gravel, soil, and sand were also measured for use in radon generation and transport modeling. Since these properties depend on both density and moisture, estimated conditions for each material were used in performing the laboratory tests. The radon diffusion coefﬁcients were measured using the transientdiffusion method reported previously [16]. The air permeability constants were determined from air pressure/ﬂow data from 10-cm diameter laboratory tubes packed with the gravel or sand samples. Air pressure and ﬂow measurements over an extended range were also used to determine the Forchheimer constant for the gravel [10]. Laboratory air permeability measurements were not attempted for the site soils, since in-situ permeability measurements were planned to coincide with the ﬁeld studies of the completed channels. 3.2.1. Soil air permeability The air permeability of soils at the channel site was initially measured in April 1995 at three depths using driven probes (Type GP, Rogers & Associates Engineering Corp., Salt Lake City, UT). The probes were connected to a permeameter that measured suction pressures and ﬂow rates of air drawn from the soil (Model MK-II Radon/Permeability Sampler, Rogers & Associates). The calibration and method for computing soil permeability from the pressure/ﬂow data are reported elsewhere [17]. Because of suspected soil drying during the April-June time period, additional soil permeability measurements were conducted in June 1995 by the same method upon completion of the channel air pressure/ﬂow tests. The air permeability in the recompacted soils over the channels was measured separately from that of the undisturbed soils. Moisture proﬁles in undisturbed soils were measured by sampling two 1.5-m boreholes in early April 1995. Soil boring utilized the same equipment and methods as for the site characterization boring described in Section 3.2. Subsequent surface samples were collected from backﬁll and undisturbed soil locations for moisture measurements at the completion of ﬁeld measurements in June 1995. 3.2.2. Channel air and radon dynamics The air ﬂow dynamics of the channels were characterized by monitoring air pressures at the various sampling tubes while a vacuum pump drew air from the main suction tube. An in-line ﬂow meter (RMA Series, Dwyer Instruments, Inc., Michigan City, IN) was used between the vacuum pump and the main suction tube to monitor total channel air ﬂow rates. Flow rates between 20 and 70 L min−1 were achieved by a carbon-vane pump or a shop vacuum cleaner to characterize the pressure/ﬂow properties of the channels. Air pressures along the channel were measured by successively attaching the air pressure gauge manifold of an MK-II unit to each tube and reading the suction pressure from the most sensitive gauge. The total radon production rates of each channel were measured by monitoring radon concentrations in the

Preferential radon transport through highly permeable channels in soils

249

efﬂuent air drawn through their main suction tubes. The radon concentrations were monitored by circulating a fraction of the air from the main suction tube through an alpha scintillation cell (110A, Pylon Electronics, Inc., Ottawa, ONT, Canada) attached to a continuous radiation monitor (AB-5, Pylon Electronics, Inc.). The radon monitor was attached between the ﬂow meter and vacuum pump on the main suction tube, and utilized its internal vacuum pump to sample a fraction of the efﬂuent air stream. Alpha scintillation counts were recorded at regular intervals of 0.5, 1, 2, or 5 minutes. Radon concentrations were calculated from the alpha count rates using the calibration method and equations of Thomas and Countess [18]. In a separate experiment with the gravel channel, radon concentrations were measured at the various locations along the channel during constant pumping of 70 L min−1 of air from the channel. For comparison, radon was also measured at several of the channel locations before the pumping had disturbed the channel radon distribution. The radon measurements involved successive connection of the radon monitor to different air sampling tubes and monitoring radon over ﬁve 1-minute intervals (1-L min−1 air sampling rate). The 70 L min−1 channel suction pump was operated for 90 minutes before radon measurements were taken to allow the radon distribution to approach a steady-state condition.

4. Results This section presents the results of the site characterization and test channel measurements. The site characterization measurements established the fundamental properties of the native soil and channel ﬁll materials at the test site. These provided an important basis for calculating the ﬂow and radon dynamics of the channels. The channel measurements of pressure/ﬂow and radon production properties provide an empirical benchmark for comparison with the model predictions. 4.1. Site characterization Results from gamma ray measurements in a borehole are presented in Fig. 6. They show relatively uniform gamma activities even to a depth of 2.2 m. Measurements near the surface are

Fig. 6. Depth in the borehole versus gamma activity.

250

R.B. Mosley

Fig. 7. Three-dimensional variations in the means of the radium, radon emanation, and moisture measurements.

lower because there is less soil above these points (2 vs. 4π solid angle), but are otherwise consistent with the relatively uniform proﬁle for the subsurface depths. The geometric standard deviation of 1.076 indicates less than 8% relative variation among the radioactivity levels at different depths. Since radium activities were found to signiﬁcantly exceed thorium activities at this site, the gamma ray log also suggests a relatively uniform radon source throughout the 2.2-m soil layer that contains the 1-m deep test channels. The radium, radon emanation, and moisture measurements on the site soils were averaged by vertical layer, by trench, and by position from the east end of the trench to estimate the site uniformity in all three dimensions. The results of these uniformity estimates are summarized in Fig. 7. As illustrated, there are no clear trends in the horizontal distributions of any of the parameters nor in the vertical distribution of radium concentrations. The site averages of all of the radium, radon emanation, and moisture measurements are presented in Table 1. The soil radium concentrations exhibit remarkable uniformity, with an overall relative standard deviation among all of the measurements of only 18%. The soil radon emanation coefﬁcients are distributed somewhat more widely, with an overall relative standard deviation of 25%. Soil moistures had a relative standard deviation of only 14%. The radium concentrations, radon emanation coefﬁcients, and moistures in the gravel and sand were considerably lower than in the clayey soil, as would be expected. The particle size distributions of the gravel, soil, and sand materials were measured. The gravel was predominantly 5 to 15 mm in diameter, with no sands or ﬁner material. The soil was mostly clay, distributed between < 0.001 and 0.1 mm in diameter. The sand was nar-

Preferential radon transport through highly permeable channels in soils

251

Table 1 Site-average properties of soil and ﬁll materials∗ Material

Radium (pCi g−1 )

Radon emanation (fraction)

Moisture (% dry mass)

Dry density (g cm−3 )

Speciﬁc gravity (g cm−3 )

Gravel Soil Sand

0.6 ± 0.1 (3) 2.1 ± 0.4 (47) 1.0 ± 0.1 (3)

0.05 ± 0.01 (3) 0.16 ± 0.04 (45) 0.07 ± 0.01 (3)

2.4 ± 0.1 (3) 27.5 ± 3.9 (35) 5.8 ± 0.2 (3)

1.51 ± 0.03 (3) 1.59 ± 0.06 (2) 1.77 ± 0.02 (3)

2.61 (1) 2.70 ± 0.02 (7) 2.68 (1)

∗ Mean ± standard deviation (number of measurements in parentheses).

rowly distributed between 0.1 and 5 mm, with approximately 5% clay. The results of the radon diffusion measurements are summarized in Fig. 8. The respective moistures used for the gravel and sand measurements (0.0 and 5.6%) are similar to ﬁeld values (2.4 and 5.8%, from Tables 1 and 4), and are not sufﬁciently different to signiﬁcantly affect the radon diffusion coefﬁcients [19]. The moistures for the soil radon diffusion measurements at saturation correspond most closely to the average ﬁeld moistures in Table 1, which also correspond to a saturation condition. Therefore, the radon diffusion coefﬁcient for the undisturbed soils is approximated from Fig. 8 by the value 1.0 × 10−5 ± 7 × 10−6 cm2 s−1 . The initial laboratory air permeability measurements for the gravel and sand materials averaged 6.2 × 10−8 ± 8.2 × 10−9 m2 and 4.1 × 10−11 ± 3.5 × 10−12 m2 , respectively. The sample moistures for these measurements were similar to those used for the diffusion measurements. The Forchheimer constant was determined from laboratory permeability measurements made on the gravel over a wide range of air pressures and ﬂow rates. These measurements were used to compute an average air permeability of 5.4 × 10−8 m2 with a Forchheimer factor of 13.7 m s−1 . The ﬁrst set of in-situ air permeability measurements made in the undisturbed site soils is summarized in Table 2. As indicated, the measurements were very narrowly distributed around the overall mean of 4 × 10−14 ± 9 × 10−15 m2 , with minimal depth and location trends. The

Fig. 8. Radon diffusion coefﬁcients as a function of moisture.

252

R.B. Mosley

Table 2 Initial in-situ air permeability of the site soils Air permeability (10−14 m2 )

Distance from east end (m)

0 18 34 46 61 Mean ± std. dev.

Mean ± std. dev.

0.5 m depth

0.9 m depth

1.2 m depth

7.0 3.9 3.7 3.9 3.4 4.4 ± 1.5

3.9 3.5 4.6 3.7 4.4 4.0 ± 0.5

3.6 3.5 3.4 3.8 3.4 3.5 ± 0.2

4.8 ± 1.9 3.6 ± 0.2 3.9 ± 0.6 3.8 ± 0.1 3.7 ± 0.6 4.0 ± 0.9

Table 3 Final in-situ air permeability of the site soils Undisturbed soils between channels

Recompacted soil over channels

Location∗

Air permeability (m2 )

Location†

Air permeability (m2 )

C1–2 @ 0.9 m C1–2 @ 3.7 m C1–2 @ 63.0 m C1–2 @ 63.0 m C1–2 @ 63.0 m C2–3 @ 0.9 m C2–3 @ 18.3 m Mean ± std. dev.

3.9 × 10−13 7.0 × 10−13 7.0 × 10−13 1.4 × 10−12 6.6 × 10−14 5.8 × 10−13 7.0 × 10−14 (5.6 ± 4.7) × 10−13

C1 @ 0.9 m C1 @ 13.7 m C1 @ 30.5 m C1 @ 61.0 m C2 @ 0.9 m C2 @ 13.7 m C3 @ 0.9 m Mean ± std. dev.

9.3 × 10−11 1.6 × 10−10 2.0 × 10−10 4.7 × 10−10 1.4 × 10−10 2.8 × 10−10 1.4 × 10−10 (2.1 ± 1.3) × 10−10

∗C m–n denotes locations between channels m and n at 1 m depth. The position is from the east end of the channels. † C denotes locations in channel n at 0.6 m depth. The position is from the east end of the channel. n

ﬁnal measurements of air permeability showed generally higher values for the undisturbed soils, and much higher values in the recompacted soils over the channels, as shown in Table 3. Soil moisture proﬁles analyzed in early April 1995 showed moisture depletion at the 0.3 m depth. Surface samples collected in June 1995 from the 0.1 m depth suggested further depletion in surface soil moisture, particularly in the recompacted soils over the test channels. These observations are consistent with the air permeability measurements in the undisturbed and recompacted soils. 4.2. Pressure measurements The results of the air pressure and ﬂow measurements in the gravel, sand, and soil channels are presented in Figs. 9 through 11, respectively. As illustrated, the air pressures were observed to decrease approximately exponentially with distance from the suction point in all of the channels. Therefore, the measurements are ﬁtted by least-squares regressions to equivalent lines in Figs. 9 through 11 to help identify the pressure-ﬂow characteristics of each channel. As illustrated by Figs. 9 through 11, the exponents measured for each channel were similar even when different suction pressures were applied. For the gravel channel, the exponents

Preferential radon transport through highly permeable channels in soils

Fig. 9. Pressure proﬁles in the gravel channel.

Fig. 10. Pressure proﬁles in the sand channel.

Fig. 11. Pressure proﬁles in the soil channel.

253

254

R.B. Mosley

averaged 0.14 ± 0.02 m−1 , corresponding to pressure inﬂuences as far away as 30 m from the suction point. The exponents for the soil channel were expectedly greater, averaging 0.70 ± 0.19 m−1 , for a more rapid pressure attenuation within approximately 5 m. However, the exponents for the sand channel were expected to be intermediate, but instead averaged 1.3 ± 0.19 m−1 . The intercepts for all of the channels (pressure at z = 0) were dependent on the amount of air being pumped from the channels. The unexpectedly large exponent for the sand channel is attributed to the high permeability of the recompacted soil. The effective permeability of the soil channel exceeded that of the sand by more than a factor of 5. The soil channel propagated suction pressures further because the channel and its cover had similar permeabilities.

5. Discussion and conclusions The comparisons of measured and calculated channel characteristics focused on the gravel channel because of its large permeability difference from the surrounding soil and its resulting potential for providing useful empirical data. The soil channel was not expected to function as a permeable conduit, but rather as a limiting reference case. The sand channel was also found not to propagate air pressure or ﬂow as far as expected because of the higher permeability of the overlying backﬁll soil. The following sections summarize the measured properties of the channels and their materials and compare the measured air and radon ﬂow characteristics with theoretically calculated values. 5.1. Values of parameters used for model analyses Several soil and channel parameters identiﬁed earlier must be deﬁned numerically to calculate comparison values for pressure and radon proﬁles. These parameter deﬁnitions are sumTable 4 Summary of channel properties Parameter (units)

Undisturbed soil

Recompacted soil

b (m) h (m) f (s m−1 ) R (pCi g−1 ) E (fraction) ρ (g cm−3 ) ρg (g cm−3 ) ε (fraction) M c (% dry mass) M d (fraction) D (cm2 s−1 ) k (m2 )

N.A. N.A. N.D. 2.1 0.16 1.59 2.70 0.411 25.8 0.998 1.0 × 10−5 5.6 × 10−13

0.172 1.07 N.D. 2.1 0.16 1.46 2.70 0.459 29.9 0.951 3.2 × 10−5 2.1 × 10−10

Gravel 0.172 1.07 13.7 0.6 0.05 1.51 2.61 0.578 2.4 0.0627 6.9 × 10−2 5.4 × 10−8

Sand 0.172 1.07 N.D. 1.0 0.07 1.77 2.68 0.340 5.8 0.302 4.6 × 10−2 4.1×10−11

N.A. means not applicable, N.D. means not determined, M c is % moisture on dry mass basis, M d = M is the fraction of moisture saturation.

Preferential radon transport through highly permeable channels in soils

255

marized in Table 4. The channel radius, b, was calculated as the radius of a circle with an area equal to that of the 30 × 30 cm channel. The depth, h, was represented by the design depth to the center of the channel. The Forchheimer factor was measured from the non-linear pressure-ﬂow data for channel gravel, but was not determined for the other materials. Radium concentrations and radon emanation coefﬁcients used in equation (14) were obtained from the measured means in Table 1. The density and speciﬁc gravity properties of each material, also obtained from data in Table 1, were used to calculate the porosities listed in Table 4. The respective radon generation rates in the site materials, calculated from equation (14), were Gsoil,un = 0.10 Bq m−3 s−1 , Gsoil,re = 0.081 Bq m−3 s−1 , Ggravel = 0.0059 Bq m−3 s−1 , and Gsand = 0.0285 m−3 s−1 . These values indicated that the soils were the primary radon sources, and that the sand and gravel ﬁll materials had less signiﬁcant radon production. Radon diffusion coefﬁcients were deﬁned directly from the measured values for the sand and gravel, and from the measured mean for the nearly saturated undisturbed soil. For the recompacted soil, the diffusion coefﬁcient was interpolated from the measured values in Fig. 8 at the indicated fraction of moisture saturation. The air permeabilities were deﬁned from laboratory measurements on gravel and sand, and from ﬁeld measurements for the soils. The soil permeability measurements in Table 4 are the means of seven measurements each in the undisturbed and recompacted regions. The measured air pressure proﬁles shown in Figs. 9 through 11 ﬁt the predicted exponential dependence of pressure on distance given by equation (8). The least-squares ﬁtting coefﬁcients for the P vs. z data in these ﬁgures were therefore used to directly estimate the effective permeability properties of the channels. By substituting the Table 5 Effective air permeability of channel and soil materials computed from air pressure proﬁles Channel air ﬂow (L min−1 ) Gravel channel 27 55 70 Mean Standard deviation Soil channel 28 50 65 Mean Standard deviation Sand channel 34 50 65 Mean Standard deviation

2α (m−1 ) 0.153 0.156 0.114

Maximum air pressure (Pa) 7.99 34.5 41.7

k1 , soil air permeability (m2 )

k2 , channel air permeability (m2 )

6.4×10−11 3.1×10−11 2.4×10−11 3.9×10−11 2.1×10−11

7.3 × 10−8 3.4 × 10−8 4.9 × 10−8 5.2 × 10−8 2.0 × 10−8

0.430 0.815 0.847

13.2 48.4 109.9

1.1×10−10 1.0×10−10 6.2×10−11 9.3×10−11 2.7×10−11

1.6 × 10−8 4.2 × 10−9 2.3 × 10−9 7.6 × 10−9 7.6 × 10−9

1.47 1.14 1.15

60.5 115.9 203.4

1.0×10−10 6.1×10−11 4.6×10−11 6.9×10−11 2.9×10−11

1.3 × 10−9 1.3 × 10−9 9.2×10−10 1.1 × 10−9 2.0×10−10

256

R.B. Mosley

mathematical deﬁnition of α and the known values of b and h from Table 4 into equation (8), expressions were derived to solve for the effective values of k1 and k2 for each measured proﬁle. The resulting effective air permeability values are presented in Table 5. The effective permeabilities in Table 5 are consistent with the more detailed site measurements. Since the recompacted soil region above the channels covers 25% of their perimeter and the undisturbed soil covers 75% of their perimeter, a 3-to-1 weighted average air permeability can be calculated using measurements from Table 4. The resulting effective permeability, 5.3 × 10−11 m2 , is within 2% of the average empirically ﬁtted value of 5.2 × 10−11 m2 for the soil around channel 1. Similar permeability values were also obtained for the soils around the other channels. The effective permeability of the gravel in Table 5 is within 4% of the measured value in Table 4. However, there were larger differences between measured and effective ﬁtted values for the soil (27×) and sand (36×) channels. These larger differences were mainly inﬂuenced by the proximity of the channel permeabilities with those of the recompacted soils covering the channels. 5.2. Comparison of measured and calculated radon parameters The radon concentration proﬁle measured along the gravel channel during an air ﬂow rate of 70 L min−1 is illustrated in Fig. 12. For comparison, the average radon concentration with standard deviation (15 910 ± 1850 Bq m−3 ) in the undisturbed channel prior to pumping is also shown by the straight solid-and-dashed lines. The radon concentrations during pumping approached the undisturbed values at distances of approximately 20 m and greater from the suction point. The depletions at locations closer to the suction point are attributed to dilution by air drawn through the soil above the channel. The vertical error bars correspond to 1 standard deviation from the measured means. The radon parameters used to compare measurements with theoretical values include radon proﬁles at different locations along the gravel channel and total radon production rates from the gravel channel. The data used in the calculations are presented in Table 5, and the measured data are shown in the empirical proﬁle in Fig. 12. The calculated radon concentration proﬁle for the gravel channel is shown

Fig. 12. Comparison of measured radon concentrations with model predictions.

Preferential radon transport through highly permeable channels in soils

257

in Fig. 12 for comparison with the measured proﬁle. Since there was considerable scatter in the data near the origin, the average radon concentration of 10 878 Bq m−3 for all of the measurements in the ﬁrst 10 m was compared to the calculated concentration of 9361 Bq m−3 for locations near the origin. In this comparison, the calculated concentration is only 14% less than the measured concentration, and is well within the range of experimental variations. At 60 m, near the far end of the channel, the calculated concentration of 14 023 Bq m−3 is within 2% of the measured concentration. The actual radon proﬁle reaches its maximum in a shorter distance from the origin than the calculated proﬁle. The experimental rate of radon production by the gravel channel was determined as the product of the measured air ﬂow rate at the origin (1.17 × 10−3 m3 s−1 ) and the measured radon concentration at the origin (10 878 Bq m−3 ). This product gave a radon entry rate of 12.7 Bq s−1 . The two alternative equations for computing radon entry give Entrya = 10.6 Bq s−1 [from equation (12)] and Entryb = 12.3 Bq s−1 [from equation (13)]. The value of Entryb is about 1% below the measured value, while the value of Entrya is about 17% below the measured value. Both values are within the estimated experimental uncertainty of the measured value (±4.07 Bq s−1 ). As expected from the theoretical derivations, Entrya is lower than Entryb .

6. Summary and conclusions Several signiﬁcant observations were made from this study of preferential radon transport through permeable channels in soils. Some observations concern channel construction, and others concern the physics of air and radon transport into and from the channels. Channels with up to 10 000 times greater air permeability than surrounding undisturbed soils may be constructed using gravel ﬁll material, but are less likely to be constructed if sand is used in the channel. Even though the channel ﬁll may have the requisite high permeability, trench construction to install the channels disturbs the natural soils enough to increase their air permeability by several orders of magnitude. This creates a semipermeable zone of recompacted soil above the channels that limits their air transport distances. Despite standard construction attempts to compact the site soils to their original condition, the air permeability of soils above the test channels increased by more than a factor of 300. The resulting effective permeability of the site soils was still more than 1000 times lower than the gravel channel material. Although other sites may offer better soil recompaction, the present experience suggests that a factor of 1000 may be more typical of utility channel permeability ratios than a factor of 10 000. The pressure distributions along the soil channels were observed to follow exponential decreases from the suction point, regardless of the preferential inﬁltration of surface air through re-compacted soils. Least-squares regressions of the pressure proﬁles also provided empirical estimates of the effective air permeability of both channel and surrounding soil materials. These estimates were consistent with direct measurements of site soil permeability and gravel permeability. Measured and calculated pressure distributions in the gravel channel agreed within an average of 4%.The radon concentrations in the gravel channel were lower than in surrounding soil pores because of the greater pore space and lower radon source strength of the gravel. With constant suction at one end, the radon concentrations in the gravel channel

258

R.B. Mosley

became diluted by inﬁltrating surface air, despite the increased advective transport of radon from surrounding soils into the channel. The resulting radon proﬁle had concentrations at the channel outlet that were about 75% of the concentrations at a 60 m distance. Most of the depletion occurred in the ﬁrst 10 to 20 m from the outlet. Calculated radon concentrations near the channel outlet averaged within 14% of measured concentrations, and calculated concentrations at the opposite end of the channel were within 2% of the measured concentration. Even though signiﬁcant surface air was drawn into the channel near its exit point, the radon levels were still sufﬁcient to cause elevated indoor radon levels. The radon production rate of the gravel channel was on the order of 12.7 Bq s−1 , which is sufﬁcient [7] to cause radon concentrations exceeding 222 Bq m−3 (the EPA action level is 148 Bq m−3 ) in a typical single-story house with a ventilation rate of 0.5 air change per hour.

References [1] US Environmental Protection Agency, A Citizen’s Guide to Radon, second edition, EPA/402/K-92-001, Ofﬁce of Air and Radiation, Washington, DC, May 1992. [2] US Environmental Protection Agency, Technical Support Document for the 1992 Citizen’s Guide to Radon, EPA/400/R-92-011 (NTIS PB 92-218395), Ofﬁce of Radiation Programs, Washington, DC, May 1992. [3] R. Nason, B.L. Cohen, Correlation between 226 Ra in soil, 222 Rn in soil gas, and 222 Rn inside adjacent houses, Health Phys. 52 (1987) 73–77. [4] C.O. Loureiro, L.M. Abriola, J.E. Martin, R.G. Sextro, Three-dimensional simulation of radon transport into houses with basements under constant negative pressure, Environ. Sci. Technol. 24 (1990) 1338–1348. [5] K.L. Revzan, W.J. Fisk, A.J. Gadgil, Modeling radon entry into houses with basements: model description and veriﬁcation, Lawrence Berkeley Laboratory report LBL-27742, Berkeley, CA, 1991. [6] M. Swami, L. Gu, V. Vasanth, Integration of radon and energy models for buildings, Florida Solar Energy Center report FSEC-CR-617-93, Cape Canaveral, FL, 1993. [7] K.K. Nielson, V.C. Rogers, V. Rogers, R.B. Holt, The RAETRAD model of radon generation and transport from soils into slab-on-grade houses, Health Phys. 67 (1994) 363–377. [8] D.J. Holford, Rn3D: A Finite Element Code for Simulating Gas Flow and Radon Transport in Variably Saturated, Nonisothermal Porous Media: User’s Manual, Version 1.0, Paciﬁc Northwest Laboratory report PNL8943, Richland, WA, 1994. [9] K. Garbesi, R.G. Sextro, W.J. Fisk, M.P. Modera, K.L. Revzan, Soil-gas entry into an experimental basement: model measurement comparisons and seasonal effects, Environ. Sci. Technol. 27 (1993) 466–473. [10] R.B. Mosley, A mathematical model describing radon entry aided by an easy path of migration along underground channels, Paper VIP-3 in: Proceedings of The 1992 International Symposium on Radon and Radon Reduction Technology, Minneapolis, MN, 1992. [11] R.B. Mosley, Model based pilot scale research facility for studying production, transport, and entry of radon into structures, Paper VI-8, in: Proceedings, The 1992 International Symposium on Radon and Radon Reduction Technology, Minneapolis, MN, 1992. [12] K.K. Nielson, V.C. Rogers, A sensitive efﬂuent method for measuring radon gas emanation from low-emanating materials, Nucl. Instrum. Methods Phys. Res. A 353 (1994) 519–523. [13] ASTM, Standard test method for density of soil in place by the drive-cylinder method, test D2937-83, American Society for Testing and Materials, Philadelphia, PA, 1984. [14] ASTM, Standard test method for speciﬁc gravity of soils, test D854-83, American Society for Testing and Materials, Philadelphia, PA, 1984. [15] ASTM, Standard method for particle-size analysis of soils, American Society for Testing and Materials, test D422-63, Philadelphia, PA, 1963. [16] K.K. Nielson, D.C. Rich, V.C. Rogers, Comparison of radon diffusion coefﬁcients measured by transientdiffusion and steady-state laboratory methods, US Nuclear Regulatory Commission report NUREG/CR-2875, Washington, DC, 1982.

Preferential radon transport through highly permeable channels in soils

259

[17] K.K. Nielson, M.K. Bollenbacher, V.C. Rogers, User’s guide for the MK-II radon/permeability sampler, Rogers & Associates Engineering Corp. report RAE-9000/9-2, Salt Lake City, UT, 1989. [18] J.W. Thomas, R.J. Countess, Continuous radon monitor, Health Phys. 36 (1979) 734–738. [19] V.C. Rogers, K.K. Nielson, Correlations for predicting air permeabilities and 222 Rn diffusion coefﬁcients of soils, Health Phys. 61 (1991) 225–230.

260

A new approach to increasing the uptake of radon remediation in England S. Scivyer Environmental Engineering Centre, BRE Environment, Building Research Establishment Limited, Garston, Watford, Hertfordshire, WD25 9XX, United Kingdom

Whilst radon protective measures are routinely installed within new buildings in areas of England affected by radon the number of building owners who remediate existing buildings is low. Despite UK government publicity initiatives and measurement campaigns it is estimated that probably no more than about 20% of houses with elevated levels have been remediated. There are many reasons why the uptake has been so poor, including a lack of understanding of the health implications, lack of knowledge of what to do, the cost of remediation, lack of conﬁdence in local builders, and the possible effect on property prices. In addition to these issues there is a further reason as to why people have not remediated. The radon issue is seen as being yet another government initiative being forced upon people in the regions. Unfortunately the approach taken to running the various radon measurement campaigns has not helped in this. Whilst the offer of free testing has been successful in encouraging some 30–40% of householders in affected areas to have their homes tested, it has not helped to encourage remediation. This is largely because invitations, and the despatch of detectors and results, have all been undertaken by post. The result is that householders receiving elevated radon readings for their homes feel that they are left on their own to resolve the problem. Radon has not been seen to be a local issue. Recognising this to be a problem, the UK government funded a pilot study between 1998– 2000, to see if local authorities could be used as a focus for raising public awareness and encouraging remediation. The project involved local authorities in three radon affected areas of England. Government funded support in the form of publicity materials and technical expertise from national organisations such as the Building Research Establishment and National Radiological Protection Board. In each area training was provided for local authority staff, builders, surveyors, estate agents, health professionals, local councillors and others to raise awareness locally. Different approaches were tried on a more localised basis to increase public awareness. The initial route was to write to homeowners with elevated radon levels to encourage them to remediate. This was then followed up by the local authority offering either home visits or local radon exhibitions and surgeries for homeowners to offer advice on remeRADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07028-7

© 2005 Elsevier Ltd. All rights reserved.

A new approach to increasing the uptake of radon remediation in England

261

dial options. The aim was to show homeowners that remedial measures are relatively simple and inexpensive to install. The pilot studies proved extremely successful in increasing the uptake of remediation – in one area the number of homes remediated was doubled. As a consequence the programme has now been rolled out to a further 32 local authorities across England. This paper describes in more detail the results of the pilot study and the public awareness raising activities now being rolled out across England. 1. Introduction Whilst radon protective measures are routinely installed within new buildings in England in areas affected by radon, the number of building owners who remediate existing buildings is low. Despite UK government publicity initiatives and measurement campaigns, it is estimated that probably no more than about 20% of houses with elevated levels have been remediated. There are many reasons why the uptake has been so poor, including a lack of understanding of the health implications, lack of knowledge of what to do, the cost of remediation, lack of conﬁdence in local builders, and the possible effect on property prices. Possibly the biggest reason however has been that radon has not been seen to be a local issue. Recognising this to be a problem, the UK government funded a pilot study between 1998– 2000, to see if local authorities could be used as a focus for raising public awareness and encouraging remediation. The pilot study proved extremely successful in increasing the uptake of remediation – in one area the number of homes remediated was doubled. As a consequence the programme is now being rolled out to a further 33 local authorities across England. This paper describes in more detail the results of the pilot study and the public awareness raising activities now being rolled out across England. 2. The radon programme in England The UK Government’s radon programme for England began in 1987 and focused upon four tasks: 1. Measurement programme – this was largely undertaken by the National Radiological Protection Board (NRPB) and was aimed at identifying homes with elevated radon levels throughout the country. Earlier studies had identiﬁed the areas most likely to be affected so measurement was generally directed towards these areas. However measurement was also carried out in the lower risk areas so that in 1996 NRPB were able to publish a detailed map of affected areas throughout England. Throughout the period 1987 to 1995 free measurements were available to householders on demand, as well as those provided as part of targeted campaigns. Since 1996 free measurements have only been available in targeted campaigns in areas most likely to be affected by radon – essentially areas with a greater than 1% probability of being above the Action Level. These campaigns achieved greater than 30% response rates. Since 1998 free measurement has been further reﬁned with invitations sent to every home in England with a greater than 5% probability of being above the Action Level. More than 400 000 measurements have been carried out since 1987.

262

S. Scivyer

2. Research – funding research looking at health risks and measurement largely undertaken by NRPB, and research into the development of practical cost effective remedial measures for existing buildings and protective measures for new buildings largely undertaken by BRE. 3. Regulations – requirements were introduced to reduce the risk from radon in the workplace. The Building Regulations were amended and supporting guidance introduced to provide adequate protective measures in new buildings in radon affected areas. 4. Grants – means tested grants for the most needy were made available for the installation of radon remedial measures.

3. A change in approach Whilst the work carried out between 1987 and 1997 had been reasonably successful in measuring homes and developing techniques for combating radon the levels of remediation remained low – probably no more than 10–20% of houses identiﬁed as having a problem being remediated. Research had shown that there were many reasons for this, including a lack of understanding of the health implications, lack of knowledge of what to do, the cost of remediation, lack of conﬁdence in local builders, and the possible effect on property prices. A signiﬁcant factor however, had been the remoteness of the campaign. The radon issue is seen as being yet another government initiative being forced upon people in the regions. Unfortunately the approach taken to running the various radon measurement campaigns has not helped in this. Whilst the offer of free testing has been successful in encouraging some 30–40% of householders in affected areas to have their homes tested, it has not helped to encourage remediation. This is largely because invitations, and the despatch of detectors and results, have all been undertaken by post. The result is that householders receiving elevated radon readings for their homes feel that they are left on their own to resolve the problem. Radon has not been seen to be a local issue.

4. The pilot studies Recognising this to be a problem, the UK government funded a pilot study between 1998– 2000, to see if local authorities could be used as a focus for raising public awareness and encouraging remediation. Local authorities provided the public face of the initiative and government provided background support. The pilot studies were organised with 3 local authorities. Two of these, Derbyshire Dales and Mendip, were long established radon affected areas whilst the third, Cherwell, was relatively new. The studies did not target all homes within each local authority district, only those areas with homes with a greater than 5% probability of being above the Action Level. Government support included providing: • “Free” radon measurements to identify homes above the Action Level, and to conﬁrm whether remedial action had worked. • Expert advice from NRPB to local authorities, local medical community and householders about the health effects of radon.

A new approach to increasing the uptake of radon remediation in England

263

• Expert advice from BRE to local authorities, local builders and housing professionals and householders about remediation methods. • Locally branded publicity materials to meet local councils’ needs. • Consultants to help local authorities develop local strategies (Action Plans). The general approach taken by the three local authorities was essentially the same although each tailored its campaign to reﬂect past activities and resources available locally. In each area two categories of householder were targeted: • All householders with elevated radon levels and thought not to have remediated (high testers). • All householders who had not had a radon test but who were in > 5% probability risk areas (ﬁrst time testers). In both cases invitations to have measurements were sent out by NRPB with a letter from the local authority encouraging testing and offering local points of contact for those with queries. To help raise public awareness generally in each area a range of approaches were tried: • “Radon Month” – concentrated public awareness activities over a one month period. • Local press, radio, and TV coverage. • Locally branded publicity material, including posters, bookmarks and leaﬂets. These general awareness initiatives were supplemented by a series of seminar and training events to improve understanding and provide technical support within in local authorities and for construction professionals. These included seminars for: • Local authority staff whether directly or indirectly involved in the local campaign. This included technical training for environmental health, building control, and housing professionals who were directly involved in advising householders of remediation options. Support staff such as receptionists were also provided with radon awareness training so that they could respond positively as the ﬁrst point of contact for householders seeking technical advice. • Local builders were given short training courses outlining the principle radon remediation techniques. • Health professionals, including local doctors, who can advise householders with health concerns. • Housing professionals including surveyors, estate agents and solicitors in order to allay fears about blight and to ensure that radon is seen to be just another building problem that is dealt with at the time of house purchase. The different local authorities, depending upon their circumstances regarding staff available to assist, used various different approaches to making direct contact with householders: • Home visits – ofﬁcers visiting householders with elevated levels to advise on possible remedial techniques and how to carry out the work. • Radon surgeries – householders invited to attend a local venue to obtain advice on possible remediation techniques, and how to carry out the work, from technical experts from BRE and local authority staff.

264

S. Scivyer

• Roadshow events – householders invited to attend a local venue to obtain advice and guidance on radon risks and possible remedial techniques from technical experts from NRPB, BRE and local authority staff. • Telephone – householders invited to phone in for advice. In each case householders were invited to participate by letter sent out by the local authority, again emphasising that radon is a local issue. 5. The results of the pilot studies The main result from the pilot studies has been a signiﬁcant increase in the number of houses measured, with a doubling of the number of houses remediated. The factors that inﬂuenced this success were: • • • • • •

Local delivery of advice and support. Effective targeting of key groups. Optimum use of technical expertise from NRPB and BRE. Deployment of simple consistent messages on health risks and remediation methods. Minimising effort demanded of householders. Sustained support and follow-up contact by the local authority.

6. The ‘roll-out’ programme With the pilot programme proving so successful the approach is now being rolled out to more local authorities across England. The main aims of the new programme, which is scheduled to run from 2000–2003, are to improve the amount of remediation amongst radon-affected households, to continue to raise awareness about radon, and to deliver this initiative at a local level. 78 Local Authorities with > 5% probability areas in their districts were invited to participate in the programme. Of these 32 Local Authorities have taken up the offer and are participating, working in 12 local groupings. Whilst only about halfway through the programme early indications look positive with a further increase of around 30% uptake of free measurement (range 20–55%). There also appears to be strong interest in carrying out remediation by householders. Interestingly the best response from the public appears to be in areas where senior local authority staff are getting involved in the programme. The programme is about to be reviewed with a view to seeing how successful it has been and to help shape any extension of the programme to include additional local authorities. Already the National Assembly for Wales are investigating running a similar programme for homes in Wales. 7. Conclusion The main conclusion is that local authorities, supported by government, can signiﬁcantly increase radon awareness and remediation in radon affected areas by establishing and maintaining a local infrastructure of help.

265

User-friendly computer programs that calculate radon cancer risks, progeny accumulation and gamma exposure S. Rydell US Environmental Protection Agency, Region 1, Ofﬁce of Ecosystem Protection, Radiation Unit, 1 Congress St., Boston, MA 02114, USA

Estimating risks from radon exposure may be inconsequential unless this information is made clear to (a) the at-risk population and/or (b) those making radon policy decisions in public health, radon education, and risk reduction programs. A decade of limited success in communicating radon risk information with tools such as nomograms, spread sheets, and perplexing computer programs has indicated the need for a different approach. In these programs, the approach is to document the risk model parameters and formulas used yet mask the technical complexity of the risk calculation engine, thus maintaining scientiﬁc rigor behind a user-friendly graphic user interface. This user interface simpliﬁes the input of information and the display of results in easy-to-read forms such as bar graphs, which has led to greater acceptance and use of risk-estimating models. Because they combine ease of use and technical accuracy/currency, these programs are useful to a wide range of technical and non-technical users, from individual homeowners to radiation scientists and radon technologists. Using the 1999 National Academy of Sciences BEIR VI health risk estimates, the models consider up to three radon sources and three exposure pathways. Since smoking is known to enhance the effects of radon exposure, smoking history is included when calculating risk in individual households. Interest in the models outside the United States encouraged internationalization through adding SI units to these computer programs originally developed for use in the United States. Two of a number of programs developed by the author will be presented: (1) UR3M, which addresses combined source radon risk evaluations and remediation decisions based on either a single house or a group of houses, and RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07029-9

© 2005 Elsevier Ltd. All rights reserved.

266

S. Rydell

(2) CARBDOSE, which addresses options concerning radon removal from water by GAC ﬁlters, the accumulation of progeny with their attendant gamma radiation, and the 210 Pb growth and ensuing ﬁlter disposal issues. A third program, the Bq–Ci calculator, is integrated into the other two programs and provides unit conversion to SI for international convenience.

1. Introduction Estimating risks from radon exposure may be inconsequential unless this information is made clear to (A) the at-risk population and/or (B) those making radon policy decisions in public health, radon education, and risk reduction programs. A decade of limited success in communicating radon risk information with tools such as nomograms, spread sheets, and perplexing computer programs indicated the need for a different approach. In the programs discussed here, the risk calculation engine comprises the appropriate risk model parameters and formulas, yet the technical complexity is masked behind user-friendly graphic user interfaces. Thus scientiﬁc rigor is maintained while user interfaces simplify information input and display calculation results in easy-to-read forms such as bar graphs. This has led to greater acceptance and use of risk-estimating models. Because they combine ease of use and technical accuracy, these programs are useful to a wide range of technical and non-technical users, from individual homeowners to radiation scientists and radon technologists.

2. Programs developed for the environmental protection agency Two of a number of programs developed by the author will be presented: (1) UR3M, which addresses combined source radon risk evaluations and remediation decisions based on either a single house or a group of houses, and (2) CARBDOSE, which addresses options concerning radon removal from water by GAC ﬁlters, the accumulation of progeny with their attendant gamma radiation, and the 210 Pb growth and ensuing ﬁlter disposal issues. These programs and additional radon-related programs developed by the author may be downloaded from the EPA Region 1 web site, www.epa.gov/region01/eco/wrsdp.html. Both UR3M and CARBDOSE use current National Academy of Sciences BEIR VI radon risk values [1]. The handling of radon concentrations in CARBDOSE and UR3M is analogous to mass balance type equations, but solved in terms of activity instead of mass. Both programs, which were developed for the US Environmental Protection Agency, can produce printed reports of the user’s calculations.

Computer programs that calculate radon cancer risks, progeny accumulation and gamma exposure

267

A program of voluntary registration by users allows us to contact them via e-mail to notify them of updated information and to have a crude estimate of who our users are. About 20% list themselves as individual users while 80% are listed with an organizational afﬁliation such as a radon mitigation business, radiation program, government agency, or university. At present, twenty ﬁve percent of users are from outside the United States, which we hope will increase as people become aware of the availability of our programs. Interest in the models outside the United States encouraged internationalization through adding SI units to the programs originally developed for use in the United States. A third program that in part accomplishes this, the Bq–Ci calculator, is included in the presentation and is integrated into the UR3M and CARBDOSE programs. The Bq–Ci calculator, described in a later section of this paper, is also available separately for international convenience. UR3M and CARBDOSE are described next.

3. UR3M (The Uniﬁed Radon Relative Risk Model) The Uniﬁed Radon Relative Risk Model (UR3M) provides a comprehensive approach to radon health risks and enables users to make informed decisions with respect to radon reduction. Risks are presented both numerically and graphically, calculated according to measured values for that instance. Individuals may customize their risk estimate based on their smoking history or the default may be accepted, calculated for the general population: an average of all smoking and non-smoking histories. The UR3M model combines three radon sources, three exposure pathways, and two risk reduction methods to give a composite assessment of health risks, attainable risk reductions, and initial cost factors associated with risk reduction. An individual’s smoking history may be included when calculating risk, thereby further personalizing the assessment. UR3M considers three signiﬁcant sources of radon in indoor air: (1) soil gas-derived radon that enters through openings, cracks, and pore spaces in basement walls or the ﬂoor slab, (2) water-derived radon, which is dissolved in water and then released into the air as a result of indoor water use, (3) ambient air radon that enters the building principally through doors, windows, and other openings. The concept behind the UR3M model is that the three signiﬁcant sources of radon risk combine and therefore should be considered jointly. When the radon concentration of indoor air is measured, the result reﬂects the combined radon sources at that time. Further, if a residence is normally occupied but has no source of soil gas radon or water derived radon (for example, a house trailer with no water supply), the long-term average of the radon content of the indoor environment should be the same as ambient air. If additional sources of radon are then introduced, they affect the indoor air/ambient air equilibrium, and as a consequence, indoor air radon content rises. This effect is cumulative and expressible as (Ambient Air Radon + Soil Gas Radon + Water Radon).

268

S. Rydell

Table 1 Risk pathways Inhalation of 222 Rn progeny (IHP) Inhalation of 222 Rn gas (IHG) Ingestion of 222 Rn gas (IGG) Total all pathways Cross-media transfer of radon from water to air (WTR)

1.60E–08 1.70E–010 1.90E–09 1.81E–08 10 000:1 or 0.0001

88% 1% 11% 100%∗

∗ NAS–NRC [2], attributes the principal inhalation risk (lung cancer risk) to inhalation of radon progeny (88%). The ingestion of radon gas is a secondary risk (11%). The radiation dose to other organs from inhalation of radon gas was estimated by NAS–NRC to be lower by a factor of 100 and accounts for about 1% of the risk.

A generally insigniﬁcant source of radon is the construction material of the building. If radon is emanating from construction materials, it is not a readily measurable quantity like ambient air radon or water-borne radon, nor is it calculable by a method similar to soil gasderived radon (see Section 5), where there is an equation with a single unknown. However, radon contribution that may be attributable to emanation from construction materials is incorporated in UR3M’s risks as an unresolvable (and unmitigable) part of the radon content of the house air measurement. UR3M uses a number of factors taken from [1–4]. These values are listed below. Table 1 above shows pathway risk factors for the general population expressed as a lifetime cancer risk per Becquerel per cubic meter (Bq m−3 ) in water. The signiﬁcance of these ﬁgures to individuals is that if they use water containing 11 000 Bq m−3 of 222 Rn over a period of 75 years, their chances of developing water radon-related cancer is about 1 in 5000. 4. Smoking history Since smoking is known to amplify the effects of radon exposure, smoking history is included when UR3M calculates risk in individual households. It is assumed that if one person in the household is a smoker, the other occupants run a similar risk. Smoking history risk factors from [2] are normalized to the general population and in this form can be substituted in the equations used by UR3M. The Never Smokers and Ever Smoker categories used in UR3M version 2.0 use the smoking history risk factors 0.312 and 1.62 respectively. NAS–NRC deﬁnes the Ever Smoker as a person who has smoked more than 100 cigarettes in his/her lifetime (Table 2). Table 2 Smoking history risk factors (multipliers of inhalation risks) General Population (GP) statistically includes all smoking and non-smoking categories listed below Never Smokers (NS) Ever Smokers∗ ∗ Combines former smoker and current smoker categories.

1.00 – No change to risk 0.312 – Lower risk than GP 1.62 – Higher risk than GP

Computer programs that calculate radon cancer risks, progeny accumulation and gamma exposure

269

5. Soil gas The model further stipulates that if values for (1) the radon content of indoor air, (2) the radon content of ambient air, and (3) the radon content of water have been measured or appropriate values have been otherwise provided, then the contribution of radon from the soil gas can be calculated through the following relationships. Calculate: Radon Content Soil Gas = RCS Given: Radon Content House (indoor) Air = RCH Radon Content Ambient Air = RCA Radon Content Water = RCW Water to Air Transfer Ratio (10 000 : 1) = 0.0001 UR3M Equation: RCS = RCH − RCA + (RCW · WTR) . The model calculates ambient air and soil gas radon cancer risks for the general population as follows: • Ambient Air Risk (AARGP): AARGP = (RCA · IHP) + (RCA · IHG) · KP · GP/WTR; • Soil Gas Risk (SGRGP): SGRGP = (RCS · IHP) + (RCS · IHG) · KP · GP/WTR where KP = 1000, GP = 1.00. To facilitate comparison, KP was set equal to 1000, thus facilitating transposing the risk to the same number of people as referenced in [5]. However, by simple multiples of this, the risk may be extended to community populations in general. The water transfer ratio is needed because the exposure pathway risks chosen are for water. If air exposure pathway risks had been chosen, the water risk would have an inverse correction. GP (General Population) is the factor 1.00 and acts as a placeholder for substitution of Never Smoker or Ever Smoker when the same formula is used in conjunction with an individual’s smoking history. The model calculates water-derived radon cancer risks for the general population as follows: Partial risks by pathway: • Inhalation of radon Progeny Risk (IHPR): IHPR = RCW · IHP; • Inhalation of radon Gas Risk (IHGR): IHGR = RCW · IHG;

270

S. Rydell

• Ingestion of radon Gas Risk (IGGR): IGGR = RCW · IGG. Total risk: • Total General Population Risk (TGPR): TGPR = (IHPR · GP) + (IHGR · GP) + IGGR. The ﬁnal equation used by UR3M caused the word “uniﬁed” to become part of the model’s name. The equation brings together the various risks and views cancer risks from radon comprehensively. • The total ambient air, soil gas, and water risk to the general population (TASWRG): TASWRG = AARGP + SGRGP + (TGPR · KP). UR3M Version 2.0 contains a number of functions dealing with radon and public water supply issues speciﬁc to the United States, which are mentioned here without elaboration: • • • • •

MCL and AMCL water risk differentials; AMCLs and Multimedia Mitigation (MMM); risk reduction equivalent to MCL treatment; calculating the number of MMM houses for equivalent risk reduction; factors to convert short term measurements to annual average living area Rn content.

For those who may be interested, there is extensive online documentation in the program help ﬁles.

Fig. 1. Input screen.

Computer programs that calculate radon cancer risks, progeny accumulation and gamma exposure

271

Figure 1 shows an input screen where measured values are entered and the house soil gas contribution calculated automatically. From entering as few as two measured values on the user input screen, the program can calculate risk estimates. While the program calculates risks expressed in conventional scientiﬁc notation, the goal of UR3M is to communicate radon risks to the public clearly and simply. It accomplishes this in two ways. • First, through selecting the focus, the risk estimate is either a single house individual risk estimate or a multiple houses public health risk estimate, thus also providing for different program functions to be enabled depending on focus choice of the user.

Fig. 2. Risk communication screen.

Fig. 3. Radon in water remediation example.

272

S. Rydell

• Second, by alternatively presenting risks in easy-to evaluate personalized bar graphs and as risk percentages, those interested in the risk, but without the technical background, are included in comprehension of the radon risk situation being presented. A screen of the UR3M program that illustrates the graphical displays appears in Fig. 2. It is a bar graph of the cancer risks corresponding to the numerical presentation in Fig. 1. Figure 3 shows the consequence in one bar of the removal of radon from water matching the values in Fig. 1. Other bars show how removal of radon from air or from both air and water can result in the reduction of radon risk.

6. CARBDOSE The CARBDOSE program was rewritten in 1999 to use the same NAS–NRC radon health risks as UR3M. This update of risk factors is particularly signiﬁcant for the ingestion risk for radon in water, which is 4 times less than thought in 1993 [6]. CARBDOSE reports risks via inhalation and ingestion separately as risks of cancer per 10 000 population. CARBDOSE reﬂects the US EPA proposed public water supply MCL and AMCL for radon. However, a ﬁnal PWS radon maximum contaminant level has not yet been issued. Health effect data in the program, including smoking history, are based on single dwelling point of entry volumes and, with the exception of gamma data, are independent of the removal method. Hence, CARBDOSE provides health effect information for radon removal by single dwelling aeration units or the larger packed tower aeration systems favored for public water supplies. CARBDOSE reports the following: 1. The amount of radon (222 Rn) removed per day. 2. The accumulation of radon on the ﬁlter. 3. The efﬂuent radon concentration and health effects, and the health effects associated with other radon concentrations. 4. The minimum Rn removal efﬁciency needed to meet the EPA proposed Public Water Supply MCL. 5. The probable gamma radiation dose at 1 meter from the tank wall and at other user speciﬁed-distances. 6. The radiation dose from an equivalent point source at any speciﬁed distance measured from the tank centerline. 7. The distance at which the probable exposure dose from the GAC tank becomes less than the residential health guideline. 8. The growth of 210 Pb and 210 Pb + progeny on the ﬁlter for the number of years speciﬁed. 9. The 210 Pb and 210 Pb + progeny activity per gram of carbon. 10. The time when the carbon exceeds the 74 Bq g−1 waste disposal category. The CARBDOSE program models a GAC ﬁlter at a state of dynamic equilibrium, which is attained after about 30 days by the linear (ﬂow rate determined) adsorption of radon followed by its exponential radioactive decay and progeny growth relationships. This assumes total retention of progeny and assigning all the radioactivity to a cylindrical disk 24 cm in diameter

Computer programs that calculate radon cancer risks, progeny accumulation and gamma exposure

273

Fig. 4. Waste disposal calculation screen.

and 13 cm thick. This is consistent with the observed gamma radiation associated with a thin layer at the top of an undisturbed GAC bed [7,8]. The program treats the disk as a cylindrically corrected 24 cm × 24 cm × 13 cm array of 1 cm cubes. The dose originating from each cubic centimeter is calculated for the 72 gamma energies of 214 Bi and 214 Pb. It is then corrected for self-absorption and the build-up factor associated with attenuation and scattering within the carbon/water volume distributed source. While decreasing the dose with distance, additional corrections for the absorption in air are made. In addition to the volume-based dose, the model calculates an equivalent point-source dose. The point-source dose is corrected for distance and absorption in air only. The validity of this model was conﬁrmed [9] by correlation between direct measurement of gamma radiation from GAC ﬁlters in use and the predictions of the theoretical CARBDOSE Model. The National Academy of Sciences [2] also found the gamma predictions of CARBDOSE in agreement with those of the more elaborate, commercially marketed Microshield program. A composite screen showing the CARBDOSE program’s waste disposal calculations, information message, and conversion to SI units using the embedded Bq–Ci calculator appears in Fig. 4. 7. Additional programs In addition, three other programs available from the EPA are described brieﬂy below.

274

S. Rydell

7.1. IONEX The IONEX (ion exchange) program performs calculations related to removing radium and/or uranium from water with the ion exchange process. IONEX quickly calculates whether a proposed ion exchange process will meet your requirements in terms of the following: (1) Producing a ﬁnished water of sufﬁcient quality to meet drinking water MCLs for radium or uranium; (2) Producing a waste product that can be disposed off in a manner consistent with EPA guidelines; (3) Operating without producing an unacceptable radiation health risk to treatment personnel or property owners. IONEX calculates radionuclide concentrations and gamma radiation doses with sufﬁcient accuracy for its intended purpose of design estimate calculation. 7.2. GARR The determination of concentrations of natural radioactivity in a public water supply begins with the measurement of its gross alpha particle activity. The GARR (Gross Alpha Radium Regulation) program follows a decision scheme where the laboratory radionuclide analysis for gross alpha activity is compared with the MCL to determine if compliance exists. If not, additional laboratory analyses for speciﬁc radionuclides might demonstrate compliance. Depending on the gross alpha value entered, the program may request additional user input regarding radium, uranium or radon before rendering a decision on COMPLIANCE vs. NONCOMPLIANCE status. Each time this program runs to its conclusion, it produces a determination of IN COMPLIANCE or NON-COMPLIANCE and writes this ﬁnding, along with the supporting data, to a ﬁle. This data ﬁle, which serves as a database for all calculations made by the GARR program, may be printed by the program, edited, or directly inserted into a report. (This program is being rewritten to reﬂect changes in the SDWA Radionuclides Rule (December 7, 2000).) 7.3. GBP CALCULATOR This program provides a fast and easy calculation of total annual dose equivalents from beta and photon emitters in water from Bq m−3 or pCi L−1 measurements. GBPcalc compares the total dose per year to the 4 mrem dose per year established as an MCL in the Radionuclides section of the Safe Drinking Water Act. It is named for the Gross Beta/Photon part of the regulation and uses the radionuclide data of the SDWA Radionuclides Rule (Federal Register, December 7, 2000). The guideline for groundwater cleanup is also 4 mrem per year at sites where beta/photon radioactive contamination has occurred. Use of GBPcalc is facilitated by online help, including sections on fundamentals of radioactivity and deﬁnitions.

Computer programs that calculate radon cancer risks, progeny accumulation and gamma exposure

275

References [1] Committee on Health Risks of Exposure to Radon, Board on Radiation Effects Research, Commission on Life Sciences, National Research Council, Health Effects of Exposure to Radon: BEIR VI, National Academy Press, Washington, DC, 1999. [2] Committee on Risk Assessment of Exposure to Radon in Drinking Water, Board on Radiation Effects Research, Commission on Life Sciences, National Research Council, Risk Assessment of Radon in Drinking Water, National Academy Press, Washington, DC, 1999. [3] Environmental Protection Agency, Radon in drinking water health risk reduction and cost analysis, Federal Register 64 (38) (February 26, 1999) 9559–9599. [4] Environmental Protection Agency, National primary drinking water regulation; radon; proposed rule, Federal Register 64 (211) (November 2, 1999) 59245–59378. [5] Environmental Protection Agency, A Citizens Guide to Radon, The Guide to Protecting Yourself and Your Family from Radon, second ed., ANR-464, EPA, 1992. [6] Environmental Protection Agency, Report to the United States Congress on radon in drinking water: Multimedia risk and cost assessment of radon, EPA 811-R-94-001, 1994. [7] J.D. Lowry, J.E. Brandow, Removal of radon from water supplies, J. Environ. Eng. 111 (1985) 511–527. [8] J.D. Lowry, Radon progeny accumulation in ﬁeld GAC units, Research Technical Report to Maine Department of Human Services, Division of Health Engineering, 1988. [9] S. Rydell, B. Keene, J. Lowry, Granular activated carbon water treatment and potential radiation hazards, J. N. England Water Works Assoc. 103 (4) (1989) 234–248.

276

Testing of radon-reducing measures under strictly controlled conditions in a laboratory house P. de Jong, W. van Dijk NRG Radiation & Environment Department, P.O. Box 9034, 6800 ES Arnhem, The Netherlands

A laboratory house has been constructed, consisting of two enclosed spaces, simulating a living area (height 2.6 m) with a crawlspace (height 1 m) beneath, both having dimensions of about 3.5 m × 2 m. The two spaces are separated by a concrete ﬂoor, 15 cm thick, the air leakage of which can be fully regulated. Phosphogypsum blocks with a total source strength of about 800 Bq radon-222 per hour are placed in the crawlspace. The house is climate controlled and the temperature, humidity, radon concentrations, airﬂow rates and pressure difference across the ﬂoor are registered every ten minutes. The facility can be used to test the effectiveness of various radon-reducing measures quickly, under precisely deﬁned and strictly controlled conditions. In this paper, consideration is given to the use of a polyethylene thinﬁlm membrane and the effects of increased crawlspace ventilation. A so-called airlock ﬂoor construction has been found to be the most effective measure, reducing the radon inﬁltration from the crawlspace by more than 95 per cent.

1. Introduction There are three established ways of studying the effectiveness of radon-reducing measures: (a) experimentally, at laboratory scale [1,2], (b) by means of mathematical modelling [3], and (c) by conducting comparative research under practical conditions [4,5]. The results of laboratory experiments can easily be entered into mathematical models in order to assess effectiveness. However, the radon concentrations depend on the type of soil underlying the dwelling and several characteristics of the building including type of house and foundation, year of construction and the building materials applied. As outlined in several studies all these factors together account for no more than 30 per cent of the total variation between dwellings [6,7], indicating that the effectiveness of countermeasures has to be veriﬁed under practical conditions. In general, this is performed by measuring the radon concentration in RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07030-5

© 2005 Elsevier Ltd. All rights reserved.

Testing of radon-reducing measures under strictly controlled conditions in a laboratory house

277

houses before and after remediation. Since the effectiveness shows a wide range, even within a same remedial measure [4,5], this option is less appropriate to compare similar measures. In the present study, we examined the applicability of a fourth option: the use of a 25 m3 test facility in which radon-reducing measures could be tested on a semi-practical scale. The advantage of such an approach is that the conditions in the laboratory house are well known and strictly controlled and the collected data will show less variation than data collected under ﬁeld conditions. Conclusions drawn from such research will therefore be more precise. The test facility has been used to assess the effectiveness of the Dutch government’s two preferred radon-reducing measures: reducing air inﬁltration through the ground ﬂoor and increasing ventilation of the crawlspace [8].

2. Materials and methods 2.1. Laboratory house The laboratory house consists of two enclosed spaces, simulating a living area with a crawlspace beneath. The heights of these spaces are 2.6 and 1.0 m, respectively, both having dimensions of about 3.5 m × 2 m. The walls of the crawlspace are made of limestone and constructed directly on the concrete ﬂoor of the laboratory hall. The walls of the crawlspace also support the ﬂoor of the house, which is made of solid concrete slabs 150 mm thick. The superstructure of the laboratory house is constructed from plywood, known to have both a low radon exhalation rate and a low air permeability. The structure features (a) four openings in the crawlspace and a further four in the living area, each with an internal diameter of 68 mm, for the in- and outlet of ventilation air, (b) an oscillating fan in the crawlspace and another in the living area, to ensure adequate mixing of the air in each space, (c) six openings in the ﬂoor, each provided with a plastic pipe with an internal diameter of 40 mm. Each pipe is threaded at the end, so that a cap may be screwed on. By using caps with perforations of different sizes, it is possible to simulate various levels of ﬂoor leakage. All connections are sealed. 2.2. Instrumentation The laboratory house is equipped with a climate control unit to control temperature and relative humidity in both the crawlspace (20 ◦ C, 90% RH) and the living area (20 ◦ C, 50%). The radon concentration in both parts of the house and the supply air are simultaneously measured using three Atmos-12 DPX ionisation chambers (Gammadata). In each of the four ventilation pipes oriﬁces of various sizes (Van Essen Instruments) can be placed to measure the set ﬂow rates. The oriﬁce induced pressure difference is determined using electronic pressure transducers made by Sensor Data. A similar transducer is used to monitor the pressure difference across the ﬂoor. All measured data items are registered every ten minutes by an acquisition

278

P. de Jong, W. van Dijk

system. Measurement continues for at least twenty hours after the establishment of equilibrium. From the measured data, the average and the standard error of the mean (SEM) are calculated. 2.3. Radon sources Phosphogypsum blocks are placed in the crawlspace as a radon source. Such blocks are known to release radon at a steady rate over time [9]. By measuring the radon concentration in the crawlspace at a known air exchange rate, it is possible to calculate the overall source strength. For the tests reported here, this was approximately 800 Bq h−1 . Before putting the phosphogypsum blocks in place, a number of experiments were performed, during which the pressure difference across the ﬂoor was varied in the range 0 to 55 Pa. At each pressure setting, the living area was ﬁrst ‘ﬂushed’ with an adequate quantity of outside air, then the ventilation ports were closed and the radon concentration monitored for twenty-four hours. The radon exhalation rate of the ﬂoor was then calculated from the linear increase in the concentration over time. The average value was 5.1 ± 0.3 Bq m−2 h−1 (n = 9). This ﬁgure is within the anticipated range [10]. The results indicate that the ﬂoor of the laboratory house did not exhibit pressure-driven exhalation. 2.4. Pressure ﬂow rate characteristic By determining the induced ﬂow rate at four to ﬁve pressure differences, a so-called ‘pressure ﬂow rate characteristic’ is obtained, which follows the equation: qv = C(P )n

(1) (m3 s−1 ),

(m3 s−1 ),

in which qv is the airﬂow rate C is the air leakage coefﬁcient deﬁned as the leak at 1 Pa, P is the pressure difference across the ﬂoor (Pa) and n is the ﬂow exponent. Both C and n can be determined by linear regression from the log-transformed measured data. Dividing C by the area in question results in the speciﬁc air leakage coefﬁcient with the dimension m3 /(m2 s). 2.5. Reducing radon inﬁltration Three different options for reducing radon inﬁltration from the crawlspace were simulated: (a) thin-ﬁlm membrane applied underneath the concrete ﬂoor, (b) as option (a), but with a small fan ﬁtted in the membrane, and (c) thin-ﬁlm membrane spread over the ground within the crawlspace. To enable a correct comparison of the effectiveness of these countermeasures, the same membrane was used. This membrane was ﬁtted as follows. A wooden rail was attached around the inside of the crawlspace wall, ﬁrmly bedded in sealant. The membrane (0.2 mm polyethylene) was then secured to the rail on one side of the crawlspace using a second strip of wood. Next, it was extended across the crawlspace and pulled taught before being secured to the other walls in the same way. Thus, the crawlspace was effectively divided into two sections: section A, containing the phosphogypsum blocks (volume 5 m3 ), and section B, between the membrane

Testing of radon-reducing measures under strictly controlled conditions in a laboratory house

279

and the concrete ﬂoor, with a volume of 0.8 m3 . For all the experiments, the air exchange rate in the living area was kept at a constant 0.5 L h−1 . For the simulation of option (b), a small variable-speed fan was used to create an under-pressure of approximately −5 Pa (relative to the living area) in section B of the crawlspace. In option (c), section B formed the ventilated crawlspace, while section A was not ventilated. 2.6. Improving crawlspace ventilation For assessment of the effect of improving crawlspace ventilation on the radon concentration in the living area, the pressure difference across the ﬂoor was kept at a constant value of about 5 Pa (range: 4.1 to 6.2). The air exchange rate in the living area was kept at 0.5 L h−1 , which is normal for a house in the Netherlands with closed windows and doors. The leakage of the concrete ﬂoor formed an additional variable.

3. Results 3.1. Characteristics of the laboratory house To enable the measured data to be interpreted properly, it was important that the shell of the upper part of the laboratory house permitted as little leakage as possible. The containment was accordingly tested by determining a pressure ﬂow rate characteristic, as described in Section 2.4, applying the same pressure in crawlspace and living area. The speciﬁc air leakage coefﬁcient was (2.9 ± 0.0) × 10−6 m3 /(m2 s), with a ﬂow exponent of 0.78 ± 0.00 (average ± SD, n = 5). For most of the experiments, the level of leakage was not relevant in relation to the measured data, since the pressure difference between the living area and the laboratory hall was kept close to zero at all times. Where this did not prove possible, the measured data were corrected to allow for the leakage on the basis of the parameter values set out above. Leakage from the crawlspace was of no relevance, since the radon concentration in the crawlspace was measured continuously. As indicated in Section 2.1, it was possible to adjust the leakage of the concrete ﬂoor by ﬁtting different sets of caps on the openings. A pressure ﬂow rate characteristic was determined for each set of (six) caps. The results obtained are set out in Table 1. By way of illustration, some of the measured data are presented in Fig. 1. Consideration was given to the extent to which the air leakage of the ﬂoor, combined with the pressure difference, inﬂuenced the radon concentration in the living area. The measured radon concentrations in the living area were corrected for the contribution of the ﬂoor and the supply air. The Table 1 Adjustable speciﬁc leakage coefﬁcient (10−6 m3 /(m2 s)) and ﬂow exponent of the ﬂoor (mean ± SD, n = 4 or 5) Parameter

Set nr. 1

Set nr. 2

Set nr. 3

Set nr. 4

Set nr. 5

Coefﬁcient Exponent

0.38 ± 0.00 0.88 ± 0.00

9.0 ± 0.3 0.54 ± 0.02

20.8 ± 0.6 0.49 ± 0.02

61.0 ± 0.8 0.50 ± 0.01

119 ± 1 0.48 ± 0.01

280

P. de Jong, W. van Dijk

Fig. 1. Correlation of ratios between living area and crawlspace radon concentrations and the air leak of the ﬂoor, given an air exchange rate of 0.5 L h−1 in the living area.

divergence between the theoretical values is just a few per cent. The ratio between the slopes of the 5 and√10 Pa regression lines is 0.69, which is close to the anticipated theoretical value of 0.71 (= 0.5). 3.2. Reducing radon inﬁltration Much attention was paid to obtaining an airtight connection between membrane and the walls of the crawlspace. As a result, the speciﬁc airﬂow rate was only (14.4 ± 0.4) × 10−6 m3 /(m2 s), with a ﬂow exponent of 0.56 ± 0.02, as determined on the basis of a pressure ﬂow rate characteristic. The results of an experiment carried out with the leak of the ﬂoor adjusted to 120 × 10−6 m3 /(m2 s) are presented in Table 2. The ratios between the living area and crawlspace radon concentrations, calculated from Fig. 1 for a speciﬁc leakage of the conTable 2 Radon concentration in the living area as a function of the pressure difference P between crawlspace section A and the living area, with a PE membrane ﬁtted beneath the concrete ﬂoor P (Pa)

0.0 5.1 9.6

Radon concentration (Bq m−3 , mean ± SEM)

Ratio∗

Supply air

Section A

Section B

Living area

3.8 ± 0.5 6.1 ± 0.6 3.5 ± 0.4

237 ± 3 234 ± 4 239 ± 4

158 ± 3 283 ± 4 276 ± 4

9.4 ± 0.5 33 ± 2 34 ± 2

0.009 0.10 0.11

∗ Ratio between the concentration in the living area and section A of the crawlspace. The concentration in the living area has been corrected for the contributions of the ﬂoor and the concentration of the supply air.

Testing of radon-reducing measures under strictly controlled conditions in a laboratory house

281

Table 3 Radon concentration (Bq m−3 ) in the living area as a function of the pressure difference between the crawlspace and the living area with an airlock ﬂoor P (Pa)†

Ratio∗

Radon concentration (mean ± SEM)

I

II

Supply air

Section A

Section B

Living area

0.0 4.2 9.8

−4.3 −5.2 −5.0

3.2 ± 0.5 8.6 ± 0.9 8.4 ± 0.6

50 ± 3 55 ± 2 55 ± 3

24 ± 2 37 ± 2 45 ± 3

5.5 ± 0.5 8.1 ± 0.6 9.4 ± 0.6

0.0 0.0 0.0

∗ As in Table 2. † P I: pressure difference between section A and living area; P II: between section B and living area.

Table 4 Effectiveness∗ (%) of the various measures involving the use of a PE membrane Radon-reducing measure

Membrane underneath the ﬂoor Same with fan (airlock ﬂoor) Membrane spread over the ground

Speciﬁc leakage in the unmodiﬁed situation (m3 /(m2 s)) 119 × 10−6

61 × 10−6

21 × 10−6

86 100 25

74 100 25

21 100 24

∗ Relative to an otherwise identical situation without the use of membrane (average for pressure differences across the ﬂoor of 5 and 10 Pa).

crete ﬂoor of 14 × 10−6 m3 /(m2 s), are 0.003, 0.080 and 0.12, respectively. Comparable ratios are presented in Table 2. Furthermore a small fan was ﬁtted in the membrane to establish a negative pressure difference across the ﬂoor of about 5 Pa, thereby creating a so-called airlock ﬂoor. This countermeasure was found to be highly effective. Table 3 shows the results of the experiments performed with the leakage of the ﬂoor adjusted to 120 × 10−6 m3 /(m2 s). The effectiveness of using a thin-ﬁlm membrane can then be assessed by comparing the two ratios (that for the situation with the membrane ﬁtted and that for the situation without the membrane) for each concrete ﬂoor air leakage ﬁgure. The results, which are shown in Table 4, prove to be dependent on the original air leakage of the ﬂoor. Comparison of the ﬁgures in Tables 2 and 3 reveals that, under the conditions used for this experiment, the concentration in section A of the crawlspace was considerably lower than it was without the fan. This observation is attributable to the high air leakage of the concrete ﬂoor in this experiment. To achieve an pressure difference across the ﬂoor of approximately −5 Pa across the ﬂoor, the fan needs to displace a considerable amount of air, thereby increasing the air exchange rate in the crawlspace. The airlock ﬂoor arrangement proved equally effective when the leakage of the concrete ﬂoor was changed (see Table 4). By covering the ground within the crawlspace with a membrane, an additional barrier to radon inﬁltration from the crawlspace was created. A summary of the results and a comparison of the various radon-reducing measures involving the use of the same membrane are presented in Table 4.

282

P. de Jong, W. van Dijk

Fig. 2. The inﬂuence of the crawlspace air change rate on the radon concentration in the living area at different levels of air leakage of the ﬂoor (P = 5 Pa).

3.3. Increasing crawlspace ventilation Increasing ventilation in the crawlspace can be expected to reduce radon inﬁltration of the living area. In Fig. 2, measured living area radon concentrations are plotted against the air exchange rate in the crawlspace. The slope of the tangent on the curves indicates the effectiveness of the measure.

4. Discussion Once the laboratory house had been set up, the performance and stability of the instrumentation were tested by conducting a number of preliminary experiments, some of which are described in Section 3.1. The method devised for simulating the air leakage of the ﬂoor proved practical and the results obtained are quite consistent with the ﬁgures suggested by transport modelling (within a few percent). The modest degree of error in the individual observations indicates that the test conditions are sufﬁciently stable. The repeatability also appears to be good. Given an air exchange rate of 0.5 L h−1 , the radon concentration in the crawlspace measured during nine different experiments spread over a period of a year, was approximately 2 per cent. The test facility was used to conduct a number of experiments to assess the reduction of radon transport from crawlspace to living area. The Dutch Building Decree [11] speciﬁes that the ground ﬂoor of a newly built house should have a speciﬁc air leakage coefﬁcient of 20 × 10−6 m3 /(m2 s) or less. Under normal weather conditions, this results in a ratio between the radon concentration in the living area and in the crawlspace of 10 to 15% (Fig. 1), which

Testing of radon-reducing measures under strictly controlled conditions in a laboratory house

283

is consistent with the averaged value obtained from a national survey involving 1500 houses [12]. Nevertheless, some houses, particularly older ones with timber ﬂoors, may require further remediation to reduce radon inﬁltration. Fitting a thin-ﬁlm plastic membrane beneath the concrete ﬂoor was found to reduce the air leakage considerably. As indicated in Table 4, the effectiveness of this measure largely depends on the original air leakage of the ﬂoor. When a small fan was additionally ﬁtted in the membrane, it was possible to achieve a living area radon concentration that did not differ signiﬁcantly from the concentration in the supply air (see Tables 3 and 4). This is consistent with the results reported by Phaff [13]. Laying a membrane over the ground within the crawlspace introduces a delay, during which some of the radon decays. As a result, less radon is released into the crawlspace, which in turn means that less radon will inﬁltrate into the living area. The effectiveness of this radon-reducing measure was calculated to be around 25 per cent, irrespective of the original air leakage of the concrete ﬂoor (Table 4). The effect of increasing crawlspace ventilation was also investigated. As indicated in Fig. 2, this measure can signiﬁcantly reduce the radon concentration in the living area. Its effectiveness depends to a considerable extent on the air leakage of the concrete ﬂoor and on the original level of the crawlspace ventilation. The most effective results are achieved in high ﬂoor leakage situations, in combination with a low crawlspace exchange rate. The test facility enables the effectiveness of various radon-reducing measures to be tested quickly, under precisely deﬁned and strictly controlled conditions. The repeatability of the experiments guarantees proper comparison of similar countermeasures, on the basis of which a selection can be made for ﬁeld-testing. The testing of countermeasures in a laboratory house thus constitutes an intermediate method, between model calculations on the one hand and testing under practical conditions on the other.

References [1] P. de Jong, W. van Dijk, Radiat. Prot. Dosim. 56 (1994) 179. [2] W.Z. Daoud, K.J. Renken, Proceedings of a Workshop on Radon in the Living Environment, Athens, Greece, April 19–23, 1999, Sci. Total Environ. 272 (1–3) (2001) 829. [3] C.E. Anderen, Proceedings of a Workshop on Radon in the Living Environment, Athens, Greece, April 19–23, 1999, Sci. Total Environ. 272 (1–3) (2001) 567. [4] S.P. Naismith, J.C.H. Miles, C.R. Scivyer, Health Phys. 75 (1998) 410. [5] G.A. Swedjemark, A. Mäkitalo, Health Phys. 58 (1990) 453. [6] J.A. Gunby, S.C. Darby, J.C.H. Miles, B.M.R. Green, D.R. Cox, Health Phys. 64 (1993) 2. [7] B. Lévesque, D. Gauvin, R.G. McGregor, R. Martel, S. Gingras, Health Phys. 72 (1997) 907. [8] Radon Policy Document, Spatial Planning and the Environment, Ministry of Housing, The Hague, The Netherlands, 1994. [9] W. van Dijk, P. de Jong, Health Phys. 61 (1991) 501. [10] P. de Jong, W. van Dijk, J.G.A. van Hulst, R.J.J. van Heijningen, Environ. Int. 22 (Suppl. 1) (1996) S287. [11] Dutch Building Decree, Staatsblad 680 (1991) 1. [12] P. Stoop, P. Glastra, Y. Hiemstra, L. de Vries, J. Lembrechts, Report 610058006, National Institute of Public Health and the Environment, Bilthoven, The Netherlands, 1998. [13] J.C. Phaff, in: Proceedings of 15th AIVC Conference, vol. 2, Buxton, UK, 27–30 September, 1994, p. 779.

284

Looping variation of correlation between radon progeny concentration and dose rate in outdoor air N. Fujinami, T. Watanabe, T. Tsutsui Kyoto Prefectural Institute of Hygienic and Environmental Sciences, 395-Murakami-cho, Fushimi-ku, Kyoto 612-8369, Japan

The absorbed dose rate in air of airborne gamma-rays and the concentration of radon progeny in outdoor air have been observed continuously in Maizuru, Japan. When data observed on ﬁne days were plotted, with dose rate as ordinate and concentration as abscissa, these points traced with a lapse of time illustrated an anticlockwise looping for each day. This result suggests that the variation of absorbed dose rate lags behind that of concentration of radon progeny; this is due to the delay time incurred as the concentration level gradually varies from ground surface to upper air. Also, internal and external effective dose rates from Rn-222 progeny outdoors derived from the observed data were 46 and 1.0 nSv h−1 , respectively. 1. Introduction It is widely known that the absorbed dose rate of airborne gamma-rays is directly proportionate to the concentration of radon progeny in air [1–6]. While some conversion factors for the concentration into the absorbed dose rate have been estimated by means of observation [3] or theoretical calculation [1–3], a time series variation of their correlation has not yet been thoroughly clariﬁed [5,6]. For this reason, the absorbed dose rate in air of airborne gammarays and the concentration of radon progeny in outdoor air have been observed every hour in Maizuru, Japan since 1994. The present paper explains results from these observations and discusses the mechanism that controls their variations. In addition, this paper deals with the comparison of ratios of internal dose to external dose from radon progeny between outdoors and indoors [7]. 2. Materials and methods The method for measuring the short-lived radon progeny concentration in outdoor air is as follows. At a height of 1.2 m above the ground, radon progeny aerosols were collected on a RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07031-7

© 2005 Elsevier Ltd. All rights reserved.

Looping variation of correlation between radon progeny concentration and dose rate in outdoor air

285

membrane ﬁlter at an airﬂow rate of 80 litre min−1 for ten minutes every hour. After sampling, the alpha spectrum from Po-218 and Po-214 on the ﬁlter was measured twice in a vacuum chamber with a silicone semiconductor detector. Each concentration of Po-218, Pb-214 and Bi-214 was determined from the alpha-ray peak area counts. The equilibrium equivalent radon concentration (EEC) was derived from the concentrations of the three nuclides by using the equation deﬁned in the UNSCEAR 2000 Report [8], and applying that value in the analysis of correlation with the absorbed dose rate. Absorbed dose rates in air have been continuously observed outdoors by means of a NaI(Tl) scintillation detector with a digital-signal-processing unit. The average dose rate during a tenminute sampling period of radon progeny was used in the analysis of the correlation. These observations have been carried out automatically at an environmental radiation monitoring station in Maizuru City, which is 60 km north west of Kyoto City, since 1994.

3. Results When data observed on ﬁne days were plotted, with the absorbed dose rate as ordinate and the concentration (EEC) as abscissa, these points traced with a lapse of time illustrated an anticlockwise looping for each day. Figure 1 shows the variation of their correlation observed during the period July 10–25, 1994, as an example. This ﬁgure reveals the typical looping variation based on the averages of dose rate and concentration for the entire time period, instead of individual looping variation for each day. Similar anti-clockwise looping variations not shown were observed during the periods April 8–16, July 18–25 1994 and October 16–25, 1997, when ﬁne weather continued respectively in spring, summer and autumn. In winter, data were not available, since ﬁne weather is rare in the observation area during this season. While this observed anticlockwise looping is consistent with the result of Tsujimoto et al. [6], it is not in agreement with that of Okabe et al. [5]. This inconsistency could be attributed to the fact that their observations of dose rate and concentration were not simultane-

Fig. 1. Correlation between radon progeny concentration and absorbed dose rate in outdoor air in Maizuru, Japan during the period July 10–25, 1994.

286

N. Fujinami et al.

ous. In addition, their result does not appear to be general, since their observation period was one month, and not sufﬁciently long. Assuming that the variation of absorbed dose rate lags by one hour behind that of radon progeny concentration, and their variations are expressed by Y = sin[π(t − 1)/12] and X = sin(πt/12) respectively, their correlation indicates an anticlockwise looping, as illustrated in Fig. 2a. On the contrary, assuming that the variation of dose rate precedes that of progeny concentration by one hour, their correlation indicates a clockwise looping illustrated in Fig. 2b. The anticlockwise looping, therefore, means that the variation of absorbed dose rate of airborne gamma-rays lags behind that of concentration of radon progeny in surface air. Fig. 3 reveals diurnal variations of radon progeny concentration and absorbed dose rate, which are normalized from data shown in Fig. 1 in such a way that both the averages are zero and both the standard deviations are one. It is clear also from this ﬁgure that the variation of dose rate lags behind that of progeny concentration. Fig. 4 shows the looping variation of correlation between normalized progeny concentration and normalized dose rate. This variation

Fig. 2. Assumed correlation between radon progeny concentration and absorbed dose rate. (a) The variation of dose rate lags behind that of progeny concentration by one hour. (b) The variation of dose rate is ahead of that of progeny concentration by one hour.

Fig. 3. Diurnal variation of normalized radon progeny concentration and normalized absorbed dose rate in air.

Looping variation of correlation between radon progeny concentration and dose rate in outdoor air

287

Fig. 4. Correlation between normalized radon progeny concentration and normalized absorbed dose rate in air.

Table 1 Conversion factors of radon progeny concentration in outdoor air into absorbed dose rate in air Author

Condition

Conversion factor (nGy h−1 )/(Bq m−3 )

Present work

Decreasing process Increasing process Strong mixing Weak mixing Inversion Uniform distribution Uniform distribution

0.39 0.23 0.47 0.36 0.17 0.40 0.41

Beck [2]

Hultzqvist [1] Kataoka et al. [3]

obtained from the observation is similar to the variation seen in Fig. 2a, which was based on the above assumption. Conversion factors of radon progeny concentration in outdoor air into absorbed dose rate in air were derived by regression analysis of the observed data in Fig. 1, which are given in Table 1, together with factors published by other researchers. Our value 0.23, which was observed in an increasing trend in the concentration at night, lies between Beck’s two values for weak mixing and inversion [2]. Furthermore, our value 0.39, which was observed in a decreasing phase in daytime, lies between his two values for strong mixing and weak mixing [2]. This value 0.39 is also similar to the values 0.41 and 0.40 observed and calculated by Kataoka et al. [3] and Hultqvist [1] for the uniform distribution of radon progeny. These results verify the validity of the obtained conversion factors. The obtained conversion factor is smaller for increasing than for decreasing processes of radon progeny concentration. This fact means that the downhill gradient of the progeny concentration with altitude is steeper for increasing than for decreasing processes, that is, the gradient is steeper for a stable than for an unstable atmosphere. This explanation is consis-

288

N. Fujinami et al.

Fig. 5. Simpliﬁed variation of radon progeny concentration in surface air and absorbed dose rate during a ﬁne day.

Fig. 6. Considered vertical proﬁles of radon progeny concentration for a ﬁne day.

tent with the view that the vertical distribution of radon and its progeny in the atmosphere is governed by turbulent mixing, which has been reported by Jacobi and Andre [9]. We notice a simpliﬁed looping variation of correlation between progeny concentration and dose rate, as shown in Fig. 5. This variation can be accounted for by a series of four vertical proﬁles of progeny concentration, which are illustrated in Fig. 6. That is to say, (1) (2) (3) (4)

after sunset, the radon progeny concentration begins to increase from near the ground, the increase of the concentration reaches the upper air, after sunrise, the concentration begins to decrease from near the ground, and the decrease of the concentration reaches the upper air.

4. Discussion M. Miller and A.C. George have calculated external gamma-ray dose rates from Rn-222 progeny indoors, and compared the effective dose equivalents for the internal versus the external pathway [7]. They concluded that the former was about three orders of magnitude greater than the latter for a given concentration of Rn-222 in indoor air under normal circumstances. For comparison, internal and external effective dose rates from Rn-222 progeny outdoors were derived from our observed data shown in Fig. 1, using the conversion coefﬁcients given in the UNSCEAR 2000 Report [8]. As a result, the average internal dose rate was 46 nSv h−1 , with a range of 16 to 81 nSv h−1 , and the average external dose rate was 1.0 nSv h−1 , with a range of 0.4 to 1.8 nSv h−1 . For Rn-222 progeny outdoors, therefore, the former is about 50 times the

Looping variation of correlation between radon progeny concentration and dose rate in outdoor air

289

latter, which indicates that there is a signiﬁcant difference between outdoors and indoors. This is mainly due to the inclusion of external effective dose from semi-inﬁnite Rn-222 progeny source in the outdoor data. For the same reason, our conversion factor of Rn-222 progeny concentration in air into absorbed dose rate is about 20 times that of Miller and George [7] for indoors, while the factor is consistent with those of Hultqvist [1] and Beck [2] for outdoors.

5. Conclusions The absorbed dose rate in air of airborne gamma-rays and the concentration of radon progeny in outdoor air have been observed continuously in Maizuru, Japan. When data observed on ﬁne days were plotted, with dose rate as ordinate and concentration as abscissa, these points traced with a lapse of time illustrated an anti-clockwise looping for each day. This result suggests the following: (1) The variation of absorbed dose rate of airborne gamma-rays lags behind that of concentration of radon progeny in surface air; this is due to the delay time incurred as the concentration level gradually varies from ground surface to upper air. (2) There is a difference in vertical distribution of atmospheric radon progeny between increasing and decreasing processes of the concentration. Internal and external effective dose rates from Rn-222 progeny outdoors were derived from the observed data. As a result, the average internal dose rate was 46 nSv h−1 , with a range of 16 to 81 nSv h−1 , and the average external dose rate was 1.0 nSv h−1 , with a range of 0.4 to 1.8 nSv h−1 . For Rn-222 progeny outdoors, the former is about 50 times the latter. Although the work presented here is only a case study in which the results were derived from local phenomena, we infer that these ﬁndings would be applicable to other places, judging from common views established in this ﬁeld. To elucidate quantitatively the mechanism of the anti-clockwise looping variation is the subject of future research.

References [1] B. Hultqvist, Studies on naturally occurring ionising radiation, with special reference to radiation doses in Swedish houses of various types, Kung. Svenska Vetenskap. Handl. (4) 6 (3) (1956). [2] H.L. Beck, J. Geophys. Res. 79 (1974) 2215–2221. [3] T. Kataoka, et al., J. Nucl. Sci. Technol. (1982) 831–836. [4] S. Minato, J. Nucl. Sci. Technol. 17 (1980) 461–469. [5] S. Okabe, et al., Nucl. Instrum. Methods Phys. Res. Sect. A 255 (1987) 371–373. [6] T. Tsujimoto, et al., Atmospheric Radon Families And Environmental Radioactivity III, Japan Atomic Energy Association, Tokyo, 1995 (in Japanese). [7] K.M. Miller, A.C. George, Health Physics 54 (1988) 203–206. [8] UNSCEAR (United Nations Scientiﬁc Committee on the Effects of Atomic Radiation), Sources and Effects of Ionising Radiation, UNSCEAR 2000 Report to the General Assembly, with Scientiﬁc Annexes, United Nations, New York, 2000. [9] W. Jacobi, K. Andre, J. Geophys. Res. 68 (1963) 3799–3814.

290

Comparative dosimetry in homes and mines: estimation of K-factors J.W. Marsh, A. Birchall, K. Davis National Radiological Protection Board, Chilton, Didcot, Oxon OX11 0RQ, UK

Comparative dosimetry has been carried out to estimate values of the so-called K-factors. The K-factor is deﬁned as the ratio of the equivalent dose to the lung per working level month (WLM) in homes to that in mines. The dosimetry was carried out using the ICRP Publication 66 Human Respiratory Tract Model (HRTM). Model parameter values that have been published and recommended by several prominent European scientists were used in the calculations. The estimated K-factor values for a typical home without smokers vary between 0.8 and 1.0 depending upon age group. The values for a home with smokers present are slightly lower ranging between 0.7 and 0.8 for different age groups.

1. Introduction K-factors are used to estimate the risk per unit exposure to radon progeny in homes from the observed risk to miners per unit exposure in mines on the assumption that the risk of excess lung cancer is directly proportional to the equivalent dose to the lung. Thus, the K-factor is deﬁned as the ratio of the equivalent dose to the lung per unit exposure in homes for a given population group to that for a miner exposed in mines. The working level month (WLM) is the unit of exposure that has been used in miner epidemiological studies as it was assumed to be proportional to the risk of excess lung cancer. Because the risk coefﬁcient for miners is usually given in terms of excess lung cancer deaths per working level month (WLM), the K-factor is conventionally deﬁned in terms of WLM: Kpopulation group =

Dosehome, population group /WLM . Dosemin er /WLM

(1)

Recently, Cavallo [1] pointed out that the BEIR VI Report [2] has (unusually) deﬁned Kfactor in terms of radon gas exposure (Bq m−3 h) and not in terms of WLM. Cavallo also points out that the two quantities termed “K-factor” are different because the equilibrium factor, a measure of the degree of radioactive equilibrium between radon gas and its progeny, RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07032-9

© 2005 Elsevier Ltd. All rights reserved.

Comparative dosimetry in homes and mines: estimation of K-factors

291

is generally different in homes from in mines. The values reported in BEIR VI indicate that the equilibrium factor in homes (mean 0.4) is greater than that in mines (mean 0.18). In this case the number of WLM corresponding to 1 Bq m−3 h is greater in homes than in mines. Thus, if the equivalent lung dose per Bq m−3 h is the same in mines and homes, as the BEIR VI report states, then the equivalent dose per WLM must be lower in homes than in mines. As a result the K-factor in terms of WLM estimated by Cavallo is 0.44 using a K-factor in terms of radon gas exposure of 1.0, as reported in BEIR VI. The BEIR VI Report gives K-factors for subjects of different ages in homes with and without smokers. The equivalent dose to the lung was calculated with the ICRP Publication 66 Human Respiratory Tract Model (HRTM) [3]. It is stated in the BEIR VI report that agedependence is accounted for in the deposition model. However, it is not stated in BEIR VI whether or not the variation in the mass of the target tissues with age is considered. This should be accounted for because equivalent dose to the target tissue is inversely proportional to the mass. The total mass of the target tissues in the bronchial and bronchiolar regions of the lung of a 1 y old is about a factor of 3 less than that in an adult [3]. This indicates that the lung dose for a 1-year-old will be a factor of 3 greater than that in the adult due to mass considerations only. However, this is largely compensated by lower breathing rates as discussed in Section 3. In this paper K-factors in terms of WLM have been calculated with the HRTM for subjects of different ages in typical houses with and without smokers. These values are compared with published values. 2. Calculation of dose and K-factors 2.1. Model used The software package RADEP (Radon Dose Evaluation Program) [4] has been developed to calculate the weighted committed equivalent dose to lung (wlung Hlung ) per unit exposure to radon progeny. Throughout the rest of this paper the weighted equivalent dose to lung, i.e. wlung Hlung is given a ‘short-hand’ symbol of Hw . RADEP implements the HRTM to calculate Hw per unit exposure to radon progeny (Hw /Pp ) in units of mSv per WLM. It is assumed that the absorption to blood can be represented with a single absorption half time of 10 h [5], and that the radon progeny are not bound to lung tissue. Typically Hw /Pp contributes more than 99% of the effective dose per WLM [4]. 2.2. Aerosol size The activity size distribution of the radon progeny can be represented by a sum of lognormal distributions. After the decay of radon gas, the freshly formed radionuclides react rapidly with trace gases and vapours, and grow by cluster formation to form particles with diameters around 1 nm. These are referred to as ‘unattached’ radon progeny in this paper. The unattached radon progeny may also attach to existing aerosol particles in the atmosphere forming the socalled attached particles. The attached particles can have a trimodal activity size distribution consisting of:

292

J.W. Marsh et al.

• the nucleation mode in which the activity median aerodynamic diameter (AMAD) or the activity median thermodynamic diameter (AMTD) is between 10 nm and 100 nm; • the accumulation mode in which the AMAD or AMTD is between 100 nm and 400 nm; and • the coarse mode in which the AMAD is between 1 μm and 4 μm. 2.3. Exposure in mines Best estimates of aerosol parameters for a typical mine, and for different exposure scenarios in a mine are given in Table 1. Values of the unattached fraction and the attached aerosol size in a ‘typical’ mine are the best estimates given by Birchall and James [6]. Aerosol values for a mine with diesel fumes and for a mine ‘without working actions’ (i.e. without diesel machinery and with good ventilation) are taken from those published by Porstendörfer and Reineking [7] based on the measurements of Butterweck et al. [8]. Generally, with diesel engines the diesel aerosol dominates the mine aerosol resulting in a low unattached fraction and an AMAD of 200 nm for the accumulation mode (Table 1). Diesel aerosols are hydrophobic [1,9]. For mines ‘without working actions’ the mine aerosol is similar to the outdoor aerosol having a higher unattached fraction and a larger AMAD for the accumulation mode [7, Table 1]. The Table 1 Typical aerosol parameter values for mine conditions and calculated values of Hw /Pp for the unattached and attached fraction obtained with the HRTM Exposure scenario

Modea

Aerosol fraction of PAECb

Parametersc AMTDd AMADe

(nm)

Hw /Pp h (mSv WLM−1 ) σg

hgff

Fg

Unatt.

Att.

Total

Typical mine [6]

u a

0.01 0.99

0.8d 250d

1.3 2.4

1 2

0.18i

0.7

11.8

12.5

Working + dieselj [7,8]

u a

0.006 0.994

0.8d 200e

1.3 2.0

1 1

0.5

0.4

8.6

9.0

Without working actions [7,8]

u a

0.02 0.98

0.8d 348e

1.3 2.3

1 2

0.35

1.5

12.0

13.5

Haulage drifts [10]

u a

0.03 0.97

0.8d 150d

1.3 2.2

1 2

0.18i

2.2

10.7

12.9

a Modes u and a represent the unattached and accumulation modes. b PAEC is the potential alpha energy concentration. c The particle density and the shape factor of unattached progeny are both assumed to be unity whereas for attached progeny the values are assumed to be 1.4 g cm−3 and 1.1, respectively. d Particle size given in terms of AMTD, activity median thermodynamic diameter. e Particle sizes given in terms of AMAD, the activity median aerodynamic diameter. f hgf is the hygroscopic growth factor. g F is the equilibrium factor. h The average breathing rate assumed is 1.2 m3 h−1 , the default value for a standard worker. Also the miner is

assumed to be a nose breather. i F value taken from the BEIR VI report [2]. j The density of the attached particles (exhaust particles from diesel engines) was taken to be 0.55 g cm−3 [11].

Comparative dosimetry in homes and mines: estimation of K-factors

293

unattached fraction and the attached aerosol size parameter values for haulage drifts in Table 1 are the values recommended by the panel of experts from the National Research Council (NRC) [10]. It is interesting to note that the equilibrium factor values (F ) measured by Butterweck et al. [8] in mines are greater than the average value given in the BEIR VI report [2] (Table 1). However, Hw /Pp is relatively insensitive to F [4]. Many of the mines in the epidemiology studies did not use diesel engines [1]. For example, most of the miners in the Colorado cohort and the Swedish cohort were not exposed to diesel fumes. Table 1 gives calculated values of Hw /Pp for each of the exposure scenarios. Hw /Pp for the ‘typical’ mine differs by less than 9% from Hw /Pp calculated for exposure scenario ‘without working actions’ and scenario ‘haulage drifts’. However, Hw /Pp for mine with diesel engines is 28% lower than that for a ‘typical’ mine. 2.4. Exposure in homes Best estimates of aerosol parameters are given in Table 2 for (1) a typical house without smokers, (2) a typical bedroom, and (3) for a living room with smokers. Several prominent European scientists recommended the aerosol values for a typical house based on their measurement data. These values have been collected and published by Marsh et al. [12]. The values for the bedroom are also based on these values. The values for a living Table 2 Typical aerosol parameter values for home conditions Exposure scenario

Modea

Aerosol fraction of PAECb

Parametersc AMTDd /AMAD (nm)

σe

Hygroscopic growth factor

Equilibrium factor

Typical home without smokers [12]

u n a c u n a u a

0.1 0.135 0.747 0.018 0.1 0.135 0.765 0.03 0.97

0.8 50 230 2500 0.8 50 230 0.8 348e

1.3 2 2.1 1.5 1.3 2 2.1 1.2 2

1 2 2 2 1 2 2 1 1.5 [13,14]

0.4

Bedroom [12]

Living room with smokers [7,10]

0.4

0.5

a Modes u, n, a and c represent the unattached, nucleation, accumulation and coarse modes, respectively. b PAEC is the potential alpha energy concentration. c The particle density and the shape factor of unattached progeny are both assumed to be unity whereas for attached progeny the values are assumed to be 1.4 g cm−3 and 1.1, respectively. d Unattached particle sizes are given in terms of AMTD whereas attached particles are given in terms of AMAD. e An AMTD of 300 nm was measured.

294

J.W. Marsh et al.

room with smokers are based on the values published by Porstendörfer and Reineking [7] and the unattached fraction value was recommended by the panel of experts from the NRC [10]. The hygroscopic growth factor value of 1.5 for radon progeny attached to cigarette smoke particles, given in Table 2, is based on the measurements of Li and Hopke [13] who measured an average value of 1.36, and Pagels et al. [14] who measured a value of 1.7. Table 3 gives the calculated values of Hw /Pp for subjects of different ages and gender exposed to radon progeny in the home. The following parameters were assumed to be subjectdependent: • The mass of target tissues. • The ventilation rates, the respiratory frequencies and tidal volumes for each level of exercise (sleeping, sitting awake, and light exercise). • The average time spent at each level of exercise. • The physiological parameters such as the functional residual capacity; the extrathoracic dead space; the bronchial dead space; the bronchiolar dead space; and the diameters of the trachea and airway generations 9 and 16. Default values for these parameters are given in ICRP Publication 66 [3] for subjects of different ages and gender. However, as in ICRP Publication 66, it is assumed here that the absorbed fractions (i.e. the fractions of energy deposited in target tissues per disintegration in source tissues) for short-range radiation are independent of age and gender. The values of Hw /Pp for subjects in houses with smokers (Table 3) were calculated assuming that the subjects were in a smoky atmosphere (living room with smokers) except while they were asleep in which case they were in the bedroom. ICRP Publication 66 gives default values for the fraction of time in the home spent asleep for each age group: 55% for adult male; 43% for adult female; 59% for 15-y-old male; 53% for 15-y-old female; 56% for 10-y-old child; 67% for 5 y-old child; 74% for 1-y-old child and 71% for 3-month-old child. Table 3 Calculated values of Hw /Pp obtained with the HRTM and calculated values of K-factors Subject typea

Adult male Adult female 15 y old male 15 y old female 10 y old child 5 y old child 1 y old infant 3 month infant

Typical home with smokersc

Typical home without smokers Hw /Pp (mSv WLM−1 ) Unatt. Att. Total

K-factorb

Hw /Pp (mSv WLM−1 ) Unatt. Att. Total

K-factorb

4.8 4.7 4.1 4.1 4.9 3.7 3.4 2.6

1.0 1.0 0.9 0.9 1.1 0.9 0.9 0.8

2.4 2.0 2.3 2.0 2.4 2.3 2.4 1.7

0.8 0.7 0.7 0.7 0.8 0.8 0.8 0.7

8.1 8.0 7.3 7.4 8.6 7.6 7.9 7.0

12.9 12.7 11.4 11.5 13.5 11.3 11.3 9.6

7.5 7.2 6.7 6.8 8.0 7.1 7.4 6.5

9.9 9.2 9.0 8.8 10.4 9.4 9.8 8.2

a Subjects are assumed to be nose breathers. b Calculated by dividing the H /P for homes with 12.5 mSv WLM−1 the value for a miner in a typical mine w p

(Table 1). c It was assumed the subjects were not in a smoky atmosphere while asleep.

Comparative dosimetry in homes and mines: estimation of K-factors

295

Table 3 also gives the calculated K-values for the different age groups obtained by dividing the dose per WLM in homes with the dose per WLM for a miner in a typical mine. Values are given for a typical home without smokers and a home with smokers.

3. Discussion The estimated K-factor values for a typical home without smokers vary between 0.8 and 1.0 depending upon age group. The mean value is 0.94 with a standard error of 0.03. The estimated K-factors for a home with smokers present are slightly lower with values ranging between 0.7 and 0.8 for different age groups. The mean value is 0.75 with a standard error of 0.02. The K-factor is lower in houses with smokers because the unattached fraction is lower and because there is no nucleation mode in the presence of cigarette smoke resulting in a lower dose [4]. Table 3 shows that Hw /Pp is relatively insensitive to age. The reason for this is that there are competing effects that tend to cancel out. For example, children have lower breathing rates so this decreases the intakes and Hw /Pp ; however, this is partly compensated by the smaller target tissue masses which increases the dose. Also children have smaller airways so this increases deposition by diffusion but this is also compensated in part by smaller residence times that decreases deposition by diffusion. Recently Butterweck et al. [15] carried out volunteer experiments to determine the absorption rate of unattached progeny from lung to blood. Assuming a single absorption rate the authors estimated absorption half times of about 65 minutes for unattached 214 Po/214 Pb and 23 minutes for unattached 214 Bi based on blood measurements. This is lower than the value of 10 hours, assumed in this paper, which is also based on volunteer experiments [5]. Assuming these ‘fast’ absorption values, the dose from the unattached fraction is reduced resulting in estimated K-values being reduced by about 10 to 15%; (0.7–0.9 for homes without smokers and 0.6–0.8 for homes with smokers). However, Butterweck et al. [13] noted from their in-vivo measurements that clearance from the head was slower than the physical radioactive half-live of radon progeny. They concluded that the unattached radon progeny are bound to respiratory tract tissue. They also conclude that rapid absorption must have occurred before the head measurements took place to account for the observed activity in the blood. Assuming the rapid fraction is absorbed with a half time of a minute the authors estimated that the fast fraction for unattached radon progeny is between 15 and 30%, and the bound fraction, fb is between 70 and 85%. Assuming the fast fraction for the unattached radon progeny is 20% and the remaining unattached fraction is bound to the lung (fb = 0.8) with a half-time of 10 h (sb = 1.7 d−1 ) then the dose from the unattached fraction increases resulting in K-values increasing by about 6% (0.8–1.1 for home without smokers and 0.7–0.9 for home with smokers). The dose is greater for bound activity as the activity is assumed to be physically closer to the target cells. The K-factors have been calculated with respect to Hw /Pp for a ‘typical’ mine (Tables 1 and 3). K-factors calculated with respect to a miner exposure scenario ‘without working actions’ or ‘haulage drifts’ yields K-factors that are 8% and 4% lower than that for a ‘typical’ mine. K-factors calculated with respect to a mine with diesel engines are about 39% higher

296

J.W. Marsh et al.

than that for a ‘typical’ mine. However, many of the miners in the epidemiology studies were not exposed to diesel fumes [1]. The value of the equilibrium factor (F ) assumed for a ‘typical’ mine (Table 1) is based on the measured values reported in Table B-5 of BEIR VI [2]. As these values are relatively low it is likely that these mines are well ventilated whereas in older mines with only natural ventilation F could be as high as 0.5. However, Hw /Pp is relatively insensitive to F [4]. Thus assuming F = 0.5 for a ‘typical’ mine results only in a negligible change in the calculated K-values (< 1%). Table 4 gives values of K-factors calculated from published values Hw /Pp . The panel of experts from NRC [10] estimated a K-factor value of 0.7, which is lower than the value calculated here (0.9 for house without smokers). This difference is mainly due to the difference in the assumed mean breathing rate for miners. The lower breathing rate value used in this paper is in close agreement with the reported mean value from the South African Chamber of Miners Research Organisation based on measurements of miners in Tajikistan and South Africa [2]. Porstendörfer and Reineking [7] have calculated values for Hw /Pp using an airway generation model developed by Zock et al. [16]. The K-values calculated from their published values of Hw /Pp are given in Table 4. The doses arising from the attached fraction calculated by Porstendörfer and Reineking are lower than the corresponding values predicted by the HRTM. This accounts for the higher K-value of 1.5, in Table 4, for home without smokers. The K-values (0.5, 0.6) given, in Table 4, for a home with active smoking can be considered as a lower limit for a house with smokers present. These values are low because the unattached fraction (fp = 0.005) is low for active smoking. Table 4 Values of K-factors calculated from published values Hw /Pp Reference

Exposure scenario

Breathing rate (m3 h−1 )

NRC [10]

Mine Home: without smoking

1.7a 0.74a

Mine: working + diesel Mine: without working actions Home: without smoking

1.7b 1.2a

9 6.7

0.75a

10.2

Home: active smoking

0.75a

4.2

Mine Home

Uncertainty analysis 17.2c Uncertainty analysis 15c

Porstendörfer and Reineking [7]

Birchall and James [6] Marsh et al. [12]

a The subject is assumed to be nose breather. b Working miner is assumed to breath through his mouth. c Mean value of the derived distribution of H /P . w p

Hw /Pp (mSv WLM−1 )

K-factor 0.7

1.1 (cf. working + diesel) 1.5 (cf. without working) 0.5 (cf. working + diesel) 0.6 (cf. without working) 0.9

Comparative dosimetry in homes and mines: estimation of K-factors

297

Marsh et al. [12] carried out a parameter uncertainty analysis to derive the probability distribution of Hw /Pp arising from radon progeny in the home. Birchall and James [6] also performed a parameter uncertainty analysis of Hw /Pp arising from radon progeny in the mine. Both analyses were performed using the HRTM. Taking the ratio of mean values of the distributions of Hw /Pp for home and mine yields a K-factor of 0.9 (Table 4). This value is in agreement with the value calculated in this paper for a typical home without smokers. In all the above calculations the equivalent dose to the lung of a miner arising from longlived alpha emitters, including 234 U, 230 Th, 226 Ra and 210 Po present in the suspended ore dusts in a mine atmosphere has been neglected. However, it is estimated that at higher concentrations of 222 Rn present in mines until 1967, the additional dose from these long-lived radionuclides is less than 10% of that from the short-lived radon progeny [10].

4. Conclusion K-factors in terms of WLM have been calculated with the HRTM for subjects of different ages in typical houses with and without smokers. The estimated K-factor values for a typical home without smokers vary between 0.8 and 1.0 depending upon age group. The values for a home with smokers present are slightly lower ranging between 0.7 and 0.8 for different age groups. The K-factors are therefore relatively insensitive to age group. The K-factors calculated here are broadly in line with other published values that range from 0.44 to 1.5.

Acknowledgements The authors thank Dr G. Butterweck, Dr M. Bohgard and Dr J. Pagels for advice on model parameter values. The authors also thank Miss F.M.G. Smith and Miss S. Thrift for writing a user-friendly interface for RADEP.

References [1] A. Cavallo, The radon equilibrium factor and comparative dosimetry in homes and mines, Radiat. Prot. Dosim. 92 (4) (2000) 295–298. [2] Committee on Health Risks of Exposure to Radon, Board on Radiation Effects Research, Commission on Life Sciences, National Research Council, Health Effects of Exposure to Radon: BEIR VI, National Academy Press, Washington, DC, 1999. [3] ICRP Publication 66: Human respiratory tract model for radiological protection, Ann. ICRP 24 (1–3) (1994). [4] J.W. Marsh, A. Birchall, Sensitivity analysis of the weighted equivalent lung dose per unit exposure from radon progeny, Radiat. Prot. Dosim. 87 (3) (2000) 167–178. [5] J.W. Marsh, A. Birchall, Determination of lung-to-blood absorption rates for lead and bismuth that are appropriate for radon progeny, Radiat. Prot. Dosim. 83 (4) (1999) 331–337. [6] A. Birchall, A.C. James, Uncertainty analysis of the effective dose per unit exposure from radon progeny and implications for ICRP risk-weighting factors, Radiat. Prot. Dosim. 53 (1–4) (1994) 133–140. [7] J. Porstendörfer, A. Reineking, Radon: characteristics in air and dose conversion factors, Health Phys. 76 (3) (1999) 300–305.

298

J.W. Marsh et al.

[8] G. Butterweck, J. Porstendörfer, A. Reineking, J. Kesten, Unattached fraction and the aerosol size distribution of the radon progeny in a natural cave and mine atmospheres, Radiat. Prot. Dosim. 45 (1–4) (1992) 167–170. [9] E. Weingartner, H. Burtscher, U. Baltensperger, Hygroscopic properties of carbon and diesel soot particles, Atmospheric Environ. 31 (1997) 2311–2327. [10] National Research Council, Comparative Dosimetry of Radon in Mines and Homes, National Academy Press, Washington, DC, 1991, ISBN 0-309-04484-7. [11] R.A. Zahoransky, E. Laile, B. Terwey, A. Konstandopouos, On-line/in-line measurements of particle emissions of diesel engines by optical multi-wavelength technique, Presented at the 4th Conference on Nanoparticle Measurement, ETH Zurich, August 7–9, 2000. [12] J.W. Marsh, A. Birchall, G. Butterweck, M.-D. Dorrian, C. Huet, X. Ortega, A. Reineking, G. Tymen, Ch. Schuler, A. Vargas, G. Vezzu, J. Wendt, Uncertainty analysis of the weighted equivalent lung dose per unit exposure to radon progeny in the home, Radiat. Prot. Dosim., in press. [13] W. Li, P.K. Hopke, Initial size distributions and hygroscopicity of indoor combustion aerosol particles, Aerosol Sci. Technol. 19 (1993) 305–316. [14] J. Pagels, E. Swietlicki, A. Gudmundsson, M. Bohgard, A set-up for ﬁeld studies of respiratory deposition of environmental particles in the size range 15 nm to 10 μm, Presented at the conference of International Society for Aerosols in Medicine (ISAM 01) in Interlaken, Switzerland, September 17–21, 2001. [15] G. Butterweck, Ch. Schuler, G. Vezzù, R. Müller, J.W. Marsh, S. Thrift, A. Birchall, Experimental determination of the absorption rate of unattached radon progeny from respiratory tract to blood, Radiat. Prot. Dosim., in press. [16] Ch. Zock, J. Porstendörfer, A. Reineking, The inﬂuence of biological and aerosol parameters of inhaled shortlived radon decay products on human lung dose, Radiat. Prot. Dosim. 63 (3) (1994) 133–140.

299

The principle of a passive, on/off, alpha track, closed radon detector J. Andru Société DOSIRAD, Villa Parc, Le Chêne, rue Lech Walesa, F-77185 Lognes, France

In measuring radon, there is no ON/OFF function for “closed” passive detectors using SSNT, “Solid State Nuclear Track detector”1 plastic. The results obtained during tests, such as the European Intercomparisons managed by NRPB, UK [1], are obtained when overprecautions, that are difﬁcult to routinely apply, are taken. These precautions allow each detector to exhibit its highest potential, but not necessarily its practical capacity in the ﬁeld. However, these overprecautions would be unnecessary for “closed” ON/OFF detectors, with a hermetic OFF position. This is already achieved with “open” detectors and can be fully transposed to a new family of “closed” detectors. An instant perfect OFF position can be easily achieved, using an adequate screen to protect an LR115 ﬁlm2 inside the closed case.

1. The source of the problem Passive “closed” detectors consist of a case with a small piece of SSNT plastic inside. In principle, before it can be used, the case is emptied of radon and sealed. It must be sealed in a radon-free atmosphere at the time of its production. When beginning a measurement, the seal must be broken to let the ambient radon to quickly enter the case. When the measurement is completed, the radon must be driven out before the case is perfectly sealed, ready for the SSNT to be processed with sodium hydroxide. In practice, this technique is certainly not easy to apply. Problems stem from the fact that the seal against radon is never perfect. Beginning from when the case is stored by the manufacturer or user, the SSNT plastic continuously records intrusive tracks. Even packaging in sealed packs does not totally resolve this problem. Socalled “closed” detectors are never actually so; they are never perfectly OFF. The air inside the case is never free of radon and the SSNT continuously gets tracks. 1 Referred to as SSNT in the remaining text. 2 The LR115 ﬁlm is a thin SSNT cellulose nitrate plastic.

RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07033-0

© 2005 Elsevier Ltd. All rights reserved.

300

J. Andru

In addition, measurement does not begin instantaneously. Several hours or days are required for the ambient radon to enter the case. When the measurement is completed, the user cannot “remove” the radon from the case. The case is sealed and immediately sent for analysis, hopefully avoiding postal strikes. Very often, in order not to risk trapping a high concentration of radon, it is better not to seal the case at all. Regardless of detector sensitivity, this radon retention prohibits short exposures. The relative measurement error then becomes too high. In certain tests, such as “European intercomparison of passive radon detectors”, extreme overprecautions are taken in order to achieve the highest “closed” detector performances. Examples of overprecautions: the use of sealed double radon proof bags, detector degassing after radon chamber output, measurement duplication by control detectors (for background noise correction), immediate dispatching, etc. For their part, detector suppliers know that the detectors are under test. They also take overprecautions: recently produced detectors, special attention to the whole process (etching, counting, calibration, etc.), urgency of analysis, etc. All the above inconveniences would be avoided if “closed” detectors functioned ON/OFF with a truly hermetic OFF position, and they would be as good in routine as in tests, even without overprecaution.

2. The principle of a “closed” ON/OFF detector To solve the “closed” detection problems, it is sufﬁcient to take inspiration from “open” detection that has an ideal, transposable ON/OFF function. The most important thing is to achieve a perfect and instantaneous “OFF” position. This can easily be obtained with a good protective screen applied to an LR115 ﬁlm. The best screen in current use is a polyester ﬁlm, possibly metallised, at least 120 μm thick. While a case is never sealed perfectly, an LR115 ﬁlm protective screen can be almost perfect, even if there is no contact between the ﬁlm and its screen. Removing the protective screen from LR115 ﬁlm is enough to set the case to position ON. Numerous ways of carrying out this manoeuvre can be devised. Some methods are yet to be invented, some are already available. Moreover, the sensitivity can be improved (amount of tracks per kBq h m−3 of exposure). It only requires replacing the protective screen by a thin polyester ﬁlm (≈ 20 to 35 μm, according to the size and shape of the case), in front of the LR115 ﬁlm, when the ON position is set. This thin ﬁlm acts as a decelerator screen. It reduces the energy of the alpha particles, emitted too close from the LR115 ﬁlm surface, and then allows more tracks to be recorded. To be certain that such ON/OFF “closed” detectors are reliable, it must be proved ﬁrst that the LR115 ﬁlm is properly adapted to the problem and that the thick protective polyester screen of a KODALPHA detector is perfect. This is the subject of the following text.

3. The qualities of the LR115 ﬁlm The European Intercomparisons seem to show that the LR115 is not necessarily good for making ON/OFF closed detectors. Firstly, because the participants who present “closed” detectors based on LR115 do not always obtain good results. Secondly, “open” detector results

The principle of a passive, on/off, alpha track, closed radon detector

301

Fig. 1. KODALPHA radon detector.

Fig. 2. Last ﬁve KODALPHA intercomparison results.

are classiﬁed according to the same criteria as for “closed” detectors, while, by design, they certainly supply offset results for low and high equilibrium factors (“F ”)3 (Fig. 2). The European Intercomparison reports do not prove that the LR115 ﬁlm, used in “open” detectors, can work as well without overprecautions, since these overprecautions are applied to the LR115 as well as to all the other detectors. To show that LR115 ﬁlm, correctly used, is perfectly adapted to the problem in question, the results obtained by the KODALPHA “open” detector (Fig. 1), from the last ﬁve European intercomparisons are described in Table 1 and in Fig. 2. In Table 1, % diff. is the difference expressed as a percentage between the average value found and the value to be found. 3 Equilibrium factor is referred to as F in the following text.

302

J. Andru Table 1 Last ﬁve KODALPHA intercomparison results (F – equilibrium factor)

F % diff. F % diff. F % diff.

1995

1997

1998

1999

2000

0.90 +29 0.40 −7 0.16 −16

0.80 +18 0.40 0 0.22 −23

0.84 +21 0.46 1 0.18 −19

0.87 +32 0.43 +2 0.22 −19

0.84 +40 0.38 1 0.22 −9

In Fig. 2, the concentrations of radon have been “standardized”. Concentration 1.00 is the concentration to be found, i.e. effective concentration within the radon chamber. We note that all the results are very good for F = 0.40. This is logical since these detectors are precisely calibrated for F = 0.40. Being always correct, the measurement dispersion is not represented in order that the graph is not complicated. LR115, correctly used, may thus be an excellent SSNT. Few detectors have such consistent results. Occasionally, we also note that the experimental sensitivity to F of a good “open” detector is a lot weaker than the theoretical sensitivity used until now. The relative dispersion of experimental points to high F is in part due to the fact that the activity and energy of daughters are not linked by a strictly bi-univocal relationship. The same total energy may be obtained with daughter combinations whose total activities are considerably different. As a reminder, the WL (Working Level, radon daughter potential energy measure) relative growth axis is described in Fig. 2. For F = 0.40, a non-standardized radon concentration of 1000 Bq m−3 corresponds to 0.10 WL. As we know that the WL is used to set limits that should not be exceeded in mines and radon gas activity (horizontal axis) is used for setting public limits, we can only ask the question, “Which is the best health risk indicator, WL or radon gas activity, or could it be somewhere between the two?” The only evidence is that for F = 0.00, the gas is not the best indicator, since the risk is therefore close to zero. 4. LR115 ﬁlm with screen protection In the European intercomparisons, no accelerated test for storing detectors has been carried out. This would entail leaving “closed” detectors in the radon chamber in their original packaging. This could, for example, allow the efﬁciency of adhesives used to seal certain “closed” detectors or sealed plastic protective packs to be checked. These tests would also conﬁrm the astounding protection of LR115 ﬁlm afforded by a simple thick polyester screen. Remember that this protection is due to the total absence of LR115 sensitivity to alpha energies higher than approximately 4.2 MeV. (This threshold depends on ﬁlm processing conditions). This allows, for example, the long-term storage of a stack of small ﬁlms (special format as seen in Fig. 1), by simple ﬁlm–ﬁlm self-shielding: no need for storing ﬁlms and detectors in a nitrogen atmosphere, etc. In our tests in radon chambers, a sealed KODALPHA detector recorded less than 2 tracks cm−2 per year in a 50 Bq m−3 atmosphere, i.e. “average” air. The 120 μm thick metallised polyester screen protection is therefore excellent. To a small extent, this ﬁgure depends

The principle of a passive, on/off, alpha track, closed radon detector

303

on ﬁlm processing and counting. It is even more surprising because in a sealed KODALPHA case, there is no contact between the LR115 ﬁlm and its screen. The distance between them is approximately 0.2 mm. All other tested materials gave poorer results. Most of the plastics “dissolve” a small amount of radon and most of the metal foils are weak emitters (e.g. aluminium used for chocolate wrapping, etc.). However, our tests are not exhaustive.

5. ON/OFF “closed” detectors It takes imagination to manufacture such detectors. DOSIRAD itself could certainly use ON/OFF “closed” detectors, but it is convinced in the use of “open” detection for a better overall estimate of health risk. However, this is another debate. 5.1. Converting “open” detectors to “closed” detectors For example, from the KODALPHA detector, we can construct the two detectors as seen in Fig. 3. For both, the OFF position is the sealed KODALPHA case itself. Conversely, in position ON, we apply an unsealed “radon lens” to the ﬁlm for the ﬁrst and we place the complete unsealed KODALPHA into a calibrated beaker for the second. 5.2. Using or gaining ideas from existing cases Certain “electret” (“e-perm” cases) have an OFF position. A cover masks the Teﬂon pellet to prevent existing ions from discharging it. All that is required is replacing the Teﬂon pellet by an LR115 ﬁlm and the cover by a thick polyester ﬁlm screen as seen in Fig. 4. To our knowledge, the only ON/OFF “closed” detector available on the market is marketed under the name DPR1 by the Algade company in France [2]. Fig. 4 shows the interior of this detector. We can see the rotary cover that masks the LR115 ﬁlm in the OFF position. This cover is operated by an external two-way ON/OFF button.

Fig. 3. “Closed” KODALPHA uses.

304

J. Andru

Fig. 4. Basing design on existing cases.

Fig. 5. Example of design of a new type of closed, ON/OFF detector.

5.3. Designing new models We can devise a great number of small ON/OFF cases. The drawing in Fig. 5 describes the principle of a detector that is traversed by a narrow strip, which is pulled or pushed to turn it ON or OFF. Only the words ON or OFF are visible on the exterior of the detector. In position ON, sensitivity is maximal due to the interposition of a thin polyester ﬁlm (∼ 20 μm). 6. Conclusion In spite of their simplicity, passive detectors have not yet achieved their optimum. It is possible to manufacture “closed” detectors that can be stored for 1 year without a signiﬁcant background noise increase: detectors that can be interrupted for several days between measurement periods; detectors that have no need for double radon proof bags, transit detectors

The principle of a passive, on/off, alpha track, closed radon detector

305

controls, degassing before dispatch and are not subject to transport strikes. . . But, the discussion is open.

References [1] European Commission, Directorate-General for Research, Unit D.II.3 – R & T speciﬁc programme “Nuclear ﬁssion safety 1994–98”, Brussels, Belgium – Tel. 32-2 29-51589. E-mail [email protected]. [2] ALGADE, BP46, F-87250 Bessines-sur-Gartempe, France – Tel. 33 5 55 60 50 00.

306

Intercomparison exercise of calibration facilities for radon gas activity concentration A. Röttger a , A. Honig a , G. Butterweck b , Ch. Schuler b , V. Schmidt c , H. Buchröder c , A. Rox d , J.C.H. Miles e , I. Burian f , N. Michielsen g , V. Voisin g , F.-J. Maringer h , A. Vargas i a Physikalisch-Technische Bundesanstalt (PTB), Bundesallee 100, D-38116 Braunschweig, Germany b Paul Scherrer Institut (PSI), CH-5232 Villigen PSI, Switzerland c Bundesamt für Strahlenschutz (BfS), Köpenicker Allee 120-130, D-10318 Berlin, Germany d Materialprüfungsamt Nordrhein-Westfalen (MPA), Marsbruchstraße 186, D-44287 Dortmund, Germany e National Radiological Protection Board (NRPB), Chilton, Didcot, Oxfordshire, OX11 0RQ, United Kingdom f Státní ústav jaderné (SUJCHBO), chemické a biologické ochrany Príbram-Kamenná, 262 31 Milín, Czech Republic g Institut de Radioprotection et de Sûreté Nucléaire (IRSN), CEA Saclay, Bât. 389, BP68,

F-91191 Gif-sur-Yvette cedex, France h ARC Seibersdorf Research GmbH (ARCS), Faradaygasse 3, Arsenal Obj. 214, A-1030 Vienna, Austria i Universitat Politècnica de Catalunya (INTE-UPC), Institut de Tèchniques Energètiques, Avda. Diagonal 647,

E-08028 Barcelona, Spain

International intercalibration and intercomparison exercises are used as an important tool to provide conﬁdence in the capability of national metrology institutes and calibration laboratories. The quality systems of national institutes provide the basis for running intercomparisons, in our case, in the ﬁeld of the radon activity concentration in air. Interchange of an electronic radon measuring instrument (intercomparison device) which demonstrated ruggedness during shipping in the past, was considered a convenient means for a relative comparison of radon reference atmospheres. 9 institutes from 7 countries calibrated the intercomparison device according to their quality system at three different activity concentrations: 1 kBq m−3 , 3 kBq m−3 and 10 kBq m−3 . The results agreed within 5%, mainly. This is the ﬁrst step in establishing this intercomparison of the radon activity concentration within the scope of Euromet and reaching the quality of a BIPM (Bureau International des Poids et Mesures) key comparison later. 1. Introduction Is this measurement result correct? You should repeat it if you are not sure. To avoid such duplication of work is the objective of an international agreement which was signed by 38 of RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07034-2

© 2005 Elsevier Ltd. All rights reserved.

Intercomparison exercise of calibration facilities for radon gas activity concentration

307

the 48 member countries of the Meter Convention and two international organisations during the 21st General Conference of the Metre Convention (CGPM) in Paris on October 14, 1999: the Mutual Recognition Arrangement (MRA) [1]. The MRA – the ‘Arrangement of mutual recognition of equivalence of national standards and of calibration certiﬁcates issued by national metrology institutes’ – is not a diplomatic treaty but a technical agreement. It deﬁnes the conditions under which the national metrology institutes mutually recognise the calibration of other institutes. One central element of the arrangement is the new regulations for worldwide comparison measurements, so-called key comparisons, to obtain information about the degree of equivalence of national measurement standards and calibration procedures. The key comparisons are also fundamental in building conﬁdence. Up to now, for radon activity concentration, neither a key comparison nor any other form of comparison has been performed within the scope of Euromet. The comparison presented here is the ﬁrst step towards fulﬁlling the requirements of the MRA for calibration certiﬁcates for the radon activity concentration. The basic concept of the intercomparison therefore is not to achieve uncertainties as small as possible by using special calibration procedures, but to use standard calibration procedures according to the quality systems of each calibration institute for the intercomparison device (named ID1 in the following).

2. Participants All participants in the intercomparison are operating a quality system. This includes the existence of a traceability chain to national or international calibration standards as well as the statement of complete results (value with assigned uncertainty). To provide conﬁdence in the results, it was agreed that all data from intercomparison measurements should be available to every participant, that the calibration procedures according to the respective quality control systems have to be communicated on demand and that the presentation of the results is always done in joint authorship. Calibration by a primary standard is based on the speciﬁcation of the activity concentration by means of a radon gas activity standard and a noble gas-tight vessel of known volume. The system under test and the radon gas activity standard are enclosed in the vessel. The activity concentration chosen for the point of calibration is calculated and compared to the reading of the system under test for the calculation of the calibration factor. After calibration, the former system under test can now be used as a secondary standard. A calibration by a secondary standard is based on a comparison of the system under test to a reference instrument, which was calibrated in or traceable to a reference atmosphere in the past. The radon activity concentration chosen for the point of calibration is established, and both systems are exposed to it. The reading of the reference instrument and the system under test are observed simultaneously for the calculation of the calibration factor. The transfer of the quantity activity concentration (created by a radon gas activity standard and a noble gastight vessel of known volume) can also be performed by the use of Lucas cells instead of enclosing a system under test in the vessel. 1 Radon monitor, type AlphaGuard, SN 1145.

308

A. Röttger et al.

The participants in this intercomparison (Table 1) were informally selected, on the basis of institutional and personal relations stemming from the ﬁrst radon intercomparison exercise in 2000 [9]. This was not intended to exclude any other potential participants, but it was rather a result of organisational requirements. The underlying idea was to have the number of

Table 1 Traceability and quality assurance of the participants of the radon intercomparison exercise 2001–2002 Institute

Traceability to a reference atmosphere cA = A · V −1

to a reference instrument cA

Quality assurance principle

ARCS, Vienna, Austria [2]

ATMOS12 DXP calibration certiﬁcate from PTB

ISO/IEC 17025, accreditation of the Austrian governmental certiﬁcation authority

BfS, Berlin, Germany [3]

AlphaGuard SN.812 calibration certiﬁcate from PTB

ISO/IEC 17025, accreditation by Deutscher Kalibrierdienst (DKD)

INTEUPC, Barcelona, Spain, following tech. [4]

A: radon gas activity standard with calibration certiﬁcate from PTB V : glass container, V traceable to mass, calibration certiﬁcate from Mettler-Toledo

ISO/IEC 17025 is followed according to the policy of the institute

IRSN, Saclay, France [5]

A: radon gas activity standard with calibration certiﬁcate from BNM-LNHB V : stainless steel container measured with a measuring tape traceable to BNM

Management quality assurance of the company (CEA-IPSN) and standard NF X 07-010

MPA, Dortmund, Germany

NRPB, Chilton, United Kingdom

A: radon gas activity standard with calibration certiﬁcate from NPL V : calibration tank measured with a tape with a calibration certiﬁcate from Birmingham City Council

PSI, Villigen, Switzerland [6]

A: radon gas activity standard with calibration certiﬁcate from PTB V : stainless steel container with calibration from METAS – no certiﬁcate for formal reasons

ATMOS12 DXP, SN.10468.01 calibration certiﬁcate from PTB

ISO/IEC 17025 and DIN, EN 45001, accreditation of Deutsches Akkreditierungs system Prufwesen GmbH

ATMOS12 SN. ATM 185

ISO/IEC 17025, self-declaration option

ISO/IEC 17025, accreditation by Bundes amt fur Metrologie und Akkreditierung (METAS)

(continued on next page)

Intercomparison exercise of calibration facilities for radon gas activity concentration

309

Table 1 (continued) Institute

Traceability

Quality assurance principle

to a reference atmosphere cA = A · V −1

to a reference instrument cA

PTB, Braunschweig, Germany [7]

A: radon gas activity standard with calibration certiﬁcate from PTB V : stainless steel container with calibration certiﬁcate from PTB

A1phaGuard SN.934 calibration certiﬁcate from PTB

ISO/IEC 17025, self-declaration option

SUJCHBO, Milin, Czech Republic [8]

A: radon gas activity standard with calibration certiﬁcate from NPL V : traceability not established

Lucas cells calibrated by NRPB and EPA

ISO/IEC 17025, accreditation by Czech Accreditation Institute

The quantity cA represents the radon activity concentration, A the radon gas activity and V the volume

participants grow slowly (4 institutes in the ﬁrst, 9 institutes in the second intercomparison) but continuously to gather more experience. New participants are welcome in the next intercomparison 2002–2004 within the registered Euromet project 657: ‘Intercomparison exercise of radon activity concentration calibration facilities’.

3. Organisation The organization of the intercomparison was split into two parts: the shipping and testing of the ID and the management of the data obtained in the calibration processes. The ﬁrst part was performed by PTB, the second one by PSI. The ID was always returned to the PTB after the calibration process was ﬁnished at an institute, and before being sent to the next one. During this stopover, the ID’s basic functions were tested, the background (due to build-up of 210 Po activity in the ionisation chamber) was measured and the completeness of the equipment (including the collection of raw data ﬁles) was checked. The progress of the intercomparison was documented by intercomparison newsletters via e-mail. The complete history of the ID exposure to natural or artiﬁcial atmospheres (activity concentration, temperature, pressure, humidity) is documented by its data ﬁles as well as in the accompanying lab book. This offers the possibility of correlating the total exposure of the ID to its background reading and of observing the amount of automatic background correction performed twice a year. The complete exposure history is given in ﬁgure block with the results. After a calibration was ﬁnished, the complete results (results of the measurements together with their assigned uncertainties) were reported to PSI, where a web page displaying the collected results was accessible to the participants. At the end of the intercomparison, a meeting for the discussion of the results, further actions and the organisation of the next intercomparison took place. As a result of this meeting, the documentation of results will be extended by

310

A. Röttger et al.

requesting the submission of the uncertainty budget according to the ‘Guide for the expression of uncertainty in measurement’ [10] for each point of calibration in the future. This will result in a deeper understanding of the calibration procedure of each participant and lead to a solid basis for mutual conﬁdence in the certiﬁcates of each participating laboratory. Nevertheless, the intercomparison in itself is not a method of establishing traceability, but is rather a tool for testing applied quality assurance.

4. Results The results of the intercomparison were reported in the form of either the calibration factor kt or the sensitivity mt as a function of reference activity concentration cr . The input quantities or this calculation are the reference activity concentration cr , the activity concentration ct observed by the ID, and its background reading activity concentration (due to contamination) ct,bg : kt (cr ) =

cr 1 = . mt (cr ) ct − ct,bg

(1)

The combined standard uncertainty for the calibration factor u(kt ) is, therefore, given by 2 2 cr 1 · u(cr ) + · u(ct )2 + u(ct,bg )2 . u(kt ) = (2) ct − ct,bg (ct − ct,bg )2 The collected results are given in Fig. 1. The value kt = mt = 1 shows the agreement of the factory calibration with the calibration by the respective participant at the different activity concentration. It is noted that this set of results is highly correlated. With the exception of the IRSN and the SUJCHBO, all participants have a traceability link to the PTB, either in the form of activity A or as activity concentration cA . A signiﬁcant deviation outside the expanded uncertainty (95% conﬁdence level) is observed only for the lower activity concentration calibrations at SUJCHBO. Other trends might become signiﬁcant during the next few years, but up to now neither the deviation of the results from each other nor the potential inﬂuence quantities like temperature, humidity and pressure have shown a signiﬁcant effect. Additionally, in the range of activity concentrations used (1 to 10 kBq m−3 ), no non-linearity or instability of the ID with time (over 3 years) was observed.

5. Conclusion and outlook The result showed agreement for the calibration of activity concentration in the range of 1 to 3 kBq m−3 of about 10% and in the range of 10 kBq m−3 of about 5%. Improvements by implementing harmonized treatment of quality assurance including traceability and the assignment of uncertainties have already been discussed. Within the scope of the registered Euromet project 657: ‘Intercomparison exercise of radon activity concentration calibration facilities’, the third intercomparison will be conducted from 2002 to 2004. All participants in the last intercomparison volunteered to continue and further participants will be welcome.

Intercomparison exercise of calibration facilities for radon gas activity concentration

Fig. 1. Calibration factor as a function of activity concentration. (The line drawn at kt = 1 has no metrological meaning, but is just a line to guide the eye. The uncertainties assigned to the values represent a conﬁdence level of 95%. The right top ﬁgure gives the background reading of the ID in comparison to the monthly integral (from 0 to 1 month) and overall integral activity concentration. The automatic background correction for this instrument is for the ﬁrst time observed in July 2001 and January 2002.)

311

312

A. Röttger et al.

The calibration points will once again be 1, 3 and 10 kBq m−3 and the information to be submitted by each participant is: 1. Quality assurance and traceability information; 2. Result of the calibration: determined measurand (e.g. calibration factor) and its combined uncertainty; 3. Uncertainty budget for each point of calibration. The intercomparison thus meets the needs of two general developments: (1) the harmonization of metrology within the scope of the MRA – the ‘Arrangement of mutual recognition of equivalence of national standards and of calibration certiﬁcates issued by national metrology institutes’ and (2) the increased demands of the EURATOM directive, transferred into national radiation protection regulations with regard to natural radioactivity and its qualityassured measurements. Nevertheless, the radon activity concentration is only one of the relevant parameters in the well-known radon problem. A comparable intercomparison dealing with the measurement of the short-lived radon progeny, either each individual activity concentration or the combined quantities equilibrium factor F and unattached fraction fp , seems not to be feasible at present: too few calibration institutes are capable of calibrating these quantities in metrological terms (including traceability and quality assurance), and a suitable intercomparison device (with established uncertainty and stability) has yet to be developed. This may be a task for the future, since the so-called radon problem is, strictly speaking, a radon progeny problem. Furthermore, this intercomparison was restricted to the activity concentration of 222 Rn. The inﬂuence of 220 Rn, which is always present in ﬁeld measurements, was excluded due to the objective of dealing with pure calibration conditions. Practical problems of measuring technique led to a restriction to the quantity of the radon activity concentration in the case of most radiation protection measurements. This leads to higher uncertainties in exposure and dose calculations, as well as in the risk estimations of epidemiology. Improvements in metrology and the extension of calibration capabilities worldwide to radon progeny measurement will therefore be of great practical importance.

Acknowledgements Organisation, running and evaluation of the data of the intercomparison was based on the personal engagement of several persons, normally under a very short schedule. The authors would like to thank the following members of the institutes named: J. Kirchmeier (ARCS), G. Hoffman (ARCS), B. Henke (MPA), S. Weinberg (MPA), C. Howarth (NRPB), G.M. Kendall (NRPB), E. Gargioni (PTB), T. Reich (PTB), R. Dersch (PTB), S. Röttger (PTB).

References [1] Arrangement on the Mutual Recognition of Equivalence of National Standards and of Calibration Certiﬁcates Issued by National Metrology Institutes (MRA), 21st Conference of the Metre Convention (CGPM), BIPM, 1999.

Intercomparison exercise of calibration facilities for radon gas activity concentration

313

[2] F.-J. Maringer, et al., in: Proc. IRPA Regional Symposium on Radiation Protection in Neighbouring Countries of Central Europe, Tech. Univ. Prague, 1997, pp. 150–153. [3] P. Hamel, V. Schmidt, Kerntechnik 66 (4) (2001) 202–206. [4] R. Falk, H. Möre, L. Nyblom, J. Res. Natl. Inst. Stand. Technol. 95 (1990) 115. [5] J.L. Picolo, D. Pressyanov, P. Blanchis, N. Michielsen, D. Grassin, V. Voisin, K. Turek, in: International Conference on Radionuclide Metrology and its Applications, ICRM’99, Prague, 1999. [6] Ch. Schuler, PSI Report Nr. 98-08, Paul Scherrer Institut, Villigen, 1998, ISSN 1019-0643. [7] A. Paul, A. Honig, S. Röttger, U. Keyser, Appl. Radiat. Isot. 52 (3–4) (2000) 369–375. [8] I. Burian, Metrologie 3 (10) (2001) 11. [9] G. Butterweck, Ch. Schuler, A. Paul, A. Honig, R. Dersch, V. Schmidt, P. Hamel, H. Buchröder, A. Rox, W. Herzog, Radiat. Prot. Dosim. 98 (2) (2002) 219–222. [10] ISO, ICS 17.020, 1995, ISBN 92-67-10188-9.

314

In-vivo measurement of deposition and absorption of unattached radon progeny G. Butterweck a , Ch. Schuler a , G. Vezzù a , J.W. Marsh b , A. Birchall b a Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerland b National Radiological Protection Board, Chilton, Didcot, Oxfordshire, OX11 0RQ, United Kingdom

Twenty-one volunteers were exposed in a radon chamber during well-controlled aerosol and radon progeny conditions, with predominantly unattached radon daughters (218 Po, 214 Pb and 214 Bi). The activity of these radionuclides deposited in the respiratory tract was measured invivo after the exposures. The results of these measurements are in agreement with predictions calculated with the ICRP Publication 66 Human Respiratory Tract Model (HRTM). Temporal analysis of the activity deposited in the head of the volunteers leads to the conclusion that a signiﬁcant amount of the deposited activity associated with particle diameters of about 1 nm is not subject to fast particle transport to the gastrointestinal tract as generally reported for larger aerosol particles. Measurements of radon progeny in blood samples of these volunteers yielded absorption half-times of 20 to 60 min. Former determinations, mainly performed with much larger aerosol particles of diameters between 100 and 1000 nm, derived absorption half-times around 10 h. This indicates that the absorption from ciliated airways into blood is dependent upon particle size and particle composition.

1. Introduction Exposure to airborne radon progeny in the domestic environment yields the largest source of exposure to ionising radiation of the general public. The radiation protection principle of overestimating the risk rather than underestimating it has limitations when applied to sources of natural radiation. As the exposure to radon progeny concerns the total population of a country, a consensus has to be found between the justiﬁed wish of the people for protection against health risks and the ﬁnancial expenditure and administrative effort necessary for protective actions. The choice of an action level for radon exposure plays an important part in the decision for remedial actions against elevated radon concentrations. Scientiﬁc research has the assignment to supply the fundamental knowledge for this choice. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07035-4

© 2005 Elsevier Ltd. All rights reserved.

In-vivo measurement of deposition and absorption of unattached radon progeny

315

The large number of persons exposed to airborne radon progeny yields the opportunity to quantify the risk with epidemiological studies. An alternative approach is to use the ICRP Publication 66 Human Respiratory Tract Model [1] (HRTM) to calculate doses to tissues and infer lung cancer risk. Employing the HRTM, a fatal lung cancer risk of 8.4 × 10−4 for radon progeny exposure of 12.97 J s m−3 (1 WLM) was predicted, whereas epidemiological studies derived a lower lung cancer risk of 2.8 × 10−4 for an identical exposure [2–4]. It was decided by ICRP not to recommend the dosimetric model for this purpose, but to base risk assessment for the inhalation of short-lived radon progeny on the epidemiological approach, because this is the most direct route from exposure to risk [3]. A unique property of radon progeny is the existence of both aerosol bound radionuclides of 100–400 nm diameter and ultra ﬁne clusters in the diameter range of 1–4 nm, called the “unattached fraction” of radon progeny [5]. The biokinetic behaviour of aerosol particles has been well studied, whereas experimental studies on the biokinetic behaviour of unattached radon progeny are scarce. The aim of this study was to carry out volunteer experiments to determine deposition of unattached radon progeny in the human respiratory tract and its absorption rate into blood. Further detail of this study is given by Butterweck et al. [6,7].

2. Material and methods After ethical approval by the “Überregionale ethische Kommission für klinische Forschung der Schweizerischen Akademie der medizinischen Wissenschaften”, 21 sitting volunteers at rest were exposed in the PSI walk-in radon chamber (10 m3 ) [8] for 30 minutes to an atmosphere enriched with radon progeny. 2.1. Exposure conditions The deposition of radon progeny in the respiratory tract is governed by the airborne activity concentrations of radon progeny, the size distribution of the inhaled particles and the amount and type of breathing. The experiments were designed with consequent regard to the ALARA principle. The detection limits of the instrumentation provided minimum exposure conditions which still yielded sufﬁciently measurable activities in air, in the respiratory tract and in blood samples of the volunteers. Airborne unattached radon progeny has a diffusion constant around 0.07 cm2 s−1 and thus signiﬁcant amounts of activity can be deposited on surfaces due to the random movement of the particles. To reduce the inﬂuence of the activity located on the surface of the volunteer on the in-vivo measurement, the volunteers were clothed in a dust protection overall during the exposure, which also covered their hair and feet. The hands of the volunteers were covered with gloves. The protective clothing was removed prior to the in-vivo measurement. With this simple method, the activity measured on the surface of the volunteer was reduced to about 10% of the original value [9]. The exposures were divided in ﬁve groups (Table 1).

316

G. Butterweck et al. Table 1 Different exposure conditions applied in the experiments at PSI. Number denotes the number of volunteers exposed to the speciﬁc exposure conditions Group No.

Type of exposure

1 2 3 4 5

Number

Standard conditions, mouth breathing Standard conditions, nose breathing Radon gas only Candle aerosol, mouth breathing Doubled concentrations, mouth breathing

7 7 3 2 2

2.2. Airborne activity concentrations The radon gas activity concentration was measured using a ﬂow-through ionisation chamber. Activity concentrations of radon progeny were measured by ﬁltering air through membrane ﬁlters and by alpha-spectrometric counting of the activity sampled on the ﬁlters during and after air sampling. Standard exposure conditions are characterized by a radon gas concentration of about 19 kBq m−3 with a large unattached fraction (fp ) of radon progeny of 0.8. In exposures for the validation of the results derived under standard conditions, two volunteers were exposed to an atmosphere with candle aerosol having a small fp of 0.08 (Group No. 4). Two further volunteers were exposed under standard aerosol conditions, but at elevated radon gas concentration (Group No. 5) (Table 2). Table 2 Average activity concentrations of airborne radionuclides, equivalent equilibrium concentrations Ceq (activity concentration of radon, in equilibrium with its short-lived progeny which would have the same potential alpha energy concentration as the existing non-equilibrium mixture), equilibrium factor F (ratio of equilibrium equivalent concentration and the radon gas concentration) and unattached fraction fp (fraction of the potential alpha energy concentration of the short-lived progeny that is not attached to the ambient aerosol) for the different exposure groups Average activity concentrations [Bq m−3 ] 3 4

Group No.

1, 2

222 Rn

18 600

18 700

18 800

49 100

3400 130 20 430

3200 130 20 410

1800 700 500 740

8000 290 30 1000

Unattached 218 Po 214 Pb 214 Bi Ceq Attached 218 Po 214 Pb 214 Bi Ceq F fp

210 80 110 105 0.03 0.80

210 90 110 110 0.03 0.79

12 700 8000 8100 8531 0.49 0.08

5

560 240 320 310 0.03 0.77

In-vivo measurement of deposition and absorption of unattached radon progeny

317

2.3. Size distributions of airborne radon progeny The size distribution of the unattached radon progeny was measured using tube diffusion batteries [10,11]. A total of 8 parallel sampling lines with 50%-cut-off diameters between 0.3 nm and 3 nm were used. Size distributions of the attached radon progeny were determined with a low-pressure cascade impactor equipped for on-line alpha-spectrometric measurement of the activity deposited on the impaction plates [12]. The size distribution of the short-lived radon progeny was assumed to consist of one or a sum of two lognormal distributions. The parameters of the lognormal modes were determined using a Monte Carlo-type iteration process [13]. For the exposures with a large unattached fraction (Groups Nos. 1, 2, 5), 218 Po was found associated with monodisperse particles of 0.9 nm diameter, whereas 214 Pb was found to have a bimodal size distribution with 22% of the unattached activity having a median diameter of 0.3 nm with a geometric standard deviation of 1.2 and 78% of the unattached activity having a median diameter of 1.1 nm with a geometric standard deviation of 1.3. In the case of candle aerosol, both exposures showed quite different size distributions, due to sooty particles in one of the experiments. Both experiments showed half the particles at median diameters around 260 nm with a geometric standard deviation of 2. The exposure of the ﬁrst volunteer with sooty particles showed a second mode at 120 nm with a geometric standard deviation of 1.4, whereas the exposure of the second volunteer yielded a much larger second mode at a median diameter of 430 nm with a geometric standard deviation of 1.2. 2.4. Inhaled amount of air during the exposure The amount of air breathed by the volunteer during the exposure was determined by measuring the tidal volume of the volunteer with a respirometer at the medical entrance examination. The number of breaths was measured during the exposure via the expansion of the chest of the volunteer. The volunteers had an average breathing rate of 0.8 m3 h−1 (range 0.4–1.4 m3 h−1 ). 2.5. Activity deposited in the respiratory tract The in-vivo counter used in this study was the whole-body counter at PSI [14], extended with two additional CsI (phoswich) detectors of 15.4 cm diameter and 5.1 cm thickness employed for the measurement of activity located in the head of the volunteers (Fig. 1). The in-vivo counter is located in a low-background steel shielding with a wall thickness of 18 cm. The activity in the chest of the volunteers was measured with a coaxial HPGe detector with 100% relative efﬁciency. The signal generated in the HPGe detector by radon progeny deposited in the head of the volunteer was reduced below 15% of the original value using a 2 cm thick lead shield positioned between the head of the volunteer and the HPGe detector. The counting efﬁciency for the activity in the head and chest of the volunteer was determined individually. Anatomical measurements were performed for each volunteer yielding the necessary data for an efﬁciency calculation using a semiempirical method [9,15,16]. The activity in the chest was assumed, for the efﬁciency calculations, to be located in the trachea and the ﬁrst large bronchi. The short-lived radon decay products 214 Pb and 214 Bi emit gamma photons with various photon energies in two distinct energy bands with a separation energy of about 500 keV. For

318

G. Butterweck et al.

Fig. 1. Schematic diagram of the in-vivo counter.

the small count rates due to the low radon progeny concentrations during the exposures, all emissions in the respective energy ranges were counted. The in-vivo activity measurement consisted of 6 counting intervals of ﬁve minutes and was started 7 minutes after the end of exposure. Because the dose to the volunteers had to be kept to a minimum, the detector signals associated with the deposited radon decay products 214 Pb and 214 Bi were only around 5 to 10 percent of the total signal. About 40% of the signal was due to the background of the invivo counter, 30% resulted from natural radionuclides in the body, 10% originated in radon progeny produced from the decay of radon gas in the body and 10% were caused from the remaining surface contamination of the volunteer [9]. Two of these sources, radon gas and surface contamination, could not be separated from the signal originating from radon progeny deposited in the respiratory tract with a single measurement. This problem was solved with an additional background exposure with identical atmospheric conditions, during which the inhalation of unattached radon progeny was prohibited with a glass ﬁbre dust mask. Possible contamination of the volunteer in the domestic environment was investigated with a third in-vivo measurement prior to the exposure. The time schedule of exposures and respective in-vivo measurements is summarised in Table 3. Table 3 Time schedule of a volunteer exposure in the PSI radon chamber relative to the start of the main exposure Relative start time of action

Action

−3 h −40 min 0 37 min 210 min 247 min

arrival at PSI in-vivo measurement 30 min main exposure 30 min in-vivo measurement 30 min background exposure 30 min in-vivo measurement 30 min

In-vivo measurement of deposition and absorption of unattached radon progeny

319

2.6. Activity in a blood sample A blood sample of 200 g was taken inside of the PSI radon chamber at the end of the exposure. Special care was taken to avoid surface contamination of the blood sample. The sample was then sealed in a vacuum tight stainless steel vessel to prevent loss of radon gas from the sample and measured for about ﬁve weeks using a 18 × 18 cm low-level NaI well detector with well dimensions of 6 × 12 cm. The components of the detector signal arise from: • absorbed radon progeny, • radon progeny produced in the blood sample by absorbed radon gas, and • background radiation. It was possible to differentiate between these components by exploiting the different radioactive half-lives of the radionuclides. However, as no radioactive equilibrium between radon gas and radon progeny can be assumed during the ﬁrst three hours of counting, the amount of radon progeny produced in the blood sample by absorbed radon gas was determined with three additional volunteer exposures (Group 3). For these exposures, the inhalation of particulate radon progeny was inhibited using a glass ﬁbre dust mask. 3. Results 3.1. Activity deposited in the respiratory tract The activity of the radon progeny in head and chest of the volunteers was predicted with the HRTM. A computer program that implements the HRTM was written speciﬁcally to do these calculations. This program uses the deposition module of LUDEP [17]. Experimentally determined input parameters for the calculations were the individual air activity concentrations of 218 Po, 214 Pb and 214 Bi, separated into attached and unattached activity, and the activity size distribution for each nuclide. Due to the low ambient aerosol concentrations in the radon chamber during the exposures, the attached activity yielded less than 5% of the total activity deposited in the respiratory tract. The activity found in the head and chest of all volunteers agreed with the HRTM within the estimated uncertainties (Figs. 2 and 3). The values obtained from the model calculations represent a best estimate of input parameters; the error bars included in the ﬁgures are based on an uncertainty analysis of the input parameters [2]. The error bars for the measured activities represent the 68% conﬁdence interval [9]. There seems to be a trend of a prediction of larger deposited activities by the HRTM compared to the results of the measurements. 3.2. Clearance of activity deposited in the extrathoracic airways The HRTM considers the extrathoracic airways (ET) to consist of the anterior nose (ET1 ) and posterior nasal passages, larynx, pharynx and mouth (ET2 ). The activity in the head of the volunteers who only breathed through the mouth decreases after an exposure due to: • the radioactive decay of the short-lived radon progeny, • the particle transport of activity from ET2 into the gastrointestinal tract, and • the absorption into blood.

320

G. Butterweck et al.

Fig. 2. Comparison between measured and predicted activity deposited in the head and chest of 7 mouth breathing volunteers. Activity normalised to average breathing rate and air concentrations.

Fig. 3. Comparison between measured and predicted activity deposited in the head and chest of 7 nose breathing volunteers. Activity normalised to average breathing rate and air concentrations.

The HRTM assumes that the half-time for particle transport of aerosol particles from ET2 is about 10 minutes [1]. Thus, it was expected that the activity measured with the head detectors for the volunteers who only breathed through the mouth would decrease with an effective half-time of less than 7 minutes. The particle transport (clearance) from the anterior nose (ET1 ) by nose blowing is difﬁcult to assess. To obtain a minimum value of the 214 Pb effective decay constant for nose breathers, no particle transport from ET1 was assumed. It was also assumed that 50% of the deposited activity in the head was located in the ET1 . Under these assumptions, an effective half-time for 214 Pb activity in the head of nose breathing volunteers was estimated to be about 13 minutes. Surprisingly, after deposition experiments with largely unattached radon progeny the measurement of 214 Pb activity neither in the head of mouth breathing volunteers nor of nose breathing volunteers showed any signiﬁcant deviation of the effective half-time from the radioactive half-life of 214 Pb (Fig. 4).

In-vivo measurement of deposition and absorption of unattached radon progeny

321

Fig. 4. Measured effective half-time of the decrease of 214 Pb activity of deposited unattached radon progeny in the head of volunteers. Volunteers P1–P7 were mouth breathers, volunteers P8–P14 were nose breathers.

3.3. Determination of the absorption half-time of unattached radon progeny from the respiratory tract into blood The absorption half-times were determined for the average blood activities of the seven mouth-breathing volunteers (Group 1) and the seven nose-breathing volunteers (Group 2). The speciﬁc activity of radon progeny measured in the blood sample was compared to model predictions using a computer code developed at NRPB based on the HRTM [1] (Table 4). It was assumed that the absorption from lung to blood could be represented by a single rate constant, sr , and that the short-lived radon progeny are not bound to the lung. Thus in HRTM terminology: fr = 1 (fraction dissolved rapidly), fb = 0 (fraction to bound state), and ss = 0 (slow dissolution rate). Absorption half-times of radon progeny from respiratory tract to blood were iterated manually until predicted activity and measured activity in the blood sample were in agreement. The measured activities of 214 Pb and 214 Bi are not sensitive to the absorption half-time of 218 Po. Thus, this value could not be determined from the measurements of the blood activity of 214 Pb and 214 Bi. The absorption half-time of 218 Po was therefore assumed to Table 4 Speciﬁc activity of radon and radon progeny found in a 200 ml blood sample taken 30 min after start of the exposure and corresponding absorption half-times Speciﬁc activity in blood sample [mBq g−1 ] Mouth breathing (Group 1) Nose breathing (Group 2) 222 Rn 214 Pb 214 Bi 218 Po/214 Pb 214 Bi

5.2 ± 0.2 5.4 ± 0.2 3.1 ± 1.0 1.9 ± 0.4 2.5 ± 0.6 2.2 ± 0.3 Absorption half-times [min] 63 (45–95) 68 (56–86) 29 (23–39) 18 (17–21)

322

G. Butterweck et al.

be equal to the corresponding value for 214 Pb. The range of values given in brackets in Table 4 was derived from the uncertainty of the average measured blood activity value.

4. Validation of results Four exposures were performed after the evaluation of the absorption half-times to test for hidden systematic errors. Two mouth-breathing volunteers were exposed under identical conditions, with the exception of radon and radon progeny activity air concentrations, which were doubled for these exposures (Group 5). Both volunteers had a low breathing rate of 0.5 m3 h−1 , which compensated partly for the increased air concentration of radon progeny. Another two mouth breathing volunteers were exposed to an atmosphere with largely attached radon progeny (Group 4). The results of the measurements of radon progeny activity in blood samples were compared to model predictions using the absorption half-time of unattached radon progeny determined for the average mouth breathing volunteer (Table 4). For the absorption of aerosol attached radon progeny into blood, a half-time of 600 min was used. Radon gas concentrations in blood were predicted for Groups 4 and 5 as follows: Firstly, the ratio of radon gas concentration in blood to airborne radon gas activity concentrations was determined from the measurements of Group 3 volunteers who inhaled radon gas alone. Secondly, the measured airborne radon gas activity concentration for the required group is multiplied by this ratio to obtain the radon gas concentration in blood. With the exception of 214 Bi for volunteer 2 of aerosol exposure Group 4, the predicted blood activity is in reasonable agreement to the measured values (Table 5). This volunteer showed in the obligatory medical examination a signiﬁcantly reduced lung function, which may be the source of the large deviation between predicted and measured activities.

5. Discussion The interpretation of the experiments presented here is based upon the assumption that the absorption rate from lungs to blood can be adequately represented by a single rate constant. As well as analysing the data with a single absorption rate the possibility was considered that part of the unattached radon progeny was absorbed rapidly while the rest (fraction fb ) was bound to respiratory tract tissues, from which it is absorbed at a rate sb . Table 5 Comparison between predicted (Pred.) and measured (Meas.) speciﬁc blood activity of the four test exposures Speciﬁc activity in blood sample [mBq g−1 ] 222 Rn

214 Pb

214 Bi

Exposure

Pred.

Meas.

Pred.

Meas.

Pred.

Meas.

Group No. 4, volunteer 1 Group No. 4, volunteer 2 Group No. 5, volunteer 1 Group No. 5, volunteer 2

5.1 5.2 13.0 13.9

5.5 5.6 13.7 12.2

3.7 3.6 4.3 4.7

3.1 5.1 5.0 5.5

6.4 5.0 3.4 3.9

5.1 1.2 4.1 3.9

In-vivo measurement of deposition and absorption of unattached radon progeny

323

5.1. Single absorption rate Assuming a single absorption rate, the evaluation of the measured blood activities yields a ten times faster absorption of unattached radon progeny deposited in the human respiratory tract than the absorption of radon progeny attached to aerosol particles. It must be emphasised that the results of the Group 4 exposures, in which 92% of the radon progeny is attached, are consistent with the attached radon progeny being absorbed with a half-time of 10 hours. In houses, typically about 10% of the radon progeny are unattached. Butterweck et al. [7] estimated that effect on the equivalent dose to the lung per WLM due to assuming the ‘faster’ absorption rates for the unattached progeny determined here is to reduce the equivalent dose by 15% for an exposure in a typical home. There was no signiﬁcant difference in the estimated absorption half-times between nose and mouth breathing volunteers. This indicates that, as assumed by the HRTM, there is no signiﬁcant absorption from deposited particles in the anterior nose (ET1 ) into blood. 5.2. Fast and slow absorption rates Results from in-vivo measurements of the deposited unattached radon progeny (Section 3.2) show that the clearance from the head is much slower than the physical half-life of radon progeny. This indicates the possibility that a large fraction of the unattached radon progeny is bound to the extra thoracic tissue. However, to account for the observed blood sample measurements, absorption must have taken place before the head measurements began. Thus, the head and blood data indicate that a small fraction of the unattached radon progeny is absorbed rapidly to blood and a large fraction is bound to tissue. This is consistent with the ﬁndings of Smith et al. [18] of a rapid absorption of plutonium oxide particles with diameters around 1 nm instilled in the lung of rats. Furthermore, Greenhalgh et al. [19] reported a rapid absorption into blood and the probability of a fraction of 212 Pb ions being retained in the epithelial tissue following instillation of 212 Pb ions into noses of rats. They estimated that about 20% of the 212 Pb ions entering the sol layer is absorbed rapidly to blood and the remaining 80% has the possibility of being bound to tissue and cleared to blood with a half-time of about 10 h. Assuming there is a fast clearance fraction, the results from the in-vivo head measurements indicate that the fast clearance half-time is much less than 10 minutes, otherwise clearance from the head would have been observed. Assuming that it is of the order of a minute, the blood measurements indicate that the fast fraction for unattached radon progeny is between 15 and 30%, and the bound fraction, fb , is between 70 and 85%. Assuming 80% is bound then the equivalent dose to the lung per WLM increases by about 7% for a typical home [7]. The dose is greater for bound activity as the activity is assumed to be physically closer to the target cells. The above results can be compared with the work of Booker et al. [20] who estimated a 10 h clearance half-time of 212 Pb from lung measurements of a volunteer who inhaled unattached 212 Pb. In view of the experimental results presented here, further work is required to see whether the data of Booker et al. [20] are consistent with an absorption model consisting of a fast and slow fraction, and a bound state for unattached radon progeny.

324

G. Butterweck et al.

6. Conclusions No signiﬁcant difference was observed between the measured activities of deposited unattached radon progeny in the human respiratory tract and the predictions with the ICRP Publication 66 Human Respiratory Tract Model. The expected fast particle transport from ET2 (posterior nasal passages, larynx, pharynx and mouth) to the GI tract was not observed. One possible explanation is that a signiﬁcant fraction of the deposited radionuclides with diameters around 1 nm is bound to the extra thoracic tissue. The present results together with literature data on the absorption of aerosol attached radon progeny show that the unattached fraction is absorbed at a faster rate into blood than the attached radon progeny. This results in the equivalent dose to the lung for a typical home being reduced by about 15%. However, if it is assumed that the unattached fraction is bound, as the head measurements indicate, then the dose increases by about 7%. The dependence of absorption half-time on particle size in the intermediate size range between 1 and 100 nm diameter is unknown and should be investigated in the future.

Acknowledgements The European Union under contract No. FI4P-CT95-0025 and the Swiss Federal Ofﬁce for Science and Education under contract No. 95 03 31 have funded this work.

References [1] ICRP Publication 65: Protection against radon-222 at home and at work, Ann. ICRP 23 (2) (1993). [2] A. Birchall, A.C. James, Uncertainty analysis of the effective dose per unit exposure from radon progeny and implications for ICRP risk-weighting factors, Radiat. Prot. Dosim. 53 (1994) 133–140. [3] ICRP Publication 65: Protection against radon-222 at home and at work, Ann. ICRP 23 (2) (1993). [4] J.H. Lubin, J.D. Boice, C. Edling, R.W. Hornung, G. Howe, E. Kunz, R.A. Kusiak, H.I. Morrison, E.P. Radford, J.M. Samet, M. Tirmarche, A. Woodward, S.X. Yao, D.A. Pierce, Lung cancer and radon: A joint analysis of 11 underground miners studies, Publication No. 94-3644, US National Institute of Health, Bethesda, MD, 1994. [5] J. Porstendörfer, Properties and behaviour of radon and thoron and their decay products in the air, J. Aerosol Sci. 25 (1994) 219–263. [6] G. Butterweck, G. Vezzù , Ch. Schuler, R. Müller, J.W. Marsh, S. Thrift, A. Birchall, In-vivo measurement of unattached radon progeny deposited in the human respiratory tract, Radiat. Prot. Dosim. 94 (3) (2001) 247–250. [7] G. Butterweck, Ch. Schuler, G. Vezzù, R. Müller, J.W. Marsh, S. Thrift, A. Birchall, Experimental determination of the absorption rate of unattached radon progeny from respiratory tract to blood, Radiat. Prot. Dosim., submitted for publication. [8] Ch. Schuler, Das Referenzlabor für Radongas-Konzentrationsmessungen am PSI. PSI Ber. 98-08, Paul Scherrer Institut, Villigen, 1998. [9] G. Vezzù, In-vivo Messung der Aktivität von im menschlichen Atemtrakt deponierten Radonzerfallsprodukten, ETH-Dissertation No. 13179, Eidgenössische Technische Hochschule Zürich, Zürich, 1999. [10] S.C. Soderholm, Analysis of diffusion battery data, J. Aerosol Sci. 10 (1979) 163–175. [11] E.O. Knutson, History of diffusion batteries in aerosol measurements, Aerosol Sci. Tech. 31 (1999) 83–128. [12] J. Kesten, G. Butterweck, J. Porstendörfer, A. Reineking, H.J. Heymel, An online α-impactor for short-lived radon daughters, Aerosol Sci. Technol. 18 (1993) 156–164.

In-vivo measurement of deposition and absorption of unattached radon progeny

325

[13] G. Butterweck-Dempewolf, Ch. Schuler, G. Vezzù, A. Reineking, Improved determination of bimodal size distributions from measurements with diffusional size classiﬁcation, Aerosol Sci. Technol. 31 (5) (1999) 383– 391. [14] M. Boschung, The high purity germanium detector whole-body monitor at PSI, Radiat. Prot. Dosim. 79 (1998) 481–484. [15] J. Moens, J. De Donder, F.X.L. Lin, F. De Corte, A. De Wispelaere, A. Simonitis, J. Hoste, Calculation of the absolute peak efﬁciency of γ-ray detectors for different counting geometries, Nucl. Instrum. Methods 187 (1981) 451–472. [16] J. Moens, J. Hoste, Calculation of the peak efﬁciency of high-purity germanium detectors, Int. J. Appl. Radiat. Isot. 34 (1983) 1085–1095. [17] N.S. Jarvis, A. Birchall, A.C. James, M.R. Bailey, M.-D. Dorrian, LUDEP2.0. Personal computer program for calculating internal doses using the ICRP Publication 66 respiratory tract model, NRPB-SR287, NRPB, Chilton, 1996. [18] H. Smith, G.N. Stradling, B.W. Loveless, G. Ham, The in-vivo solubility of plutonium-239 dioxide in the rat lung, Health Phys. 33 (1977) 539–551. [19] J.R. Greenhalgh, A. Birchall, A.C. James, H. Smith, A. Hodgson, Differential retention of 212 Pb ions and insoluble particles in nasal mucosa of the rat, Phys. Med. Biol. 27 (6) (1982) 837–851. [20] D.V. Booker, A.C. Chamberlain, D. Newton, A.N.B. Stott, Uptake of radioactive lead following inhalation and injection, Br. J. Radiol. 42 (1969) 457–466.

326

Assessing radiation dose for tour guides in Australian show caves from radon monitor measurements S.B. Solomon, J. Peggie, R. Langroo Australian Radiation Protection and Nuclear Safety Agency, Lower Plenty Road, Yallambie, Melbourne, Victoria 3085, Australia

This paper describes an extensive study to assess size-dependent radon progeny dose conversion factors for selected Australian show caves, and presents the results of the analysis for the 6 cave systems with radon levels exceeding current radon action levels. The measured size-weighted dose conversion factors for all caves in this study were found to be signiﬁcantly greater than the radon conversion convention recommended by International Commission for Radiological Protection in its Publication 65. The measurement results also show that a dose conversion factor based on radon exposure of 5 μSv h−1 per kBq m−3 , provides a more consistent estimate of effective dose than a dose estimates based on radon progeny exposure for the range of aerosol conditions in these caves.

1. Introduction In a survey of radon levels in Australian show caves, carried out in 1995, it was shown that 14 of a total of the 67 caves investigated had yearly average radon concentrations exceeding the action level of 1000 Bq m−3 [1]. Most of these problem caves were in southeastern Australia. While the radon levels in these caves represent a negligible risk to members of the public during a short caves tour, the same cannot be said for exposure to the tour guides who conduct these tours. The radon levels in these caves cannot be reduced without major modiﬁcation to the cave conditions and potential damage to the cave decoration. If the radon levels in these problem caves cannot be reduced, then Australian guidelines recommend that the employers should implement radiation protection measures, including the monitoring of worker exposures [2,3]. Personal monitoring using passive radon monitors can provide a simple and cost effective method for assessing individual exposure to radon progeny, provided appropriate conversion factors can be determined. The derivation of dose estimates from these radon exposures requires information about the ventilation conditions in each cave, in order to relate radon levels RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07036-6

© 2005 Elsevier Ltd. All rights reserved.

Assessing radiation dose for tour guides in Australian show caves from radon monitor measurements

327

to radon progeny levels. Further, based on current respiratory tract dosimetric models, the low aerosol concentrations and high values of radon progeny unattached fractions in these caves would lead to a signiﬁcant increase to the radon progeny dose estimates [4]. This paper describes a systematic study of radon and radon progeny levels and relevant aerosol parameters, at a number of representative sites in each of the Australian show caves with radon levels exceeding the action level. These measurements have been analysed to assess appropriate factors to convert from radon and radon progeny exposure to effective dose.

2. Inhalation dose assessment The health risk associated with radon arises from the inhalation of the short-lived decay products, i.e. radon progeny. The radiation dose delivered to the respiratory system, and the resultant health detriment, is a complex function of the radon progeny aerosol characteristics and the physiological parameters related to the deposition of radon progeny in the respiratory tract of the exposed individual. The magnitude of this risk has been quantiﬁed in a number of epidemiological studies, many dealing with the increased rate of lung cancer amongst uranium miners [5]. The International Commission for Radiological Protection (ICRP) in its Publication 65 recommends the use of a single factor, the conversion convention, determined from these epidemiological studies, as the preferred method converting radon progeny exposure to effective dose. While dosimetric modelling of inhaled radon progeny has demonstrated the effect of aerosol and physiological parameters in modifying the dose conversion factors, the ICRP argued that use of the conversion convention value of 1425 mSv per J h m−3 was appropriate for most occupational exposure situations. The Human Respiratory Tract Model described in ICRP Publication 66 [6] provides an alternative method for assessing radon progeny inhalation dose that takes account of aerosol conditions. The effective dose per unit intake of 222 Rn progeny as a function of particle size, were derived using the computer program RADEP, which implements the ICRP66 Human Respiratory Tract Model for exposure to short-lived radon progeny [7]. In general, the dosimetric models yield larger values for the radon progeny DCF than the epidemiological approach, and for this study the RADEP-derived values have been normalised using a factor of 0.3 to provide consistency with the epidemiologically-derived risk estimates in ICRP65 [5]. The results from the RADEP program indicate that the size dependence of the dose per unit exposure. DCF values for ultraﬁne (unattached) radon progeny with activity diameters ∼ 1 to 3 nm, are up to 20 times higher than for attached radon progeny, with diameters in the range 100 to 300 nm.

3. Field measurements Over the period March 1996 to October 1998, the Australian Radiation Protection and Nuclear Safety Agency (ARPANSA, formerly the Australian Radiation Laboratory) carried out a detailed investigation of radon, radon progeny and aerosol related parameters in all show caves in Australia with radon levels exceeding the workplace action level.

328

S.B. Solomon et al.

3.1. Measurement sites The Australia-wide survey of radon levels in show caves identiﬁed 14 caves in six locations with high radon levels. The measurement sites in each cave were selected after consultation with the tour guides and the cave management. The caves measured in this study were at: Jenolan. The Jenolan Caves are situated in the Blue Mountains, 180 km to the west of Sydney. There are more than three hundred caves in the area, of which nine are open for conducted tours by a workforce of 15 part-time and full-time tour guides. Seven sampling sites with access to mains power were selected for detailed measurement over the period 11 to 14 March 1996. These sites were in the Baal cave, the Cathedral chamber, Queen’s Canopy, Nellies Grotto, Sink Hole and Katies Bower chambers. Buchan. The Buchan caves are located in eastern Victoria, 260 km from Melbourne. At the time of the measurements a workforce of four tour guides conducted tours through the Royal and the Fairy caves. Ten sites were selected for the study: Skeleton, Reid, Font of Gods, Octopus, and the Princess Royal in the Royal cave, and at the entrance, the Eastern Chamber, the Ivory Palace, the Amber Bank and the King’s Chamber in the Fairy cave. Measurements were carried out for at least 24 hours at each site during the period 16 to 20 June 1997. Nelson. The Princess Margaret Rose Cave is located at Nelson, 350 km to the west of Melbourne. It is a small cave, with a workforce of four tour guides. Three sites were selected for the study, one near the entrance, a second at the mid-point of the cave and the third at the cave end. Measurements were carried out over a 24-hour period starting 29 July 1997. Gunns Plains. The Gunns Plains cave is a small limestone cave situated on the North West coast of Tasmania. It is operated by a single tour guide. Three sites were selected for the study, one near the entrance, a second at the mid-point of the cave and the third at the cave end. Measurements were carried out over a 24-hour period starting 20 October 1998. Mole Creek. Two show caves are located at Mole Creek Karst National Park in the central highlands of Tasmania. They are managed by a workforce of ten guides. Of the two caves, only King Solomon’s Cave had radon levels exceeding the action level. Three sites were selected for the study, one near the entrance, a second at the mid-point of the cave and the third at the cave end. Measurements were carried out over a 24-hour period starting 21 October 1998. Hastings. The Newdegate cave is a small limestone cave situated in the south of Tasmania. In 1997 in had a workforce of seven tour guides. Three sites were selected for the study, one near the entrance, a second at the mid-point of the cave and the third at the cave end. Measurements were carried out over a 24-hour period starting 22 October 1998. 3.2. Measurement methods A minimum of three sets of measurements were carried out in each cave system, with measurement data recorded automatically at 10 or 20 minute intervals over a period of at least 24 hours at each site. Three measurement systems were used for these studies. System A comprised a ﬂow-through scintillation cell for radon measurement and a multistage wire screen diffusion battery for radon progeny and activity size distribution measurement. Systems B and C comprised an AlphaGUARD radon monitor (Genitron GmbH, Frankfurt am Main, Germany) for radon measurement and a two-stage radon progeny sampler for radon progeny and dose conversion factor measurement. All monitors were calibrated in the Radon Reference Facility at ARPANSA prior to their use in the ﬁeld.

Assessing radiation dose for tour guides in Australian show caves from radon monitor measurements

329

3.2.1. Analysis of radon progeny size distribution for System A The potential alpha energy concentration (PAEC) and the PAEC activity size distributions were measured using a six-stage wire screen diffusion battery, which was operated with a continuous sampling rate of 1.36 × 10−5 m3 s−1 (0.8 lpm) per stage. The diffusion battery used in-situ counting of alpha particles from the radon progeny activity deposited on the collector (ﬁlter or screen) in each stage. The mode of operation of the diffusion battery has been described previously [8]. The collection and analysis of the diffusion battery data were carried out automatically using a purpose-written computer program running on a portable computer. For each 20 minute integration period, the set of six alpha activities were converted to PAEC and deconvoluted using both the Twomey [9] and the Expectation Maximisation (EMax) algorithms [10] to derive two independent particle size distributions in 40 logarithmically spaced size intervals between 0.6 nm and 1375 nm. For each sample, a size-weighted dose conversion factor was derived from measured of radon progeny size distribution, combined with the particle-size dependent DCF values, calculated from RADEP. 3.2.2. Analysis of radon progeny effective dose for Systems B and C The two stage radon progeny samplers were operated as Effective Dosimeters [11], with a 105 mesh, 0.7 cm diameter pre-separator and a 400 mesh, 4.0 cm diameter collector. It has been shown that two screen sampler where the size-dependent collection efﬁciency has been matched to the DCF response function, provides a reliable estimate of the radon progeny DCF that takes account of the radon progeny particle size distribution [11].

4. Results The wire screen diffusion battery measurements were initially analysed to derive the geometric mean (GM) and fraction of each of the modes in the radon progeny activity size distribution. The results for seven cave sites are summarised in Table 1. The GM diameters for the accumulation mode (attached radon progeny) were relatively consistent across all but one site, with a range of 130 to 160 nm. The fraction of activity in the ultraﬁne mode (unattached radon progeny) was more variable, ranging from 1 to 47%. Table 1 Fraction and geometric mean (GM) diameter of radon progeny particle size modes, and the size-weighted dose conversion factor using RADEP, estimated from diffusion battery measurements in the caves in this study Site

Ultraﬁne mode % GM (nm)

Accumulation mode % GM (nm)

Size-weighted DCF from RADEP mSv/(J h m−3 ) (mSv/WLM)

Baal Cave, Jenolan Katies Bower, Jenolan Kings Chamber, Fairy Cave, Buchan Princess Margaret Rose Cave, Nelson Gunns Plains Cave King Solomons Cave, Mole Creek Newdegate Cave, Hastings

46.6 1.0 15.5 6.2 18.5 16.8 7.4

52.5 99.0 83.7 52.5 74.1 76.7 67.8

3950 (13.9) 1670 (5.9) 2720 (9.5) 1960 (6.9) 2930 (10.3) 2920 (10.3) 2620 (9.2)

0.9 0.6 5.7 1.5 1.3 4.1 2.6

140.9 159.1 143.2 130.2 138.2 155.8 251.4

330

S.B. Solomon et al.

The derived size-weighted dose conversion factors for these size distributions were a factor of 1.2 to 2.8 times the ICRP65 conversion convention value of 1425 mSv/(J h m−3 ). Radon and radon progeny levels from the System B and C measurements are summarised in Table 2. The measured radon levels for the electronic monitors were consistent with the seasonal values from the previous radon survey for most sites. Reﬂecting the wide variation in cave ventilation conditions, the radon progeny equilibrium ratios ranged from 0.19 to 0.75, with an average value of 0.4. The fraction of activity collected on the screen collector of Table 2 Average values for radon concentration, radon progeny potential alpha energy concentration (PAEC), equilibrium ratio (F ) and percent fraction of unattached progeny as determined from the Effective Dosimeter, for all cave sites in this study Site Jenolan Baal Cathedral Queens Canopy Nellies Grotto Sink Hole Royal Skeleton Reid Font of Gods Octopus Cave Fairy Fairy Entrance Eastern Chamber Ivory Palace Amber Bank Princess Margaret Rose Entrance Center End Section Gunns Plains Entrance Center End Section King Solomon’s Cave Entrance King’s Chamber End Chamber Newdegate Entrance Main Chamber Top Chamber Average 1 Standard deviation Maximum value Minimum value

Rn (Bq m−3 )

PAEC (μJ m−3 )

(mWL)

F

% fHe

2063 1275 3573 4808 646

3.1 3.4 5.6 11.1 2.5

(147) (166) (271) (535) (118)

0.22 0.48 0.35 0.40 0.75

17.4 6.3 15.6 9.8 4.0

1770 1857 1968 1289

4.4 4.7 3.8 2.5

(212) (228) (181) (119)

0.44 0.34 0.34

7.6 7.9 11.4 8.3

2339 2398 2277 1848

4.9 5.1 3.4 3.3

(234) (243) (161) (158)

0.38 0.38 0.26 0.32

8.1 9.8 11.6 8.7

500 531 516

1.3 1.4 1.8

(61.3) (66.5) (85.4)

0.46 0.47 0.63

8.1 6.4 6.3

67 194 164

0.2 0.2 0.3

(8.0) (11) (15

0.50 0.22 0.36

10.3 11.3 8.6

216 1266 2426

0.2 2.0 3.8

(11) (98) (182)

0.19 0.30 0.28

16.2 13.3 12.8

514 740 1076

1.1 1.3 2.8

(54) (64) (133

0.43 0.34 0.46 0.39 0.13 0.75 0.19

9.3 9.6 9.1 9.9 3.2 17.4 4.0

Assessing radiation dose for tour guides in Australian show caves from radon monitor measurements

331

the Effective Dosimeter were in the range 4 to 17% and show an inverse relationship to the equilibrium ratios; the higher the equilibrium ratio, the lower the collected fraction. Table 3 summarises the derived radon progeny dose conversion factors for all measurement sites in this study. The value for the Baal Cave in Table 3 was derived from the Effective Dosimeter data, while the value in Table 1 was derived from the diffusion battery data; nevertheless the derived factors are consistent. The average value across all sites Table 3 Average radon progeny and radon dose conversion factors (DCF) for each cave site, determined from measurement data using dosimetric modelling (RADEP) and ICRP Publication 65 conversion convention % fHe

Jenolan Baal Cathedral Queens Canopy Nellies Grotto Sink Hole Royal Skeleton Reid Font of Gods Octopus Cave Fairy Fairy Entrance Eastern Chamber Ivory Palace Amber Bank Princess Margaret Rose Entrance Center End Section Gunns Plains Entrance Center End Section King Solomon’s Cave Entrance King’s Chamber End Chamber Newdegate Entrance Main Chamber Top Chamber Average 1 Standard Deviation Maximum value Minimum value

Progeny DCF from RADEP

Radon DCF (μSv per kBq h m−3 )

(mSv/(J h m−3 ))

(mSv/WLM)

RADEP

ICRP 65

17.4 6.3 15.6 9.8 4.0

3634 1824 3306 2508 1454

(12.8) (6.4) (11.6) (8.8) (5.1)

5.5 5.0 6.3 5.8 6.0

2.2 3.9 2.7 3.3 5.9

7.6 7.9 11.4 8.3

2078 2138 2786 2204

(7.3) (7.5) (9.8) (7.7)

5.2 5.3 5.3 4.1

3.6 3.5 2.7 2.7

8.1 9.8 11.6 8.7

2172 2504 2820 2292

(7.6) (8.8) (9.9) (8.0)

4.5 5.3 4.0 4.0

3.0 3.0 2.0 2.5

8.1 6.4 6.3

2167 1841 1818

(7.6) (6.5) (6.4)

5.5 5.1 6.5

3.6 3.9 5.1

10.3 11.3 8.6

2571 2929 2265

(9.0) (10.3) (7.9)

6.9 3.5 4.6

3.8 1.7 2.9

16.3 13.3 12.8

3615 2923 3034

(12.7) (10.3) (10.6)

3.8 4.8 4.7

1.5 2.3 2.2

9.3 9.6 9.1 9.9 3.2 17.4 4.0

2399 2617 2353 2490 538 3634 1454

(8.4) (9.2) (8.3) (8.7) (1.9) (12.8) (5.1)

5.5 4.9 6.0 5.1 0.8 6.9 3.5

3.3 2.7 3.6 3.1 1.0 5.9 1.5

332

S.B. Solomon et al.

was 2490 mSv/(J h m−3 ) with a standard error of 540 mSv/(J h m−3 ), and a range 1454 to 3634 mSv/(J h m−3 ). This average value is a factor of 1.7 times the ICRP65 conversion convention. For each site, the radon progeny DCF values were combined with the measured radon values at that site to derive an estimate of the inhalation dose per unit radon exposure. The derived values, listed in the fourth column of Table 3, were found to be relatively uniform across the measurement sites, with a range of values from 3.5 to 6.9 μSv per kBq h m−3 and a mean value and standard error of 5.1 ± 0.8 μSv per kBq h m−3 . For comparison, the ﬁfth column in Table 3 lists the dose per unit radon exposure for from radon progeny DCF values derived from the ICRP65 conversion convention.

5. Discussion The estimates of the size-weighted dose conversion factors in this study were on average 1.7 times the ICRP65 recommended conversion convention. In the previous study of radon levels in Australian caves, the inhalation dose estimates for tour guides employed in these caves, calculated using the ICRP65 radon conversion convention, ranged from less than 0.01 to ∼ 9 mSv per year [1]. Use of the conversion convention signiﬁcantly underestimates the inhalation for the aerosol conditions found in these caves, by up to a factor of 2.8. Applying the size-weighted dose conversion factor to the highest tour guide exposure in the survey gives an inhalation dose estimate of 16 mSv per year. Most of the caves in this study had high unattached fractions, implying low aerosol concentrations. The measurement results show a range of a factor of two about the mean equilibrium value of 0.4, and a range of ±50% about the mean radon progeny DCF value. In contrast, the radon-based DCF values had a range of ±35% and a standard deviation of ±16% for the average value of 5 μSv per kBq h m−3 . A dose conversion factor based on radon exposure of 5 μSv h−1 per kBq m−3 provides a more consistent estimate of effective dose than a dose estimates based on radon progeny exposure for the range of aerosol conditions in these caves. These results in this work would suggest that, rather than estimating equilibrium ratios and applying radon progeny-based DCF values or using site speciﬁc values, an approach applying a single, radon-based DCF value to radon monitor results will give reliable estimates of tour guide inhalation dose.

Acknowledgements The authors thank the management and staff of the various show caves studied in this work for their advice and assistance and for providing access to the caves. The assistance of ARPANSA staff, in the development of measurements systems and in the collection and analysis of ﬁeld data, and the input of Dr A.C. James and Dr A. Birchall in providing the RADEP results is acknowledged.

Assessing radiation dose for tour guides in Australian show caves from radon monitor measurements

333

References [1] S.B. Solomon, R. Langroo, J.R. Peggie, R.G. Lyons, J.M. James, Occupational exposure to radon in Australian tourist caves: an Australia-wide study of radon levels, Australian Radiation Laboratory technical report ARL/TR119, 1996. [2] National Health and Medical Research Council (NHMRC), National Occupational Health and Safety Commission (NOHCS), Recommendations for Limiting Exposure to Ionising Radiation [Guidance note NOHSC:3022 (1995)], Radiation Health Series 39, Australian Government Publishing Service, Canberra 1995. [3] Occupational Health and Safety Commission (NOHCS), National Standard for Limiting Occupational Exposure to Ionising Radiation [NOHSC:1013 (1995)], Australian Government Publishing Service, Canberra, 1995. [4] S.B. Solomon, R. Langroo, R.G. Lyons, J.M. James, Occupational exposure to radon in Australian show caves, Environ. Int. 22 (Suppl. 1) (1996) 409–413. [5] ICRP Publication 65: Protection against radon-222 at home and at work, Ann. ICRP 23 (2) (1993). [6] ICRP Publication 66: Human respiratory tract model for radiological protection, A report of Committee 2 of the ICRP, Ann. ICRP 24 (1–4) (1994). [7] A. Birchall, A.C. James, Uncertainty analysis of the effective dose per unit exposure from radon progeny and implications for ICRP risk-weighting factors, Radiat. Protect. Dosim. 60 (4) (1995) 321–326. [8] S.B. Solomon, M. Wilks, Characterisation of airborne radioactivity in an Australian dwelling, Radiat. Prot. Dosim. 56 (1994) 109–111. [9] S. Twomey, Comparison of constrained linear inversion and an iterative algorithm applied to indirect estimation of particle size distribution, J. Comp. Phys. 18 (1975) 188–200. [10] E.F. Maher, N.M. Laird, Algorithm reconstruction of particle size distributions from diffusion battery data, J. Aerosol Sci. 16 (1985) 557–570. [11] S.B. Solomon, A radon progeny sampler for the determination of effective dose, Radiat. Protect. Dosim. 72 (1997) 71–142.

334

Inﬂuence of variable stress on underground radon concentrations A. Kies, F. Massen, Z. Tosheva Centre Universitaire, 162a avenue de la Faïencerie, L-1511 Luxembourg

For several years, radon gas concentrations have been measured continuously in boreholes drilled through the concrete bottom of galleries situated under the upper water storage reservoir of the pumping storage power plant in Vianden (Luxembourg). Water levels in the reservoir vary daily over some 14 m, depending on the amount of water needed for the power output. It can be shown that radon concentrations are highly affected by these daily, often irregular variations of the water level. The water reservoir is divided into two parts I and II. Due to maintenance works, for some period either reservoir I or reservoir II was emptied leading to a quite different radon response. The draining effect of the concrete bottom of the overlying reservoir and the typical geologic features of the rock under the reservoir are proposed for the interpretation of the important inﬂuence of variable volumetric strain and stress on the radon concentrations. Two competing inﬂuences, one short-scale and another long-scale, are used in a preliminary model for interpretation of the data.

1. Introduction and experimental set-up The present contribution presents the results of underground radon monitoring in relation to local stress and strain changes. After an investigation of the inﬂuence of earth tides on underground radon concentrations [1], we were interested in the inﬂuence of changing loads on radon levels [2]. Changing water loads and thus changing volumetric strains can be observed under large water reservoirs with varying water levels. For the present study we used the Vianden pumped storage facility to the North of Luxembourg. The Vianden electric storage power station is situated in the Devonian part in the North of the Grand Duchy of Luxembourg. The pumping storage plant serves as a buffer unit within the West-European electricity network. The lower basin has an effective volume of some 10 Mm3 , with water levels ranging between 219 m and 228.5 m. The upper reservoir consists of two connected reservoirs with a useful water capacity of 3.0 Mm3 for reservoir I RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07037-8

© 2005 Elsevier Ltd. All rights reserved.

Inﬂuence of variable stress on underground radon concentrations

335

Fig. 1. Situation of the two upper reservoirs and some geological features like variscan faulting and fractures observed, important for the interpretation of the data [3].

and 3.8 Mm3 for reservoir II, occupying an area of 50 ha (Fig. 1). The minimum and maximum water heights vary between 494.0 m and 510.3 m, equivalent to a water level variation of 16.3 m. In general the connection between reservoirs I and II is open, such that in normal use water levels are nearly the same. During the two reported investigation periods, due to routine works, either reservoir I or reservoir II was emptied for some months. The underground radon measurements were performed in the narrow tunnel galleries under the upper reservoir II and recently also under reservoir I. One meter deep vertical boreholes were drilled through the concrete ﬂoor of the galleries into the bedrock. Radon gas concentrations were measured continuously in the packer tightened boreholes, either by ﬂow through radon monitors (Radim), or by passive Geiger–Mueller counters (Aware). The latter show an excellent correlation with radon concentrations measured by classical radon monitors.

2. Results and discussion For more than 3 years radon has been measured continuously in a borehole under reservoir II. Fig. 2 gives an example of radon gas concentrations in the air of the packer-tightened borehole when both reservoirs I and II were working simultaneously. The effect of changing water loads on radon concentrations is very important; radon concentrations are correlated directly with those loads. In general, a response time of 2 to 3 hours is observed between the cause (decrease of water level) and the effect (variation of radon concentrations). Very often a non-linear effect was observed, in the sense that if the water level decreased less than 6 m, spikes of very low radon levels could be observed. It seems that decreasing common loads from reservoirs 1 and 2 are able to open radon pathways that have a diluting effect on radon concentrations measured in the vicinity of the borehole. Another explanation of the observed behaviour could be changing water levels in a possible fracture system under the reservoirs. A quite different radon response is observed if either reservoir I or reservoir II is empty.

336

A. Kies et al.

Fig. 2. Radon concentrations measured in the borehole. For the measuring time water level variations of both reservoirs were the same.

Fig. 3. Radon dose-rate (μR h−1 ) in the borehole under reservoir 2 and water level variations in reservoir II. During the investigated period, reservoir I was empty.

Fig. 3 shows that if reservoir I is empty, i.e. only reservoir II is working, only a small inﬂuence of water level on radon concentrations is observed. Furthermore we notice an anticorrelation between water level and radon concentration. This effect was expected when we planned the experience: decreasing water levels lower the vertical load in the rocks. As a consequence, supplementary pathways for radon migration may open and induce an increased radon transport through the rocks. Reservoir II is empty. Fig. 4 shows the radon concentrations before, during and after the emptying process of reservoir II. Once reservoir II is empty, radon concentrations experienced an exponential ingrown with a 10 h time constant to stable equilibrium concentrations, the same as the highest measured under full load. It is interesting to note that now changing water levels of the reservoir I (not represented in Fig. 4) do not inﬂuence radon concentrations at all. For the interpretation of the results we may use a model of stress-energy framework, which suggests that the induced stress situations decouple between different scale lengths [4]. We can assume spatial heterogeneity in the rocks with different correlation lengths: small-scale grain-grain interactions and a macro scale network of clustered cracks and fractures for longscale interactions.

Inﬂuence of variable stress on underground radon concentrations

337

Fig. 4. Inﬂuence of the emptying process of reservoir II on radon concentration in the borehole under reservoir II. Actually a dose-rate (in μR h−1 ) proportional to radon concentrations was measured.

Long scale effects are dominant when both reservoirs are working. The density and orientation of cracks is essential for the possibility of fracture–fracture communication of ﬂuids. Long-scale effects are much more important than small-scale effects, as shown by the differences in the variations of radon concentrations from Figs. 2 and 3. The anti-correlation between radon and load observed in Fig. 3 is masked by the strong direct correlation of Fig. 2. When reservoir II is emptied, the long-scale effect is inhibited as are the small scale interactions. If only reservoir II over the measuring point was working, radon concentrations are inﬂuenced by small-scale grain–grain interactions where grain-scale defects interact with stress energy. Recently, measurements have been initiated in boreholes drilled in different positions of the galleries. For the moment it is too early to report on the results of these investigations.

3. Conclusion We show that variable water levels induce a signiﬁcant change in radon concentrations in the air of boreholes under a large water reservoir of a pumped storage station. Changing water levels of both reservoirs has a surprisingly high inﬂuence on radon concentrations that exhibit a direct correlation with water levels. Long-scale interactions through a macroscale network of clustered cracks and fractures are thought to be responsible for this behaviour. In the case of localized loads when only the reservoir over the measuring point is working, a small effect, radon concentrations anti-correlated to water levels, is observed. Emptying the reservoir above the borehole leads to constant radon concentrations unaffected by water levels of the reservoir. Unfortunately, until now, no simultaneous measurements of radon concentrations and other geophysical parameters have been performed. Further measurements are planned, in parallel with other geophysical investigations that may permit a better modelling of the stress induced variations of radon concentrations under the reservoir of the Vianden pumped storage station.

338

A. Kies et al.

Acknowledgements The authors express their cordial thanks to the management, especially to director Weis, and the co-workers of the Vianden power station, as well as to the Société Electrique de l’Our (SEO), Luxembourg.

References [1] A. Kies, J. Majerus, N. d’Oreye, Underground radon concentrations related to earth tides, Nuovo Cimento 22C (3–4) (1999) 287–293. [2] M. Trique, P. Richon, P. Ferrier, J.P. Avouac, J.C. Sabroux, Radon emanation and electric potential variation associated with transient deformation near reservoir lakes, Nature 399 (1999) 137–141. [3] J. Bintz, Die Geologie und der varistische Gebirgsbau im Bereich des Pumpspeicherwerkes Vianden, Publication XIV du Service Géologique de Luxembourg, 79–100, 1964. [4] P.C. Leary, Rock as a critical-point system and the inherent implausibility of reliable earthquake prediction, Geophys. J. Int. 131 (3) (1997) 451–466.

339

Measurement of the unattached radon decay products with an annular diffusion channel battery N. Michielsen a , V. Voisin a , G. Tymen b a Institut de Radioprotection et de Sûreté Nucléaire, IRSN/DPEA/SERAC, CEA-Saclay, Bâtiment 389,

BP 68, 91192 Gif-sur-Yvette cedex, France b Laboratoire de Recherches Appliquées Atmosphère – Hydrosphère, Université de Bretagne Occidentale,

6 avenue Le Gorgeu, 29285 Brest cedex, France

This paper describes a new method to determine the size distribution of short-lived 222 Rn decay products of nanometer size. In order to eliminate certain problems induced by the instruments classically used, a new device has been built. It consists of ﬁve Annular Diffusion Channels of different lengths equipped with a reference ﬁlter; they operate in parallel. During sampling, the nanometer-range particles are partly trapped by diffusion onto the inner channel walls, whereas the others are collected on downstream ﬁltering membranes. From activity data measured in each unit, the size distribution of nanometer size radon daughters can be reconstructed through a non-linear inversion method, EVE. The size range of concern is within 0.3 and 5 nm and corresponds to diffusion coefﬁcients in the range 0.2–22 mm2 s−1 . 1. Introduction It is currently admitted that about half the effective dose from exposure to natural radiations results from short-lived 222 Rn daughter inhalation [1]. To assess the dose delivered to the target tissues of the airways, the size distribution of inhaled particles must be accurately known. The classical models of particle deposition in the lungs have evidenced that the radiation dose per unit exposure of nanometric short-lived radon daughters, i.e. the unattached fraction, is higher than that of the attached radon daughters. Moreover, the diffusion properties of this fraction affect both its attachment to the aerosols being present and its deposition onto surfaces which are the basic processes used in the “room model” to calculate the indoor activities of radon daughters [2]. Most of the measuring devices used to determine the size distribution of the unattached fraction are based on the wire screen method. This technique, however, has two drawbacks: on the one hand, the grid may collect a part of the attached fraction, and on the other hand, the ﬁne radioelements can leave the screen by recoil effect. In order to avoid these problems RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07038-X

© 2005 Elsevier Ltd. All rights reserved.

340

N. Michielsen et al.

we have thus developed a new technique using an annular diffusion channel (ADC) [3]; this channel was modiﬁed to allow continuous measurement of the unattached fraction [4].

2. Materials and methods In this section the measurement system, the ADC, and the data deconvolution algorithm EVE are presented and discussed. 2.1. Measurement system The ADC diffusion battery consists of six units: ﬁve ADCs of different length plus a reference unit, all operated in parallel (Fig. 1). The sampled air is drawn through the ADC where diffusive particles are deposited; the remaining particles are collected onto a membrane ﬁlter (Poretics, polycarbonate, 0.8 μm). The alpha particles emitted by the 218 Po and 214 Po collected, or formed on the ﬁlter, are detected by an alpha PIPS detector (Canberra 450) placed in the inner tube of the ADC opposite the ﬁlter. Each unit of the diffusion battery system is characterised by the particle penetration curve; this parameter was the subject of theoretical studies [5] that led to the following expression: P=

∞

Ai exp(−2Bi μ).

(1)

i=1

Ai and Bi depend on the R1 /R2 ratio where R1 is the inner radius and R2 is the outer one, and μ=

DL 4U (R1 − R2 )2

Fig. 1. Schematic diagram of one ADC and the reference unit.

(2)

Measurement of the unattached radon decay products with an annular diffusion channel battery

341

where D is the diffusion coefﬁcient of nanometer particles, L is the channel length and U is the mean velocity of the air within the channel. The penetration curve of nanometer-sized particles of the annular channel was compared to that of a screen. Figure 2 shows that the slope of the channel penetration curve is steeper than that of the screen, which is indicative of the channel’s higher selectivity. Comparison of an ADC with a tube diffusion battery of alike penetration capability provided similar penetration curves. However, because of its geometry, the length required to obtain a developing ﬂow is much shorter for a channel than for a tube, and then, most of the time, the entrance effect can be neglected in an ADC. These two examples highlight why the use of an annular diffusion battery is beneﬁcial for such a study. Table 1 gives the operating parameters where D (50%) is the diffusion coefﬁcient for a 50% collection efﬁciency and dp (50%) the corresponding particle diameter. Let us note that, in this nanometer size range, the diffusion coefﬁcient of the particle should also ﬁt to the gas kinetic theory. Then, a correction for particle under 2 nm has been applied [6]. The penetration curves of the various ADC including the reference unit are plotted in Fig. 3. The activity concentration of the short-lived radon decay products collected on the various channel ﬁlters is determined using the method developed by Tremblay et al. [7] with a 20minute sampling and one-hour counting time. The detection and analysis procedure involves the measurement of the number of alpha particles decayed from 218 Po and 214 Po over two

Fig. 2. Comparison of the penetration curve of an annular channel with that of a wire screen. Table 1 Operating parameters of the ADC for R1 = 1.8 cm and R2 = 2 cm

ADC1 ADC2 ADC3 ADC4 ADC5 Reference

Length (cm)

Flow rate (L min−1 )

D (50%) (mm2 s−1 )

dp (50%) (nm)

2.5 4.5 10 25 30 –

12.8 12.4 12.4 12.5 5.3 11.7

11.5 6.4 2.9 1.1 0.37

0.46 0.7 1.2 2.1 3.7

342

N. Michielsen et al.

Fig. 3. ADC penetration curves.

successive counting time-periods. Because of the instrument design, this method allowed us to simultaneously carry out sampling and counting, which increased the precision in the calculations of concentration activities of 218 Po in particular. Each measurement cycle is separated from the next by a three-hour period in order to allow decay of the 214 Pb and 214 Bi atoms trapped on the ﬁlters. 2.2. Deconvolution algorithms The non-linear data inversion method Extreme Value Estimation, EVE [8], has been used in order to analyse the size or the diffusion coefﬁcient distribution of the nanometer radon daughters. This method, based on the minimisation of the χ 2 function, consists in ﬁnding a set of possible solutions included in a conﬁdence interval. The main advantage of EVE are that it takes into account the measurement uncertainties and gives conﬁdence intervals. We currently use the version 11.0 released in January 1994 by Pentti Paatero. For the example presented in the next section, the window of interest was taken from 0.1 to 10 nm; a lognormal distribution was assumed and 50 mean diameters were taken, equidistant on a log scale, with geometric standard deviation of 1.2.

3. Results and discussion of a series of measurements Illustration of the results obtained with the ADC is provided by a series of measurements carried out in a sealed radon chamber at the INTE-UPC [9]. Seven measurements were carried out at a constant radon concentration under constant environmental conditions. Particle concentration was measured by a CNC 3025 (Condensation Nuclei Counter) and remained below 20 part cm−3 throughout the experiment. The analysis of raw data from the ﬁlter of each ADC showed a decrease in the activity concentration of radon decay products, mainly 218 Po, from channel 1 to channel 5. Only a few percent of activity passed through channel 5; this indicated that nearly almost all radon daughters were unattached (Fig. 4).

Measurement of the unattached radon decay products with an annular diffusion channel battery

343

Fig. 4. Penetration data obtained for each unit of the ADC.

Considering the low concentration activity of 214 Pb and 214 Bi over these experiments, only the size distribution of the nanometer-sized 218 Po was reconstructed. As shown in Fig. 4, variability in measurements on each channel was very low, which is indicative of good reproducibility under stable conditions. Mean values and the standard deviations of the penetration efﬁciency through each channel were as follows: – – – – –

ADC1: 6.32E−01 ± 1.23E−02, ADC2: 5.12E−01 ± 1.72E−02, ADC3: 3.28E−01 ± 1.28E−02, ADC4: 7.13E−02 ± 7.90E−03, ADC5: 1.88E−02 ± 2.70E−03.

These values were used with EVE to reconstruct the size distribution together with its conﬁdence interval (Fig. 5). This size reconstruction gave a Chi-2 of 8.8 and the following calculated penetration for each unit: – – – – –

ADC1: 6.38E−01, ADC2: 5.14E−01, ADC3: 3.00E−01, ADC4: 8.66E−02, ADC5: 1.78E−02.

One should note that solutions with diameter under 0.3 nm are physically impossible since that is the limit of the size of the atom itself. Then, one can conclude that the size distribution of cluster 218 Po found during this experiment is mono-modal with a diameter of about 0.84 nm, which corresponds to a diffusion coefﬁcient of 5 × 10−6 m2 s−1 (0.05 cm2 s−1 ).

344

N. Michielsen et al.

Fig. 5. Size distribution results.

4. Conclusion A prototype ADC battery to measure size distribution of short-lived radon decay products in nanometer form was designed to have a better selectivity than current devices. The EVE deconvolution algorithm was chosen because it takes the uncertainty of the measurement into account and produces conﬁdence intervals on the results. Experiments with this apparatus showed a good reproducibility of the results. The calculated particle size distribution of the 218 Po showed a mode centred at 5 × 10−6 m2 s−1 (diffusion coefﬁcient) corresponding to a thermodynamic diameter of 0.84 nm in agreement with data previously reported in the literature. The ADC battery needs now to be assessed over ﬁeld experiments in indoor environments where unattached and attached radon decay products are mixed. Acknowledgement The authors thank Dr. Christelle Huet for help and fruitful discussion on the EVE deconvolution method. References [1] UNSCEAR, in: Exposures from Natural Radiation Sources, 48th Session of UNSCEAR, Vienna, 12 to 16 April 1999. [2] A.M. Gouronnec, F. Goutelard, N. Montassier, D. Boulaud, A. Renoux, G. Tymen, Behaviour of radon and its daughters in a basement: model–experiment comparison, Aerosol Sci. Technol. 25 (1996) 73–89. [3] G. Tymen, D. Kerouanton, C. Huet, D. Boulaud, An annular diffusion channel equipped with a track detector ﬁlm for long-term measurements of activity concentration and size distribution of 218 Po particles, J. Aerosol Sci. 30 (1999) 205–216. [4] C. Huet, G. Tymen, D. Boulaud, Long-term measurements of equilibrium factor and unattached fraction of shortlived radon decay products in a dwelling – comparison with PRADDO model, Aerosol Sci. Technol. 35 (2001) 553–563.

Measurement of the unattached radon decay products with an annular diffusion channel battery

345

[5] D. Kerouanton, G. Tymen, D. Boulaud, Small particle diffusion penetration of an annular duct compared to other geometries, J. Aerosol Sci. 27 (1996) 345–349. [6] M. Ramamurthi, The detection and measurement of the activity size distribution (dp > 0.5 nm) associated with radon decay products in indoor air, Thesis from the University of Illinois, 1989. [7] R.J. Tremblay, A. Leclerc, C. Mathieu, R. Pepin, M.G. Townsend, Measurement of radon progeny concentration in air by a-particle spectrometric countiilg during and after air sampling, Health Physics 36 (1979) 401–411. [8] P. Paatero, The extreme value estimation deconvolution method with application to aerosol research, University of Helsinki, Report Series in Physics, HU-P-250, 1990. [9] A. Vargas, N. Michielsen, C. Le Moing, M. Rio, G. Tymen, X. Ortega, Determination of 218 Po nanometer size distribution in controlled environment by two new systems, in: J.P. McLaughlin, S.E. Simopoulos, F. Steinhäusler (Eds.), The Natural Radiation Environment VII, Proc. VIIth International Symposium on the NRE, Rhodes, Greece, 20–24 May 2002, in: Radioactivity in the Environment, vol. 7, Elsevier, Amsterdam, 2004, this volume.

346

Intercomparison of the radiosensitivity of different measuring devices for environmental levels of radon A.C. George, N. Bredhoff Radon Testing Corporation of America, 2 Hayes Street, Elmsford, NY 10523, USA

Different types of radon measuring devices have been developed and used in the last twenty years to meet the demand of the radon industry. At ﬁrst, most of the radon measuring methods and techniques were developed in government laboratories to address the needs of the government programs such as in the characterization of uranium mine atmospheres, in contaminated US Atomic Energy Commission Excess Sites, and in indoor environments. With the discovery of elevated indoor radon concentrations throughout most of the US, a great deal of awareness was generated about the health risks from exposure to radon/thoron and their decay products. This increased interest on the health risk from radon exposure, stimulated instrument research and development to meet the need for short-term and long-term radon measurements. In the US, unlike other countries, most of the radon measurements are performed with short-term passive integrating or continuous monitoring devices over periods of 2–7 days. Although, several types of instruments for measuring the concentration of radon decay products were developed, they are hardly used for routine monitoring of the exposure risk from radon decay products. For this reason, only the sensitivities of radon monitoring devices will be discussed in this paper. During the last 20 years more than 13 million radon measurements were made in the US alone, with results showing that about 8% of the homes have radon levels above the US EPA action level of 150 Bq m−3 (4 pCi L−1 ). There are more than 60 millions homes yet to be tested initially. Since, high radon levels have been found in every State, there is a need to encourage testing in all homes. The selection of the radon measuring instrument, is based on the manufacturer’s claim what their instrument can do without comparison with other competing instrument developers. Comparison of the radiosensitivity among different radon measuring methods and techniques can make it easier for radon testing ﬁrms to choose the most appropriate one for their ﬁeld measurements. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07039-1

© 2005 Elsevier Ltd. All rights reserved.

Intercomparison of the radiosensitivity of different measuring devices for environmental levels of radon 347

This paper compares the sensitivity of every radon device we could get information from its manufacturer in net counts min−1 per 150 Bq m−3 . This value was selected from the fact that most of the measurements in the US are below 150 Bq m−3 . In all, nineteen different devices were intercompared. They represent active and passive continuous radon monitors such as scintillation monitors pulse ionization chambers, solid state alpha detectors, current ionization chambers and several types of diffusion barrier passive activated carbon collectors. The sensitivities of short-term electret ion chambers and longterm alpha track detectors are also presented for comparison with the other standard shortterm devices used in the US radon industry. The results of the intercomparisons indicate that in most cases the sensitivity of each device is roughly proportional to its sensitive volume or sensitive mass. The sensitivities of most portable continuous radon monitors range from 0.8 to 24.0 net cpm/150 Bq m−3 whereas the sensitivity of the most frequently used continuous radon monitors is < 5 net cpm/150 Bq m−3 . By comparison, the sensitivities of the six diffusion barrier passive activated carbon collectors range from 48–145 net cpm/150 Bq m−3 . At radon levels below 150 Bq m−3 most of the continuous radon monitors will provide very poor results if counts are acquired at one minute intervals. For this reason EPA, requires that continuous radon monitors accumulate counts based on one hour basis to improve counting statistics. Passive activated carbon collectors are usually counted for a ten minute interval yielding 480–1450 net counts/150 Bq m−3 . If counted for one hour as the continuous radon monitors are required to do, they yield 2880–8700 net counts/150 Bq m−3 .

1. Introduction According to USEPA, in the last ﬁfteen years, more than eighteen million radon measurements have been made in the US. The results show that 6–8% of the homes have radon levels equal to or greater than 150 Bq m−3 (4 pCi L−1 ). It is estimated that about 15% of the housing stock in the US has been tested, averaging about one million measurements per year. Less than 10% of the homes that tested above the action level have been mitigated. It is estimated that approximately 500 000 residential mitigation systems have been installed, averaging about 50 000 mitigations per year. If these numbers are accurate the radon measurement and mitigation ﬁrms will be active for many years. According to the National Association of Home Builders Research Center, about 200 000 new homes are built radon resistant per year, or about 20% of all the newly constructed homes. In the early days of the US Radon program, most of the short-term measurements for radon were made with open face and diffusion barrier activated carbon collectors, which were analyzed primarily using the gamma counting technique [1]. Open face activated carbon collectors were found to be very reliable for exposure intervals of two to three days, even in environments with relative humidity of 70% and where the radon concentration changed by a factor of two to four. Open face canisters containing 50–90 grams of carbon, exposed for 2 days and analyzed 3 days later, were found to be very sensitive yielding 1.3–4.2 dps/150 Bq m−3 (80–250 cpm/4 pCi L−1 ), depending on the gamma counting system and the type and size of the radon collector.

348

A.C. George, N. Bredhoff

To extend the exposure period up to seven days or longer under extreme conditions of humidity and varying radon concentrations, the open face collector was modiﬁed to improve its response by adding a diffusion barrier. This modiﬁcation extended the time of exposure from 2–4 days up to 7–10 days for the activated carbon method. The ﬁrst diffusion barrier activated carbon collector, developed and fully evaluated by USDOE, demonstrated very accurate radon measurement results in radon chambers and in residential buildings [2]. The USEPA also developed and tested a diffusion barrier collector [3], similar in design to the USDOE radon collector, and achieved excellent results from both radon chamber tests and from ﬁeld environments. The evaluation and standard operating procedures for the EPA diffusion barrier charcoal canister was published in an EPA, National Air and Radiation Environmental Laboratory Report [4]. The measurement period for the diffusion barrier radon collector was extended from 2–4 days up to 2–7 days. Excellent measurement results were achieved both in the laboratory and in ﬁeld situations with the diffusion barrier collector, even when the radon concentration varied by more than 10 to 1 and the humidity ranged from 20 to 85%. Currently, the EPA diffusion barrier activated carbon collector is being used by the EPA Las Vegas Laboratory in its short-term radon measurements. Measurements conducted with diffusion barrier activated carbon collectors developed by the Radon Testing Corporation of America (RTCA), demonstrated very good agreement with continuous radon monitors during evaluation in radon calibration facilities where the radon concentration was varied by a factor of 120. The Pennsylvania DEP Bureau of Radiation Protection also manufactures a diffusion barrier activated carbon collector similar in design to the EPA device. The results of ﬁeld tests showed very good agreement between the DEP carbon collectors, continuous radon monitors and electret ion chambers when exposed simultaneously in environments where the radon varied from 75 to 225 Bq m−3 (26.1 pCi L−1 ) and from 260 to 850 Bq m−3 (7–23 pCi L−1 ), respectively.

2. Test results and intercomparisons The design characteristics and sensitivities of ﬁve different size, type and conﬁguration diffusion barrier activated carbon collectors and an open face activated carbon collector are shown in Table 1. Listed are those collectors for which information was available in previous publications or publicized results. The radon collectors were exposed to radon at 150 Bq m−3 (4 pCi L−1 ) for 2 days and analyzed 3 days after exposure. Analysis was performed by the gamma counting technique. Included also in the table, are the characteristics of a small RTCA diffusion barrier vial containing 2 grams of activated carbon using alpha counting by the liquid scintillation (LS) technique. The results show that the sensitivity of all activated carbon collectors is high and is roughly proportional to the mass of carbon, the type of counting method used in the analysis and design – open face or diffusion barrier type. The high net sensitivity of these collectors is due to the accumulation of radon on the collector. The counting rate decreases only with the decay rate of radon from the end of exposure till counted. In contrast, continuous radon monitors detect radon passing actively or passively through their sensitive volume without

Intercomparison of the radiosensitivity of different measuring devices for environmental levels of radon 349 Table 1 Characteristics and sensitivities of different activated carbon collectors Facility id.

EML, DOE (7.6 cm) DB US, EPA (10.2 cm) DB EPA, DER (10.2 cm) DB RTCA ( 7.6 cm) DB RTCA (10.2 cm) DB RTCA (10.2 cm) OF RTCA LS (20 ml) DB

Carbon mass (9-)

Counting efﬁciency (cps/Bq)

(dps/150 Bq m−3 )

Net sensitivity (cpm/4 pCi L−1 )

50 70 75 50 90 90 2

0.30 0.28 0.36 0.48 0.48 0.48 2.54

0.5 0.8 1.0 1.5 2.4 4.2 0.9

30 48 62 90 145 250 53

DB = Diffusion barrier type, OF = Open face type.

Table 2 EPA, activated carbon canisters versus continuous radon monitors Test No.

Continuous mean (Bq m−3 )

Radon monitor range (Bq m−3 )

Concentration variation factor

EPA canisters mean (Bq m−3 )

Difference (%)

1 2. 3 4 5 6 7

352 111 233 181 192 66 207

19−630 44−289 130−463 37−344 55−389 19−150 111−352

33 7 4 9 7 8 3

325 96 230 181 192 52 218

7 13 2 0 0 22 5

the advantage of accumulation. The evaluation of the EPA activated carbon collectors under varying radon concentration is listed in Table 2. The diffusion barrier type carbon collectors were exposed simultaneously from 6–7 days with continuous radon monitors in seven residential structures. The agreement by the two techniques, shown as performance ratio, is very good even when the radon concentration varied by a factor of 3–33. The higher performance ratios of tests 2 and 6 are probably due to the poorer sensitivity of the continuous radon monitor at radon concentration levels below 150 Bq m−3 (4 pCi L−1 ) and to the detrimental effect of accumulated background when making measurements in low concentrations of radon. At high radon levels, the error from the background contribution is less signiﬁcant. The sensitivities of different continuous radon monitors were obtained from instrument manufacturers and from an earlier publication [5]. Table 3 lists active and passive devices of different detection techniques with various sensitive volumes. Comparing the results from Table 3, the sensitivities of activated carbon collectors at radon levels of 150 Bq m−3 (4 pCi L−1 ) are 15–50 times higher than the sensitivities of the most frequently used continuous radon monitors. Measurement results with continuous radon monitors acquiring data on minute intervals are very inaccurate. Accuracy improves when hourly

350

A.C. George, N. Bredhoff

Table 3 Sensitivity of active and passive devices for measuring radon in air Device/method

Volume (L)

dps/150 Bq m−3

cpm/4 pCi L−1

Active scintill. cell (Radonics) Active scintill. cell (Pylon AB-5) Active scintill. cell (EML, DOE) Active scintill. cell (Eberline RGM-3) Passive pulse ion. chamber (femto-TECH) Passive pulse ion. ch. (alpha Guard MC50) Active pulse ion. ch. (ATMOS-12D) Passive solid state det. (Sun Nuclear1027) Passive solid state det. (Sarad RTM-210) Active solid state det. (Durridge RAD7) Passive current ion. cham.(RTCA E-Smart) RTCA–Liquid scintillation vial RTCA (7.6 cm DB canister) RTCA (10.2 cm DB canister) RTCA (10.2 cm OF canister) US EPA (10.2 cm DB canister) PA, DER (10.2 cm DB canister)

0.10 0.27 0.46 3.00 0.20 0.55 0.60 ND 0.27 0.75 0.47

0.033 0.093 0.140 0.400 0.020 0.130 0.050 0.003 0.025 0.033 0.015 0.90 1.50 2.40 4.20 0.80 1.00

2.0 5.7 8.4 24.0 1.2 7.6 3.0 0.17 1.5 2.0 0.9 54.0 90.0 145.0 250.0 48.0 60.0

Table 4 Sensitivities of alpha track detectors and electret ion chambers Volume (ml)

Number of tracks per cm2 per 150 Bq m−3 h−1

per 4 pCi L−1 d−1

Landauer CR-39 closed REM CR-39 closed NYU CR-39 closed NYU CR-39 220 Rn and 222 Irish CR-39 closed Swedish CR-39 closed Germany makrofol closed Japanese polycarbonate closed Slovenia CR-39 closed French LR-115 open French LR-115 type II open Italian LR-115 closed

22.0 7.0 100.0 100.0

0.07 0.33 0.48 0.23 0.51 0.39 0.15 0.19 0.97 0.20 0.60 0.92

1.6 8.0 11.4 5.6 12.2 9.4 3.6 4.5 23.4 4.8 14.4 22.0

Rad Electret Ion Ch. (short-term) 200 Rad Electret Ion Ch. (long-term) 200 RTCA Radome Ion Ch. (long-term) 65

8.0 V/150 Bq m−3 d−1 0.7 V/150 Bq m−3 d−1 0.6 V/150 Bq m−3 d−1

Device id., type

counting data is used. In the US, the minimum recommended counting interval for continuous radon monitors is one hour, to improve counting statistics. In Table 4, the sensitivities of alpha track detectors and electret ion chambers are listed for comparison purposes. Although these methods cannot be compared directly with the devices

Intercomparison of the radiosensitivity of different measuring devices for environmental levels of radon 351

listed in Tables 1 and 3 in terms of net counts per 150 Bq m−3 , the information is useful for selecting a particular device for long-term measurements of radon.

3. Conclusions The radon measurement results from diffusion barrier activated carbon collectors show that the carbon collectors integrate accurately in ﬁeld situations with high humidity and extreme variation of radon concentration. When results are needed in a hurry (2–7 day test duration), activated carbon collectors are the most sensitive and most accurate devices yielding the highest net counting rate at radon levels of 150 Bq m−3 (4 pCi L−1 ) (the US Action Level). Radon measurements made with continuous radon monitors report the average value for a 2–3 day exposure at generally more expensive. Their use requires a trained operator. In contrast, the activated carbon method is the simplest, most cost effective and most sensitive technique for short-term radon tests. At radon levels below 150 Bq m−3 , most continuous radon monitors will provide poor results if counts are acquired at one minute intervals. For this reason, EPA requires that electronic continuous radon monitors accumulate counts based on hourly counting intervals to improve accuracy. Alpha track detectors exhibit similar sensitivities and are more suitable for long-term testing, providing a more accurate annual average concentration of radon. The two types of electret ion chambers have similar sensitivities and can be used for short-term or long-term measurements depending on the thickness of electret used with the device.

References [1] A.C. George, Passive integrated measurement of indoor radon using activated carbon, Health Phys. 46 (1984) 767–872. [2] A.C. George, T. Weber, An improved passive activated carbon collector for measuring environmental 222 Rn in indoor air, Health Phys. 58 (1990) 583–589. [3] D.J. Gray, J.F. Burkhart, A.P. Jacobson, An evaluation of the performance of the EPA diffusion barrier charcoal adsorber for Radon-222 measurements in indoor air, in: The 1992 International Symposium on Radon and Radon Reduction Technology, vol. 2, Session 5, Minneapolis, MN, September 22–25, 1992. [4] US Environmental Protection Agency, NAREL, Standard Operating Procedures for 222 Rn measurement using diffusion barrier charcoal canisters, EPA 520/590-032, November 1990. [5] A.C. George, State-of-the-art instruments for measuring radon/thoron and their progeny in dwellings. A review, Health Phys. 70 (1996) 451–463.

352

Integrating measurements of radon and thoron and their deposition fractions in the respiratory tract W. Zhuo, S. Tokonami, H. Yonehara, Y. Yamada Radon Research Group, National Institute of Radiological Sciences, 4-9-1 Anagawa, Inage, Chiba 263-8555, Japan

For simultaneous measurements of indoor 222 Rn and 220 Rn and estimation of their deposition fractions in the respiratory tract, a new type of passive integrating 222 Rn and 220 Rn monitor and a portable bronchial dosimeter were developed. The passive 222 Rn and 220 Rn monitor was rebuilt from a commercially available passive 222 Rn monitor. Besides its simple construction, the volume and weight of the new monitor are only 110 cm3 and 20 g. Calibration factors of 222 Rn and 220 Rn for the new monitor were systematically studied through calibration experiments. The results indicated that indoor 220 Rn could be discriminated from 222 Rn by using the new passive monitor. The bronchial dosimeter consists of three sets of progeny integrating sampling units (PISUs) with different conﬁgurations of sampling heads. Multiple metal screens are used to mimic the penetrating and deposition behaviour of 222 Rn/220 Rn progeny in the nasal and tracheo-bronchial (T-B) regions of the human respiratory tract. The potential alpha energy concentrations (PAEC) of 222 Rn/220 Rn progeny are directly measured with the allyl diglycol carbonate (CR-39) detectors inside the PISUs. The deposition fractions of 222 Rn and 220 Rn progeny in the T-B region were measured, with averages of 4.5 and 4.0% for ordinary room conditions, in general agreement with other reported values. Both the new monitor and device are simple and compact as well as of low cost, and they are considered to be practical for large-scale and long-term surveys of indoor 222 Rn and 220 Rn.

1. Introduction Public exposure to 222 Rn and its progeny has been of worldwide concern in the past decades. It is well known that the health detriment associated with radon (222 Rn) mainly arises from the inhalation of its progeny and the resultant alpha particle exposure to the epithelium of the respiratory tract [1]. Previous studies have shown that the tracheo-bronchial (T-B) region is the most radiosensitive in the respiratory tract, and the radiation dose in the region due to 222 Rn progeny constitutes a dominant fraction of the total lung dose [2]. Recent surveys revealed that the exposure to thoron (220 Rn) and its progeny could equal or even exceed that RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07040-8

© 2005 Elsevier Ltd. All rights reserved.

Integrating measurements of radon and thoron and their deposition fractions in the respiratory tract

353

of 222 Rn and its progeny in some areas of the world [3–5]. Therefore, measurements of 220 Rn are indispensable for more precise assessment of public exposure to 222 Rn and 220 Rn. For simultaneous measurements of indoor 222 Rn and 220 Rn, several types of 222 Rn and 220 Rn discriminative monitors have been developed. Doi et al. [6] developed a uniﬁed 222 Rn and 220 Rn discriminative monitor with polycarbonate ﬁlm as detectors. Besides its large volume (about 560 cm3 ), the construction of monitor and the etching processes of detectors were complicated. The 222 Rn and 220 Rn monitors developed by Guo et al. [7] and Iida et al. [8] were somewhat inconvenient for ﬁeld applications because they were two separated hemispherical cups. Moreover, it was also complicated and time-consuming to exchange the detectors. The traditional methods for measuring the deposition of 222 Rn progeny in the T-B region are tedious and difﬁcult, and generally require complex equipment [9,10]. For simulating and measuring the deposition of 222 Rn progeny in the respiratory tract, several type of sampling systems consisting of multiple layers of wire screens with appropriate wire factors and air ﬂow velocity have been developed, and the deposition fractions of 222 Rn progeny in the N and TB regions have been measured under certain conditions [11–16]. However, the inadequacy of the grab sampling method used in the measuring systems may cause large measurement uncertainties [12]. Moreover, the whole measuring systems were generally complicated, largesized and expensive, and so they are unpractical for ﬁeld surveys. In this paper, a new type of passive 222 Rn and 220 Rn monitor and a portable device for measuring the deposition fractions of 222 Rn and 220 Rn progeny in the respiratory tract are described. Because of their simple construction and low cost, they are considered to be suitable for large-scale and long-term ﬁeld surveys.

2. Materials and methods 2.1. Passive 222 Rn and 220 Rn monitor Figure 1 shows the construction of the passive 222 Rn and 220 Rn monitor. The 222 Rn monitor made in Hungary is commercially available. The diffusion chamber is a small cylindrical pot (φ = 35 mm, H = 55 mm) made of anti-statically treated plastic. The allyl diglycol carbonate (CR-39) detector is ﬁxed on the middle of the pot cover with its sensitive side towards the pot interior. The cover is screwed to match the pot bottom ﬁrmly during measurements. 222 Rn in air can penetrate into the pot through the non-visible air gaps between its cover and bottom through diffusion. The mean 222 Rn concentration during the exposure period can be derived from the track density on the CR-39, the 222 Rn calibration factor and the exposure time. Because of its simple construction and low cost, the 222 Rn monitor has been widely used for 222 Rn surveys worldwide. For this study, in order to measure 220 Rn concentrations, the air exchange rate in the original pot needed to be enhanced. Therefore, four holes (φ = 12 mm) were evenly made on the wall of the pot bottom and covered with the cellulose ﬁber ﬁlter (Whatman® No. 41) to form a 220 Rn monitor for this study. The ﬁlter paper used here was chosen for its low cost, high mechanical strength and high permeability. Filtration prevents particulate 222 Rn/220 Rn progeny from entering the pot, thus ensuring the establishment of radioactive equilibrium between 222 Rn/220 Rn and their short-lived progeny. In order to discriminate 222 Rn from the 220 Rn monitor, the 222 Rn monitor is attached to the 220 Rn monitor

354

W. Zhuo et al.

Fig. 1. Construction diagram of the remodelled passive 222 Rn and 220 Rn monitors.

using a piece of double-sided tape. The rebuilt 220 Rn and 222 Rn monitor has a total volume of only 110 cm3 and a weight of as low as 20 g. The etched-track densities on the CR-39 detectors set in the 222 Rn and 220 Rn monitors (NRn and NTn ) can be expressed by the following equations: NRn = QRn CF Rn1 T + QTn CF Tn1 T + B,

(1)

NTn = QRn CF Rn2 T + QTn CF Tn2 T + B,

(2)

where QRn and QTn are the mean concentrations of 222 Rn and 220 Rn in the exposed period in Bq m−3 , CF Rn1 and CF Tn1 are the 222 Rn and 220 Rn calibration factors for the 222 Rn monitor in tracks cm−2 (Bq m−3 h−1 )−1 , CF Rn2 and CF Tn2 are the 222 Rn and 220 Rn calibration factors for the 220 Rn monitor in tracks cm−2 (Bq m−3 h−1 )−1 , T is the exposure time in h, and B is the background track density of the CR-39 detector in cm−2 . Therefore, provided the etched-track densities, calibration factors and the exposure time were known, 222 Rn and 220 Rn concentrations could be derived from the simultaneous equations. 2.2. Bronchial dosimeter A conceptional diagram of the measuring system is shown in Fig. 2. The whole system consists of three sets of progeny integrating sampling units (PISUs) and ﬂow meters as well as a mini air pump. The single mini pump with combined ﬂows was used for air sampling. The sampling ﬂow rate is adjusted to be 13.6 cm3 s−1 for each PSIU by the ﬂow meters, resulting in a face velocity of 12.0 cm s−1 on the wire screens. PISU-A collects the total airborne activity (Aa ) on an AA type MF membrane ﬁlter with a pore size of 0.8 μm. PISU-B containing

Integrating measurements of radon and thoron and their deposition fractions in the respiratory tract

355

Fig. 2. Schematic diagram for simultaneous measurements of the concentrations of 222 Rn and 220 Rn progeny and their deposition in the respiratory tract.

a single 100-mesh wire screen sampling head simulates the collection characteristics of the human nose. Its back-up ﬁlter collects the activity penetrating the nasal region (Ab ). The percentage of deposition fractions in the N region is then given as 100(Aa − Ab )/Aa . PISU-C contains one 100-mesh and four 400-mesh wire screens in the sampling head, which mimics the combined collection characteristics of the N and T-B regions. Its back-up ﬁlter collects the activity penetrating the N and T-B regions (Ac ). Then, the percentage of deposition fraction in the T-B region can be expressed as 100(Ab − Ac )/Aa . The three sets of PISUs are identical cylinders made of stainless steel. The detailed construction of the PSIU has been reported by the author [17]. Four chips of CR-39 detectors are set beneath the absorbers at the sites demonstrated as four dotted circles in each PISU. Among the four CR-39 detectors, two are set beneath a thicker Al Mylar absorber (4.8 mg cm−2 ), and it was conﬁrmed that only alpha particles (8.78 MeV) emitting from 212 Po of 220 Rn progeny collected on the ﬁlter can be detected. While the other two detectors set beneath another thinner absorber (0.29 mg cm−2 ) can detect all of the alpha particles emitting from 222 Rn/220 Rn progeny collected on the ﬁlter. After a predetermined period of sampling, the sampling heads were sealed, and the PISUs are left for 3 days to allow all 212 Bi atoms collected on the ﬁlters to decay. From the densities of etched tracks and the integrating volume of sampling ﬂow, the PAEC of 222 Rn and 220 Rn expressed by the equilibrium-equivalent concentrations (EECRn

356

W. Zhuo et al.

and EECTn , in Bq m−3 ) were derived as follows [17]: EECRn = 4.33 × 10−2 DRn · A · V −1 ,

(3)

EECTn = 7.30 × 10−3 DTn · A · V −1 ,

(4)

where DRn and DTn are the net etched-tracks in cm−2 produced by 222 Rn and 220 Rn progeny, respectively; A is the counting area of the CR-39 in cm2 ; and V is the integrated air volume passing through the ﬁlter of each PISU in m3 .

3. Results and discussion 3.1.

222 Rn

and 220 Rn calibration factors for the new passive monitors

The calibration experiments were performed in an airtight stainless chamber with a volume of 150 L. For ﬁve 222 Rn exposure periods, the integrated exposure ranged within 367 ∼ 508 kBq h m−3 . The averaged 222 Rn calibration factors for 12 sets of 222 Rn and 220 Rn monitors were estimated to be 2.62 ± 0.20 and 2.64 ± 0.16 tracks cm−2 (kBq m−3 h−1 )−1 for the 222 Rn and 220 Rn monitors, respectively. In the 220 Rn calibration experiments, 220 Rn with minimal 222 Rn gas was continuously supplied through an airtight circulation system [18]. An interior electric fan was switched on during experiments in order to make the 220 Rn uniformly distributed in the chamber. The concentrations of 220 Rn and 222 Rn during the exposure periods were measured with scintillation cells by the grab sampling method [19]. Averaged concentrations of 220 Rn and 222 Rn were measured to be about 3200 Bq m−3 and 220 Bq m−3 during the calibration experiments which usually lasted for a week. Subtracting the track densities due to 222 Rn exposure, the derived 220 Rn calibration factors for 12 sets of 222 Rn and 220 Rn monitors are shown in Fig. 3. The mean 220 Rn calibration factors are 1.32 ± 0.14

Fig. 3. Experimental results of 220 Rn calibration factors for the 222 Rn and 220 Rn monitors. The error bars represent ±1 standard deviation from 5 time exposures.

Integrating measurements of radon and thoron and their deposition fractions in the respiratory tract

357

tracks cm−2 (kBq m−3 h−1 )−1 and 0.10 ± 0.01 tracks cm−2 (kBq m−3 h−1 )−1 for the 220 Rn and 222 Rn monitors, respectively. The 220 Rn calibration factor of the 220 Rn monitor is more than 10 times that of the 222 Rn monitor. Therefore, it is expected that 220 Rn can be evaluated from the different etched-tracks on the detectors in the 222 Rn and 220 Rn monitors, provided measurements were simultaneously performed at the same site. The 220 Rn calibration factor being less than 4% of the 222 Rn calibration factor in the 222 Rn monitor, the inﬂuence of 220 Rn on the 222 Rn measurements may be neglected in most environments. However, in some environments, such as in some Japanese wooden houses where 220 Rn concentrations are much higher than 222 Rn concentrations, the inﬂuence of 220 Rn on the 222 Rn measurements might be signiﬁcant if only the 222 Rn monitor was used for measurements. In this case, the inﬂuence of 220 Rn on the 222 Rn measurements could be quantiﬁed if the rebuilt monitor was used. Therefore, the rebuilt monitor is not only useful for 220 Rn measurements, but also helpful for more precise measurements of 222 Rn. 3.2. Main characteristics of some passive 222 Rn and 220 Rn monitors The main characteristics of the four types of passive 222 Rn and 220 Rn monitors are summarized in Table 1. Our present monitor has the highest calibration factor and lowest detection limits for 220 Rn measurements. It indicates that 220 Rn concentrations can be more precisely and easily measured with our present monitors. Because the variance of background etch-pits on the CR-39 detectors used in this work is relatively high, the lower detection limit of 222 Rn is slightly higher than for other monitors. However, it is still considered sensitive enough for indoor measurements. Besides its simple construction, the present monitor is small and light, and it is considered as the most practical for large-scale surveys among these monitors. 3.3. PAEC measurements In order to test the PAEC measurements with the PISU, measurements were performed along with a working level monitor (WLx, Pylon Electronic, Canada). Except in a radon chamber, Table 1 Main characteristics of the 4 types of passive 222 Rn and 220 Rn monitors Researcher

Doi et al. [6]

Guo et al. [7]

Iida et al. [8]

This study

Reported year Detector Polycarbonate 222 Rn calibration factora 220 Rn calibration factora LDL of 222 Rn (Bq m−3 )b LDL of 220 Rn (Bq m−3 )c Volume (cm3 ) Weight (g)

1994 CN 1.65 0.53 1.3 > 5.3d 560 200

1995 CR-39 2.15 0.51 6.4 31 260 200

1996 CR-39 4.17 0.62 2.3 35.6 260 200

2.63 1.32 3.5 11 110 20

a Unit in tracks cm−2 (kBq m−3 h−1 )−1 . b Lower detection limits estimated for an exposure of 90 d. c Estimated for an exposure of 90 d in an environment of 10 Bq m−3 of 222 Rn. d The uncertainty of 222 Rn measurements was not considered.

358

W. Zhuo et al.

Table 2 222 Rn and 220 Rn progeny concentrations measured with the WLx and PISU-A in 4 rooms Room

EECRn (Bq m−3 )

Air conditions

WLx Storage Ofﬁce

Living Rn Chamber

Low air exchange Air conditioner ON Air conditioner ON/OFF Air conditioner OFF Air conditioner ON/OFF Air cleaner OFF Air cleaner OFF Air cleaner ON

47.9 ± 2.74 5.58 ± 2.86 7.72 ± 1.78 9.21 ± 2.88 4.25 ± 1.93 3610 ± 430 5540 ± 659 262 ± 31.6

EECTn (Bq m−3 ) PISU-A

43.2 ± 3.05 5.94 ± 0.45 6.89 ± 0.43 8.76 ± 0.53 4.56 ± 0.39 3360 ± 265 5280 ± 337 403 ± 32.5

WLx

PISU-A

NDa

ND 0.21 ± 0.02 0.37 ± 0.02 0.39 ± 0.02 1.21 ± 0.08 ND ND ND

0.18 ± 0.07 0.38 ± 0.08 0.34 ± 0.07 1.07 ± 0.12 ND ND ND

a Lower than the detection limits.

measurements ran continuously for about 2 days in order to obtain a large number of etched tracks on the CR-39 detectors. The results of the 222 Rn and 220 Rn progeny concentrations measured in different rooms and air conditions are summarized in Table 2. Both EECRn and EECTn measured with the two devices in the same room and air conditions agreed well with each other in most cases. In the condition that the air cleaner and the dehumidiﬁer were switched on, the unattached fraction of PAEC was about 80% in the radon chamber, and the EECRn measured with the PISU-A was much higher than that measured by the WLx. It is considered that a large amount of sampling loss might occur in the sampling system of the WLx, because the gap between the ﬁlter and detector is very narrow in its ﬁlter holder assembly. To examine this possibility, samples were also taken with open-faced ﬁlters at 5 min intervals in the same period. The individual concentrations of 222 Rn progeny were calculated using the Raabe–Wrenn method [20], and the average EECRn was 434 Bq m−3 , similar to the results measured with PISU-A. From the measurements, it was also observed that the concentrations of both 222 Rn and 220 Rn progeny changed with the operating status of the air conditioning. The concentrations were generally higher when the air conditioning was off, and lower when it was on. 3.4. Deposition fraction measurements Measurements of deposition fractions of 222 Rn and 220 Rn progeny in the N and T-B regions were performed in the same 4 rooms as the previous measurements. The results including total particle concentrations measured by PORTACOUNT Plus (Model 8020, TSI, USA) are summarized in Table 3. In ordinary room air conditions, the deposition fractions of 222 Rn and 220 Rn progeny in the N region averaged 5.3 and 3.9%, respectively. The fractions of 222 Rn and 220 Rn progeny in the T-B region were 4.5 and 4.0%, respectively. For the deposition fractions of 222 Rn progeny, our results are similar to the values reported by other researchers [12,16]. For these measurements, the deposition fractions of 222 Rn progeny in the N and T-B regions generally decrease with increasing aerosol concentration. However, variations of the 220 Rn progeny deposition are not clear. The fp of 222 Rn progeny is generally known to be inversely proportional to the aerosol concentration indoors, and most of the unattached progeny are

Integrating measurements of radon and thoron and their deposition fractions in the respiratory tract

359

Table 3 Deposition of 222 Rn and 202 Rn progeny in the nasal and tracheobronchial regions in 4 roomsa Room

Aerosol concentration (×103 cm−3 ) 5.8 ± 2.0 12.0 ± 4.6 21.3 ± 6.1 32.5 ± 8.9 12.2 ± 5.8 2.4 ± 0.7 3.2 ± 0.6 0.4 ± 0.2

Storage Ofﬁce

Living Rn chamber

222 Rn progeny deposition

220 Rn progeny deposition

N (%)

N (%)

T-B (%)

7.3 5.8 4.3 3.1 5.9 7.9 8.4 65.3

T-B (%)

–b 4.4 2.8 3.9 4.3 – – –

6.0 4.6 3.6 3.1 5.2 6.5 7.0 14.8

– 4.5 3.5 3.3 4.6 – – –

a All data are calculated based on the PAEC. b No data.

Table 4 PAEC weighted size distribution of 222 Rn progeny in a radon chamber Run

1 2 3

Unattached progeny

Attached progeny

%

GMD (nm)

GSD

%

GMD (nm)

GSD

77.3 84.7 79.1

1.05 1.00 0.97

1.51 1.49 1.50

22.7 15.3 20.9

172.14 147.56 123.43

2.80 3.23 3.65

likely to deposit in the N and T-B regions according to the diffusion mechanism. For 220 Rn progeny, its ﬁrst decay product 216 Po has an extremely short half-life (0.15 s), but the half-life of the second decay product 212 Pb is relatively long (10.6 h). It is expected that the unattached fractions of 220 Rn progeny are negligible in the general living environment. Therefore, the variations of the 220 Rn progeny deposited in the N and T-B regions may not be signiﬁcant for different aerosol concentrations. Also shown in Table 3, the deposition fractions of 220 Rn progeny in both N and T-B regions are somewhat different from those of 222 Rn progeny. This may be due to the different fp and size distributions of 222 Rn and 220 Rn progeny [21]. The results above suggest that, for precise dose assessment, not only the concentrations but also the deposition fractions in the T-B region should be measured for both 222 Rn and 220 Rn progeny. In order to compare the measured results with theoretical estimations, the PAEC-weighted particle size distributions of 222 Rn progeny were also measured in the radon chamber when the air cleaner and the dehumidiﬁer were on. The graded screen array (GSA) including ﬁve graded screens (30, 145, 200, 400 and 635 mesh) and a back-up glass ﬁber ﬁlter (GF/F) was used in the measurements. The results calculated with the code developed by the Environmental Measurements Laboratory [22] are shown in Table 4. Based on the equation derived by Ramamurthi and Hopke [23], the deposition fractions in the N and T-B regions were calculated

360

W. Zhuo et al.

to be about 63 and 16%, respectively. The theoretical values are in good agreement with our observed values of 65.3 and 14.8%, respectively.

4. Conclusion A commercially available passive 222 Rn monitor was redesigned and rebuilt for both 222 Rn and 220 Rn measurements. The new monitor is not only useful for 220 Rn measurements, but also useful for more precise measurements of 222 Rn. Using the conceptual designs of multiple wire screen samplers proposed by previous researchers, a portable integrating bronchial dosimeter for both 222 Rn and 220 Rn progeny was developed. Its measuring results are in general agreement with other reported values in ordinary room conditions. The main merits of the new bronchial dosimeter and the associated passive integrating monitors are their compactness, portability and low-cost. These devices are expected to be useful for large-scale dose surveys.

References [1] NCRP, Evaluation of occupational and environmental exposures to radon and radon daughters in the United States, NCRP report No. 78, National Council on Radiation Protection and Measurements, Bethesda, MD, 1984. [2] W.W. Nazaroff, A.V. Nero (Eds.), Radon and Its Decay Products in Indoor Air, Wiley, New York, 1988. [3] F. Steinhäusler, Environ. Int. 22 (1996) S1111. [4] W. Zhuo, T. Iida, X. Yang, Radiat. Prot. Dosim. 87 (2000) 137. [5] Q. Guo, J. Sun, W. Zhuo, J. Nucl. Sci. Technol. 37 (2000) 716. [6] M. Doi, S. Kobayashi, Hoken Butsuri 29 (1994) 155. [7] Q. Guo, T. Iida, K. Okamoto, T. Yamasaki, J. Nucl. Sci. Technol. 32 (1995) 794. [8] T. Iida, R. Nurishi, K. Okamoto, Environ. Int. 22 (1996) S641. [9] G. Batterweck, G. Vezzù, Ch. Schuler, et al., Radiat. Prot. Dosim. 94 (2001) 247. [10] K.W. Tu, E.O. Knutson, A.C. George, Aerosol Sci. Technol. 15 (1991) 170. [11] N. Jonassen, B. Jensen, Radiat. Prot. Dosim. 45 (1992) 669. [12] A.C. George, E.O. Knutson, Radiat. Prot. Dosim. 45 (1992) 689. [13] S. Oberstedt, H. Vanmarcke, Radiat. Prot. Dosim. 59 (1995) 285. [14] S.B. Solomon, Radiat. Prot. Dosim. 72 (1997) 31. [15] K.N. Yu, Z.J. Guan, Health Phys. 75 (1998) 147. [16] K.N. Yu, Z.J. Guan, E.C.M. Young, M.J. Stokes, Health Phys. 75 (1998) 153. [17] W. Zhuo, T. Iida, Health Phys. 77 (1999) 584. [18] S. Tokonami, M. Yang, T. Sanada, Health Phys. 80 (2001) 612. [19] S. Tokonami, M. Yang, H. Yonehara, Y. Yamada, Rev. Sci. Instrum. 73 (2001) 69. [20] O.G. Raabe, M.E. Wrenn, Health Phys. 17 (1969) 593. [21] K.W. Tu, E.O. Knutson, A.C. George, Aerosol Sci. Technol. 20 (1994) 266. [22] E.O. Knutson, EML Report 517, USDOE, New York, 1989. [23] M. Ramamurthi, P.K. Hopke, Health Phys. 56 (1989) 189.

361

Determination of 218Po nanometer size distribution in a controlled environment by two new systems A. Vargas a , N. Michielsen b , C. Le Moing c , M. Rio a , G. Tymen c , X. Ortega a a Institut de Tècniques Energètiques, Universitat Politécnica de Catalunya, Avda. Diagonal 647,

08028 Barcelona, Spain b Institut de Radioprotection et de Sûreté Nucléaire, IRSN/DPEA/SERAC, CEA-Saclay, Bâtiment 389,

BP 68, 91192 Gif-sur-Yvette cedex, France c Laboratoire de Recherches Appliquées Atmosphère–Hydrosphère, Université de Bretagne Occidentale,

6 avenue Le Gorgeu, 29285 Brest cedex, France

During the European research programme RARAD (1996–1999), the group formed by the Institut de Radioprotection et de Sûreté Nucléaire (IRSN) and Brest University (UBO), and the group from the Institut de Tècniques Energètiques at the Technical University of Catalonia (INTE-UPC) have developed new measurement systems for the determination of the radon progeny nanometer-size distribution based on diffusion tubes instead of the wire screen technique due to a lack of knowledge of the characterisation of the screen diffusion batteries in this small diameter range. Moreover, during this project, different deconvolution techniques were checked in an intercomparison study. Thus, an intercomparison with these two systems and different deconvolution algorithms were planned under controlled conditions in a large radon chamber. The inﬂuence of radon concentration on size was studied. The data show good agreement in the results with a most probable mean diameter of 0.85 nm; however, the inﬂuence of environmental conditions such as radon concentration and relative humidity cannot be ascertained.

1. Introduction An important parameter affecting the 222 Rn radiological risk is the 218 Po nanometer size distribution [1]. There are indications of three discrete modes depending on atmospheric conditions, such as relative humidity, radon gas concentration and atmospheric trace gas composition. Therefore, the neutralisation rate, which depends on these environmental conditions, could indicate that the electric charge state of 218 Po clusters determines the diffusion equivalent diameter. In the European research project RARAD [2], ﬁeld measurements showed two small modes at approximately 0.55 and 0.85 nm which could be correlated with neutral RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07041-X

© 2005 Elsevier Ltd. All rights reserved.

362

A. Vargas et al.

species and a larger mode at 1.35 nm with charged clusters. On the other hand, measurements carried out in large radon chambers only show a single mode at approximately 0.85 nm and a size mode greater than 1.00 nm has never been found. During the last few years, wire screen diffusion batteries have been the most commonly used method for the determination of this size distribution. Deconvolution of the data obtained with diffusion batteries leads to an ill-posed inversion problem. Direct inversion of ill-posed equations rarely produces acceptable solutions and the development of different inversion algorithms has become hard work for researchers. Published data of this size distribution are quite limited and the results obtained are not deﬁnitive, mainly because of the loss in accuracy due to: (a) the lack of knowledge in the characterisation of the screen diffusion batteries in this small diameter range, (b) the weakness in the mathematical algorithms used in data deconvolution, and (c) the disturbance produced by submicrometric particles within the nanometric size range. In order to diminish these uncertainties, an intercomparison study using two new measurement systems has been carried out in a large radon chamber. The group formed by members of the Institut de Radioprotection et de Sûreté Nucléaire (IRSN) and Brest University (UBO) has developed a diffusion battery based on a new technique using an Annular Diffusion Channel (ADC). The technique used by the group at the Institut de Tècniques Energètiques of the Technical University of Catalonia (INTE-UPC) is the Cylindrical Diffusion Tube (CDT). Before each exposure, submicrometer particles were cleaned in the radon chamber in order to avoid uncertainty (c). Extreme Value Estimation (EVE), Random Walk (RW) and Expectation Maximisation (EMAX) data deconvolution algorithms were used in the intercomparison to analyse the sensitivity of the technique. An intercomparison of these and other deconvolution techniques with simulated data have been carried out by Butterweck et al. [3].

2. Materials and methods In this section, a description of the two new measurement systems of the 218 Po nanometersize distribution and the data deconvolution algorithms used is presented. Afterwards, the environmental conditions in the radon chamber during the different experiments in the intercomparison study are described. 2.1. Measurement systems A new technique using an Annular Diffusion Channel (ADC), which allows continuous measurements, was designed by the ISRN-UBO group [4]. The ADC diffusion battery consists of six units: ﬁve ADCs of different length plus a reference unit, all operated in parallel (Fig. 1). The sampled air is drawn through the ADC where diffusive particles are deposited; the remaining particles are collected onto a membrane ﬁlter (Poretics, polycarbonate, 0.8 μm). The alpha particles emitted by the 218 Po and 214 Po collected, or formed on the ﬁlter, are detected by an alpha PIPS detector (Canberra 450) placed in the inner tube of the ADC opposite the ﬁlter.

Determination of 218 Po nanometer size distribution in a controlled environment by two new systems

363

Fig. 1. Top: schematic diagram of one ADC and the reference unit. Bottom: schematic diagram of one diffusion tube and the reference unit, together with typical alpha spectra.

A Cylindrical Diffusion Tube (CDT), which also allows continuous measurements, was designed by the INTE-UPC group [5]. It consists of six units: ﬁve of them include 2-cm-diameter glass cylindrical diffusion tubes of different length and the remaining unit is the reference, all operated in parallel. In order to minimise the error in the particle penetration curves, the CDT units were designed with the ﬁlter directly mounted at the end of the tube. A very thin polycarbonate ﬁlter Millipore 1.2 μm type RTTP has been chosen to allow alpha particles to pass through it. Therefore, the alpha particles can reach the PIPS detector (Canberra 300) located at the back of the ﬁlter. A scheme of the reference unit and one unit with a diffusion tube are shown in Fig. 1. Each unit of the diffusion battery systems is characterised by the particle penetration curve and the radioactive efﬁciency. In both systems, radioactive efﬁciency is easily determined by numerical or semi-analytical methods.

364

A. Vargas et al.

Table 1 Operating parameters of the ADC and CDT diffusion batteries Unit

1 2 3 4 5 ref.

ADC (IRSN-UBO)

CDT (INTE-UPC)

ﬂow rate (l min−1 )

length (cm)

d50 (nm)

ﬂow rate (l min−1 )

length (cm)

d50 (nm)

12.8 12.4 12.4 12.5 5.3 11.7

2.5 4.5 10.0 25.0 30.0 –

0.47 0.69 1.30 2.10 3.70 –

4.6 4.6 4.5 2.2 2.2 4.6

30 50 70 50 70 –

0.50 0.72 0.91 1.17 1.45 –

Fig. 2. On the left: penetration curves for the ADC system. On the right: penetration curves for the CDT.

On the other hand, penetration curves were not so easily determined since it is necessary to take into account the developing ﬂow at the entrance, the inﬂuence of the ﬁlter, in-ﬂight formation, removal by decay, axial diffusion and non-uniform particle deposition on the ﬁlter. Table 1 shows the operating parameters for both systems, where d50 is the diameter of the particle for a 50% collection efﬁciency. The penetration curves are shown in Fig. 2. A measurement cycle takes 3 hours with 20 minutes air sampling at the start of each cycle. 2.2. Deconvolution algorithms Three different data deconvolution algorithms have been used in order to analyse their inﬂuence on 218 Po size distribution. The EVE [6] method is used by the IRSN-UBO group and basically consists in ﬁnding a set of possible solutions included in a conﬁdence interval. The method is based on the minimisation of the χ 2 function. The main advantages of this method are that it takes into account the measurement uncertainties and that it gives conﬁdence intervals. The version 11.0 of January 3rd of Pentti Paatero was used. The window of interest was

Determination of 218 Po nanometer size distribution in a controlled environment by two new systems

365

taken from 0.1 to 10 nm, a log-normal distribution was assumed and 50 mean diameter were taken equidistant on a log scale with geometric standard deviation of 1.2. The RW [7] is a method which probes every combination of size distribution parameters within the size range interest to obtain a bimodal distribution. An empirical condition is introduced which eliminates cells from the parameter space. The advantage of the RW method is that the iteration is based on eliminating areas with large error functional instead of trying to locate a better set of parameters in the vicinity of the actual best ﬁt. The disadvantage of this method is the assumption of log-normal distribution and a maximum of two distributions. The RW algorithm used in the intercomparison was developed in FORTRAN by the INTE group. The mean diameter range of interest was from 0.1 to 6 nm and the geometric standard deviation from 1.00 to 2.60. The EMAX [8] algorithm yields iterative solutions from the approximation of the general reconstruction integral using the division of the size distribution into discrete reconstruction intervals within the size range of interest. The EMAX algorithm used in the intercomparison was implemented in FORTRAN by the INTE group. The number of reconstruction intervals was chosen to be 16 with midpoint diameters in geometric progression within the size range from 0.4 to 1.8 nm. The ﬁt criterion used in the algorithm was the weighted least square (WLS) parameter. The iterations are repeated until the WLS is lower than 10−5 or a maximum of 200 000 iterations are done. 2.3. Environmental conditions The intercomparison was carried out in a 20 m3 sealed chamber at the INTE-UPC in Barcelona [9]. Radon gas concentration inside the chamber was measured every 10 minutes by an ATMOS 12 DPX from the Gammadata Company. The radon activity was generated from a solid 2100 kBq 226 Ra source from the Pylon Company. When necessary, a pump with

Fig. 3. Temporal evolution of radon concentration and relative humidity during the intercomparison exercise.

366

A. Vargas et al.

adjustable frequency controlled the ventilation rate in the chamber. Particle concentration was measured by a condensation nuclei counter CNC 3025 from the TSI Company. In order to ﬁnd out if the level of the radon concentration affects the 218 Po nanometersize distribution, two experiments at different radon concentration levels were planned. In the ﬁrst experiment, seven measurements were carried out at a constant radon concentration of 5.5 kBq m−3 . In the second, six measurements at 2 kBq m−3 were done. Figure 3 shows the temporal evolution of radon concentration and relative humidity inside the chamber. Since particle concentration should be maintained as low as possible in order to avoid the disturbance produced by submicrometric particles within the nanometric size range, in both experiments the radon chamber air continuously circulated in a closed system through a HEPA ﬁlter. Moreover, the air conditioning system of the radon chamber was not used during experiments since it can generate particles. Thereby, particle concentration was lower than 20 part cm−3 .

3. Results and discussion Figure 4 shows the mean value and its standard uncertainty estimated from the seven experimental penetration data at 5.5 kBq m−3 and the six experimental data at 2 kBq m−3 radon concentration levels. Both systems show the same results for units 1 and 2 since they have almost the same penetration curves indicating a good agreement. The experimental penetration data from unit 3 of the ADC system for the two radon concentration levels are signiﬁcant different by about 20%. It is interesting to point out that the experimental penetration data are always lower for the higher concentration in both systems. Finally, data from units 4 and 5 of the ADC system conﬁrm that particle size greater than 1 nm are negligible.

Fig. 4. Penetration data mean value and standard uncertainty for each unit and radon concentration level. On the top for the ADC system and on the bottom for the CDT.

Determination of 218 Po nanometer size distribution in a controlled environment by two new systems

367

Fig. 5. Size distribution results obtained from the three deconvolution algorithms: EVE (top), RW (centre) and EMAX (bottom) at two concentration levels, 2 (left) and 5 kBq m−3 (right) for the ADC measurement system.

368

A. Vargas et al.

Fig. 6. Size distribution results obtained from the three deconvolution algorithms: EVE (top), RW (centre) and EMAX (bottom) at two concentration levels, 2 (left) and 5 kBq m−3 (right) for the CDT measurement system.

Determination of 218 Po nanometer size distribution in a controlled environment by two new systems

369

Table 2 Comparison of the log-normal distribution of the 218 Po nanometer size obtained with the three inversion algorithms Group

IRSN-UBO INTE-UPC

Radon level kBq m−3

Estimated distribution (mean, geometric deviation) EVE

RW

EMAX

2 5 2 5

0.86 ± 1.26 0.84 ± 1.25 0.85 ± 1.20 0.84 ± 1.20

0.86 ± 1.58 0.69 ± 1.78 0.84 ± 1.00 0.83 ± 1.00

0.89 ± 1.07 0.76 ± 1.08 0.82 ± 1.10 0.82 ± 1.09

Figures 5 and 6 show the 218 Po nanometer size distribution for the ADC and CDT systems, respectively. For the EVE algorithm the conﬁdence interval for each situation is graphically presented. The RW method directly shows the parameters of the log-normal distributions and for the EMAX an estimate value for each interval diameter was obtained. From these ﬁgures, it can be observed that the size distributions ﬁt quite well in all cases to a log-normal function. In some cases a very small diameter, which is physically impossible can be seen. Table 2 shows the log-normal distributions of the 218 Po nanometer size ﬁtted with the commercial software GRAPHER. The size distributions obtained with the CDT system are quite similar for both radon concentration and deconvolution systems. However, when the RW and EMAX algorithms are used, the mean diameter obtained with the ADC system is signiﬁcantly lower than for the EVE in the 5.5 kBq m−3 radon concentration level. Moreover it can be appreciated that the distribution obtained with the RW is too wide for a real interpretation since the physical impossibility to have lower diameters than about 0.3 nm. From these results, it is not possible to conclude if there is any difference in the size diameter due to the effect of radon concentration or relative humidity.

4. Conclusions An intercomparison with two new diffusion battery systems based on a new technique using an Annular Diffusion Channel for one Laboratory and a Cylindrical Diffusion Tube for the other, was carried out under controlled conditions in a large radon chamber. The results from the two systems are in good agreement and a log-normal distribution of 0.85 nm mean equivalent diffusion diameter and a geometric standard deviation less than 1.3 has been obtained as the most probable distribution. However, it can be observed that the experimental penetration data obtained from the two diffusion batteries is always slightly lower when radon concentration and relative humidity are both high. The estimated distributions using three different deconvolution algorithms does not clearly show a difference in the size distribution indicating a low sensitivity of these inversion systems. Therefore, it would be interesting to carry out a set of measurements in more extreme conditions, such as in the range from 10 to 90% for relative humidity and from 2 to 80 kBq m−3 in order to analyse the inﬂuence of radon concentration level and relative humidity to the 218 Po nanometer size distribution.

370

A. Vargas et al.

References [1] J.W. Marsh, A. Birchall, Sensitivity analysis of the weighted equivalent lung dose per unit exposure from radon progeny, Radiat. Prot. Dosim. 87 (3) (2000) 167–178. [2] G. Monchaux, Final report of the CEC Contract F14P-CT-95-0025, Risk Assessment of Exposure to Radon Decay Products (RARAD), Ed. Direction de l’information scientiﬁque et technique, Report CEA-R-5882(2), 1999. [3] G. Butterweck-Dempewolf, Ch. Schuler, G. Vezzú, A. Reineking, C. Huet, G. Tymen, J.C. Strong, E.O. Knutson, A. Vargas, Intercomparison of approximation algorithms for the determination of the size distribution of the unattached fraction of radon progeny, Aerosol Sci. Technol. 33 (2000) 261–273. [4] N. Michielsen, V. Voisin, G. Tymen, Measurement of the unattached radon decay products with an annular diffusion channel battery, in: J.P. McLaughlin, S.E. Simopoulos, F. Steinhäusler (Eds.), The Natural Radiation Environment VII, Proc. VIIth International Symposium on the NRE, Rhodes, Greece, 20–24 May 2002, in: Radioactivity in the Environment, vol. 7, Elsevier, Amsterdam, 2004, this volume. [5] A. Vargas, M. Rio, X. Ortega, Set up of a diffusion tube battery to measure continuously the radon progeny nanometric size distribution, in: Abstracts of the VII Congress of Natural Radiation Environment (NRE-VII), Rhodes, Greece, 2002. [6] P.K. Hopke, P. Paatero, Extreme value estimation applied to aerosol size distributions and related environmental problems, J. Res. Natl. Inst. Stand. Technol. 99 (1994) 361–366. [7] G. Butterweck-Dempewolf, Ch. Schuler, G. Vezzú, A. Reineking, Improved determination of bimodal size distribution from measurements with diffusional size classiﬁcation, Aerosol Sci. Technol. 31 (1999) 383–391. [8] E.F. Maher, N.M. Laird, EM algorithm reconstruction of particle size distributions from diffusion battery data, J. Aerosol Sci. 16 (6) (1985) 557–570. [9] A. Vargas, X. Ortega, M. Rio, J.L. Martín Matarranz, The Spanish radon Reference Chamber, in: Abstracts of the VIIth International Symposium on Natural Radiation Environment (NRE-VII), Rhodes, Greece, 2002.

371

Measurements and modelling of combined diffusive and advective radon transport in porous building materials M. van der Pal a , W.H. van der Spoel a,b , R.J. de Meijer c,d , N.A. Hendriks a , E.R. van der Graaf c a Faculty of Architecture, Building and Planning, Group Building Physics, Eindhoven University of Technology,

PO Box 513, 5600 MB Eindhoven, The Netherlands b Faculty of Civil Engineering and Geosciences, Section Building Engineering, Delft University of Technology,

Stevinweg 1, 2628 CN Delft, The Netherlands c Kernfysisch Versneller Instituut, Nuclear Geophysics Division, Zernikelaan 25, 9747 AA Groningen,

The Netherlands d Faculty of Physics, Group Applications of Ion Beams, Eindhoven University of Technology, PO Box 513,

5600 MB Eindhoven, The Netherlands

Measurements of combined diffusive and advective radon transport in aerated concrete have been made and compared with model calculations. The experimental set-up consists of a leaktight vessel containing a hollow aerated-concrete cylinder that divides the vessel into an inner (inside the cylinder) and outer (outside the cylinder) compartment [1]. The measurements were conducted with a radon source in the inner compartment, for a range of ventilation rates in the inner and outer compartment and for various ﬂows through the concrete. Pressure, temperature and relative humidity in the set-up were kept constant. It was found that a combined diffusion–advective ﬂow (Fick–Darcy) model does not describe the data. A much better ﬁt is obtained by assuming an advective-velocity-dependent diffusivity in the model calculations. The dependence of the diffusivity on the ﬂow velocity may result from mechanical and/or hydrodynamic dispersion. Alternatively, a model assuming two parallel porous media with distinct volume and permeability also describes the data quite well. The results of this parallel-media model indicate a preferential airﬂow. About 90% of the airﬂow is transported through 7% of the material. 1. Introduction Soil and building materials are considered the main contributors to radon (222 Rn) entry in dwellings. To calculate radon concentrations in dwellings, a detailed and accurate model deRADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07042-1

© 2005 Elsevier Ltd. All rights reserved.

372

M. van der Pal et al.

scribing production and transport of radon in those materials is required. Various parameters inﬂuencing the production and transport of radon include, among others, radium content, porosity, moisture content, adsorption and differences in concentration and pressure. For sand, a model that includes those parameters has been validated. Van der Spoel [2–4] showed that radon concentrations in crawl spaces can be well calculated based on the mentioned parameters. The deviations between model calculations and measurements of diffusive and advective transport of radon were within 15% for dry sand and between 15 and 40% for (partly) wetted sand. The model represents sand as a homogenous material and describes diffusive and advective ﬂows using Fick’s and Darcy’s law, respectively. Research on production and transport of radon in building materials remains limited and results are often contradictive. An extensive radon survey in the Netherlands by RIVM (National Institute of Public Health and the Environment) including more than 1000 dwellings showed that 60 to 70% of the average 29 Bq m−3 radon concentration in Dutch dwellings is due to radon produced by building materials [5]. On the basis of the average radon concentration and the measured ventilation rate in the dwellings, the calculated in-situ radon-exhalation rate of the building materials was found to be 4 to 5 times higher than the measured exhalation rate in various laboratories for similar building materials. A probable cause for this difference between in-situ and laboratory results is that building materials in dwellings are exposed to weather conditions such a varying temperature, humidity and pressure (differences) where the building materials in the laboratories were measured under static conditions. To gain more insight, research into more complex transport processes such as combined diffusive and advective transport in building materials is desired. This paper describes measurements of diffusive and advective transport of radon in autoclaved aerated concrete (AAC) under welldeﬁned conditions. The measurement results are compared with three radon transport models, all based on the laws of Fick and Darcy. 2. Materials and methods 2.1. Experimental set-up The experimental set-up, shown in Fig. 1, consists of a 1-m-diameter stainless-steel cylinder with air inlets and outlets. In the set-up, a hollow autoclaved aerated-concrete (AAC) cylinder with 0.6 m inner and 0.8 m outer diameter is placed on the bottom. The top of the aeratedconcrete cylinder is closed with a stainless-steel lid. AAC was chosen as building material because its fast curing excludes ageing effects and it is assumed quite homogeneous and isotropic. The aerated-concrete cylinder was milled from a 1-m-side cube, made by Ytong (Meppel) in The Netherlands. The dimensions of the set-up are given in Table 1. The aerated concrete cylinder is glued to its steel lid (inside) and the inner ring below. This way, two compartments are created in the set-up, which are separated by the aerated concrete cylinder. In addition, radon transport between the two compartments through the aerated concrete cylinder is reduced to one dimension: radial transport only. The glue connection has been checked for leaks by measuring the diffusivity of a sample of AAC as function of sample thickness relative to glue thickness. As the diffusivity was related to the sample thickness only, it was concluded that there was no leak.

Measurements and modelling of combined diffusive and advective radon transport

373

Fig. 1. Schematic cross-section of the radon cylinder.

Table 1 Dimensions of set-up and AAC cylinder Property

Value

Total volume set-up Volume inner compartment Volume outer compartment Height aerated concrete cylinder Inner diameter aerated concrete cylinder Outer diameter aerated concrete cylinder Volume aerated concrete cylinder Porosity AAC

0.887 ± 0.005 m3 0.224 ± 0.003 m3 0.483 ± 0.006 m3 0.743 ± 0.002 m 0.606 ± 0.002 m 0.798 ± 0.002 m 0.157 ± 0.003 m3 0.778 ± 0.013

Relative humidity, ventilation rate, pressure and airﬂow through the aerated concrete in the radon cylinder are controlled by a series of mass-ﬂow controllers (MFC). Radon concentrations are measured with two quasi-continuous radon meters with a time resolution of 30 minutes [6], one for each compartment. A more detailed description of the set-up can be found in Van der Pal et al. [1]. 2.2. Experimental procedures To measure the effects of (combined) diffusive and advective radon transport through AAC, the following experiment has been conducted: A Pylon type 2000A radon source with a nominal activity of 22.2 kBq 226 Ra is placed in a ﬁve-litre stainless steel container that is connected with the inner compartment of the set-up via tubes. The air in the container is refreshed at a

374

M. van der Pal et al.

rate of 2 litres per minute via a continuous operating pump to provide good mixing of the generated radon throughout the inner compartment. This way a high radon concentration is created in the inner compartment compared to the outer compartment and results in a diffusive ﬂow from the inner to outer compartment. Air with a relative humidity of 50% is introduced in the outer compartment of the set-up (Flow in) and leaves the set-up via the inner and/or outer compartment (Flow out inner and Flow out outer). These three ﬂows are controlled by a series of mass-ﬂow controllers. As a fraction of the Flow in leaves the set-up from the inner compartment, an airﬂow through the aerated concrete is induced, resulting in advective transport of radon in opposite direction to its diffusive transport. A schematic overview of the experimental set-up is shown in Fig. 2. The experiment starts with a Flow in of 3 L min−1 and is increased in steps of 3 L min−1 up to 15 L min−1 . Each value of Flow in is maintained during ﬁve days. The fraction of air leaving the set-up via the inner compartment is increased from 0 to 0.8 in steps of 0.2. Each step is maintained during 24 hours. The experiment takes 25 days to complete. An overview of the ﬂow values is given in Table 2. The pressure in the vessel is kept at a constant value of 1060 h Pa by adjusting the setting of the Flow out outer. The temperature in the room was set at 21.5 ◦ C and controlled by an air-conditioning unit. During the measurement, airﬂows, temperature, pressure, pressure difference, relative humidity and the radon concentrations in both compartments were recorded and stored on a PC.

Fig. 2. Schematic overview of the experimental set-up.

Table 2 Flow settings Flow In (L min−1 )

Fraction of Flow In through AAC 0

0.2

0.4

0.6

0.8

3 6 9 12 15

day 1 day 6 day 11 day 16 day 21

day 2 day 7 day 12 day 17 day 22

day 3 day 8 day 13 day 18 day 23

day 4 day 9 day 14 day 19 day 24

day 5 day 10 day 15 day 20 day 25

Measurements and modelling of combined diffusive and advective radon transport

375

3. Experimental results Figure 3 shows the measured radon concentrations as a function of time. The “peaks” in the radon concentration of the inner compartment indicate the intervals with diffusion only (fraction of ﬂow through AAC is zero). Also the 24-hour intervals with combined diffusive and advective transport are clearly visible. The pattern meets expectations: high radon concen-

Fig. 3. Measured radon concentrations in the inner and outer compartment as a function of time.

Fig. 4. Flow through AAC as a function of pressure difference over AAC.

376

M. van der Pal et al.

trations in the inner compartment for diffusive transport where for combined advective and diffusive transport decreasing radon concentrations are measured. Figure 4 shows the ﬂow through the AAC as a function of the pressure difference between the inner and outer compartment. The ﬂow through the AAC has been calculated as the outgoing air of the inner compartment (Flow out inner) as well as the difference in ingoing and outgoing ﬂow of the outer compartment (Flow in – Flow out outer). It can be seen from Fig. 4 that the relation between pressure difference and ﬂow is not linear, indicating higher air-velocities in the pores than expected. 4. Model calculations The measurements are compared with three types of model calculation, all based on a macroscopic description for radon transport. 4.1. Macroscopic description Rogers and Nielson [7] developed a macroscopic formalism to describe diffusive and advective multi-phase transport of radon in porous materials. This formalism includes four processes that determine the pore–air radon concentration: diffusion, advective ﬂow, radon production and radon decay. The time-dependent partial differential equation, assuming incompressible air, can be written as: β

∂Ca K = ∇(βDe ∇Ca ) + ∇Pa · ∇Ca − βλCa + S, ∂t μ

(1)

where: β = partition-corrected porosity, Ca = radon concentration in air-phase (Bq m−3 ), De = effective diffusivity (in m2 s−1 ), K = intrinsic permeability (m2 ), μ = dynamic viscosity of air (Pa s), Pa = air pressure (Pa), λ = decay constant of radon (= 2.1 × 10−6 s−1 ), S = radon production rate per unit bulk volume (Bq m−3 s−1 ). Effects of water-ﬁlled pores and adsorption of radon to solid surfaces are accounted for via the multiphase-corrected porosity, β: β = εa + Lεw + ρb ka where: εa = air-ﬁlled porosity, εw = water-ﬁlled porosity, L = Ostwald coefﬁcient for radon (0.26 at 293 K), ka = surface adsorption coefﬁcient for radon (m3 kg−1 ), ρ = density of the porous medium (kg m−3 ).

(2)

Measurements and modelling of combined diffusive and advective radon transport

377

For the 50% RH, adsorption and water-ﬁlled porosity are negligible so β = εa . For solving equation (1), a numerical ﬁnite-difference model has been developed and validated [4]. 4.2. Model calculations The following models are considered: 1. Classic description. This model describes the transport in AAC based on equation (1) with constant diffusivity D. Calculations are carried out using the numerical model. 2. Flow-dependent diffusivity (FDD) description. Here the effect of dispersion is simulated using equation (1) and an advective-velocity dependent diffusivity D(ν). Also these calculations are based on the numerical model. 3. Parallel-media model. A schematic representation of this model is given in Fig. 7. In the parallel-media model, radon is transported from the outer to the inner compartment via two different media. Each medium has its own permeability, diffusivity (in this paper assumed to be the same for both media to reduce the number of ﬁt-parameters) and volume. It is also assumed that there is no exchange of radon between the two media. Advective and diffusive transport of radon takes place according to the macroscopic formalism described above. For this model, steady-state analytical calculations were performed. The ﬁrst two models require input values for various material parameters. The porosity and the adsorption rate of the AAC were determined independently. The porosity of AAC was measured with the gas-displacement method [8]. The adsorption of radon by AAC was determined together with the radon exhalation rate of AAC in a series of ingrowth measurements [9]. The diffusivity (Classic model) was determined from the measurements without advective ﬂow. These measured input parameters are given in Table 3. In the FDD-model, a least-squares regression ﬁt was conducted to ﬁnd the best estimate for the diffusivity De for each ﬂow velocity (ν) through the aerated concrete cylinder. The ﬁtted diffusivity values as a function of the ﬂow velocity are shown in Fig. 5. The measured and modelled radon concentration using these two models are shown in Fig. 6. This ﬁgure shows that the calculated concentrations in the outer compartment using the classic model are much lower than the measured concentrations. As these differences exceed the uncertainties in both model calculations and measurements, it is concluded that the model is not an accurate representation of the real situation. The results of the FDD model show that by modelling dispersion effects as shown in Fig. 5 a good match with the data is obtained. Although the increase in diffusivity with increasing ﬂow-rates is described in literature by both theoretical models as measurements [10], the ﬁtted dispersion effects are Table 3 Input parameters modelling Parameter

Value

K β De

(0.1 ± 0.6) × 10−4 m3 kg−1 (0.788 ± 0.013) m3 m−3 (0.63 ± 0.02) × 10−6 m2 s−1

378

M. van der Pal et al.

Fig. 5. Bulk diffusivity as a function of the ﬂow velocity through the AAC.

Fig. 6. Radon concentrations in the outer compartment as a function of time.

much larger than expected. A possible explanation for this mismatch may be that large pores are connected with each other via small pores creating high air velocities in the small pores. It is however not clear whether such high velocities occur in the AAC. In the third model, the AAC is schematically divided in two separate media. Each medium has three unknown parameters: volume, permeability and diffusivity; thus 6 unknown parameters in total. The number of degrees of freedom is however less. A ﬁrst restriction is that the sum of the volumes must equal the total volume of the AAC. Secondly, the combined “effective” diffusivity of the two media must ﬁt the measured diffusivity. For simplicity, we assume the same diffusivity for both media. The third restriction is that the sum of the ﬂows

Measurements and modelling of combined diffusive and advective radon transport

379

through the two media must equal the airﬂow through the AAC. These constraints reduce the number of degrees of freedom of this model to two: the volume ratio and the permeability ratio. The results of this model were compared with steady-state concentrations as they would be found with the second model. A comparison with measurements was not possible since steady-state conditions were not attained in each 24 h interval. The ﬁtted values for the volume and permeability ratio are given in Table 4. These indicate one relative large compartment with a relative small permeability together with a small compartment with a relative Table 4 Results of the parallel-media model Parameter

Ratio

Volume Permeability

0.08 ± 0.03 153 ± 40

Fig. 7. Schematic representation of the parallel-media model.

Fig. 8. Equilibrium radon concentrations extrapolated from measured values vs. radon concentrations modelled with parallel-media model.

380

M. van der Pal et al.

high permeability. According to this model, approximately 90% of the airﬂow passes through 7% of the volume. In Fig. 8, the equilibrium radon concentrations as found by extrapolation of measured data using the second model and radon concentrations modelled by the parallelmedia model are shown. The parallel-media model describes the data just as well as the FDD model. Unlike the FDD-model, a characteristic of this model is a large ﬂow-velocity (for medium B). This results in larger Reynolds numbers and therefore the measured non-linearity between ﬂow and pressure difference is more likely. However, if the length of the cracks is limited to a few centimetres, the effect of preferential paths will be limited and the assumption of (little or) no interaction between the compartments invalid. For a further validation of this model, the transport parameters such as permeability and diffusivity of the preferential ﬂow paths should be measured independently.

5. Discussion and conclusions The conclusion drawn from the measurements is that a simple radon transport model cannot describe the data. A better description of these data is found when including ﬂow dependent diffusivity or inhomogeneity (resembled by the parallel-media model) in the model calculations. Future experiments will focus on distinguishing these two effects. Such experiments may include dyeing experiments to show possible preferential ﬂow paths, combined advective and diffusive transport measurements with ﬂow and diffusion in the same direction and measurement and modelling of the pore structure of AAC. Whether or not the described effects occur in dwellings under normal circumstances will depend on the building materials used and their properties for permeability, porosity and poresize distribution and on the weather conditions the dwelling is exposed to. However, if these effects occur, the radon exhalation rates of building materials or soil may be larger than expected from calculations using a diffusivity measured under no-ﬂow conditions.

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10]

M. van der Pal, et al., Sci. Total Environ. 272 (2001) 315. W.H. van der Spoel, E.R. van der Graaf, R.J. de Meijer, Health Phys. 72 (5) (1997) 765. W.H. van der Spoel, E.R. van der Graaf, R.J. de Meijer, Health Phys. 74 (1) (1998) 48. W.H. van der Spoel, Radon transport in sand, PhD thesis, TU Eindhoven, Eindhoven, The Netherlands, 1998. P. Stoop, P. Glastra, Y. Hiemstra, L. de Vries, J. Lembrechts, Results of the second national survey on radon in dwellings, RIVM report 610058006, 1998. F.J. Aldenkamp, P. Stoop, Sources and transport of indoor 226 Radon – measurements and mechanisms, PhD thesis, Rijksuniversiteit Groningen, The Netherlands, 1994. V.C. Rogers, K.K. Nielson, Health Phys. 60 (1991) 807. E.R. van der Graaf, I. Cozmuta, M. van der Pal, Calibration of the NGD-KVI Porosimeter and results of initial experiments on aerated concrete samples from TUE, KVI-report S75, 2000. M. van der Pal, W.H. van der Spoel, A simple method for measuring radon adsorption in porous materials, in: Proc. Workshop Radon in the Living Environment, Liege, Belgium, 2001. M. Verlaan, Dispersion in heterogeneous media, PhD thesis, Delft University of Technology, Delft, The Netherlands, 2001.

381

Czech study on lung cancer risk from residential radon A. Heribanová a , L. Tomášek b a State Ofﬁce for Nuclear Safety, Senovážné nám. 9, 100 00 Prague, Czech Republic b National Radiation Protection Institute, Šrobárova 48, 100 00 Prague, Czech Republic

Lung cancer risk from residential radon can be extrapolated from studies of miners. Direct evidence in the general population based on a cohort study of 12 000 subjects is reported. A total of 218 cases of lung cancer were observed by 1999. The excess relative risk depends linearly on average radon concentration in the past 5–34 year period. The estimated excess relative risk per 100 Bq m−3 is 8%. The effect from exposures in the distant past (before 34 years) is lower by one order in comparison to the period 5–34 years. The combined effect of smoking and radon progeny analyzed in the study is consistent with a sub-multiplicative interaction: the excess relative risk per unit exposure among non-smokers was roughly twice in comparison to smokers. In comparison, the risk of lung cancer from occupational and residential radon shows considerable agreement, both in temporal patterns and smoking interaction. 1. Introduction According to the International Agency for Research on Cancer, there is signiﬁcant evidence to classify radon as a carcinogen. Using extrapolations from occupational studies, it can be shown that for some countries environmental exposure to radon is the second most important cause of lung cancer in the general population after cigarette smoking. The Czech study on lung cancer and residential radon, established in 1989 by Josef Ševc, aims to contribute to knowledge on the risk from radon in the general population, particularly by evaluating temporal factors and the interaction of radon exposure and smoking. 2. Methods 2.1. Deﬁnition of cohorts The study was designed as a retrospective follow-up covering the period since 1960. The study area – the Middle Bohemian Pluton – is mostly granitoid with considerable breaks. The RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07043-3

© 2005 Elsevier Ltd. All rights reserved.

382

A. Heribanová, L. Tomášek

Fig. 1. Arithmetic means of radon progeny concentrations in houses by administrative units of the Czech Republic.

area of the study covers 241 km2 (0.3% of the country). The levels of radon concentration in the selected area are considerably higher than the mean for the country (Fig. 1). The study population includes inhabitants of the area (80 villages) who had resided there for at least 3 years, who were alive by 1960 or were born later. The collected data included date of birth, past residences, smoking habits, occupation, and housing characteristics. Data on 12 003 subjects were collected by trained interviewers who also installed radon detectors. A total of 201 people were excluded from the present analyses for reasons of unsatisfactory personal data, e.g., inability to identify them in the national population registry or for large gaps in their residence histories. 2.2. Follow-up Information on vital status and causes of death were obtained from the Czech population registry and diagnoses from registries of deaths at local administrative ofﬁces. Follow-up for each subject started at the latest of the following dates: three years after ﬁrst year of recorded residence in the study area or 1 January 1961. The follow-up ended at the earliest of date of death, emigration, 85th birthday or 1 January 2000. The condition of excluding person–years of people older than 85 was used in order to eliminate the uncertainties in causes of death and also to stay clear of potential errors in the population registry. Observed numbers of deaths were compared to expected numbers of deaths that were derived from annual publications of the Czech Statistical Ofﬁce by stratifying on age, gender, and calendar year. 2.3. Estimation of exposure to radon and tobacco smoke The exposure assessment in the cohort was based on measurements of equivalent equilibrium concentrations of radon (radon progeny) in most houses (80%) of the study area. During the period 1991–1992, two integral detectors (Kodak LR115) were usually installed for one year

Czech study on lung cancer risk from residential radon

383

Table 1 Strata for estimation of radon concentrations in the study PY † Direct Commune mean∗ Greater commune mean∗ District mean∗ Total

Measurement in the study area 5 391 042 748 899 Mapping of the Czech Republic 615 426 604 264 7 359 631

%

Bq m−3

74 10

508 542

8 8 100

282 164 463

Distribution of PY of exposure and mean concentrations in the strata. ∗ Missing data replaced by respective means. † PY of relevant residences in the period 5–34 years prior to each year of follow-up (1961–1999).

in the two most occupied rooms of the house. In order to compare results to other residential studies, which are related to radon gas rather than radon progeny, it was necessary to establish the conversion factor between these two quantities. This was done on the basis of 652 simultaneous measurements of radon progeny by passive track detectors and radon gas by electrets. All results are thus given in terms of radon gas concentrations. Exposure estimates in residences outside the area were derived from a large scale mapping of radon in the country (Fig. 1). For houses in the study area that were not available for measuring, the community means were used instead of missing values. Concentrations corresponding to residences outside the study area (16% of respective residence person–years) were estimated by larger community means for inhabitants in the neighboring four districts and by district means for the residences in other districts, where concentrations were usually much lower (Table 1). 2.4. Statistical analyses The statistical analyses were based on the relative risk model. The observed numbers (O) of lung cancer were assumed to have the Poisson distribution: O∼ (1) = iE 1 + b(C − C0 ) , where E is the number of cases expected from national mortality data, parameter i is an intercept term that allows the background mortality rate for the ‘unexposed’ cohort to differ from that in the general population. The variable C is the time-weighted average concentration of radon corresponding to the past 5–34 years, C0 is the background concentration (country mean), and parameter b denotes the coefﬁcient of relative risk per unit radon concentration. The evaluation of temporal modiﬁcation was based on the model: O∼ = iE(1 + b1 W1 + b2 W2 ),

(2)

where W1 and W2 represent the exposures cumulated during the past periods, 5–19 and 20–34 years or 5–34 and 35–49 years. Parameters b1 and b2 denote time-speciﬁc risk coefﬁcients. In order to express one-sided statistical tests, the conﬁdence intervals related to risk coefﬁcients are given in terms of 90%.

384

A. Heribanová, L. Tomášek

3. Results During the follow-up of 297 901 person–years, a total of 4012 subjects (34%) of the cohort died, 390 of them over 85 in age, and 17 subjects emigrated. A total of 218 lung cancers were observed (Table 2). The present ﬁgures suggest increased mortality from lung cancer (O/E = 1.15) in contrast to generally low numbers corresponding to cancers other than lung (O/E = 0.87). The relative risk of lung cancer in relation to average radon concentration in the past 5– 34 years shows a linear trend (Fig. 2). The estimate of excess relative risk (ERR) per unit radon concentration (100 Bq m−3 ) was 0.083 (90%CI: 0.014–0.199). This value did not substantially change when calendar period, gender or smoking was adjusted for. The smoking adjusted estimate was 0.078, but the risk coefﬁcient for non-smokers (0.130) was higher in Table 2 Speciﬁc mortality in the Czech residential cohort by 1999

All causes Lung cancers Other cancers Other diseases§ Violent deaths

O∗

O/E †

3622 218 600 2576 195

0.95 1.15 0.87 0.98 0.75

RR‡ 1.35

∗ O: number of observed cases (age below 85). † O/E: ratio of observed numbers to numbers expected

from national mortality data. ‡ RR: relative risk – ratio of observed cases to numbers expected at zero exposure in the cohort. § Causes other than cancers and violent deaths.

Fig. 2. Relative risk of lung cancer (RR) by average concentration (C) of radon (Bq m−3 ); dashed line – model from meta-analysis of 8 indoor studies.

Czech study on lung cancer risk from residential radon

385

comparison to that for smokers (0.069), but not signiﬁcantly different. The estimated intercept (0.86) is consistent with mortality from cancers other than lung. The estimates of excess relative risk per unit concentration are consistent with estimates derived in the meta-analysis of eight indoor case–control studies [1] (see Fig. 2). 3.1. Comparison to results from the Czech study of uranium miners In the Czech study of 9960 uranium miners, a total of 922 lung cancers have been observed by the end of 1999, representing a relative risk of 3.29 [2]. In order to compare the risk in the residential and occupational studies, the indoor exposure was converted into another quantity in terms of kBq m−3 y integrating both the concentration of radon (kBq m−3 ) and the duration of residence in years (y), similar to the unit of WLM (Working Level Month), which is used in occupational studies. The WLM integrates the concentration of radon progeny in air in terms of working levels (WL) and the duration of exposure in working months (170 hours). One WL corresponds to 130 000 MeV of potential alpha energy released by the short-lived progeny in equilibrium with radon in one liter of air (3.7 kBq m−3 ). Relative risk in relation to cumulated exposures below 150 WLM (kBq m−3 y) is shown in Fig. 3. Risk coefﬁcients related to cumulated exposure in mines and houses are compared in Table 3. The considerable effect of time since exposure seen in the occupational study can also be observed in the residential study. The analysis of temporal modiﬁcation of the risk suggests that the risk results mainly from exposures experienced in the previous period of 5–19 years. Time speciﬁc risk coefﬁcients (ERR/(kBq m−3 y)) in the residential study decline with time since exposure similarly as in the occupational study, where the statistical power to detect such an effect is substantially higher. The risk coefﬁcient from exposures in the distant past (35–49 years) is negligible and the time window of 5–34 years is appropriate.

Fig. 3. Relative risk (RR) by exposure accumulated in the preceding period of 5–34 years. Comparison of the residential () and occupational () studies.

386

A. Heribanová, L. Tomášek Table 3 Time since exposure (TSE) speciﬁc risk coefﬁcients Residential (218 cases)

Uranium miners (922 cases)

TSE

ERR/(kBq m−3 y)

90%CI

ERR/WLM∗

5–19 20–34 35–49 5–34 35–49

0.045 0.013 not estimated 0.028 0.003

−0.000–0.153 −0.000–0.123

0.059 0.015 0.008 0.028 0.004

0.005–0.069 −0.000–0.071

90%CI 0.043–0.075 0.010–0.021 0.003–0.012 0.020–0.035 0.001–0.008

Estimates from occupational and residential studies. ∗ For exposure rates below 8 WL.

3.2. Smoking Smoking data were collected in person among people living in the study area by the time of interview in 1991–1992. Information on deceased subjects and people who moved from the study area were obtained from relatives and from medical records. So far, smoking data are available for 71% of subjects older than 15 years. The proportion of ever smokers is 53% among males and 13% among females, which is less than estimated ﬁgures in the general Czech population of similar ages [3]. Since the relative risk among long term ex-smokers was not signiﬁcantly different from never smokers (RR = 0.93), subjects who quited smoking more than 20 years before (6 cases) were combined with never smokers. Preliminary results (Table 4) exhibit a considerable consistency in risk coefﬁcients. Results from the study of uranium miners [2] are given for comparison. Because of the low numbers of non-smoking lung cancer cases, the comparison of the risk coefﬁcients between smokers and non-smokers has not resulted in signiﬁcant differences in either cohorts, however, higher relative risk is suggested among non-smokers. For instance, exposure to radon at 1 kBq m−3 roughly corresponds to the relative risk among non-exposed smokers at 2 cigarettes a day. However, the simultaneous exposure to radon and tobacco smoke does not result in the product of the risks, but is about 3/4. These patterns are consistent with sub-multiplicative interaction (Table 5). Table 4 Excess relative risk per unit exposure among smokers and non-smokers UE∗ kBq m−3 y Smokers Non-smokers§ Occupational WLM Smokers Non-smokers§ Residential

Cases ERR/UE† 90%CI 173 42 362 43

0.023 0.043 0.018 0.040

0.000–0.064 −0.006–0.444 0.010–0.030 0.014–0.140

∗ UE: unit of exposure. † Excess relative risk per unit exposure. ‡ Test for heterogeneity of risk coefﬁcients between smokers and non-smokers. § Including 6 and 14 ex-smokers who quited smoking before more than 20 years.

p-value‡ 0.768 0.214

Czech study on lung cancer risk from residential radon

387

Table 5 Relative risk of lung cancer from radon and smoking (sub-multiplicative interactions)

Not exposed Exposed (25 years at 1 kBq m−3 )

Non-smoker

Smoker (20 c/d)

Smoker (2 c/d)

1 2

10 15

2 3

4. Discussion In evaluating the observed numbers of lung cancer deaths in the study, comparisons were made with the expected numbers that were derived from the country statistics. Potential differences between the study population and the Czech population were accounted for by using the multiplicative correction term (parameter i) in the model. This intercept term can also account for potential loss of follow-up and missing causes of deaths. The estimated value of this correction factor in the residential study (0.86) corresponds to lower mortality from cancers other than lung in the study (O/E = 0.87), which is not believed to be inﬂuenced by ionizing radiation from radon. These lower relative ﬁgures reﬂect the mostly rural character of the study population. Therefore, the observed elevated mortality from lung cancer in the study (O/E = 1.15) has to be compared in fact to this estimated lower mortality (0.86), which corresponds to zero exposure. Then the resulting observed excess relative risk in the study would be 35%. The evidence of lung cancer risk from radon is based mainly on studies of men employed underground in mines. Direct estimation of the risk from residential radon is more complex than in occupational studies. In addition, exposure estimates in residential studies show higher uncertainty than in studies of miners. The concentrations of radon vary substantially in time and location. Moreover, in most studies recent exposures are estimated with higher accuracy than those in the past. Estimates of cumulative exposure in occupational studies are generally more precise; not only because the radon measurements in mines were conducted in the past, but also because the duration of stay of workers in the radon environment was recorded with a higher precision than in houses. In spite of these weak points, the results of the Czech occupational and residential studies exhibit considerable similarities. Virtually identical coefﬁcients were estimated when exposures were limited to a preceding period of 5–34 years and very similar coefﬁcients were obtained in the studies for smokers and non-smokers. The exposure units used in the occupational and residential studies are different though. According to ICRP-65 [4], 1 WLM = 3.5 mJ h m−3 and 1 kBq m−3 y = 15.6 mJ h m−3 (assuming 2000 and 7000 hours of exposure annually, respectively). The important point in the estimation of potential alpha energy intake is the estimation of breathing rates at various activities. Using breathing rate estimate [5], it can be shown that the annual intake is approximately 9.5 mJ and 7.5 mJ corresponding to 1 kBq m−3 and 1 WLM per year, respectively. It can be concluded that the same risk estimated from the studies at annual rates of 1 WLM and 1 kBq m−3 can also be justiﬁed by the dosimetric account. Smoking prevalence among miners is usually higher than in the general population. For instance, the proportion of ever smokers is 53% among males and 13% among females, which is less than in the Czech male population (62% and 22%) of similar age [3] in the period 1960– 1999. The assessment of the combined effects of smoking and radon is an obvious issue in

388

A. Heribanová, L. Tomášek

occupational studies. However, only few had been evaluated. The crucial point is a sufﬁcient number of cases in the non-smoking group. Results obtained in six combined uranium miners studies [4] (64 non-smoking cases) are in line with our ﬁndings. Generally, it is presumed that the difference in risk coefﬁcients is due to different patterns of lung deposition and clearance in smokers and non-smokers.

Acknowledgements The author acknowledges the contribution of Drs. Josef Ševc and Václav Plaˇcek, who initiated these studies, and thanks to the subjects who took part in the study and also to many ofﬁcials of local authorities for assistance in ascertaining causes of death. The present research was supported by the Internal Grant Agency of the Ministry of Health (IGA NJ/6768-3).

References [1] J.H. Lubin, J.D. Boice, J. Natl. Cancer Inst. 89 (1997) 49. [2] L. Tomášek, Czech study on uranium miners – evaluation of temporal factors in a 50 year follow-up, in: J.P. McLaughlin, S.E. Simopoulos, F. Steinhäusler (Eds.), The Natural Radiation Environment VII, Proc. VIIth Internation Symposium on the NRE, Rhodes, Greece, 20–24 May 2002, in: Radioactivity in the Environment, vol. 7, Elsevier, Amsterdam, 2004, this volume. [3] Z. Škodová, Z. Píša, et al., Cor Vasa 38 (1996) 11. [4] ICRP Publication 65: Protection against radon-222 at home and at work, Ann. ICRP 23 (2) (1993). [5] National Research Council, Comparative Dosimetry of Radon in Mines and Homes, National Academy Press, Washington, DC, 1991.

389

Czech study on uranium miners – evaluation of temporal factors in a 50 year follow-up L. Tomášek National Radiation Protection Institute, Šrobárova 48, 100 00 Prague, Czech Republic

Epidemiological evidence of lung cancer risk from radon is based mainly on studies of miners. Two such studies among 9960 Czech uranium miners were established in 1970 and 1980. A total of 922 cases of lung cancer were observed by 1999. The excess relative risk in both studies depends linearly on cumulative exposure experienced in the past 5–34 year period. The effect from exposures in the distant past before 34 years is less than 1/13 in comparison to the 5–19 year period. The combined effect of smoking and radon progeny analysed in the study is consistent with a sub-multiplicative interaction: the excess relative risk per unit exposure among non-smokers was roughly twice relative to smokers. In comparison, the risk of lung cancer from occupational and residential radon, shows considerable agreement, both in temporal patterns and smoking interaction. If breathing rates in miners and the general population are accounted for, the intake of potential alpha energy from indoor radon of 1 kBq m−3 is approximately equal to the risk in mines at an exposure rate 1 WLM per year, which is in line with observed risks in the studies.

1. Introduction

According to the International Agency for Research on Cancer, there is signiﬁcant evidence to classify radon as a carcinogen. Using extrapolations from occupational studies, it can be shown that for some countries environmental exposure to radon is the second most important cause of lung cancer in the general population after cigarette smoking. Czech studies among uranium miners, established in 1970 by Josef Ševc [1], aim to contribute to knowledge on the risk from radon, particularly by evaluating temporal factors and interaction of radon exposure and smoking. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07044-5

© 2005 Elsevier Ltd. All rights reserved.

390

L. Tomášek

2. Methods 2.1. Deﬁnition of cohorts The study population of occupational cohorts consists of uranium miners exposed in two different periods. The older cohort (S) involves 4339 miners from the Jáchymov region (West Bohemia), who began underground work in 1948–1959 [2]. The newer cohort (N) consists of 5621 miners who entered the Pˇríbram mines (Central Bohemia) during 1968–1974, when hygienic measures had been already fully introduced [3]. 2.2. Follow-up Information on vital status and causes of death were obtained from the Czech population registry and the diagnoses from registries of deaths at local administrative ofﬁces. Follow-up for each subject started according to the respective deﬁnition of each cohort [4]. The present follow-up is limited by the end of 1999. 2.3. Estimation of exposure to radon and tobacco smoke Exposure estimates in the S study were derived from extensive measurements of radon commencing already in 1949 [2]. Each man’s annual exposures to radon progeny were estimated by combining measurement data with the men’s registered employment details, including duration of underground work at different shafts and job category. In the N study, the exposure estimates were based on personal dosimetric records [3]. Occupational exposures are given in working level months (WLM) integrating the concentration of radon progeny in air in terms of working levels (WL) and the duration of exposure in working months (170 hours). One WL corresponds to 130 000 MeV of potential alpha energy released by the short-lived progeny in equilibrium with radon in one litre of air (3.7 kBq m−3 ). Information on smoking in cohorts was different. In the S cohort, data were collected from only 332 cases since 1970 and 502 controls matched by year of birth and attained age. This information was obtained from medical records, relatives, and from living miners. In the N study, smoking data were routinely collected (85%). As smoking details from these sources were different, the collected smoking data were limited to numbers of cigarettes smoked per day and the year of cessation. 2.4. Statistical analyses The statistical analyses were based on the relative risk model. Numbers of cases (O) observed at given levels of cumulated exposures and modifying variables (Wk ) were assumed to have the Poisson distribution with parameter iE(1 + bk Wk ), where E is the number of cases expected from national mortality data, parameter i is an intercept term that allows the background mortality rate for the ‘unexposed’ cohort to differ from that in the general population, parameters bk represent speciﬁc excess relative risk per unit exposure (ERR/WLM) in each level (k) of time since exposure (5–19, 20–34, 35–49), age at exposure (–29, 30–39, 40–), and exposure rate in terms of WL (< 1, 1–2, 2–4, 4–8, > 8). The observed relative risk (RR) in the cohort can be estimated by the ratio O/(iE).

Czech study on uranium miners – evaluation of temporal factors in a 50 year follow-up

391

3. Results By 1999, a total of 4008 deaths were observed in both cohorts of uranium miners. Among these deaths, a total of 922 cases of lung cancer were diagnosed (Table 1). Increased mortality from lung cancer in occupational cohorts corresponds to cumulative exposures experienced in the two cohorts. Lung cancer risk in the cohorts shows a linear trend (Fig. 1). The obvious departure from linearity in exposures over 300 WLM is attributed to the so-called inverse exposure rate effect. This is accounted for by including exposure rate speciﬁc terms in the model (Table 2). Similar effects were observed in most studies of highly exposed miners. In our analyses, the total cumulated exposure for each miner was differentiated according to the exposure rate in each year of employment. The exposure rate effect is present only at exposure rates over 4 WL. In Table 1 Cohorts of Czech uranium miners by 1999

Number of miners Mean cumulated exposure Alive by 1999 Person-years of follow-up Lung cancers Standardised mortality (O/E ∗ ) Relative risk (RR† )

S

N

S+N

4339 152 WLM 20% 116 366 841 4.64 3.93

5621 7 WLM 84% 143 709 81 1.43 1.23

9960 70 WLM 56% 260 075 922 3.88 3.29

∗ O/E: ratio of observed numbers to numbers expected from national mortality data. † RR: relative risk – ratio of observed cases to numbers expected at zero exposure.

Fig. 1. Relative risk of lung cancer (RR) in the cohort of U miners by categories of cumulative exposure (WLM); solid line: RR = 1 + 0.02 WLM for exposure rates < 8 WL.

392

L. Tomášek

Table 2 Effect of exposure rate in relative risk model RR = 1 + ck Wk

ERR/WLM‡ Deviance§ ERR/WLM ERR/WLM ERR/WLM ERR/WLM ERR/WLM Deviance§ ERR/WLM ERR/WLM ERR/WLM Deviance§ ERR/WLM ERR/WLM Deviance§

Estimate∗

90%CI

Estimate†

90%CI

overall (WL)

0.012

0.009–0.016 7090.23

0.017

0.012–0.023 7053.42

0–1 1–2 2–4 4–8 8–

0.020 0.026 0.022 0.009 0.005

0.015 0.027 0.027 0.011 0.007

0–4 4–8 8–

0.024 0.010 0.005

0–8 8–

0.019 0.003

0.009–0.035 0.017–0.037 0.015–0.032 0.003–0.018 0.001–0.010 7056.23 0.017–0.034 0.003–0.017 0.002–0.010 7057.03 0.014–0.026 0.000–0.007 7065.45

0.004–0.031 0.018–0.041 0.018–0.039 0.004–0.021 0.002–0.013 7023.93 0.021–0.041 0.005–0.024 0.003–0.014 7027.21 0.018–0.035 0.001–0.011 7033.38

0.029 0.013 0.008 0.025 0.006

∗ Crude estimates. † Estimates adjusted for time since exposure and age at exposure. ‡ Excess relative risk per WLM. § Deviance represents measure of ﬁt of each model.

Table 3 Model of relative risk and modifying effect of time since exposure, age at exposure and exposure rate RR = 1 + b ti aj ck Wij k Modifying factor

Parameter of model ERR/WLM∗ b

Estimate 0.111

Time since exposure

5–19 20–34 35–49 –29 30–39 40– 0–4 WL 4–8 WL 8– WL

1.000 0.174 0.073 1.000 0.676 0.388 1.000 0.697 0.360

Age at exposure

Exposure rate

t1 t2 t3 a1 a2 a3 c1 c2 c3

95%CI 0.077–0.160

χ2

p-value

264.02

0.0001

40.52

0.0001

15.08

0.0005

0.127–0.221 0.036–0.110 0.449–0.903 0.282–0.494 0.285–1.110 0.098–0.621

∗ Excess relative risk per WLM corresponding to ﬁrst categories of each modifying factor.

general, the effect from exposures received at very high rates (> 8 WL) is roughly 1/3 when compared to rates < 8 WL. Lung cancer risk from cumulated exposure to radon is strongly inﬂuenced by time since exposure (chi-squared 264, 2 degrees of freedom), being less than 1/13 after more than 34 years since exposure in comparison to the period of 5–19 years (Table 3).

Czech study on uranium miners – evaluation of temporal factors in a 50 year follow-up

393

The evaluation of modifying effect of age at exposure is related to the effect of time since exposure. For instance, when exposures are considered in time windows 5–34 and 35–49, the effect of age at exposure is diluted in this longer period. If time since exposure factor is omitted, the risk is paradoxically higher for miners exposed at older ages, which at the same time actually represents recent exposures. The evaluation of age at exposure depends both on numbers of miners employed at young ages (say below 30) and simultaneously on the critical period about two decades after such exposure. As the background rates below age 50 are generally relatively low in comparison to older age categories, the observed absolute excess is low too and so the detection of age at exposure effect is limited. 3.1. Comparison to results from the Czech residential study In the Czech residential study among 12 000 inhabitants of a radon prone area [5,6], a total of 218 lung cancers had been observed by the end of 1999, representing RR = 1.35. The relative risk of lung cancer depends linearly on average radon concentration. In order to compare the risk in the occupational and residential studies, the indoor exposure was converted into another quantity in terms of kBq m−3 y integrating both the concentration of radon (kBq m−3 ) and the duration of residence in years (y). Relative risk in relation to cumulative exposures below 150 WLM is shown in Fig. 2. Risk coefﬁcients related to cumulative exposure in mines and houses are compared in Table 4. The considerable effect of time since exposure seen in the occupational study can also be observed in the residential study. The analysis of temporal modiﬁcation of the risk suggests that the risk results mainly from exposures experienced in the previous 5–19 year period. Time speciﬁc risk coefﬁcients (ERR/(kBq m−3 y)) in the residential study decline with time since exposure similarly as in the occupational study, where the statistical power to detect such an effect is substantially

Fig. 2. Relative risk (RR) by exposure accumulated in the preceding period of 5–34 years in the occupational (WLM) and residential (kBq m−3 y) studies.

394

L. Tomášek

Table 4 Time since exposure (TSE) speciﬁc risk coefﬁcients Uranium Miners (922 cases)

Residential (218 cases)

TSE

ERR/WLM∗

90%CI

ERR/(kBq m−3 y)

90%CI

5–19 20–34 35–49 5–34 35–49

0.059 0.015 0.008 0.028 0.004

0.043–0.075 0.010–0.021 0.003–0.012 0.020–0.035 0.001–0.008

0.045 0.013 not estimated 0.028 0.003

0.000–0.153 0.000–0.123 0.005–0.069 0.000–0.071

Estimates from occupational and residential studies. ∗ For exposure rates below 8 WL.

higher. The risk coefﬁcient from exposures in the distant past (35–49 years) is negligible (about 1/7 in comparison to the 5–34 period). 3.2. Smoking With the exclusion of the S cohort, smoking data were collected from most cohort members. When the S cohort of miners was identiﬁed in 1970, the mines had already closed and so direct investigation in the study was not possible. As most informative results in this respect can be expected in highly exposed miners, a retrospective investigation was recently initiated. Data were collected from medical records and from relatives of miners. So far, information on 834 miners from the S cohort is available. Since the relative risk among long term exsmokers was not different from never smokers (RR = 0.80, 95%CI: 0.39–1.65), subjects who quit smoking before more than 20 years ago (14 cases) were combined with never smokers. Preliminary results (Table 5) exhibit a considerable consistency in risk coefﬁcients. Results from the Czech residential study [5,6] are given for comparison. Because of the low numbers of non-smoking lung cancer cases, the comparison of the risk coefﬁcients between smokers and non-smokers has not resulted in signiﬁcant differences in either cohorts; however, higher relative risk is suggested among non-smokers. In addition to Table 5 Excess relative risk per unit exposure among smokers and non-smokers UE∗ Occupational

WLM

Residential

kBq m−3 y

Smokers Non-smokers§ Smokers Non-smokers§

Cases

ERR/UE†

90%CI

p-value‡

362 43 173 42

0.018 0.040 0.023 0.043

0.010–0.030 0.014–0.140 0.000–0.064 −0.006–0.444

0.214

∗ UE: unit of exposure. † Excess relative risk per unit exposure. ‡ Test for heterogeneity of risk coefﬁcients between smokers and non-smokers. § Including 14 and 6 ex-smokers who quited smoking before more than 20 years.

0.768

Czech study on uranium miners – evaluation of temporal factors in a 50 year follow-up

395

Table 6 Relative risk of lung cancer from radon progeny and smoking

Not exposed Exposed (200 WLM)

Non-smoker

Smoker

1 10

11 55

smoker speciﬁc estimated risk, the relative risk of lung cancer for smokers in comparison to non-smokers in the cohort was estimated (RR = 11.1). This value is consistent with the amount of cigarettes smoked per day among miners. Using these estimates, the combined effect from radon and smoking can be summarised (Table 6). For instance, exposure to radon progeny at 200 WLM roughly corresponds to the relative risk among non-exposed smokers (assuming 20 cigarettes a day). However, the simultaneous exposure to radon progeny and tobacco smoke does not result in the product of the risks, but is about half. These patterns are consistent with sub-multiplicative interaction.

4. Discussion In comparison to our previous publications [4], results based on the extended follow-up show that factors that modify the general linear model became more discernible. The present analyses conﬁrmed the decreasing effect with time since exposure, observed in studies of miners [7]. A similar pattern was observed in the residential study [6]. Although the conﬁdence intervals for the time speciﬁc risk coefﬁcients in the Czech residential study are relatively wide, the estimates suggest that using a time window of 5–34 years is appropriate in such studies. The issue of smoking in the evaluation of lung cancer risk is essential. Although the design in the S and N cohorts for smoking is different, the results are consistent with general knowledge on the risk of smokers compared to non-smokers. Smoking prevalence among miners is usually higher than in the general population. For instance, in the N study, the proportion of ever smokers is 77%, which is more than in the Czech male population (65–70%) of similar age [8]. The assessment of the combined effects of smoking and radon is an obvious issue in occupational studies. However, only few had been evaluated. The crucial point is a sufﬁcient number of cases in the non-smoking group. Results obtained in six combined uranium miners studies (64 non-smoking cases) are in line with our ﬁndings [7]. Generally, it is presumed that the difference in risk coefﬁcients is due to different patterns of lung deposition and clearance in smokers and non-smokers. The evidence of lung cancer risk from radon is based mainly on studies of men employed underground in mines. Direct estimation of the risk from residential radon is more complex than in occupational studies. In addition, exposure estimates in residential studies show higher uncertainty than in studies of miners. Estimates of cumulative exposure in occupational studies are generally more precise; not only because the radon measurements in mines were already conducted in the past, but also because the duration of stay of workers in the radon environment was recorded with a higher precision than in houses. In spite of these weak points, the results of the Czech occupational and residential studies exhibit considerable similarities.

396

L. Tomášek

Virtually identical coefﬁcients were estimated when exposures were limited to the preceding period of 5–34 years and very similar coefﬁcients were obtained in the studies for smokers and non-smokers. The exposure units used in the occupational and residential studies are different though. According to ICRP-65 [9], 1 WLM = 3.5 mJ h m−3 and 1 kBq m−3 y = 15.6 mJ h m−3 (assuming annually 2000 and 7000 hours of exposure, respectively). The important point in the estimation of potential alpha energy intake is the estimation of breathing rates for various activities. Such values were estimated by the Panel on dosimetric assumptions affecting the application of radon risk estimates [10]. Using [10, Table 9–5] and assuming 50% heavy exercise, 40% light exercise, and 10% rest for miners and 40% sleep, 40% rest, and 20% light exercise within the annual 7000 hours for residents, we can estimate the annual intake of 7.5 mJ and 9.5 mJ corresponding to 1 WLM and 1 kBq m−3 , respectively. In conclusion, it is likely that the risk corresponding to 1 WLM and 1 kBq m−3 is practically the same.

Acknowledgements The author acknowledges the contribution of Drs. Josef Ševc and Václav Plaˇcek, who initiated these studies, and thanks to the subjects who took part in the study and also to many ofﬁcials of local authorities for assistance in ascertaining causes of death. The present research was supported by the Internal Grant Agency of the Ministry of Health (IGA NJ/6768-3).

References [1] [2] [3] [4] [5] [6]

[7]

[8] [9] [10]

J. Ševc, V. Plaˇcek, J. Jeˇrábek, in: Proc. 4th Conf. on Radiat. Hyg., Part II, 1971, p. 315. L. Tomášek, S.C. Darby, et al., Radiat. Res. 137 (1994) 251. V. Plaˇcek, L. Tomášek, et al., Pracov. Lék. 49 (1997) 14 (in Czech). L. Tomášek, V. Plaˇcek, Radiat. Res. 152 (1999) S59. L. Tomášek, T. Müller, et al., Cent. Eur. J. Publ. Health 9 (2001) 150. A. Heribanová, L. Tomášek, Czech study of lung cancer from residential radon, in: J.P. McLaughlin, S.E. Simopoulos, F. Steinhäusler (Eds.), The Natural Radiation Environment VII, Proc. VIIth International Symposium on the NRE, Rhodes, Greece, 20–24 May 2002, in: Radioactivity in the Environment, vol. 7, Elsevier, Amsterdam, 2004, this volume. Committee on Health Risks of Exposure to Radon, Board on Radiation Effects Research, Commission on Life Sciences, National Research Council, Health Risks of Exposure to Radon, National Academy Press, Washington, DC, 1999. Z. Škodová, Z. Píša, et al., Cor Vasa 38 (1996) 11. ICRP Publication 65: Protection against radon-222 at home and at work, Ann. ICRP 23 (2) (1993). National Research Council, Comparative Dosimetry of Radon in Mines and Homes, National Academy Press, Washington, DC, 1991.

397

Activity concentrations of the thoron and radon progenies Pb-212 and Pb-210 in the healing gallery of Badgastein, Austria G. Wallner, P. Pany, S. Ayromlou Institut für Anorganische Chemie, University of Vienna, Währingerstraße 42, A-1090 Vienna, Austria

In the healing gallery of Badgastein (Austria) aerosol samples were collected with cascade impactor devices and Pb-210 and Pb-212 activity concentrations were measured by liquid scintillation counting. Simultaneously Rn-222 was determined with a Si surface barrier detector. At different ventilation rates of the gallery, Pb-210 concentrations ranged between 4.6 and 730 mBq m−3 , while Pb-212 values were between 760 and 7230 mBq m−3 . The highest Pb-212/Pb-210 ratio of 165 was found during full ventilation, when Rn-222 and Pb-210 levels dropped considerably. The lowest ratio of 10 was measured after a few months of patients’ treatment, when very high Rn-222 levels of up to 350 kBq m−3 and the maximum Pb-210 level were reached. Taking into account the particle sizes of the aerosols, we observed that the Pb-212 activity distribution was shifted towards smaller particle sizes compared to the Pb-210 distribution.

1. Introduction Badgastein is a famous radon spa in the Austrian Alps (Hohe Tauern), where patients with rheumatic complaints have been treated in the so-called “Heilstollen” (a thermal gallery) in Böckstein near Badgastein since 1952. More than 2 km of tunnels, originally blasted for prospection of gold, are a natural source of air with a temperature of 36–41 ◦ C and a relative humidity of 70–100%, bearing an average radon activity concentration of 44 kBq m−3 [1]. In this gallery at a distance of about 2 km from the entrance, Rn-222 was measured continuously at different ventilation rates, and simultaneously aerosols were collected on aluminium foils with a cascade impactor in order to enable the measurement of the particle size distribution of the long-lived daughter product Pb-210 (half-life 22 a) and its progeny [2]; on these foils the Rn-220 daughter product Pb-212 (half-life 10.6 hours) was also found. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07045-7

© 2005 Elsevier Ltd. All rights reserved.

398

G. Wallner et al.

Activity concentrations of both lead isotopes will be presented for different radon concentrations as well as isotope ratios, and special emphasis will be placed on the different particle size distributions of the lead isotopes.

2. Methods Rn-222 was determined continuously with a Si surface barrier detector of a battery operated working level monitor (BWLM-plus™ , Tracerlab) via its α-emitting daughters Po-218 and Po-214 collected on a cellulose ﬁlter. Po-214 (half-life 162 μs) and Po-218 (half-life 3.05 min) are assumed to be in radioactive equilibrium with Bi-214 and Rn-222, respectively. The differential equations for formation and decay of the nuclides were integrated during the collecting and measuring period (usually 0.5 hours) and so air activity concentrations of the short-lived radon progenies as well as the radon equivalent concentration were calculated. The counter was calibrated by the manufacturers by measurements in atmospheres of known radon activity concentrations. The aerosols for Pb-210 and Pb-212 determination were collected on aluminium foils in a 6- or 8-stage cascade impactor device [3]. Particle sizes collected on the ﬁrst-stage ﬁlter of the 6-stage device were between 0.15 and 0.30 μm, on the second stage between 0.30 and 0.60 μm and so on, until 5 to 10 μm on the sixth stage. With the 8-stage device, the smallest particles were between 0.062 and 0.125 μm and the largest between 8 and 16 μm. Volumes between 33 and 131 m3 of air were collected within sampling times of 18 to 26.7 hours. The aerosols were leached from the Al-foils with hot concentrated nitric acid, the acid was evaporated and the residue was taken up in a mixture of H3 PO4 and HCl. Then this solution was shaken with the extractant cocktail POLEX™, by which Po as well as Bi isotopes were extracted completely [4]. After phase separation, an aliquot of the scintillator phase was measured in a liquid scintillation counter using α/β-separation. The much more abundant Rn-220 progeny Pb-212 was measured immediately after extraction via the alphas (36.2%) of its daughter product Bi-212 (half-life 60.6 min), which was in radioactive equilibrium with its predecessor because of a delay of about 24 hours between sampling stop time and extraction time. Bi-210 (half-life 5.0 days) and Po-210 (half-life 138 days) were measured the following day after decay of Bi-212 (these results are published elsewhere: [2]); Pb-210 was measured via extraction of the ingrown Bi-210 after a waiting period of at least two weeks. The extraction and counting efﬁciencies were determined by extracting the isotopes of interest from Pb-210 and thorium solutions [4]. The lower limits of detection with counting times of 1000 min were 0.5 mBq for Pb-210 (measured via Bi-210) and 0.2 mBq for Pb-212 (via Bi-212). The chemical blank values (±1 σ ) of 1.4 ± 0.2 mBq Pb-210 and 0.2 ± 0.08 mBq Pb-212 were negligible compared with the high activities of our samples. Figure 1 shows the α- and β-spectrum of an aerosol sample originating from the thermal gallery (sampling date 2001.10.31; particle size 0.25–0.5 μm), measured immediately after sample preparation. The double peak in the α-spectrum consists of Po-210 (5.30 MeV) and Bi-212 (6.05 MeV); the broad high-energy α-peak, which has partly spilled over into the β-spectrum, is the short-lived Po-212 (8.78 MeV). The β-spectrum consists of the betas of Bi-210 (Emax = 1.2 MeV) and of Bi-212 (Emax = 2.3 MeV).

Activity concentrations of the thoron and radon progenies in the healing gallery of Badgastein

399

Fig. 1. α- and β-spectrum of an aerosol sample originating from the thermal gallery (sample October 31, 2001; particle size 0.25–0.5 μm), measured immediately after sample preparation (see text).

3. Results and discussion When taking the ﬁrst and second aerosol sample (January 12 and 13, 1999) the healing gallery was closed for therapeutic treatment. The main door was open (so the mine atmosphere is changed completely once in about 3 hours; these conditions are referred to as so-called full ventilation) and loose rocks were removed from the mine walls with water jets. The temperature was 36 ◦ C and the relative humidity about 60%. During the ﬁrst day, the simultaneously measured radon concentration fell from 10 000 to 150 Bq m−3 . Pb-210 and Pb-212 activity concentrations (summed over all impactor stages) were 13.2 and 815 mBq m−3 , respectively, giving a Pb-212/Pb-210 ratio of 62. During the second day with full ventilation, the radon concentration was constant at about 150 Bq m−3 . Pb-210 activity concentration fell to 4.6 mBq m−3 , while Pb-212 remained nearly constant at 757 mBq m−3 . Here the maximum Pb-212/Pb-210 ratio of 165 was found. The patients’ treatment started on January 18th, and so the following samples were collected under standard ventilation conditions (the air exchange is 10% of the above given between 7:30 a.m. and 6:00 p.m.; overnight fresh air supply is shut off). The temperature was 40 ◦ C and the relative humidity 85%. The radon concentration increased up to 48 kBq m−3 , while the Pb-210 and Pb-212 concentrations rose from 76 to 103 and 3454 to 5409 mBq m−3 , respectively, giving isotope ratios of 45 and 52. In a second measuring campaign in May 1999, also during therapeutic operation with standard ventilation conditions, activity concentrations of 728 and 7226 mBq m−3 were found for Pb-210 and Pb-212, respectively, giving a lower Pb-212/Pb-210 ratio of 10. In Autumn 2001, we collected aerosols at the end of the therapeutic season. On the last day with standard conditions, activity concentrations of 302 and 5756 mBq m−3 were observed, while on the next day during full ventilation the values fell to 38 and 1591 mBq m−3 . So the Pb-212/Pb-210 ratio rose from 19 to 42. The results are compiled in Table 1.

400

G. Wallner et al.

Table 1 Particle size distribution of Pb-212 and Pb-210 activity concentrations and isotope ratios; also the average Rn-222 concentrations during the collecting interval is given Date

Particle size (μm)

99 01 12

0.15–0.3 0.3–0.6 0.6–1.2 1.2–2.4 2.4–5.0 5.0–10.0 Sum

279 359 140 20 12 5 815

51 70 43 13.8 12 5 99

3.0 6.3 2.9 0.75 0.22 0.02 13.19

0.08 0.13 0.09 0.04 0.02 0.01 0.18

93 57 48 27 55 250 62

5.7 3.7 5.0 6.2 18 93 2.5

10–0.15

0.15–0.3 0.3–0.6 0.6–1.2 1.2–2.4 2.4–5.0 5.0–10.0 Sum

242 369 127 15.4 1.2 2.2 757

24 32 19 7.0 1.2 2.2 45

1.2 1.97 0.85 0.3 0.18 0.09 4.59

0.07 0.08 0.05 0.03 0.02 0.01 0.12

202 187 149 51 7 24 165

7.7 6.0 8.0 8.0 2.2 8.2 3.6

0.15

26.4 37.2 10.6 1.32 0.34 0.09 76

0.43 0.54 0.28 0.06 0.03 0.01 0.75

51 38 55 73

1.7 1.1 3.1 11

45

0.9

0.64 0.73 0.33 0.07 0.03 0.03 1.03

60 50 43 49 39 27 53

1.1 0.6 0.9 2.4 6.0 3.9 0.5

5–50

20–350

99 01 13

99 01 19

99 01 20

99 05 26

01 10 30

Pb-212 (mBq m−3 )

Error (3 σ )

Pb-210 (mBq m−3 )

Error (3 σ )

Ratio 212/210

Error (1 σ )

0.15–0.3 0.3–0.6 0.6–1.2 1.2–2.4 2.4–5.0 5.0–10.0 Sum

1355 1423 579 97

136 126 97 42

3454

213

0.15–0.3 0.3–0.6 0.6–1.2 1.2–2.4 2.4–5.0 5.0–10.0 Sum

2008 2821 482 79 9 10 5409

100 90 28 11 4 4.3 138

33.6 55.9 11.3 1.6 0.23 0.37 103

0.125 0.25 0.5 1 2 Sum

1221 3737 1675 537 56 7226

72 144 91 51 15 192

97.3 345 210 66 9.9 728

5.8 21.0 13.0 4.0 0.6 26

13 11 8 8 6 10

0.4 0.3 0.2 0.3 0.5 0.1

0.062–0.125 0.125–0.25 0.25–0.5 0.5–1.0 1.0–2.0 Sum

759 2054 2555 359 29 5756

93 230 198 74 15 326

20.7 119.3 153.5 17.0 1.6 312

2.1 10.4 12.7 1.33 0.15 17

37 17 17 21 18 18

1.9 0.8 0.6 1.6 3.2 0.5

Rn-222 (kBq m−3 )

(continued on next page)

Activity concentrations of the thoron and radon progenies in the healing gallery of Badgastein

401

Table 1 (continued) Date

Particle size (μm)

Pb-212 (mBq m−3 )

Error (3 σ )

01 10 31

0.062–0.125 0.125–0.25 0.25–0.5 0.5–1.0 Sum

379 668 438 106 1591

60 104 81 40 150

01 10 23

0.062–0.125 0.125–0.25 0.25–0.5 Sum

205 297 397 899

33 71 159 177

Pb-210 (mBq m−3 ) 6.1 16.4 13.2 1.8 37.5 0.55 0.63 0.40 1.59

Error (3 σ )

Ratio 212/210

0.62 1.41 1.25 0.16 2.0

62 41 33 59 42

0.06 0.07 0.05 0.10

371 469 988 566

Error (1 σ )

Rn-222 (kBq m−3 )

3.9 2.4 2.3 7.6 1.5 24 41 138 39

The only slight reduction of Pb-212 levels during full ventilation is striking as an airexchange within three hours should also reduce Pb-212 with its half-life of 10.6 hours (its precursor thoron reaches steady-state activity rapidly and probably depends little on the ventilation rate because of its very short half-life (55.6 s)). So we conclude that the large variation of the isotope ratios is due not only to the different half-lives of the isotopes (resulting in a much quicker reaching of the saturation level for Pb-212 compared to Pb-210), but also to the different origin of their precursors Rn-222 and Rn-220. Rn-222 originates from the thermal water horizons deep down in the mountain and reaches the healing gallery through clefts and fractures. The nearest water level, about 300 m below, contains 758 ± 22 Bq L−1 Rn-222 when reaching the surface via different springs (e.g., the so-called Doktorquelle) [5]. In this water 150 ± 25 mBq L−1 Ra-226 were found, while Ra-228, one of the precursors of Rn-220, could not be detected [6]. However, in rock samples originating from the healing gallery, the prominent γ-lines of Ac-228 as well as Pb-212 and Bi-212 were found. This means that the walls of the gallery contain thorium and its decay products, including also Rn-220, which might escape from the rock surface. The probability of a successful escape and as a consequence the possibility to generate aerosol-bound Pb-212 seems to increase with the ventilation rate due to the air-duct in the gallery. Taking into account the particle sizes of the aerosols, we found in all samples taken in the healing gallery that the Pb-212 activity distribution was shifted towards smaller particle sizes compared to the Pb-210 distribution. This is reﬂected in the decreasing Pb-212/Pb-210 ratios with increasing particle diameters (see Table 1). The observed effect might be due to the fact that Pb-210 had been detached from the aerosol due to recoil after the 7.7 MeV α-decay of its predecessor Po-214; this detachment is more probable for smaller particles. Pb-212, on the contrary, is the ﬁrst Rn-220 progeny to become attached due to the short half-life of its predecessor Po-216 (0.15 s); no recoil detachment should change its distribution [7]. The calculation of the activity median aerodynamic diameter (AMAD, see Table 2) is only an estimation and therefore the real errors would be clearly larger than those calculated from the errors of the measured activities, which are in the order of magnitude of only a few percent (see Table 2); in our opinion, realistic errors would be in the order of 10%. So the difference of the AMADs is signiﬁcant only for half of the samples, while the Pb-212/Pb-210 ratios

402

G. Wallner et al. Table 2 Activity median aerodynamic diameter (AMAD) for the Pb-212 and Pb-210 distribution Date

AMAD (micrometer) Pb-212 Pb-210

1999 01 12 1999 01 13 1999 01 19 1999 01 20 1999 05 26 2001 10 30 2001 10 31

0.41 0.41 0.38 0.37 0.21 0.26 0.20

0.47 0.47 0.39 0.40 0.22 0.27 0.23

2001 10 23

0.23

0.17

Samples January 12 to May 26, 1999, were collected with a 6-stage impactor, the others with an 8-stage impactor; sample October 23, 2001 is from the vapor bath.

show the decline with increasing particle size for all samples with the exception of the sample October 23, 2001, from the so-called vapor bath. Here the AMAD of Pb-210 was 0.17 μm, clearly lower than the AMAD of Pb-212, 0.23 μm (October 23, 2001). We do not have an explanation for this contrary shift.

4. Conclusions Due to different ventilation rates the activities of the lead isotopes Pb-210 and Pb-212 changed considerably. The values of Pb-210 varied by a factor of 150, while Pb-212 concentrations altered only by a factor of 10. This is probably due to the different origin of the respective gaseous predecessors of the isotopes: Rn-222 originates from a water horizon deep down in the mountain, while Rn-220 emanates directly from the walls of the gallery. Looking at the particle size distribution, in all samples from the healing gallery a shift of the Pb-212 activity distribution towards smaller particle diameters was found compared to the Pb-210 distribution. This conﬁrms the assumption that Pb-210 had been detached mainly from the smaller aerosols due to recoil effects after Po-214 decay, while Pb-212 is the ﬁrst Rn220 progeny which can become attached to aerosols, while its short-lived predecessor Po-216 stays in the gaseous phase.

Acknowledgements For supplying the aerosol samples and the working level monitor, we are indebted to A. Berner, Institute for Experimental Physics, University of Vienna, and H. Kelm, Tracerlab Germany, respectively. We thank the Erzbergbau Radhausberg Ges.m.b.H., especially P. Brandmaier, for permission and technical assistance. Parts of this work were performed under contract to the Austrian Federal Chancellery (Project GZ 353.019/0-VI/9/98).

Activity concentrations of the thoron and radon progenies in the healing gallery of Badgastein

403

References [1] P. Brandmaier, Strahlenschutz im Gasteiner Heilstollen, in: K. Mück, A. Hefner, N. Vana (Hsgb.), Strahlenschutz für Mensch und Gesellschaft im Europa von Morgen, Tagung des ÖVS und FS, Gmunden, September 2001, pp. 297–300. [2] K. Irlweck, G. Wallner, A. Berner, Langlebige Radon-Folgeprodukte im Gasteiner Heilstollen, in: K. Mück, A. Hefner, N. Vana (Hsgb.), Strahlenschutz für Mensch und Gesellschaft im Europa von Morgen, Tagung des ÖVS und FS, Gmunden, September 2001, pp. 301–304. [3] A. Berner, Design principles of the AERAS low pressure impactor, in: B.Y.H. Liu, D.Y.H. Pui, H.J. Fissan (Eds.), Aerosols, Elsevier, Amsterdam, 1984, pp. 139–142. [4] G. Wallner, Simultaneous determination of 210 Pb and 212 Pb progenies by liquid scintillation counting, Appl. Radiat. Isot. 48 (4) (1997) 511–514. [5] C. Katzlberger, G. Wallner, K. Irlweck, Determination of 210 Pb, 210 Bi and 210 Po in natural drinking water, J. Radioanal. Nucl. Chem. 249 (1) (2001) 191–196. [6] F. Schönhofer, G. Wallner, Very rapid determination of Ra-226, Ra-228 and Pb-210 by selective absorption and liquid scintillation counting, Radioact. Radiochem. 12 (2) (2001) 33–38. [7] C. Papastefanou, E.A. Bondietti, Aerodynamic size associations of 212 Pb and 214 Pb in ambient aerosols, Health Phys. 53 (5) (1987) 461–472.

404

Characterization of thoron and radon ﬂow-through sources R. Rolle, M. Gründel, R. Schulz, J. Porstendörfer Institute of Physical Chemistry, Georg-August University Goettingen, Tamman Street 6, 37077 Goettingen, Germany

Different types of Rn gas measurement systems are compared regarding their sensitivity and discrimination against interferences (differentiation), to arrive at a reference system suitable for the characterization of a thoron (220 Rn–Tn) and a radon (222 Rn) ﬂow-through source. The source manufacturer’s calibration of these sources was found to be unacceptable.

1. The need for a 220 Rn/222 Rn reference system All extended measurements of x Rn (222 Rn + 220 Rn) decay product concentrations in air (analogous to air ﬁltration by the lung) show appreciable contributions from 220 Rn decay products. The majority of current measurements of radon decay product potential alpha energy (PAE) exposure (units 3.54 mJ h m−3 = 1 WLM) are, however, performed by Rn gas measurement, assuming an associated equilibrium factor (F = 0.4). The degree of inﬂuence of 220 Rn on the presumed 222 Rn measurement is largely unknown. A reference system was thus sought to determine such interference in commercial ‘Rn’ instruments and dosimeters. This paper outlines our choice of monitoring system and measurements of a 222 Rn and a 220 Rn ﬂow-through source.

2. Selection criteria for a 220 Rn/222 Rn reference system An ideal reference system would consist of convenient 222 Rn and 220 Rn sources and monitors with calibration traceable to international or absolute standards. Desirable features for the monitors are high measurement sensitivity, i.e. a preferably high signal rate per gas activity concentration, and high discrimination against interferences, e.g. decay products in various non-equilibrium situations may enhance or impede a x Rn gas measurement. A useful measure for comparing the sensitivities of gas measurement systems is the ‘effective volume’ gas EV x . This can be deﬁned as a detection system’s equivalent volume, which at RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07046-9

© 2005 Elsevier Ltd. All rights reserved.

Characterization of thoron and radon ﬂow-through sources

405

continuous, steady-state monitoring effectively contributes for each gas nuclide decay in that volume 1 count or charge packet (or of a collected nuclide’s chain member x at equilibrium). This count needs to be differentiated from others not originating from the gas to be measured. For enhancement of the quality of a measurement (reﬂected by the covariance matrix of the evaluated measurement data) partial spectral differentiation and time-variant equilibrium differentiation of the 222 Rn and 220 Rn decay chains is employed in measurement procedures, the former generally being more effective. 3. Types of x Rn gas concentration measurement systems Various types of α(β)γ (or decay charge packet) measurement for x Rn gas concentration are in use – determination of the gas concentration from difference measurements of a gas source, non-concentrating and gas-concentrating measurements. Non-equilibrium calculations of the decay chain saturation (coefﬁcients S) are required for practically all of these but are not detailed here. 3.1. Source differential x Rn concentration measurement Here a 226 Ra or (228 Th/)224 Ra source container is measured by γ-spectrometry before and after supplying x Rn. The starting point is usually decay chain equilibrium: x

Ra + x Rn + decay products (at equilibrium) − x Ra + x Rn + decay products (at partial equilibrium) = x Rn + decay products (part)

Evaluation of the before/after ratio measurement yields the degree of Rn emanation εRn , i.e. the fraction of Ra activity in the source actually releasing Rn to the total activity of Ra. If the x Ra activity is known, then the difference measurement yields the supplied x Rn. 3.2. Non-concentrating x Rn α-measurement (of gas + decay-product air concentrations) A range of detectors is used: SSD Si, electret, ZnS, track etch surfaces, thin window counters, (skin). For such detectors in open air, the effective volume of a 1 cm2 detector surface (at detection efﬁciency εd = 1) can be computed from the α-particle range in air: 3 3 air 222 Rn EV 4.8 MeV α = 1.01 cm and air 220 Rn EV 6.3 MeV α = 1.25 cm . The α-spectrum shape at the detector window is an intensity continuum ranging from 0 to the α energy of the x Rn. Air concentrations of 218 Po, 214 Po, 216 Po and 212 Bi/Po compound the detectable α-spectral continuum, while detection efﬁciency εd < 1 via detector windows and electronic signal thresholds further reduce the effective volumes. The sensitivity cm−2 (EV cm−2 ) of such detection systems is thus rather low compared with other types of Rn detecting systems (see Section 3.3), and discrimination, if needed, is ineffective. In systems with detector surfaces in enclosed air volumes, bounding surfaces within α range (4 to 7 cm) of the detector surface reduce the above x Rn sensitivity/cm2 still further

406

R. Rolle et al.

(EV < 1 cm3 per 1 cm2 surface). However, for ﬁltered x Rn inlet air, the gas ratios to decay products in the enclosed volume are better deﬁnable, and with an adequate diffusion barrier 220 Rn can be suppressed. This type of Rn detection system, calibrated for 222 Rn exposure, has found wide application in screening measurements where the low sensitivity is compensated by long exposure; diffusion barrier systems are available for simultaneous 222 Rn and 220 Rn concentration measurement. For charge collection in single- or multi-wire ionization chambers the effective volumes 3 3 x air Rn EV range from a few cm to > 20 000 cm . Current measurement, pulse counting and pulse shape processing offer from no to fair (low resolution spectroscopy) differentiation of the individual decay events of the two x Rn decay chains; the decay products, at respective equilibrium with x Rn are collected on chamber (wire) surfaces. Speciﬁc calibrated systems are in use as subordinate standards. For x Rn measurement systems employing electrostatic collection of α recoil nuclei on the detector, the effective volume EV is speciﬁed via an additional collection efﬁciency εc term for the recoil nuclei, besides the detection efﬁciency εd and (non-steady-state) saturation Ss Sm of the decay chain member(s): x Rn EV = V · εc · εd · Ss Sm . These systems generally have good a sensitivity EV ∼ 200 cm3 (range 2 cm3 to 3000 cm3 ) and with Si detectors very good spectral discrimination (bar 218 Po and 212 Bi 6 MeV α). It should be noted that the collection efﬁciency εc amongst instruments is not uniformly consistent and the systems require calibration. 3.3. Concentrating x Rn and subsequent γ-measurements When concentrating x Rn at efﬁciency εc and ﬂow rate f the effective volume for non-steadystate measurement is given by air x Rn EV = εc · f · λ−1 · Ss Sm · εd (see also associated paper N089 of these proceedings). For example, concentrating at εc · f = 3 L min−1 into a suitable activated charcoal vessel, we have εc · f · x Rn λ−1 = 24 × 106 and 4 × 103 cm3 , respectively, for 222 Rn and 220 Rn, so that at fair values of Ss Sm × εd , EVs of many liters can effectively be attained for 222 Rn and EVs in the region of 10 to 100 cm3 for 220 Rn. By sensitively monitoring outﬂow, an εc = 1 can be veriﬁed. High-resolution gamma spectrometry of the vessel offers excellent differentiation of individual nuclides of the decay chains. After the collection of Rn, the charcoal needs to be homogenized (by shaking) for quantitative γ determination. 222 Rn concentration by freezing (< −71 ◦ C) and γ-spectrometry has been carried out at standards laboratories with elaborate instrumentation and is not dealt with here. 222 Rn concentration onto activated charcoal, and transfer to a liquid scintillator for high efﬁciency αβ-spectrometry is well established. Normally, however, only small charcoal volumes are used and the collection εc · f , i.e. Rn uptake to saturation of the charcoal, is poorly quantiﬁable. Transfer of the 222 Rn to a scintillation ﬂask has also been used. These systems are 222 Rn speciﬁc and ﬁnd application where high accuracy is not required.

4. Flow through source measurements Repeated source differential type measurements (Section 3.1), shown diagrammatically in Fig. 1, were carried out on a dry powder 228 Th and on a dry powder 226 Ra ﬂow-through source. In order to reduce the relative variability of the ‘unknown’ position of the powder

Characterization of thoron and radon ﬂow-through sources

407

Fig. 1. Source differential type x Rn measurements. Table 1 Degree of emanation and yield of a 220 Rn and 222 Rn ﬂow-through source Measurements∗ εx Rn /x Rn available†

Manufacturer speciﬁcation ID/calibration certiﬁcate Th-1025, SN B-157, ε220

=1

0.17 σ 15% (11)

Rn-1025, SN A-349, ε222

=1

7.9 × 103 σ 3% (4) 220 Rn atoms s−1 0.40 σ 5% (5)

Rn 18.21 × 103 Bq ± 8% 228 Th‡ − 18.21 × 103 atoms 220 Rn s−1 Rn

20.3 × 103 Bq ± 4% 226 Ra − 20.3 × 103 atoms 222 Rn s−1

20.5 × 103 σ 4% (5) 222 Rn atoms s−1

∗ Number of runs in brackets; multiple measurements per run; yield standard deviation σ excludes uncertainty of γ reference standards. † At calibration certiﬁcate reference date. ‡ In the 220 Rn source transient equilibrium was established: Bq 224 Ra = 0.9985 × Bq 228 Th.

inside the source, a ﬁxed long-distance geometry of source to detector was used. Since the true source activity was unknown, only the degree of emanation could be determined and is listed in Table 1. The source manufacturer’s data sheet quotes ‘The material is in a form which emanates 100% of the gas produced. . . .’ For quantiﬁcation of the ﬂow-through sources, the 222 Rn or 220 Rn was pumped onto 160 g activated charcoal in a vessel as shown diagrammatically in Fig. 2. A γ-calibration curve was

Fig. 2. x Rn concentration on activated charcoal + subsequent γ-spectroscopic measurement.

408

R. Rolle et al.

established with nine reference nuclides on 160 g charcoal and this was complemented with a 222 Rn standard transferred to the charcoal in the vessel. Initial experiments had shown that the top charcoal layers collected no activity. For later measurements sensitive electrostatic recoil nucleus collecting monitors (effective volumes EV > 200 cm3 ) were routinely connected to the exit of the vessel to verify that all the x Rn had been retained by the charcoal (meaning that the collection efﬁciency is εc = 1) or to reject measurements. This monitoring removed the uncertainty of non-quantitative transfer. The precise volumes of the source, drier, ﬁlter and pre-cooler were used to evaluate the decay of 220 Rn in the transfer. The charcoal in the vessel was subsequently homogenized by shaking before HPGe γ-measurement. Results of these measurements are given in Table 1. 5. Results and discussion Table 1 shows the determined degree of emanation and the x Rn yield in atoms s−1 , which is equivalent to the Bq x Ra in the source providing the x Rn (all converted to the manufacturer’s source calibration reference date). The measured yield of the 220 Rn source is about 0.4 and the degree of emanation is about 0.2 of the alleged value, thus grossly disagreeing with the values stated on the supplied source calibration certiﬁcate. For the 222 Rn source, the measured yield agrees with the yield on the supplied calibration certiﬁcate, however, the degree of emanation is about 0.4 of manufacturer speciﬁed value of ε220 Rn = 1. With the measured ε220 Rn = 0.40 and yield of 20.5 × 103 222 Rn atoms s−1 the total 226 Ra activity of the source must be 20.5 × 103 ÷ 0.4 = 51 × 103 Bq contrasting with the 20.3 kBq 226 Ra stated on the manufacturer’s source calibration certiﬁcate. With the experimental setup shown in Fig. 2, cooling to −40 ◦ C was generally used to enhance the capacity of the activated charcoal to absorb Rn from a few m3 of air. This necessitated thorough drying of the air to prevent ice formation; several runs had failed due to ﬂow blockage. Surface adsorption drying agents also adsorb Rn, while this was not observed with the chemical drying agent CaCl2 . Two of the 220 Rn measurement runs listed in Table 1 were made at room temperature without drying. Conventional Rn monitors need to be calibrated at normal room humidity where they are to be used, and further work is in progress to determine the dependence of source yield and of monitors on air humidity.

6. Conclusions Considerable care is needed in the evaluation of radon measurement systems. Sensitive instrumentation needs to be chosen that can differentiate between 220 Rn, 222 Rn and other interferences. Care needs to be taken to evaluate all the disequilibrium situations as the measurements generally involve also the Rn decay products of the two decay chains. The adsorption behavior of the miniscule (carrier-free) quantities of Rn is largely dominated by co-adsorbants – thus in systems involving Rn adsorption the quantitative transfer needs to be carefully assessed. Dry ﬂow-through sources of 220 Rn and of 222 Rn are convenient for many experiments and for testing other equipment. Two such sources were calibrated via γ-spectrometry. Our measurements showed that the commercial supplier’s calibration of these was grossly in error.

409

Some aspects of the radon problem in Kazakhstan V.N. Sevostyanov Scientiﬁc company “SOLO Ltd.”, Dostyk St. 192B, 480051 Almaty, Republic of Kazakhstan

The work summarizes data from radon studies conducted in Kazakhstan in 1980–2001. For the purpose of preliminary identiﬁcation of radon risk level in Kazakhstan, geological data were analyzed regarding the regularity of the natural radionuclide distribution in rocks, as well as the abundances of radon and radionuclides in underground waters. As a result, we identiﬁed a number of radon risk areas in Kazakhstan with high dose rates to the population due to natural radiation sources.

1. Radon safety standards It should be noted that in Kazakhstan the following norms have been successively implemented: NRB-76 [1] starting in 1976, NRB-96 (1996) [2] and NRB-99 (1999) [3]. The allowable average annual potential alpha energy concentration (PAEC) of radon and thoron in air, in residential and public buildings, was established in Kazakhstan for the ﬁrst time following the introduction of NRB-96 [2]. The allowable levels for radon concentration were deﬁned as 100 Bq m−3 for constructing new buildings, and as 200 Bq m−3 for existing buildings on the hypothesis that the time for which the public stays indoors is 7000 h y−1 . Furthermore, according to NRB-99 [3], for rating, control and reduction of the exposure of the occupationally exposed, the annual average PAEC value of radon (Rn) and its isotopes, i.e. 222 Rn (radon) and 220 Rn (thoron), in the breathing zone, should be used. The radon isotope actinon 219 Rn is not considered, because of its rare occurrence only in speciﬁc production, due to the contamination of premises by treatment of uranic primary products that contain the selectively extracted radionuclides of 227 Ac and 231 Pa. NRB-99 suggests that the average annual value of PAEC of radon isotopes Aequiv is constituted of the annual average value of PAEC of radon (222 Rn) − ARn equiv plus the annual average value of PAEC of thoron (220 Rn) − ATn equiv , following the formula: Aequiv = ARn equiv + 4.6ATn equiv .

RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07047-0

(1)

© 2005 Elsevier Ltd. All rights reserved.

410

V.N. Sevostyanov

Table 1 Radon and its decay daughter product measurement devices Type

Measured physical magnitude

Measuring principle

RAMON-01O RAMON-02 RRA-01O

PAEC

Alpha-spectrometry

VA

TREK-REI-1 “Camera” equipment set: RRG-20P1

VA VA

Alpha-spectrometry of RaA sediment on electrostatic detector Tracking Gross gamma measurement of daughter products adsorbed in coal

Measuring accuracy, ± (%)

Made in

4–5 × 105

±30

Kazakhstan

20–2 × 104

±30

Russia

20–2 × 103 10–1 × 106

±50 ±30

Russia Russia

Measuring range (Bq m−3 )

For the occupationally exposed, two categories of personnel are deﬁned: A: facility operating personnel and B: workers who do not belong to the facility operating personnel. For personnel of category A, the allowable volumetric activity (AVA) of all radioactive daughter products of (222 Rn) and thoron (220 Rn) in the breathing zone is AVAA = 1240 Bq m−3 , while for personnel of category B, AVAB = 0.25AVAA = 310 Bq m−3 . The duration of occupational exposure is assumed to be 2000 h y−1 .

2. Radioecological survey criteria and instrumentation When designing a radiation-ecological investigation in building construction areas, the physical magnitude of radon-ﬂux density from the ground surface, measured in mBq m−12 s−1 , is used for estimating the potential radon risk. In the technical literature, this physical magnitude is referred to as radon “exhalation”. According to the Law of the Republic of Kazakhstan “On population radiation safety” [4], when constructing new residential and occupational buildings, it is necessary to investigate the construction area in order to determine the radon risk using soil gas radon and radon exhalation measurement. The radon safety criteria are the following [5]: at the construction area, radon exhalation should not exceed 80 mBq m−2 s−1 and soil gas radon content should not exceed 50 Bq L−1 . In case the above limits are exceeded, remedial measures should be applied. Table 1 summarizes the equipment used for the related measurements.

3. The problem in Kazakhstan 3.1. Kazakhstan zoning by natural radiation sources Based on geological data on the regularity of natural radionuclide distributions in rocks as well as on the distribution of radon and natural radionuclides in underground waters as presented in Tables 2–4, a variety of possible radon risk areas are identiﬁed in Kazakhstan. In these areas,

Some aspects of the radon problem in Kazakhstan

411

dose to the population due to natural radiation is considered to be elevated. These areas are primarily the following: • high-activity concentration rock areas with high radon exhalation in uranium ore and raremetal regions (Kokshetau, Akmola, Karaganda, Zhezkazgan regions) as well as in mountain and submountain regions of south and south-east of Kazakhstan (Zhambyl, Almaty, TaldyKorgan, Semipalatinsk, Eastern Kazakhstan regions); • areas of artesian basin waters with high contents of uranium and radium in Shu-Sarysu and Syrdarya uranium ore regions (Kzyl-Orda, Southern Kazakhstan regions); • separate sources of underground (thermal) waters of Western Kazakhstan, areas of several oil ﬁelds with high content of natural radionuclides in formation waters, creating radioactive contamination of the earth’s surface during oil production processes (Mangystau, Atyrau, West-Kazakhstan, Aktyubinsk region). Thus, the basic radon sources are: • • • • •

crude ore (uranic, rare-metal ores and some others); tectonic faults; water from water sources containing radon; construction materials from naturally radioactive rocks; soils and grounds, formed from naturally radioactive rocks or overlapping fault systems.

Taking into consideration the widespread of naturally radioactive rocks, uranium ore and rare-metal ores in several Kazakhstan regions, a great number of inhabited localities where high ground radon emission occurs, may be identiﬁed. 3.2. General geological characteristics of Kazakhstan territory The territory of Kazakhstan is a massif platform, parts of which – Tien Shan and Altai mountain systems – are subject to neotectonic movements. Folded foundation appears on original ground only within Kazakh shield and Ural ridge. According to the time of continental Earth crust formation, the foundation is divided into Caledonian (Kokshetau-North Tien Shan, Chingiz-Tarbagatay) and Hercynian (Ural, Jungar-Balkhash, Zaisan) folded systems. They are saturated with polyphase intrusive and volcanic formations, separate phases of which are characterized by increased contents of natural radionuclides (uranium, thorium, potassium) and are main sources of dose due to natural environmental radiation. 3.3. Rules of radionuclide distribution in rocks and soils of Kazakhstan The mechanisms determining the main natural radioactive element distributions in soils and rocks of Kazakhstan were studied in the ﬁfties–sixties [6,7]. It is well known that among rocks granites contain the largest contents of radioactive elements – uranium, thorium and potassium. In Kazakhstan, two stages of Paleozoic magmatism are identiﬁed: Low and MiddleUpper Paleozoic. Low Paleozoic magmatism and basic stages of Middle Paleozoic magmatism appeared in Kazakhstan–Tien Shan Caledonian middle massif. As regards time and composition, acidic intrusions are classiﬁed into three rock groups (Table 2), where radioactive rock features are given as well.

412

V.N. Sevostyanov

Table 2 Content of natural radionuclides in Caledonian magmatic cycle rocks Rocks

Quartz diorites, diorites, granodiorites (Ordovician–Silurian) Plagiogranites, granodiorites (Silurian–Middle Devonian) Biotite-corniferous and leucocratic granites

Number of studied intrusions

Uranium-238 (10−4 %) from–to average

Thorium-232 (10−4 %) from–to average

8

1.2–3.8

2.3

6–15

9

7

1.0–4.7

2.9

5–20

15

2.4–6.8

4.4

11–43

Potassium-40 (10−4 %) from–to average 1.2–2.3

1.8

15.3

2–4

2.3

27

2–3.5

2.7

Table 3 Content of natural radionuclides in Jungar–Balkhash region Rocks

Diorites, granodiorites and biotite granites Plagiogranites, granodiorites, alkaline granites Alaskite and leucocratic granites

Uranium-238 (10−4 %)

Thorium-232 (10−4 %)

Potassium-40 (10−4 %)

from–to

average

from–to

average

10−5 from–to

6

3.0–4.3

3.6

12–20

17

2.4–3.6

3.0

10

3.0–5.2

4.0

10–30

20

2.9–3.7

3.3

12

3.1–5.7

4.6

19–44

29

Number of studied intrusions

19–44

average

29

Middle and Upper Paleozoic granite intrusions are most developed in the Jungar–Balkhash region of Early Variscian folding. Among them, three rock groups are also identiﬁed (Table 3), where radioactive rock features are given as well. Granites of various stages of magmatic activity are characterized by the general rule of radioelement accumulation from early phases to late ones. This means that leucocratic granites are most radioactive among other granitoids. Clear-cut spatial inclination of intrusion with higher radionuclide content toward the areas of the greatest tectonic mobility and magmatic intensity is observed. On the general background of effusive and intrusive formations, ore zones of most elements are pronounced, characterized by radiochemical abnormalities, often with highly increased concentrations of uranium, thorium, potassium, that enhances the entry of radon into underground waters and soils. Soil radioactivity is associated with concentrations of uranium, thorium and potassium in mother beds and is usually directly correlated with them. Geochemical processes in soils lead either to enrichment or removal of chemical elements from them. The analysis of radioelement distributions in soils showed that the radionuclide contents depend on their location relative to intrusive–effusive rock complexes. Contents of radium, thorium and potassium are highest in Kazakh small mounds, where in vast areas acidic igneous rocks are developed, characterized by high radioactivity; the lowest contents are in the Caspian lowland, where sea-sands contain

Some aspects of the radon problem in Kazakhstan

413

Table 4 Natural radionuclide distribution in chestnut soils Distribution area

Caspian lowland Common syrt (plain) Kazakh small mounds

Radioelement content uranium (10−4 %)

thorium (10−4 %)

potassium (10−4 %)

1.6 1.9 2.2

9.0 9.2 10.0

1.9 1.9 2.0

little radionuclide as a result of washing up during migration and resedimentation of these sands (Table 4). The characteristics of radioelement distributions in other Kazakhstan soil types are similar. Moreover, the content of radioelement in saline clayey soils is higher than in sandy soils. At the same time, the content of radioelements in soils is much lower than in granites, because soils are derivative of all drift zone rocks. 3.4. Radon and natural radionuclides in underground waters of Kazakhstan The biggest part of Kazakhstan is characterized by a very dry climate and therefore water resources and primarily underground waters are very important for the country. The availability of water resources was the primary factor for the geography of towns and villages. Underground waters are the source of potable water supply in most regions of Kazakhstan and many settlements have no alternative water sources. However, near-surface groundwaters are in most cases saturated with uranium, radium, radon, the concentrations of which often exceed the allowable level and therefore water location inspection and water quality control are vital to preventing (reducing) population radiation exposure risk. In Kazakhstan there are many radon-rich water sources, the location of which usually is identiﬁed by the presence of highly radioactive rocks (enriched with uranium, radium) of magmatic formation. Some such sources are used in sanatoria and health resorts. Water sources with high radon content (over 370 Bq L−1 ) are located in submountain and mountain areas in the south and south-east of Kazakhstan, in uranium ore and rare-metal areas of southern, central and northern Kazakhstan, covering practically all the eastern half of Kazakhstan, enriched with high activity magmatic rocks. These provinces include Kendyktas–Chuili–Betbakdala uraniferous area (Jambyl, Almaty oblasts) and North Kazakhstan (Kokshetau and Akmola oblasts) as well as North-Balkhash rare-metal region (Zhezkazgan and Karaganda oblasts). In these areas, particularly in ore villages, radiological services registered the most frequent cases of above average radon concentrations in residential and production premises air. Artesian underground waters with high uranium and radium concentrations in Shu-Sarysu and Syrdarya uraniferous areas are also an important vector for radiation exposure risk (Kzylorda and South Kazakhstan oblasts). There, underground waters are the only source of potable water supply and radionuclide and radon content control is very important. Underground and deep thermal waters of West Kazakhstan basins also carry high concentrations of uranium and radium radionuclides. Large radioactive pollution areas formed in the course of oil production in a number of oil ﬁelds play a vital role there.

414

V.N. Sevostyanov

Radioecological surveys of several oil ﬁelds (total area 2528 km2 ) of Mangystau and Atyrau oblasts identiﬁed 230 zones of radioactive contamination (with equivalent dose over 1 mSv h−1 ) of total area 517 ha. An average value of equivalent dose of several mSv h−1 is found at naturally radioactive polluted areas, reaching 100–170 mSv h−1 in certain local spots. The basic pollutant here is 226 Ra, other less active pollutants being other uranium and thorium decay products. According to test data, ground alpha activity in anomaly epicenters reaches 338 000 Bq kg−1 . Increased concentrations of radionuclides (226 Ra) and occurrence of radioactive pollution are associated with oil production processes and occur in repeated formation water discharge locations in low relief areas, on pipe work walls, re-injection ﬁlters, oil slime discharge locations, at oil sump peripheries, etc. Taking into account the large radiation pollution areas with comparatively high equivalent dose rates in oil ﬁelds, the risk of dose-rate increase for oil production workers and nearby village populations arises. It should also be noted that radioactive impact risk in oil production areas, also polluted by hydrocarbons, multiplicatively increases due to “synergism” factors, the importance of which is beyond question nowadays. Potable water makes a signiﬁcant contribution to population exposure rate from natural sources of ionizing radiation. The results of independent studies show that the levels of natural uranium and thorium in several sources of potable water supply in South Kazakhstan, Kzylorda and Taldy-Korgan oblasts signiﬁcantly exceed the allowable levels: 3.1 Bq L−1 and 0.6 Bq L−1 for 238 U and 232 Th, respectively. However, there are still no complete studies of this issue by the state sanitary supervision bodies and it is impossible nowadays to give an objective country-wide evaluation of the radioactivity of potable water used by the population. In conclusion, an attempt was made, using the data of the previous paragraphs, to describe the radiation zoning of Kazakhstan by geological data taking into account information on the radioactivity of rocks, waters and uranium ore, rare-metal and other ore-bearing regions associated with natural and man-made radioactive anomalies and contamination (Table 5). The performed zoning should be considered as regional, since only large regions (oblasts) were identiﬁed, where more detailed radioecological surveys are required, using previously conducted special aerogeophysical, radiogeochemical and hydrogeological studies. Further investigation planning includes radioecological mapping to identify high radiation risk regions and within these regions, detailed radioecological studies in inhabited localities and industrial zones characterized by high population radiation loads. 3.5. Level of radon problem in Kazakhstan Today, in the course of investigating the radon concentration indoors in residential buildings, state supervision bodies and other radiological organizations identiﬁed a number of villages and towns where the PAEC values are higher than those allowable. These localities include Zhezkazgan, Akchatau, Aktogay villages in Zhezkazgan oblast and Aksai in West Kazakhstan oblast, and Gorny village in North Kazakhstan oblast. Also, according to surveys conducted in 1994–1995 in the framework of a state approved project for planning of procedures on improvement of the radiation situation in the Republic of Kazakhstan, 64 settlements (regional centers, regional villages) in Kokshetau, Akmola, Karaganda and Zhezkazgan oblasts are identiﬁed as potential radon risk zones

Some aspects of the radon problem in Kazakhstan

415

Table 5 Radiation risk zone areas in Republic oblasts Area (thousand km2 ) Oblast

West Kazakhstan Aktyubinsk Atyrau Mangystau Kostanay North Kazakhstan Kokshetau Turgay Akmola Pavlodar Karaganda Zhezkazgan Kzylorda South Kazakhstan Jambyl Almaty Taldy-Korgan Semipalatinsk East Kazakhstan Total

Population million people

Total

Radioactive anomalies (including occupied)

High-activity under-ground waters

Uranic ore

Rare-metal, polymetal

0.595 0.651 0.376 0.268 0.962 0.580 0.624 0.285 0.818 0.836 1.273 0.456 0.582 1.617 0.952 1.829 0.675 0.784 0.890

151.2 298.7 112.0 166.2 114.5 44.3 78.1 111.9 124.6 127.5 854 313.4 228.1 116.3 144.6 104.7 118.5 179.6 97.3

– 33.6 (3.4) – – 0.5 (0.3) 1.3 (1.2) 9.4 (3.2) 2.0 (0.4) 7.4 (2.0) 77.7 (2.7) 257 (3.4) 67.3 (18.0) – 0.2 (0.2) 16.2 (7.2) 5.5 (2.7) 8.1 (4.7) 57.1 (5.8) 34.5 (13.1)

1.0 2.0 2.0 3.0 – – 30.0 – 10.0 – 20.0 10.0 60.0 60.0 10.0 10.0 40.0 40.0 –

– – – 1.0 – 1.0 70.0 – 10.0 – 15.0 10.0 60.0 60.0 30.0 10.0 – – 30.0

– – – – – 10.0 – – – – – – – – – – – – –

15.053

2717.3

256.6 (68.3)

298.0

251.0

10.0

according to geological criteria. Soil radon concentrations in the areas where high natural radionuclide contents and tectonic faults occur are high. Cases were registered where soil radon concentration reached up to 300 000 Bq m−3 and indoor concentration was found to be 6000 Bq m−3 or higher [8]. Flux density reached up to several hundreds mBq m−2 s−1 . According to the geological criteria, the number of inhabited localities in Kazakhstan, where soil radon concentration is over 5 kBq m−3 with ﬂux density over 80 mBq m−2 s−1 is quite large. Therefore, during construction of new buildings in such areas it is necessary to consider radon preventive construction at the building foundation. In some towns and villages, radon concentrations indoors in up to 70% of surveyed buildings exceeded the allowable level of 200 Bq m−3 . These towns are: Zhezkazgan, Makinsk, Schuchinsk, Akchatau, Aktogay, Aryk-Balyk, Balkashino. It should be noticed that in kindergartens of Aryk-Balyk village radon concentrations range from 510 to 4500 Bq m−3 . Full-scale studies which allowed obtaining generalized data on radon inﬂuence on population sickness rate in residential and occupational areas were conducted in Akchatau village in Zhezkazgan oblast in 1985 [9]. An elevated natural gamma-background was discovered in the residential area along with high radon concentrations due to its emanation from the ground and local construction materials. All housing stock was surveyed in detail in Akchatau village

416

V.N. Sevostyanov

and based on the results a governmental decision on removal of people from poor quality ﬂats to new ones was taken and realized. As regards the other settlements, the number of conducted investigations is insufﬁcient for complete and objective evaluation of population exposure rates in order to take reasonable decisions on their reduction. The reduction of population radon exposure level greatly depends on the construction characteristics of building foundations – subsurface space and indoor air exchange. In existing foundation constructions, the radon concentration reduction factor relative to its concentration in soil air according to preliminary analysis of village single-storey house rooms without basements is on average 8 ± 4, for houses with basements 36 ± 14. Basement being an intermediate link between the ground and the above ground section of the building reduces the concentration of radon 5 ± 3 times. Reduction factors do not depend on the soil radon concentration. Besides soil, other indoor air radon sources are the construction materials. According to NRB-99 [3], all construction materials are divided into four radiation risk categories, among which only category I is acceptable for use in residential and public building construction. At the present time, not all of the enterprises engaged in production, manufacturing and sale of construction materials and concentrates have certiﬁcates of compliance, which indicate the radiation risk category of materials for residential building construction. Category II products are manufactured in ﬁve Republic oblasts, category III products are manufactured by Makinsk quarry. The use of their products in construction increases radon concentrations in buildings. Moreover, there are no production test laboratories in Kazakhstan for control over construction materials radioactivity. Further to their exposure due to radon and building materials, a substantial group of the population is exposed to radon in workplaces. During 1985–1987 ore mining and processing enterprises of Akchatau, Zhezkazgantsvetmet, Eastern-Kaunrad mine were surveyed. It was found that the most exposed workers in the Republic of Kazakhstan are the non-ferrous and ore metal miners; not the workers of uranium mining and uranium treatment enterprises. This is explained by the fact that there exists no control not only of radon concentrations in the mining and processing work places but also of the high concentrations of naturally radioactive dust in the air. In some cases, exposure doses for lungs of underground miners exceeded 5 Sv y−1 , while the allowable level is 0.015 Sv y−1 .

4. Radon studies in Almaty In 1998–2001 a radon survey was conducted in Almaty [10], at 200 sites, primarily in preschool and school buildings; some private houses were examined as well. Almaty city area is characterized by the fact that it is located on vast zones of tectonic faults. The city itself, former capital of Kazakhstan, is situated in the south-east of Kazakhstan, on the northern foothills of Zailisky Alatau. The city population is over 1 000 000 people. The ﬁrst measurements showed that increased levels of radon emission are associated with existing tectonic faults. Therefore, the buildings located in the tectonic fault zones were selected for studies. Also, some buildings outside the faults were studied.

Some aspects of the radon problem in Kazakhstan

417

The survey results led to the following conclusions: (a) Zones of tectonic ruptures in the territory of Almaty City are considered as radon risk areas and buildings inside these zones are characterized by an increased PAEC content. (b) The measurement of the unattached radon progeny concentration in air shows that not considering this parameter may result in serious undervaluation of exposure from radon. (c) For buildings with high radon concentrations, the application of ordinary preventative measures, such as more extensive ventilation, succeeded in sufﬁciently decreasing the PAEC levels to allowable values.

5. Some general results The results of the studies conducted in Kazakhstan on radon volumetric activity in buildings and soil are given as charts in Figs. 1 and 2, respectively.

Fig. 1. Radon distribution in dwellings (Kazakhstan), log-normal, GM = 204 Bq m−3 , GSD = 1.31.

Fig. 2. Radon distribution in soils (Kazakhstan), log-normal, GM = 14 700 Bq m−3 , GSD = 1.4.

418

V.N. Sevostyanov

6. Summary and conclusion Preliminary analysis shows that the population of Kazakhstan lives primarily in potentially dangerous areas from the radioactivity exposure point of view. According to selected studies, radon isotope PAEC values in premises where people stay for long times may exceed allowable levels by several times. At the same time, on account of insufﬁcient organization of work on a national scale, lack of funds, standard methods and metrology equipment, no radon studies are actually performed even in those regions where geological rock structure and availability of tectonic faults suggest that an excess radon concentration in residential and public buildings may be anticipated. Preliminary recorded data from independent studies show that increased radon concentrations in living premises are typical of the Republic. The survey of population exposure from natural sources of ionizing radiation in Akchatau village in Zhezkazgan oblast and the survey done in Almaty are insufﬁcient to understand and solve the problem in Kazakhstan in general. In light of the above, a conclusion may be drawn that for the complete study and solution of the problem it is necessary to elaborate a Comprehensive Republic Program on the reduction of the rate of population exposure from natural sources of ionizing radiation and to apply the following appropriate procedures: – Zoning of Kazakhstan according to radon exposure risks and preparation of detailed radioecological maps. – Study of radon emanation from soil and sedimentary rocks in radon risk regions and development of forecast methods of radon penetration in buildings. – Perform large-scale surveys of the housing stock in potential radon risk regions of Kazakhstan and of the uranium and non-uranium mines. Identify population exposure rates and risk groups. – Work out forecast methods of exposure consequences and their prevention. – Provide project operators with necessary equipment, techniques and measurement accuracy traceability. – Develop engineering and construction of radon preventive structures for building foundations and include these in construction standards. – Create a compulsory control network over the level of dust-radiation exposure factor for enterprises dealing with extraction and treatment of phosphorous ores, nonferrous and rare metals. – Establish a network of certiﬁcation test laboratories to monitor and control radioactivity in construction materials.

References [1] [2] [3] [4] [5]

Radiation safety standards NRB-76, Atomizdat, Moscow, pp. 19–20. Radiation safety standards (NRB)-96, GN 2.6.1.054-96, Almaty, 1997, p. 85. Radiation safety standards (NRB)-99, SP 2.6.1.758-99, Almaty, 2000, p. 80. Act of the Republic of Kazakhstan “On population radiation safety” N 219-1 ZRK, Akmola, April 23, 1998. Potential radon risk area criteria, LLP “SOLO”, Almaty, Guideline, p. 41. Approved by the Ministry of Ecology and Natural Resources and Ministry of Health of RK, 1997.

Some aspects of the radon problem in Kazakhstan

419

[6] Regional aerogamma-spectral studies experience, granitoid radioactivity in Kazakhstan, in: Field GammaSpectrometry Methods, Hydrometeoizdat, Leningrad, 1965. [7] L.I. Boltneva, L.I. Buyanova, et al., Regarding some rules of radioelement distribution in USSR soils, in: Field Gamma-Spectrometry Methods, Hydrometeoizdat, Leningrad, 1965. [8] H. Sirazhet, I.V. Kazachevsky, G.I. Krasnov, T.N. Madiyanov, T.S. Saibekov, L.N. Smirin, Determination of radon concentration in soil, premises and construction area classiﬁcation, Preprint of the Nuclear Physics Institute of the State Nuclear Center N2-96, Almaty, 1996. [9] B.S. Baiserkin, Comprehensive hygiene-sanitary and clinical evaluation of chronic inﬂuence of various radon gas concentrations on population and involved workers’ health, Thesis for Candidate of Medical Sciences, Almaty, 1996. [10] V.V. Abelentsev, V.N. Sevostyanov, G.D. Vdovichenko, G.Ch. Oghahliev, The results of radon survey in Almaty (Kazakhstan), in: Proc. Third Eurosymposium on Protection against Radon, Liege, Belgium, 10 and 11 May 2001, pp. 229–232.

420

Studies on the short-lived radon decay products: The inﬂuence of the unattached fraction on the measurement of the activity size distribution M. Gründel a , A. Reineking b , J. Porstendörfer c a Institute of Physical Chemistry, Georg August University Göttingen, Tammannstr. 6, 37077 Göttingen, Germany b Research Center Forest Ecosystems, Büskenweg 1, 37077 Göttingen, Germany c Am Hirtenberg 8, 37136 Waake, Germany

The activity size distributions of the short-lived radon decay products were examined with a low-pressure online alpha cascade impactor. The inﬂuence of the unattached fraction on the measurement was checked. In this study, the coarse mode of the activity size distribution inferred in earlier work [2] is examined more exactly.

1. Introduction For the determination of the natural radiation dose from the short-lived radon decay products 218 Po, 214 Po and 214 Bi/214 Po information about the activity size distribution with reference to the diameter of the aerosols is important. The diameter inﬂuences the diffusion constant and therefore the probability of deposition in the lung. For this reason the activity size distribution is of interest for the dose calculation. Due to this relevance many investigations of the activity size distribution have already been conducted. These measurements however resulted in partially contradictory predicates. With some a bimodal size distribution of the short-lived radon decay products was found [1], with others additionally a coarse mode within the μm-area was found (e.g. [2]). It was assumed in a previous paper [3] that the signal of the coarse mode is a result of the unattached fraction which separates in the upper impactor stages. This hypothesis is to be checked in this paper. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07048-2

© 2005 Elsevier Ltd. All rights reserved.

Studies on the short-lived radon decay products

421

2. Measurement technique In order to receive an exact signal about the activity size distribution, the activity size distribution of the aerosol associated activities was measured by a low-pressure online alpha cascade impactor (OLACI), which was developed and built at the institute [4]. This impactor has a good resolution in the area of the accumulation mode and the coarse mode. Each separating stage consists of an injector plate with drillings, a foil carrier and a detector including alpha spectroscopy. The measuring air is accelerated in the nozzles and steered onto the foil surface. Depending on the inertia of the aerosol particles these can follow the airﬂow or become separated. The decays on the foil are registered by α-spectroscopy. In order to get an even distribution of aerosols on the foil, the foil is rotated. This leads to a better energy resolution of the spectra. With this arrangement the activity size distribution of the short-lived 218 Po mainly can be measured. The impactor system permits a continuous measurement of the short-lived radon decay products in outside air. Thereby sufﬁcient statistics can be collected in order to be able to meet predicates about the examined diameter range of the aerosols. In Table 1 the 50% cut-off diameters of the impactor stages are shown. In order to examine the inﬂuence of the unattached fraction on the measurement, it is important to be able to remove these well without inﬂuencing the aerosols. For this reason, a tube diffusion battery was used as input stage of the impactor. This tube diffusion battery removes the unattached fraction. This is important, since the unattached fraction separated in the upper stages of the impactor [3]. Thus the measuring signal of the coarse particles was falsiﬁed. So that the coarse particles did not get stuck on the entrance of the impactor, the tube diffusion battery must be accurately perpendicular. The length of the diffusion tube depends on the 50% cut-off diameter of the particles. With an exactly perpendicular tube, the deposition of the coarse particles on the tube sheets can be neglected. A screen cannot be used for the removal of the unattached fraction. With a screen, coarse particles are also collected. Thus an inﬂuence of the relative activity size distribution was present and the proportion of the coarse particles was underestimated. The measurement setup can thus be represented as in Fig. 1.

Table 1 50% cut-off diameters of the impactor stages (additionally, the velocity in the jets and the pressure ratio of the stage with reference to the external pressure are shown) Stage

v (m s−1 )

p/p0

d50 (nm)

8 7 6 5 4 3 2 1

2.8 8.9 15.6 28.1 72.6 159.8 279.6 288.7

1.000 0.999 0.998 0.993 0.960 0.812 0.459 0.220

8845 3513 1833 966 505 275 130 58

422

M. Gründel et al.

Fig. 1. Schematic setup of an impactor measurement with the tube diffusion battery.

3. Results The data on the measured size distributions were approximated by the sum of two lognormal distributions characterised by the activity median aerodynamic diameters (AMAD) and geometric standard deviations (σg ) and using the simplex algorithm taking into account the aerosol deposition probabilities of the impactor stages. The measurements were carried out with the tube diffusion battery at the entrance of the impactor. These measurements for 218 Po and 214 Po resulted in a bimodal structure of the activity size distribution (nucleation and accumulation mode). The accumulation mode with an activity median aerodynamic diameter (AMAD) near 340 nm was the main mode with a fraction of 80%. With an AMAD near 30 nm was a nucleation mode with 20%. This area cannot be resolved with the impactor used, since the activity is completely on the backup ﬁlter. A signiﬁcant coarse mode in the μm-size range was not found. For an example, the activity size distribution of the 218 Po is illustrated in Fig. 2. In test measurements without the tube diffusion battery, the accumulation mode and the nucleation mode are identiﬁed again. In addition, an apparently coarse mode could be indicated for the 218 Po (Fig. 3), which does not occur with the 214 Po. The median and the standard deviation were similar to the distribution received from the measurements with the tube diffusion battery. This result seems to be a misinterpretation of the unattached decay products. In a previous publication [3], it was shown that the unattached fraction of the radon decay products is collected in the upper stages of the impactor. This behaviour would explain the deviation from earlier measurements [2], where a coarse mode was found. The unattached fraction is now removed in the tube diffusion battery.

Studies on the short-lived radon decay products

423

Fig. 2. Activity size distribution of the 218 Po. At the entrance of the impactor was a tube diffusion battery to separate the unattached fraction. The distribution shows two modes with median at 25 nm and 320 nm (nucleation mode and accumulation mode). A coarse mode in the μm-size range was not found.

Fig. 3. Without the tube diffusion battery at the entrance, the activity size distribution of the 218 Po has a nucleation mode, an accumulation mode and an additionally coarse mode.

4. Conclusion The activity size distribution of the short-lived radon decay products 218 Po and 214 Po were measured by an online alpha cascade impactor including tube diffusion battery. In the measurements, a bimodal structure were found without a coarse mode. Without the tube diffusion battery, the unattached fraction of the 218 Po was collected in the upper stages. For this reason, misinterpretation of the unattached fraction as a coarse mode was possible.

424

M. Gründel et al.

References [1] G. Butterweck, Natürliche Radionuklide als Tracer zur Messung des Turbulenten Austausches und der trockenen Deposition in der Umwelt, Dissertation, Universität Göttingen, 1991. [2] Ch. Zock, Die Messung der Aktivitätsgrößenverteilung des radioaktiven Aerosols der Radonfolgeprodukte und deren Einﬂuß auf die Strahlendosis beim Menschen, Dissertation, Universität Göttingen; Cuvillier Verlag, Göttingen, 1996. [3] A. Reineking, K.H. Becker, J. Porstendörfer, Measurements of activity size distributions of the short-lived radon daughters in the indoor and outdoor environment, Radiat. Prot. Dosim. (1988) 245–250. [4] J. Kesten, G. Butterweck, J. Porstendörfer, A. Reineking, H.-J. Heymel, An online α-impactor for short-lived radon daughters, Aerosol Sci. Technol. 18 (1993) 156–164.

425

Radon exposure of the Greek population D. Nikolopoulos, A. Louizi, A. Serefoglou, J. Malamitsi Medical Physics Department, University of Athens, Mikras Asias 75, Goudi, 11527 Athens, Greece

A radon survey has been carried out from 1995 to 1998 in Greece in order to study the exposure of the Greek population. The total sample size was 1137 dwellings. Values of residential potential alpha energy exposure ranged between (0.024 ± 0.009) and (8 ± 1) WLM per year (p < 0.05). Values of effective dose ranged between (0.09 ± 0.04) and (28 ± 4) mSv (p < 0.05). The mean lifetime risk of the Greek population was found to be equal to 0.4%.

1. Introduction Small-scale measurements of radon gas concentration within dwellings in Greece have been reported [1–5]. Continuing the radon investigation conducted by the Medical Physics Department of University of Athens (MPD-UOA) since 1988, a large-scale nation-wide residential radon survey in Greece was designed and performed, using dosimeters of MPD-UOA construction, fully calibrated and tested by the MPD-UOA [6]. The main scope was to obtain an adequate estimation of the annual radon concentration distribution indoors, to assess the average risk, to determine the percentage of dwellings in which radon concentrations exceed certain reference levels and to investigate some factors that affect indoor radon concentrations.

2. Materials and methods 2.1. Statistical data The most recently published statistical data is the 1991 census, according to which Greece had a population of 10.4 million people [7]. The population was highly concentrated in urban areas, and mainly in Athens where about 37% of the total population resided. The total number of buildings was 3.8 million. Approximately 75% (2.85 million) of the buildings were used as dwellings. The buildings were classiﬁed by the National Statistical Service of Greece (NSSG) according to their use. Building data were provided nation-wide according to an administrative RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07049-4

© 2005 Elsevier Ltd. All rights reserved.

426

D. Nikolopoulos et al.

partition proposed by the NSSG. For those used as residencies, no data about the number of stories, building attributes and occupational status of each separate dwelling was provided by the NSSG, except for a part of the capital (Athens). For the rest of the country, the data were given only in an accumulative manner. Moreover, the NSSG provided no maps through which the geographical coordinates of each house could be found. 2.2. Sampling design Due to the limitations placed by the statistical data and taking into account the ﬁnancial and work power of the MPD-UOA, the radon survey was not based on a grid division but was administratively designed. The sampling design was based on the partition proposed by the NSSG. Ten [10] administrative districts, called regions, were divided into prefectures, called departments, which were subdivided into provinces. These were organized in municipalities and communes, which included villages and city quarters. The sampling design was based on the following procedure: a sampling density of 1 per 1000 dwellings was adopted, balancing both feasibility and precision of estimation. The number of samples was calculated for a department, according to the total number of its dwellings. This number was further allocated to each province, in proportion to the fraction of the number of dwellings of the province over the total number of dwellings of the department. Continuing, the sample number of each province was allocated to its municipalities and communes, in proportion to the number of the dwellings of the municipality or commune over the total number of dwellings of the province. The procedure was followed so as to allocate an adequate number of samples to each village or city quarter [8]. 2.3. Experimental apparatus The experimental apparatus was the MPD radon dosimeter [9]. The dosimeter consisted of a cylindrical non-conductive plastic cup of 5 cm height and 1.5 cm radius. The cover had a 3 mm hole in the center and a ﬁlter that prevented radon daughters from entering. Radon was detected by a 2 × 2 cm CR-39 nuclear track detector placed at the bottom of the cup. The overall uncertainty of radon measurement in the 95% conﬁdence interval was below 10% [6]. The 12-month exposure period was selected due to the best estimation of the average value it provides. One detector was installed in each sampled dwelling, placed in the bedroom 1 m above the ground, near the wall. 2.4. Measurement procedure Detectors were installed by trained personnel. A door-to-door approach was selected, so as to minimize non-response and bias. This scheme was generally followed and changed only by restrictions placed at the implementation stage (i.e. refusals and other difﬁculties). Within every sampling location, dwellings were selected by the personnel, so as to sample nation-wide all types of buildings. In each case, a questionnaire was ﬁlled and the inhabitant was given informative brochures. At the end of the 12-month period, the dosimeters were collected, either via door-to-door approach or by post.

Radon exposure of the Greek population

427

3. Results The survey was carried out between July of 1995 and August 1998 with the installation of 1500 MPD dosimeters in 834 locations (i.e. villages and city quarters) resulting in the sampling of 950 dwellings in 722 locations. The data included an additional 216 samples in 12 locations collected between 1988 and 1994 by other MPD-UOA investigators [1–3], resulting in a total of 1137 samples in 734 locations within Greece. Broad sampling was performed in South Greece, i.e. Attica Department, Peloponnese and the island of Crete covering about 40% of the Greek territory and about 50% of the Greek population, while local sampling occurred in all other investigated areas. Sample density ranged between 1/271 dwellings and 1/10003 dwellings with an average of 1/2405 dwellings excepting the capital province within the West Attica prefecture where the sample density was 1/46018 dwellings and the four provinces where no dosimeters were ﬁnally collected. With the above exceptions, the sample density is representative according to [10] and is comparable to that of the other similarly designed surveys based on statistically representative sampling [11–15]. The results are given as a frequency distribution histogram in Fig. 1. Introducing the χ 2 test, the overall results follow the log-normal distribution (p < 0.01). Residential radon concentration ranged between 200 and 400 Bq m−3 in 22 dwellings (1.9%), 400 and 1000 Bq m−3 in 8 (0.7%) dwellings, and above 1000 Bq m−3 in 4 (0.4%) dwellings. In the full data set, the arithmetic mean was found to be 55 Bq m−3 and the geometric mean 44.0 Bq m−3 with a geometric standard deviation of 2.4 Bq m−3 . In only a small percentage (1.1%) of dwellings in Greece did the measured radon concentrations exceed the action level proposed by [16] (400 Bq m−3 ). Through the questionnaires it was found that from the full data set 527 dwellings located were on the ground ﬂoor, 334 on the ﬁrst ﬂoor, and 89 above the ﬁrst ﬂoor of a building. Among these categories the one-way analysis of variance (ANOVA) method was applied to the logarithms of the radon concentrations, which follow a Gaussian distribution. Ground ﬂoor dwellings presented statistically signiﬁcant higher radon concentrations but for the dwellings

Fig. 1. Frequency distribution histogram of indoor radon concentrations in Greece. Sample size 1137.

428

D. Nikolopoulos et al.

of the ﬁrst ﬂoor and above, the differences were not signiﬁcant (p < 0.001). Moreover, 529 dwellings were constructed from brick and concrete, 109 from concrete only, 254 from stone and 71 from mixed building materials. By application of the same method it was found that the dwellings constructed with stone presented statistically signiﬁcant higher concentrations (p < 0.005). Applying the same method to the ground ﬂoor dwelling, radon concentration data of each surveyed area it was found that some areas presented statistically signiﬁcant differences in radon concentrations (p < 0.001). In some of these areas, the residential radon concentrations lie in the tail (p < 0.01) of the log-normal distribution (Fig. 1). From these only two (2), i.e. Arnea Chalkidikis and Vrisses Apokoronou Chanion, are “radon prone” areas according to [17]. Residential Potential Alpha Energy Exposure (PAEE) and effective dose values may be calculated from the above data set by using appropriate values for the equilibrium, occupancy and dose conversion factor. Since no such values are available for Greece, a mean equilibrium value of 0.4 [18] and an occupancy factor of 0.8 [10], respectively, were used to estimate risks. PAEC values were calculated using 72 WLM y−1 /(Bq m−3 ) of mean annual equivalent radon concentration, and effective doses using 6 nSv h−1 /(Bq m−3 ) of mean annual equivalent radon concentration as a dose conversion factor. In South Greece, where broad area sampling was performed, residential PAEC values ranged between (0.024 ± 0.009) and (2.8 ± 1.0) WLM per year (p < 0.05) with a mean of 0.2 WLM per year. Effective doses were between (0.09 ± 0.04) and (11 ± 4) mSv per year (p < 0.05), with a mean of 0.8 mSv per year. These mean values lie far beyond the maximum values of (8 ± 1) WLM per year and (28 ± 4) mSv per year (p < 0.05) that occurred in the radon prone area of Arnea Chalkidikis. Using the risk factor of 2.8 × 10−4 per WLM [18] according to epidemiological data, a mean PAEC value of 0.2 WLM per year assuming that this is representative for Greece, due to the broad sampling performed there and the mean life expectancy of 74 y for men and 77 y for women in Greece [19], the mean lifetime risk in Greece due to residential radon is 0.4% (0–1.1% in the 95% conﬁdence interval). This means that on average 40 out of every 10 000 inhabitants of Greece would die due to lung cancer caused by residential radon exposure. Since Greece has a population of 10.4 million people, it may be calculated that on average 400 mortal lung cancers due to residential radon are expected to occur each year in Greece. Mean lifetime risk was calculated excluding the data from the rest of the country because the sampling there was not statistically representative according to [10]. The uncertainties of PAEC and effective dose values were calculated taking into account the instrumental uncertainty of the MPD radon dosimeter and the statistical ﬂuctuations of the recorded concentrations within every surveyed area. Mean lifetime risk uncertainty was calculated taking into account the ﬂuctuations of the calculated PAEC values in South Greece. Both uncertainties are biased by the uncertainties in the dosimetric conversion factors [20,21]. Moreover, mean lifetime risk is biased by age, smoking habits [20] and by uncertainties in the mean life expectancy in Greece.

4. Conclusions According to the results obtained, it was found that only a small percentage of dwellings appeared to have annual average radon levels above 400 Bq m−3 , which is the action level

Radon exposure of the Greek population

429

proposed by the European Community. The survey supports the recommendation of testing mainly ground ﬂoor or ﬁrst ﬂoor dwellings, since there were not found to be signiﬁcant differences in radon concentrations among the dwellings of the upper ﬂoors. In addition, the radon estimates have shown geographical differences, leading to support for the strategy of focusing on areas with high radon potential. The survey is still in progress, because on the one hand, different results may be obtained elsewhere and on the other hand, a better estimation of the national average should be determined. Moreover, geological and other relevant data are being collected by MPD-UOA. These will be combined in future with the questionnaire data, in order to investigate the factors that affect indoor radon concentrations in Greece. The calculations of the mean nation-wide annual risk due to residential radon were based only on broad area sampling (about 40% of the Greek territory and about 50% of the Greek population) because the use of data from local sampling may have introduced a systematic error if these had represented over- or under-estimations of the mean radon concentration of each surveyed area. Nevertheless, elevated residential radon concentrations may be found in the non-broadly surveyed part of Greece.

References [1] C. Proukakis, M. Molfetas, K. Ntalles, E. Georgiou, E. Serefoglou, Indoor radon measurements in Athens, Greece, in: British Nuclear Energy Society Conference on Health Effects of Low Dose Ionizing Radiation – Recent Advances and Their Implications, BNSE, London, 1988, pp. 177–178. [2] E. Georgiou, K. Ntalles, A. Molfetas, A. Athanassiadis, C. Proukakis, Radon measurements in Greece, in: Radiation Protection Practice, IRPA 7, vol. I, Pergamon Press, New York, 1988, pp. 125–126. [3] E. Georgiou, K. Ntalles, G. Anagnostopoulos, C. Proukakis, A. Athanassiadis, Comparative study of 222 Rn concentrations in ancient and contemporary mixed sulﬁde (silver) mines at Laurium (Attica Greece), Nucl. Med. 25 (A) (1988) 136–141. [4] C. Papastefanou, S. Stoulos, M. Manolopoulou, A. Ioannidou, S. Charalambous, Indoor radon concentrations in Greek apartment dwellings, Health Phys. 66 (3) (1994) 270–273. [5] K.G. Ioannides, K.C. Stamoulis, C.A. Papachristodoulou, A survey of 222 Rn concentrations in dwellings of the town of Metsovo in North-Western Greece, Health Phys. 79 (6) (2000) 697–702. [6] D. Nikolopoulos, A. Louizi, N. Petropoulos, S. Simopoulos, C. Proukakis, Experimental study of the response of cup-type radon dosemeters, Radiat. Prot. Dosim. 83 (3) (1999) 263–266. [7] National Statistical Service of Greece, The 1991 Census, National Press, Athens, 1995 (in Greek). [8] D. Nikolopoulos, S. Maddison, A. Louizi, C. Proukakis, Radon survey in Kriti–Greece. Design implementation and results, in: Proceedings of the European Conference on Protection against Radon at Home and at Work, Praha, 1997, pp. 156–159. [9] D. Nikolopoulos, A. Louizi, D. Papadimitriou, C. Proukakis, Study of the calibration of the Medical Physics Department Radon Dosimeter in a radon facility, in: Extended Abstracts of the Second Regional Mediterranean Congress on Radiation Protection, Tel-Aviv, 1997, pp. 130–133. [10] UNSCEAR, Sources and Effects of Ionizing Radiation, Report to the General Assembly, United Nations, New York, 1993, Sales Publication E.94.IX.2. [11] P. McLaughlin, J. Wasiolek, Radon levels in Irish dwellings, Radiat. Prot. Dosim. 24 (1) (1988) 383–386. [12] K. Ulbak, B. Stenum, A. Sorensen, B. Majborn, L. Botter-Jensen, S. Nielsen, Results from the Danish indoor radiation survey, Radiat. Prot. Dosim. 24 (1/4) (1988) 401–405. [13] K. Langroo, N. Wise, C. Duggleby, H. Kotler, A nationwide survey of 222 Rn and α radiation levels in Australian homes, Health Phys. 61 (6) (1991) 753–761. [14] F. Marcinowski, M. Lucas, M. Yeager, National and regional distributions of airborne radon concentrations in U.S. homes, Health Phys. 66 (6) (1994) 699–706.

430

D. Nikolopoulos et al.

[15] F. Bochicchio, G. Campos-Venuti, C. Nuccetelli, S. Piermattei, S. Risica, L. Tommasino, G. Torri, Results of the representative Italian national survey of radon indoors, Health Phys. 71 (5) (1996) 741–748. [16] European Commission, Recommendation 90/143/Euratom of the 21 February 1990 on the protection of the public against indoor exposure to radon, Ofﬁcial J. Eur. Commun. Ser. L 080 (1990) 26–28. [17] M. Kendall, H. Miles, D. Cliff, R. Green, R. Muirhead, W. Dixon, R. Lomas, M. Goodridge, Exposure to Radon in UK Dwellings, NRPB-R272, HMSO, London, UK, 1994. [18] ICRP Publication 65: Protection against radon-222 at home and at work, Ann. ICRP 23 (2) (1993). [19] K. Katsougianni, M. Koyevinas, N. Dontas, P. Maisonneuve, P. Boyle, D. Trichopoulos, Mortality Due to Malignant Neoplasms in Greece 1960–1985, National Anticancer Union Publication, Athens, 1990 (in Greek). [20] W.W. Nazaroff, A.V. Nero, Radon and Its Decay Products in Indoor Air, Wiley, New York, 1988. [21] A. Louizi, D. Nikolopoulos, Health risks due to radon, Iatriki 73 (4) (1998) 341–345 (in Greek).

431

Preliminary study of two high radon areas in Greece A. Louizi, D. Nikolopoulos, A. Serefoglou, J. Malamitsi Medical Physics Department, University of Athens, Mikras Asias 75, Goudi, 11527 Athens, Greece

Following the survey research carried out by our Department, we have further investigated two areas in Greece, which had shown relatively high residential radon concentrations. These areas are the city of Arnaia in the department of Chalkidikis in northern Greece and Vrisses Apokoronou in the department of Chania in Crete. In Arnaia Chalkidikis, we carried out radon measurements in 47 (∼ 5%) of about 1000 dwellings and in Vrisses Apokoronou in 13 (∼ 4%) of about 350 dwellings. In dwellings with maximum radon concentration, active measurements using two modern monitors (ALPHA GUARD, Genitron Instruments & EQF3023, Sarad Instruments) were performed, estimating the variation of indoor radon and daughter nuclei concentrations with time and other parameters. The measurements were performed for a 1-month period in the case of Arnaia and a 10-day period in the case of Vrisses Apokoronou. In both cases, we have further carried out soil gas measurements and potable water measurements using Alpha Guard. The recorded soil gas concentrations lie mainly in the range of 10–200 kBq m−3 in both cases. The recorded values lie far above the 10–40 kBq m−3 soil gas concentration usually recorded. Moreover, potable water concentrations were found to be in the range 1–450 Bq L−1 .

1. Introduction Up to this day, in Greece, radon concentration measurements in indoor air have been reported by various researchers [1–3] and our Laboratory [4,5]; they are relatively low (of about 40 Bq m−3 ). From the measurements of our Laboratory it has been established that some areas show increased radon concentration [6]. Areas of increased radon concentrations have been characterized as those for which log(C) > 1.65 σ , where C is the measured 222 Rn concentration and σ the geometrical standard deviation. These areas are shown in Fig. 1. From those deﬁned as “radon prone areas” have been considered those in which at least 1% of the dwellings have shown concentrations above 200 Bq m−3 . Those are Arnea Chalkidikis with 19 dwellings RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07050-0

© 2005 Elsevier Ltd. All rights reserved.

432

A. Louizi et al.

Fig. 1. Areas of increased radon concentration in Greece.

(2.3% of the total) and Vrises Apokoronou Chanion with 3 dwellings (1.0% of the total) with concentrations higher than 200 Bq m−3 [7]. The aim of this research was the study of areas with increased radon concentrations that can be characterized as “high radon areas” [7]. The study includes indoor radon and daughter nuclei concentration measurements as well as soil gas, soil radon and potable water concentration measurements. 2. Materials and methods The indoor air concentration measurements were carried out by Alpha Guard monitor of Genitron Ltd and EQF 3023 monitor of SARAD INSTRUMENTS. The Alpha Guard monitor with special ﬁttings has been used for soil gas measurements (soil gas, unit, Genitron Ltd). This unit consists of a tubeless rod, which opens holes in the ground. Inside the rod a small-diameter probe is placed. Special ﬁlters obstruct any moistness and daughter products from entering. An air proof air pump “pumps” soil gas and leads it to the interior of the Alpha Guard with the help of plastic tubes, which do not allow 222 Rn to escape. The monitor has a side outlet, which either remains open or is linked to a bicycle inner tube of 2 L or to the input through a tube. The Alpha Guard monitor was also used for radon potable water measurements. The monitor was ﬁtted with a degassing vessel, a security vessel, an air proof pump connected each other and to the Alpha Guard in closed circle with the help of plastic tubes which do not allow the radon to escape. By switching the pump on, the air within the unit circulates, forcing the diluted radon in the water to degas.

Preliminary study of two high radon areas in Greece

433

In dwellings with the highest 222 Rn concentration in Arnaia and Vrisses Apokoronou Chanion as determined by past survey [6,8,9], Alpha Guard and EQF 3023 monitors have been installed simultaneously between 1999 and 2002 for the constant surveillance of concentration. Samples of water have been taken from two dwellings of Arnea and three from dwellings of the area around Apokoronas Chanion. One water sample has been taken from the public spring in Arnea. In eight (8) dwellings in Arnaia and ﬁve (5) in the area around Apokoronas Chanion with known 222 Rn concentrations in indoor air, soil gas concentration has been measured.

3. Results – discussion The results of Alpha Guard and EQF measurements for the dwelling in Arnaia Chalkidikis are presented in Figs. 2 and 3. In Figs. 4 and 5 the relevant diagrams for the dwelling of Vrises Apokoronou Chanion are presented. Figure 6 presents a typical diagram of soil gas concentration measurements in Arnaia Chalkidikis in an experiment, in which the outlet of Alpha Guard was linked to the inner tube of a bicycle. From 13:26 to 13:46 the pump was continuously supplying the unit with soil gas from a depth of 90 cm and at a rate of 1 L min−1 . From 13:46 pumping was stopped and the side outlet was connected to the input of Alpha Guard via a radon proof pipe. Soil gas contains both 222 Rn and 220 Rn (thoron) and Alpha Guard is equally sensitive to thoron when this is present in its interior [10]. After turning off the pump, thoron is completely decayed after approximately 5 minutes and the unit practically measures radon. The average concentration of soil gas is (102 ± 18) kBq m−3 whereas the radon concentration is (32 ± 12) kBq m−3 . Thoron concentration is the subtraction of radon from soil gas concentration. Figure 7 presents the concentration at another experiment, in which after a 15-min soil gas pumping from 90 cm depth and at a 1 L min−1 rate with Alpha Guard’s sideoutlet open to the atmosphere, the outlet was subsequently linked to Alpha Guard’s input. The maximum corresponds to a (102 ± 18) kBq m−3 soil gas concentration. The data in Fig. 7

Fig. 2. Alpha Guard indoor radon concentration measurements in Arnaia Chalkidikis.

434

A. Louizi et al.

Fig. 3. EQF3023 indoor radon and daughter concentration measurements in Arnaia Chalkidikis.

Fig. 4. Alpha Guard indoor radon concentration measurements in Vrisses Apokoronou Chanion.

were ﬁtted to the exponential model C(t) = C0 e−λt , 40 min after the beginning and for 24 hours after. The λ value found was (0.00744 ± 0.00011) h−1 while the value C0 was found to be (34.0 ± 0.6) kBq m−3 . The above experimental value of λ is not signiﬁcantly different from the theoretical 222 Rn decay constant, which is 0.0075536 h−1 . Also the C0 value is almost the same as the value of C, which was deﬁned with the previous method. The second method is more accurate as it includes both radon decay and more measurements. However, it lacks in the fact that it is much more time consuming. The results for indoor radon concentrations of the dwellings, in which soil gas has been measured, are presented in Table 1. In the same table potable water concentrations in the areas of interest are presented. In the case of Vrisses Apokoronou, soil 222 Rn was not estimated. In Table 1, 226 Ra and 222 Rn concentration values in soil from other researchers [11,6] are also presented.

Preliminary study of two high radon areas in Greece

Fig. 5. EQF3023 indoor radon and daughter concentration measurements in Vrisses Apokoronou Chanion.

Fig. 6. Typical soil gas concentration measurements in Arnaia Chalkidikis.

Fig. 7. Soil gas concentration measurements in Arnaia Chalkidikis.

435

436

A. Louizi et al.

Table 1 Total measurement results Area

(1)

(2)

1 2 3 4 5 6 6 7 8 9 10 11 12 13

ARNAIA ARNAIA ARNAIA ARNAIA ARNAIA ARNAIA ARNAIA ARNAIA ARNAIA ARNAIA (TAP) VRISSES NIPPOS STILOS STILOS

Dwelling C ± σ (C) (Bq m−3 ) (3) (4)

Radium C (Bq kg−1 ) (5)

Rad + Thor C ± σ (C) (kBq m−3 ) (6) (7)

Thoron C ± σ (C) (kBq m−3 ) (8) (9)

Radon C ± σ (C) (kBq m−3 ) (10) (11)

Water C ± σ (C) (Bq L−1 ) (12) (13)

1300 37 89 1700 178 900 1400 700 1100

180 8 12 220 23 120 180 100 140

31

19 2 7 21 14 15 17 16 18

79 5 18 90 59 64 71 51 70

37 3 8 43 27 31 32 24 32

320

40

31 31 31 31 31 31

116 8 26 133 86 95 103 75 102

122

16

540 95 34 34

80 29 7 8

28 28 14 14

189 88 56 154

31 15 10 26

410 15

45 4

23 2 8 25 18 19 21 19 22

13 1 3 13 11 12 12 11 12

1.3 1.3

0.5 0.5

1st column represents the measurement number, 2nd the measurement area, 3rd the measured indoor 222 Rn concentration, 4th the 226 Ra soil content measured by others [11,6], 5th the measured soil gas concentration, 6th through 11th the estimated thoron and radon soil concentrations and 12th the measured 222 Rn concentration in water samples. C represents the concentration value and σ (C) the uncertainty in the 95% conﬁdence interval.

4. Conclusions Reported soil gas concentrations may be considered to be elevated according to our database for Greece but lie in the intermediate range according to the literature [12–14]. According to our database, commonly measured soil gas concentration values in Greece lie in the range 5–15 kBq m−3 . Radon content in drinking waters in Arnea Chalkidikis may also be considered as elevated. The areas of Arnaia Chalkidikis and Vrisses Apokoronou require further investigation.

References [1] A. Geranios, M. Kakoulidou, P. Mavroidi, M. Moschou, S. Fischer, I. Burian, J. Holecek, Radon survey in Kalamata (Greece), Radiat. Prot. Dosim. 93 (1) (2001) 75–79. [2] K.G. Ioannides, K.C. Stamoulis, C.A. Papachristodoulou, A survey of 222 Rn concentrations in dwellings of the town of Metsovo in North-Western Greece, Health Phys. 79 (6) (2000) 697–702. [3] C. Papastefanou, S. Stoulos, M. Manolopoulou, A. Ionnidou, S. Charalambous, Indoor radon concentrations in Greek apartment dwellings, Health Phys. 66 (3) (1994) 270–273. [4] E. Georgiou, K. Ntalles, A. Molfetas, A. Athanassiadis, C. Proukakis, Radon measurements in Greece, in: Radiation Protection Practice, IRPA 7, vol. I, Pergamon Press, New York, 1988, pp. 125–126. [5] C. Proukakis, M. Molfetas, K. Ntalles, E. Georgiou, A. Serefoglou, Indoor radon measurements in Athens, Greece, in: British Nuclear Energy Society Conference on Health Effects of Low Dose Ionizing Radiation – Recent Advances and Their Implications, BNSE, London, 1988, pp. 177–178.

Preliminary study of two high radon areas in Greece

437

[6] D. Nikolopoulos, PhD thesis, Medical Physics Department, Athens, Greece, 2001 (in Greek). [7] M. Kendall, H. Miles, D. Cliff, R. Green, R. Muirhead, W. Dixon, R. Lomas, M. Goodridge, Exposure to Radon in UK Dwellings, NRPB-R272, HMSO, London, UK, 1994. [8] A. Louizi, D. Nikolopoulos, E. Lobotesi, A. Serefoglou, K. Gogos, C. Proukakis, Factors affecting indoor radon concentrations in Greece, Proceedings of a Workshop on Radon in the Living Environment, Athens, Greece, April 19–23, 1999, Sci. Total Environ. 272 (1–3) (2001). [9] D. Nikolopoulos, A. Louizi, V. Koukouliou, E. Lobotesi, K. Gogos, C. Proukakis, Exposure of the Greek population to indoor radon-risk assessment, Proceedings of a Workshop on Radon in the Living Environment, Athens, Greece, April 19–23, 1999, Sci. Total Environ. 272 (1–3) (2001). [10] Alpha Guard Manual, Genitron Ltd, 1993. [11] P. Probonas, PhD thesis, Medical Physics Department, University of Athens, 1991 (in Greek). [12] W.W. Nazaroff, A.V. Nero, Radon and Its Decay Products in Indoor Air, Willey, New York, 1988. [13] J. Kemski, A. Siehl, R. Stegemann, M. Valdivia-Manchego, Mapping the geogenic radon potential in Germany, Sci. Total Environ. 272 (2001) 217–230. [14] I. Barnet, Reliability of radon risk maps – Czech Republic, in: Book of Abstracts of the 7th International Symposium on Natural Radiation Environment (NRE-VII), Rhodes, Greece, May 2002, p. 396.

438

Year-to-year variations in radon levels in a sample of UK houses with the same occupants N. Hunter, C.B. Howarth, J.C.H. Miles, C.R. Muirhead National Radiological Protection Board, Chilton, Didcot, Oxon, OX11 ORQ, UK

In 1993 and 1994 the householders of 96 UK houses were recruited for a long-term study of the year-to-year variation in radon levels. The houses were chosen because they had been found in earlier surveys (in 1991 and 1992) to have radon levels around 100 Bq m−3 . This level was chosen because it was high enough to give reasonable statistical accuracy in estimating radon concentrations, but well below the UK radon Action Level of 200 Bq m−3 above which householders are advised to implement remedial measures to reduce the concentration. The initial measurements were carried out using passive radon detectors exposed for three months in both the main living area and an occupied bedroom in the house. Repeat measurements were carried out using the same type of detector in the same locations for the same three months in successive years until 1996. After that, the measurements for all houses were carried out simultaneously each year, to reduce the administrative burden of the survey. Each year a few houses dropped out of the study, reducing the number of participants to 70 by 2001. The annual average radon exposure was estimated for each house using two methods: correcting for the typical seasonal variation in radon levels found in earlier studies, or correcting for the change in levels caused by the average outdoor temperature during the measurement period [a]. The variations in corrected and uncorrected results from year to year are being analysed using the BUGS and S-Plus statistical packages, and the results will be presented at the Symposium. Initial results using uncorrected log transformed data give an estimated mean within dwellings year-to-year standard deviation of 0.40, with a 95% conﬁdence interval of 0.37 to 0.43. 1. Introduction Passive radon detectors exposed over limited periods are used to estimate long-term average radon concentrations in houses. It is important to know how accurate these estimates are. The sources of uncertainty in these estimates include inaccuracies in the results of the detectors themselves, inaccuracies in extrapolating from a limited period to a full year, and inaccuracies in extrapolating from a single year to the long-term average. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07051-2

© 2005 Elsevier Ltd. All rights reserved.

Year-to-year variations in radon levels in a sample of UK houses with the same occupants

439

The aim of this study is to investigate the variability of radon measurements in houses from year to year. This study is complicated by the fact that measurements took place over a period of only 3 months, rather than 12 months. However, there are ways of normalising these 3-month measurements to obtain an annual average, by using either seasonal or temperature [4] correction factors. Some housing characteristics also affect the variability in measurements [1,3] such as solid ﬂoors, double-glazing, draught prooﬁng and structural alterations such as sealing of cracks, reduced ventilation and blocking of a chimney. Multiple regression has been used to investigate these factors. The analysis of variance (ANOVA) examines yearly radon concentrations and allows for variation between individuals. This model is ﬁtted using the “Bayesian inference Using Gibbs Sampling” (BUGS) [5] program and the S-Plus statistical package. The model is used to study repeated measurements of radon concentrations in houses and to discuss the comparison between concentrations measured over 3 months (uncorrected data) and concentrations corrected to annual averages using seasonal and temperature factors (corrected data).

2. Materials and methods The data were collected by the National Radiological Protection Board throughout the United Kingdom; information on house characteristics were also recorded. The homes selected to be included in the study all had an initial seasonally corrected radon level of around 100 Bq m−3 . This radon level was chosen because it was high enough to ensure that the response from the detectors would be less susceptible to statistical error whilst being well below the UK Action Level of 200 Bq m−3 . None of the homes selected had undertaken any kind of remedial measure to reduce the radon level. Ninety-six householders agreed to take part. The houses were initially measured in 1991 and 1992 and recruited to the study in 1993 and 1994. For the ﬁrst four years of the study the repeat measurements were made in the same three months as the initial measurement had been made in each house. In 1996 a decision was made to carry out all measurements simultaneously each year in order to reduce the administrative burden. One difﬁculty with carrying out this type of long-term survey is maintaining the level of participation. Since 1993 the number of householders participating declined from 96 to 70 for a number of reasons including moving away, death of the householder or a loss of interest in participating in the study. All measurements were carried out using alpha etched track detectors exposed for three months in both the main living area and a sleeping area in the house. The initial measurements were followed by seven repeat measurements in the same locations in the study houses during 1991–1999. In the year 1997, administrative problems led to measurements being completed in only a few houses, so the results collected in that year were not included in this study. For each house, the average indoor radon concentration was estimated using a weight of 0.55 for the bedroom value and 0.45 for the living area [6] and then combined. Table 1 shows summary statistics for the radon concentrations over six years from 96 houses for these uncorrected and corrected data. The arithmetic and geometric means, and geometric standard deviation of the measured concentrations using uncorrected data are 105, 90 and 1.77 Bq m−3 , respectively. For seasonally and temperature corrected data, the mean values are increased to 111, 96 and 109 and 94 Bq m−3 , respectively. The geometric standard deviation

440

N. Hunter et al.

Table 1 Summary statistics for the radon concentration data (Bq m−3 ) over six years from 96 houses in the study

Uncorrected data Seasonally corrected Temperature corrected

Number of dwellings

Min.

Max.

Mean

Std. deviation

Geometric mean

Geometric std. deviation

96 96 96

12.00 14.00 13.00

757.79 727.00 690.00

105.04 111.45 109.39

66.77 75.37 73.24

90.02 95.58 93.69

1.77 1.72 1.73

for both the seasonal and temperature corrected data remain almost the same, namely 1.72 and 1.73, respectively. Distributions of radon concentrations in houses are frequently found to be lognormal. However, in this case the study houses were chosen to have a small range of initial radon concentrations, so there is no expectation that results of measurements in these houses should conform

(a)

(b) Fig. 1. Normal probability plot on radon data over six years from 96 houses for uncorrected data (a) raw data; (b) after log transformation.

Year-to-year variations in radon levels in a sample of UK houses with the same occupants

441

to the lognormal distribution. Figure 1 shows a normal probability plot of the uncorrected results of all measurements in all houses before (Fig. 1a) and after (Fig. 1b) log transformations. In Fig. 1a, the data are reasonably near to a normal distribution as most of the plotted points fall in a straight line. However, there are a few at the lower end and too many values at the upper end, i.e., it is a positively skewed distribution. The normal probability plot of the log radon concentrations (Fig. 1b) is reduced by the inﬂuence of outlying values. The normal probability plots of the temperature-corrected data and the seasonally corrected data over six years from 96 houses have also been studied – not shown here – they have also indicated too many values in both tails of the distribution. A possible explanation for those upper tail values that these ﬁgures show a combination of two distributions: variation between dwellings and variation from year-to-year within dwellings.

Fig. 2. Histogram of variances of log radon concentration over six years using unconnected data from 96 houses.

Fig. 3. Geometric mean radon concentrations over 6 years for 96 houses.

442

N. Hunter et al.

Figure 2 shows the histogram of variances of log radon concentrations over six years for 96 houses using uncorrected data. Clearly the individual variances vary greatly from house to house. Figure 3 shows the estimates of the geometrically averaged indoor radon concentrations based on six repeated measurements. Separate estimates are given for the uncorrected data, the temperature corrected data and the seasonally adjusted data. Clearly in Fig. 3, the ﬁrst four measurements demonstrate close agreement between the corrected and uncorrected measurements. However, the results for the ﬁfth and sixth measurement periods show that corrected measurements are signiﬁcantly higher than uncorrected measurements. The explanation of this is that the ﬁfth and sixth sets of measurements were mainly made only during April–July 1998 and August–November 1999, respectively, whereas the rest of the earlier sets of measurements were taken in various seasons throughout the year. Overall, there are no notable differences between the seasonal and temperature corrected measurements. 2.1. Selection of covariates Various covariates were recorded, of which a subset most likely to affect radon concentrations was selected for this study. The variables chosen were: building year, house type, living room and bedroom ﬂoor type, storey level for living room and bedroom, double-glazing, draught prooﬁng and building changes. All covariates were categorised into different groups: – – – – –

House type: detached, semi-detached, terraced, ﬂat and other; Floor type: suspended and solid; Double-glazing: no and yes; Draught prooﬁng: no and yes; Building year: before 1900, 1900–1919, 1920–1944, 1945–1964, 1965–1976 and after 1976; – Bedroom storey level: basement, ground ﬂoor, 1st ﬂoor and 2nd ﬂoor; – Living room storey level: basement, ground ﬂoor and ﬁrst ﬂoor; – Changes in buildings: no and yes. 2.2. Model The statistical model uses logarithmically transformed data for the radon concentrations and was obtained using backward stepwise regression. The house type, double-glazing, ﬂoor type and draught prooﬁng factors were deﬁned in the model as ﬁxed effects. The results of the regression implied that the rest of the covariates, such as building year, bedroom and living room storey could be dropped. None of the interaction effects was statistically signiﬁcant. About 31% of the houses recorded changes in the building condition over the period for which measurements were available. Major changes involved the installation of doubleglazing, insulation (e.g., cavity wall insulation) and sealed ﬂoors. Unexpectedly, there was no signiﬁcant difference found in radon concentrations between homes with and without changes in building structure. Figure 4 shows geometric mean of radon concentrations for the six sets of repeated measurements, both for buildings with and without changes. Clearly there is an increasing trend with year in those dwellings with changes compared to those without.

Year-to-year variations in radon levels in a sample of UK houses with the same occupants

443

Fig. 4. Geometric mean radon concentrations with and without changes in houses.

However, the average measurement for the fourth period appears to be the same for both dwellings with and without changes. This possibly explains the insigniﬁcant effect above. Included in the model equation (1) is a parameter which represents the effect of each individual house. We also allow for differences between individuals in their year-to-year variability Mij = μ + αk + βm + δn + γt + θi + Eij .

(1)

Here Mij represents log-radon concentration of house i (i = 1, 2, . . . , 96) in year j (j = 1, 2, . . . , 6), μ represents the true average radon level for all houses in the study, k represents house type category (k = 1, . . . , 5), m represents the double glazing category (m = 1, 2), n represents the ﬂoor type category (n = 1, 2) and t represents the draught-prooﬁng category (t = 1, 2). The parameters μ, α, β, δ and γ are estimated by regression using the BUGS statistical package, which estimates the parameters using the Bayesian Markov Chain and Monte Carlo technique [2], and using the S-Plus statistical package. θi represents the difference be2 ). tween log radon levels of the ith house (assumed normal with mean 0 and variance σbetween Eij represents random variations within house i between years (assumed normal with mean 0 2 2 2 and variance σwithin ). The parameters σbetween and σwithin are also estimated by regression using the BUGS and S-plus statistical packages. The parameters σbetween and σwithin represents between and within house standard deviations, respectively. Fitting such a model allows us to estimate two variances: 2 ) which measures the extent to which mean radon con(1) between-house variance (σbetween centrations of individual houses vary about the mean radon estimated from the entire study population; 2 ) which measures the amount by which houses’ yearly (2) within-house variance (σwithin radon measurements vary about their individual means.

444

N. Hunter et al.

3. Results The results of the analysis of variance are shown in Table 2 which shows the percentage of the total variation explained by each of the factors from the ﬁnal model (equation (1)). The factors selected for the ﬁnal model for uncorrected radon measurements are: house type, double-glazing, ﬂoor type for living room and draught prooﬁng. Floor type for living room and draught prooﬁng were the only factors that were signiﬁcant at the P = 0.05 level or less when house type and double-glazing were included. The percentage of total variation explained by the ﬁnal model for uncorrected data is 15.1%. The remaining four factors (building year, storey for bedroom and living room and building changes) only accounted for an extra 4.0% of the total variation in the data. The factors selected for the ﬁnal model (equation (2)) for temperature corrected radon concentrations are house type and draught prooﬁng. The percentage of total variation explained by the ﬁnal model for the temperature corrected data is 7.0% (Table 2). The remaining six factors (building year level, storey for bedroom and living room, changes in mitigation, ﬂoor type for living room, and double-glazing) only accounted for an extra 4.5% of the total variation Mij = μ + αk + γt + θi + Eij .

(2)

The same factors were again selected for the ﬁnal model for seasonally corrected radon concentrations, i.e., house type and draught prooﬁng (equation (2)). The percentage of total variation explained by the ﬁnal model for the seasonal corrected data is 8.3% (Table 2). The remaining six factors (building year, storey for bedroom and living room ﬂoor, changes in mitigation, ﬂoor type for living room, and double-glazing) only accounted for an extra 6.0% of the total variation. The estimates of the effects of these factors, together with 95% conﬁdence intervals (CI) based on the model derived from the uncorrected data, are given in Table 3. This also shows the number of houses falling in each category for house type, ﬂoor type, double-glazing, and draught prooﬁng. It should be noted that the numbers of houses with missing data for house type and ﬂoor type are 4 and 9, respectively. Table 2 Percentage of total variation explained by each of the four factors selected for the ﬁnal model using uncorrected and corrected (seasonal and temperature) data Source of variation

# of categories

House type

5

Double-glazing

2

Floor type living room Draught prooﬁng

2

∗ Signiﬁcance level.

2

% variation explained in uncorrected data (p-value)∗ 11.0 (< 0.0001) 2.1 (0.002) 1.2 (0.03) 0.8 (0.04)

% variation explained in temperature corrected data (p-value)∗ 5.9 (< 0.0001) – –

% variation explained in seasonal corrected data (p-value)∗ 6.7 (< 0.0001) – –

1.1 (0.02)

1.6 (0.005)

Year-to-year variations in radon levels in a sample of UK houses with the same occupants

445

Table 3 Estimated effects for the categories of each factor under the ﬁnal model with 95% conﬁdence interval (CI) in parenthesis, based on the whole data set Parameter

Number of houses

Estimated value for uncorrected data (95% CI)

Estimated value for temperature corrected data (95% CI)

Estimated value for seasonal corrected data (95% CI)

μ, intercept (Bq m−3 ) House type α1 , detached α2 , semi-detached α3 , terraced α4 , ﬂat α5 , other Double glazing β1 , none β2 , available Floor type living room δ1 , suspended wooden δ2 , solid Draught prooﬁng γ1 , none γ2 , available σbetween σwithin

96

104.48 (84.52, 129.15)

105.74 (93.13, 118.98)

109.72 (96.45, 124.84)

46 30 10 2 4

1.00 0.73 (0.60, 0.88) 0.72 (0.54, 0.96) 1.07 (0.56, 2.02) 0.87 (0.56, 1.36)

1.00 0.81 (0.69, 0.96) 0.79 (0.62, 1.01) 0.84 (0.49, 1.45) 0.77 (0.53, 0.90)

1.00 0.78 (0.66, 0.94) 0.79 (0.61, 1.01) 0.87 (0.49, 1.54) 0.76 (0.52, 1.17)

48 48

1.00 0.89 (0.73, 0.93)

–

–

37 50

1.00 1.12 (0.92, 1.35)

– –

– –

66 30 96 96

1.00 0.89 (0.73, 1.09) 0.37 (0.30, 0.44) 0.40 (0.37, 0.43)

1.00 0.89 (0.76, 1.04) 0.29 (0.22, 0.35) 0.46 (0.43, 0.49)

1.00 0.87 (0.73, 1.02) 0.32 (0.26, 0.38) 0.43 (0.40, 0.46)

In Table 3, the same baseline categories are referred as the ones were used in a study by Gunby et al. [3], for example, no double glazing, no draught prooﬁng and suspended ﬂoor type. All other categories have an associated multiplicative increase or decrease in the baseline indoor increment of the quantity given in the table. The results for uncorrected data in Table 3 show that the most signiﬁcant correlation with radon concentrations was for the house type. Relative to detached houses, radon concentrations in semi-detached houses were a factor of 0.73 lower with a 95% CI of (0.60, 0.88); the corresponding factor for terraced houses was 0.72 with a 95% CI of (0.54, 0.96). Radon concentration in ﬂats were a factor of about 1.07 above those in detached houses with a wide 95% conﬁdence interval of (0.56, 2.02); however, the estimated effect is not statistically signiﬁcant (P = 0.36) because the study included only 2 ﬂats. For other type of houses, radon concentrations were found to be about 19% higher than semidetached houses, but again this difference is not statistically signiﬁcant and there were only 4 dwellings in the other type category in the study. After house type, double-glazing was found to be the second most signiﬁcant factor. Houses with double-glazing were a factor of 0.89 lower relative to these without double glazing with a 95% CI of (0.73, 0.93). As was expected, the radon concentration in solid type ﬂooring houses was higher than suspended wooden ﬂoor houses. The concentration was increased in houses with solid type of ﬂooring by a factor of about 1.12 with a 95% CI of (0.92, 1.34). Houses with draught prooﬁng had concentrations lower by a factor of about 0.89 relative to these without draught prooﬁng with a 95% CI of (0.73, 1.09).

446

N. Hunter et al.

Table 4 Estimated between and within house standard deviation values in six-year and four-year measurements for uncorrected, seasonal and temperature corrected data Six-year

Uncorrected data Seasonally corrected Temperature corrected

Four-year

σbetween (95% CI)

σwithin (95% CI)

σbetween (95% CI)

σwithin (95% CI)

0.37 (0.30, 0.44) 0.32 (0.26, 0.38) 0.29 (0.22, 0.35)

0.40 (0.37, 0.43) 0.43 (0.40, 0.46) 0.46 (0.43, 0.49)

0.34 (0.27, 0.42) 0.26 (0.20, 0.33) 0.22 (0.15, 0.29)

0.38 (0.34, 0.41) 0.38 (0.34, 0.41) 0.41 (0.37, 0.45)

The results for data with seasonal and temperature corrections, based on the model given in equation (2), are also given in Table 3. The most signiﬁcant correlation with radon concentrations was for house type; draught prooﬁng was the second most signiﬁcant factor for both seasonal and temperature corrected data. Overall there was no indication of differences between the ﬁndings using seasonally and temperature corrected data. Using the uncorrected data, the estimated standard deviation between houses in log radon concentrations is 0.37 with a 95% CI of (0.30, 0.44) and the within house variation from year to year is 0.40 with a 95% CI of (0.37, 0.43). The corresponding values for the corrected temperature and seasonal data are 0.29 (95% CI: 0.22, 0.35) and 0.32 (0.26, 0.38), respectively for between house variation, and 0.46 (95% CI: 0.43, 0.49) and 0.43 (95% CI: 0.40, 0.46), respectively for variation within houses between years. 3.1. Sensitivity analysis As indicated in Fig. 2, the repeated measurements were carried out in the same locations for three months in varying parts of the year, during 1991–1996. After that, the measurements in 1998 and in 1999 were carried out for all houses simultaneously. It may be advisable to treat the ﬁrst four sets of measurement data separately from the whole data set to see if there are any signiﬁcant differences. Table 4 shows the estimated values for the between and within house standard deviations using all of the data and for the ﬁrst four sets of measurements, both for uncorrected and corrected data. Compared to the whole data set, there is no marked difference in the between and within house variability using ﬁrst four sets of measurements which, in fact, contain much of the total data. For both the seasonally and temperature corrected data, the variability between and within houses decreases by about 20% and 10%, respectively, when using the ﬁrst four sets of measurements, relative to the total data set. For both the uncorrected and seasonally corrected data, the within house variability is the same based on the ﬁrst four sets of measurements, 0.38 (95% CI: 0.34, 0.41).

4. Conclusion It is generally accepted that uncertainties in estimates of long-term mean radon concentrations are affected by a number of factors. One of these factors is the year-to-year variability, thought to be mainly due to variations in the weather. This study investigated variations in radon concentrations within and between dwellings using the results of measurements over periods of

Year-to-year variations in radon levels in a sample of UK houses with the same occupants

447

only 3 months, corrected to annual averages using seasonal and temperature factors. Some other factors that affect the variability in measurements such as house type, double-glazing and draught prooﬁng were also investigated. For the uncorrected data, the variation explained by the factors included in the model is 15%. These factors were house type, double-glazing, living room ﬂoor type and draught prooﬁng. The variation due to random effects is 77% and is the sum of the between and within house variations. For both the seasonally and the temperature corrected radon concentrations, variations explained by the model are 8% and 7%, respectively, due to ﬁxed effects and 75% for both due to random effects. Overall there is no indication of differences in the ﬁndings using seasonal and temperature corrected data. This study shows that in general a three month measurement of radon in a home using etched track detectors combined with the use of appropriate correction factors for seasonal effects will give a good estimate of the long term average radon level in a home. This being said, it must be acknowledged that several homes in this study have shown radon levels that deviate markedly from the initial measurement. Analysis of the data provided on house characteristics does not show any obvious reasons why some houses are more variable than others. In a postal survey such as this the possibility that data may be incomplete or inaccurate cannot be ruled out.

References [1] M. Gerken, L. Kreienbrock, J. Wellmann, M. Kreuzer, H.E. Wichmann, Models for retrospective quantiﬁcation of indoor radon exposure in case–control studies, Health Phys. 78 (3) (2000) 268–278. [2] W.R. Gilks, S. Richardson, D.J. Spiegelhalter, Markov Chain Monte Carlo in Practice, Chapmann & Hall, London, 1996. [3] J.A. Gunby, S.C. Darby, J.C.H. Miles, B.M.R. Green, D.R. Cox, Factors affecting indoor radon concentrations in the United Kingdom, Health Phys. 64 (1) (1993) 2–12. [4] J.C.H. Miles, Mapping radon-prone areas by lognormal modelling of house radon data, Health Phys. 74 (1998) 370–378. [5] D.J. Spiegelhalter, A. Thomas, N.G. Best, W.R. Gilks, BUGS: Bayesian Inference Using Gibbs Sampling, Version 0.50, Technical report, Medical Research Council Biostatistics Unit, Cambridge, 1996. [6] A.D. Wrixon, B.M.R. Green, P.R. Lomas, J.C.H. Miles, K.D. Cliff, E. Francis, C.M.H. Driscoll, A.C. James, M.C. O’Riordan, Natural Radiation Exposure in UK Dwellings, NRPB R190, HMSO, London, 1988.

448

The charged fraction of the 218Po ions in air under environmental conditions P. Pagelkopf, M. Gründel, J. Porstendörfer Institute for Physical Chemistry, Georg-August-University Göttingen, Tammannstr. 6, 37077 Göttingen, Germany

The charged fraction of unattached 218 Po ions was measured quantitatively under normal environmental conditions. The experiments can be divided into two stages. First, the determination of the charged fraction directly after decay. The experimental results agree with the value given in the literature and show that 87% of the generated 218 Po ions are positively charged. Second, the determination of the charged fraction after neutralisation. Under indoor conditions with relative humidity between 30 and 96% and ionisation rates between 60 and 2500 μR h−1 (this is equivalent to radon gas concentrations between 80 and 5400 Bq m−3 ) the charged fraction is mainly inﬂuenced by the variation of the ionisation rates by radon and its decay products.

1. Introduction The greatest part of the natural exposure of the population is a result of inhaling the short-lived decay products of 222 Rn. The physical processes (cluster formation, neutralisation, attachment to aerosols and plate-out on surfaces) of the ﬁrst product in this decay chain, the 218 Po, mainly determine the characteristics of the other decay products. The electrical charge of the 218 Po ions has a strong inﬂuence on these physical processes and so on the dose relevant parameters [1]. After the decay of radon most of the formed 218 Po ions are positively charged while the rest is neutral. The ions react rapidly with trace gases and vapours in the air and become small particles called “unattached” decay products or clusters (size spectrum, diameters between 0.5 and 3 nm). At the same time the charged 218 Po ions get neutralised by the following processes [2,3]: (1) recombination with small air ions; (2) electron scavenging by OH-radicals formed by radiolysis of water vapour; (3) charge transfer by molecules of lower ionisation potential. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07052-4

© 2005 Elsevier Ltd. All rights reserved.

The charged fraction of the 218 Po ions in air under environmental conditions

449

Simultaneously, the 218 Po attaches to aerosol particles and is plated out on surfaces. All these processes are inﬂuenced by the electrical charge of the 218 Po, which also affects the mobility of the 218 Po. This mobility is described by the diffusion coefﬁcient and determines the attachment on aerosols and surfaces. The aim of this study was to measure the charged fraction of the 218 Po ions in air under environmental conditions by changing the relative humidity and the ionisation rate. In addition the fraction of freshly formed 218 Po ions, directly after the decay of the radon, was determined.

2. Measurement technique To measure the unattached charged fraction two conditions had to be met: (1) separation of the unattached and the on-aerosols-attached 218 Po ions; (2) separation of the charged unattached from the neutral unattached 218 Po ions. To fulﬁl the ﬁrst condition we separate the attached from the unattached fraction by using the screen diffusion method. To record the decays on this screen the online-back-screentechnique (OBST) was used (Fig. 1). All the unattached decay products in the laminar airﬂow were collected on the screen by diffusive deposition, without entry losses. The decay of the collected material is detected simultaneously through the screen with an alpha detector and analysed by alpha spectroscopy. The energy resolution of the system allows a separation between the 218 Po and the 214 Po peaks. The error from the amount of decay product aerosol, which was also collected on the screen, was adjusted by a method suggested by Porstendörfer and Reineking [4]. With this technique it is possible to measure simultaneously during sampling and continuously, without entrance losses of the unattached decay products in air. To meet the second condition of separating the charged from the neutral unattached fraction, the electrostatic ﬁeld in a cylindrical condenser was used (Fig. 2). In the electrical ﬁeld of the condenser all the charged unattached 218 Po ions were drawn to the electrode wire in the centre of the cylinder. At the end of the condenser the 218 Po clusters were collected on a screen alternatively with applied voltage (which means that only neutral 218 Po clusters were collected

Fig. 1. The online-back-screen-technique (OBST) for the measurement of the unattached radon decay products.

450

P. Pagelkopf et al.

Fig. 2. Cylindrical condensers with the online-back-screen-technique for the separation of the charged from the neutral 218 Po.

on the screen) and without applied voltage (collecting all of the unattached 218 Po clusters on the screen). For measuring the charged fraction of the freshly formed 218 Po ions, there was a cellulose ﬁlter mounted at the entrance of the condenser to prevent pre-formed decay products from entering the condensers. The geometrical dimensions of the condenser and the ﬂow rate were optimised to minimise diffusional wall losses and to minimise a signiﬁcant generation of 218 Po ions by decay in the cylinder volume. For measuring the charged fraction of the Po ions direct after the decay of the radon the ﬂow rate was reduced to obtain a higher generation of decay products in the cylinder volume. In addition the following parameters were measured to calculate theoretically the “neutralised” charged fraction with a room model [5] together with neutralisation rates measured by Dankelmann et al. [6]: – the radon gas concentration using a monitor based on electrostatic precipitation on the alpha detector with alpha spectroscopy; – the radon decay products were measured using the ﬁlter method (membrane ﬁlter) and alpha spectroscopy; – the dose rate of gamma radiation and cosmic rays were measured with a dose meter; – the aerosol concentration was measured with a condensation nuclei counter; – the temperature and the relative humidity were monitored by the appropriate monitor. The production rate of the air ions, which showed a strong inﬂuence on the neutralisation rate in air, was expressed in the units 1 R h−1 = 2.54 × 10−4 C kg−1 h−1 ).

The charged fraction of the 218 Po ions in air under environmental conditions

451

The ionisation rate was estimated taking into account the alpha radiation of radon and its short-lived decay products. For the generation of an air ion pair an average energy of 33 eV was assumed.

3. Results 3.1. Measurement of the charged fraction of freshly formed 218 Po ions The two condensers (one for measuring the total fraction, the other for measuring the neutral fraction of the unattached clusters) with the OBST-devices were mounted on a small container (about 1 m3 ) in which a high radon gas concentration was established. At the entrance of each of the condensers a cellulose ﬁlter was ﬁtted so that all previously formed 218 Po ions were unable to enter the condenser volume. Only the freshly formed Po ions in the condensers were separated with the electrical ﬁeld and collected on the screen. At least 15 runs of about 4 h each were made. The results of these measurements are shown in Fig. 3. The mean value of all the 15 measurement is 87.3% with a standard deviation of 1.6%. This is in agreement with the value of 88% obtained by Wellisch in 1913 [7]. 3.2. Measurement of the charged fraction of neutralised 218 Po clusters under environmental conditions These measurements were carried out in a Radon-Test-Chamber (volume ∼ 8 m3 ) with natural indoor aerosol. Varying the radon gas concentration changed the ionisation rate. The variation of the relative humidity was obtained with a humidiﬁer. The results of these measurements are illustrated in Figs. 4–6. In Figs. 4 and 5 a comparison is shown between the measured charged unattached fraction (CFmeas ) and the calculations by model of the estimated charged

Fig. 3. Measured fraction of the positive 218 Po clusters directly after the decay of the radon in relation to the literature value of 88%.

452

P. Pagelkopf et al.

Fig. 4. Fraction of the positive 218 Po clusters in air as function of the ionisation rate at an r.h. of 30–50%. Measured (CFmeas ) and calculated (CFcal ) values.

Fig. 5. Fraction of the positive 218 Po clusters in air as function of the ionisation rate at an r.h. of 85–96%. Measured (CFmeas ) and calculated (CFcal ) values.

unattached fraction (CFcal ) under variation of the ionisation rate (60–2500 μR h−1 ) by changing the radon concentration (80–5400 Bq m−3 ) and r.h. of 30–40% and 85–96%, respectively. The variation of the radon concentration caused the CFmeas to vary between 9 and 33%. All measured fractions of the positive 218 Po clusters are on average about 20% higher than the estimated values. The prediction by the model calculations is thus satisfactory. Comparing the measurements at the different relative humidities it can be deduced that the variation in the charged fraction of the unattached 218 Po clusters is mainly caused by the variation of the ionisation rate and not by the changing of the relative humidity.

The charged fraction of the 218 Po ions in air under environmental conditions

453

Fig. 6. Fraction of the positive 218 Po clusters in air as function of the ionisation rate. Comparison between the values measured at an r.h. of 30–40% and at an r.h. of 85–96%.

References [1] J. Porstendörfer, Physical parameters and dose factors of the radon and thoron products, Radiat. Prot. Dosim. 94 (2001) 365–373. [2] S.D. Goldstein, P.K. Hopke, Environmental neutralisation of polonium-218, Environ. Sci. Technol. 19 (1985) 146. [3] K.D. Chu, P.K. Hopke, Neutralisation kinetics for polonium-218, Environ. Sci. Technol. 22 (1988) 711–717. [4] A. Reineking, J. Porstendörfer, Unattached fraction of short-lived Rn decay products in indoor and outdoor environments: An improved single-screen method and results, Health Phys. 58 (1990) 715–727. [5] J. Porstendörfer, A. Wicke, A. Schraub, The inﬂuence of exhalation, ventilation and deposition upon the concentration of radon and thoron and their decay products in room air, Health Phys. 34 (1978) 465–473. [6] V. Dankelmann, A. Reineking, J. Porstendörfer, Determination of neutralisation rates of 218 Po ions in air, Radiat. Prot. Dosim. 94 (2001) 353–357. [7] E.M. Wellisch, The distribution of active deposit of radium in an electric ﬁeld, Philos. Mag. 28 (1913) 623–635.

454

Activity size distribution in outdoor air: Short-lived ( 214Po, 214Bi/214Po) and long-lived ( 210Pb, 210Po) radon and thoron ( 212Pb, 212Po) decay products and 7Be M. Gründel a , A. Reineking b , J. Porstendörfer c,* a Isotope Laboratory of the Institute of Physical Chemistry, Georg-August-University, Tammannstr. 6,

37077 Göttingen, Germany c Research Center Forest Ecosystems, Büskenweg 1, 37077 Göttingen, Germany b Am Hirtenberg 8, 37136 Waake, Germany

The activity size distribution of the short-lived (118 Po, 214 Po) and long-lived (210 Pb, 210 Po) radon and the thoron (212 Pb, 212 Po) decay products and 7 Be in indoor air over a longer time period were measured. The radionuclides with longer half-lives show signiﬁcantly greater activity median aerodynamic diameters (AMAD) than the short-lived nuclides. This increase in size of the original generated radioactive aerosol during its residence time in the atmosphere could be explained by coagulation with the non radioactive aerosol particles.

1. Introduction There are a number of radioactive isotopes of different elements in the atmosphere near ground level. Most of them are decay products from the 238 U- and 232 Th-chains with the highest activity concentrations being for the short-lived 222 Rn-(radon) (1–100 Bq m−3 ) and 220 Rn(thoron) (0.01–1 Bq m−3 ) decay products. The long-lived radon decay products (210 Pb/210 Po) have concentrations between 1 and 5 × 10−4 Bq m−3 . In contrast to these nuclides, the activity concentration of 7 Be (1–7 × 10−3 Bq m−3 ) is produced by interaction of cosmic rays with atmospheric gases. Most of these airborne radionuclides are adsorbed on the surface of aerosol particles and form radioactive aerosol. Therefore the behaviour of the airborne radionuclides is determined by the behaviour of the aerosol particles in the atmosphere. * E-mail address: [email protected] (J. Porstendörfer).

RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07053-6

© 2005 Elsevier Ltd. All rights reserved.

Activity size distribution in outdoor air

455

Besides the generation rate the activity concentration of the radionuclides in air is inﬂuenced by their transport in the atmosphere and by radioactive decay. In addition, for the longerlived radionuclides (half-life > 1 d) removal processes (dry deposition, wet deposition) from the atmosphere have an inﬂuence on the activity concentration and the coagulation process will signiﬁcantly change the activity size distribution of the original radioactive aerosol. The activity size distribution of the radioactive aerosol is the dominant parameter for the behaviour of the radionuclides in the environment. The removal processes from the atmosphere, the deposition rate on ground and vegetation and the deposition probability in the lung during inhalation depend on the particle size. In this paper the measurement results of the activity size distribution of the short-lived (218 Po, 214 Po) and the long-lived (210 Pb, 210 Po) radon and thoron (212 Pb, 212 Po) decay products and 7 Be are summarised, obtained from measurements over a longer period in outdoor air during recent years. In particular, the aim was to ﬁnd out the differences between the size distributions of these radionuclides. Therefore it was important to use measurement techniques which made it possible to register simultaneously almost all of these radionuclides during one measurement run over a longer time period.

2. Measurement techniques For the measurement of the activity size distribution two measurement techniques were used: • A low-pressure Cascade Impactor (CI) (type BERNER) with 9 stages and 50% cut-off diameters between 60 and 16 000 nm and a back up-ﬁlter. After a sampling time of about one week the deposited size fractionated activities were measured by gamma-spectroscopy with low-level Ge-detectors. The size distributions of 212 Pb, 210 Pb, and 7 Be were determined by this technique. • A low-pressure Online Alpha Cascade Impactor (OACI) (SARAD, Dresden), which was developed to measure continuously the size distribution of the short-lived and long-lived radon decay products and the thoron decay products over longer time periods (1–2 weeks) at low activity concentrations under environmental conditions [1]. The OACI with nine stages and 50% cut-off diameters between 60 and 15 500 nm makes it possible to measure the alpha activity of 218 Po, 214 Po, 210 Po, and 212 Po during sampling by alpha-spectroscopy. After the on-line alpha measurement the foils of the stages were removed and the activities of 210 Pb and 7 Be were measured by gamma-spectroscopy with a Ge-detector in a low background shielding system. Each measurement-run lasted about 2–3 weeks. The unattached clusters of the radon decay products were removed from the entrance air by a tube diffusion battery mounted in front of each impactor system. The data on the measured size distributions were approximated by the sum of two lognormal distributions characterised by the activity median aerodynamic diameters (AMAD) and geometric standard deviations (σg ) and using the Simplex algorithm taking into account the aerosol deposition probabilities of the impactor stages.

456

M. Gründel et al.

3. Results The average values of the parameters of the activity size distributions obtained from all measurements are summarised in Table 1. The short-lived radon and the thoron decay products show activity size distributions with a fraction of about 10–20% in the nucleation mode and a fraction of 80–90% in the accumulation size range. As, examples for the short-lived decay products the results for 214 Po and 212 Po are illustrated in Fig. 1. The activity median aerodynamic diameters of the accumulation mode (AMADa ) varied between 330 and 350 nm for the short-lived radon and between 380 and 420 nm for the thoron decay products. The longer-lived (> 1 d) radionuclides are almost all (93–97%) adsorbed on aerosol particles in the accumulation size range. Only 3–7% of the activity is attached on nuclei with diameters < 60 nm. For the log-lived radon decay products, 210 Pb and 210 Po, AMADa -values of 558 and 545 nm were measured, signiﬁcantly bigger than those of the short-lived nuclides. Figure 2 presents the activity size distributions of 210 Po and 7 Be.

Fig. 1. Relative activity size distribution of radionuclides with shorter half-lives in outdoor air averaged over a three week measurement campaign.

Activity size distribution in outdoor air

457

Table 1 Values of the activity size distributions of the short-lived and long-lived radon and thoron decay products and 7 Be in outdoor air averaged over the measurement period. The measurements were carried out with the CI and OACI technique Nuclide

218 Po 214 Po 210 Pb 210 Po 212 Pb 212 Po 7 Be

Impactor

OACI OACI CI OACI CI OACI OACI CI

Measurement time (days)

Nuclei mode

56 56 77 56 252 56 56 77

23 20

2.4 2.5

26

2.6

AMADn (nm)

Accumulation mode σgn

fn (%)

AMADn (nm)

σga

fa (%)

16 19 4 7 12 19 3 4

332 347 558 545 421 382 702 767

2.2 2.1 1.8 2.3 2.2 2.1 1.7 2.1

84 81 96 93 88 81 97 96

Fig. 2. Relative activity size distribution of long-lived radionuclides in outdoor air averaged over a three week measurement campaign.

458

M. Gründel et al.

Due to their longer residence time in the atmosphere the size distributions of the long-lived radionuclides change to greater particle sizes. This increase in size of the originally generated radioactive aerosol during their residence time in the atmosphere was explained by coagulation with the nonactive aerosol particles [2]. The coagulation of the radioactive aerosol with the inactive aerosol can be described by a nonlinear integro-differential equation, which was numerically calculated according to the example of 210 Pb by Butterweck [2]. His calculation showed that the increase of the median diameter of the freshly generated 210 Pb aerosol which has the same size distribution as 214 Po, from 350 to 550 nm, of the real 210 Po aerosol size distribution can be obtained after a coagulation time of about 15–20 days. This time is about the average residence time of an aerosol of that size distribution in the atmosphere [3].

References [1] J. Kesten, G. Butterweck, J. Porstendörfer, A. Reineking, H.J. Heymel, Aerosol Sci. Technol. 18 (1993) 156–164. [2] G. Butterweck, Dissertation, Faculty of Physics, Georg-August-University Göttingen, Germany, 1991. [3] S.E. Poet, H.E. Moore, E.A. Martell, J. Geophys. 77 (1972) 6515–6527.

459

A comparison of the precision of (quasi-) continuous methods of measuring radon/thoron and decay product concentrations in air R. Rolle Institute of Physical Chemistry, Georg-August University Goettingen, Tamman Street 6, 37077 Goettingen, Germany

Many types of systems are in use for measuring Rn and decay product concentrations in air. A comparison of measurement methods is attempted which should be helpful in selecting optimum, cost-effective methods for speciﬁc applications. The concept of effective volume (EV) of a measuring system is used to compare the sensitivities of different measurement systems. Minimal and optimal methods for differentiating individual 222 Rn- and 220 Rn-decay product concentrations are discussed. The optimal methods can be the most cost effective for the end result that is sought. With improved calibration signiﬁcant new applications, such as determining the air exchange rate from ratios of decay product concentrations, and the Rn source of a room, may become feasible.

1. New direction in radon and decay product monitoring Exposure to radon decay product concentrations constitutes about half of the radiation burden to man and poses a challenge for cost-effective reduction, which includes efﬁcient metrology for a variety of applications. The tracking down of sporadically occurring high-potential exposure enclosed spaces will remain an ongoing task. For some of these spaces remediation may require critical, cost-effective decisions pending the results of (Rn and) decay product measurements. At an enhanced accuracy improved evaluation of the Rn/air-exchange dynamics of enclosed spaces appears feasible. It remains to be seen whether appropriate Rn decay product measurement systems can be applied to air-exchange determinations in rooms with nonelevated decay product concentrations or, possibly with a convenient portable source of Rn. Over the past decades many millions of α-counting measurements have been made for Rn concentration exposure. Most of these measurements were made by passive Rn gas monitoring, where the detection evaluation occurs after an extended exposure period. A mean equilibRADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07054-8

© 2005 Elsevier Ltd. All rights reserved.

460

R. Rolle

rium factor F is normally assumed for conversion to Rn decay product potential alpha energy exposure, the principal parameter related to radiation dose. Far fewer ﬁltration measurements of the decay product concentrations directly have been made, and mostly using (inefﬁcient) short batch measurement procedures. In all these measurements the accuracy seldom reached as low as 20%, or even near that value. An analysis of more efﬁcient measurement systems, commensurate with current technology – Si detectors, miniaturised electronics and computational evaluation – is attempted.

2. Sensitivity of radon and decay product measuring system Many measurements have shown that in all longer measurements (analogous to breathing) the decay products of 220 Rn (thoron) are also present. When measurement accuracy better than 20% is sought for decay product concentrations, then unattached and aerosol-attached concentrations should be evaluated and all members of the two decay chains, listed below, ought to be considered in the evaluations. For practically all measurements decay chain disequilibrium, i.e., saturation calculations are required, but these will not be detailed here. 222

5.5 MeV α

6 MeV α

βγ

3.62 d

3.1 min

26.8 min

Rn −−−−−−→ 218 Po −−−−−→ 214 Pb −−−−−→ 214 Bi/Po βγ, 7.68 MeV α

βγ, 5.3 MeV α

19.9 min

22.3 a

−−−−−−−−−−→ 210 Pb/Bi/Po −−−−−−−−−→ · · · , 220

6.3 MeV α

6.8 MeV α

βγ

βγ, 6&7.68 MeV α

55.6 s

0.14 s

10.6 h

60.55 min

Rn −−−−−−→ 216 Po −−−−−−→ 212 Pb −−−→ 212 Bi/Po/208 Tl −−−−−−−−−−−−→ · · · .

These nuclides generally occur in air at relatively low concentrations and ﬂuctuate widely in time and space. Enhancement of measurement sensitivity and differentiation of the various concentrations is sought for more demanding applications. Advantage is usually taken of the high energy of α-particles and associated low background count in the high-energy region, due to the high absorption and consequently short range of the α-particles. Since β-radiation can also readily be counted with Si α-detectors, its inclusion with the α-measurement improves the accuracy particularly in continuous measurement. The sensitivity of a system measuring air concentrations can be deﬁned as its effective volume air EVx . It is the system’s equivalent air volume effectively at steady state contributing for each nuclide decay (of the speciﬁc air concentration considered) 1 count, or 1 count of a collected nuclide’s decay chain member x. This count, furthermore, needs to be differentiated from others not originating from the speciﬁc air concentration. For enhancement of the quality of a measurement (reﬂected by the covariance matrix of the evaluated measurement data), partial spectral differentiation and timevariant equilibrium differentiation of the partial decay chains is employed in measurement procedures, the former generally being the more effective agent. Time-variant differentiation uses noncontinuous sampling, manipulating the (dis)equilibrium of the partial decay chains, so that for a particular sampling/measurement procedure the EV needs to be multiplied by sampling/decay saturation coefﬁcients Ss Sm . For a ﬁxed on/off sampling cycle air EVx × Ss Sm

A comparison of the precision of (quasi-) continuous methods

461

can be considered as the EV of a quasi-continuous system, or similarly for the single cycle of a batch measurement procedure. Since air concentrations vary in time and space (nonsteady state) the (dis)equilibrium of the partial decay chains needs to be evaluated in any case over all selected periods. The EV of Rn gas monitoring systems is presented in [1]. When continuously collecting a particular decay chain member x at efﬁciency εc and ﬂow rate f from air, e.g., on a ﬁlter or diffusion screen, and continuously measuring member y of its partial decay chain at a detection efﬁciency y εd , then the effective volume is given by airx EVy

= x εc · f · x λ−1 · y εd .

With an efﬁcient Rn decay product instrument (typically ec = 1, εdα ≈ 0.25, εdβ ≈ 0.15, f 50 cm3 s−1 , quasi-continuous Ss Sm 0.1) the following sensitivities are readily achieved: 3 air EVα ≈ 300Po 218 , 3000Pb 214 , 2000Bi 214 , 0.26Po 216 , 70 000Pb 212 , 6000Bi 212 cm . This shows that 216 Po concentrations are, in relation, negligible while even relatively low 212 Pb concentrations make signiﬁcant contributions with longer measurements; when considering relative EV’s of the lung (f ≈ 300 cm3 s−1 , cancer initiating εd ?) the clearance and transfer mechanisms dominate the biological t1/2 of 212 Pb, somewhat reducing the EV of 212 Pb in the lung. 3. Differentiating 218 Po, 214 Pb, 214 Bi, 212 Pb, 212 Bi concentrations (covariance matrix quality) The simultaneously sampled activity concentrations (and instrument 210 Pb/Bi/Po accumulation) produce α and β signals, partially overlapping in spectral regions of interest (ROIs), that require differentiation. When sampling steady concentrations for less than 1 h the longer-lived 212 Pb and 210 Bi concentrations (and 210 Pb/Bi/Po) can be neglected. Differentiation in (1) α- and (2) β-spectral ROIs sufﬁces and provides good differentiation. This is generally augmented by time-variant equilibrium differentiation (providing additional degrees of freedom for evaluation). With only α measurement, time-variant differentiation of the successive Pb and Bi/Poα is poor due to the shorter half-life of Bi after the longer half-life Pb precursor. When sampling (quasi-) continuously for longer than 1 h then ﬁve independent decay product concentrations need to be evaluated. Here the spectrum essentially offers only 1 β- and 3 α-, i.e., 4 ROI’s, so that time-variant differentiation is required and provides additional degrees of freedom for evaluation of the concentrations. Additionally 210 Pb/Bi/Po ought to be taken care of by instrument + collector background measurement.

4. Calibration Principally a measurement is not better than the calibration of the measurement system. System calibration of α-detection efﬁciencies has conventionally been carried out with (< tertiary) standards of a long-lived α-emitter, such as an 241 Am disc. This is poorly suited for the measurement sample-speciﬁc Rn decay product calibration of different α energy and

462

R. Rolle

spectral region-of-interest (ROI). Usually the α-calibration is referenced to (better) established γ-standards. However, of late absolute α-calibration based on fundamental parameter, absorption calculation can be veriﬁed to 1σ < ±1% by α-spectral peak shape matching; even a varying ﬁlter depth penetration of Rn decay products can be evaluated online. For diffusion-collection wire screens the α-detection efﬁciencies have to date been scaled approximations, to 1σ ∼ ±7%, of the efﬁciency fundamental parameter calculation. In preparation is a screen fundamental parameter calculation for α efﬁciencies with 1σ < ±2%. To date β calibrations of ﬁlters and screens were evaluated via time-variant analysis, of suitable α-quantiﬁed ﬁlter measurements, to 1σ ∼ ±5%. In preparation is an approximated fundamental parameter calculation for β with 1σ < ±5%. Enhanced accuracy of individual decay product concentrations is desired particularly for the evaluation of ratios of the concentrations. For this purpose conventional ﬂow calibration to 1σ < ±5% is deemed adequate, more stringent calibration is, however, available.

5. Simple model of radon source, decay product attachment and air exchange in a room The diagram in Fig. 1 indicates the various rate constants used in simple mathematical modelling of Rn and decay products in a room. The speciﬁc attachment-, recoil- and air exchange rate constants signiﬁcantly inﬂuence the ratios of the various concentrations. A better understanding of some of the attachment rates will still be obtained from small chamber experiments. From these the enhanced accuracy of β, α decay product measurement, outlined in the foregoing, makes the evaluation, inter-alia of air exchange and source term, possible and should lead to better decisions on cost effective remediation.

Fig. 1. Transformation rate constants effective in a room with a Rn source and air exchange (short arrows); λu−a – attachment rate to aerosol, λu−w to wall; ra , rw – recoil fractions.

A comparison of the precision of (quasi-) continuous methods

463

6. Conclusions Important metrology applications in Rn radiation protection are locating ‘buildings’ with (potentially) high indoor (Rn) Rn decay product levels and making proper remediating decisions. The ﬁrst requires cost-effective screening measurements of adequate PAEC precision. The second in critical situations requires cost-effective determination of Rn sources and air exchange rate, evaluated via room modelling with input of decay product measurements of enhanced accuracy. Subsequently remediating options can be modelled more objectively for the most cost-effective reduction of concentrations, and can be reevaluated after (stepwise) remediation. Conventional track-etch, charcoal and electret Rn gas determination, at an assumed gas to potential alpha energy equilibrium factor F , largely ﬁll the screening need. For a remediation contract the conventional Rn decay product determination merely provides fair PAEC information to test before/after situations. There is room for more accurate decay product measurement, as outlined above, for source and air exchange rate modelling in critical remediation situations, and possibly in ‘non-radon-related’ air-exchange determination.

Reference [1] J.P. McLaughlin, S.E. Simopoulos, F. Steinhäusler (Eds.), The Natural Radiation Environment VII, Proc. VIIth International Symposium on the NRE, Rhodes, Greece, 20–24 May 2002, Radioactivity in the Environment, vol. 7, Elsevier, Amsterdam, 2004, this volume.

464

Correlation between radon concentration and geological structure of the Kraków area K. Kozak a,* , J. Swako´n a , M. Paszkowski b , R. Gradzi´nski b , J. Łoskiewicz a , M. Janik a , J. Mazur a , J. Bogacz a , T. Horwacik a , P. Olko a a The Henryk Niewodnicza´nski Institute of Nuclear Physics, Radzikowskiego 152, 31–342 Kraków, Poland b Polish Academy of Sciences, Institute of Geological Sciences, Senacka 1, 31-002 Kraków, Poland

The aim of this paper is to investigate the inﬂuence of the geological structure of bedrock on the concentration of radon in soil gas and in the buildings. The radon (222 Rn and 220 Rn) concentration in soil gas was measured using the AlphaGUARD PQ2000PRO ionization chamber and CR-39 detectors. Also natural radioactive isotopes (radium, thorium and potassium) in soil samples were determined using gamma spectrometry with NaI and HPGe detectors. The selection of measurement areas was based on the study of geological maps of Kraków. Geophysical methods (ground penetrating radar and shallow acoustic seismic) were applied to detail the geological structures.

1. Introduction Radon is the most important radioactive factor which can harmfully inﬂuence the human population. The contribution of radon to the average annual dose rate is as high as 40% for the Polish population [1]. It is, therefore, the important subject of investigations aiming at recognition and diminishing the hazard of lung cancer. For many years, surveys have been conducted in different countries in order to recognize the real health hazards that radon and its daughters may cause. Radon-potential maps have been produced (e.g., in Sweden and the Czech Republic [2,3]) which allow determination of the areas with higher risk. In Poland, the “Radiological Atlas” was prepared [1], which includes data on indoor radon concentrations in different parts of the country and the distribution of radium concentrations in the soil. The Polish Geological Institute also prepared a map of surface concentrations of natural radionuclides [4]. Higher radon concentrations are, in general, connected with areas where uranium- and radium-enriched acidic crystalline rocks and black shales occur. There are, however, other * E-mail address: [email protected] (K. Kozak).

RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07055-X

© 2005 Elsevier Ltd. All rights reserved.

Correlation between radon concentration and geological structure of the Kraków area

465

reasons for the occurrence of elevated radon concentrations in soil gas. These are faults, karstiﬁed carbonate rocks and ﬁssured rocks [5,6]. In the paper some investigations in such zones in the Kraków region are presented. The research areas (chosen on the basis of the attainable geological data) are located in the vicinity of the fault zones. The preliminary measurements of indoor radon and radon concentrations in soil gas were made earlier near the faults at St. Bronislawa Hill [7]. 2. Geological setting and site selection criteria The geological structure of the region is rather complex [8]. Its main characteristic is a thick (150–300 m) cover of relatively permeable carbonate Mesozoic rocks, where Jurassic limestones are predominant, which override the Cadomian crystalline basement of the Upper Silesia Massif, with locally acid, uranium-enriched rock subcrops. This cover is subdivided by the alpine presence of fault systems into several horsts, usually exhumed as limestone hills and grabens, ﬁlled with impermeable marine Miocene clays, acting as a barrier to radon migration [7]. Using the maps and other available geological data [9], a research polygon was chosen which is the quite representative of the geological structure of the Kraków region. On that area 8 measurement proﬁles were marked which are perpendicular to the probable courses of the faults (see Fig. 1). On each proﬁle radon concentration in soil gas was measured at several points. Soil samples were also collected from each point for the determination of concentrations of natural radioactive elements – 226 Ra, 232 Th and for completeness also 40 K. 3. Methods 3.1. Geophysical investigation of the geological structure of the research polygon In order to recognize reliably the bedrock structure in the research area, some seismic and radar investigations were performed. For recognition of shallow subsurface structures (to a

Fig. 1. Sketch map of the research area.

466

K. Kozak et al.

depth of 20 m) the RAMAC/GPR ground penetrating radar was used. The total investigated distance on the surface was 2140 m. The method of shallow acoustic seismic logging was used for the geological identiﬁcation of subsurface bedrock structure. The special seismic equipment TERRALOC MK6 (made by ABEM – Sweden) was applied and the investigation covered a total distance of 1900 m. This study was carried out by experts from the Department of Geology, Geophysics and Environmental Protection of the University of Mining and Metallurgy, Kraków. 3.2. Methods of radon measurement The concentrations of radon in soil gas were determined using two complementary methods: – the active method uses the AlphaGUARD gauges [10]. Both isotopes (220 Rn and 222 Rn) were measured together. Here we obtain the average value of about thirty single, 1-min measurements recorded by AlphaGUARD PQ2000 PRO, or of about ﬁve single 10-min measurements recorded by the PQ2000 model; – the passive method uses diffusion chambers with CR-39 detectors [11]. The method was developed at the Institute of Nuclear Physics, Kraków. The diffusion chamber with a CR-39 detector was mounted at the end of an aluminum pole, which was then driven into the soil to the depth of about 1 m and remained here for 15–21 days. As a result the average radon concentration in soil gas over the exposure time was obtained. With this method mainly 222 Rn is measured. The access of 220 Rn (thoron) was signiﬁcantly limited. 3.3. Determination of 226 Ra, 232 Th, 40 K concentrations in soil samples The soil samples were collected from the depth of 0.15–1 m while drilling holes for radon measurements. For all those samples the lithology and humidity were determined. The latter corresponded with soil humidity during the radon measurement. By means of scintillation spectrometry with a NaI(Tl) detector and computer code based on the “3-windows” method, the concentrations of the natural radionuclides – 40 K, 226 Ra, 232 Th – were determined in all soil samples.

4. Results In the year 2001 (spring–summer season) 73 measurements of radon (222 Rn) concentration in soil gas were made. In the same 73 sites thoron (220 Rn) concentrations were measured with the AlphaGUARD gauge. The concentrations of the natural radioactive elements (40 K, 226 Ra, 232 Th) were determined in soil and ground samples collected from all sites. Signiﬁcant spatial variability of radon 222 Rn concentration was observed. An average radon concentration of 38.4 kBq m−3 was obtained using the active method with median of 33.3 kBq m−3 . The same average concentration obtained using CR-39 detectors (the passive method) amounted to 37.2 kBq m−3 . The maximum value of radon concentration was 89 kq m−3 . Along with the measurements of 222 Rn, measurements of thoron (220 Rn) concentrations in soil gas were carried out. Its average value equaled 11.4 kBq m−3 with median of 10.9 kBq m−3 . Only in two locations were the concentrations higher than 30 kBq m−3 .

Correlation between radon concentration and geological structure of the Kraków area

467

Fig. 2. Cross section of the alpine fault zone rimming the northern slope of Sowiniec Hill, obtained by the shallow acoustic method and the distribution of radon concentration (average values) in soil gas obtained by the active method along the section line.

For most of the collected soil samples the lithology was loess or loesslike deposit. The data indicated that all collected samples have very similar values of 226 Ra, 232 Th, 40 K concentrations. The average concentrations amounted to 472 ± 26 Bq kg−1 for potassium 40 K, 35 ± 4 Bq kg−1 for radium 226 Ra and 49 ± 5 Bq kg−1 for thorium 232 Th. These values are in good agreement with such results obtained for loesses in other regions of Poland [12] and are close to the average concentrations of natural radioactive elements in soil given in the UNSCEAR report [13]. The homogeneity of loess samples, conﬁrmed by natural radionuclide concentrations, does not explain the variability of 222 Rn concentrations in soil gas. Figures 2 and 3 show the results of interpretation of shallow acoustic seismic logging with the average values of radon (222 Rn) and thoron (220 Rn) concentrations in the respective areas. The x axes in the ﬁgures represent the length of a measurement proﬁle, which was investigated with shallow acoustic seismic. Higher radon concentrations were observed in the vicinity of faults on most proﬁles. The obtained results may indicate that radon is produced mainly in other, probably deeper, geological layers. In the loess dense cover, which may constitute a barrier for radon migration, it is accumulated. The relatively thick loess layer is generally impermeable, however the dense host loess may be penetrated by numerous vertical ﬁssures. As a result a porous, secondary structure is obtained where relatively easy vertical migration of both descended and ascended geoﬂuids will be possible. Thoron (220 Rn) concentrations were up to 37 kBq m−3 , but 75% of the measured concentrations varied from 5.5 to 23.5 kBq m−3 . The concentration of thoron does not show the signiﬁcant correlation with geological structure of the bedrock. This may

468

K. Kozak et al.

Fig. 3. Cross section of the alpine fault zone rimming the northern slope of Astronomical Observatory Hill, obtained by the shallow acoustic method and the distribution of radon concentration (average values) in soil gas the obtained by the active method along the section line.

mean that thoron is locally produced mainly in the loess. The mean distance of thoron migration is not big because of its short half-life (55.6 s), much shorter than that of radon 222 Rn.

5. Conclusions The results of measurements of radon in soil in the vicinity of fault zones are presented. The investigation area was chosen using the available geological data and maps and the probable courses of faults were determined. The application of geophysical methods allowed the geological structure of the investigated area to be identiﬁed more exactly. The average radon concentrations in soil gas within this area were found to be about three times as high as in other regions of the Kraków agglomeration. A connection between elevated radon concentrations and the geological structure of the deep basement was found. The concentrations of 222 Rn change signiﬁcantly along lines perpendicular to the fault zones and reach maximum values in their direct vicinity. The results of thoron (220 Rn) measurements in that region are also presented. Unlike 222 Rn concentrations, they do not show essential variability along lines perpendicular to faults. This may indicate that thoron is produced only in the sub-surface loess layer which is the homogeneous structure consisting of poorly diverse material. The concentrations of natural radioactive elements (40 K, 226 Ra, 232 Th) in soil samples, collected from sites where radon was measured, are characterized by very small changeability. Loess layers can play an important role in processes of radon transport, but their signiﬁcance is still insufﬁciently known. The loess layer, from several to a few dozen meters thick, can be a barrier for radium (produced in deeper layers) and radon migration.

Correlation between radon concentration and geological structure of the Kraków area

469

Radium and radon, when migrating towards the ground surface and encountering a loess layer, are probably accumulated in it. This may cause locally higher radon concentrations around fault zones. More precise identiﬁcation of the mechanisms which are responsible for higher radon concentrations and the determination of radon migration routes in shallow geological layers require further research in situ as well as a more complete investigation of the physical characteristics of the ground.

Acknowledgement This work was partly supported by Research Grant 6 P04D 026 19 from the Polish Committee for Scientiﬁc Research.

References [1] J. Jagielak, M. Biernacka, J. Henschke, A. Sosi´nska, Radiological Atlas of Poland, 1998, Warszawa. [2] I. Barnet, J. Miksova, R. Tomas, J. Karenova, Radon risk mapping of the Czech Republic on a scale 1:50 000, in: Radon Investigations in the Czech Republic VIII and the Fifth International Workshop on the Geological Aspects of Radon Risk Mapping, Czech Geological Survey and Radon corp., Prague, 2000. [3] G. Akerblom, Investigation and mapping of radon risk areas, in: F.C. Wolff (Ed.), Geology for Environmental Planning, in: Geological Survey of Norway, 1986, pp. 96–106. [4] R. Strzelecki, S. Wołkowicz, J. Szewczyk, P. Lewandowski, Radioecological Maps of Poland, 1994, Warszawa (in Polish). [5] J. Kemski, R. Klingel, H. Schneiders, A. Siel, J. Wiegand, Radiat. Prot. Dosim. 45 (1/4) (1992) 235–239. [6] A. Vasarhelyi, I. Hunyadi, I. Csige, J. Hakl, E. Hertelendi, J. Borossay, K. Torkos, Radon enriched deep earthgas upﬂow in a seismically active inhabited area, in: Proc. ICRGG, vol. 3, India, 1997. [7] D. Mazur, M. Janik, J. Łoskiewicz, P. Olko, M. Paszkowski, J. Swako´n, Identiﬁcation of the areas with an elevated level of radon in the Kraków agglomeration, in: I. Hunyadi, et al. (Eds.), Proc. of the 5th Int. Conf. on Rare Gas Geochemistry, Debrecen, Hungary, 30 August–3 September, 1999, 2001, p. 127. [8] R. Gradzi´nski, Folia Geographica Polonica, vol. VIII, 1974 (in Polish). [9] J. Rutkowski, Geological Map of Poland 1:5000, Kraków Region with Explanations, Warszawa 1989 (in Polish). [10] J. Swako´n, J. Mazur, in: Measurements of Radon Concentration in Soil Gas, PAA Jelenia Góra, 2000, p. 140 (in Polish). [11] D. Mazur, M. Janik, J. Łoskiewicz, P. Olko, J. Swako´n, Radiat. Measur. 31 (1999) 295–300. [12] A. Solecki, Radiometric anomalies in the middle part of Fore-Sudetic region and their connection with geological environment, in: Geological–Mineralogical Papers, vol. LXIX, 2000, Wrocław. [13] UNSCEAR (United Nations Scientiﬁc Committee on the Effects of Atomic Radiation), Sources and Effects of Ionizing Radiation, Report to the General Assembly, with Scientiﬁc Annexes, United Nations, New York, 1993, Sales No. E. 94.IX. 2.

470

Application of a portable LS counter for measurements of radon progeny in air S. Chalupnik a , A. Kies b a Laboratory of Radiometry, Central Mining Institute, Katowice, Poland b Centre Universitaire, Luxembourg

One of the possible applications of liquid scintillation counting is measurement of radon progeny. There are certain advantages of this method, especially its high counting efﬁciency for alpha and beta particles, emitted by 218 Po, 214 Pb, 214 Bi and 214 Po. This advantage has been pointed out several years ago, when this method has been applied for calibration of portable monitors of radon progeny, especially due to the fact that in the case of radon progeny no standard atmosphere exists. The main reason is that alpha and beta particles are counted in liquid scintillator with an efﬁciency close to 100%. Radon progeny might be collected on the ﬁlter, which after immersion in the liquid scintillator becomes transparent and can be counted without signiﬁcant quenching. Therefore this method can be stated as an absolute one and could be used very widely for radon progeny monitoring. But the main drawback for a long time was the lack of portable LS counters. In the last few years new portable instruments appeared and, therefore, the monitoring of radon progeny in different environments is possible. One of the portable monitors is the Triathler, produced by the Hidex Company. The version of the Triathler with lead shielding has signiﬁcantly lower background; moreover, the counter is equipped with the feature of alpha/beta separation. Such an LS spectrometer has been used in our investigations in the Centre Universitaire in Luxembourg for radon progeny monitoring. Application of a pump with high ﬂow-rate enabled measurements of radon progeny in outdoor air. Investigations in an underground, old gypsum mine have been performed as well. Although the method seems useful, it will have rather limited application because LS readers are expensive. Moreover, now we have several alternative sampling and counting methods that can provide radon progeny measurements at lower cost with good accuracy. On the other hand, the calibration of the method based on LS counting is much easier and reliable. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07056-1

© 2005 Elsevier Ltd. All rights reserved.

Application of a portable LS counter for measurements of radon progeny in air

471

1. Introduction The radiation risk caused by radon (radon progeny) was primarily recognised in uranium mines. Therefore, at ﬁrst, methods of measurement of radon progeny concentration and instrumentation were developed for the uranium industry [1–3]. Later, elevated radon and radon daughter concentrations were found in other types of underground mines [4–6]. In Poland monitoring of radon in ore mines was started in the late 1970s by means of passive track detectors and RGR portable monitors. At the beginning of the 1980s a research programme, on radon and radon progeny in coal mines, was started by the Laboratory of Radiometry in the Central Mining Institute [7]. Sources of radon and the inﬂuence of exploitation and ventilation conditions on radon and radon progeny concentrations have been investigated. Due to the special requirements of instrumentation intrinsic safety in Polish coal mines a special integrating monitor of PAEC was developed [8]. In this device, called the ALFA-31 sampling probe, thermoluminescence detectors (CaSO4 : Dy) were applied for alpha particle counting [9]. Three independent heads are placed over the membrane ﬁlter (FM-1) in the dust sampler’s microcyclone. This solution enables simultaneous measurements of PAEC and dust content. Moreover, the information stored in the TLD chips is the energy of the alpha particles and not the number of counted particles. Therefore, the readout of TL detector shows directly the potential alpha energy, with no dependence on equilibrium factor, etc. Moreover, the implementing of that device within coal mines was approved by the mining authority [10]. The application of a convenient and reliable method of calibration of ALFA-31 monitors appeared to be difﬁcult, as references concerning methods of calibration are rather rare. Even as complete a compilation as NCRP report No. 97 Measurement of radon and radon daughters in air [11] in its chapter Calibration contains only a kind of framework of guidelines, without any references. Usually, descriptions of a method of calibration of PAEC monitors are internal reports only. A lack of primary standards of radon progeny leads to the conclusion that the best calibration method should be an absolute one. Therefore, our efforts have been focused on the liquid scintillation technique, in which measurements of alpha and beta particles with an efﬁciency close to 100% are possible. The biggest problem could be the quenching effect of a ﬁlter in the scintillator, but a membrane ﬁlter become transparent in a toluene based liquid scintillator [12]. This method has been applied as a calibration method only, because at that time there were no portable LS counters.

2. Theoretical background The time regime of the calibration method is the classical Thomas method [13]. It has been chosen because it can give not only the PAEC concentration but also concentrations of any single nuclide – 218 Po, 214 Pb and 214 Bi (the same as the concentration of 214 Po). The disintegration of nuclides from the same series, as, for instance, radon progeny, are described by Bateman’s equations [14] dNi = λi−1 · Ni−1 − λi · Ni , dt where:

(1)

472

S. Chalupnik, A. Kies

• Ni is number of atoms of ith nuclide in the series; • λi is decay constant (s−1 ). To describe the changes of activities during collection of nuclides on the ﬁlter an additional term must be added. If the ﬂowrate through the ﬁlter is V (L min−1 ), and a concentration of atoms of the ith nuclide in air ni , than we have dNi = λi−1 · Ni−1 − λi · Ni + V · ni . dt

(2)

Typical solutions are cited in many publications [11,13], but only for alpha decays. As in liquid scintillators alpha and beta particles are counted simultaneously, we must recalculate it for this case. At ﬁrst, a solution of equations (2), for sampling time tp links concentration of short lived radon daughters in air with their concentrations collected on the ﬁlter AA (tp ) =

CA · V 1 − e−λA ·tp , λA

CA · V CB · V λB 1− 1 − e−λB ·tp + e−λA ·tp λB λA λB − λA λA + e−λB ·tp , λB − λA CB · V CC · V λB λC AC (tp ) = 1− 1 − e−λC ·tp + e−λB ·tp + e−λC ·tp λC λB λC − λB λC − λB CA · V λC · λA λB · λC e−λA ·tp + e−λB ·tp + 1− λA (λB − λA )(λC − λA ) (λB − λA )(λC − λB ) λA · λB (3) e−λC ·tp , − (λC − λA )(λC − λB ) AB (tp ) =

where: • CA , CB and CC are concentrations of 218 Po, 214 Pb and 214 Bi in air; • AA , AB and AC are activities of 218 Po, 214 Pb and 214 Bi on the ﬁlter. These results are the input data to equations (1). The solution of these equations is similar to the solution of equations (2): AA (t) = AA (tp ) · e−λA t , AB (t) = AB (tp ) · e−λB t + AA (tp ) ·

λB −λA t e − e−λB t , λB − λA

Application of a portable LS counter for measurements of radon progeny in air

AC (t) = AC (tp ) · e−λC t + AB (tp ) · + AA (tp ) · λC · λB ·

473

λC −λB t e − e−λC t λC − λB

1 e−λA t (λB − λA ) · (λC − λA )

1 1 −λB t −λC t . e e + − (λB − λA ) · (λC − λB ) (λC − λA ) · (λC − λB )

(4)

The solution for the Thomas method (sampling time tp = 10 min, and three consequent counting periods, between 2 and 5, 6 and 20 and 21 and 30 min after the sampling period) is as follows [12]: N1 523.0 2256.4 2684.4 CA (5) N2 = 1470.0 11279.0 9054.8 · CB · V , CC N3 520.0 7012.8 3723.4 where N1 , N2 and N3 are numbers of counts, obtained during ﬁrst, second and third counting period, respectively. To calculate concentrations of 218 Po, 214 Pb and 214 Bi, a reverse matrix to the matrix of coefﬁcients (equation (5)) must be found. This gives us as a result CA 52.690 −25.570 24.175 N1 10−4 . (6) CB = −4.775 0.580 2.030 · N2 · V C N −2.605 4.530 −6.450 C

3

Now it is possible to calculate the potential alpha energy concentration Cα : Cα = 0.579 · CA + 2.850 · CB + 2.087 · CC ,

(7)

where: • Cα is in μJ m−3 ; • CA , CB and CC are in kBq m−3 . Finally, Cα can be expressed as a function of N1 , N2 and N3 : (11.430 · N1 − 3.650 · N2 + 6.250 · N3 ) × 10−4 . (8) V Using these formulae on the results of the measurements of the ﬁlter in the liquid scintillation counter, the concentrations of the radon progeny and potential alpha energy concentration can be calculated. Cα =

3. Application of the Triathler portable LS counter for radon progeny monitoring in air One of the possible applications of liquid scintillation counting is in measurement of radon progeny. There are certain advantages of this method, especially the high counting efﬁciency for the alpha and beta particles emitted by 218 Po, 214 Pb, 214 Bi and 214 Po. This advantage has been pointed out some years ago, when such a method was applied as a calibration method of

474

S. Chalupnik, A. Kies

portable monitors of radon progeny [12], especially due to the fact that in the case of radon progeny no standard atmosphere exists. On the other hand, alpha and beta particles are counted in a liquid scintillator counter with an efﬁciency close to 100%. Radon progeny are collected on the ﬁlter and the ﬁlter after immersion in the liquid scintillator becomes transparent and can be counted without signiﬁcant quenching. Therefore this method can be stated to be an absolute one and used very widely for radon progeny monitoring. But the main drawback for a long time was the lack of portable LS counters. In the last few years new portable instruments appeared and, therefore, the monitoring of radon progeny in different environments is possible. One of the portable monitors is the Triathler, produced by the Hidex Company (Finland). As the Centre Universitaire in Luxembourg has two counters, it is possible to use these instruments for radon progeny monitoring. Much better is the version of the Triathler with lead shielding due to its lower background. This LS spectrometer has been used in our investigations.

4. Preparation for ﬁeld measurements A very important problem was to ﬁnd a pump with high ﬂowrate, because the detection limit was dependent on the total amount of air pumped through the ﬁlter during sampling. The pump, which has been applied in our experiments was completely clogged, when membrane ﬁlters have been tested. Therefore ﬁbreglass ﬁlters, from Muster and Nagel, have been used. Application of ﬁbreglass ﬁlters ensured the transparency of ﬁlters immersed in the scintillator (counting efﬁciency close to 100%) and enabled high ﬂowrate, of about 300 L min−1 . The next step was to choose a proper scintillation cocktail. In the Central Mining Institute a toluene-based scintillator is used during calibration procedures. But due to strict rules in the EU, application of a toluene cocktail is forbidden. Fortunately, the cocktail OptiScint “HiSafe 3”, produced by Perkin-Elmer and based on DIN, provides the same opportunities and performance as the toluene one. We use typical glass vials with 12 ml of the cocktail. Measurements of the background in a wide window in the Triathler have been done. The clean ﬁlter was immersed into the scintillator and a blank sample was measured. We found a rather high background of the instrument, ranging from 150 to 250 cpm. This is the main reason that a high volume pump must be used for air sampling. The next step was the calculation of the lower limit of detection (LLD). Assumptions have been made for a ﬂowrate of 300 L min−1 , a counting time of 10 min, and a background of about 200 cpm. The detection limit for alpha potential energy is estimated to be approximately 0.0008 μJ m−3 . Despite the relatively high background the detection limit is very low, because of the very good counting efﬁciency (100%) and high ﬂowrate, which compensate for the background inﬂuence.

5. Results of ﬁeld experiments Our measurements have been performed in the building of the Centre Universitaire (CU) in Luxembourg, in laboratories, cellars and outdoors. Additionally some experiments have been made in an abandoned gypsum mine nearby. Most of the results are shown in Table 1.

Application of a portable LS counter for measurements of radon progeny in air

475

Table 1 Results of measurements in the Centre Universitaire, performed with application of the Triathler portable LS spectrometer cpm1

cpm2

cpm3

Backgr. (cpm)

v (L min−1 )

C1 Po-218 conc. (Bq m−3 )

C2 Pb-214 conc. (Bq m−3 )

C3 Bi-214 conc. (Bq m−3 )

Cα (μJ m−3 )

Laboratories of CU 1438 1126 632 566

1022 512

230 230

310 310

13.72 2.15

1.44 1.15

0.45 0.49

0.0130 0.0055

Cellars in CU building 60 780 49 180 71 430 58 480 80 200 66 450 87 400 68 300

41 320 48 710 53 560 57 200

360 280 350 280

310 300 255 240

318.03 315.17 209.38 750.73

90.08 111.84 142.44 142.33

79.44 108.06 188.06 137.22

0.6068 0.7271 0.9206 1.1269

Old gypsum mine 93 430 73 450 8 980 300 7 789 200

61 675 4 963 400

240 240

20 240

9424.77 −120 399

1894.00 10 533.66

1745.16 56 535.78

14.4996 78.7825

3420 1054

270 200

255 245

49.99 12.90

3.93 2.08

4.85 1.67

0.0618 0.0169

Repeated measurements in gypsum mine 109 443 86 797 74 320 200 47 910 39 200 32 340 200

20 7

11 973.53 8188.00

2462.17 3149.03

1677.91 3354.28

17.4504 20.7279

Outdoors 5500 1505

3984 1207

The measurements in the CU laboratories showed very low concentrations of potential alpha energy of radon progeny (0.005–0.013 μJ m−3 ). This corresponds to the value for outdoors, but measurements were done during a hot summer with open windows in laboratories. This is probably the explanation of the results gathered. This thesis can be supported by strong disequilibrium amongst particular radon progeny, which seems to be due to the high ventilation rate. Other measurements were done in the cellars and underground car park of the CU building. Concentrations of total alpha energy concentrations varied in a range 0.5–1.2 μJ m−3 , while concentrations of 218 Po were in a range 300–700 Bq m−3 . A radon chamber of CU is located in one of these cellars. Results from conﬁned rooms, with limited ventilation and exhalation of radon from the ground underneath the building, were in agreement with our expectations. The next step was an attempt to measure radon progeny concentrations outside the building. We found that such measurements must be performed very carefully; especially important is to avoid the exposure of the vial with scintillator to the sun. Luminescence, induced in the scintillator, signiﬁcantly increased the background and the results of the ﬁrst measurement were wrong. The solution is to keep the vial with the scintillator in a place with no direct sunlight. Further measurements done in the same and following days showed that such monitoring can be done with very good precision and accuracy. The ﬁnal step was the monitoring of radon progeny concentrations in an abandoned gypsum mine. At ﬁrst, a high volume pump for air sampling was used, and the results were completely

476

S. Chalupnik, A. Kies

erroneous. Too high concentrations of radon progeny, collected on the ﬁlter, produced a very, too high, count-rate in the LS counter. Therefore, we decided to use a pump with much lower ﬂowrate (7 L min−1 ). Also ﬁlters from personal dust masks were measured. We would like to point out the very good agreement of the results obtained for ﬁlters from a personal respirator. This respirator has been used for inhalation during a period of 10 min and later the ﬁlter was transferred into liquid scintillator and counted. When we assumed the ﬂow rate through the mask at the level 20 L min−1 , the results were close to those from the mechanical pump. We found that the concentrations of potential alpha energy concentration in the mine were at the level 20 μJ m−3 . Unfortunately, we were not able to compare our results with any other instrument for radon progeny monitoring. At the same time, radon monitoring in the mine was done with application of AlphaGuard instrumentation, showing results at the level 10– 12 kBq m−3 , in quite good agreement with calculated concentrations of 218 Po.

6. Summary We have tested the possibility of application of a portable LS counter (Triathler, Hidex Oy, Finland) for measurement of radon progeny concentration in air. Tests were performed in the Centre Universitaire, Luxembourg. Results of preliminary measurements were satisfactory (see Table 1). The very good sensitivity of the method leads sometimes to very peculiar effects. During our ﬁrst measurements in the gypsum mine the count rates were too high for the counter to make a proper correction of the count rate – more than 8 000 000 CPM in the ﬁrst counting period. Results calculated on this basis were completely wrong, especially for 218 Po concentration which was found to be negative. Therefore, special attention must be paid to applying a proper ﬂow rate to avoid such problems. Additionally, outdoor measurements near CU showed the signiﬁcant inﬂuence of luminescence on the count rate, especially during the ﬁrst counting period. The application of a high volume pump enables application of the method for outdoor measurements. We would like to point out that the method could be regarded as the absolute one; therefore, problems with calibration can be avoided.

7. Final conclusion On the basis of our preliminary results the ﬁnal conclusion can be drawn that application of a portable LS counter for radon progeny monitoring in the environment seems to be a very useful and sensitive method.

References [1] E.C. Tsivoglou, H.E. Ayer, D.A. Holaday, Occurrence of non-equilibrium atmospheric mixtures of radon and its daughters, Nucleonics 11 (9) (1953) 40. [2] H.L. Kusnetz, Radon daughters in mine atmospheres: A ﬁeld method for determining concentrations, Am. Ind. Hyg. Assoc. Quarterly 17 (1956) 85.

Application of a portable LS counter for measurements of radon progeny in air

477

[3] R. Rolle, Improved radon daughter monitoring procedure, Am. Ind. Hyg. Assoc. J. 30 (1969) 153. [4] E. Stranden, L. Berteig, Radon daughter equilibrium and unattached fraction in mine atmospheres, Health Phys. 42 (1982) 479. [5] G. Schiocchetti, Grab sampling and continuous monitoring of radon and its progeny in air, in: G. Furlan (Ed.), Proceedings of the Second Workshop on Radon Monitoring, Radioprotection, Environmental and/or Earth Sciences, World Scientiﬁc, Singapore, 1991, p. 123. [6] J. Lebecka, I. Tomza, J. Skowronek, K. Skubacz, S. Chalupnik, Monitoring of radiation exposure from different natural sources in Polish coal mines, in: Proceedings of Int. Conference on Occupational Safety in Mining, Toronto, 1984, p. 408. [7] J. Lebecka, J. Skowronek, I. Tomza, S. Chalupnik, K. Skubacz, A thermoluminescent monitor of low radon daughters concentrations in air, J. Appl. Radiat. Isot. 39 (9) (1988). [8] J. Lebecka, I. Tomza, K. Skubacz, T. Niewiadomski, E. Ryba, Monitoring of radon daughters in coal mine atmosphere, in: Proceedings of 3rd Int. Mine Ventilation Congress, Harrogate, 1983, p. 121. [9] T. Niewiadomski, E. Ryba, Radon and its daughter products in mines, Report of Institute of Nuclear Physics No. 28/81/NPP, Kraków, 1982 (in Polish). [10] WUG, Approval for ALFA-31 sampling probe for application in coal mines, State Mining Authority (Wyzszy Urz¸ad Górniczy), Katowice, 1988 (in Polish). [11] NCRP, Measurement of radon and radon daughters in air, NCRP report No. 97, National Council on Radiation Protection and Measurements, Bethesda, MD, 1988. [12] S. Chalupnik, A method of calibration of PAEC monitors, PhD thesis, Central Mining Institute, Katowice, 1996 (in Polish). [13] J.W. Thomas, Modiﬁcation of the Tsivoglou method for radon daughters in air, Health Phys. 19 (1970) 691. [14] H. Bateman, The solution of a system of differential equations occurring in the theory of radioactive transformations, Proc. Cambridge Philos. Soc. 15 (1910) 423.

478

Calibration and quality assurance of radon measurements in Sweden. Activities in 1990–2000 N. Hagberg Department of Environmental and Emergency Assessment, Swedish Radiation Protection Authority, SE-171 16 Stockholm, Sweden

Calibration of instruments and detectors used for indoor radon measurements has been carried out at the Swedish Radiation Protection Authority (SSI) since the late seventies. Protocols for measurements of radon in dwellings have been in force since 1980, with revisions due to changes in the regulations, new measuring methods and other factors. The measurement protocols require regular calibration of the measurement devices. Instruments for continuous measurements have to be calibrated at least every 12 months. Passive detectors such as tracketch detectors are calibrated by exposure of a selection of the total production of detectors at each laboratory offering measurements. In this report the equipment used for calibrations is described and the calibration activities during the last 10 years are reported. International intercomparisons are an important part of the quality assurance used as a tool to verify the overall performance of a measurement system. Results from NRPB intercomparisons of passive radon detectors from SSI and other Swedish laboratories are reported as well as results from some national Swedish intercomparisons. 1. Introduction The calibrations are based on standards traceable to the National Institute of Standards and Technology (NIST) through the use of a 226 Ra standard reference solution. The reference activity is used to calibrate secondary standard instruments in order to establish the traceability for the calibration of ﬁeld instruments and detectors illustrated in Fig. 1. An essential part of the facilities necessary for calibration of ﬁeld instruments for 222 Rn and 222 Rn progeny is a “radon room”, where an atmosphere with known activity per volume (Bq m−3 ) of 222 Rn and 222 Rn progeny can be maintained during different atmospheric conditions. The radon room at SSI is an air-tight construction of stainless steel with thermal insulation in the double wall construction. The room is equipped with a separate ventilation system so that a constant air-exchange rate can be maintained. The 222 Rn source used is either a set of dry 222 Rn emanation sources or uranium ore. The sources are placed outside the RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07057-3

© 2005 Elsevier Ltd. All rights reserved.

Calibration and quality assurance of radon measurements in Sweden. Activities in 1990–2000

479

Fig. 1. Traceability to international standards.

chamber in smaller enclosures. By using different sources and different air-exchange rates the 222 Rn concentration in the chamber can be chosen between outdoor concentration and more than 20 kBq m−3 . The standards and reference instruments used at SSI are described in detail in Falk et al. [1]. 2. Standards for 222 Rn activity concentration in air 2.1. Primary standard The primary standard for 222 Rn consists of an NIST 226 Ra standard reference solution in a set of glass containers. Figure 2 shows a schematic diagram of the 222 Rn reference bubbler. The three gas washing bottles are made of glass and fused to form a unit. Also, the stoppers and the valves are made of glass to ensure a radon-tight system. The ﬁrst glass washing bottle contains distilled water, the second the NIST 226 Ra standard reference material and the third glass wool. When ﬂushing the system with air, the air is ﬁrst moistened in the water bottle, the air passes through the reference radium solution and ﬁnally through the glass wool, where water droplets will be removed from the 222 Rn-laden air stream. The build-up of 222 Rn in the system is determined by the 226 Ra activity of the reference solution and the decay constant of 222 Rn. For calibration purposes a build-up time of between 7 h and 14 days has been used. To obtain a reference air concentration of 222 Rn, the activity in the bubbler is ﬂushed into an evacuated metal container with an accurately known volume. The volume of the container is 4740 cm3 and a complete removal of the accumulated 222 Rn occurs when the bubbler is

480

N. Hagberg

Fig. 2. Primary standard for 222 Rn.

ﬂushed with the same volume of air. The container used serves three purposes: as a reference volume, a transfer container and a sampling vessel. 2.2. Secondary standard I The secondary standard for 222 Rn in air is an instrument based on α-spectrometric measurement of 218 Po and 214 Po electrostatically collected on a surface barrier detector. Figure 3 shows a schematic diagram of the instrument. The sensitive volume of the instrument is a sphere of aluminium with a volume of 10.8 L that contains the 222 Rn sample to be analysed. A surface barrier detector of 150 mm2 active area is mounted on the inside surface of the sphere, electrically isolated from the metal sphere. A potential of 8 kV is applied between the detector surface and the metal sphere generating an electrostatic ﬁeld that draws the charged 218 Po atoms to the detector surface. The decay of 222 Rn within the sphere will generate 218 Po at a rate proportional to the 222 Rn activity. A build-up of 218 Po activity takes place on the detector surface, and after approximately 20 min of collection a steady state is reached. The

Fig. 3. The secondary standard I for 222 Rn.

Calibration and quality assurance of radon measurements in Sweden. Activities in 1990–2000

481

activity detected by the α-detector is then proportional to the 222 Rn activity in the sphere. The calibration of the secondary standard is carried out using a procedure whereby a part of the activity in the reference volume is transferred into the sensitive volume of the instrument in a controlled and reproducible way. The response of the instrument is thus directly related to the 222 Rn concentration in the reference volume and the traceability to the NIST standard is thereby achieved. The secondary standard is the reference instrument used for grab sample measurements and for the calibration of other instruments. For grab sample measurements the same container, or one identical to the reference volume, is used. The container is evacuated and a short tube containing a drying agent is connected to the inlet of the container. The sample is taken by opening the valve on the container and leaving the valve open long enough to ensure pressure and temperature equilibrium with the ambient air. For grab samples, the same procedures as for the calibration are used to transfer the 222 Rn from the container to the sensitive volume of the instrument.

218 Po

2.3. Secondary standard II For 222 Rn concentrations above 10 000 Bq m−3 another instrument is used as the secondary standard. The method used is a direct measurement of 222 Rn in a Marinelli beaker by means of γ-spectrometry. A vacuum-tight metal Marinelli beaker with a volume of 3.4 L ﬁlled with a metal sponge matrix has been designed. The matrix is necessary to give a homogeneous distribution of 222 Rn decay products within the beaker. A HPGe detector with a background shield is used for the γ-measurements. This secondary standard is calibrated by ﬂushing the 222 Rn activity from the bubbler directly into the evacuated Marinelli beaker. A detailed description of the method including calibration and measuring procedures and uncertainty analysis is given by Möre and Huang [2]. 2.4. The working standard Continuous monitoring of the 222 Rn concentration in the radon room is made with an instrument identical to the α-spectrometric instrument for grab samples, but it works in a ﬂowthrough mode. Figure 4 shows a schematic diagram of the instrument. Air from the radon room is continuously pumped through a ﬁlter, to remove aerosols and 222 Rn progeny, and a drying agent, to reduce the humidity to a low and stable level, before it enters the sensitive volume of the detector. 218 Po will be generated at a rate proportional to the 222 Rn concentration in the air. A build-up of activity on the detector surface takes place and it reaches a steady state if the 222 Rn concentration in the room is constant. This instrument is connected to a computer, which keeps control of the measuring intervals, performs the necessary calculations and stores the data and results in ﬁles. Measured and calculated hourly mean values are used to assess the 222 Rn exposure of other instruments and detectors exposed in the radon room. The calibration of the working standard is done indirectly. While keeping the 222 Rn concentration in the room at different but constant levels several grab samples are taken and measured by the secondary standard. Traceability to the NIST standard is thereby maintained.

482

N. Hagberg

Fig. 4. The working standard for 222 Rn in the radon-room.

A detailed description of the instrument, the calibration procedure and the uncertainty estimation are given by Möre [3]. 3. Standards for short-lived 222 Rn progeny in air The activity concentration (Bq m−3 ) of the airborne short-lived 222 Rn progeny 218 Po, 214 Pb, and 214 Bi and 214 Po or the potential alpha energy concentration in air (J m−3 ) is usually determined by collection of the 222 Rn progeny on a ﬁlter and subsequent measurement of the α-activity of the ﬁlter. 3.1. The primary standard for 222 Rn progeny The primary standard is a sealed NIST 226 Ra standard reference material source with 222 Rn and its short-lived progeny in equilibrium with the 226 Ra activity. The reference source is a sealed Plexiglass container with a known amount of 226 Ra activity. The cavity in the Plexiglass container has the same diameter as the active surface of a ﬁlter to make the geometry of the activity distribution of the reference source as close as possible to that of the activity on a ﬁlter sample. After sealing the reference solution into the Plexiglass, leakage tests have been performed to ensure radon tightness. 3.2. The secondary standard for 222 Rn progeny The secondary standard or the reference instrument for 222 Rn progeny is based on the collection of 222 Rn progeny on a ﬁlter and subsequent measurement of the α-particles from 218 Po

Calibration and quality assurance of radon measurements in Sweden. Activities in 1990–2000

483

and 214 Po on the ﬁlter. An automated system, controlled by a computer, has been developed which includes sampling, measurement, calculation, displaying and storage of the result every hour. To avoid errors associated with ﬁlter changes, the same ﬁlter is used for the subsequent hourly measurements. A correction then has to be made for the remaining activity on the ﬁlter, from the previous sampling periods. During the sampling period, the ﬁlter in an open face ﬁlter holder is placed 25 cm from the detector to obtain a good sampling geometry. After the completion of the sampling, the ﬁlter holder is moved to the measuring position under a surface barrier detector (150 mm2 ), which is connected to a multichannel analyser. The α-activity on the ﬁlter is measured during two counting intervals and when the measurements and calculations are ﬁnished the same ﬁlter is moved away from the detector and reused in a new measuring cycle. The calculations are described in detail in [1]. The counting efﬁciency calibration of the α-detector is done indirectly via a γ-detector. The calibration technique used is based on simultaneous α- and γ-measurements of 214 Bi and 214 Po in an air ﬁlter sample. The calibration source is a ﬁlter sample where the collection of 222 Rn progeny has been carried out using the same ﬂow rate, ﬁlter type, etc., as those intended for later measurements. By using an actual ﬁlter sample as the calibration source as well as the same counting geometry, no correction for self-absorption, etc., has to be made. The determination of the 214 Bi activity of the sample is performed by γ-spectrometry measurement simultaneously with the measurement of the α-particles from 214 Po by the α-detector to be calibrated. Due to its short half-life, the 214 Po activity on the ﬁlter is in equilibrium with, and practically identical to, the 214 Bi activity. The calibration of the γ-detector is performed with the primary standard for 222 Rn progeny. The calibration of the α-detector is thus traceable to an NIST standard.

4. Calibration of secondary/working standards for 222 Rn Calibration of the secondary standard for grab samples and of the working standard in the radon room is performed on a regular basis. The procedures were described earlier, in Sections 2.2 and 2.4, respectively, and the results are illustrated in the diagrams in Figs. 5 and 6.

Fig. 5. Calibration of secondary standard I.

484

N. Hagberg

Fig. 6. Calibration of working standard.

Values shown are the calibration factors used in the calculations when the instruments are used for measurements of 222 Rn concentrations. Each point represents a mean value of repeated calibrations and the uncertainties (1 S.D.) are indicated. The discontinuity in the data for the grab sample instruments (secondary standard) is the result of a mechanical change in the surface barrier detector position in the chamber.

5. Calibration of ﬁeld instruments and detectors Measurement devices used for measurements in dwellings should be calibrated according to the requirements in the measurement protocols. Different types of instruments and detectors are used ranging from advanced continuous recording instruments to passive detectors for long duration integration. The measurement protocols require regular calibration of the measurement devices, at least every 12 months. The calibration of all types of devices is performed

Fig. 7. Numbers of instruments and detectors exposed for calibration at SSI.

Calibration and quality assurance of radon measurements in Sweden. Activities in 1990–2000

485

in the radon room at SSI as intercomparison measurements. Results from individual instruments are directly compared to the readings from the secondary/working standards in the radon room. Solid-state track detectors are exposed to known levels, and a calibration factor, including the chemical work at the processing laboratory, can be determined. Such calibration exposures are made on a selection of the total production of detectors. The numbers of instruments and detectors exposed in the radon room during the period 1985–2001 are shown in Fig. 7. In this diagram, different types of instruments and detectors are separated as indicated.

6. Intercomparison measurements and tests 6.1. Intercomparison measurements, 222 Rn samples To verify the performance of the secondary standards for 222 Rn measurements, intercomparisons have been made with international reference laboratories. The results from the intercomparisons carried out with NIST, USA, Environmental Measurements Laboratory (EML), USA and National Physical Laboratory (NPL), UK are shown in Table 1. 6.2. National intercomparison tests Besides the normal calibrations, intercomparison tests for quality control are made, for example, on track-etch detectors. A number of detectors from participating laboratories are exposed to a level unknown to the laboratories and reported results are compared with data from the working standard. In a ﬁrst intercomparison run in 1992, mean values from the six participating laboratories showed deviations from the SSI reference value between −20 and +40%. The distribution Table 1 International intercomparison measurements Reference laboratory

Measured concentration or activity

Ratio

SSI/Ref. lab.

SSI instrument used

NIST June 1990 EML April 1991 EML Nov. 1991 EML April 1992 EML Nov. 1992 EML April 1993 EML April 1994 EML April 1996 NPL July 1992 NPL May 1994

5500∗ 620 805 2050 1426 886 545 425 14 900† 10 400†

Bq Bq m−3 Bq m−3 Bq m−3 Bq m−3 Bq m−3 Bq m−3 Bq m−3 Bq Bq

0.995 1.02 0.97 0.99 0.96 0.96 0.96 0.97 0.999 1.00

222 Rn sec. std. I

∗ Mean value of 5 samples. † Mean value of 3 samples.

222 Rn sec. std. I 222 Rn sec. std. I 222 Rn sec. std. I 222 Rn sec. std. I 222 Rn sec. std. I 222 Rn sec. std. I 222 Rn sec. std. I 222 Rn sec. std. II 222 Rn sec. std. II

486

N. Hagberg

Table 2 Results from “blind-test” of track etch detectors on the Swedish market in 1995 Laboratory

Difference from reference exposure (%)

Deviation in reported values (1 S.D.) (%)

Type of detector

MRM Konsult Gammadata Strålskyddstjänst SP Fys. Inst. LTH Thunberg Production (Kodalpha detectors)

+5 −5 −19 +12 −7 −11

7 15 19 11 20 5

CR-39, Filtered CR-39, Filtered CR-39, Filtered CR-39, Filtered LR-115, Open LR-115, Open

between ten detectors from each laboratory was between ±3 and ±24% (1 S.D.). The detectors were both open and ﬁltered types and they were exposed to two different levels, 100 and 1000 kBq h m−3 , respectively. A second intercomparison run was carried out during 1995 [4]. A difference from the ﬁrst intercomparison was that the whole procedure was totally “blind” for the laboratories. Detectors were bought from the laboratories by a private person. They were exposed in the radon room at SSI, returned to the laboratories and processed. The results were ﬁnally reported to the person who bought the detectors. The results are presented in Table 2. The radon exposure level was 300 kBq h m−3 and the open detectors were exposed with an equilibrium factor F = 0.4. 6.3. International intercomparison tests International intercomparison tests of passive radon detectors have been arranged annually at the NRPB since 1997 starting with a series of three sponsored by the European Commission. The NRPB intercomparisons normally include about 50 different participating laboratories using them as a part of their quality control or for calibration purposes. The results are reported as deviations from the reference exposures (%) and the distribution in the results from detectors exposed to the same level (1 S.D.) Normally three different exposure levels are used and a mean value is used to form a “ranking list” of the participating laboratories. Results from the intercomparisons show that the Swedish laboratories, Swedish Radiation Protection Authority (SSI) and the two commercial laboratories Gammadata Mätteknik AB and MRM Konsult AB achieve high quality measurements at a stable level. Table 3 shows total mean values for laboratories participating in the intercomparisons 1997–2000. The table is limited to the ten laboratories showing the best results ranked by the sum in the last column. Only the Swedish laboratories are identiﬁed in the table. There are also other participating Swedish laboratories showing lower ranking. They are not accredited and they are subjected to only limited control by the authorities but the number of measurements carried out by those laboratories is small.

Calibration and quality assurance of radon measurements in Sweden. Activities in 1990–2000

487

Table 3 Results in NRPB intercomparison of passive radon detectors 1997–2000 Laboratory

Mean % diff.

Mean % S.D.

Mean sum

SSI Gammadata A B MRM Konsult C D E F G

3.1 3.7 3.6 4.5 7.1 8.0 5.2 7.4 8.2 9.2

4.6 5.0 6.1 6.6 4.4 4.9 7.9 7.4 8.5 7.8

7.7 8.7 9.7 11.1 11.5 13.0 13.1 14.8 16.7 17.0

7. Radon in water To measure 222 Rn in water, the two most common methods are either γ-spectrometry measurement directly on the sampled water in a radon-tight sample bottle or the use of a liquid scintillating (LSC) system. Factors important for the quality of the analyses are the sampling procedure, the bottles used and the calibration of the instrument. The sampling procedure must be described in instructions to be followed by the person taking the sample. A draft measurement protocol has been issued by SSI for the measurement procedures including requirements regarding the sampling vessels. Calibration of the instrument at the laboratory is normally performed using a reference 226 Ra solution. Commercial laboratories shall have their calibration veriﬁed by intercomparison measurements with SSI. Double water samples are taken; one is measured at the actual laboratory, the other is measured at SSI using a HPGe reference γ-spectrometer. On each occasion intercomparisons are made at two different levels, e.g., 100 and 1000 Bq L−1 , respectively, and deviations up to ±15% from the reference value are accepted. During the period 1995–2000, 100 water samples from 20 different laboratories were analysed in such intercomparisons.

8. Accreditation and training Since 1991 it has been possible for Swedish companies and laboratories to be accredited for measurement of 222 Rn in indoor air and in water. This accreditation is issued by the Swedish Board for Accreditation and Conformity Assessment, SWEDAC. It is based on the international standard EN ISO/IEC 17025 and requirements in the measurement protocols. The number of accredited laboratories can vary from one year to another, the actual number in 2001 was two laboratories accredited for measuring radon in indoor air and four for radon in water. There are no formal requirements for proﬁciency for 222 Rn measurements organisations in Sweden. However, most companies of good repute have technicians who have attended training courses organised by SSI. The radon training courses at SSI are divided into two

488

N. Hagberg

stages, both involving a total of four or ﬁve days of training. The second stage is concentrated on one of the following topics: – – – –

indoor air – measurement techniques; indoor air – remedial measures; radon in water – measurements and remedial measures; radon in soil.

These courses include both theoretical and practical tests and the individuals that have passed the tests are listed in a catalogue that is distributed to national and local authorities and potential customers. Today, there are more than 200 persons who have received training in one or more of the above-mentioned courses listed in the catalogue. For persons responsible for the radon measurements at accredited laboratories having passed a test at the corresponding training course is one of the requirements concerning their competence.

9. Future plans Established environmental objectives for Sweden can lead to an increasing demand for radon measurements in dwellings and at workplaces. This will also lead to an increasing number of calibrations of instruments and detectors. The equipment involved in calibration activities at SSI, e.g., standards and reference instruments, has been in use for many years and it is timeconsuming to keep it in the condition required for high quality measurements. A change to other instruments and methods is therefore under consideration. There is no National Reference Laboratory for activity measurements in Sweden. In practice, SSI has worked as a reference laboratory for radon measurements for many years. The possibility of giving the radon calibration laboratory at SSI the status of National Reference Laboratory is under discussion.

References [1] R. Falk, N. Hagberg, L. Mjönes, H. Möre, L. Nyblom, G.A. Swedjemark, Standards, calibration and quality assurance of 222 Rn measurements in Sweden, Nucl. Instrum. Methods Phys. Res. A 339 (1994) 254–263. [2] H. Möre, Z. Huang, Radon measurement using gamma spectrometry, Appl. Radiat. Isot. 43 (1992) 103. [3] H. Möre, SSI-report 93-04, Swedish Radiation Protection Institute, Stockholm, 1993. [4] N. Hagberg, SSI publication i 96-02, Swedish Radiation Protection Institute, Stockholm, 1996.

489

The diurnal change in the vertical distribution of atmospheric 222Rn due to the growth and rise of the stable stratiﬁcation height in the atmospheric boundary layer K. Yoshioka a , T. Iida b a Shimane-Prefectural Institute of Environmental Science, 582-1, Nishihamasada, Matsue,

Shimane 690-0122, Japan b Department of Nuclear Engineering, Graduate School of Engineering, Nagoya University, Furo-cho, Chikusa-ku,

Nagoya 464-8603, Japan

To analyze the time-change of 222 Rn vertical distribution in the atmospheric boundary layer, 222 Rn was continuously measured at 3 different heights – 30, 150 and 520 m – within a 6-km radius. The air temperature was synchronously observed at 4 different heights – 1, 33, 160 and 520 m – at the same site and then the vertical distribution of temperature lapse rates calculated from these measurements. The time-change of 222 Rn vertical distribution could be attributed to the time-change of the vertical distribution of the temperature lapse rate. On the basis of this relationship, the mechanism of 222 Rn vertical mixing has been generalized as a mixing model in the atmospheric boundary layer.

1. Introduction 222 Rn

is a radioactive inert gas, which usually escapes out of the earth’s crust. It cannot be eliminated from the atmosphere by deposition apart from radioactive decay. It is widely known that 222 Rn concentration varies extensively with time and location in the atmospheric boundary layer. In previous papers, the time-change of 222 Rn concentration has been analyzed in relation to wind direction, wind speed, atmospheric stability and/or various meteorological parameters [1–3]. However, in most reports, 222 Rn concentration has been observed in the surface boundary layer, while the spatial distribution of 222 Rn by observation of the atmospheric boundary layer is not yet reported. 222 Rn could be a useful tracer in the study of the transport of continental air masses due to the same movements as the atmosphere. The stability of an air mass, which moves vertically RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07058-5

© 2005 Elsevier Ltd. All rights reserved.

490

K. Yoshioka, T. Iida

in an adiabatic process, depends on the dry-adiabatic lapse rate in the unsaturated-air mass, or the moist-adiabatic lapse rate in the saturated-air mass. The stability of vertical motion of an air mass is inﬂuenced by the relative balance between the temperature lapse rate of the peripheral atmosphere and the dry- or moist-adiabatic lapse rate [4–8]. As a feasibility study on characteristics of 222 Rn as a tracer, we have continuously measured 222 Rn concentration synchronously at three sites at measurement heights 30, 150 and 520 m, within a 6-km radius around Shimane peninsula, and obtained the time-change of 222 Rn vertical distribution in the atmospheric boundary layer. At the same sites, we have synchronously continuously observed four air temperatures with measurement heights 1, 33, 160 and 520 m, and obtained the time-change of vertical-distribution of temperature lapse rate to analyze the relationship between the time-change of 222 Rn concentration and temperature lapse rate. We have analyzed the time-change of 222 Rn-vertical distribution by assuming a vertical-diffusion model, which depends on the diurnal change of the height of stable stratiﬁcation.

2. Measurement method The three measurement sites – Hamasada, Kashima, and Misakayama – are located within a 6-km radius on the Shimane peninsula on the main island of Japan, as shown in Fig. 1. The measurement height of 222 Rn at Hamasada is 30 m above sea level on the rooftop of our laboratory. The measurement heights at Kashima and Misakayama are respectively 150 and 520 m on mountaintops. Figure 2 shows a schematic diagram of the 222 Rn monitor [9]. First, moisture in the sample air is removed by a ﬁlter for eliminating aerosols and a phosphorus pentoxide (P2 O5 ) dehumidiﬁer. Then, using a diaphragm pump, the sample air ﬂows at 1 L min−1 into a 16.8 L chamber in which an alpha ray detector is ﬁtted. 218 Po and 214 Po, disintegrated out of 222 Rn in the sample air, were collected on the electrode with a high electric ﬁeld applied (the ZnS(Ag) scintillation detector is ﬁtted on the back), and the α-particles emitted when they decay were counted to calculate the concentration of radon every hour. In the natural atmosphere, most short-lived daughter nuclides of radon such as 218 Po and 214 Po are attached to aerosols and therefore they are all removed when sample air goes through a membrane ﬁlter (pore size: 0.8 μm). Also, because 220 Rn half-life is extremely short, it

Fig. 1. Location map of radon measurement.

The diurnal change in the vertical distribution of atmospheric 222 Rn

491

Fig. 2. Schematic diagram of radon monitor.

mostly decays in sample air in which its concentration is 10% or lower than that of 222 Rn, by radioactive disintegration while it ﬂows from the inlet to the chamber. The air temperature was measured at the same sites at 1, 33, 160, and 520 m. The air temperature lapse rate is calculated for 1–33, 33–160, and 160–520 m.

3. Results and discussion 3.1. Comparison of diurnal change of

222 Rn

concentration due to elevation

Figure 3 shows three diurnal changes of 222 Rn concentration in each month at three sites. concentration at Hamasada (henceforth H-Rn) is larger in the night than the daytime. This diurnal change shows a maximum at sunrise and a minimum at sunset through the year. The maximum value is the largest in autumn, the smallest in summer.

222 Rn

Fig. 3. 3-Diurnal changes of 222 Rn during Dec. 1996 to Nov. 1997.

492

K. Yoshioka, T. Iida

The diurnal change of 222 Rn concentration at Kashima (henceforth K-Rn) is generally smaller than that of H-Rn, showing a clear maximum in the night in autumn and winter, but indistinctly in spring and summer. The maximum appears 1–2 h after sunrise and the minimum 2–3 h after sunset. The diurnal change of 222 Rn concentration at Misakayama (henceforth M-Rn) is smaller than that of H-Rn and K-Rn, with a larger value in the daytime and lower at night. This diurnal change is delayed in these phases of H-Rn and K-Rn. These maxima and minima appear during the periods 900–1300 h and 2000–100 h, respectively.

Fig. 4. The correlation between 222 Rn and temperature lapse rate.

The diurnal change in the vertical distribution of atmospheric 222 Rn

3.2. Correlation of

222 Rn

493

concentration and temperature lapse rate

As shown in Fig. 4, it became clear through comparative analysis of their time-changes that concentration and temperature lapse rate are negatively correlated. The correlation coefﬁcient of H-Rn to TLR1 is −0.9; this is a good negative correlation; and to TLR2 is 0.4; this is not a good correlation. The coefﬁcient of K-Rn to TLR2 is −0.9; this is a good negative correlation and to TLR3 0.2; this is not a good correlation. The coefﬁcient of M-Rn to TLR3 is −0.8; this is a good negative correlation. These correlations between H-Rn and TLR2, and K-Rn and TLR3 are not good. This is because their heights are not the same. A phase delay occurs in the time-change of temperature lapse rate due to the height-interval rate. It was clear that 222 Rn vertical distribution is strictly controlled by the temperature lapse rate. 222 Rn

3.3. The relationship between the phase-change of temperature lapse rate due to height interval and the diurnal change of 222 Rn vertical distribution After sunset, as TLR1, the lowest layer, changes to temperature inversion from the normal proﬁle and the inversion layer then grows in the surface boundary layer, H-Rn begins to increase, as shown in Fig. 5. K-Rn begins to increase a few hours later, as no 222 Rn is vertically

Fig. 5. The diurnal change of vertical distributions of 222 Rn and temperature lapse rate.

494

K. Yoshioka, T. Iida

Table 1 Time zone of the maximum or minimum of 222 Rn

Winter (Dec.–Feb.) Spring (Mar.–May) Summer (Jun.–Aug.) Autumn (Sept.–Nov.)

Time zone of maximum H-Rn K-Rn

M-Rn

Time zone of minimum H-Rn K-Rn

M-Rn

07–08 06–07 05–06 06–07

10–13 09–12 09–11 09–11

15–16 15–17 17–18 16–17

20–23 23–01 22–23 22–23

08–10 07–09 06–07 08–09

15–16 15–17 17–18 16–17

transported from the lower layer because the height of K-Rn is higher than H-Rn; the weak diffusion continues till TLR2 changes to temperature inversion. In the night, H-Rn and K-Rn continue to increase due to the growth of the inversion layer and reach maxima before sunrise. The height of the inversion layer barely exceeds the altitude of Misakayama, so M-Rn shows a tendency to decrease due to the lack of vertical transport of 222 Rn from the lower layer and the continuity of weak diffusion upward. After sunrise, the inversion layer begins to disappear upward from the lowest layer due to thermal-convection with the solar radiation. Because of the change of TLR1 from temperature inversion to normal proﬁle, the mixing layer grows upward, so that H-Rn rapidly decreases and then reaches a minimum before sunset. TLR2 changes to the normal temperature proﬁle from the inversion 2–3 h behind TLR1, so that K-Rn begins to decrease 2–3 h behind H-Rn. This change of TLR2 means thermal convection becomes stronger and active and then enters the Ekman layer. 222 Rn in the surface boundary layer is thus vertically transported to the Ekman layer, so that M-Rn begins to increase. In the daytime, this thermal convection is most active, so that 222 Rn vertical distribution becomes M-Rn > K-Rn > H-Rn, in contrast to the night. Before sunset, H-Rn, K-Rn, and M-Rn have nearly the same distribution. As we presume that the vertical distribution of temperature lapse rate becomes the normal proﬁle in the whole atmospheric boundary layer, this suggests that the mixing depth has grown to the highest altitude in this time. As shown in Table 1, the time zones of the maximum and minimum of 222 Rn vertical distribution are delayed in phases depending on the altitude. 3.4. The vertical mixing model due to the time-change of

222 Rn

vertical distribution

Based on the time change of 222 Rn vertical distribution depending on the time-change of the vertical distribution of temperature lapse rate, the mechanism of 222 Rn-vertical mixing can be generalized as the diurnal model shown in Fig. 6. After sunset, the temperature inversion layer grows upward from the ground surface and its height increases gradually. Because the inversion layer suppresses upward diffusion of 222 Rn, it stays in the lower layer and its concentration increases. In the upper layer, 222 Rn concentration decreases due to the lack of vertical transport of 222 Rn from the lower layer and the continuity of diffusion upward. After sunrise, the mixing layer grows upward from the ground surface and therefore the stable stratiﬁcation is pushed upward. In the daytime, the mixing layer grows further upward

The diurnal change in the vertical distribution of atmospheric 222 Rn

495

Fig. 6. Vertical mixing model due to rise of stable stratiﬁcation height after growth of inversion layer.

and erodes the stable stratiﬁcation. Before sunset, the whole atmospheric boundary layer becomes the mixing layer and the boundary of the capping inversion becomes indistinct. In this case, the mixing layer sometimes reaches the free atmosphere. The phase of 222 Rn vertical distribution is delayed with increasing height, but it corresponds well to the diurnal change of the height of stable stratiﬁcation.

4. Conclusions 222 Rn

vertical distribution and the vertical distribution of temperature lapse rate have been observed in the atmospheric boundary layer and analyzed by these time-series measurements. The above analyses have clearly revealed the following: (1) the periodic changes of the time-change of 222 Rn vertical distribution and the vertical distribution of temperature lapse rate are the diurnal change and an annual change; (2) the diurnal change of 222 Rn vertical distribution depends on the diurnal change of the vertical distribution of temperature lapse rate; (3) the amplitudes of diurnal changes of 222 Rn and the temperature lapse rate become small as the height increases. They also make seasonal changes; (4) based on the diurnal change of the vertical distribution depending on the temperature lapse rate, a general model of the mechanism of atmospheric vertical mixing is proposed.

References [1] Y. Ikebe, Variation of radon and thoron concentrations in relation to the wind speed, J. Meteorol. Soc. Jpn. 48 (5) (1970) 461–468. [2] N. Fujinami, S. Esaka, Inﬂuence of solar radiation and heat emission on radon-222 daughter concentrations in surface air, J. Geophys. Res. 93 (D10) (1988) 12627–12629. [3] K. Yoshioka, The vertical proﬁle of 222 Rn concentration in the lower atmospheric boundary layer at Shimane peninsula, in: Radon and Thoron in the Human Environment, World Scientiﬁc, 1998.

496

K. Yoshioka, T. Iida

[4] T. Asai, in: Science of Atmospheric Convection, Tokyodo Press, 1983, pp. 160–185. [5] J. Kondo, in: Science of Atmospheric Boundary Layer, Tokyodo Press, 1982, pp. 125–153. [6] J.C. Kaimal, in: The Atmospheric Boundary Layer – Its Structure and Measurement, Gihodo Press, 1993, pp. 1– 22. [7] T. Asai, in: Local Meteorology, University of Tokyo Press, 1996, pp. 11–37. [8] F. Pasquill, F.B. Smith, in: Atmospheric Diffusion, Ellis Horwood, 1983, pp. 26–41. [9] T. Iida, Y. Ikebe, et al., Continuous measurements of outdoor radon concentrations at various locations in East Asia, Environ. Int. 22 (1) (1996) s139–s147.

497

A new method for supplying low radon air by using a hollow ﬁber module T. Iida, T. Kato Department of Nuclear Engineering, Graduate School of Engineering, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8603, Japan

The aim of this study is to develop a new technique for supplying air with low radon concentration by applying the method of membrane permeation using a hollow ﬁber module. This method is applicable to reduce high radon levels in indoor air. Two hollow ﬁber set-ups were tested: a MERASILOX-S module made of silicon rubber membrane and a MHF module made of thin segment polyurethane membrane. If radon containing air is forced through these ﬁlters, output air may have an up to 40% reduced radon concentration. These results suggest that it is possible, with the application of this technology, to substantially reduce radon concentrations indoors.

1. Introduction On average the effective dose, due to radon and its decay products, in dwellings accounts for about half of the dose due to all natural radiation sources [1]. Radon concentration has a tendency to be elevated in dwellings or underground space with low ventilation. Furthermore, radon concentration is the main cause of increased background for most radiation measurements. Know-how on the reduction of radon and its decay product concentrations in indoor air is therefore essential not only for dwelling inhabitants but also for low background radiation measurements and even for clean rooms for the production of semiconductor devices and research for molecular biology. The most common method for reducing the effective dose due to radon and its progeny is to remove radon and decay product concentrations using increased ventilation rates [2]. For airborne radon progeny, Jonassen [3] had already studied the effects of ﬁltration and electric ﬁelds. Since radon is chemically inactive, the charcoal trap method has been the only method for indoor radon adsorption so far. However, this method is affected by air humidity. In the present study, a membrane permeation method has been tested for radon removal from air [4]. In principle, air containing radon is forced through the membrane. During the RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07059-7

© 2005 Elsevier Ltd. All rights reserved.

498

T. Iida, T. Kato

process radon permeates through the membrane thus resulting in a lower radon concentration in output air. Similar methods have already been applied to the production of oxygen and hydrogen, and separation of krypton and xenon from nuclear reactor atmospheres [5]. The characteristics of the membrane permeation method for radon have been investigated and the validity for radon removal has been considered.

2. Experimental method 2.1. Hollow ﬁber module When there exists a radon concentration gradient in a membrane, radon permeates along the gradient. The level of permeability depends mainly on the membrane materials. The criteria for selecting the membrane materials is as follows: (i) high permeability, (ii) low membrane thickness, and (iii) acceptable resistance to pressure. There are few measurements of the permeability of some membrane materials for radon [6]. For the radon removal experiments, a MERASILOX-S module (HSO-80: Senko Ika Kogyo Co., Ltd.) and a MHF module (MHF0504: Mitsubishi Rayon Co., Ltd.) were used. The membrane material of MERASILOX-S module is made of silicon rubber. It is well known that silicone rubber membranes present high permeabilities. Moreover, they have a unique nature of increasing permeability with gas molecular weight [7]. This MERASILOX-S module is used as an artiﬁcial lung that is composed of 9000 ﬁne hollow ﬁbers made of silicon rubber. Each hollow ﬁber has a 200 μm inner diameter and is 100 μm thick. This module of silicon rubber hollow ﬁbers has an effective surface membrane area of 0.8 m2 , and is set in a cylindrical polycarbonate housing of 270 mm in length and 45 mm in diameter. The MHF0504 hollow ﬁber module is made of segment polyurethane and is used for the degassing of oxygen. Each hollow ﬁber has 200 μm inner diameter and is 25 μm thick. The segment polyurethane membrane is sandwiched within microporous membrane made of polyoleﬁn. The three-ply membrane has therefore good resistance to pressure. This module of ﬁne hollow ﬁbers has an effective surface membrane area of 0.6 m2 , and is set in a cylindrical polycarbonate housing of 200 mm in length and 48 mm in diameter. 2.2. Permeability and permeance Let C1 and C2 be the radon concentrations in Bq m−3 on the left and right side of the membrane, respectively (Fig. 1). A radon ﬂux through a thin membrane could be given by Fick’s ﬁrst law as follows: J =D

C1 − C 2 dC = DT , dx h

(1)

where J is radon ﬂux in Bq m−2 s−1 , D is diffusion coefﬁcient of radon in membrane in m2 s−1 , h is membrane thickness in m, and T is solubility coefﬁcient. The gaseous pressure,

A new method for supplying low radon air by using a hollow ﬁber module

499

Fig. 1. Diffusion of gas through membrane.

radon content per standard volume, and the permeability are given by p in atm (= 1.013 × 105 Pa), C¯ in Bq m−3 (STP), and P = DT in m3 (STP) m m−2 s−1 Pa−1 , respectively. Then, the radon ﬂux is expressed by p1 C¯ 1 − p1 C¯ 2 . (2) h If C1 is the high radon concentration in the radon permeation measuring system and C2 is very small compared to C1 , the time variation of radon concentration in the system is given by J =P

JS dC1 =− − λC1 , dt V

(3)

where S is the surface area of hollow ﬁber in m2 , V is the total volume of the radon permeation measuring system in m3 , and λ is radon decay constant in s−1 . Deﬁning the radon exchange rate λR in the system as p1 SP , Vh the following solution is obtained from equation (3): C1 = C10 exp −(λR + λ)t , λR =

(4)

(5)

where C10 is initial radon concentration in the system in Bq m−3 . The concentration C1 decreases exponentially with elapsed time. Since the concentrations C1 and C10 are being measured, the radon exchange rate λR could be calculated from equation (5). Thus, the permeability P could be determined. The permeability is the main characteristic of the membrane material. The radon ﬂux depends on not only the permeability of the membrane but also on the thickness. Therefore, we deﬁne the permeance R in m3 (STP) m−2 s−1 Pa−1 as follows: R=

P . h

(6)

500

T. Iida, T. Kato

The permeance could express the permeation performance of the material. In this study, permeability and permeance were evaluated and compared under deﬁned experimental conditions. 2.3. Experimental system and procedure Figure 2 shows the schematic diagram of the set-up for radon permeation measurement with hollow ﬁber modules. The system consists of the hollow ﬁber module, ﬂow meters, a 247.8-L vessel, a diaphragm pump, and a 0.9-L buffer vessel, which are connected forming the “bottom” and the “top” air circuit. At the “bottom” circuit (L1), air containing high radon concentration is created using a radium solution. This air is then forced through the circumference of the hollow ﬁber module and subsequently introduced to the 247.8-L vessel. In the “top” circuit (L2), outdoor air with low radon concentration air is circulated through the core of the hollow ﬁbers module. The radon concentration in the “bottom” circuit air decreases exponentially since radon from this air permeates from the circumference to the core of the hollow ﬁber membrane, in which the outdoor air of the “top” air circuit runs. The radon concentration of the “bottom” circuit air is measured with an electrostatic radon monitor [8] set-up in the 247.8-L vessel. This radon monitor measures only the alpha counts from 218 Po with a half-life of 3.05 min. Such a radon monitor has good response for the time variation of radon concentration in the system. The ﬁrst experiment was carried out under atmospheric pressure. The ﬂow-rates in both “bottom” and “top” air circuits were changed from 0.5 to 6.0 L min−1 at intervals of 0.5 L min−1 . Flow directions of the “bottom” radon containing air through the circumference of the module and the “top” outdoor air through the core of the module were opposite to each other. The radon concentration changes in the “bottom” air circuit were followed for 120 min. Based on the concentration measurement results over these 2 h, the permeability of the hollow ﬁber module was determined using equations (4) and (5). Since the total radon ﬂux through a membrane depends on the apparent difference of radon concentrations of the two radon exchanging circuits and this difference increases with pres-

Fig. 2. Schematic diagram of the system for radon permeation measurement with hollow ﬁber module.

A new method for supplying low radon air by using a hollow ﬁber module

501

Fig. 3. Experimental arrangement for applying pressure on the inside of hollow ﬁbers.

sure, a second experiment was carried out under pressure at constant ﬂow. Both “bottom” circuit radon containing air and “top” circuit outdoor air were equally pressurized. Experimental pressure was maintained constant by the use of regulating gauges. The gauges were placed between the 0.9-L buffer vessel and the hollow ﬁber module for the “bottom” air circuit and between the pump and the hollow ﬁber module for the “top” air circuit in Fig. 2. The air pressure was reduced to atmospheric pressure at the ﬂow measuring points using suitable pressure controllers. Based on the concentration measurement results over 2 h, the permeability of the hollow ﬁber module was again determined using equations (4) and (5). In order to test whether a hollow ﬁber module is capable of supplying air containing lower radon concentrations in case that indoor air is forced through the module core the experimental set-up was modiﬁed to operate like a “radon ﬁltering” device as per Fig. 3. Radon containing air collected from the 247.8-L vessel at atmospheric pressure is pressurized via a pump and run through the core of the module. The air depressurizes at the exit of the module down to atmospheric pressure. Its radon concentration is subsequently measured by another radon monitor noted as the detector. Then this air – still in atmospheric pressure – returns back to the 247.8-L vessel. The 247.8-L vessel, being itself also at atmospheric pressure, is connected to the circumference inlet and outlet of the module through a second closed circuit and another in-line pump. Therefore, it is expected in this set-up that the radon concentration in the 247.8-L vessel will be decreased only by physical radon decay constant. The dilution ratio obtained using this module set-up is calculated from the ratio of the radon concentrations measured with the “detector” and the radon monitor within the 247.8-L vessel.

3. Results and discussion 3.1. Atmospheric pressure experiment at different ﬂow rates Table 1 shows the measured values of permeability and permeance for radon, obtained with ﬂow rates of 6.0 L min−1 for L1 and 3.5 L min−1 for L2. The permeability was almost constant with ﬂow rates of more than 4.0 L min−1 for L1 and more than 2.5 L min−1 for L2. The silicone rubber membrane of the MERASILOX-S presents higher permeability for radon gas

502

T. Iida, T. Kato Table 1 Permeability and permeance of radon for MERASILOX-S and MHF Module type

Permeability (m3 (STP) m m−2 s−1 Pa−1 )

Permeance (m3 (STP) m m−2 s−1 Pa−1 )

MERASILOX-S MHF

(2.06 ± 0.13) × 10−14 (6.35 ± 0.55) × 10−15

(2.06 ± 0.13) × 10−10 (2.54 ± 0.22) × 10−10

than the polyurethane membrane of the MHF. However, the thickness of the polyurethane membrane of MHF is only 250 μm. Consequently the MHF hollow ﬁber module presents higher permeance. 3.2. Higher pressure experiment at equal ﬂow rates The second experiment has been carried out using only the MHF module, the reason being that the MERASILOX-S silicon rubber hollow ﬁbers do not effectively resist pressure. If the same pressure is applied on both sides of the membrane (circumference and core) of the MHF module, the apparent radon concentration difference between the two radon exchanging circuits increases. It is obvious that if pressure is applied only on one side of the hollow ﬁber module (in this case either only in the module circumference circuit or only in the module core circuit) the whole system will reduce itself to a ventilating device. Since permeability depends on the pressure differences, it is inadequate for the comparison of radon permeance rate in equally pressurized circuits. Therefore, for this experiment, the permeation rate is deﬁned as follows: Q = pP ,

(7) m3 m m−2 s−1 ,

where Q is permeation rate in p is the applied pressure in Pa and P is the apparent permeability in m3 (STP) m m−2 s−1 Pa−1 . So, in this case, the permeation rate corresponds to the apparent permeability. The experiment was carried out under the following pressures: 0.5 × 105 , 1.0 × 105 , and 1.5 × 105 Pa and also at atmospheric conditions. Table 2 presents the values of permeation Table 2 Results of high pressure experiments at equal ﬂow rates Run

Pressure (×105 Pa)

Flow rate∗ (L min−1 )

Flow rate† (L min−1 )

Permeation rate (×10−10 m3 m m−2 s−1 )

Permeability (×10−15 m3 (STP) m m−2 s−1 Pa−1 )

1 2 3 4 5 6

+0.5 +1.0 +1.5 0 0 0

3.0 2.75 2.0 3.0 2.0 6.0

2.0 2.75 2.0 2.5 2.5 3.5

5.63 ± 1.08 8.59 ± 1.12 9.42 ± 1.13 3.41 ± 0.41 4.33 ± 0.41 6.43 ± 0.56

3.75 ± 0.72 4.30 ± 0.56 3.77 ± 0.45 4.27 ± 0.41 3.37 ± 0.40 6.35 ± 0.55

∗ Apparent ﬂow rate of high 222 Rn concentration air ﬂowing in module circumference. † Apparent ﬂow rate of low 222 Rn concentration air ﬂowing in module core.

A new method for supplying low radon air by using a hollow ﬁber module

503

rate and permeability obtained. To compare the permeation rates reported under the different pressures applied, the ﬂow rates inside the module were kept approximately equal and are given in Table 2 in the form of apparent ﬂow rate under atmospheric pressure as measured in the experimental set-up. The results indicate that the permeation rate increases with applied pressure. The values of permeability obtained in runs 1, 2 and 4 or runs 3 and 5 agree well. The permeability in run 6 represents the value obtained in the saturated condition. The comparison of the permeation rates of runs 3 and 6 indicates that the permeation rate may increase up to 45% if the circuit pressure is equal to 1.5 × 105 Pa. 3.3. Radon ﬁltering experiment The experiment was carried out with pressurized air with elevated radon concentration. The experimental parameters were radon concentration, applied pressure and ﬂow rate. The ﬂow rates through the module are given by apparent ﬂow rate under atmospheric pressure similarly to the second experiment. The obtained results are shown in Table 3. The dilution ratios of 222 Rn concentration in Table 3 decrease with increasing pressure, and increase with ﬂow rate. The results obtained at a pressure of 2.0 × 105 Pa and a ﬂow rate of 0.6 L min−1 indicate that the radon concentration could be reduced up to about 40% by passing once through the hollow ﬁber module. This means that about 60% of radon could be removed. The dilution ratios were also calculated by a model of radon permeation in a membrane. Following this model the hollow ﬁber was divided to n cells, and the permeation rate in each cell was calculated by using the value of radon permeability in Table 1. As shown in Table 3, the experimental values agree well with the calculated ratios. The ﬂow meters installed in the experimental set-up also measured the total ﬂow in integrating mode, thus permitting the calculation of the permeating air quantity from the module core towards the module circumference. Total ﬂows were measured at a pressure of 2.0 × 105 Pa Table 3 Experimental results under applied pressure on only the inside of hollow ﬁbers Pressure (×105 Pa)

Flow rate (L min−1 )

Initial 222 Rn concentration (Bq m−3 )

Final 222 Rn concentration (Bq m−3 )

Ratio of 222 Rn concentration

Calculated ratio

+0.5 +1.0 +1.5 +2.0 +2.0 +2.0 +2.0 +0.5 +1.0 +1.0 +1.5 +1.5 +2.0 +2.0

0.6 0.6 0.6 0.6 0.6 0.6 0.6 2.0 2.0 2.0 2.0 2.0 2.0 2.0

6.25 × 103 6.62 × 103 4.91 × 103 1.40 × 104 1.08 × 104 34.6 10.2 8.10 × 103 7.21 × 103 6.76 × 103 9.36 × 103 4.94 × 103 1.38 × 104 1.14 × 104

5.20 × 103 6.62 × 103 2.52 × 103 5.75 × 103 4.33 × 103 14.5 3.54 7.15 × 103 5.59 × 103 5.30 × 103 6.55 × 103 3.70 × 103 8.49 × 103 6.76 × 103

0.82 0.62 0.49 0.39 0.38 0.40 0.33 0.87 0.76 0.77 0.68 0.73 0.60 0.57

0.79 0.62 0.50 0.41 0.41 0.41 0.41 0.90 0.82 0.82 0.74 0.82 0.66 0.66

504

T. Iida, T. Kato

and a ﬂow rate of 0.6 L min−1 after 20 h of operation. Under these conditions, the quantity of air that entered the module core was 1163.14 L and the quantity of air permeating towards the module circumference was 695.85 L. Therefore, about 40% of the air permeates through the membrane tested. 3.4. Estimation of radon removal The radon concentration in an airtight room made of concrete, which has a volume of 50 m3 and a 50 m2 surface area, is saturated at about 1000 Bq m−3 , assuming that the radon exhalation rate from the concrete surfaces is 2.3 × 10−3 Bq m−2 s−1 . Using an MHF module with an effective surface membrane area of 34 m2 , set in a cylindrical polycarbonate housing of 774 mm in length and 165 mm in diameter, such a radon concentration could be reduced to about 1/10 if air is passing through the module core at atmospheric pressure. When the high pressure of 6.0 × 105 Pa is applied both at the core side and the circumference side of the hollow ﬁbers with a ﬂow rate of about 300 L min−1 , the radon concentration in this airtight concrete room could be reduced below about 1/50, thus reaching 20 Bq m−3 after 20 h.

4. Conclusion Radon passing through hollow ﬁber membranes permeates along the concentration gradient. The applicability of the membrane permeation method has been investigated for radon removal from air. The experiments have been carried out for the MERASILOX-S hollow ﬁber module and the MHF hollow ﬁber module. The values of apparent permeability for MERASILOX-S module membrane and MHF module membrane were (2.06 ± 0.13) × 10−14 and (6.35 ± 0.55) × 10−15 m3 (STP) m m−2 s−1 Pa−1 , respectively, for a ﬂow at atmospheric pressure. The application of the same, higher than atmospheric, pressure on air ﬂowing at both sides of such a membrane, substantially increases the permeation rate of radon. Forcing pressurized radon containing air through the core of the hollow ﬁber cylinder of the experimental set-up, the radon concentration in this air could be reduced to about 40% at a pressure of 2.0 × 105 Pa and a ﬂow rate of 0.6 L min−1 . The dilution ratios obtained by these experiments agree well with the ratios calculated by modeling radon permeation through a membrane. Following this evaluation of the MHF module radon removal capacity from air, it is estimated that the radon concentration in a 50 m3 room, as elevated as 1000 Bq m−3 , could be reduced below about 20 Bq m−3 using a hollow ﬁber module of moderate dimensions. Further investigation is necessary concerning: (i) testing for other types of hollow ﬁbers, (ii) experimenting with higher pressure, and (iii) practical usage of the ﬁndings.

References [1] UNSCEAR, Sources and Effects of Ionizing Radiation, United Nations, New York, 2000. [2] F.T. Najaﬁ, Health Phys. 75 (1998) 514.

A new method for supplying low radon air by using a hollow ﬁber module

505

[3] N. Jonassen, Radon and its decay products in indoor air, in: ACS Symp. Ser., vol. 331, American Chemical Society, Washington, DC, 1987, p. 264. [4] T. Iida, T. Kato, H. Mocizuki, Proceedings of a Workshop on Radon in the Living Environment, Athens, Greece, April 19–23, 1999, Sci. Total Environ. 272 (1–3) (2001). [5] S.A. Stern, S.M. Leone, AIChE J. 26 (1980) 881. [6] G. Jha, M. Raghavayya, N. Padmanabhan, Health Phys. 42 (1982) 723. [7] O. Barrer, O. Chio, JPS C 10 (1965) 111. [8] S. Tasaka, Y. Sasaki, H. Okazawa, M. Nakagawa, Radioisotopes 43 (1994) 125 (in Japanese).

506

Dose evaluation of indoor thoron progeny in some areas in China Q. Guo a,1 , J. Cheng b , Y. Chen b a Department of Technical Physics, School of Physics, Peking University, Beijing 100871, China b Department of Engineering Physics, Tsinghua University, Beijing 100084, China

In order to evaluate the risk from thoron progeny indoors in China, measurements of thoron progeny have been performed by 24 h integrating measurement in Beijing (13 dwellings) and Zhuhai (54 dwellings) areas, whereas in Guangdong (220 dwellings) and Fujian (204 dwellings) provinces 10 min ﬁlter sampling was adopted. The highest average EECTn was 4.0 ± 2.3 Bq m−3 , which was found in Zhuhai city, the thorium-rich area, some 3 times higher than that of Beijing city. The range of mean EECTn /EECRn ratio was 0.07–0.10 in all the areas, the dose ratio of thoron progeny to radon progeny being 0.31–0.47 accordingly. It is suggested that attention should be paid to the exposure from the inhalation of thoron progeny in thorium-rich areas in China. Elevated indoor radon progeny levels were also found in Zhuhai areas.

1. Introduction Compared with radon (222 Rn) and its progeny, thoron (220 Rn) and its progeny have not been well studied in the natural radiation ﬁelds, and our knowledge on thoron and its progeny in the environment around us is very limited. In recent years, however, high contributions of thoron and its progeny to the total exposure of radon, thoron and their progeny have been reported in some areas in the world [1–3]. It was reported in China that the national average of thorium (232 Th) concentration in soil is 49.1 ± 27.6 Bq kg−1 (area weighted) according to the nationwide survey of environmental radioactivity level which was carried out from 1983 to 1990 [5], which is 1.6 times higher than the world average 30 Bq kg−1 [4]. The provinces or regions with average soil 232 Th concentrations higher than 50 Bq kg−1 are shown in Fig. 1. Thorium concentrations in soil in the South are generally higher than those in the North of China just as Fig. 1 shows. In some provinces of the heavily populated South such as Fujian, 1 This work is funded by National Natural Science Foundation of China (No. 10175007).

RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07060-3

© 2005 Elsevier Ltd. All rights reserved.

Dose evaluation of indoor thoron progeny in some areas in China

507

Fig. 1. The geographical distribution of Th-232 contents in soil in China.

Guangdong and Guangxi, the 232 Th concentrations in soil are signiﬁcantly higher than the national average value. The main building material in both urban and rural areas in China is brick. Most kinds of bricks are baked from soil. Adobe, the sun-dried mud brick, is still widely used as the building material in some rural areas. In some provinces such as Gansu and Shanxi provinces in northwestern China, about 30% of the population lives in cave dwellings with soil or mud walls. It is speculated that the elevated natural thorium concentrations, the building materials widely used with high 232 Th concentrations, and the special building structure, like cave-dwellings in some areas, should be essential factors affecting the levels of thoron and its progeny in dwellings in China. The purpose of our study is the estimation of thoron exposure to the public in China in different kinds of dwellings, and in different areas. For a more precise assessment of the exposure from thoron, direct measurements of thoron progeny are desirable because of the different spatial distribution pattern between thoron gas and its decay products [6]. Even though there are several reports on thoron gas measurements in China, especially in cave-dwellings, this paper only summarizes the preliminary results of local surveys on thoron/radon progeny which were carried out in recent years in China. The annual effective dose due to thoron progeny is also evaluated. 2. Materials and methods 2.1. Integrating measurement of radon/thoron progeny An instrument for measuring the equilibrium equivalent of 222 Rn and 220 Rn concentrations with allyl diglycol carbonate (CR-39) plastic detectors [7] was adopted during our surveys in

508

Q. Guo, et al.

both the Beijing and Zhuhai areas. The monitor gives the average equilibrium equivalent of 222 Rn and 220 Rn concentrations (EEC Rn and EECTn ) during sampling intervals. The detection efﬁciencies of the alpha particles were calculated by the Monte Carlo method. The lower limits of detection for EECRn and EECTn are estimated to be 0.57 and 0.07 Bq m−3 for 24 h continuously sampling at a ﬂow rate of 0.8 L min−1 . The instrument was improved on and manufactured in the author’s lab and was adopted for the local survey in Beijing and Zhuhai areas. 2.2. Short time sampling and measurement of radon/thoron The ﬁlter method was adopted for sampling in the survey of Guangdong [8] and Fujian [9] provinces, the sampling time being 10 min in all the surveys. Determination of 218 Po, 214 Bi, 214 Pb and 212 Pb was carried out by total alpha particle counting. The ﬂow rate being 30 L min−1 during the measurement in Guangdong province, the collection efﬁciency of the membrane ﬁlter was 0.7 including self-absorption; the detection efﬁciency was 32.5% using a uranium standard source. While in Fujian province, airborne radon and thoron progeny were collected on glass ﬁber ﬁlters at 60 L min−1 , and a piece of 241 Am standard source was used for the calibration of alpha particle detection. 2.3. Protocols of local surveys on radon/thoron progeny Beijing city and Zhuhai city, located in the North and South of China, respectively, were chosen as survey areas for the comparison of 222 Rn/220 Rn progeny levels in the areas in which 232 Th concentrations are quite different. The monitor which could give the average EEC and Rn EECTn during sampling intervals (24 h with ﬂow rate of 0.8–1.0 L min−1 ) was adopted in the surveys of the two cities. Guangdong and Fujian are the two provinces with both large population and high 232 Th concentrations in soil located in the Southeast of China. Their geological structure is mainly formed by Mesozoic igneous rocks, and granite is widely distributed. The ﬁlter method for short time measurement was adopted both in the biggest city Fuzhou of Fujian province, as well as in the biggest city Guangzhou and the Yangjiang area of Guangdong province. Sampling was principally conducted 1.5 m height above the ﬂoor/ground surface. The outline and information in detail of the local surveys in the areas mentioned above is shown in Table 1.

3. Results and discussion Measurements have been performed in 13 dwellings, including 9 apartment building rooms and 4 single ﬂat houses around the authors’ university in Beijing city for one year, at a frequency of twice monthly. The measurements are still going on. The building materials of both the apartment building and ﬂat houses were red-brick. The walls of the apartments were well painted or decorated, in ﬂat houses, however, they had a ﬂoor slab or brick with almost no decorations.

Dose evaluation of indoor thoron progeny in some areas in China

509

Table 1 The outline and information of the local surveys reviewed in this paper Number of dwellings

Soil 232 Th contents (area weighted) (Bq kg−1 )

Areas

Measuring period

Beijing

Apr. 2001–Mar. 2002

13

34.1 (17.0–63.0)

Zhuhai Fujian province

Mar.–Apr. 2001 Jun. 1993–Aug. 1994

54 204

159.9 (21.7–263.0) 96.3 (19.5–260.1)

Guangdong province

July 1984–July 1986

220

57.2 (1.0–152.7)

Sampling & measuring 24 h sampling ﬂow rate: 0.8–1.0 L min−1 CR-39 detector 10 min sampling ﬂow rate: 60 L min−1 10 min sampling ﬂow rate: 30 L min−1

Table 2 Measurement results of radon/thoron progeny in Beijing and Zhuhai areas Area and type of dwelling

Number of dwellings

EECRn (Bq m−3 ) (arith. mean ± SD)

EECTn (Bq m−3 ) (arith. mean ± SD)

EECTn /EECRn (mean)

Beijing area Apart. building Flat house Total

9 4 13

10.4 ± 5.9 16.3 ± 8.9 12.8 ± 7.2

0.7 ± 0.5 1.5 ± 0.7 1.0 ± 0.7

0.07 0.09 0.07

Zhuhai area Sealed room Ventilated room Total

31 23 54

72.2 ± 36.9 25.0 ± 31.3 52.9 ± 39.1

4.5 ± 1.6 3.2 ± 3.0 4.0 ± 2.3

0.06 0.13 0.08

In Zhuhai city, 54 dwellings were measured during a 2 weeks survey in the spring of 2001. According to the ventilation conditions, 54 dwellings were classiﬁed into 31 sealed rooms and 23 ventilated rooms. The measurement results for Beijing and Zhuhai areas are shown in Table 2. It is seen that the levels of both EECRn and EECTn in Zhuhai, the thorium, radium-rich area, were 52.9 ± 39.1 and 4.0 ± 2.3 Bq m−3 , respectively, some 3 times higher than in the Beijing area, namely 12.8 ± 7.2 and 1.0 ± 0.7 Bq m−3 for EECRn and EECTn , respectively. It is necessary to carry out an investigation in more detail in the thorium-rich areas in the South of China in future according to the results. It was quite like radon progeny that EECTn was higher in ﬂat house than in the apartment dwellings, and the EECTn of sealed rooms was higher than that of ventilated rooms. It is worth mentioning that in the Zhuhai area when comparing the levels between sealed rooms and ventilated rooms, for EECRn it was 3 times higher in sealed rooms than in ventilated rooms, but for EECTn , it was not so different. It was suggested that the inﬂuence of ventilation on indoor levels of thoron and its progeny was quite minor relative to that on radon and its progeny. Table 3 presents the results of short-term measurements of indoor EECRn and EECTn carried out in Guangdong and Fujian provinces, both thorium-rich areas located in the Southeast of China. There were 220 dwellings measured in Guangdong province, with average EECRn

Q. Guo, et al.

510

Table 3 Short-term measurements of EECRn and EECTn in Guangdong and Fujian provinces Area

Number of dwellings

EECRn (Bq m−3 ) (arith. mean ± SD)

EECTn (Bq m−3 ) (arith. mean ± SD)

EECTn /EECRn (mean)

Guangdong province Fujian province

220 204

10.6 ± 3.8 12.9 ± 5.9

1.1 ± 0.8 0.9 ± 0.5

0.10 0.07

Table 4 Estimated annual effective dose (E) in each area Areas

Rn progeny EECRn

Beijing Zhuhai Guangdong Fujian

(Bq m−3 )

12.8 52.9 10.6 12.9

Tn progeny E (mSv)

EECTn

0.77 3.20 0.64 0.78

1.0 4.0 1.1 0.9

(Bq m−3 )

E (mSv) 0.27 1.08 0.30 0.24

Dose ratio of Tn progeny/Rn progeny 0.35 0.34 0.47 0.31

and EECTn 10.6 ± 3.8 and 1.1 ± 0.8 Bq m−3 , respectively. The results for Fujian province were 12.9 ± 5.9 and 0.87 ± 0.5 Bq m−3 for average EECRn and EECTn , respectively. No differences of both indoor EECRn and EECTn levels between Beijing city and Guangdong or Fujian province were found in comparing Tables 2 and 4. The limited number of dwellings in Beijing area and the different methods adopted for the surveys should be considered as interfering factors, and further investigations are necessary. All the mean indoor EECTn /EECRn ratios shown in both Tables 2 and 3 were from 0.07 to 0.10, which is higher than the range (0.02–0.04) report by UNSCEAR in 2000 [4].

4. Dose evaluation A brief calculation of exposure from the inhalation of radon/thoron progeny was applied as follows, according to the preliminary results of the surveys above: ERn = EECRn × fRn × T , ETn = EECTn × fTn × T , where the conversion factors fRn and fTn were adopted as 9 and 40 [nSv/(Bq h m−3 )] for radon and thoron, respectively [4]; The exposure time T is 6720 h for an occupancy factor indoors assumed to be 0.8. The results are shown in Table 4. The exposure from the inhalation of Rn/Tn progeny was evaluated to be 1.04–4.28 mSv in the areas above, and the dose ratios of Tn progeny to Rn progeny were around 1/3, which were signiﬁcantly higher than the value given by the UNSCEAR 2000 report [4].

Dose evaluation of indoor thoron progeny in some areas in China

511

5. Conclusions To obtain an outline of the exposure from the inhalation of thoron progeny in China, this paper reviews 4 primary local surveys carried out in different areas by both 24 h integrating sampling and 10 min short term sampling. High levels of EECRn and EECTn , of 52.9 ± 39.1 and 4.0 ± 2.31 Bq m−3 , respectively, were found in 54 dwellings in the survey in Zhuhai city, the radium/thorium-rich area, by 24 h integrated measurements. This was 3 times higher than that of Beijing city. It is suggested that a detailed investigation of the levels of thoron progeny in thorium-rich areas is necessary. The range of mean EECTn /EECRn ratio was 0.07–0.10 in all the areas examined, the dose ratio of thoron progeny to radon progeny being 0.31–0.47. From these limited data, the levels of thoron progeny appear to be somewhat higher than the typical value in the UNSCEAR 2000 report. This suggests that attention should be paid to the exposure from the inhalation of thoron progeny in some thorium-rich areas in China. Further work has to be done on the study of thoron and its progeny, where integrated measurements of thoron, especially of thoron progeny, are necessary for a precise assessment of radiation exposure.

References [1] [2] [3] [4] [5]

[6] [7] [8] [9]

G. Sciocchetti, et al., Radiat. Prot. Dosim. 45 (1992) 509–514. Q. Guo, T. Iida, K. Okamoto, J. Nucl. Sci. Technol. 32 (8) (1995) 794–803. L. Mjones, R. Falk, H. Mellander, et al., Radiat. Prot. Dosim. 45 (1/4) (1992) 349–352. Sources and Effects of Ionizing Radiation, UNSCEAR 2000 Report to the General Assembly, with Scientiﬁc Annexes, United Nations, New York, 2000. The Writing Group for the Summary Report on Nationwide Survey of Environmental Radioactivity Level in China, Investigation of natural radionuclide contents in soil in China, Radiat. Prot. 12 (2) (1992) 122 (in Chinese). Q. Guo, J. Sun, W. Zhuo, J. Nucl. Sci. Technol. 32 (8) (2000) 716–719. W. Zhuo, T. Iida, Health Phys. 77 (5) (1999) 584–587. Z. Wu, J. Zeng, Radiat. Protect. 9 (1989) 454–459 (in Chinese). W. Zhuo, T. Iida, X. Yang, Radiat. Prot. Dosim. 87 (2) (2000) 137–140.

512

EU-Concerted Action for a survey on radon exhalation rate measurements for building materials and soils L. Roelofs a , R. Wiegers b , K.-H. Puch c , G. Keller d a NRG Arnhem P.O. Box 9035, 6800 ET Arnhem, The Netherlands b IBR Consult BV, De Giesel 12-14, 6081 PH Haelen, The Netherlands c VGB PowerTech e.V., Klinkestraße 27-31, 45136 Essen, Germany d Universität des Saarlandes, Institut für Biophysik, Universitätsklinik, Geb. 76, 66421 Homburg, Germany

Measurements in dwellings in different EU member states showed a large variation of the radon concentrations. In some EU member states governments already use action levels in order to control high radon levels. The main sources of radon have been identiﬁed to be the soil and the building materials. A reliable and cost-effective radon exhalation rate measuring method is thus of great importance especially for the building materials industry. Different measurement methods of radon exhalation from soils and building materials have been recently developed in several EU member states. These methods are (i) based on different measuring principles and, (ii) partially still under development. As a consequence, both the accuracy of the results and their comparability are limited. In order to increase the necessary reliability as well as the comparability, the main methods and their backgrounds were inventoried and discussed.

1. Introduction In the last decades the quality of the indoor environment has become an important issue for human health. One of the indoor risks to human health is the exposure to radiation of naturally occurring nuclides. The radiation dose from inhaled radon (222 Rn) and radon decay products and to a lesser extend from thoron (220 Rn) and its decay products is in general the main component of natural radiation exposures of the human population. Measurements in dwellings in many EU member states showed a large variation of radon concentration indoors. In some EU member states governments already use action levels in order to control high radon concentrations indoors. Mitigation methods for reducing radon concentrations should be based upon the identiﬁcation of the sources responsible for the enhanced RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07061-5

© 2005 Elsevier Ltd. All rights reserved.

EU-Concerted Action for a survey on radon exhalation rate measurements

513

levels. The main sources of indoor radon have been identiﬁed to be the soil and the building materials. Radon produced in the soil can reach indoor space by pores, cracks and small holes in the building materials or in the construction structure. Since raw materials for building materials contain a certain level of natural radioactivity, including 226 Ra, building materials also produce radon, which contributes to the indoor radon concentration. This contribution is sometimes increased due to the use of regional soil-based raw materials for the building materials with a relatively high content of 226 Ra. The contribution of both sources (soil and building materials) to indoor radon concentrations can be quantiﬁed by measuring the radon exhalation rates. In most European member states, (national) institutes have measured radon exhalation rates of both building materials and soils. The results of these measurements can vary greatly for reasons such as type of material, test set-up and measuring conditions.

2. Scope and approach The main objective of this EU Concerted Action was to bring together researchers and building industries in order to create an inventory of the different measuring methods recently developed and employed in the EU, for radon exhalation rates of both building materials and soils. Furthermore, given the fact that these methods are based on different measuring principles and as a consequence, both the accuracy of the results and their comparability are limited, the Concerted Action aimed at increasing the necessary reliability as well as the comparability, by discussing all the main methods. In order to gain a good overview of the existing information on the measuring methods of radon exhalation for building materials and soils as well as on the demands of users, data were gathered by means of questionnaires and were subsequently evaluated. Based on this data collection, conclusions and recommendations for the development of a European standard for radon exhalation measurements were discussed, within a selected group of experts in a closed workshop.

3. Questionnaires and feedback Two questionnaire target groups were initially identiﬁed: the radon exhalation researchers and the materials producers. Each target group received their own type of questionnaire. The researchers’ questionnaire covered ﬁelds concerning the employed exhalation measurement techniques as well as experience gained using them. The questionnaire sent to building industries aimed to reveal mainly the demands of material producers for radon measurements. The questionnaire for researchers was sent by mail to ca 90 institutes in 23 different countries. The questionnaire for industries was sent by mail to approximately 70 building industries in 15 countries. The industries involved cement and concrete, aerated concrete, gypsum, calcium silicate brick and ceramic brick producers. A total number of 43 questionnaires were returned by the research institutes. The answers represented 19 different countries amongst which were most EU countries. However, some more uncompleted questionnaires were retrieved from research institutes, which pointed out

514

L. Roelofs et al.

that they do not carry out exhalation measurements (any more) and that no information was now available. Only 10 questionnaires were received back from the industries. Production industries from the following EU countries answered: France (1), Germany (5), Netherlands (2), United Kingdom (1) and Sweden (1). Before going into the analysis of the questionnaire results it must be stated that not every respondent covered all aspects of the question ﬁelds.

4. Results of research institutes’ questionnaires The ﬁrst set of questions covered some general aspects of the used measurement methods. From the results it can be concluded that for each type of application (e.g., soil, bricks, in situ, etc.) several radon exhalation measurement methods are available. Further analyzing into the measurement principles used by the institutes, one can establish that these involve differences in (a) set-up, (b) detectors and (c) procedures. Most respondents use a closed can system. Some of them mix the air in the can. Others use a radon free nitrogen ﬂow in order to create a ventilation rate. Also a system with an open area (open chamber) is frequently used. Most respondents sample air in which radon concentration is measured. The most frequently used absorber is activated charcoal (8 methods). The main detector type used in most methods for actively measuring radon concentration is the Lucas cell. Also gamma detection of radon decay products with a germanium or NaI detector is carried out. Passive monitoring (track etch) is mentioned twice. Considering reproducibility as well as repeatability, it can be seen that (as far as the respondents were able to supply information on this question) the difference between the averages of the reproducibility and repeatability is small. The detection limits vary between 3 × 10−10 and 8 × 10−4 Bq s−1 . It is also reported that in most cases skilled personnel (at least a scientiﬁc technician) was required to carry out the radon exhalation tests. Considering the sample size and sample numbers used for a measurement it could be seen that the smallest samples are 5 × 10 × 20 cm3 for calcium silicate brick, ceramic brick and aerated concrete. The maximum size is 50 × 45 × 10 cm3 . The exhaling surface varies between 11.25 cm2 and 4500 cm2 . It is well known that the sample preparation is of eminent importance for the result. From the replies it can be deduced that (as far as sample storage conditioning is applied) the temperature ranges from 18 to 24 ◦ C. One exception of 105 ◦ C concerns a method for the determination of radon emanation. The relative humidity (RH) varies to a much higher extent. The median is 50% RH. However, 60% RH is also used as default condition. The duration of the conditioning period ranges between 1 hour and 84 days. Another important sample conditioning parameter is to consider when the sample is ready for measurement. Nearly 50% of the respondents use a ﬁxed storage period. Few respondents start the measurement after a storage period in which they have observed a constant mass for the samples. However, 40% of the respondents use no criterion according to which they start the measurement procedure. The covering of the samples by radon tight layers (for example in order to achieve onedimensional radon transport), can largely inﬂuence the results. Therefore, a question on this subject was included. From the response it can be seen that, in most cases, samples are not

EU-Concerted Action for a survey on radon exhalation rate measurements

515

covered with a radon-tight material (i.e. foil). 22% of the respondents (sometimes) cover the samples with a radon-tight material. In order to check the radon tightness of radon tight layers two respondents indicated that they use a control procedure. Approximately 50% of the respondents use a sampling period of less than 1 day. The other 50% use a sampling period of more than 1 day and up to 14 days. The counting time generally does not exceed a period of 1 day. The total measuring time – including preparation – required for one measurement is less than 1 day for the majority of the presented methods. The majority of the described methods do not employ an adsorber. However, charcoal is the most commonly used adsorber. Regarding the type of active method used for radon concentration measurement, the Lucas Cell is the most commonly used detector. If activated charcoal is used as adsorber, gamma radiation is detected by germanium detector in nearly all cases. In Table 1 an overview of the above results is summarized. Table 1 Overview of the type of used detector Type of detector

Number of respondents

Germanium detector (gamma) NaI (gamma) LSC (alpha’s and beta’s) Lucas cell Proportional counter Track etch Other

5 1 2 13 3 3 8

Table 2 Overview of the nuclide, which are actually measured Nuclides measured

Number of methods

Nuclides individually 222 Rn 218 Po 218 Po + 214 Po 214 Bi + 214 Pb 214 Bi + 214 Pb + 214 Po 222 Rn + 218 Po + 214 Bi + 214 Pb + 214 Po

6 1 1 3 1 1

Combination of nuclides 222 Rn + 218 Po 222 Rn + 218 Po + 214 Po 214 Bi + 214 Pb 222 Rn + 218 Po + 214 Bi + 214 Pb + 214 Po 214 Bi + 214 Pb + 214 Po 218 Po + 214 Bi + 214 Pb + 214 Po 218 Po + 214 Po

1 7 2 3 1 1 1

516

L. Roelofs et al. Table 3 Inﬂuencing factors on the results of measurement Factor

Number of respondents

Leakage of radon Back diffusion Non-equilibrium between radon and its progeny Conditions during the measurement No inﬂuence/no response Others

14 8 2 13 4 5

Table 4 Advantage of method Advantage

Number of respondents

Time (quick method) Simple Low costs Low detection limit

5 8 3 3

Another important aspect investigated by this questionnaire was the nuclides by which the trapped radon is measured (Table 2). The detection of radon for the majority of the methods is based on a combination of nuclides. Especially, the combination of the alpha-emitters 222 Rn + 218 Po + 214 Po is often used. Of course, it is very interesting to know what the researchers themselves consider as the main factors which have an inﬂuence on the results of the measurement. More than 80% of the respondents believe that one or more of the factors mentioned in Table 3 inﬂuence the results of the measurements. Leakage of radon and conditions during the measurement were marked by 50% of the respondents. In 25% of the cases back diffusion is seen as an inﬂuencing parameter. The questionnaire results further indicate that most measurements are carried out on building materials. The respondents report radon exhalation results which differ by factors of 10 to 100 for each type of material. Most respondents gave their data in Bq m−2 s−1 or Bq kg−1 s−1 . Therefore, the data of the different respondents are not suitable for comparison. The respondents were also asked to supply information on what they believed to be the main advantages of their own method (Table 4). This question was followed by another on the greatest drawback of their method. The time necessary for exhalation rate determination (including preparation time) was considered to be the major problem in the practical use of most methods. The possibility of the occurrence of leakage was also mentioned. Other inﬂuencing parameters were noted as well (meteorological conditions, wind, etc.) 5. Results of industry questionnaire Industries (as potential users of the conclusions from this Concerted Action) were asked to supply information on their demands for radon exhalation measurements. The industries, as

EU-Concerted Action for a survey on radon exhalation rate measurements

517

far as a full response was returned, represented the following product groups: aerated concrete (2), ceramic brick (1), calcium silicate brick (1), cement (1) and gypsum (5). The ﬁrst question was on their preferences for sample conditioning (as being the experts on both the application and the behaviour of their materials). From this, it could be seen that storage before measurement, with or without drying or soaking, is their main preference for sample preparation in order to ensure a realistic moisture content of the product. There were however some differences in the suggested storage conditions as described by the different product groups. The representatives of the aerated concrete industry suggested 80% RH in the conditioning cabinet. The ceramic brick and calcium silicate industries mentioned 50%. Depending on the product, 95% RH is marked by the respondent of the ceramic brick industry. The gypsum industry shows diversity in opinions on the relative humidity for gaining a relevant water content: 50, 65 and 5%. The next question considered the rest of the storage parameters. From the response it could be derived that the required temperature in most cases is ca 20 ◦ C and the drying time varies between 4 hours and 4 weeks (Table 5). Information was also collected not only on the suggested RH during conditioning but also on the RH during the measurement itself. From the answers it can be seen that the standard conditions during measurement depend on the type of product. For aerated concrete 50% RH and 23 ◦ C are suggested. For other products the suggested RH is 50 or 65% and the suggested temperature is ca 20 ◦ C. Considering aging of the products over a longer period, most respondents expect no aging effect on the exhalation rate of their product. The cement industry indicated that concrete properties change with time, but they did not give a suggestion for the time period after production before starting exhalation measurements. One respondent of the gypsum industry suggested a period of 3 months. From another question it could be concluded that most industries have experience with exhalation measurements, except the aerated concrete industry. The respondent of the cement industry indicates that they are aware of the extensive literature on this subject. Since the radon exhalation rate depends also on the pore structure of the product/material it is important to know what factors (besides time dependent changes) are considered to have a signiﬁcant inﬂuence on the pore structure. Therefore, the industry was asked to indicate these (expected) main factors for their own product group. A diversity of factors which signiﬁcantly inﬂuence the pore structure are mentioned: Table 5 “Could you give the storage conditions which could be applied on your type of product/material?” Type of industry

Drying temperature (◦ C)

Drying time

Aerated concrete Ceramic brick Calcium silicate Gypsum Cement

23, 105 105 20 45, 40, 45–50, 40, 40 No information

28 h, 24 h 24 h 2–4 weeks 48 h, 4 h, 48 h, 24 h, 24 h No information

518

• • • • •

L. Roelofs et al.

water content, density, raw material composition, moulding pressure, method of shaping.

6. Conclusions The variety of methods and the variety in units used (e.g. activity per area per time) lead to the conclusion that each measurement method has its own approach. Furthermore, even within one institute several variants of the same method are used. This makes the interpretation and comparison of the methods difﬁcult if not impossible. So any further conclusions drawn are limited by this difﬁculty. There is a huge number of measuring instruments, methods and labs within the EU available for radon exhalation rate measurements. For every type of sample (bricks, soil, etc.) different measurement methods are available. Most methods can be applied on more than one type of sample. The set-up types can be basically divided into a closed-can and open area method. Within both set-up types several variants have been developed. Regardless of type of sample, application, set-up and detection device, most methods have a detection limit of approx. 10−6 – 10−5 Bq s−1 . There seems to exist no relation between the total measurement period and the detection limit or between the number of nuclides measured and the detection limit, although this could have been expected. The most frequently used quantities for reporting the radon exhalation rate are activity per area per time and activity per mass per time depending on the purpose of the measurement. There exist institutes practicing radon exhalation tests on many types of materials. However, no knowledge is available on subjects such as reproducibility and repeatability of the test methods used. As far as the pre-treatment for building materials is concerned, in some cases no conditioning is applied. The requirements of the industry regarding sample conditioning are in many cases different from the conditions applied in the Laboratories (especially for the relative humidity and the sample size). More than 80% of the respondents believe that one or more factors like leakage of radon, back diffusion, non-equilibrium and conditions during measurement inﬂuence the results. The selected group of experts evaluated the presented data during a closed workshop. The main comments stated are gathered as follows: It was generally felt that it is necessary to use several types of radon exhalation measurements due to the fact that these methods have to meet the requirements of a wide range of both sample types and purposes. In addition the industry signalled their interest in standardisation of the measurement methods. It was seen as a major advantage that the institutes have at least a basic knowledge of the materials they are testing. Assurance in application of a standard sound sample preparation method is needed. Hence, as far as building materials are concerned, it is of great importance to have an intensive and regular exchange of information between the industry involved and the institutes. The aim of standardisation is to obtain a method with

EU-Concerted Action for a survey on radon exhalation rate measurements

519

a sufﬁcient accuracy, repeatability and reproducibility. These aspects are evaluated as being more important than other aspects such as the overall measuring time and costs. There was no agreement on the importance of the overall measuring time. The participants stressed the utmost importance of conducting a round robin test involving several methods (c.q. institutes) and different types of samples. The practical value of measuring radon exhalation rate of standard types of building materials was discussed during several meetings. No consensus on this item could be achieved.

7. Recommendations Factors such as leakage of radon, back diffusion, non-equilibrium and conditions during measurement should be considered during the development of a (European) radon exhalation measurement method. From the point of view of radiation protection the quantity activity per area per time is the most relevant. However, it must also be stated that such quantities as activity per mass per time can be derived for speciﬁc purposes. The fact that samples of the same type of material can easily vary their exhalation rate by a factor ranging from 10 to 1000 implies that the measurement method must be capable of detecting levels ranging by the same amount. It is of great importance to carry out a round robin test on radon exhalation measurements comprising both several test methods and several types of samples. It is necessary to bring industries and institutes together to discuss the optimum sample preparation methods. The conditions during measurement should only represent the typical indoor environment.

520

Radon concentration survey in schools of the Friuli-Venezia Giulia region, North-East Italy C. Giovani a , C. Cappelletto a , M. Garavaglia a , R. Villalta b a ARPA (Regional Environmental Protection Agency) Friuli-Venezia Giulia, Dipartimento di Udine, Via Colugna 42,

33100 Udine, Italy b ARPA (Regional Environmental Protection Agency) Friuli-Venezia Giulia, Direzione Centrale, P.zza Grande 1,

33057 Palmanova (Ud), Italy

In the past, several surveys were made in the Region in order to locate inhabited areas with high radon risk. In the last two years the Regional Environmental Protection Agency (ARPA) of Friuli-Venezia Giulia carried out a survey to determine the radon concentration in the schools of that area. Geological information was collected, both on a regional scale and, where possible, in detail. The purpose of this study is to consider the possibility to use all the data collected to ﬁnd out some radon prone areas and to protect people from radon exposure. Even if the data gained in this survey refer to spot checks made in buildings with different constructional characteristics and with an heterogeneous distribution, nonetheless the results have allowed us to recognize geographically and geologically well deﬁned zones, in which the percentage of buildings that exceed the reference levels is signiﬁcantly higher than the Regional average. The ﬁrst results of this study seem to locate some radon-prone areas in karstic regions or in regions where the cover consists of very permeable gravel deposits.

1. Introduction Since 1989 radon monitoring surveys in buildings and schools have been carried out in the Italian region of Friuli-Venezia Giulia [1,2]. The results of these surveys showed that many buildings (from 3 to 8%) had an average radon concentration higher than 400 Bq m−3 . During the years 2000–2001 a new study to monitor all schools in the provinces of Pordenone and Trieste was carried out. This study involved several measurement surveys with different measurement techniques. The survey was conducted to ﬁnd schools that show high radon concentrations. The aim is to offer practical advice, speciﬁc for each school, concerning remedial action where necessary. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07062-7

© 2005 Elsevier Ltd. All rights reserved.

Radon concentration survey in schools of the Friuli-Venezia Giulia region, North-East Italy

521

In this work, the results related to the analyses of radon concentration, together with the geological information, are presented. The gathered data can be useful to identify radon prone areas. 2. Geological features 2.1. Pordenone province The Province of Pordenone stretches from the Carnian Prealps to the Friuli plain. It is bounded to the North by the Southern Tagliamento river divide, to the East by the river itself. It is characterized by a rich and complex hydrographic system; the rivers, e.g. Cellina and Meduna, ﬂow in a N–S direction, and, secondarily, NE–SW, as a consequence of tectonics. Geomorphologically, the Province is composed of two well deﬁned regions: to the N the mountains, the plain to the S. Focusing on lithologies, the mountain region is mainly represented by carbonate rocks, e.g. the well-known Dolomia Principale (Upper Triassic) and minor terrigenous rocks, in the younger complex, e.g. the Scaglia Rossa (Upper Cretaceous). In the Piedmont, there is a clear prevalence of terrigenous rocks (Flysch and Molassa, Cretaceous and Tertiary). The high plain consists of a huge deposit of alluvial material: it is the “zona delle conoidi ”, that are the Cellina, Meduna, Tagliamento river fans: they can reach 200 m thickness. This deposit displays a N–S variation in granulometry, as a consequence of the slight inclination of the plain itself, with selective deposition of waterﬂow loads in decreasing granulometric classes: boulders and gravels in the high-plain (“zona dei magredi ”) down to sand, clay and mud South of the “spring line” (Fig. 1). The low-plain is constituted by semi- or non-permeable sediments, and spreads 40 km grading to the sea; going down- and southward from the “linea delle risorgive” there is a gradual prevalence of sands and clays, as shown by the stratigraphic cores obtained from wells. The ﬁrst clays in the North, e.g. Maniago, are found at a depth of 200 m, while in the South, e.g. Azzano Decimo, at a depth of 2 m [3].

Fig. 1. Cross section of the plain region, from the Cellina river fan to the “spring line”: the granulometric grading of the alluvium is clearly shown. The dashed line indicates the mean ground-water level.

522

C. Giovani et al.

The mountain area is highly stressed: the stratigraphic units are organized in bands with an E–W axis, the contacts are tectonic, due to several structural lines of regional scale and importance, mostly thrust faults and overthrusts. The plain is affected by the buried Sequals Line, striking NE–SW: it has been identiﬁed and localized by seismic surveys; it is the only extensional feature of the area; this would explain the plain subsiding [4,5]. 2.2. Trieste province The province of Trieste is situated in the South-East part of Friuli-Venezia Giulia and has a surface area of 200 km2 . Geomorphologically, the province corresponds with the western end of the calcareous Karst Plateau, which abruptly descends towards the sea. Concerning lithologies three main areas can be distinguished: (1) the Karst Plateau characterized by limestone as the main substrate; (2) the hilly region, in the south, with a clear prevalence of terrigenous rocks e.g. Flysch; (3) the coastal plain constituted by semi- or non-permeable sediments [6]. The Karst is characterized by a rich and complex hydrological system. It has the characteristic of being formed of limestone, and therefore permeable to water, because of its high solubility and ﬁssuration. This causes the formation of innumerable subterranean cavities. Karst features, particularly small caves or channels, are difﬁcult to detect. Microkarstic features, poljes and dolines are also present. Faults, perpendicular to the coastal line, are present in the north and in the south of the province [7]. 3. Methods and data collection In the province of Pordenone, from February to June 2000 more than 900 alpha track detectors with Kodak-LR 115 ﬁlm [8] were exposed in 293 schools. In the province of Trieste from January to June 2001, 270 schools were monitored with more than 830 alpha track detectors. The main structural and constructional features of the buildings, and data relative to the occupancy time of the schools were collected in speciﬁc forms. The number of dosemeters– between 2 and 6–that were located in each school, depends on the constructional features of the building and on its dimensions, so that the collected data are adequate for assessment of the dose, as well as of the gas concentration distribution inside the building. The radon concentration measurement technique is that used in previous surveys [1]. The detectors were calibrated at NRPB in London. 4. Results The mean value calculated from the Pordenone and Trieste schools is 125 ± 140 Bq m−3 and 80 ± 90 Bq m−3 , respectively. A selection of the data was made, extracting those collected only for the ground ﬂoor of the schools. The mean value calculated from this restricted data set is 130 ± 150 Bq m−3 in the province of Pordenone and 95 ± 150 Bq m−3 in the Province of Trieste. The standard deviation is relatively high. It could be due to the geographical variability of the samples.

Radon concentration survey in schools of the Friuli-Venezia Giulia region, North-East Italy

523

Fig. 2. Percentage distributions of the schools’ mean values in the provinces of Trieste and Pordenone.

Fig. 3. Map of the radon concentrations measured in the school buildings of the province of Pordenone: (1) the Piedmont zone, (2) drainage basin of the Cellina–Meduna rivers (3) drainage basin of the Tagliamento river.

524

C. Giovani et al.

Figure 2 shows the different shape of the percentage distributions of the schools’ mean values in the provinces of Trieste and Pordenone. Figure 3 shows the distribution of the radon concentration in school buildings of the province of Pordenone. The data interpolation clearly shows the existence of higher concentration areas, which are not homogeneously distributed. Zone 1 in the ﬁgure corresponds geographically to the Piedmont, while zones 2 and 3 ﬁt respectively the drainage basin of the Cellina-Meduna and the Tagliamento rivers. The lowest concentrations are recorded in the low plain, South of the “spring line”, characterized by the presence of muds and clays. The data have been grouped in 5 sets, following the geographical subdivision in Health Districts. These sets consist respectively of 40, 70, 58, 42 and 80 schools. The results are reported in Table 1 in terms of percentages of schools where mean values of radon concentration in the ground ﬂoor are above the reference values of 200, 400 and 500 Bq m−3 , respectively. The highest, the intermediate and the lowest concentrations have been found respectively in the Northern, Western and Southern districts. Table 2 shows the mean value of radon concentration for each Health District and the relative standard deviation. The high concentration checked in the North and West zones is not imputable to the building types; in fact a similar variety of buildings was found in all the districts. The mean values of indoor radon concentrations are grouped in two sets: in the ﬁrst set there are North and West district data and in the second South and urban data. The Kolmogorov– Smirnov test (KS-test) was performed in order to determine whether the two data sets differ Table 1 Number and percentage–on the total of each health district–of schools where the mean radon concentration exceeds respectively 200, 400 and 500 Bq m−3 Health district

No. of schools

Percentage (%)

Schools with average radon concentration higher than 200 Bq m−3 East 3 North 27 West 9 South 0 Urban 8 All districts 47

8 39 16 0 11 17

Schools with average radon concentration higher than 400 Bq m−3 East 0 North 9 West 3 South 0 Urban 3 All districts 15

0 13 5 0 4 5

Schools with average radon concentration higher than 500 Bq m−3 East 0 North 7 West 1 South 0 Urban 1 All districts 9

0 10 2 0 1 3

Radon concentration survey in schools of the Friuli-Venezia Giulia region, North-East Italy

525

Table 2 Mean values of radon concentration (Bq m−3 ) for the ground ﬂoor and for all levels and relative standard deviations of different health districts of the province of Pordenone Health district

Ground ﬂoor

Standard dev.

All levels

Standard dev.

East North West South Urban

90 230 120 60 105

60 230 115 30 105

85 220 120 60 95

50 230 120 30 85

Table 3 Number and percentage of schools in different municipalities of the province of Trieste where the mean radon concentration exceeds respectively 200, 400, 500 Bq m−3 Municipality

No. of schools

Percentage (%)

Schools with average radon concentration higher than 200 Bq m−3 Trieste Duino Aurisina Muggia S. Dorligo della Valle Sgonico-Monrupino Total

18 8 0 0 1 27

9 32 0 0 14 10

Schools with average radon concentration higher than 400 Bq m−3 Trieste 6 Duino Aurisina 1 Muggia 0 S. Dorligo della Valle 0 Sgonico-Monrupino 0 Total 7

3 4 0 0 0 3

Schools with average radon concentration higher than 500 Bq m−3 Trieste 3 Duino Aurisina 1 Muggia 0 S. Dorligo della Valle 0 Sgonico-Monrupino 0 Total 4

1 4 0 0 0 1

signiﬁcantly [9]. The maximum difference between the cumulative distributions, D, is 0.62. Thus D = 0.62 suggests a signiﬁcant difference. The data for the Province of Trieste have been grouped in 5 sets, following the geographical subdivision in the municipality. The results are reported in Table 3 in terms of percentages of schools where mean values of radon concentration are above the reference values of 200, 400 and 500 Bq m−3 , respectively. Figure 4 shows the map of the province of Trieste with the schools’ locations and the substratum lithology. We can see that the schools with high radon concentration are located in the zone with a calcareous substratum. The mean values of indoor radon concentrations are

526

C. Giovani et al.

Fig. 4. Map of the province of Trieste with the schools’ locations and the lithology substratum: (1) the Karst plateau characterized by limestone; (2) the hilly region with a clear prevalence of terrigenous rocks e.g. Flysch; (3) the coastal plain constituted by semi- or non-permeable sediments.

Table 4 Mean values of radon concentration and relative standard deviations of different lithologic groups of the province of Trieste Lithology

Limestone Flysch Total

Mean value of radon concentration (Bq m−3 ) Ground ﬂoor Standard dev. All levels

Standard dev.

200 55 95

130 40 90

170 50 150

180 50 80

grouped in two sets: in the ﬁrst set there are data for the schools with a calcareous substratum and in the second data set for the schools with a substratum of Flysch. Table 4 shows the mean value of radon concentration for each group and the relative standard deviation. The Kolmogorov–Smirnov test (KS-test) was again carried out in order to determine whether the two data sets differ signiﬁcantly [9]. The maximum difference between the cumulative distributions, D = 0.64, suggests a signiﬁcant difference.

5. Discussion and conclusions Even if the data reported in this survey refer to measurements inside buildings with different constructional characteristics and with heterogeneous distribution, nonetheless the results make it possible to identify well deﬁned zones geographically and geologically, where the percentage of buildings that exceed the reference levels is signiﬁcantly high.

Radon concentration survey in schools of the Friuli-Venezia Giulia region, North-East Italy

527

In the province of Pordenone, these zones (Sequals, Maniago, Aviano, Sacile) deﬁne a belt, corresponding to the piedmont and high-plain zones, where the cover consists of highly permeable gravel deposits, which favour radon diffusion from soil to air. Moreover, this area is the transition belt from the compressive strain ﬁeld of the mountain region to an extensional strain ﬁeld characteristic of the ﬂat region. It does not seem, however, that the tectonic regime is to be considered the key factor in the distribution of radon emissions: if it were true, higher radon concentrations would be expected in the Southern part of the province. The quaternary cover and soils of the Southern area (silt and clay) have low permeabilities; so that the main factor that determines the low emissions in this area is the kind of cover. In the province of Trieste the zone with high radon concentration coincides with the Karst region, characterized by limestone substratum. Karst features, particularly small caves, channels, microkarstic features, poljes and dolines, give a very unhomogeneous radon distribution and the high radon concentrations are strictly correlated to these kinds of features. On a regional scale a correlation between radon concentration in buildings and the geological characteristics seems acceptable. On a local scale, however, other factors have to be taken into account, such as the characteristics of the building and those of the ground on which they are built. For the above reasons a single and unique risk criterion cannot be identiﬁed.

Acknowledgements Part of this study was ﬁnancially supported by the Project rep. 4158 “Studio sulla radioattività ambientale dovuta alla presenza di gas radon negli ediﬁci scolastici” with the Environmental Department of province of Pordenone.

References [1] M.R. Malisan, R. Padovani, Assessment of Radon Exposure in kindergartens in North-East Italy, Radiat. Prot. Dosim. 56 (1–4) (1994) 293–297. [2] M.R. Malisan, R. Padovani, C. Foti, The regional survey of indoor radon in Friuli-Venezia Giulia (Italy), in: Proc. Int. Workshop on Radon Monitoring in Radioprotection, Environmental Radioactivity and Earth Sciences, ICTP Trieste, World Scientiﬁc, Singapore, 1991. [3] S. Stefanini, Litostratigraﬁe e caratteristiche idrologiche di pozzi della pianura friulana, dell’anﬁteratro morenico del Tagliamento e del campo di Osoppo e Gemona, Università degli Studi di Trieste,Istituto di Geologia e Paleontologia, Trieste, 1986. [4] A. Zanferrari, G. Bollettinari, L. Carotene, A. Carton, G.B. Carulli, D. Castaldini, A. Cavallin, M. Panizza, G.B. Pellegrini, F. Pianetti, U. Sauro, Evoluzione neotettonica dell’Italia nord-orientale, Mem. Sc. Geol. 35 (1982) 355–376. [5] A. Castellarin, G.B. Vai, Guida alla geologia del Sudalpino centro-orientale, Guide Geologiche Regionali, Società Geologica Italiana, Bologna, 1982, 386 pp. [6] Carta di sintesi dei substrati geolitologici, Regione Autonoma Friuli-Venezia Giulia, Direzione Regionale delle Foreste, Servizio della Selvicoltura, Udine, 1998. [7] G.B. Carulli, F. Cucchi, Proposta di interpretazione strutturale del Carso triestino, Atti Ticinensi di Scienze della Terra 34 (1991) 161–166. [8] L. Tommasino, Passive sampling and monitoring of radon and other gases, Atti del convegno: Radon tra natura e costruito Venezia-Mestre 24-26/11/1997. [9] S. Siegel, N.J. Castellan Jr., Statistica Non Parametrica, McGraw–Hill, Milano, 1992.

528

210 Po

implanted in glass surfaces: Calibration and improved performance for retrospective radon reconstruction D.J. Steck a , J.A. Berglund a , R.W. Field b a Schaefer Environmental Radiation Laboratory, Physics Department., St. John’s University,

Collegeville, MN 56321, USA∗

b Department. of Epidemiology College of Public Health, N222 Oakdale Hall University of Iowa,

Iowa City, IA 52242, USA∗

We studied the implanted 210 Po–210 Pb activities and cumulative radon gas exposure of glass surfaces in two groups of homes to establish an authentic calibration for retrospective radon gas reconstruction and to identify factors that affect the calibration. Sixty-seven glass objects from 24 Minnesota homes were analyzed. The radon gas concentrations in these homes were measured repeatedly with annual, yearlong detectors over a two-decade span while the glass was being exposed to radon progeny. The surface activity–radon exposure correlation was better (R 2 ∼ 0.8) than in studies where only contemporary radon concentrations had been measured. For example, in our study of a large group of Iowa homes, cumulative radon exposure based on contemporary radon gas measurements showed an R 2 ∼ 0.5 for 2147 glass surfaces. The lower correlation probably results from the inaccuracy of the contemporary radon gas based exposure estimates and the more diverse set of deposition conditions, including many homes with a smoker. The correlation between implanted activity and radon gas exposure was improved by 10 to 20% when the measured contemporary radon progeny deposition rates were included in the regression as a variable. We investigated new techniques to reduce the signiﬁcant interference in our detectors from intrinsic alpha activity in glass. We did not observe any signiﬁcant change in the activity–exposure ratio with the age of the glass, suggesting that chronic losses over time are small. Neither data set showed any signiﬁcant effect due to the size, or type of the glass; washing frequency, existence of surface ﬁlms, or air movement. Implanted surface activity measurement proved to be as good a predictor for long-term radon concentrations as yearlong radon gas measurements in the Minnesota homes. * Supported by grant R01 CA859422 from the National Cancer Institute, National Institutes of Health. This report

is solely the responsibility of the authors and does not necessarily reﬂect the ofﬁcial views of the NCI or NIH. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07063-9

© 2005 Elsevier Ltd. All rights reserved.

210 Po implanted in glass surfaces

529

1. Introduction Accurate cumulative radon exposure assessments are crucial given the signiﬁcant role that radon progeny plays in some lung cancers [1]. The highly variable temporal and spatial behavior of indoor radon gas and progeny concentrations makes assessments of long-term radon exposure difﬁcult. Radon concentrations in a room may vary signiﬁcantly over time periods as long as a year [2]. Hence, contemporary radon measurements, even those that average over a year, may not be an accurate measure of the long-term average radon concentration. The existence in the radon decay chain of a long-lived radon decay product. 210 Pb, followed by an alpha-emitter 210 Po, provides a convenient method to estimate cumulative radon gas exposure [3,4] and long-term average radon concentrations. The 210 Po activity must be trapped in a stable matrix for years for this technique to be successful. Glass has been shown to be a good matrix when the activity is implanted in the surface through recoil following alpha decay. Laboratory experiments conﬁrm a high correlation between implanted 210 Pb and cumulative radon exposure for glass surfaces exposed to high radon concentrations in a clean stable environment [5]. Previous ﬁeld studies in homes show lower correlation between implanted activity and contemporary radon gas concentrations in a wider variety of atmospheres [6–10]. Numerous factors have been suggested for the lower correlation in homes as compared to the lab: long-term changes in radon concentration; variation in deposition conditions due to aerosols or air movement; impediments to implantation from surface ﬁlms or nearby objects, chronic activity loss from washing or leaching, and the variation of the natural uranium and thorium content of glass. We studied the implanted 210 Po activity and cumulative radon gas exposure on glass surfaces in a group of homes to establish a simple, authentic calibration for retrospective radon gas reconstruction under realistic exposure conditions. In a larger, more diverse group of homes, we investigated the effects of some of the factors that may affect the ratio of implanted activity to cumulative radon exposure.

2. Methods A calibration exercise was conducted in 24 Minnesota homes where radon gas concentrations were measured repeatedly over two decades [11]. Yearlong radon gas measurements were made with alpha track detectors (ATDs) starting in 1983. Most homes have one or more measurements in the early 1980s and continuous measurements from 1990 to 1999. Multiple imputations were used to cover missing measurement periods. Imputation was needed only 40% of the time, on average. Hence, the cumulative radon exposure, deﬁned as the product of the radon concentration times the age of the glass, was primarily based on direct measurements. Glass surfaces were selected in these houses whose age and location best matched the radon gas measurements. Implanted surface alpha activity was measured for 60 days by track registration material (LANTRAK) incorporated in a retrospective radon detector (RRD) module. The surface activities were corrected for background and intrinsic alpha activity emanating from the glass to calculate a 210 Po activity density. The activities were further adjusted to calculate the 210 Pb activity surface density by correcting for the age of the glass to account for the decay of 210 Pb and disequilibrium between 210 Po and 210 Pb. All subsequent references

530

D.J. Steck et al.

to implanted surface activity are the corrected 210 Pb surface activity densities The RRDs were also able to measure contemporary radon gas concentration and deposited alpha activity using two additional track registration chips [10–12]. Implanted surface activities and contemporary radon gas concentrations were measured in more than one thousand homes as part of the Iowa Radon Lung Cancer Study [12,13]. The detectors were placed and retrieved by technicians who also collected detailed home, occupant, and glass surface information. The response of the track registration material to alpha particles of varying energy was measured using thin, calibrated sources of 230 Th, 210 Po, 214 Po, and 212 Po. We were able to measure the efﬁciency from less than 1000 keV to 8600 keV using energy-reducing ﬁlters whose thickness ranged from 0.4 to 3.2 mg cm−2 . The natural alpha-emitting contamination (intrinsic) alpha activity in glass was studied by placing RRDs on 21 different new glass samples that had not been exposed to radon in homes. We measured the activity of a calibrated thick source (NIST SRM610) that has equal activity of the uranium and thorium series uniformly distributed in a glass matrix with a semiconductor diode detector (300-mm2 ) and track registration detectors.

3. Results and discussion Two separate studies, one in Minnesota and one in Iowa, are described below. In the Minnesota homes, the correlation between age-corrected 210 Pb activity and the measured cumulative radon gas exposure was very good (see Fig. 1). The coefﬁcient of determination (R 2 ) was 0.81 for a simple linear regression. This R 2 was substantially larger than in our previous ﬁeld results [5,6,10,12] where cumulative radon exposure was based on a single yearlong radon measurement and the glass’ age. However, the slope of the activity–exposure line, 1.8 m ky−1 , was similar to previous work in groups of homes where active smoking was rare. Only 19% of the glass surfaces in the Minnesota homes were exposed to more than 1 pack-year.

Fig. 1. Age corrected implanted 210 Pb activity and measured cumulative radon gas exposure in 67 glass surfaces from 24 homes.

210 Po implanted in glass surfaces

531

In the 990 Iowa homes, the cumulative radon gas exposure had to be estimated from a single contemporary radon gas measurement. A simple regression analysis of the 2147 glass surfaces had a lower R 2 (0.5), than in the Minnesota homes. The slope (1.0 m ky−1 ) was also lower. This lower slope probably reﬂects the deposition effectiveness, and possibly implantation impediments, due to the higher aerosol atmospheres in these homes where smoking was often present. In addition, previous theoretical and experimental work predicts more variation in the activity–exposure ratio where aerosol concentrations can be quite variable [14–16]. Since our detectors measure deposited contemporary radon progeny activities, we added an extra term that included this deposited activity–radon exposure ratio. The R 2 improved for both the Minnesota homes (0.81 to 0.89) and the Iowa homes (0.5 to 0.6). In the Minnesota homes, we compared the accuracy of surface activity measurements using the regression calibration to predict cumulative radon exposure with the accuracy of the prediction based on a single yearlong ATD measurement. We randomly selected an ATD result from one year during a 3-year window. The time window simulates the length of many epidemiologic studies. We observed similar results with a 5-year window. Both ATD and RRD measurement methods gave a prediction that was within 25% of the measured cumulative exposure in these homes. Since surface activity measurements can be performed much more rapidly, they may be the technique of choice for future home screening assessment or epidemiologic studies. The clear trend of the data in Fig. 1 is towards a non-zero intercept on the activity axis. One possible explanation for this behavior, non-zero intercept, is the variation of natural alpha-emitting contamination, the intrinsic alpha activity, in the glass. Our detectors (RRDs) were designed to measure intrinsic activity during a yearlong exposure using a thick ﬁlter (5000 keV loss) between the glass and the detector. The 210 Po (5310 keV) was determined from a region of the detector covered by a thin ﬁlter (400 keV loss), corrected for intrinsic activity from the thick ﬁlter region. Fig. 2 shows that the bulk of the intrinsic activity

Fig. 2. Alpha activity energy spectrum (dots + solid line) of a glass sample with equal U and Th contamination compared with the detection efﬁciency of our track registration material (long dashed line).

532

D.J. Steck et al.

(dotted-solid line) from a simulated natural glass source falls in the 0 to 6000 keV range. The RRD detection efﬁciency has a broad peak from 2000 to 6000 keV. Hence, when the RRDs were exposed for only 60 days in the Minnesota homes, too few tracks were produced in the thick region to provide an accurate measure of the intrinsic activity. Since the unexposed glass samples exhibited an average contamination rate equivalent to 1.3 Bq m−2 of 210 Po, we subtracted that constant rate from the Minnesota results. However, the samples set had a standard deviation of 1.0 Bq m−2 . Such a variation hinders accurate interpretation of glasses with low exposure. We are currently testing a technique that uses, thin, medium, and thick ﬁlters. Other researchers use a LR115 companion detector to measure intrinsic activity (9). However, LR115 is much more sensitive to light than our material. We found that windows are often the only suitable glass surface available. We conducted a preliminary statistical analysis of the Iowa home results to look for some of the factors that might affect the activity–exposure slope. Deposited household aerosol can form a ﬁlm that impedes implantation. However, we found no signiﬁcant difference in the activity–exposure ratio based on observable ﬁlms, the amount of surface material removed from the glass when installing the RRD, or the homeowners cleaning frequency habits. Some researchers have avoided using windows because of the possibility of increased air movement or thermophoresis. However, windows had the same activity–exposure ratio distribution as the other major categories of glass (mirrors, cabinets doors, pictures) that we measured. Two recent works have suggested that the implanted activity may be lost through various mechanisms over long periods of time [17,18]. Losses would reduce the activity-exposure ratio for older glasses. As Figs. 3 and 4 show, we observe no signiﬁcant reduction in the ratio is evident in either the Iowa study or Minnesota study homes. Neither a linear nor a log-linear regression analysis had an R 2 > 0.05.

Fig. 3. 210 Pb activity–radon gas exposure ratio and age of glass surfaces from the Iowa study.

210 Po implanted in glass surfaces

533

Fig. 4. 210 Pb activity–radon gas exposure ratio and age of glass surfaces from the Minnesota homes.

4. Conclusion Implanted surface activities provide a good measure of cumulative radon exposure. A measurement of the contemporary deposition rate and radon gas can improve the predictive accuracy of the technique. Since surface activities can be measured rapidly, they are an attractive alternative for future home radon screening assessments or epidemiologic studies involving lung cancer where estimates of retrospective radon exposure are required.

References [1] Committee on Health Risks of Exposure to Radon, Board on Radiation Effects Research, Commission on Life Sciences, National Research Council, Health Effects of Exposure to Radon: BEIR VI, National Academy Press, Washington, DC, 1999. [2] D.J. Steck, Spatial and temporal indoor radon variations, Health Phys. 62 (1992) 351–355. [3] R.S. Lively, E.P. Ney, Surface radioactivity from the deposition of 222 Rn daughter products, Health Phys. 52 (1987) 411–415. [4] C. Samuelsson, Retrospective determination of radon in houses, Nature 334 (1988) 338–340. [5] R.S. Lively, D.J. Steck, Long-term radon concentrations estimated from 210 Po embedded in glass, Health Phys. 64 (1993) 485–490. [6] D.J. Steck, R.S. Lively, E.P. Ney, Epidemiological implications of spatial and temporal radon variations, in: Proceedings of Hanford Symposium on Health and the Environment: Indoor Radon and Lung Cancer: Reality or Myth, Richland, WA, 1990, pp. 889–904. [7] C. Samuelsson, L. Johansson, M. Wolff, 210 Po as a tracer for radon in dwellings, Radiat. Prot. Dosim. 45 (1992) 73–75. [8] J.A. Mahaffey, M.A. Parkhurst, A.C. James, F.T. Cross, M.C.R. Alvanja, J.D. Boice, S. Ezrine, P. Henderson, R.C. Brownson, Estimating past exposure to indoor radon from household glass, Health Phys. 64 (1993) 381– 391. [9] R. Falk, K. Almren, I. Ostergren, Experience from retrospective radon exposure estimations for individuals in a radon epidemiological study using solid-state nuclear track detectors, Sci. Total Environ. 272 (2001) 61–66.

534

D.J. Steck et al.

[10] R.W. Field, D.J. Steck, M.A. Parkhurst, J.A. Mahaffey, M.C.R. Alavanja, Intercomparison of retrospective radon detectors, Environ. Health Perspect. 107 (1999) 901–904. [11] D.J. Steck, M.C.R. Alavanja, R.W. Field, M.A. Parkhurst, D.J. Bates, J.A. Mahaffey, 210 Po implanted in glass surfaces by long term exposure to indoor radon, Health Phys. (to be published August 2002). [12] D.J. Steck, R.W. Field, The use of track registration detectors to reconstruct contemporary and historical airborne radon (222 Rn) and radon progeny concentrations for a radon-lung cancer study, Radiat. Measur. 31 (1999) 401–406. [13] R.W. Field, D.J. Steck, B.J. Smith, C.P. Brus, J.S. Neuberger, E.F. Fisher, C.E. Platz, R.A. Robinson, R.F. Woolson, C.F. Lynch, Residential radon gas exposure and lung cancer: the Iowa radon lung cancer study, Am. J. Epidemiol. 151 (2000) 1091–1102. [14] J. Porstendörfer, Properties and behaviour of radon and thoron and their decay products in the air, J. Aerosol Sci. 25 (1994) 219–263. [15] J. Porstendörfer, A. Reineking, Radon: characteristics in air and dose conversion factors, Health Phys. 76 (1999) 300–305. [16] B. Fitzgerald, P.K. Hopke, A prospective assessment of the 210 Po surface collection for estimating 222 Rn exposure, J. Environ. Radioac. 51 (2000) 79–98. [17] P. Cauwels, A. Pofﬁjn, An improved model for the reconstruction of past radon exposure, Health Phys. 78 (2000) 528–532. [18] R.L. Fleischer, R.H. Doremus, Uncertainties in retrospective radon exposure of glass: possible effects of hydration and leaching, Health Phys. 81 (2001) 110–113.

535

Development of a method for estimating the airborne concentration of radon progeny, using an imaging plate T. Iimoto, T. Kosako, N. Sugiura, K. Kawashima Research Center for Nuclear Science and Technology, The University of Tokyo, 2-11-16 Yayoi, Bunkyo-ku, Tokyo 113-0032, Japan

The airborne 222 Rn (radon) progeny is a primary natural radioactive source generating an effective dose. It is important to identify high concentration areas for the radon progeny and to estimate these concentrations in air. However, most of general commercial instruments for measuring radon progeny concentrations using a precise electric detector are not perfect, because they are not suitable for use in high humidity environments that can reach up to 100% such as caves, utility mains, spas, and underground facilities. Therefore we are proposing the method demonstrated here using an imaging plate (IP) as a radiation detector that has strong characteristics against humidity. An IP type named BAS-III was selected mainly because of its high detection stability in varying degrees of humidity. Individual concentrations of shortlived radon progeny were measured with a combination of a ﬁlter and grab-sampling and decay methods. A glass-ﬁber ﬁlter (GF/F) was used. The sampling ﬂow rate was 65 lpm, and the grab-sampling time was 5–10 minutes. One minute after air sampling, the surface of the ﬁlter was attached directly to the IP. This was repeated six times for 10-minute-each, for a total exposure of 60 minutes. PSL (Photo-Stimulated Luminescence) values from the six repetitions were analyzed to establish the concentration of radon progeny using a least squares calculation that was based on the ﬁtting procedure due to the theoretical decay curve of progeny. The lower detection limit of the equilibrium-equivalent concentration of radon (EEC Rn) using this system is estimated to be about 20 Bq m−3 . According to inter-comparison tests with a representative commercial instrument using a ZnS(Ag) detector, the EEC Rn ratio of the proposed instrument was 0.98 ± 0.04. Therefore, this system was successfully used to conduct measurements in places such as caves, utility mains, spas, and underground facilities.

1. Introduction The airborne 222 Rn (radon) family is a primary natural radioactive source generating an effective dose. This family is responsible for about half of the personal effective dose due to RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07064-0

© 2005 Elsevier Ltd. All rights reserved.

536

T. Iimoto et al.

natural radiation. The important target nuclides in the radon family are decay products called radon progeny, which are ﬁne metal airborne particles with short half-lives. Speciﬁcally, they are 218 Po, 214 Pb, 214 Bi, and 214 Po. Therefore, it is extremely important to estimate precisely the airborne concentration of radon progeny. Commercial instruments for measuring radon concentrations are, however, not the most suitable for measurements in certain natural environments such as caves, utility mains, spas, and underground facilities, since most of them are affected by air humidity, which can reach 100% in such environments. These special environments are naturally enhanced radiation places, where investigation of the effective dose is needed. A new system that could overcome these problems is proposed according to the method demonstrated here, which uses an imaging plate as a radiation detector for estimating the airborne concentration of individual radon progeny. 2. Materials and methods 2.1. IP selection and sensitivity The imaging plate (IP) is a relatively new and simple tool that spatially detects radiation levels. Several kinds of IP have been developed. For example, Fuji Photo Film Co., Ltd. has developed BAS-III for high-sensitivity measurements, BAS-SR for precise position analysis, and BAS-TR for 3 H measurements. These IPs have different levels of sensitivity for radiation measurements. In this study, high sensitivity for the alpha particle is one of the most important features; therefore, the BAS-III was selected because it is about twice as sensitive as the other IPs. BAS-III also shows high detection stability in varying degrees of humidity. This feature is extremely advantageous when conducting measurements in some special environments such as caves or spas, where the humidity can be around 100%. IP measurement uses a special quantitative unit of PSL (Photo-Stimulated Luminescence) to indicate the exposed level of radiation. ‘100 PSL mm−2 ’ is deﬁned to be equivalent to the detection value gained after exposure to 0.15 mR by X-ray with a W target. The PSL is approximately proportional to the product value between the source activity and the exposure time. The proportional factor is, however, different for different type and energy of radiation. Radon short-lived progeny consisting of four nuclides irradiates alpha, beta, and gamma rays with a wide range of energy. Therefore, an important task in this study was to investigate the sensitivity of the BAS-III for each type of radiation. In order to investigate the sensitivity for different types of radiation, the following point sources were utilized. For an alpha ray, 5.49 MeV of 241 Am was selected. In the same way, 1.71 MeV of 32 P for a beta ray and 60 keV of 241 Am and 1.17/1.33 MeV of 60 Co for a gamma ray were selected. When the energy of the radiation was changed, the sensitivity of the IP, that is, PSL mm−2 /radiation, could be changed as mentioned above. However, in this study, we wanted to investigate the rough sensitivity ratio among the types of radiation in order to apply the IP to the radon progeny measurements. Therefore, these energies were selected as the representative energies for this test. Each radiation source is directly attached to the surface of the BAS-III for 10–20 minutes. Twenty-four hours after the end of exposure, i.e., the ﬁrst fading period discussed in the next section, the PSL value was analyzed. The PSL analyzer is the BAS-1800 with settings of sensitivity 4000, latitude 5, steps 256, and resolution 10−4 m in this study. These settings

Method for estimating the airborne concentration of radon progeny

537

Table 1 Average intensity (PSL mm−2 ) per single incident according to the type of radiation Radiation type

Nuclide

Energy [MeV]

(PSL mm−2 )/(single radiation)

Alpha Beta Gamma

241 Am

5.49 1.71 0.06 1.17/1.33

0.48 0.098 0.011 0.0015

32 P 241 Am 60 Co

were not changed for any analysis of PSL in the study. Table 1 shows the results of these sensitivity tests. The sensitivity ratio of PSL mm−2 /radiation is approximately alpha : beta : gamma = 50 : 10 : 1. This result is quite similar to that reported by Mori [1] after the fading effect is considered. According to this, the gamma sensitivity was judged to be negligible for the purpose of this study when compared with that for alpha and beta. Therefore, we decided not to consider the gamma effects on PSL from the radon progeny in the measurements. 2.2. Fading of PSL intensity When using IPs, it is also important to investigate the PSL fading feature of BAS-III. This feature is used to determine the start time for analyzing the PSL after the end of exposure in order to obtain stable measurements. Related information has been reported by Ohuchi [2], who has demonstrated that environmental temperatures have the potential to dramatically change the fading speed. In this study, however, it is assumed that the IP would be used under a comparatively stable environment at a temperature of 15–20 ◦ . Therefore, the fading speed is assumed to be constant. PSL fading is also individually investigated for various types of radiation. The results of the fading feature are shown in Fig. 1. In this experiment, 241 Am for alpha, 32 P for beta, and

Fig. 1. PSL fading feature of BAS-III, standardized at 24 hours after the end of exposure.

538

T. Iimoto et al.

241 Am

for gamma are used. The length of exposure for each result is 10–20 minutes. As seen in Fig. 1, at about 10 hours after the end of the exposure, the ﬁrst strong fading phenomenon ﬁnishes for all types of radiation. Therefore, the PLS value fully stabilizes about 24 hours after exposure to radiation stops. In this study, the standard time for analyzing PSL has been determined to be 24 hours from the end of exposure, and all PSL data would be corrected accordingly. 2.3. Measuring technique for individual radon progeny concentration Individual concentrations of radon progeny were measured with a combination of a “ﬁlter method” and “grab sampling and decay method” [3]. A glass-ﬁber ﬁlter (GF/F, manufactured by Whatman Co.) was selected for the sampling ﬁlter. The pore size of the ﬁlter is equivalent to 0.8 micrometer. The collection efﬁciency for the radon progeny is around 100% and the pressure drop is lower than that of other similar ﬁlters [4]. The sampling ﬂow rate was 65 lpm, and the grab-sampling time was 5–10 minutes. One minute after the air sampling, the surface of the ﬁlter was attached directly to the IP detection surface inside of the IP cassette. The exposure was repeated six times for 10 minutes each as the position of the attachment was changed for a total exposure of 60 minutes. This procedure is shown in Table 2. The PSL intensity from the six exposures was analyzed to establish the concentration of radon progeny using a least squares calculation that was based on the ﬁtting procedure due to the theoretical decay curve of progeny. The basic equation is: Pa+b Ka = Kf V ηa Iaa (i) + ηc Iac (i)Xa + ηc Ibc (i)Xb + ηc Icc (i)Xc ; (1)

• Pa+b is PSL intensity corrected at 24 hours after the end of exposure [PSL mm−2 ]; • Ka is correction factor for the PSL value changing into the value inﬂuenced only by alpha rays from radon progeny [–]; • Kf is correction factor for fading feature standardized at 24 hours after the end of exposure; • V is sampling ﬂow rate [m3 s−1 ]; • ηa , ηc are conversion factors including detection efﬁciency of 218 Po (a) and 214 Po (c) from the alpha decay number to the PSL value of the alpha ray; • Inm (i) is decay number as an ‘m’ nuclide in the exposure time of ith for an ‘n’ nuclide collected in the sampling time when the ﬂow rate is 1 [m3 s−1 ], Xj is 1 [Bq m−3 ], and n/m is 218 Po (a), 214 Pb (b), and 214 Po (c); • Xj is radon progeny concentration [Bq m−3 ]; j denotes 218 Po (a), 214 Pb (b), and 214 Bi (= 214 Po) (c). Table 2 Time sequence for sampling radon progeny and exposing them to IP (min) Sampling progeny

Preparing Exposing IP to radiation from ﬁlter IP 1st 2nd 3rd 4th

5th

6th

0–10∗

10–11

51–61

61–71

11–21

∗ Example of sampling time; 10 min.

21–31

31–41

41–51

Method for estimating the airborne concentration of radon progeny

539

Fig. 2. An example of the difference in the measurements between alpha and beta.

Six equations are obtained from the six results in the PSL data. The correction factor Ka and conversion factors ηa , ηc in the basic equation are discussed in the next section. When the concentration in the target environment is predicted to be about 10 times higher than the natural level, more PSL value can be obtained, and the measurement error can be decreased. In this case, a more precise measurement can be obtained by covering one half of the ﬁlter surface with two sheets of paper in order to distinguish between alpha and beta information. In this case, it is not necessary to use the correction factor Ka , which is slightly changeable depending on environmental conditions and can cause estimation errors, after calculating the PSL intensity due to alpha rays only by the following equation: Pa+b Ka = Pa+b − Pb

(2)

Pa+b , Pb = PSL intensity of the bare IP region (a + b) or beta region (b) covered with two sheets of paper corrected at 24 hours after the end of exposure. An example of the difference in the measurements between alpha and beta is shown in Fig. 2. The left half of each ﬁlter position is the alpha and beta region (a + b), which is the bare IP part. The right half is the beta region (b), which is covered with two sheets of paper in order to cut the alpha rays. 3. Discussion 3.1. Correction factor Ka In order to estimate the effective Ka , the differences between alpha and beta are measured 20 times. The representative correction factor Ka is determined as the average value of (Pa+b − Pb )/Pa+b from among the 20 measurements. In these experiments, the individual concentration ratio among radon progeny was artiﬁcially changed over a large range by varying the environmental conditions, especially the aerosol concentration in the air. The individual ratio of 218 Po : 214 Pb : 214 Bi (= 214 Po) would have some effect on the value of Ka because the ratio would change the number ratio between the alpha and beta rays. The individual ratio

540

T. Iimoto et al.

of radon progeny was changed here between approximately 1 : 1 : 1 and 1 : 0.4 : 0.2 from an equilibrium condition to a special condition. As a result, Ka = 0.87 was obtained as the mean value by the experiments. The margin of error was about 10%. The value of Ka would only be used in the case of measurements for a low concentration. Therefore, the errors for the other parameters or factors, for example, the PSL intensity itself, would be larger than that of Ka in the low concentration environments, and the 10% error would not be effective for the calculation of the radon progeny concentration. 3.2. Conversion factor The conversion factors of 218 Po and 214 Po from the alpha decay number to the PSL value of Pa+b , ηa and ηc , respectively, are determined by the following experiment with a Si semiconductor detector. The conversion factor includes the detection efﬁciency. A GF/F ﬁlter holding the radon progeny on its surface was prepared. The mutual procedure for attaching the ﬁlter to the IP surface with the method involving the difference in alpha and beta for 5 minutes and the measurement by alpha spectroscopy using the Si detector for 5 minutes was repeated for 7 sets at 30-s intervals. The airborne concentrations of radon progeny sampled on the source ﬁlter were calculated by the least squares method [5] using the alpha decay curve obtained by the Si detector. The concentrations were ﬁtted to 7 equations of combinations of (1) and (2) using the PSL intensity of IP due to only alpha. Then, the two unknown factors of ηa and ηc were determined by the least squares calculation. This experiment was repeated 10 times to estimate the average values. As a result, the values of ηa = 7.4 × 10−5 and ηc = 1.3 × 10−4 were obtained. These values have a margin of error of 5% each. 3.3. Detection sensitivity The background PSL value due to natural radiation during the IP storage period was the main factor contributing to the detection sensitivity of the method. This feature was also estimated experimentally. Here, the following three representative methods for storing the IP during the 24-hour fading period are discussed. The ﬁrst method involves the use of a commercial IP cassette (cassette storage). Its natural environment background is 70–80 μSv per hour. The second method involves placing the IP in a lead shield that is 5 cm thick (cassette + Pb (5 cm) storage). The background is estimated to be 10–15 μSv per hour. The third involves placing the IP in a cassette with an iron shield that is 25 cm thick (cassette + Fe (25 cm) storage). The background is estimated to be about 5–7 μSv per hour. The estimated equation for the background PSL value depending on the storage duration is as follows: • Cassette storage: P = 0.17H , • Cassette + Pb (5 cm) storage: P = 0.038H , • Cassette + Fe (25 cm) storage: P = 0.022H , with P the PSL intensity due to background [PSL mm−2 ] and H the storage duration under each condition [hours]. For example, in the case of IP storage with a 5cm-thick lead shield, the background rate was 0.038 PSL mm−2 h−1 . Using this background PLS value, the lower detection limit concentration for the IP system was estimated. The estimation conditions are as follows:

Method for estimating the airborne concentration of radon progeny

541

(1) the cassette + Pb (5 cm) storage method is adopted, (2) the fade time is 24 hours, and (3) the net PSL value must be at least two times the background of the PSL. Then, the lower detection limit of the equilibrium-equivalent concentration of radon (EEC Rn) using this system was estimated to be about 20 Bq m−3 . The sensitivity of this system for EEC Rn is not so high as those of general conventional systems using a ZnS(Ag) scintillation detector or a Si semiconductor detector. For example, when a combination of ZnS(Ag) detector system with 50 mm-diameter and the data analysis by the Thomas method is adopted for EEC Rn measurements, the lower detection limit might be about 3–5 Bq m−3 [3]. However, as we mentioned in the introduction section, these precise electric devices such as ZnS(Ag) scintillation detectors or Si semiconductor detectors are not suitable for use in high humidity environments. Therefore, the measurement system using IP as the radiation detector has special advantage in this respect. Finally, inter-comparison measurements were performed between this IP system and the 50-dia. ZnS(Ag) scintillation system (Model ZDS-451B manufactured by ALOKA Co.) in a radon test chamber. Test conditions were 1500–2500 Bq m−3 of EEC Rn, 20 ◦ C of temperature and 60% of relative humidity. As a result, the estimation ratio on EEC Rn, which is the value of IP/ZnS(Ag), is 0.98 ± 0.04 as the average of three such tests. 4. Conclusion A new technique for estimating the airborne concentration of individual radon progeny has been developed. This method uses an imaging plate (IP) of BAS-III as a radiation detector. This system has a strong advantage due to the special IP feature for measurements in environments with extremely high humidity. In addition, the mobility to sample the environmental data is high because of the simple construction and light weight of the IP and cassette. An electric device is not needed to record the alpha information from the sample ﬁlter into the IP. Therefore, this system was successfully used to take measurements, especially in natural environments such as caves, utility mains, spas, and underground facilities, which are difﬁcult places for estimating airborne concentrations of radon progeny. References [1] C. Mori, A. Matsumura, Radioactivity and geometrical distribution measurements of alpha-emitter specimens with the imaging plate, Nucl. Instrum. Methods Phys. Res. A 312 (1992) 39–42. [2] H. Ohuchi, A. Yamadera, T. Nakamura, Functional equation for the fading correction of imaging plates, Nucl. Instrum. Methods Phys. Res. A 450 (2000) 343–352. [3] J.W. Thomas, Measurement of radon daughters in air, Health Phys. 23 (1972) 783–789. [4] T. Iimoto, Characteristics of major ﬁlters used for 222 Rn progeny measurements, Radiat. Measur. 29 (2) (1998) 161–164. [5] D.A. Martz, D.F. Holleman, D.E. McCurdy, K.J. Schiage, Analysis of atmospheric concentration of RaA, RaB, and RaC by alpha spectroscopy, Health Phys. 17 (1969) 131–138.

542

Concentration, distribution and transportation of 222Rn and its decay products in the environment of coastal Karnataka and Kaiga in southwest India H.M. Mahesh∗ , D.N. Avadhani, K. Siddappa Department of Studies in Physics, Mangalore University, Mangalagangotri 574 199, India

Studies on radon and its decay products in the atmosphere are useful for determining the radiological risk to the general public. Concentrations of 222 Rn and its decay products in ground level air are dependent on its exhalation rate and atmospheric diffusion depending on meteorological parameters. In the present investigation, extensive studies on 222 Rn exhalation rate, 222 Rn and its decay product concentrations in air at ambient temperature and humidity were carried out simultaneously at each sampling location in coastal Karnataka and Kaiga. Correlations of 226 Ra in soil, 222 Rn exhalation rate, equilibrium factors, and annual effective dose to the general public were estimated.

1. Introduction In the natural radiation environment, 222 Rn and its airborne decay products – 218 Po, 214 Pb and 214 Bi-cause signiﬁcant health hazards due to inhalation. They are the largest contributors to the average effective dose received by human beings [1]. 222 Rn enters into the atmosphere through its exhalation from soil. Meteorological parameters like humidity, temperature, wind direction, wind velocity, etc. inﬂuence the 222 Rn concentration, transportation and distribution in the environment [2–5]. In view of all these, systematic investigations were carried out on the concentration, distribution and transport of 222 Rn exhalation rate, 222 Rn and its decay products in air and their diurnal behavior with meteorological parameters. * Presently: PDRA, ETL, EERC, School of Civil Eng., Queen’s University Belfast, Belfast BT95AG, Northern

Ireland, UK; e-mail: [email protected]; [email protected]. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07065-2

© 2005 Elsevier Ltd. All rights reserved.

Concentration, distribution and transportation of 222 Rn

543

2. Materials and methods 2.1. Rate of 222 Rn exhalation from soil The area covered under the present investigation is as shown in Fig. 1. The rate of 222 Rn exhalation from soil was measured using the accumulation chamber method [6,7]. The accumulation chamber is of dimensions 25 cm height and radius 10.5 cm with an open end on one side and the other connected to a glass tube at the center with a valve. The open end of the accumulation chamber was driven into the surface of the soil. The sensitive volume available after ﬁxing the chamber to 3–5 cm depth was about 6900 cm3 covering an area of about 340 cm2 . Radon emanated was allowed to collect in the chamber for 60 min. The radon accumulated inside the chamber was then transferred into an evacuated scintillation cell. The scintillation cell was counted in an assembly set up for alpha counts any time after 200 minutes so that the radon inside the cell attains equilibrium with its decay products. The radon concentration was estimated and the exhalation rate of radon extrapolated. 226 Ra in soil was determined by gamma spectrometry employing the methods detailed elsewhere [8]. 2.2. Radon concentration Radon concentration in air was measured using a Low Level Radon Detection System (LLRDS). Radon decays to 218 Po atoms, which on formation are positively charged and continue to carry the charge for a while after formation, until neutralized by free electrons or plate

Fig. 1. Area covered in present investigation.

544

H.M. Mahesh et al.

out on condensation nuclei or nearby surfaces. This property of the daughter atoms of radon decay products is made use of in the LLRDS [9]. The LLRDS consists of a sample collection chamber of 24 cm diameter, 11.5 cm height and volume ∼ 8 liters. It can be evacuated for collecting samples by vacuum transfer. The electrostatic ﬁeld can be applied externally to the metallic plate, which is an integral part of the chamber cover. To start with, an air sample was collected into the chamber, at least 10 min. delay being allowed for the complete decay of thoron. A negative voltage (∼ 800 VDC) was then applied to the metallic plate for about 90 minutes for the collection of 218 Po atoms. The plate was then removed and counted for its alpha activity immediately. 2.3. Radon decay product concentrations Air samples were collected on to a glass ﬁber ﬁlter paper at a ﬂow rate of 30 lpm for a period of 15 minutes. Hourly samplings were collected for 24 hours at a height of one meter above the ground level. After sampling, the glass ﬁber ﬁlter paper was counted for alpha activity using a ZnS (Ag) alpha counter (efﬁciency 30% and background 0.2 cpm) for three counting intervals i.e. 2–5, 6–20 and 21–30 minutes. From these counts, the activities of 218 Po, 214 Pb and 214 Bi were obtained using a modiﬁed Tsivoglou method [10,11]. 3. Results and discussion The average concentrations of temperature, relative humidity, 226 Ra in soil, 222 Rn exhalation, 222 Rn and its decay product concentrations were measured for a day at seven locations of coastal Karnataka and Kaiga and the results are presented in Table 1. In addition, the mean values of prevailing temperature and humidity at these sampling locations are presented in the same table. The equilibrium factor (F ) and inhalation dose from radon and decay products are also computed and presented in Table 1. 3.1. Rate of 222 Rn exhalation The exhalation rate was found to be higher at Kaiga and signiﬁcantly higher at Udupi than at other sampling stations and lower at Honnavar (Table 1). The higher exhalation rate in the Kaiga region can be ascribed to the possible inﬂuence of regional ecology [7]. It has also been pointed out that the higher 222 Rn exhalation rate in Kaiga is due to higher values of 222 Rn diffusion coefﬁcients in Kaiga soil [12]. The minimum 222 Rn exhalation rate observed at Honnavar is mainly due to the geology of the region as the region is mainly a lateral formation and contains minimum concentrations of 226 Ra in soil. A typical diurnal variation of 222 Rn exhalation rate measured along with ambient temperature and relative humidity is presented in Fig. 2. Rate of exhalation was maximum at 14.00 hr and minimum at 22.00 hr. Relatively higher 222 Rn exhalation rates were observed to correspond to higher temperature and lower humidity and vice versa. This may be due to a relatively higher turbulence in the atmosphere due to high temperature and lower humidity. Cross correlations between exhalation rate and ambient temperature and relative humidity are presented in Table 2. There is no one-to-one correspondence between exhalation rates and

Location

Temp. (◦ C)

Humidity (%)

226 Ra

222 Rn

226 Ra

222 Rn decay products

concentration exhalation (Bq kg−1 ) (mBq m−2 s−1 )

concentration (Bq kg−3 ) (Bq kg−3 ) 218 Po 214 Bi

214 Pb

PAEC (mWL)

Coastal Karnataka Mangalore 28.0 Surathkal 25.8 Udupi 25.3 Kundapur 25.3 Honnavar 26.4 Karwar 25.1

80.2 79.0 80.8 83.6 69.8 76.0

46.0 24.0 74.3 11.2 7.0 22.7

89.0 34.1 150.2 42.7 10.6 28.9

10.5 9.0 27.7 13.5 9.2 13.0

7.1 8.4 17.2 6.7 9.7 10.4

3.8 5.4 11.7 6.2 5.4 6.4

3.1 4.9 8.3 8.0 5.7 5.5

1.0 1.4 2.9 1.8 1.5 1.6

Kaiga Range Mean Median

89.0 69.8–89.0 79.8 80.2

37.0 7.0–74.3 31.7 24.0

156.1 10.6–156.1 73.1 42.7

24.1 9.0–27.7 15.3 13.0

11.8 6.7–17.2 10.2 9.7

8.8 3.8–11.7 6.8 6.2

8.3 3.1–8.3 6.3 5.7

2.3 1.0–2.9 1.8 1.6

20.8 20.8–28.0 25.2 25.3

F

Annual effective dose (mSv y−1 ) DFp

DFu

0.35 0.58 0.39 0.49 0.60 0.45

0.198 0.298 0.568 0.305 0.271 0.307

0.331 0.284 0.874 0.426 0.290 0.410

0.35 0.35–0.6 0.46 0.45

0.469 0.20–0.57 0.34 0.31

0.760 0.28–0.87 0.48 0.41

Concentration, distribution and transportation of 222 Rn

Table 1 Diurnal averages of temperature, humidity, 222 Rn decay products, 226 Ra concentration in soil, 222 Rn exhalation rate, 222 Rn, equilibrium factor, annual effective dose due to inhalation of 222 Rn in the environment of coastal Karnataka and Kaiga

F – estimated equilibrium factor. DFP – dose estimated from the equilibrium factor of present work. DFU – dose estimated from the equilibrium factor suggested in [16].

545

546

H.M. Mahesh et al.

Fig. 2. Diurnal variation of 222 Rn exhalation rate with temperature and humidity in the environment of Udupi during December 23–24, 1998. Table 2 Cross-correlation between 222 Rn exhalation rate, relative humidity and temperature Sampling locations

Mangalore Surathkal Udupi Kundapur Honnavar Karwar Kaiga

No. of samples

24 23 23 22 22 23 22

Values of r Humidity

Temperature

0.48 −0.18 −0.22 0.19 −0.15 0.25 0.14

−0.39 0.09 0.18 −0.12 0.03 −0.30 −0.26

temperature and humidity. Similar trends in variations were observed at all sampling stations. Many investigators reported almost similar variations for other environs [13–15]. 3.2.

226 Ra

activity in soil

226 Ra activity in soil plays a major role in 222 Rn exhalation rates and therefore 226 Ra in surface

soils was measured in the corresponding sampling locations and the results are presented in Table 1. In coastal Karnataka, maximum activity of 226 Ra was observed in Udupi where the 222 Rn exhalation rate is signiﬁcantly higher. It was observed that though the observed 222 Rn exhalation rate is higher compared to other regions, the 226 Ra activity in Kaiga soil is at minimum. This is because Kaiga soil has higher values of emanation coefﬁcients due to the tropical evergreen forests [7,12]. A positive correlation was observed (r = 0.81, n = 7) between 222 Rn exhalation rate and 226 Ra in soil (Fig. 3). This indicates that, in general, 222 Rn exhalation rate depends on 226 Ra in addition to the meteorological parameters and local geology.

Concentration, distribution and transportation of 222 Rn

547

Fig. 3. Correlation between 226 Ra concentrations in soil and 222 Rn exhalation rate.

3.3.

222 Rn

concentration in the atmosphere

The results of 222 Rn concentration in air are presented in Table 1. Radon concentrations were found to be higher at Kaiga and Udupi than in other locations. It is interesting to note that 222 Rn exhalation rates measured in Udupi and Kaiga environs are also signiﬁcantly higher than at other sampling locations of the region. This may be one of the main reasons for the observed higher concentrations of 222 Rn in atmospheric air. The diurnal variations of 222 Rn concentrations vary by a factor of 3 to 20 in the different sampling locations. This range is slightly higher than the value (10) reported in UNSCEAR [16]. A typical diurnal variation of 222 Rn concentration with temperature and humidity is presented in Fig. 4. 222 Rn concentration in air reaches its maximum during early morning hours (3.00 to 5.00 hr) and thereafter it decreases. Similar trends of variations were observed for all sampling stations of the region. Similar variations have been reported earlier for different environs [13,17]. Detailed explanations for the diurnal variation in atmospheric radon can be found elsewhere [2].

Fig. 4. A typical diurnal variation of 222 Rn concentrations with temperature and humidity at Kundapur during December 24–25, 1998.

548

H.M. Mahesh et al. Table 3 Cross-correlation between atmospheric 222 Rn, relative humidity and temperature Sampling locations

Mangalore Surathkal Udupi Kundapur Honnavar Karwar Kaiga

No. of samples

08 11 09 11 08 10 07

Values of r Humidity

Temperature

0.56 0.86 0.87 0.87 0.89 0.57 0.82

−0.81 −0.88 −0.90 −0.82 −0.78 −0.76 −0.79

Cross-correlations between 222 Rn and ambient temperature and 222 Rn and humidity are presented in Table 3. The cross-correlation coefﬁcient between 222 Rn and relative humidity is positive conﬁrming the direct relationship between them. However, there was a negative correlation coefﬁcient between radon and ambient temperature indicating that the two parameters are incoherent. The relative humidity is usually associated with periods when wind speed is low and the atmosphere stable. The low wind speed and the atmospheric stability are probably the most important factors resulting in increased concentrations of radon. These results are analogous to those reported for the environment in Spain [14,18]. Positive correlation (r = 0.89, n = 7) between 222 Rn exhalation rate in soil and 222 Rn concentration in atmosphere can be observed from Fig. 5. Further, good correlation (r = 0.73, n = 7) was observed between 226 Ra in soil and the 222 Rn concentration in atmospheric air (Fig. 6). This suggests that the concentration of 222 Rn in the atmosphere depends on 226 Ra in the soil and 222 Rn exhalation rate. It can also be noted here that a good correlation was observed between 226 Ra in soil and 222 Rn exhalation rate.

Fig. 5. Correlation between 222 Rn exhalation rate and 222 Rn concentration in air.

Concentration, distribution and transportation of 222 Rn

549

Fig. 6. Correlation between 226 Ra concentration in soil and 222 Rn concentration in air.

3.4.

222 Rn

decay product concentrations in atmospheric air

Mean values of diurnal variations of concentrations of 218 Po, 214 Pb, 214 Bi and PAEC prevailing at various sampling stations of the region are presented in Table 1. The concentrations of all the three radon decay products and PAEC due to the above-mentioned radionuclides were maximum at Udupi and minimum at Mangalore. It can be noted here that higher concentrations of both 222 Rn exhalation rate and 222 Rn concentration in air were observed at Udupi. This may be the main reason for the observed higher concentrations of 222 Rn decay products at Udupi. A typical diurnal variation of 222 Rn decay products in the atmosphere is presented in Fig. 7. The atmospheric concentrations of 218 Po, 214 Pb, 214 Bi, and PAEC showed signiﬁcant diurnal variations of the order of 40, 8, 63, and 13 respectively. This is within the limits of the reported factor of 10 to 100 [19–22]. The concentration of 218 Po is higher compared to the other two decay products. This is due to the fact that, positively charged 218 Po, which is formed by the

Fig. 7. A typical diurnal variation of 222 Rn decay products and PAEC in air in the environment of Udupi during December 22–23, 1998.

550

H.M. Mahesh et al.

decay of 222 Rn, attaches immediately to the atmospheric aerosols contributing signiﬁcantly to its atmospheric concentration. The 222 Rn decay product concentrations were found to be minimum during 12.00 to 20.00 hr and maximum during 5.00 to 10.00 hr. Similar trends of diurnal variation were observed for all other sampling stations. Further, similar patterns of variations have also been observed and reported for the environment on Bombay, India [23,24], in USA [4] and in Germany [5,22]. The latter reported that great variations of 222 Rn decay products in the atmosphere during a single day are caused by changes in the eddy diffusivity in the boundary layer. The observed trend of diurnal variation is also consistent with the trends observed for 222 Rn exhalation rates and 222 Rn concentrations in atmospheric air. There exists a poor correlation (r = 0.63, n = 7) between 222 Rn exhalation rates and its decay products (Fig. 8). This suggests that 222 Rn atoms after exhalation attach to the aerosols and will migrate depending on the meteorological parameters (it is to be noted that both 222 Rn exhalation and 222 Rn decay products were sampled simultaneously at the sampling location).

Fig. 8. Correlation between 222 Rn exhalation rate and PAEC.

Fig. 9. Correlation between 222 Rn in air and PAEC.

Concentration, distribution and transportation of 222 Rn

551

However, it can be observed from Fig. 9 that there exists a good correlation (r = 0.93, n = 7) between 222 Rn and its decay products in the atmosphere of the study region. 3.5.

222 Rn

equilibrium factor

A typical diurnal variation of the 222 Rn equilibrium factor with ambient temperature and humidity is shown in Fig. 10. 222 Rn concentration reaches its maximum around 02.00 hr and minimum at 12.00 hr. In the case of PAEC, the concentration reaches maximum around 6.00 hr and minimum at 18.00 hr. But it is interesting to observe that the equilibrium factor remains constant during 18.00–02.00 hr. The main reason for the variation of equilibrium factor may be due to unattached decay products lost to the earth’s surface due to variations in the meteorological conditions [5,17]. The diurnal variation of equilibrium factors in atmospheric air is also reported for Mumbai [25], Taiwan [26], Japan [17] and Germany [5,22]. They have reported a general trend of variation of equilibrium factors being minimum during daytime and maximum during nighttime. The latter [22] reported the variation of equilibrium factor with meteorological parameters and observed the modest trend of increasing equilibrium factors with increasing atmospheric stability. The observed trend of variations in the present study is in agreement with this reported trend. The equilibrium factor varied in the range 0.35–0.63 with a mean of 0.45 (Table 1). Interestingly, the F varied by a factor of 4 to 5 at each sampling station. The estimated average value of 0.45 in the present study is slightly lower than the value (0.6) suggested in UNSCEAR [16]. The frequency distribution of outdoor 222 Rn equilibrium factors (Fig. 11) showed that the equilibrium factor lies in the range between 0.4 and 0.5. However, a Gaussian ﬁt gives a peak value of 0.42, which is close to the average 0.45. 3.6. Inhalation dose due to 222 Rn and its decay products The dose to the general public of coastal Karnataka and Kaiga regions due to the inhalation of and its decay products has been estimated. In order to estimate the inhalation dose, the 222 Rn concentrations, the equilibrium factors obtained in the present investigation and an exposure time of 8760 hr were considered. Comparison of the inhalation doses estimated by the 222 Rn

Fig. 10. A typical diurnal variation of the 222 Rn equilibrium factor in the environment of Kaiga during January 4–5, 1999.

552

H.M. Mahesh et al.

Fig. 11. Frequency distribution of the 222 Rn equilibrium factor in air.

equilibrium factors of the present investigation and the equilibrium factor (0.6) as suggested in UNSCEAR [16] is presented in Table 1. The inhalation dose estimated using the equilibrium factor suggested by UNSCEAR [16] gives higher values. This suggests the importance of measurement of individual 222 Rn concentrations and its decay products and of establishing the equilibrium factor for the region independently instead of considering the suggested values for the estimation of inhalation dose to the general public. Maximum inhalation dose was observed at Udupi and minimum at Mangalore. It can be recalled here that the 222 Rn exhalation rate, 222 Rn and its decay product concentrations in the atmosphere were higher in Udupi. Because of this, the dose was also found to be higher in Udupi. However, the overall annual inhalation dose observed in the study region was found to be in good agreement with the estimated global average value of 0.32 mSv y−1 (222 Rn = 10 Bq m−3 and for 8760 hr) resulting from the exposure to outdoor 222 Rn decay products. The variations in annual effective doses due to inhalation of radon and its decay products suggest the importance of establishing the equilibrium factors.

4. Conclusions • Maximum 222 Rn exhalation rates were observed corresponding to the highest atmospheric temperatures and minimum humidities, and vice versa. Correlation studies showed that there exists a good correlation between 222 Rn exhalation rate and 226 Ra in soil. These results suggest that the 222 Rn exhalation rate is supported by 226 Ra but is also affected by the meteorological parameters and local geology. • Correlation between 222 Rn exhalation rates and its decay products was found to be poor. However, there exists a good correlation between atmospheric 222 Rn and its decay products. • The 222 Rn equilibrium factors showed diurnal variations and varied by a factor of 4–5 at each sampling station. This indicates that not all the decay products are attached to the aerosols. However, the average value of the equilibrium factor, 0.5, is comparable with the literature values reported for other environs and also with the value recommended by UNSCEAR [16] for the outdoor 222 Rn equilibrium factor.

Concentration, distribution and transportation of 222 Rn

553

• The estimated dose to the general public due to inhalation of outdoor 222 Rn and its decay products using the equilibrium factor suggested by UNSCEAR [16] yielded higher values than the dose calculated from the equilibrium factors obtained here. This suggests the need and importance of measurement of individual 222 Rn concentrations and its decay products, and of establishing the equilibrium factor individually for a larger number of areas so that a more realistic equilibrium factor can be derived for global use. Acknowledgements The authors are grateful to Sri M. Raghavayya, Formerly Scientist, RMP, Mysore and to Dr S. Sadasivan, Former Head, Environmental Assessment Division, BARC, Mumbai for their useful suggestions. The authors thank the Board of Studies in Nuclear Science, Department of Atomic Energy, Government of India for sponsoring the research project under which this work has been carried out. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26]

UNSCEAR, Forty-second session of UNSCEAR, A/AC, 82/R. 526, United Nations, New York, 1993. F.T. Gessel, Health Phys. 45-2 (1983) 289. E. Stranden, A.K. Kolstad, B. Lind, Health Phys. 47 (1984) 480. M. Wilkening, Radon in the environment, in: Studies in Environmental Science, vol. 40, Elsevier, 1990. J.J. Porstendörfer, J. Aerosol Sci. 25 (1994) 219. K.G. Vohra, U.C. Mishra, K.C. Pillai, S. Sadasivan, (eds.) Wiley Eastern Ltd. New Delhi, 1982, p. 584. H.M. Somashekarappa, Y. Narayana, A.P. Radhakrishna, K. Siddappa, V.B. Joshi, R.V. Kholekar, A.M. Bhagwat, Radiat. Measur. 26 (1996) 35. N. Karunakara, H.M. Somashekarappa, D.N. Avadhani, H.M. Mahesh, Y. Narayana, K. Siddappa, Health Phys. 80 (2001) 470. G.K. Srivastava, M. Raghavayya, A.H. Khan, Health Phys. 46 (1984) 225. J.W. Thomas, P.C. LeClare, Health Phys. 18 (1970) 113. M. Raghavayya, Radiat. Prot. Environ. 4 (2001). N. Karunakara, Studies on radionuclide distribution and uptake in the environment of Kaiga, PhD thesis, Mangalore University, Mangalagangotri, India, 1997. 5th International Symposium on the Natural Radiation Environment, CEC Report EUR 14411 EN, Commission of the European Communities, Luxembourg, 1993, ISSN 1018-5593. C. Duenas, M. Perez, M.C. Fernandez, J. Carretero, J. Environ. Radioactivity 31 (1996) 87. H. Kojima, N. Nagano, in: Proceedings of Workshop on Radon in the Living Environment, Athens, 2000. UNSCEAR, Exposures from natural radiation sources, Annex A in: Sources and Effects of Ionizing Radiation, UNSCEAR 2000 Report to the General Assembly, with Scientiﬁc Annexes, United Nations, New York, 2000. H. Kojima, Environ. Int. 22 (1996) S187. C. Duenas, M.C. Fernandez, M. Senciales, Atmos. Environ. 24 (1990) 1255. W. Jacobi, K. Andre, J. Geophys. Res. 68 (1963) 3799. R.L. Grasty, Health Phys. 66 (1994) 185. M.C. Robe, A. Rannou, Le Bronnec, Radiat. Prot. Dosim. 45 (1992) 455. S. Raviart, P. Richon, D. Haristoy, M.C. Robe, Y. Belot, M. Kummel, C. Dushe, W. Ullman, Environ. Int. 22 (1996) S279. T.F. Gesell, H.M. Prichard, in: The Natural Radiation Environment III, DOE CONF 780422, National Technical Information Service, Springﬁeld, VA, 1980, p. 327. C. Rangarajan, Smt.S. Gopalakrishnan, C.D. Eappen, Pure Appl. Geophys. 112 (1974) 941. M.C. Subbaramu, K.G. Vohra, Tellus XXI (1968) 395. C. Jiang Chen, P. Shan Weng, T. Chi Chu, Health Phys. 64 (1993) 74.

554

Natural radiation levels in Tamil Nadu and Kerala, India S. Tokonami a , H. Yonehara a , S. Akiba b , M.V. Thampi c , W. Zhuo a , Y. Narazaki d , Y. Yamada a a Radon Research Group, National Institute of Radiological Sciences, 4-9-1 Anagawa, Inage,

Chiba 263-8555, Japan b Department of Public Health, Kagoshima University, 8-35-1 Sakuragaoka, Kagoshima 890-8520, Japan c Low Level Radiation Research Laboratory, Bio-Medical Group, Bhabha Atomic Research Centre, IRE Campus,

Beach Road, Quilon, Kerala 691 001, India d Fukuoka Institute of Health and Environmental Sciences, 39 Mukaizano, Dazaihu, Fukuoka 818-0135, Japan

Natural radiation measurements were preliminarily carried out in high background areas, in India, so as to understand the biological effects on human beings due to natural radiation exposures. The following parameters were taken into account for dose assessment: (1) 222 Rn (radon) concentration, (2) 220 Rn (thoron) concentration, (3) equilibrium equivalent thoron concentration (EETC), and (4) indoor/outdoor gamma dose rates. These measurements were made at 15 sites in Tamil Nadu and 5 sites in Kerala, India. The sites include houses and schools. The radon concentration, thoron concentration and EETCs ranged between 2–70, 6–690 and 0.1–1.6 Bq m−3 , respectively. The indoor and outdoor gamma dose rates ranged from 0.3 to 3.9 and from 0.4 to 6.2 μGy h−1 for radon and thoron, respectively. After classifying the dose by exposure, the annual effective dose at each site for radon and thoron was calculated with several assumptions. From the sampled data, the annual effective dose ranged from 4 to 22 mSv with an arithmetic mean of 9.3 mSv and the dose contribution was signiﬁcantly due to external exposure.

1. Introduction It is well known that high natural radiation areas are located in the southwest coast of the Indian peninsula [1]. Such situations of high radiation levels result from the presence of heavy minerals containing monazite, which have high levels of thorium. The monazite sands are RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07066-4

© 2005 Elsevier Ltd. All rights reserved.

Natural radiation levels in Tamil Nadu and Kerala, India

555

speciﬁcally distributed along several beaches, where people are mining out. In addition, many people are living near those beaches, and are exposed to natural radiation sources. In order to understand the biological effect due to natural radiation, information on the relevant dosimetric quantities is indispensable. If accurate dose assessment is successful, people in these areas will be suitable for cases and controls in an epidemiological study. Although several studies have been carried out already [2–4], natural radiation measurements using our up-to-date instruments were preliminarily but systematically carried out in Tamil Nadu and Kerala, India in the present study. The following parameters were taken into account for dose assessment: radon, thoron and thoron decay products as internal exposures and gamma dose rates indoors and outdoors as external exposures. The present paper describes the preliminary results of the radiation survey program.

2. Materials and methods In total, twenty houses or schools were chosen in our radiation survey program: 15 sites in Tamil Nadu (Manavalakurichy) and 5 sites in Kerala (Needendakara), India. For the determination of radon and thoron concentrations, two types of alpha-track detector (CR-39) were used. One detector was designed to detect radon effectively. It is called RADOPOT, is made in Hungary and is commercially available. The other detector is a modiﬁed RADOPOT. The air exchange rate of the detector has been enhanced so as to detect thoron as well as radon. Using two readings, both radon and thoron concentrations can be evaluated. On the other hand, a deposition rate measurement of the thoron decay products was applied for the EETC determination [5]. A CR-39 detector was incorporated into the measuring system. The CR-39 detector was covered with an energy absorber such as aluminium-evaporated Mylar® ﬁlm to detect alpha particles emitted from 212 Po (8.78 MeV) only. The thickness of the ﬁlm was set to about 71.5 mm as an air-equivalent value. The EETC (Bq m−3 ) can be given by the following equation: EETC = 0.87 × N/T

(1) cm−2 )

where N is the track density (tracks and T is the exposure period (day). Note that the above three devices measured indoor concentrations and no measurements were made outdoors in the present study. These three devices were placed indoors for 3–5 months. Information on housing structure, exposure period and position of the device is tabulated in Table 1. The gamma dose rate is measured at 1 m height above ground with a 1 × 2 NaI (Tl) scintillation spectrometer (commercial name: SS-γ) indoors and outdoors. The energy resolution of the scintillation spectrometer is 11% for the 662 keV gammas of 137 Cs. Readings on the spectrometer were corrected with another well-calibrated instrument with an empirical equation [6]. Soil samples were taken in mined beaches to measure their radioactivities. They were determined by gamma spectrometry using a pure Ge detector. A HPGe detector manufactured by ORTEC was used. The relative efﬁciency and FWHM resolution were 36% and 1.76 keV, respectively, for the 1.33 MeV gammas of 60 Co. Counting time for each sample was generally 8 × 104 s or more.

556

S. Tokonami et al.

Table 1 Information on measurement sites Site ID

Exposure period beginning, end

Housing structure (roof/wall/ﬂoor)

Positions∗ of the device location and remarks Gas & DP (both): on the table. Gas: 30 cm, DP: 0 cm, H = 2.5 m Both: 0 cm, H = 2 m Gas: 20 cm, DP: 0 cm. H = 2 m; Near mining beach Both: 0 cm, H = 2 m; Near mining beach Both: 0 cm, H = 2 m Both: 0 cm, H = 2 m; Near mining beach Both: 0 cm, H = 2 m Both: 0 cm, H = 2 m Both: 0 cm, H = 2 m; School Both: 0 cm, H = 2.2 m Middle school Both: 0 cm, H = 2.2 m; Secondary school Both: 0 cm, H = 2.2 m; High school Both: 0 cm, H = 2.5 m; Elementary school Both: 0 cm, H = 2.2 m Both: 0 cm, Height = 2 m for gas, H = 1.9 m for DP Gas: center of room, DP: 0 cm, H = 2 m for gas, Height = 1.9 m for DP Both from ceiling, Height = 2 m; Soil sample taken Both: 0 cm, H = 1.9 m Both: 0 cm, H = 2.1 m

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

2/18/01, 7/24/01 2/19/01, 7/25/01 2/19/01, 7/25/01 2/19/01, 7/24/01 2/19/01, 7/24/01 2/19/01, 7/25/01 2/19/01, 7/24/01 2/19/01, 7/24/01 2/19/01, 7/24/01 2/19/01, 7/25/01 2/20/01, 7/25/01 2/20/01, 7/25/01 2/20/01, 7/25/01 2/20/01, 7/25/01 2/20/01, 7/25/01 2/23/01, 5/24/01 2/23/01, 5/24/01

Concrete/Granite/Cement Concrete/Brick/Cement Tile/Brick/Cement Concrete/Brick/Cement Tile/Brick/Cement Tile/Brick/Cement Tile/Brick/Cement Concrete/Brick/Cement Tile/Brick/Cement Concrete/Brick/Cement Concrete/Brick/Cement Concrete/Brick/Cement Concrete/Brick/Cement Concrete/Brick/Cement Thatch/Coconut leaves/Sand Concrete/Brick/Cement Asbestos/Wood/Cement

18 19 20

2/23/01, 5/24/01 N.D. 2/23/01, 5/24/01

Asbestos/Brick/Cement Concrete/Brick/Cement Asbestos/Wood/Cement

∗ In most cases, the distance from wall and height (H ) from the ﬂoor.

3. Results and discussion Results of the radiation survey are listed in Table 2. Only the data from one site (No. 19) was invalid because the detector was lost. The radon concentration ranged from 2 to 70 Bq m−3 , and an arithmetic mean (AM) was estimated to be 17 Bq m−3 , as low as expected due to open housing structures. On the other hand, thoron concentration ranged widely from 6 to 690 Bq m−3 , and the AM was 168 Bq m−3 . Our thoron detectors were generally suspended on walls and were close to the wall. It is well known that the thoron concentration varies greatly with distance from the source. However, our thoron gas measurements were justiﬁed on the following grounds: The result of thoron concentrations would be useful for dose assessment from thoron gas itself due to inhalation because the position of the device was clariﬁed at the site. Since there is little information on thoron in the environment as the UNSCEAR 2000 report has pointed out, such data should also be accumulated wherever possible. From the thoron decay product measurements using their deposition rate measurements, the AM of EETC was estimated to be 0.5 Bq m−3 , which was fairly low (range: 0.1–1.6 Bq m−3 ). The reason can be also explained as being because of open housing structures, as for radon. If an equilibrium factor (F ) between thoron and thoron decay products is evaluated using two relevant data, though meaningless in a sense, an average F is obtained as 0.007. Since the thoron concentration varies greatly in space, the equilibrium factor will subsequently change according to the position of those devices. Using the data of Table 2, an average thoron to

Natural radiation levels in Tamil Nadu and Kerala, India

557

Table 2 Results of the radiation survey Site ID

Radon (Bq m−3 )

Thoron (Bq m−3 )

EETC (Bq m−3 )

Indoor dose rate (μGy h−1 )

Outdoor dose rate (μGy h−1 )

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

56 3 8 9 33 20 19 10 19 8 7 5 12 70 2 9 18 9 N.D. 5

45 16 143 13 690 411 480 31 441 38 83 34 94 480 76 30 6 67 N.D. 17

0.4 0.2 0.6 0.2 1.6 0.8 0.9 0.1 0.7 0.2 0.4 0.4 0.2 0.9 0.3 0.1 0.03 1.4 N.D. 0.5

1.7 1.3 0.6 3.9 2.3 0.7 1.3 0.4 0.8 0.8 0.9 0.4 0.4 0.9 1.6 0.3 1.1 3.8 0.8 1.8

1.1 1.71 0.4 2.5 1.8 0.4 0.5 0.6 0.4 0.9 2.0 1.6 1.7 1.5 2.8 2.4 2.0 6.2 1.6 4.6

radon ratio was estimated to be 11.0. Special attention should be paid to selection of the device when radon measurements are accurately made [7]. The indoor and outdoor gamma dose rates ranged from 0.3 to 3.9 and 0.4 to 6.2 μGy h−1 , respectively. Their AM were estimated to be 1.3 and 1.8 μGy h−1 , respectively. The indoor to outdoor ratios ranged from 0.1 to 2.4, with an AM of 0.97. As far as the sampled data are concerned, there is little correlation amongst them and they were independent. When an epidemiological study has to be carried out, individual doses should be considered together with general information on their living activities. Natural radioactivities in soil at some locations (four samples on the mining beach in Tamil Nadu and one sample in Kerala, India) were measured by gamma-ray spectrometry using the Ge detector. In the site where the highest gamma doses were given, 238 U and 232 Th activities in soil were 2.4 and 12.0 kBq kg−1 , respectively, with some assumptions of radioactive equilibrium between those nuclides and their decay products. Regarding the activities in beach sands, they are widely distributed in the range of 0.03–7.0 kBq kg−1 for 238 U and 0.14–50.8 kBq kg−1 for 232 Th, respectively, even on the same beach. To understand the variation of individual doses among the sampled data, the preliminary dose assessment was made with our limited data based on the UNSCEAR 2000 Report approach. Several assumptions were made as follows: outdoor radon concentration was equal to 2 Bq m−3 because all the sites were near beaches where the outdoor radon concentration seems to be low; the equilibrium factors indoors and outdoors were given as 0.4 and 0.6, respectively, because no measurements were made but they were reasonable values after taking their surrounding conditions into account; outdoor thoron concentrations and EETCs were equal to 10 and 0.1 Bq m−3 , respectively, because there were no means to assign these values

558

S. Tokonami et al. Table 3 Summary of the annual effective dose (mSv y−1 ) under several assumptions based on the UNSCEAR 2000 Report approach Items

Radon and its decay products

Thoron and its decay products

Internal dose (subtotal)

External dose

Total

Min Max Mean

0.07 1.9 0.5

0.02 1.0 0.3

0.2 2.5 0.8

3.7 21.6 8.6

4.0 21.9 9.3

at the moment. Occupancy factors were assumed to be the same at all the sites, as for outdoor gamma dose rates, a mean value of 1.8 μGy h−1 being assigned. After classifying the exposure doses, the annual effective dose at each site was calculated via the several assumptions mentioned above. The results are summarised in Table 3. When the dose from radon and its decay products is compared with that from thoron and its decay products, the dose from the former is more signiﬁcant than that from thoron as long as the dose was assessed based on the UNSCEAR 2000 Report methodology. From the sampled data, the annual effective dose ranges from 4 to 22 mSv with an AM of 9.3 mSv and the dose contribution is signiﬁcantly due to the external exposure. The external dose accounts for 90% of the total dose in the present study.

4. Conclusion There are a lot of problems to be solved in order to understand the biological effects due to very low-dose radiations. If an epidemiological study is effectively conducted with a small population in the future, high background radiation areas will be suitable for the purpose. The present radiation survey was preliminarily conducted so as to understand the relevant dosimetric quantities taking this future project into account. There was little correlation amongst the data and they seemed to be independent. The most signiﬁcant dose arises from gamma radiation because the other dosimetric quantities are much less than the gamma dose. Since there are no epidemiological data on thoron exposure, many problems remain unsolved. The dose assessment for thoron and its decay products has not yet been established as well as that for radon. It can be pointed out that the dose contribution from thoron gas itself should be considered because residents lie almost directly on/near the ground/ﬂoor (main source of thoron) when they sleep. Note that there is a large variation in radiation dose even in such high background radiation areas. Therefore, individual doses should be accurately evaluated in order to understand the biological effects on human beings of natural radiation exposures.

References [1] Sources and Effects of Ionizing Radiation, UNSCEAR 2000 Report to the General Assembly, with Scientiﬁc Annexes, United Nations, New York, 2000. [2] A.C. Paul, et al., in: Proc. 2nd National Symposium on Environment, 1993, p. 33. [3] A.C. Paul, et al., J. Environ. Radioact. 22 (1994) 243.

Natural radiation levels in Tamil Nadu and Kerala, India [4] [5] [6] [7]

A.C. Paul, et al., J. Environ. Radioact. 40 (1998) 251. W. Zhuo, T. Iida, Jpn. Health Phys. 35 (2000) 365. M. Furukawa, S. Tokonami, Jpn. Health Phys. 36 (2001) 195. S. Tokonami, et al., Health Phys. 80 (2001) 612.

559

560

Natural radiation exposures for cave residents in China S. Tokonami a , Q. Sun b , S. Akiba c , T. Ishikawa a , M. Furukawa a , W. Zhuo a , C. Hou b , S. Zhang b , Y. Narazaki d , H. Yonehara a , Y. Yamada a a Radon Research Group, National Institute of Radiological Sciences, 4-9-1 Anagawa, Inage,

Chiba 263-8555, Japan b Department of Biology and Medicine, Laboratory of Industrial Hygiene, Ministry of Health, PRC,

2 Xiankang St., Deshengmenwai, PO Box 8018, Beijing 100088, China c Department of Public Health, Faculty of Medicine, Kagoshima University, 8-35-1 Sakuragaoka,

Kagoshima 890-8520, Japan d Fukuoka Institute of Health and Environmental Sciences, 39 Mukaizano, Dazaihu, Fukuoka 818-0135, Japan

In order to understand the biological effects on human beings due to natural radiation exposures as a ﬁnal goal, the following radiation measurements were preliminarily carried out in cave dwellings and their surrounding areas in China (Shanxi and Shaanxi provinces): (1) indoor and outdoor gamma doses and (2) radon, thoron and their decay product concentrations. 202 cave dwellings were subject to investigation. The gamma dose rates were measured with a compact gamma dose meter with a semiconductor detector at all the sites. Mean gamma dose rates indoors and outdoors were 150 and 110 nSv h−1 , respectively. They were also measured with a 1 × 2 NaI scintillation spectrometer at 32 dwellings. The indoor and outdoor gamma dose rates ranged from 121 to 182 nSv h−1 indoors and 99 to 142 nSv h−1 outdoors, respectively. After making some corrections on conversion, the arithmetic means (AM) were eventually estimated to be 145 and 107 nGy h−1 , respectively. After only four dwellings were rejected due to invalid data, statistical analyses have shown indoor radon concentrations ranging from 18 to 224 Bq m−3 with an AM of 64 Bq m−3 , indoor thoron concentrations ranging from 8 to 1176 Bq m−3 with an AM of 219 Bq m−3 and the indoor equilibrium equivalent thoron concentration (ETTC) ranging from 0.05 to 7.6 Bq m−3 with an AM of 1.8 Bq m−3 , respectively. The present survey has revealed the two following points: (1) gamma dose rates and radioactivities in soil were homogeneous in these areas; (2) the presence of thoron cannot be negligible from the viewpoint of accurate dose assessment. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07067-6

© 2005 Elsevier Ltd. All rights reserved.

Natural radiation exposures for cave residents in China

561

1. Introduction Radon and its decay products in air are the most important contributors to human exposure from natural sources though dose assessment due to radon decay product inhalation has not yet been established [1]. For the time being, two approaches were proposed but both do not agree well. There is a three-time difference between both of them. One approach was based on epidemiological study. Most epidemiological data were related to radon exposures in mines. The second approach was by dosimetric study, in which the dose was derived from laboratory investigation. The greatest interest is how the information gap will be ﬁlled. Many studies on underground miners have shown that the radon exposure increases the lung cancer risk [2]. In particular, signiﬁcant risks have been observed for miners exposed to low levels and receiving cumulative exposures comparable to those obtained by long-term residence with high radon levels. Therefore, it is necessary to demonstrate the lung cancer risk from residential radon exposure. Several case–control studies on residential radon have been carried out. Although some of them have found no risk with indoor radon exposure, others are consistent with increasing risk with increasing indoor exposure. It is easy to imagine that the environmental conditions in mines and homes are quite different. Although the presence of thoron can be negligible in mines, it should be considered if the radon level is low. Unfortunately, there is no epidemiological data for lung cancer risk following thoron exposure. Prior to conducting a case–control study on residential radon exposure, it is indispensable to understand what is present and how much is in the study area. There are many cave dwellings where the radon concentration seems to be high in the Chinese loess plateau. Previous studies in those areas can be referenced [3–5]. Since the residential mobility is low, this area will be suitable for conducting a case–control study on lung cancer risk and residential radon exposure. Within our careful strategy, natural radiation measurements were preliminarily made in Shanxi and Shaanxi provinces as the ﬁrst step. The present study demonstrates the characteristics of natural radiation exposures for cave residents. Note that a case–control study in Gansu province has been carried out by the National Cancer Institute (NCI) in USA, and has concluded that the lung cancer risk increased with increasing radon level [6]. Their study area is also located on the Chinese loess plateau. 2. Materials and methods The present survey was conducted in Luliang, Shanxi and in Yan’an, Shaanxi provinces. The following items were assigned to be measured: (1) (2) (3) (4) (5) (6)

radon and thoron concentrations for a long-term period; thoron decay product concentrations for a long-term period; individual radon decay product concentrations by grab sampling; thoron decay product concentrations by grab sampling; radon concentrations for a short-term period; gamma dose rates indoors and outdoors.

For the long-term measurements, passive radon and thoron monitors paired (CR-39) were used in the survey. The thoron decay product concentrations were estimated by their deposition rate measurement, as developed by Zhuo and Iida [7]. Those passive devices were placed

562

S. Tokonami et al.

in 202 caves (101 caves in each province) for a 6-month exposure (August 2001 through February 2002). Since these devices were generally suspended on the wall or ceiling in the middle of the cave, the distances from the wall and the ground were estimated to be 5–30 and 150– 300 cm, respectively. Grab sampling and measurements were made by ourselves during the observation period. The individual radon decay product concentrations were determined by the Tremblay method [8]. The thoron decay product concentrations were as estimated by grab sampling and the alpha-track registration technique using CR-39. The detailed information is available in Tokonami et al. [9]. The radon concentration was continuously measured with the “AlphaGUARD” in some cave dwellings. The device was set to the diffusion mode with a 60min time interval. The sensitivity is 1 cpm at 20 Bq m−3 . Gamma dose rates were measured using an electronic pocket dosimeter (PDR-101 manufactured by Aloka Co., Ltd., ﬂuctuation coefﬁcient: 5.0, measurable energy: 60 keV–3 MeV, calibration coefﬁcient for standard gamma radiation for 1 MeV gammas: 0.87 ± 0.01 Gy Sv−1 in air) in all the cave dwellings and a 1 × 2 NaI(Tl) scintillation spectrometer (energy resolution: 11% for 662 keV gammas of 137 Cs) at some sites. In addition to the above measurements, several interview tests on their lifestyles/habits were carried out for the coming study.

3. Results and discussion In total, 21 data on individual radon decay product concentrations were obtained with the equilibrium equivalent radon concentration (EERC) ranging from 0.7 to 94.4 Bq m−3 with an AM of 17.8 Bq m−3 . Although only three data on the unattached fraction (fp ) were available, the unattached fraction was estimated to be 0.023, 0.033 and 0.035, and was lower than that in the general indoor environment. In some caves, the radon concentration varied greatly in time in the range 40 to 320 Bq m−3 with an AM of 188 Bq m−3 . The concentration level was high during the night and low in the daytime. Figure 1 exempliﬁes the time variation of radon concentration with the AlphaGUARD in a cave dwelling. Thirty two air grab samples were taken so as to measure thoron decay product concentrations with the alpha-track registration

Fig. 1. Time variation of radon concentration with the AlphaGUARD in a cave dwelling.

Natural radiation exposures for cave residents in China

563

Table 1 Results of radon, thoron and thoron progeny concentrations in the present survey Items

Number of dwelling

Shanxi

98

Shaanxi

100

All

198

Radon (Bq m−3 )

Thoron (Bq m−3 )

EETC (Bq m−3 )

52 (18–199) 76 (22–224) 64 (18–224)

182 (8–609) 255 (17–1176) 219 (8–1176)

1.4 (0.15–4.2) 2.2 (0.05–7.6) 1.8 (0.05–7.6)

Result as the arithmetic mean. Range in parentheses.

technique on CR-39. The 212 Pb concentration consequently ranged from 0.4 to 3.6 Bq m−3 with an AM of 1.4 Bq m−3 . Based on the long-term measurements, Table 1 shows the results of the present survey. Only four cave data were invalid among all the data. Thus, 98 cave data in Shanxi and 100 cave data in Shaanxi were available. The statistical analyses on all the data have shown that the indoor radon concentration ranged from 18 to 224 Bq m−3 with an AM of 64 Bq m−3 , the indoor thoron concentration ranged from 8 to 1176 Bq m−3 with an AM of 219 Bq m−3 and the indoor equilibrium equivalent thoron concentration (ETTC) ranged from 0.05 to 7.6 Bq m−3 with an AM of 1.8 Bq m−3 , respectively. These values are consistent with the results of the activity measurements mentioned above. In Shanxi province, arithmetic means of radon, thoron and thoron decay products concentrations were 52, 182 and 1.4 Bq m−3 , respectively. In Shaanxi province, they were 76, 255 and 2.2 Bq m−3 , that is higher than those in Shanxi. The reason can be explained by the ratio of housing structure. In our survey, about 90% of houses consist of loess cave dwellings in Shaanxi. However, 50% of houses were stone cave dwellings with low radon level (37 Bq m−3 with 52 data) in Shanxi. In the loess cave dwellings, the mean radon concentrations were almost the same (74–80 Bq m−3 ) in the two provinces. Such cave dwellings provided the highest radon concentration in the present survey. Surprisingly, the presence of relatively high thoron was conﬁrmed in all the sites. In fact, most of devices were placed near the wall or ceiling. This implies that special attention should be paid to the location of radon monitors when single use of the device is made. Figures 2–4 illustrate the frequency distributions of radon concentration, thoron concentration and EETC. It appears that all these concentrations obey the log-normal distribution. Figures 5 and 6 show correlations amongst the three concentrations. As far as the data are concerned, each concentration seems to be independent. Even the relationship between the thoron concentration and EETC has a weak positive correlation. Table 2 summarizes a comparison of the present survey with other studies. The NCI study area is close to our study area and is also located on the Chinese loess plateau. The Yan’an data in the present study have shown relatively high thoron concentrations and were similar to the previous study conducted by Wiegand et al. [5]. On the other hand, only the NCI study showed high radon concentrations, compared to two other studies. From the EETC measurements, the contribution from thoron decay products will not be signiﬁcant. However, the dose contribution ratio should be clear for accurate dose assessment if a lung cancer study is conducted. The dose from thoron gas itself should also be

564

S. Tokonami et al.

Fig. 2. Frequency distribution of the radon concentration in the present survey.

Fig. 3. Frequency distribution of the thoron concentration in the present survey.

Fig. 4. Frequency distribution of the EETC in the present survey.

Natural radiation exposures for cave residents in China

565

Fig. 5. Correlation between radon and thoron concentrations.

Fig. 6. Correlation between thoron concentration and EETC. Table 2 Comparison of the survey result with other studies Items

NCI, 2002

Wiegand et al. [5]

NIRS (Present study)

NIRS (Present study)

Study area Province Radon (Bq m−3 ) Thoron (Bq m−3 ) EETC (Bq m−3 )

Pingliang, Qingyang Gansu 223 None None

Yan’an Shaanxi 76 255 2.2

Luliang Shanxi 52 182 1.4

Excess odds ratio (Lung cancer risk)

0.19 at 100 Bq m−3 (95% CI: 0.05, 0.47)

Yan’an Shaanxi 92∗ 215∗ 21.5† (F of 0.1 used) None

To be studied

To be studied

∗ Median. † An equilibrium factor of 0.1 was assigned to obtain the thoron decay product concentration.

566

S. Tokonami et al.

considered because residents lie directly on the ground when they sleep. Mean gamma dose rates indoors and outdoors were 150 and 110 nSv h−1 , respectively, at all the sites. They were also measured with a 1 × 2 NaI scintillation spectrometer at 32 dwellings. The indoor and outdoor gamma dose rates ranged from 121 to 182 nSv h−1 indoors and 99 to 142 nSv h−1 outdoors, respectively. These readings in nSv h−1 were corrected with another well-calibrated instrument using an empirical equation [10]. The AMs were eventually estimated to be 145 and 107 nGy h−1 , respectively. Soil samples were taken at 12 sites for gamma spectrometry with a Ge detector. The radioactivities of 238 U, 232 Th and 40 K in loess soils ranged from 30 to 37, 41 to 47 and 578 to 670 Bq kg−1 with AMs of 34, 44 and 614 Bq kg−1 , respectively. The soil radioactivity was homogeneously distributed on the Chinese loess plateau.

4. Conclusion There are many cave dwellings where the radon concentration seems to be high on the Chinese loess plateau. Since the residential mobility is low, this area will be suitable for conducting a case–control study on lung cancer risk and residential radon exposure. According to our careful strategy, natural radiation measurements were preliminarily made in Shanxi and Shaanxi provinces. The case–control study in Gansu province conducted by NCI has shown that the lung cancer risk increased with increasing radon level. According to survey results from the present study, however, the presence of thoron cannot be negligible. In fact, the radon concentration was much lower than that given by NCI. It seems that there are several problems to be solved and new ﬁndings to be of interest. After adding the results of another 6-months exposure, all these data will be discussed in detail.

References [1] UNSCEAR, Sources and Effects of Ionising Radiation, UNSCEAR 2000 Report to the General Assembly, with Scientiﬁc Annexes, United Nations, New York, 2000. [2] Committee on Health Risks of Exposure to Radon, Board on Radiation Effects Research, Commission on Life Sciences, National Research Council, Health Effects of Exposure to Radon: BEIR VI, National Academy Press, Washington, DC, 1999. [3] Z. Wang, et al., Health Phys. 70 (1996) 192. [4] B. Shang, et al., Radon and Thoron in the Human Environment, World Scientiﬁc, Singapore, 1998. [5] J. Wiegand, et al., Health Phys. 78 (2000) 438. [6] Z. Wang, et al., Am. J. Epidemiol. 155 (2002) 554. [7] W. Zhuo, T. Iida, Jpn. Health Phys. 35 (2000) 365. [8] R. Tremblay, et al., Health Phys. 36 (1979) 401. [9] S. Tokonami, et al., Jpn. Health Phys., in press. [10] M. Furukawa, S. Tokonami, Jpn. Health Phys. 36 (2001) 195.

567

A study of atmospheric radon transport as a tracer of pollutants over the Japan Sea H. Aoshima a , Y. Hashiguchi a , J. Moriizumi a , K. Yoshioka b , Y.S. Kim c , T. Iida a a Department of Nuclear Engineering, Graduate School of Engineering, Nagoya University, Fro-cho, Chikusa-ku,

Nagoya 464-8603, Japan b Shimane Prefectural Institute of Environmental Science, 582-1, Nishihamasada, Matsue,

shimane Pref. 690-0122, Japan c Institute of Environmental and Industrial Medicine Hanyang University, 17 Haengdang-dong, Seongdong-Gu,

Seoul, 133-791, South Korea

The ratio of the air pollutants generated on the Chinese continent and transported to Japan across the Sea of Japan was estimated quantitatively by the 222 Rn concentration changes at Tonghae (South Korea) and Oki (Japan). It was found that diurnal variations were clearer at Tonghae than Oki. By comparing the baseline trends, which are day-to-day trends of daily minima at these two stations, pollutants in the continental air mass were found to be diluted to 40% during their transport over the Sea of Japan in the winter season.

1. Introduction and measurement In some studies of air transportation, 222 Rn is effectively used as a tracer [1,2]. In recent years, atmospheric concentrations of 222 Rn have been measured in combination with aerosol observations in the framework of the project titled the East Asia and Northwestern Paciﬁc Ocean (“ACE-Asia”). This was decided because 222 Rn is a noble gas and chemically and physically inert. In addition, since its half-life of 3.82 days is comparatively long, 222 Rn is advantageous for such observations over large areas including continents and oceans. In this study, 222 Rn is used as a tracer of air mass transport in order to evaluate quantitatively the ratio of transport. Measurements of 222 Rn concentrations have been carried out at Seoul and Tonghae (South Korea), Beijing (China), Oki, Kanazawa, Nagoya and Hachijo Island (Japan) to demonstrate the advantages of using 222 Rn as a natural atmospheric tracer. Furthermore, 222 Rn concentrations at Tonghae and Oki during winter were analyzed to evaluate air mass transport over the Sea of Japan. These locations are shown in Fig. 1. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07068-8

© 2005 Elsevier Ltd. All rights reserved.

568

H. Aoshima et al.

Fig. 1. Measuring locations across the Sea of Japan.

Fig. 2. 222 Rn detector system.

The 222 Rn detector system is shown in Fig. 2. 222 Radon decays by emission of alpha particles to 218 Po, and 218 Po atoms are positively charged [3]. The positive 218 Po ions are collected electrostatically on an electrode of aluminized Mylar coated with ZnS(Ag) scintillator [2,4]. Part of the alpha particles emitted from 218 Po and 214 Po atoms on the electrode are incident on the underlying ZnS(Ag) layer through the Al Mylar. The scintillations due to the alpha particles are detected by the photomultiplier tube. The scintillation pulse, which is ampliﬁed and processed, is fed into a personal computer. The computer automatically calculates the 222 Rn concentration from the alpha-counts accumulated every hour. Atmospheric concentrations of 222 Rn on the ground show diurnal variation. The concentration is low in the daytime when 222 Rn near the ground is transported upwards by vertical mixing. On the other hand, the concentration becomes high at night. There are two components of the actual measured 222 Rn concentrations at Oki, one being the local component, which is strongly affected by nearby sources and the other the continental component. The diurnal variation in 222 Rn concentrations is mainly caused by the local component. In this study, only the 222 Rn from the continent of China is taken into consideration. In general, it is easier to measure the continental component on a small island in an ocean inasmuch as the local component has a small contribution. The steps mentioned below were followed in order to evaluate the continental components at Tonghae and Oki. 1. The 222 Rn concentrations at Oki were corrected for radioactive decay during an estimated travel time of 24 h, between the two locations.

A study of atmospheric radon transport as a tracer of pollutants over the Japan Sea

569

Fig. 3. The “baselines” (Bq m−3 ) of 222 Rn concentrations at Tonghae and Oki in January 2001.

2. Daily minimum concentrations were attributed to the continental component. The temporal variations in the daily minima are called “baselines” in this paper. Figure 3 illustrates examples of the baselines at Tonghae and Oki.

2. Meteorological conditions During the Japanese winter, a northwestern monsoon blows constantly. It was checked that the back trajectory of air mass transport during the investigated season had passed through Oki from the Korean Peninsula. The wind direction and wind speed at Oki in January 2001 are shown in Table 1. The northwest to west wind blew throughout the month. The average Table 1 Daily mean of wind direction and speed at Oki in January 2001 Day

Direction Speed (m s−1 )

Day

Direction Speed (m s−1 )

1-Jan. 2-Jan. 3-Jan. 4-Jan. 5-Jan. 6-Jan. 7-Jan. 8-Jan. 9-Jan. 10-Jan. 11-Jan. 12-Jan. 13-Jan. 14-Jan. 15-Jan. 16-Jan.

NW W W W NW NW E NNE SSE NW W W NW WNW W W

17-Jan 18-Jan. 19-Jan. 20-Jan. 21-Jan. 22-Jan. 23-Jan. 24-Jan. 25-Jan. 26-Jan. 27-Jan 28-Jan. 29-Jan. 30-Jan. 31-Jan.

W W NE SE NW NW W E NE NE NE WSW WSW WSW SW

2.9 7.5 7.7 5.8 4.3 2.3 7.0 4.3 3.7 3.1 5.6 5.3 3.5 4.3 3.4 3.8

2.8 3.6 3.3 3.2 2.3 1.8 2.9 3.1 3.1 7.0 4.1 5.5 3.9 2.9 3.2

570

H. Aoshima et al.

wind speed at Oki was 4.8 m s−1 , which implies that the travel time from Korea to Japan can be calculated to be approximately 1 day since Tonghae is 400 km away from Oki.

3. Discussion 3.1. Correlation between Tonghae and Oki It can be assumed that there is a time lag of the traveling time of c.a. 1 day between the variations in the 222 Rn concentrations at Tonghae and Oki under the monsoon condition in winter. The new baselines are shown in Fig. 4 in which the baseline of Tonghae is shifted ahead by 24 h. Better correlation can be seen at the positions of the peaks of the variations than in Fig. 3. The correlation coefﬁcient in Fig. 3 is 0.21 for the non-shifted data, while the correlation coefﬁcient in Fig. 4 is 0.39 for the shifted data. The improvement of the correlation coefﬁcient implies that the baseline concentration is strongly affected by the continental component of 222 Rn. However, the relatively low correlation coefﬁcient of 0.39 can be attributed to the fact that the wind direction is not always steady from Tonghae to Oki. 3.2. Dilution of contaminants during long-range transport To quantitatively evaluate air mass transport, a better correlation coefﬁcient than R = 0.39 calculated for the whole of January is preferable (Fig. 5). Instead of analyzing the data for the whole month, a particular period is considered when the correlation coefﬁcient is high. This period (“analysis period” in Fig. 4) is from 14 to 23 January 2001. The reason for choosing this period is based on the relatively constant wind directions from the west to the northwest, which is typical of air mass transport in the Japanese winter. During this period a much better correlation coefﬁcient R = 0.84 was obtained, as shown in Fig. 6. Since the slope of the regression line is 0.40, it can be said that 40% of the continental surface air mass arrived at ground level in Japan.

Fig. 4. The “baselines” (Bq m−3 ) of 222 Rn concentrations at Tonghae shifted by 24 h ahead and at Oki in January 2001.

A study of atmospheric radon transport as a tracer of pollutants over the Japan Sea

571

Fig. 5. The correlation of 222 Rn “baseline” (Bq m−3 ) concentrations at Tonghae and at Oki in January 2001, and their correlation coefﬁcient R.

Fig. 6. The correlation between 222 Rn “baseline” (Bq m−3 ) concentrations at Tonghae and those at Oki during the analysis period from 14 to 23 January 2001, and their correlation coefﬁcient R.

3.3. Factors affecting the reduction of concentration There is a difference in the 222 Rn concentrations between Tonghae and Oki. Several reasons can be considered for of 222 Rn concentration decrease. The following are most the plausible: (1) radioactive decay of 222 Rn, (2) mixing to the upper atmosphere, (3) dissolving in sea water. The ﬁrst factor has been taken into consideration in this analysis. It is difﬁcult to take into account the last factor, and therefore we have not estimated its contribution. It must be considered in future studies. It is therefore considered that the second factor is the main factor leading to a concentration decrease. The vertical proﬁles of 222 Rn concentration measured over the windward sea of Oki in December 2000 by airplane are presented in Fig. 7. As shown in the ﬁgure, 222 Rn had arrived at a height of 3000 m or higher over the sea. These results are consistent with the

572

H. Aoshima et al.

Fig. 7. The vertical proﬁle (altitude in m) of 222 Rn concentrations (Bq m−3 ) in the air over Oki in December 2000. The open circles, triangles and squares are the observations at the ground surface on the Oki Islands.

interference (2) mentioned above, i.e., air masses originating from the continental surface are mixed high up in the atmosphere.

4. Conclusions The aim of this study was to evaluate the ratio of air mass transport quantitatively using radon as a tracer, for Oki and Tonghae. It was found that the contribution of transport from the surface of the continent of China to the surface of Japan reaches 40% under winter monsoon conditions in Japan. In this paper, some results were presented only for Oki and Tonghae. A similar analysis for Nagoya and Oki was also made. However, 222 Rn concentrations measured at Nagoya are strongly affected by local sources, which override the variations in the continental component. This makes the correlation coefﬁcient between Nagoya and Oki much poorer than that between Tonghae and Oki. The method presented for the estimation of the transport ratio is still very approximate. The following issues have to be considered in future for a more precise and quantitative evaluation: 1. Estimation of more precise concentrations of the non-local component when it is quite different from the actual local minimum value of 222 Rn concentrations. 2. Evaluation of the effect of 222 Rn solution to sea water. 3. More accurate evaluation of air mass travel time. 4. Comparison with the measurements of other contaminants.

References [1] [2] [3] [4]

S. Whittlestone, Atmos. Environ. 26 (1992) 251. T. Iida, Y. Ikebe, K. Suzuki, K. Ueno, Z. Wang, Y. Jin, Environ. Int. 22 (1996) 139. S.K. Dua, P. Kotrappa, P.C. Gupta, Health Phys. 45 (1983) 152. T. Iida, Y. Ikebe, K. Tojo, Res. Lett. Atmos. Electr. 11 (1991) 55.

573

Non-destructive evaluation of concrete condition using radon exhalation monitoring: a feasibility study E.R. van der Graaf, R.J. de Meijer Nuclear Geophysics Division, Kernfysisch Versneller Instituut, Zernikelaan 25, 9747 AA Groningen, The Netherlands

In this paper we study the feasibility of developing a new non-destructive in situ monitoring technique for concrete that is based on measurements of the radon exhalation of the concrete surface. The achievable sensitivity of such a technique has been assessed based on calculations with a radon transport code for concrete that has recently been successfully validated through comparison with experiments. This assessment showed that the surface radon exhalation rate is in principle sensitive enough to reﬂect anticipated changes in the concrete condition and consequently, a non-destructive evaluation of concrete condition by means of radon exhalation monitoring is feasible. However, as the surface exhalation rate is also sensitive to the ambient relative humidity, preconditioning of the concrete surface before measurements may be needed.

1. Introduction Within the concept of sustainable development the durability of concrete constructions becomes increasingly important [1]. In most instances it is still a priori unknown over which time span concrete construction elements keep their fundamental properties needed to preserve the integrity of the construction. In situ non-invasive monitoring of parameters that are associated with these properties is often used to obtain information on the actual condition of the concrete. Decisions on the necessity of replacement or remediation of constructions can be largely inﬂuenced by the outcomes of such monitoring programmes. Especially in cases of large constructions like, tunnels, underground parking facilities, bridges, hydroelectric dams, etc., a decision for replacement or remediation will have a huge economical impact. Furthermore, for most of these constructions, undetected damage will result in a signiﬁcant risk for human safety. In situ condition assessment of concrete should therefore be very reliable. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07069-X

© 2005 Elsevier Ltd. All rights reserved.

574

E.R. van der Graaf, R.J. de Meijer

Concrete, as all cement based materials is subject to continuous ageing. The actual effects of this ageing process are determined by a multitude of parameters such as concrete composition, presence of reinforcing materials, environmental conditions, position in the construction, internal and external stresses etc. Degradation due to ageing may result in both changes in the microstructure, e.g., altered porosity due to carbonation, occurrence of micro cracks due to static or cyclic loads and larger scale changes such as holes, delaminations and spalling. Recently, the properties of near-surface concrete (cover-zone concrete; cover-concrete) have been identiﬁed as being essential for assessment of durability [2,3]. Near-surface concrete actually is the ﬁrst line of defence against attack by deterioration mechanisms such as, chloride ingress, carbonation, freezing and thawing, alkali–silica reactions, etc. Consequently, changes in the cover-zone concrete are the ﬁrst indication of possible failure of the construction. Most of the deterioration mechanisms are controlled by the rate at which deleterious substances can enter the concrete. Information on the transport properties of the near-surface concrete is thus essential to estimate the remaining service life of the structure. As a consequence of the natural presence of trace amounts of 238 U (uranium) all stony building materials exhale radon (222 Rn) that is formed in the decay of 226 Ra, which is part of the 238 U decay series. Both the generation and the transport of radon in concrete are intrinsically related to the concrete (micro) structure [4,5]. Consequently, the amount of radon that exhales from a certain area of the surface of a concrete structure is a telltale signal of the condition inside the structure. Moreover, due to the limited half-life of radon (3.8 days) only radon that originates from the near-surface zone will have the possibility to reach (by diffusion, advection, etc.) the surface of the concrete construction. Radon exhalation monitoring is thus especially suitable to assess the condition of cover-zone concrete. In this study the sensitivity of the radon exhalation rate to changes in the physical properties of concrete is assessed using a 1D-version of a radon release model that was recently developed and validated.

2. Radon release model and parameters for standard concrete 2.1. Radon release model The model for radon release is based on the multi-phase radon transport equation [6,7]. In this initial assessment we consider a steady-state situation with no pressure differences, then this equation simpliﬁes to ∇ · (D∇Ca ) − βλCa + S = 0,

(1)

where D is bulk radon diffusion coefﬁcient (m2 s−1 ); Ca is radon concentration in the airﬁlled pore volume (Bq m−3 ); β is partition corrected porosity, β = ε(1 − m + Lm) + ρb ka ; ε is porosity; m is fraction of water saturation of the pores; L is Ostwald coefﬁcient (L = 0.26, value at 293 K); ρb is bulk dry density of concrete (kg m−3 ); ka is radon surface-adsorption coefﬁcient (m3 kg−1 ); λ is decay constant of radon (2.1 × 10−6 s−1 ); S is radon production rate per unit bulk volume (Bq m−3 s−1 ); S = ηρb λCRa ; η is radon emanation factor; CRa is radium content (dry mass basis) (Bq kg−1 ).

Non-destructive evaluation of concrete condition using radon exhalation monitoring: a feasibility study

575

Fig. 1. Radon release rate R of a 15 × 15 × 15 cm3 test cube as a function of m. The experimental values are given by ﬁlled circles; the model calculations using equation (1) are represented by the solid line. Calculations are based on input parameters for standard concrete.

The 1D-equivalent of equation (1) has been solved numerically for Ca using a grid of 500– 1500 control volumes for a medium extending from −L tot + L with the NGD-KVI radon transport code with boundary conditions of a zero concentration at the surfaces. The radon exhalation rate was then deduced from the slope of the concentration proﬁle at the surfaces. The NGD-KVI radon transport model (1D- and 2D-versions) and its implementation have been extensively validated for both steady state and transient situations against experiments with dry and wet sand [8,9] and against analytical solutions [7] and other numerical codes [10]. Recently, the 3D-version has been compared with experiments in which the radon release rate R (Bq s−1 ) of a concrete cube (15 × 15 × 15 cm3 ) was measured as a function of its moisture content m [4]. The model calculations were in good agreement with the experimental data (Fig. 1). Especially in the range 0 < m < 0.8 the correspondence is excellent. In the range 0.8 < m < 1.0 the calculations slightly overestimate the date. However, in the view of the still limited amount available for the input parameters (see the next section) it was concluded that the moisture dependence of radon release from concrete modelled on basis of the multi-phase radon transport equation is consistent with the experimental results. 2.2. Parameters for standard concrete Test cubes (sides 15 cm) of standard Dutch concrete (composition on mass basis, blast furnace cement: 14.4%, sand: 28.4%, gravel: 50.5%, water: 6.7%) that were allowed to cure for at least six months were used in the experiments described above. To solve equation (1) and obtain quantitative results, values for a multitude of parameters of the concrete that occur in the equation have to be determined. This section presents the parameters that were used in the model calculations presented in Fig. 1. These parameters can be ordered in two groups namely those that do not depend on m and those that are m-dependent.

576

E.R. van der Graaf, R.J. de Meijer Table 1 Values of m-independent parameters for standard concrete Parameter

Value

CRa ε ρb ka

21.7 ± 0.4 Bq kg−1 0.115 ± 0.005 2260 ± 30 kg m−3 set to zero

2.2.1. m-independent parameters Radium content CRa of the concrete was determined by high-resolution γ-spectroscopy on dried crushed parts of one of the concrete cubes. Porosity ε was calculated on basis of the mass difference of completely water saturated cubes and of dry cubes (dried to constant mass at 200 ◦ C). Bulk dry density ρb was calculated from the dry mass and the dimensions of the cubes. The radon surface-adsorption coefﬁcient ka was not measured. For wet surfaces this adsorption is usually assumed to be negligible. For dry surfaces the value of the coefﬁcient might be non-zero. Lacking any experimental information for the type of concrete used in this study, ka was set to zero. The values of the m-independent parameters for standard concrete are presented in Table 1. 2.2.2. m-dependent parameters The radon-diffusion coefﬁcient in water is four orders of magnitude smaller than the coefﬁcient for diffusion in air. This implies that with increasing water saturation of the concrete the bulk radon diffusion coefﬁcient D will decrease. The most consistent description of experimental values of this m-dependency of D for standard concrete was found [4] to be given by a three-parameter function (equation (2)). This function with D(0) = (1.77 ± 0.06) × 10−8 m2 s−1 , a = 2.57 ± 0.08 and b = 2.55 ± 0.08 is depicted in Fig. 2 and shows that the diffusion coefﬁcient decreases more rapid at higher m-values:

Fig. 2. m-Dependence of bulk diffusion coefﬁcient D for standard concrete.

Non-destructive evaluation of concrete condition using radon exhalation monitoring: a feasibility study

577

Fig. 3. m-Dependence of emanation factor η for standard concrete.

D(m) = D(0)e−a(m+bm ) . 5

(2)

It is well known that the radon-emanation factor depends on the moisture content of the porous medium. For soil-like materials [9] the emanation factor seems to increase with increasing moisture content up to m ≈ 0.1–0.5 at which point the emanation factor reaches a constant maximum value for higher moisture contents. The moisture dependency of the radon-emanation factor of concrete was calculated from radon-release rates on a batch of crushed parts and the radium content of the concrete under the assumption that all radon produced into the pore space also was released [4]. The emanation factor of concrete showed an almost linear increase over the entire range 0 < m < 1 (best estimate η(m) = η(0) + cm; η(0) = 0.01 ± 0.01; c = 0.35 ± 0.02, see Fig. 3) which is in contrast to the behaviour of, e.g., sand that shows a fast increase at low m-values and thereafter obtains an constant value [9].

3. Sensitivity analysis In order for radon exhalation monitoring to be a viable technique for studies of concrete condition, changes in parameters that are associated with this condition should be reﬂected in preferably large variations in the radon exhalation rate. The porosity and the diffusion coefﬁcient are known indicators of the degree of carbonation of concrete. Porosity values between 0.05 and 0.2 are reported [11] while, e.g., the chloride diffusion coefﬁcient may increase by more than an order of magnitude [3]. To assess the sensitivity of the surface radon exhalation rate Es (Bq m−2 s−1 ) for these two parameters, Es was calculated as a function of moisture content m for the standard concrete parameters (Section 2.2) and for some variations in porosity ε and diffusion coefﬁcient D. Calculations were done for a 1D-situation (inﬁnite concrete slab) with a thickness of 20 cm (L = 0.10 m, Section 2.1). This thickness is of the same order as walls and ﬂoors in constructions. The 1D-calculation with the standard concrete parameters (Fig. 4, solid line) results in a similar m-dependence as for the 3D-calculations for the cube (Fig. 1). A decrease of the porosity with approximately a factor 2 increases Es for all m-values. However, the increase

578

E.R. van der Graaf, R.J. de Meijer

Fig. 4. 1D-model calculations of m-dependence of Es for three porosities.

is only 2–5% for low moisture contents (m < 0.5), for higher moisture contents the increase can be as large as 50%. Increasing the porosity to ε = 0.2 leads to a 2–5% decrease in Es for m < 0.5 and to a 5–30% decrease for larger m-values. Physically these effects are caused by ﬁrstly, increasing/decreasing porosity will decrease/increase the amount of solid material (and thus the amount of radium) and consequently increase/decrease the total amount of radon produced. This results in a lower/higher radon exhalation rate. Secondly, changes in ε will result in changes in the partition corrected porosity β and thus change the concentration proﬁle calculated by using equation (1). The effect of changes in D was simulated by considering, besides the standard value of D(0) = 1.77 × 10−8 m2 s−1 in equation (2), also values that were an order of magnitude higher and lower. Figure 5 shows that, as to be expected, larger diffusion coefﬁcients lead to higher radon exhalation rates. Moreover, especially for the larger m-values, differences of a factor 2–3 seem to be possible.

Fig. 5. 1D-model calculations of m-dependence of Es for three values of D(0) (equation (2)).

Non-destructive evaluation of concrete condition using radon exhalation monitoring: a feasibility study

579

4. Discussion In general, the surface radon exhalation rates from standard concrete and the variations thereupon as considered in this study, range from 0.1–3 mBq m−2 s−1 . These exhalation rates can easily be measured in situ by state-of-the-art radon measurement techniques with relative uncertainties less than 10% [12]. This implies that, in principle, for the situations considered in this study the radon exhalation rate is sensitive enough to change in the concrete structure to provide information on the condition of the concrete. One of the consequences of the radon exhalation rate of concrete being dependent on its moisture content is that Es will dependent on the relative humidity (RH) of the surrounding atmosphere. This implies that concrete of similar condition in a structure might give a different response with respect to its radon exhalation rate due to a difference in environmental conditions. To estimate the magnitude of this effect a relation is needed that links RH and the moisture content (m) of concrete. Such a relation was developed on basis of the pore-size distributions of the gel and capillary pores of the standard concrete (after 160 days of curing) calculated by the concrete structure modelling code DuCOM [5,13]. Figure 6 shows that the fraction of water saturation m of the total porosity increases relatively fast at low RH-values (RH < 10%). Thereafter the decrease is more gradual up to RH = 80%, the ﬁnal increase until saturation is again more steep (RH > 80%). The ﬁgure also shows that the gel porosity is already 80% (m = 0.8) saturated at RH = 20%, this is in sharp contrast with the behaviour of the capillary porosity that gradually increases to m = 0.2 unto RH = 80%, thereafter the increase to saturation is very steep. The relation between RH and m has been used to calculate the surface exhalation rate Es of standard concrete that is in equilibrium with a certain relative humidity (Fig. 7). Es is found to increase gradually in going from RH = 0 to 60% where a plateau is reached that extends from RH = 60 to 80%. Thereafter, Es decreases very rapidly, reﬂecting the fast ﬁlling of the porosity at the higher end of the m-RH curve (Fig. 6), which will result in a fast decrease of the diffusion coefﬁcient.

Fig. 6. Relation between RH and m for standard concrete. Solid line: fraction of water saturation of total porosity; dotted line: idem, for gel porosity and dashed line: idem, for capillary porosity.

580

E.R. van der Graaf, R.J. de Meijer

Fig. 7. Surface exhalation rate Es as a function of relative humidity.

The RH-dependence of Es as given in Fig. 7 implies that one has to be careful in comparing radon exhalation rates of parts of concrete structures that are exposed to different environmental conditions. Especially, in the range RH > 80% large differences in Es may be induced by relatively small changes in RH. For precise measurements it may be needed to precondition the concrete to a certain relative humidity. A value around RH = 70%, in the middle of the plateau seems to be an obvious choice for such preconditioning.

5. Conclusion The results presented in this paper indicate that non-destructive in situ monitoring of concrete condition based on measurements of the radon exhalation of the concrete surface is feasible. An assessment of the sensitivity of the radon exhalation rate on changes in concrete porosity and diffusivity showed that difference of a factor 2–3 in the surface exhalation rate are possible. However, as the surface exhalation rate is also sensitive to the ambient relative humidity, preconditioning of the concrete surface before measurements may be needed.

References [1] [2] [3] [4] [5] [6] [7] [8]

P.-C. Aïtcin, Cem. Concr. Res. 30 (2000) 1349. A.E. Long, G.D. Henderson, F.R. Montgomery, Constr. Build. Mater. 15 (2001) 65. L. Basheer, J. Kropp, D.J. Cleland, Constr. Build. Mater. 15 (2001) 93. E.R. van der Graaf, I. Cozmuta, R.J. de Meijer, in: Proc. Third Eurosymposium on Protection against Radon, Liège, Belgium, 10 and 11 May 2001, p. 95. I. Cozmuta, Radon generation and transport – a journey through matter, Thesis, University of Groningen, The Netherlands, 2001. V.C. Rogers, K.K. Nielson, Health Phys. 74 (1991) 807. W.H. van der Spoel, Radon transport in sand. A laboratory study, Thesis, Eindhoven University of Technology, The Netherlands, 1998. W.H. van der Spoel, E.R. van der Graaf, R.J. de Meijer, Health. Phys. 74 (1998) 48.

Non-destructive evaluation of concrete condition using radon exhalation monitoring: a feasibility study

581

[9] W.H. van der Spoel, E.R. van der Graaf, R.J. de Meijer, Health. Phys. 77 (1999) 163. [10] C.E. Andersen, D. Albarracín, I. Csige, E.R. van der Graaf, M. Jiránek, B. Rehs, Z. Svoboda, L. Toro, ERRICCA Radon model intercomparison exercise, Risø rapport nr. R-1120(EN), Risø National Laboratory, Roskilde, Denmark, 1999. [11] W.P.S. Dias, Cem. Concr. Res. 30 (2000) 1255. [12] F.J. Aldenkamp, R.J. de Meijer, L.W. Put, P. Stoop, Radiat. Prot. Dosim. 45 (1992) 449. [13] K. Maekawa, Modeling of Concrete Performance, Hydration, Microstructure Formation and Mass Transport, EFN Spon, London, 1999.

582

Study of Radon-222 exhalation of phosphogypsum blocks used as building materials. Comparison with modeling F. Fournier a , J.E. Groetz a , F. Jacob b , H. Lettner c , A. Chambaudet a , J.M. Crolet b a Nuclear Microanalysis Laboratory, LRC CEA M07, University of Franche-Comté, 16, route de Gray,

25030 Besançon cedex, France b Laboratory of Mathematics, UMR CNRS 6623, University of Franche-Comté, 16, route de Gray,

25030 Besançon cedex, France c Institute for Physics and Biophysics, University of Salzburg, Hellbrummer Strasse 34, 5020 Salzburg, Austria

Phosphogypsum is a by-product of the phosphate fertilizer industry, which is stockpiled in large quantities worldwide. It consists mainly of dihydrate gypsum but contains elevated concentrations of 226 Ra and other inorganic species, which originate from the processing of phosphate rocks. 222 Rn is the ﬁrst decay product of 226 Ra and 222 Rn gas evolution has been identiﬁed as one of the major environmental concerns associated with phosphogypsum. We present here a laboratory method for the determination of radon exhalation rate from gypsum used in the fabrication of building materials. In this study, the phosphogypsum slabs are placed under a hemispherical chamber in order to determine the radon concentration in samples by alpha spectroscopy and the uni-directional radon exhalation. The radon exhalation rate from phosphogypsum at various moisture contents is also studied: with increasing moisture content, a signiﬁcant increase of 222 Rn exhalation can be shown. On the other hand, we performed a 3D simulation of the transport and exhalation of radon in phosphogypsum samples using a ﬁnite-element method. The theoretical study characterized the equation of radon transport in a building material, which combines the diffusive migration of radon and the effects of radon decay. This simulation follows the evolution of 222 Rn concentration in the phosphogypsum and we obtain an evaluation of the theoretical 222 Rn exhalation rate in the chamber. Finally, the experimental and theoretical results are compared. 1. Introduction Phosphogypsum (PG), a waste by-product derived from the wet process production of phosphoric acid, represents one of the most serious problems facing the phosphate industry RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07070-6

© 2005 Elsevier Ltd. All rights reserved.

Study of Radon-222 exhalation of phosphogypsum blocks used as building materials

583

today [1]. Naturally occurring uranium (U) and its radioactive decay series are associated with phosphate mineral deposits. Indeed, phosphogypsum consists mainly of gypsum, but also contains elevated concentrations of 226 Ra. Consequently, 222 Rn gas, the ﬁrst decay product of 226 Ra, has been identiﬁed as one the major environmental concerns associated with PG [2]. Inhalation of 222 Rn and its short-lived progeny (218 Po and 214 Po) has been linked to lung cancer in humans [3]. Because of that, the uses of phosphogypsum should be limited. The objective of this laboratory study was to determine experimentally the radon concentration in phosphogypsum blocks by alpha spectroscopy in order to obtain an evaluation of the radon exhalation rate, which represents the ﬂux at which 222 Rn leaves the surface of a material such as phosphogypsum. These experimental measurements allowed us to better understand the risks associated with 222 Rn in phosphogypsum and to highlight the signiﬁcant effect of humidity on the 222 Rn exhalation. At the same time, a three-dimensional model for radon migration in phosphogypsum sample was established. This model uses the radon transport equation by molecular diffusion in building material. The modeling allows us to achieve numerical results of the theoretical 222 Rn exhalation rate in the surface of the material. After that, it was possible to compare the experimental and numerical results in order to investigate whether our 3D radon transport model in building materials is well calibrated.

2. Radon transport in a building material, the phosphogypsum After the radon generation into the porous medium from the emanation process, a part of the produced radon atoms can be transported by a ﬂuid. Radon transport in porous media such as building materials is controlled mainly by molecular diffusion. This mechanism allows for the radon atoms to reach the atmosphere by exhalation process. 2.1. Equation for diffusive radon transport in phosphogypsum Rogers and Nielson [4,5] have developed a formalism to describe the diffusive and advective multi-phase transport of radon in porous materials. For a small phosphogypsum sample, the advection mechanism can be neglected because the pressure and/or temperature gradients are very low throughout the material. This formalism includes three processes, which determine the radon concentration [6]: molecular diffusion, radon emanation and radon decay. The transport phenomenon is given by the following time-dependent partial differential equation: β

∂Ca = ∇(D∇Ca ) − βλCa + S ∂t

(1)

where β is the partition-corrected porosity, D the bulk diffusion coefﬁcient (m2 h−1 ), Ca the radon concentration in the air phase (Bq m−3 ), S the radon production rate per unit volume (Bq m−3 h−1 ) and λ the decay constant of radon (2.1 × 10−6 s−1 ). The partition-corrected porosity β depends directly on the water content [4]. It can be expressed by the equation: β = ε(1 − m + Lm)

(2)

584

F. Fournier et al.

where ε is the porosity of material, L the Ostwald coefﬁcient for radon (0.26 at 293 K) and m the relative humidity saturation rate. The production rate of radon S is assumed to be constant in space and time [7], i.e., S = λCRa ρη

(3)

where CRa is the 226 Ra concentration per unit mass of the material (Bq kg−1 ), ρ the density of the material (kg m−3 ) and η the radon emanation coefﬁcient of the material. 2.2. Numerical tool For solving this time-dependent, partial differential equation in three dimensions, a threedimensional ﬁnite element method for the spatial discretization was developed and a ﬁnite differences method for the temporal discretization was used [8]. Programs are written in Fortran 77 and the code is running under the Linux system. The software Modulef99 is used to create and display the area mesh. The resolution of equation (1) makes it possible to obtain the radon concentration in all points of the porous building material. The model equations assume an isothermal medium and no adsorption of radon on the medium grain surfaces [9]. 2.3. Values of parameters Radon diffusion depends on different chemical parameters as shown in equations (1)–(3). The different parameter values chosen for the numerical simulation are: • ε = 0.62 for phosphogypsum; • ρb = 924.9 kg m−3 ; • S = 6.685 Bq m−3 h−1 . Diffusion coefﬁcients for 222 Rn in porous materials can be measured by standard laboratory methods [10]. If measurements are unavailable, generic values can be estimated for modeling purposes from an empirical equation depending on material porosity and material water content [11]. The diffusion coefﬁcient D can be obtained from equation (4): D = D0 ε exp −6mε − 6m14ε (4) where D0 is the diffusion for 222 Rn in air (1.1 × 10−5 m2 s−1 ). For example, for a dry material (m = 0), D = 2.45 × 10−3 m2 h−1 . 3. Results of simulation 3.1. Area built and boundary conditions imposed In order to simulate radon diffusive transport in phosphogypsum, a phosphogypsum cubeshaped sample was built with a 10 cm edge. A cubic accumulation chamber of the same size was placed above the sample, as shown in Fig. 1. For the numerical resolution of equation (1), boundary conditions were ﬁxed for radon transport by molecular diffusion. These boundary conditions (as Dirichlet conditions) are:

Study of Radon-222 exhalation of phosphogypsum blocks used as building materials

585

Fig. 1. Modeling of the phosphogypsum sample and the accumulation chamber.

• C = 500 Bq m−3 on the bottom side of the PG sample; • C = 0 Bq m−3 on the top side of the accumulation chamber; • C = 0 Bq m−3 at t = 0 h allover the sample. These Dirichlet conditions make it possible to follow the changes in the radon concentration in the phosphogypsum sample. 3.2. Numerical results of radon diffusion in phosphogypsum From the previous data, simulations of the radon activity in the PG sample and the accumulation chamber were performed. Fig. 2 shows the results of the calculated radon concentration versus time for three various depths in the PG sample. Results of the numerical calculations show that the radon activity builds over time, reaching a plateau after about 10 hours. Thus, it can be concluded that an equilibrium between the radon concentration in the PG sample and the radon concentration in the accumulation chamber is reached. The phenomenon called “back diffusion” could take place if the radon concentration in the chamber becomes higher than the concentration in the PG sample [12]. In that case, radon could collect at the top of the PG sample.

Fig. 2. Evolution of radon concentration versus time for various depths in the phosphogypsum sample at x = y = 0.05 m for the three curves.

586

F. Fournier et al.

4. Experimental results for radon concentration in PG 4.1. Experimental protocol Some experimental measurements were performed on the PG blocks in order to evaluate the radon concentration and the radon exhalation rate as a function of humidity [13]. For the determination of the all-round 222 Rn exhalation rate, phosphogypsum blocks were placed on a cubic measurement chamber. For the measurements of the uni-directional exhalation, a hemispherical chamber of volume V = 7 L was placed on a single PG sample, as it shows in Fig. 3. The accumulation chamber was used only for the radon progeny 218 Po, a short-lived alpha emitter (3.05 min). The radioactive equilibrium between 222 Rn and 218 Po is reached after about 1000 seconds in the chamber. Once reached, the radon activity becomes equal to the polonium activity. So, it will be possible to obtain the experimental radon concentration in the chamber after 1000 seconds. 4.2.

222 Rn

concentration and exhalation rate

The concentration of radon emanated from the PG block inside the exhalation container was allowed to build up with time and was measured in 30 minutes cycles for a time between 600 hours and 1300 hours. The build-up of radon activity inside the exhalation chamber follows the well-known equation: A(t) =

E ·S 1 − exp(−λt) λ·V

(5)

where E is the 222 Rn exhalation rate, S the cross section of the hemisphere on the sample, and V the volume of the exhalation chamber. The results of 222 Rn concentration versus time for different humidity percents in the PG sample are shown in Fig. 4. It shows an increase of radon activity versus time in the accumulation chamber for the four cases. A plateau of radon concentration is obtained after about 7 hours even if the water content in the PG block increases. So, after 7 hours, the radon concentration in the sample is in equilibrium with the radon concentration in the exhalation chamber. Moreover, Fig. 4 shows

Fig. 3. Experimental setup for the determination of unidirectional radon exhalation rate.

Study of Radon-222 exhalation of phosphogypsum blocks used as building materials

587

Fig. 4. 222 Rn concentration versus time for a PG sample of different water saturation rate.

that the higher the water content is, the higher the radon concentration in the accumulation chamber is. The second stage of the experiments is to evaluate the radon exhalation rate from the PG sample for different water saturation rates. The radon exhalation rate per unit area of the PG, E, is deﬁned as the ﬂux of radon released from the surface of the material. From equation (5), E can be calculated only if small times are considered. However, for time higher than 2–3 hours, the two mechanisms that yield loss of radon, which are: the back diffusion into the sample and the loss through leakage cannot be neglected. The contribution of both phenomena could prevent to estimate accurately the radon exhalation rate. So, for time less than 3 hours, E can be estimated using the following formula [14]: E=

AT · V Bq m−2 s−1 . S ·T

(6)

The results of radon exhalation rate for different levels of the water content in the PG sample are shown in Fig. 5. It shows an increase in radon exhaled from PG block up to 10% of water saturation rate and a small decrease in the measurements at 15%. Indeed, the reduction of the water content in the pores of the PG sample leads to a decrease in the probability of retaining radon within the pores (i.e. a lower probability of the emanated radon atom to be stopped within the pore volume) and the probability of radon emanation from the pores. Moreover, raising the water content in a material contributes to an increase in the direct recoil fraction and in the capillary pore volume, thus in the liberated amount of radon. These phenomena could explain the radon exhalation from PG blocks versus the water content. However, this cannot explain the decrease in the radon exhalation rate for 15% of humidity.

588

F. Fournier et al.

Fig. 5. Results of 222 Rn exhalation rate as a function of humidity in the PG sample.

4.3. Comparison with numerical calculations The comparison between the calculations and the experimental measurements shows that both are in close agreement in the entire time interval, as shown in Fig. 6. Note that, for the experimental measurements, the radon concentration was obtained for a PG sample of 5% humidity. For the calculation, we applied the experimental conditions in our model and the radon concentration values have been calculated at the interface, namely between the sample and the accumulation chamber. The very good agreement between simulation and experiments highlights the good calibration of our three dimensional radon transport model for building materials.

Fig. 6. Comparison between experimental measurements and simulation calculations in the interface of the simulated PG sample for a saturation rate of 5%.

Study of Radon-222 exhalation of phosphogypsum blocks used as building materials

589

5. Conclusion A radon transport equation has been established from a preliminary theoretical study of radon migration through porous media such as building materials. The equation has been applied to the case of phosphogypsum because this kind of material represents a signiﬁcant source of radon. With our three-dimensional model developed for simulating radon transport through different porous media, the radon concentration in 3D versus the depth in a phosphogypsum block has been studied. Experimental measurements of radon activity have also been performed by alpha spectroscopy and radon exhalation rate has been calculated for different water contents in the PG sample. The results reveal a build-up of radon exhaled from the PG sample, with a radon concentration plateau after about 7 hours even if the water saturation rate varies. By determination of the radon exhalation rate with different water contents in the sample, the inﬂuence of water has been characterized. An increase in water content provides an increase in the radon exhalation rate. The results of experiments have been compared with calculation, showing a very good agreement that demonstrates the good calibration of our three-dimensional model of radon transport. However, in the future, this model should be improved in order to take into account the inﬂuence of other phenomena such as non-linear ﬂow and other physical characteristics of the medium such as heterogeneity. References [1] F. Ferguson, Phosphogypsum – an overview, in: Proc. Second Internat. Sympos. on Phosphogypsum, vol. 1, Miami, FL, December 1986, Publication 01-037-055, Florida Institute of Phosphate Research, 1988, p. 117. [2] P.M. Rutherford, M.J. Dudas, R.A. Samek, Environmental impacts of phosphogypsum, Sci. Total Environ. 149 (1994) 1. [3] NCRP, Exposure of the population in the United States and Canada from natural background radiation, NCRP Report No. 94, National Council of Radiation Protection and Measurements, Bethesda, MD, 1997. [4] V.C. Rogers, K.K. Nielson, Generalized source term for the multiphase radon transport equation, Health Phys. 64 (1993) 324. [5] V.C. Rogers, K.K. Nielson, Multiphase radon generation and transport in porous materials, Health Phys. 60 (1991) 807. [6] M. Van der Pal, E.R. Van der Graaf, Experimental set-up for measuring diffusive and advective transport of radon through building materials, Sci. Total Environ. 272 (2001) 315. [7] M.S. Gadd, T.B. Borak, In-situ determination of the diffusion coefﬁcient of 222 Rn in concrete, Health Phys. 68 (1995) 817. [8] F. Fournier, J.E. Groetz, F. Jacob, H. Lettner, A. Chambaudet, Three-dimensional model for radon transport in porous soil: application to soil and building materials, Comput. Phys. Commun., in press. [9] S.D. Schery, Modeling radon transport in dry, cracked soil, J. Geophys. Res. 98 B1 (1993). [10] K.K. Nielson, V.C. Rogers, D.C. Rich, Comparison of radon diffusion coefﬁcients measured by transientdiffusion and steady-state laboratory methods, US Nuclear Regulatory Commission report No. NUREG/CR2875, Washington, DC, 1982. [11] V.C. Rogers, K.K. Nielson, Correlations for predicting air permeabilities and 222 Rn diffusion coefﬁcients of soils, Health Phys. 61 (1991) 225. [12] Y.H. Chao, C.W. Tung, W.T. Chan, Determination of radon emanation and back diffusion characteristics of building materials in small chamber tests, Build. Environ. 32 (1997) 355. [13] H. Lettner, F. Steinhäusler, Radon exhalation of waste gypsum recycled as building material, Radiat. Prot. Dosim. 24 (1–4) (1988) 415. [14] R. Mustonen, Natural radioactivity in and radon exhalation from Finnish building materials, Health Phys. 46 (1984) 1195.

590

15 years repeated investigations of the radon levels in 105 remediated Swedish dwellings, mostly single-family buildings B. Clavensjö Bjerking AB, Box 2006, SE-750 02 Uppsala, Sweden

The object of this project was to investigate the long time effectiveness of different radon remedial methods. The ten years project started in 1991. From the start the investigation comprised 105 dwellings (91 single-family houses and 14 ﬂats in multi-family buildings). In all of the dwellings remedial measures were carried out in the eighties. Before and immediately after the reduction the local authorities measured the radon concentrations. New measurements of the radon concentrations have been made every third year: in 1991, 1994, 1997 and in 2000. Twelve different radon remedial methods and method combinations were used. The radon sources were building materials as well as sub-soils. In all of the dwellings the radon concentrations were measured by nuclear track ﬁlms over 3 month (January–March) measurement periods [1] and in half of them the air change rates by passive tracer gas methods [2].

1. Remedial measures Remedial actions [3] that have been studied in this project are: (1) (2) (3) (4) (5)

converting a natural draught system to a mechanical exhaust air system; installation of a mechanical supply and exhaust air ventilation system; sealing the structure above the crawl space; sealing leakage points in the basement ﬂoor; installation of a system for sucking out indoor air and compressing it into the soil below the house to increase the air pressure there; (6) a combination of installation of mechanical exhaust air system and sealing of leakage points in the basement ﬂoor; (7) a combination of installation of a mechanical supply and exhaust air ventilation system and sub-slab suction; RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07071-8

© 2005 Elsevier Ltd. All rights reserved.

15 years repeated investigations of the radon levels in 105 remediated Swedish dwellings

591

(8) a combination of installation of a mechanical exhaust air ventilation system and sub-slab suction; (9) installation of sub-slab suction; (10) installation of a radon well; (11) additional natural ventilation; (12) a combination of converting a natural draught system to a mechanical exhaust air system and sealing round about the culvert intake. The reduction actions have been done in multi-family buildings.

2. Results The result of the 1991 study showed that the radon concentration was up to 200 Bq m−3 in 54 single-family houses and 7 ﬂats, between 210 and 400 Bq m−3 in 18 single-family houses and 6 ﬂats, and higher than 400 Bq m−3 in 18 single-family houses and 1 ﬂat. The study showed also that in about 40% of the cases the radon concentration had increased by more than 30% only a few years after reduction actions had been taken (see Fig. 1). In 19 dwellings the radon concentration was at least doubled. The result of the 1994 study showed that the radon concentration was up to 200 Bq m−3 in 47 single-family houses and 9 ﬂats, between 210 and 400 Bq m−3 in 24 single-family houses and 4 ﬂats, and higher than 400 Bq m−3 in 17 single-family houses and 1 ﬂat. As in 1991 the radon concentration was higher than 400 Bq m−3 in 12 of the 17 single-family houses. In the calculation values less than 100 Bq m−3 have not been used. Such values have been increased to 100 Bq m−3 in order to reduce the proportion at low values. An example: In a house the radon concentrations were 60 Bq m−3 in 1991 and 370 Bq m−3 in 2000. We increase from 60 to 100 Bq m−3 in order to reduce the proportion at low values.

Fig. 1. The diagram presents the number of houses in which the radon concentration on one measuring occasion (A) divided by the measured value on another occasion (B) was < 0.7 (the lower part of the column), 0.7–1.3 (the intermediate part), > 1.3 (the upper part) as follows: (1) A measurements in 1991, B measurements immediately after the reduction action; (2) A measurements in 1994, B measurements in 1991; (3) A measurements in 1997, B measurements in 1991; (4) A measurements in 2000, B measurements in 1991.

592

B. Clavensjö

370 (A) divided by 100 (B) is 3.7. The result is to be found in the upper part of column 4 (1 of 15 houses). The result of the 1997 study showed that the radon concentration was up to 200 Bq m−3 in 46 single-family houses and 1 ﬂat, between 210 and 400 Bq m−3 in 34 single-family houses and 1 ﬂat, and higher than 400 Bq m−3 in 7 single-family houses and 10 ﬂats. As in 1991 the radon concentration was higher than 400 Bq m−3 in 5 of the 7 single-family houses. The result of the 2000 study showed that the radon concentration was up to 200 Bq m−3 in 54 single-family houses and 7 ﬂats, between 210 and 400 Bq m−3 in 23 single-family houses and 5 ﬂats, and higher than 400 Bq m−3 in 12 single-family houses and 2 ﬂats. As in 1991 the radon concentration was higher than 400 Bq m−3 in 7 of the 12 single-family houses. In 4 single-family houses the radon concentration was higher than 400 Bq m−3 in every measurement. In Sweden the highest permissible annual mean value of radon gas concentrations indoors in existing dwellings is 400 Bq m−3 . In comparison with this value, 12 single-family houses and 2 ﬂats had radon concentrations over that in 2000. 35 single-family houses and 7 ﬂats had over 200 Bq m−3 , the limit for new buildings. In no fewer than 38 dwellings has the radon level been over 400 Bq m−3 on at least one of the four measuring occasions. The change in radon concentrations was not speciﬁc to any given method but seemed to be evenly distributed over all of them. The investigation results showed the necessity for repeated measuring where countermeasures have been taken. Figures 2–10 show the individual radon levels from the measures immediately after the reduction actions and from the check measures in 1991, 1994, 1997 and 2000. Every dwelling is indicated by a square turned round 45 degrees representing the radon level in 1994, a regular square indicating the radon level in 1997 and a triangle indicating the radon level in 2000.

Fig. 2. Radon levels after remedial action. Converting a natural draught system to a mechanical exhaust air system. 6 ﬂats in multi-family buildings and 6 single-family houses.

15 years repeated investigations of the radon levels in 105 remediated Swedish dwellings

593

Fig. 3. Radon levels after remedial action. Installation of a mechanical supply and exhaust air ventilation system. 17 single-family houses.

Fig. 4. Radon levels after remedial action. Sealing leakage points in the basement ﬂoor. 6 single-family houses.

3. Causes of increasing radon levels 3.1. Less suitable measures Radon is transported from the soil into a house with soil gas drawn in through leakage points in the foundation structure. The driving force for the transport is the difference between the air pressures just above and below the lowest ﬂoor in the building. The air pressure in the house

594

B. Clavensjö

Fig. 5. Radon levels after remedial action. Installation of a fan assisted sub-slab pressurisation system. 3 single-family houses.

Fig. 6. Radon levels after remedial action. Installation of a mechanical supply and exhaust air ventilation system and a sub-slab suction. 12 single-family houses.

is determined by the temperature difference between the outdoors and indoors, the effect of the wind and by any mechanical ventilation system. In many houses with radon problems from the soil it is not a good remedial action to install a normal mechanical ventilation system or to improve the natural draught ventilation. The ﬂow of radon-containing air from the soil into the house will not stop, only become more diluted in the room. The radon concentration in the house can even increase by installing an exhaust

15 years repeated investigations of the radon levels in 105 remediated Swedish dwellings

595

Fig. 7. Radon levels after remedial action. Installation of a sub-slab suction. 18 single-family houses.

Fig. 8. Radon levels after remedial action. Installation of a radon well. 14 single-family houses.

ventilation system, because of lowering of the air pressure indoors. These occurrences are relatively common in houses built on ground with high permeability and high radon content in the soil air. A single-storey house without a basement has a very low warmed-up volume of air. If the house is ventilated by a mechanical supply and exhaust system, the air pressure indoors is only about 1 Pa below the air pressure outdoors caused by the ventilation system. The thermal force causes about 2 Pa sub-atmospheric pressure just over the lowest ﬂoor, when the outdoor

596

B. Clavensjö

Fig. 9. Radon levels after remedial action. Improving the natural draught ventilation. 13 single-family houses.

Fig. 10. Radon levels after remedial action. A combination of converting a natural draught system to a mechanical exhaust air system and sealing round about the culvert intake. These reduction actions have been done in multi-family buildings. 7 ﬂats in multi-family buildings.

temperature is 5 ◦ C above zero, in total about 3 Pa. If the temperature drops to 5 ◦ C below zero, the sub-atmospheric pressure will increase to a bit over 4 Pa or 40% but the rate of air change will not change. The result will be an increasing ﬂow of radon-containing soil air into the house and an increasing radon content indoors.

15 years repeated investigations of the radon levels in 105 remediated Swedish dwellings

597

3.2. Measurement effects Measurement effects, of course, have not increased the annual mean value of radon content in reality. By measuring during too short a time and the wrong time the annual mean value of radon progeny content can be underestimated. Converting radon progeny levels to radon gas levels by using a too high F -factor will also result in an underestimated radon gas level. 3.3. Lifestyle related causes If the radon contents become very low after sanitising, many house-owners would use that as a justiﬁcation for reducing the number of revolutions of the fan or fans. The purpose of the reduction could be energy saving or fan noise abatement. Different use of the windows for airing and varying the open area of intakes during measurement can result in changed radon levels. 3.4. Insufﬁcient operating instructions The very high radon levels in a lot of ﬂats in 1997 were a result of bad operating instructions. The staff did not know there was a radon well in the area of the multi-family buildings. The fan of the radon well was out of order and nobody had detected it.

4. Conclusion Remedial measures against increased radon concentrations indoors can result in durable low radon concentrations if: • the remedial measure is the most suitable measure in each separate case; • the remedial measure is carried out in a workmanlike manner and by persons who know why the work is being done; • the house owner receives operating and maintenance instructions; • the installation is operated by the house owner in accordance with these instructions. After mitigation the radon concentrations have to be measured over a long period of time (at least 2 months) and during the right period of the year.

References [1] Swedish Radiation Protection Authority, Strålning i bostäder. Metodbeskrivning: Långtidsmätning för uppskattning av radongashaltens årsmedelvärde. Rådgivande korttidsmätning. Mätmetoder för radon: Metodblad nr 1-8, 1994 (in Swedish). i 94-05. [2] R.N. Dietz, R.W. Goodrich, E.A. Cote, R.F. Wieser, Detailed description and performance of a passive perﬂuorocarbon tracer system for building ventilation and air exchange measurements, in: H.R. Trechsel, P.L. Lagus (Eds.), Measured Air Leakage of Buildings, in: ASTM STP, vol. 904, American Society for Testing and Materials, Philadelphia, 1986, pp. 203–264. [3] B. Clavensjö, G. Åkerblom, The Radon Book – Measures against Radon, The Swedish Council for Building Research, Stockholm, Sweden, 1994.

598

Comparison between long-term and short-term measurements for indoor radon risk mapping F. Tondeur, I. Gerardy Institut Supérieur Industriel de Bruxelles, 150, rue Royale, B1000 Brussels, Belgium

Two databases of indoor radon measurements in the Walloon region (Belgium), collected respectively by long-term track-etch measurements and by short-term charcoal measurements, are compared at the county level. They are shown to agree reasonably for geometrical mean values, whereas short-term data show a higher variability. Methods to deﬁne affected areas from short-term measurements are discussed. 1. Introduction Indoor radon has been studied in the southern part of Belgium (the Walloon region) since the late 1980s. Indoor radon data have been collected by several laboratories, both with long-term and short-term measurements. The two most important databases have been collected by the Federal Ministry of Health (FMH) on the one hand, and by the “Institut Supérieur Industriel de Bruxelles (ISIB)” on the other. The recent publication of the analysis of the FMH database [1] allows us to make a systematic comparison with the ISIB database. In particular, the existence of the two databases for the same region gives an opportunity to compare long-term and short-term measurements when applied to indoor radon mapping. Long-term measurements are known to give a better indication of the average situation in a given house, and should be the basis on which the need for mitigation can be evaluated. Short-term measurements show more variability. They were also often used with a protocol for exposure, not applied by ISIB, that is likely to enhance the radon concentration (exposure in the basement under closed conditions) and may explain the large ratios between short-term and long-term measurements observed by some authors [2]. However, their lower cost could make them a good choice for radon mapping projects when large numbers of measurements are averaged. 2. Measurement methods The FMH data are long-term indoor radon data collected from 1995 to 1999 in a national survey, using closed track-etch detectors of the so-called “old-Karlsruhe” type [3] (i.e., a MakroRADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07072-X

© 2005 Elsevier Ltd. All rights reserved.

Comparison between long-term and short-term measurements for indoor radon risk mapping

599

Table 1 Number of data per region and in each Walloon province

Flemish region Brussels region∗ Walloon region Walloon Brabant Hainaut Namur Liège Luxembourg

FMH database

ISIB database

Population (millions)

334 85 5188 154 588 802 1788 1856

122 420 4796 1104 1236 468 1300 584

5.94 0.96 3.34 0.35 1.28 0.44 1.02 0.25

∗ Including data from the Public Health Institute [5].

fol foil in a diffusion chamber), installed for three months in the main living area on the ground ﬂoor of randomly chosen houses, usually during spring. The survey was concentrated in the Walloon region, i.e., the Southern part of the country. No signiﬁcant radon risk was indeed expected in the Northern part, on the basis of the previously available data and of geological information. The ISIB data are short-term measurements collected between 1990 and 2001 in other houses with charcoal canisters (with diffusion barrier [4]) exposed for 3 to 4 days, usually on the ground ﬂoor, in partially conﬁned conditions (no opening of windows in the test room), at any season except summer. Radon is measured in equilibrium with its short-lived progeny by gamma-spectrometry with a NaI(Tl) detector.

3. The databases The two databases cover the Walloon region as well as the Brussels region; they also include a few measurements from the Flemish region, generally considered as containing no affected areas. They are globally of similar size, as shown in Table 1. It is seen that the FMH database is denser in radon-affected provinces (Luxembourg, Namur, Liege) whereas the ISIB database is somewhat closer to the population distribution, and could be a better tool for studying indoor radon in globally unaffected provinces. Hence, the two databases are not redundant but complementary.

4. Comparison between the databases 4.1. Method Because of the important variability of radon measurements from house to house, comparing long-term to short-term measurements in different houses is only meaningful if enough data from the same area are grouped and averaged, as is done for radon mapping. In the present work, we shall compare FMH and ISIB databases by grouping the ground ﬂoor data according

600

F. Tondeur, I. Gerardy

to the “arrondissement”, i.e., the Belgian equivalent to counties. In this way, we shall compare groups of at least 40 measurements, except in one case (the Mouscron county which in the FMH database only has 14 data). The discussion is limited to the 21 Walloon counties. It is well known that indoor radon data are usually well described by a log-normal distribution. A good way to compare different distributions is thus to show their logarithmic mean and logarithmic standard deviation (or equivalently, the exponential of these quantities, known as the geometrical mean and geometrical standard deviation). Published FMH data [1] give the geometrical mean for each county, as well as the percentage of houses with more than 400 Bq m−3 . From these data, and the corresponding ones from the ISIB database, we calculate the logarithmic standard deviation for the log-normal distribution giving the same percentage of houses above 400 Bq m−3 . 4.2. Mean values Figure 1(a) shows the ratio between short- and long-term logarithmic means for the 20 counties. The error bars correspond to one standard deviation. Globally, no signiﬁcant systematic increase of the mean concentration per county is observed in the short-term data. Hence, both databases should produce rather similar radon maps of the geometrical mean concentration, despite noticeable discrepancies for each county, and they could be merged for such a purpose. Note that a large part of the discrepancies may be explained by the statistical ﬂuctuations, but not all. Indeed, the ratio displayed in Fig. 1 differs from 1 by more than one standard deviation in 10 counties instead of the 6 or 7 expected from purely random ﬂuctuations. We have tried to avoid ﬂuctuations related to the differences between ISIB and FMH in the geographical distribution of the measurements in each county, by weighting each commune by surface or by population, but this was not found to improve the overall agreement between the databases. A further analysis of the discrepancies between them at the level of the communes has rather revealed spatial correlations associated with counties: applying a t-test to the differences between communal geomeans, most of the communes with high differences

(a)

(b)

Fig. 1. (a) Ratio between short-term and long-term logarithmic means radon concentration for the 20 Walloon counties (counties are in the same order as in Table 2) and (b) plot of short-term vs. long-term logarithmic mean.

Comparison between long-term and short-term measurements for indoor radon risk mapping

601

(t > 1) of a given sign are spatially grouped according to the counties. One possible origin of this correlation might be associated with the practical organisation of the FMH screening campaign on an administrative basis: all measurements from a given county were usually made in the same 3-month period by the same laboratory, but different counties may have been measured in different years and by different laboratories. Hence this unexpected spatial correlation could in fact reﬂect yearly variations and/or small calibration errors of the laboratories. The possible amplitude of these variations and of the differences in calibration of a few percents noted in intercomparisons [6] are big enough to largely justify the part of the variability noted in Fig. 1(a) that is not explained by random ﬂuctuations. However, the fact that time-related correlations might also affect the short-term measurements cannot be ruled out. 4.3. Standard deviations It is less obvious to use the short-term data to determine the affected areas on the basis of the usual deﬁnition (more than 1% of the houses above the reference level). Short-term data show more variability and thus larger percentages of high values, as shown in Table 2. A ﬁrst examination shows that counties having more than 1% long-term measurements above 400 Bq m−3 have more than 7% short-term measurements above this limit, with one exception (Marche), and all counties with more than 7% short-term measurements above the limit have more than 1% short-term measurements above it, with one exception (Philippeville). One method to determine affected areas could thus be a simple modiﬁcation of the threshold percentage above the reference level. Assuming log-normal distributions with identical logarithmic means and the average logarithmic standard deviations of longterm (0.77) and short-term data, respectively (1.07), it is straightforward to show that the value (67 Bq m−3 ) of the geometrical mean leading to 1% of long-term data above the reference level would give approximately 5% of short-term data above it. With this reference percentage, the overlap between the predictions of the two databases is a bit worse, as two additional counties would be predicted to be affected on the basis of ISIB data (Huy, Virton) but unaffected on the basis of FMH data. There is, of course, no reason why the two databases Table 2 Percentage of houses above 400 Bq m−3 according to the county and the database County Nivelles Ath Charleroi Mons Mouscron Soignies Thuin Tournai Huy Liege

FMH (long term) 0.4 0.04 0.15 0.04 0.00 0.11 0.28 0.01 0.73 0.65

ISIB (short term) 4.4 2.3 1.1 1.3 0.0 0.0 2.2 0.0 6.8 3.2

County

FMH (long term)

ISIB (short term)

Verviers Waremme Arlon Bastogne Marche Neufchateau Virton Dinant Namur Philippeville

2.34 0.82 0.79 14.82 1.13 9.46 0.32 1.16 0.38 0.27

10.5 1.7 2.7 13.6 4.0 24.2 6.8 8.6 2.3 10.8

602

F. Tondeur, I. Gerardy

Fig. 2. Logarithmic standard deviation by county: short-term vs. long-term.

Fig. 3. Logarithmic standard deviation vs. logarithmic mean for short-term data.

should ﬁnally lead exactly to the same predictions or even to the best overlap between the predictions. Another approach would be to evaluate the percentage of affected houses starting with the short-term logarithmic mean – supposed to be correct – and making some reasonable hypothesis for the logarithmic standard deviation. There is no strong correlation between the logarithmic standard deviations of the two databases for the different counties, as seen in Fig. 2. On the other hand, the short-term data show some correlation between the logarithmic standard deviation and the logarithmic mean, i.e., higher variability in affected counties (Fig. 3). The same trend is observed, but is less clear, in the long-term data. From ﬁrst principles, assuming that the time variability included in short-term measurements is statistically independent of the spatial variability present in both kinds of data, we could expect the data of Fig. 2 to satisfy a model like 2 2 2 ≈ σlong + σtime , σshort

where the last term could be approximated by a constant (the value of which is 0.6 from a least-squares ﬁt). However, the data do not allow discrimination between this assumption and 2 2 , where C 2 is close to 2. The two formulae allow evaluation ≈ C 2 · σlong other ones like σshort of a value of σlong equivalent to the observed σshort . Combined with the observed short-term logarithmic mean, these values allow us to evaluate the percentages above 400 Bq m−3 from the short-term data. A third option would be to use the average long-term logarithmic standard deviation (0.77, corresponding to a geometrical standard deviation of 2.15) together with the short-term logarithmic mean value. The latter option would lead to a particularly simple appli-

Comparison between long-term and short-term measurements for indoor radon risk mapping

603

cation, as the evaluated percentage of affected houses only depends on the geometrical mean: all areas with a geometric mean higher than 67 Bq m−3 would be predicted to be affected. The second option is found to be the best one. The ﬁrst one tends to amplify the contrast between affected and unaffected counties with respect to FMH data. The last one has the opposite effect. Simply reducing the short-term logarithmic standard deviation proves to be more effective. With this hypothesis, three counties are predicted to be affected whereas they are not in the long-term database, and one county is predicted not to be affected whereas it is in the same database, i.e., the same result as with the “5% rule” discussed above.

5. Conclusion The comparison of the two existing indoor radon databases for Southern Belgium shows that the short-term measurements (3–4 day exposure of charcoal canisters) may be adequate to evaluate the geometrical mean concentration, at least with the measurement procedure that has been applied by ISIB. As for the prediction of affected areas from these short-term measurements, two procedures are proposed: the ﬁrst is to deﬁne them as areas where more than 5% of short-term measurements are above the reference level (instead of the usual 1% for long-term measurements); the other is to evaluate the percentage of houses above the reference level by reducing the logarithmic standard deviation of the short-term measurements by a factor of 1.4 and assuming a log-normal distribution. As the variance of the short-term data is about twice that of the long-term data, a survey based on the short-term data should collect twice as much measurements to achieve the same statistical quality (this is a rough ﬁrst evaluation, as all the variability noted in Fig. 1 has not been explained). This may only be cost-effective if the short-term measurements are half the cost of the long-term measurements.

References [1] H.C. Zhu, J.M. Charlet, A. Pofﬁjn, in: 3rd Eurosymposium on Protection against Radon, Liège, Belgium, AIM, 2001. [2] P.N. Price, Health Phys. 69 (1995) 111. [3] M. Urban, E. Piesch, Radiat. Prot. Dosim. 1 (1981) 97. [4] B.L. Cohen, R. Nason, Health Phys. 50 (1986) 457. [5] F. Tondeur, I. Gerardy, D. Christiaens, S. Hallez, J.M. Flemal, Health Phys. 77 (1999) 697. [6] F. Tondeur, Ann. Assoc. Bel. Radioprot. 23 (1998) 175.

604

Cancer risks from radon in indoor air and drinking water in Sweden. The Swedish Radiation Protection Authority’s risk assessment L. Mjönes, R. Falk Swedish Radiation Protection Authority, SE-171 16 Stockholm, Sweden

In February 2001 the Swedish Radiation Protection Authority, SSI, presented an updated risk assessment for radon in indoor air and in drinking water. SSI now estimates that about 500 cases of lung cancer, 18 percent of the total incidence, are related to exposure to indoor radon in dwellings in Sweden. SSI deems it unlikely that the annual number of cases related to radon in Sweden should exceed 1000. The estimation is based primarily on residential case– control studies, particularly the Swedish national epidemiological radon study presented in 1993 and the recent study of non-smokers. There is a synergistic effect between tobacco smoking and radon, which implies that most of the radon-related lung cancer cases, almost 90 percent, occur among smokers. The risk for non-smokers is consequently much smaller than for smokers. Calculations show that up to 150 lung cancer cases per year can be avoided if remedial measures are taken in existing dwellings with radon concentrations exceeding 400 Bq m−3 (the action level for radon in existing dwellings in Sweden). Giving up smoking is the most effective measure an individual can take to reduce the risk of lung cancer from radon. The average effective dose to the Swedish population from consumption of waterborne radon is estimated to about 0.01 mSv per year. At 100 Bq L−1 , the action level for radon in public water supplies in Sweden, the annual effective dose is about 0.02 mSv per year and at 1000 Bq L−1 , the recommended upper level for private wells, the annual dose is about 0.2 mSv per year. Radon released from household water into the indoor air might be the origin of some tens of the annual number of radon-related lung cancer cases in Sweden. The risk following exposure to short-lived radon progeny present in the drinking water is estimated to be small. The risk estimation for intake of radon in drinking water is based primarily on the 1999 National Research Council report “Risk assessment of radon in drinking water”. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07073-1

© 2005 Elsevier Ltd. All rights reserved.

Cancer risks from radon in indoor air and drinking water in Sweden

605

1. Radon in indoor air 1.1. Introduction It is well known that radiation can cause lung cancer. This is evident from, for instance, the reports of the United Nations Scientiﬁc Committee on the Effects of Atomic Radiation (UNSCEAR). The latest UNSCEAR report was published in 2000 [1]. Several studies have shown that radon can cause lung cancer in animals, with or without the addition of tobacco smoke. Investigations of miners who have been exposed to radon, in uranium mines and other types of mines, show that radon can cause lung cancer and that the relationship between exposure and excess risk ﬁts a linear dose-response model, except for the highest exposure levels, at which the curve tends to level off. Mean exposure to radon in dwellings is considerably less than in mines, but high exposure levels in dwellings overlap lower levels in the miner studies. There is a reasonably sound scientiﬁc basis for assuming that the results from the miner studies can be extrapolated to the radon levels that normally occur in dwellings. A Swedish study of the relationship between radon in dwellings and lung cancer presented in 1993 shows, at levels that normally occur in dwellings, radon can cause lung cancer [2,3]. Results from the case–control studies in dwellings and the miner studies tally well. Since 1993 the results from the Swedish study have been conﬁrmed by several other residential case– control studies of indoor radon, including studies from the UK and Germany [4,5]. A metaanalysis of eight large indoor case–control studies presented in 1997 also supports the Swedish results [6]. 1.2. Earlier Swedish risk assessments In 1993 SSI assessed the health risks from radon in indoor air and in drinking water, as part of a programme of intensiﬁed measures against high radon levels in dwellings proposed to the Swedish Government. The number of expected lung-cancer cases related to radon was estimated at between 300 and 1500, with 900 as the most probable number. Most cases – some 85 percent – were thought to occur among smokers. 1.3. Swedish epidemiological study When it was presented in 1993 the Swedish study was the largest and most detailed study of its kind in the world. It included 1360 lung-cancer cases from the Swedish Cancer Register and two control groups of 2847 persons altogether. Exposure was assessed by means of radon measurements in the dwellings where the subjects had lived for two years or more between 1947 and to three years before diagnosis. Altogether, radon measurements were performed in almost 9000 dwellings. The results of the study showed that 16 percent of lung-cancer cases in Sweden were attributable to radon, with an uncertainty interval of 8–32 percent. At that time the incidence of lung cancer in Sweden was 2600 cases a year. Consequently, the number of radon-attributed lung-cancer cases in the Swedish population was estimated at 400 a year. The results also suggested strong interaction between radon exposure and smoking with regard to lung cancer [2]. The relative excess risk of lung cancer from exposure to radon from the study is seen in Fig. 1 [3].

606

L. Mjönes, R. Falk

Fig. 1. Relative excess risk of lung cancer from exposure to radon in the national Swedish residential-radon study [3].

1.4. Other residential-radon studies Over the last ﬁve or six years a fairly large number of residential case–control studies have been performed, in both Europe and North America. Eight case–control studies, each including more than 200 lung-cancer cases and assessing radon exposure by means of long-term radon measurements, have been subjected to a meta-analysis. The eight studies comprised a total of more than 4200 lung-cancer cases and 6600 controls [6]. The results from the metaanalysis were consistent with the results from the Swedish study. Since the presentation of the meta-analysis in 1997, results from some extensive European residential case–control studies have been presented. One of these was performed in the UK, in Cornwall and Devonshire, where the average indoor radon concentrations are much higher than in most other parts of the country. More than 1000 lung-cancer cases were included in the study [4]. The results are in accordance with the Swedish study. Two investigations, including a total of more than 2500 lung-cancer cases have been performed in Germany, one in the east and one in the west of the country [5]. The trends from these studies also tally with the Swedish results. The Iowa lung-cancer study that was presented in 2000 shows a higher excess relative risk than the residential case–control studies from Europe. This study included more than 400 female lung-cancer cases and 600 controls [7]. A research project to pool the most important European residential case–control studies is in progress. In total, 10 000 lung-cancer cases are included in the pooled analysis. 1.5. Miner studies In 1998 a committee BEIR VI (Committee on Biological Effects of Ionizing Radiation) from the US National Academy of Sciences presented the report Health effects of exposure to

Cancer risks from radon in indoor air and drinking water in Sweden

607

radon [8]. The report gives a comprehensive overview of the scientiﬁc basis for the health effects from residential radon. According to BEIR VI, the residential case–control studies still do not provide a sound statistical basis for a risk assessment of indoor radon. On the other hand, the BEIR committee has used the meta-analysis of eight case–control studies to support the assumption that the results from the miner studies can be extrapolated to the radon levels normally occurring in residential buildings. The International Commission on Radiological Protection (ICRP) assessed the health risks from indoor radon 1993 in its report Protection against radon at home and at work [9]. ICRP based its risk assessment on miner studies, but the Swedish study was thought to support the conclusion that radon constitutes a health risk for the general public. The Committee behind the BEIR VI report presents two risk models for indoor radon, based primarily on a pooled analysis of eleven studies of miners exposed to radon. If these two models were applied to Swedish data for residential radon it would mean that 23 or 33 percent of lung-cancer cases in Sweden were attributable to indoor radon exposure. Of the more than 2700 lung-cancer cases that occur in Sweden each year, roughly 600 or 900, depending on which of the models is used, would be attributable to radon in dwellings. The risk coefﬁcient presented by ICRP would imply more than 1100 expected lung-cancer cases annually related to indoor radon in Sweden. 1.6. The SSI standpoint Since its presentation in 1993, the Swedish epidemiological study has been supported by several other large residential case–control studies. The results from the meta-analysis of eight case–control studies mentioned above tally with the results of the Swedish study. One important result is that none of the studies involved predominates, i.e., excluding any one of the studies has no signiﬁcant impact on the estimated excess risk. In the last two years, results from major residential case–control studies in the UK and Germany have been presented that support the results from the Swedish study. Residential studies yield more relevant risk estimates for radon in dwellings than miner studies, since they relate directly to the population of interest. Most of the miners investigated were exposed to much higher radon levels than normally occur in dwellings. The duration of their exposure was also much shorter. Furthermore, the population in the miner studies is different from the general public. Almost all the miners investigated were males, and most of them were smokers. Unfortunately the results from the pooled European residential studies are not yet available. SSI nonetheless concludes that the residential case–control studies now collectively sufﬁce as a basis for assessing the health risks from indoor radon. According to the Swedish study, 16 percent of lung-cancer cases are attributable to residential radon. In 1998 the incidence of lung cancer in Sweden was 2720, indicating some 440 radon-attributed lung-cancer cases. The 95 percent conﬁdence interval would be roughly 200–900 cases. With regard to the higher risk estimates from the miner studies, the number of cases expected is rounded to 500 a year. SSI deems it unlikely that the annual number of cases related to radon should exceed 1000. The radon-related lung-cancer cases that occur are attributable to radon exposure since the 1960s. The number of lung-cancer cases in Sweden is still slowly increasing although smoking has decreased substantially in the past few decades. In the 1980s some 40 percent of the population were smokers. The proportion is now below 20

608

L. Mjönes, R. Falk

percent. If smoking continues to decrease, the number of lung-cancer cases related to radon exposure may be expected to fall in the future. 1.7. Risk for non-smokers Data from the miner studies indicate a higher excess relative risk from radon exposure for non-smokers than for smokers. However, assessment of risk for non-smoking miners is based on a small number of cases and is consequently highly unreliable. The case–control studies published to date provide no evidence for such a difference between non-smokers and smokers. In 2000 a Swedish study, Residential radon and lung cancer among never-smokers in Sweden [10], was presented by the Institute of Environmental Medicine at Karolinska Institute in Stockholm. The study included 436 “never-smoker” lung-cancer cases and 1650 controls. The study shows that exposure to radon elevates lung-cancer risk in never-smokers. The excess relative risk for never-smokers is estimated at 0.10 per 100 Bq m−3 . This is roughly the same ﬁgure as in the Swedish study from 1993, where most of the cases were smokers. However, an increase in radon-related risk was observed only for those subjects who had been exposed to environmental tobacco smoke at home. For those with no environmental tobacco smoke at home, no excess relative risk with increasing radon concentrations was found. These ﬁndings are, however, still uncertain and need conﬁrmation through other studies. It is important to note that the absolute risk of lung cancer from exposure to radon is much smaller for non-smokers than for smokers, Fig. 2. The majority of radon-related lung cancers, almost 90 percent, occur among smokers. ICRP has calculated the risk of late effects at ﬁve percent per sievert. Using the risk coefﬁcients for non-smokers and smokers derived from the Swedish case–control studies, the corresponding “dose” can be calculated. At 108 Bq m−3 , the average radon concentration in Swedish dwellings, a non-smoker’s risk of death from lung cancer corresponds to an effective dose of 0.3 mSv a year. For a smoker, the risk at 108 Bq m−3 corresponds to about 5 mSv a year. At 400 Bq m−3 the corresponding doses would be 1.3 mSv a year for a non-smoker and 18 mSv a year for a smoker. In this example we have deﬁned a “smoker” as a person who smokes some 10 cigarettes per day. For heavy smokers the risk is higher.

Fig. 2. Lifetime lung-cancer risk for smokers and never-smokers.

Cancer risks from radon in indoor air and drinking water in Sweden

609

1.8. Lung-cancer cases that can be avoided by remedial action SSI has calculated that some 35 percent of radon-related lung-cancer cases are caused by exposure in dwellings with average radon concentrations exceeding 400 Bq m−3 , and 20 percent by exposure in dwellings with radon levels between 200 and 400 Bq m−3 . Further calculations have shown that if remedial measures were taken in all dwellings with radon levels exceeding 400 Bq m−3 , roughly 150 lung-cancer cases a year could be avoided. Through remedial action in all dwellings with radon levels between 200 and 400 Bq m−3 a further 50 cases annually could be avoided. For the calculations it has been assumed that the mean radon concentration in the remediated dwellings is 100 Bq m−3 . Many lung-cancer deaths can be avoided each year by means of radon-reducing measures in dwellings. The majority of deaths occur among smokers. Giving up smoking is the most effective measure an individual can take to reduce the risk of lung cancer from radon. Simultaneously, risks of a number of other smoke-related diseases are reduced. 2. Radon in drinking water 2.1. Introduction The dose calculations and risk assessments concerning consumption of radon in drinking water that have been performed are based on a small number of investigations of adults after a unit intake of radon-rich water. To our knowledge, no epidemiological studies of the relationship between intake of waterborne radon and cancer have been conducted. It is unlikely that such a study could show excess risk. 2.2. Previous Swedish risk assessments In its 1997 Drinking Water Ordinance, based on a risk assessment by SSI, the Swedish Food Administration issued action levels for radon in drinking water. The SSI risk assessment, in turn, was based on a risk model from the UK National Radiation Protection Board (NRPB), presented in the 1993 UNSCEAR report [11]. According to this model the effective doses from intake of radon in drinking water are signiﬁcantly higher for infants and children than for adults, mostly owing to differences in body weight. At a radon concentration of 1000 Bq L−1 in household water, an adult’s annual effective dose would be approximately 0.5 mSv while a 10-year-old child’s dose would be some 1.5 mSv, and that to an infant as much as 7 mSv. SSI pointed out in 1993 that these dose calculations were based on a small number of experimental investigations and that no epidemiological studies had been published. 2.3. US National Academy of Sciences risk assessment In 1999 a National Academy of Sciences (NAS) committee issued its Risk assessment of radon in drinking water [12]. The committee had made a thorough survey of the literature on waterborne radon. Although they performed no new experiments or new measurements, they presented two new theoretical models for dose calculations – one for the diffusion of radon in the stomach and one for the behaviour of radon dissolved in the blood and other tissues.

610

L. Mjönes, R. Falk

The report conﬁrms earlier knowledge. For example, it conﬁrms that the mean transfer coefﬁcient from radon in drinking water to indoor air is 10−4 . Accordingly, if the radon concentration in drinking water is 1000 Bq L−1 , the average contribution to indoor air is 100 Bq m−3 . Radon from drinking water is absorbed in the stomach, and most of the radon leaves the body through exhalation air. This process is surprisingly fast: most of the radon has left the body within an hour of consumption. The major site of long-term retention is adipose tissue. The committee estimates the dose coefﬁcient for waterborne radon at 3.5 nanosievert per Bq, roughly one-third of the estimate for adults in the UNSCEAR report from 1993 (10 nanosievert per Bq). The lower effective dose in the NAS model than in the NRPB model follows from the assumption that the radon is absorbed in the stomach wall and then distributed among other organs according to the blood ﬂow they receive. Earlier models have assumed that the radon is absorbed in the small intestine. Based on the new models, the committee has also investigated the risk from short-lived radon progeny in water. The analysis shows that the radon progeny cannot diffuse into the stomach wall and, consequently, that the alpha particles cannot reach the cells in the stomach wall that are at the greatest risk of developing into cancer cells. This analysis is supported by experiments at SSI. In mice that had been fed water containing no radon gas but only radon progeny, no radon progeny was found in the blood, kidney or liver. This means that radon progeny present in the water at intake probably poses no elevated cancer risk. The key conclusions are that the estimated risk from intake of waterborne radon is somewhat smaller than previous assessments indicate, and that short-lived radon progeny probably constitutes no risk. However, it is important to bear in mind that the NAS estimation is not based on any new experimental studies. 2.4. SSI’s risk assessment SSI concludes that the NAS report is the best available basis for an assessment of cancer risk from consumption of waterborne radon. The mean radon concentration in drinking water in Sweden has been estimated at roughly 40 Bq L−1 [13]. The average effective dose to the Swedish population would then be less than 0.01 mSv a year. Theoretically, a few cancer deaths a year from stomach cancer might be expected in Sweden from consumption of waterborne radon. Radon released from household water into indoor air constitutes a few percent of total exposure to radon in indoor air in Sweden, and may be the cause of some tens of the radon-related lung-cancer cases occurring each year in this country. The basis for assessing the risk from consumption of waterborne radon is still inadequate – far more so than for radon in indoor air. To obtain a more reliable risk assessment, more research in this ﬁeld is desirable. References [1] UNSCEAR (United Nations Scientiﬁc Committee on the Effects of Atomic Radiation), Sources and Effects of Ionizing Radiation, UNSCEAR 2000 Report to the General Assembly, with Scientiﬁc Annexes, United Nations, New York, 2000. [2] G. Pershagen, G. Åkerblom, O. Axelson, B. Clavensjö, L. Damber, G. Desai, A. Enﬂo, F. Lagarde, H. Mellander, M. Svartengren, G.A. Swedjemark, Residential radon exposure and lung cancer in Sweden, N. Engl. J. Med. 330 (1994) 159–164.

Cancer risks from radon in indoor air and drinking water in Sweden

611

[3] F. Lagarde, G. Pershagen, G. Åkerblom, O. Axelson, U. Bäverstam, L. Damber, A. Enﬂo, M. Svartengren, G.A. Swedjemark, Residential radon and lung cancer in Sweden: Risk analysis accounting for random error in the exposure assessment, Health Phys. 72 (2) (1997). [4] S. Darby, E. Whitley, P. Silcocks, B. Thakrar, M. Green, P. Lomas, J. Miles, G. Reeves, T. Fearn, R. och Doll, Risk of lung cancer associated with residential radon exposure in south-west England: a case–control study, Br. J. Cancer 78 (3) (1998) 394–408. [5] H.E. Wichmann, J. Heinrich, M. Gerken, M. Kreuzer, J. Wellmann, G. Keller, L. Kreienbrock, Domestic radon and lung cancer – current status including new evidence from Germany, in: Excerpta Med., Int. Congress Ser., vol. 1225, 2002, pp. 247–252. [6] J.H. Lubin, J.D. Boice, Lung cancer risk from residential radon: Meta-analysis of eight epidemiological studies, J. Natl. Cancer Inst. 89 (1) (1997). [7] R.W. Field, D.J. Steck, B.J. Smith, C.P. Brus, E.L. Fisher, J.S. Neuberger, C.E. Platz, R.A. Robinson, R.F. Woolson, C.F. Lynch, Residential radon gas exposure and lung cancer, The Iowa Lung Cancer study, Am. J. Epidemiol. 151 (11) (2000). [8] Committee on Health Risks of Exposure to Radon, Board on Radiation Effects Research, Commission on Life Sciences, National Research Council, Health Effects of Exposure to Radon: BEIR VI, National Academy Press, Washington, DC, 1999. [9] ICRP Publication 65: Protection against radon-222 at home and at work, Ann. ICRP 23 (2) (1993). [10] F. Lagarde, G. Axelsson, L. Damber, H. Mellander, F. Nyberg, G. och Pershagen, Residential radon and lung cancer among never-smokers in Sweden, Epidemiology 12 (4) (2001). [11] UNSCEAR (United Nations Scientiﬁc Committee on the Effects of Atomic Radiation), Sources and Effects of Ionizing Radiation, UNSCEAR 1993 Report to the General Assembly, with Scientiﬁc Annexes, United Nations, New York, 1993. [12] Committee on Risk Assessment of Exposure to Radon in Drinking Water, Board on Radiation Effects Research, Commission on Life Sciences, National Research Council, Risk Assessment of Radon in Drinking Water, National Academy Press, Washington, DC, 1999. [13] J. Kulich, H. Möre, G.A. Swedjemark, Radon and radium in household water (in Swedish, summary in English), SSI Report 88-11, Swedish Radiation Protection Authority, Stockholm, Sweden, 1988.

612

Review of seasonal variation in residential indoor radon concentrations H. Arvela Radiation and Safety Authority – STUK, Box 14, 00881 Helsinki, Finland

The study explored the theoretical basis of radon entry into buildings and processes affecting the seasonal variation in indoor radon concentration. The radon source of the model comprises two components, diffusive and convective entry rate. The force driving the convective ﬂow is the pressure differential caused by the indoor–outdoor temperature difference. The physics of diffusive radon entry into buildings, of airﬂow from soil, driven by the pressure differences, and of air inﬁltration explain to a considerable extent the observed seasonal variation in indoor radon concentration. The seasonal variation in houses with natural ventilation has been estimated for both typical Northern European climate conditions and for warmer conditions in Central and Southern Europe. In houses with soil gas radon as the dominant radon source a summer minimum is typically 50% of the winter maximum indoor radon concentration. Ventilation through open windows in summertime may result in signiﬁcantly lower concentration relative to theoretical calculations. The cold climate in North Europe increases radon concentrations by 50% compared to countries in South Europe. The annual average to winter radon concentration ratio was typically 0.7–0.8. This is in agreement with the measured seasonal variations.

1. Introduction Building materials and radon-bearing soil gas are the dominant sources of indoor radon. The main removal process decreasing indoor radon concentration is the air-exchange of indoor spaces. Both indoor–outdoor temperature difference and wind contribute to the pressure difference in the envelopes of residential buildings which is the driving force for air leakage providing ventilation to those buildings. It is noteworthy that the same pressure difference is the driving force also for airﬂow from soil into indoor spaces. Modelling these phenomena provides a tool for understanding the seasonal variation in indoor radon concentration. The aim of this study is to review observations on seasonal variation as well as to compare the measurements with simple model predictions. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07074-3

© 2005 Elsevier Ltd. All rights reserved.

Review of seasonal variation in residential indoor radon concentrations

613

2. Literature review Studies of seasonal and diurnal variation in indoor radon concentration have been carried out, ﬁrst, in order to reach a general comprehension of the radon problem, entry mechanisms and reasons for elevated radon concentrations; and, second, in order to evaluate the validity of short-term measurements and to estimate the annual average radon concentration. Seasonal variations are affected by many factors, which vary from country to country. House characteristics, building soil and geology, ventilation practices and climatic conditions are the most important factors. Generally in conditions of a temperate or cold climate, winter concentrations are higher than summer concentrations. Patterns of air movement on karstic limestone [1] and on eskers [2] may result in anomalous seasonal variations. An annual average radon concentration is the main quantity used in estimates of the radiological effects of radon in homes. It reﬂects an average of the radon concentrations in all living spaces in a home. However, measurements taken over the course of an entire year have seldom been possible in practice. Short-term measurements are simpler to take and provide faster results. The expectations of the home-owners as well as the measurement methods used and recommendations by the authorities have led to national practices of using a measurement period of 2 days–3 months in most countries. Correction factors for annual average radon concentration are needed to adjust measurements taken over periods other than twelve months. Majborn has reported highly signiﬁcant variation in slab-on-grade houses [3]. The mean annual average/winter concentration ratio in all 67 houses with varying types of foundation was 0.77. The ratio was lowest in houses with slab-on-grade. Statistical studies made using radon measurements from the radon-affected area of southwest England were used to estimate seasonal correction factors [4,5]. The factors obtained result in a change, in the estimated annual average, of up to 35% for measurements taken over a six-month period and 56% for measurements taken over a three-month period.

3. Materials and methods The radon source of the model [6] comprises two components, diffusive and convective entry rates, Sd and Ssoil . The force driving the convective ﬂow is the pressure differential p caused by the indoor–outdoor temperature difference T , this phenomenon designated as the stackeffect. Ssoil is proportional to the soil gas radon concentration Asoil and ﬂow rate of soil gas into the house Qsoil , which in turn is the quotient of the stack pressure difference ps at ﬂoor level and the ﬂow resistance Rt . Wind and the resulting pressure difference also affect the radon entry rate. Compared to the stack effect the contribution of this steady-state wind effect is of minor importance and has been discounted in this study. The Lawrence Berkeley Laboratory model [7] was used in air inﬁltration predictions. It adds in quadrature the inﬁltration rates due to indoor–outdoor temperature differences (stack effect) and wind speed (v), as follows: 2 1/2 Q = ELA fsr T 0.65 + (fwr v)2 (1) , N = QV −1 ,

(2)

614

H. Arvela

where Q is the total inﬁltration, ELA is the effective leakage area, T is the indoor–outdoor temperature difference, fsr is the stack parameter and fwr is the wind parameter. N is the air exchange rate and V the house interior volume. The fraction of leakage in the ﬂoor and ceiling affects both the stack and wind parameters. The detailed choice of the parameters has been presented in Ref. [6]. Equation (3) gives the radon concentration, ARn , used in model calculations. ARn = Asoil Fsoil T (Tin Tout )−1 + Sd (λ + N )−1 V −1 + Aout . (3) The factor Asoil Fsoil (Tin Tout )−1 demonstrates the total convective radon entry rate per unit temperature difference, Asoil being the radon concentration of the leakage air from soil. λ is the radioactive decay constant of 222 Rn, 0.0076 h−1 . Aout is the outdoor air radon concentration, the annual average being about 5 Bq m−3 . The effective leakage area, ELA, is a standard measure of building tightness, which is measured by pressurising a building with a fan. ELA is deﬁned assuming that in the pressure range characteristic of natural inﬁltration (−10 to +10 Pa) the ﬂow versus pressure behaviour of a building more closely resembles the square root (turbulent rather than viscous ﬂow). The model calculations were made using the parameters estimated for a typical Finnish house [6]. The diffusive radon entry for the house is 1500 Bq h−1 . The parameters for the convective radon entry have been set to a value, which results in the Finnish climate in an annual average radon concentration of 90 Bq m−3 , where the contribution of the diffusive and convective radon entry is 20 and 70 Bq m−3 . Model calculations were carried out also for this house located in different European areas. In this simpliﬁed study indoor temperature has been set at a constant level of 20 ◦ C, irrespective of the location and outdoor temperature. The monthly average temperatures were used. A wind speed of 3 m s−1 has been used for locations throughout the year.

4. Results Figure 1 shows the calculated seasonal variation for the Finnish model house located in different European areas. Due to the dominance of convective radon entry, radon concentrations are higher the lower outdoor temperature is. Figure 2 shows the diffusive and convective contributions to indoor radon concentrations for the model house in Helsinki and Athens. In Athens outdoor temperature exceeds the supposed constant indoor temperature of 20 ◦ C in summertime. Therefore, the convective entry is reduced to zero level. The air exchange is in this model house proportional to the absolute value of the indoor–outdoor temperature difference. In the case of a pure diffusion source the decreasing outdoor temperature decreases the indoor radon concentration. In the case of a ﬂow driven by pressure difference, the decreasing outdoor temperature increases the indoor radon concentration. It is noteworthy that in the case of pressure difference driven ﬂow the indoor radon concentration increases simultaneously when air inﬁltration increases. Soil airﬂow is proportional approximately to the indoor–outdoor temperature difference, T , whereas the ventilation rate is proportional to the envelope leakage exponent of T , which has a range of 0.5–0.7; a value of 0.65 has been used in this study. Table 1 shows the annual average indoor radon concentrations for the climatic conditions of several European cities. The annual average in southern Europe is 30–40% lower than in

Review of seasonal variation in residential indoor radon concentrations

615

Fig. 1. Seasonal variation in indoor radon concentration of a Finnish model house in different European climates.

Fig. 2. Diffusive and convective contribution to indoor radon concentration for a Finnish model house in Helsinki and in Athens.

Helsinki. Table 1 also gives the quotient of the annual average radon concentration to the December–March average radon concentration. This correction factor is needed to obtain the annual average radon concentration from wintertime results. Although the annual average radon concentrations vary signiﬁcantly, the seasonal correction factor varies only in the range of 0.68 to 0.81, being lowest in the warm areas. For Central and North Europe the factor is approximately 0.80. Similar experimental results on seasonal variation have been observed in Norway [8], in Denmark [3] and in the radon-affected areas of England [4,5]. In Norway an average correc-

616

H. Arvela Table 1 Annual average indoor radon concentrations and correction factors for the annual average for a Finnish model house located in different European climates City

Annual average indoor radon concentration (Bq m−3 )

Annual average/December–March, radon concentration ratio

Athens Madrid Milan Paris Frankfurt Hamburg Stockholm Helsinki

54 67 71 79 81 86 93 96

0.68 0.72 0.71 0.79 0.77 0.81 0.80 0.81

tion factor of 0.75 has been applied for winter measurements [8]. In Finland a factor of 0.8 has been used as a representative average [6].

5. Conclusions The results demonstrate the theoretical seasonal variation for a model house with a dominating convective radon entry, located in different European areas. The annual average to winter radon concentration ratio was typically 0.7–0.8. This is in agreement with the measured seasonal variations. The summer minimum is typically 40–50% of the winter maximum. Ventilation through open windows in summertime further lowers summertime radon concentrations, which has not been taken into account in the model. The simpliﬁed calculations were made using a constant wind speed of 3 m s−1 in all areas although local wind conditions may vary signiﬁcantly. The cold climate in North Europe increases radon concentration by 50% compared to countries in South Europe. Generally the physics of diffusive radon entry into buildings, of airﬂow from soil, driven by the pressure differences, and of air inﬁltration explain to a considerable extent the observed seasonal variation in indoor radon concentration.

References [1] D.L. Wilson, R.B. Gammage, C.S. Dudney, R.J. Saultz, Summertime elevation of Radon-222 levels in Huntsville, Alabama, Health Phys. 60 (2) (1991) 189–197. [2] H. Arvela, Voutilainen, T. Honkamaa, A. Rosenberg, High indoor radon variations and the thermal behaviour of eskers, Health Phys. 67 (3) (1994) 254–260. [3] R. Majborn, Seasonal variation of radon concentrations in single family houses with different sub-structures, Radiat. Prot. Dosim. 45 (1) (1992) 443–447. [4] J.C.H. Miles, B.M.R. Green, P.R. Lomas, Radon affected areas: Derbyshire, Northamptonshire and Somerset, Docs NRPB 3 (4) (1992) 18–28. [5] J. Pinel, T. Fearn, S.C. Darby, J.C.H. Miles, Seasonal correction factors for indoor radon measurements in the United Kingdom, Radiat. Prot. Dosim. 58 (2) (1995) 127–132.

Review of seasonal variation in residential indoor radon concentrations

617

[6] H. Arvela, Seasonal variation in radon concentration of 3000 dwellings with model comparisons, Radiat. Prot. Dosim. 59 (1) (1995) 33–42. [7] M. Sherman, M. Modera, Comparison of measured and predicted inﬁltration using the LBL inﬁltration model, LBL-17001, Lawrence Berkeley Laboratory, California, 1984. [8] NRPA, Radon in dwellings. Recommendations for measurements indoors and recommendations for investigation on building site, NRPA Radiation Protection Series No. 2, Norwegian Radiation Protection Authority, Osteraas, 1992.

618

Developments in radon-safe building in Finland H. Arvela a , J. Bergman b , R. Yrjölä b , J. Kurnitski c , M. Matilainen c , P. Järvinen d a Radiation and Safety Authority – STUK, Box 14, 00881 Helsinki, Finland b Laboratory of Structural Engineering, Helsinki University of Technology, Box 2100, 02015 Espoo, Finland c Laboratory of HVAC, Helsinki University of Technology, Box 4400, 02015 Espoo, Finland d Katepal Inc., Box 33, 37501 Lempäälä, Finland

Slab-on-grade is the prevalent substructure in Finnish low-rise residential buildings. The gap between the foundation wall and ﬂoor slab is the main entry route for radon-bearing air from soil into living spaces. A new construction for an airtight joint between the foundation wall and ﬂoor slab was developed for houses with slab on ground or houses with basement. The construction prevents the ﬂow of radon-bearing air from soil into the house. Based on the results, new guidance for wide use in Finland is under preparation. In the new sealing practice, bitumen felt will be installed underneath the ﬂoor slab in direct contact with a concrete slab. Both laboratory and ﬁeld tests showed that bitumen felt adheres well to concrete. Low indoor radon concentrations (20–60 Bq m−3 ) were achieved in seven of the eight test houses although most houses were located in areas where 50% of houses exceed 200 Bq m−3 . The underpressure in houses with mechanical exhaust ventilation was typically 7–10 Pa, which was remarkably higher than in houses with supply and exhaust ventilation, 1–5 Pa. This difference affects also the indoor radon concentrations. Underpressures of 10–40 Pa were measured in apartments. This signiﬁcantly increases ﬂow of radon-bearing air from soil into indoor spaces. In ﬂats with no air intakes, signiﬁcant reductions in radon concentration are achievable through installation of new air intakes. Mechanical supply/exhaust ventilation seemed to be the only safe ventilation solution in order to control the indoor radon concentration in radon prone areas.

1. Introduction High radon concentrations of indoor air in low-rise residential buildings create an important national health problem in Finland. The target level for new buildings is 200 Bq m−3 . Decrease in the use of crawl space in house foundations has increased the indoor radon concentration of Finnish housing stock during recent decades. The prevailing type of foundation is slab-onRADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07075-5

© 2005 Elsevier Ltd. All rights reserved.

Developments in radon-safe building in Finland

619

Fig. 1. Installation of bitumen felt to the joint of foundation wall and ﬂoor slab.

grade (Fig. 1). In this type of foundation the ﬂow of radon-bearing air from soil into living spaces through gaps between the foundation wall and ﬂoor slab should be prevented through special measures. This study presents results of a development project “Radon-safe foundation, moisture prevention and air exchange in a healthy building”, funded by the Finnish Technology Agency. The aim of the project was to develop simple construction practices which can be utilised for prevention of these leakage ﬂows and which simultaneously provide quality moisture prevention for the foundation structures [1–3]. Several building companies participated in the project. The project aimed at development of radon-related technical practices by the companies. The effect of activated sub-slab radon piping on the drying process of the ﬂoor slab was tested in two houses. The study focuses also on the applicability of different ventilation strategies, the control of depressure in dwellings and the use of fresh air vents for reduction of indoor radon concentration. One of the objectives of the study was to ﬁnd improvements to air intake vents in ﬂats with mechanical exhaust ventilation and to develop the radon-safe foundation construction practices of blocks of ﬂats.

2. Materials and methods The construction of the foundations of 20 houses was followed up. In the ﬁrst phase 10 houses were studied. In 8 of these the house owner participated actively in the building work. In the second phase the studies were made in Cupertino with building companies and suppliers of element based single family houses. All measures relating to the air-tightness of the founda-

620

H. Arvela et al.

Fig. 2. Bitumen felt installed to the joint of foundation wall and ﬂoor slab before casting of the ﬂoor.

tion as well as difﬁculties and drawbacks were documented. When the houses were ready and in normal use indoor radon concentration, air exchange rate and depressure were measured. The effect of activated sub-slab radon piping on the drying process of the ﬂoor slab was tested in two houses. As a sub-project, air exchange rate, depressure and indoor radon concentration in 80 new houses built in the Building Fair 2000 area were measured. Air exchange results will be considered in this paper. The effect of activated sub-slab piping on the drying process of the ﬂoor slab was tested in two houses. Humidity probes were installed both into the ﬂoor slab and sub-ﬂoor gravel layer. The moisture content of the slab and sub-slab gravel was measured over 3–20 months. The inﬂux of radon-bearing soil air was studied in 7 recently constructed ﬂats on the lowest ﬂoor and with ﬂoor slab in ground contact. In some of the houses, sub-ﬂoor radon piping had been installed as a preparatory method in order to be able to reduce the radon levels. Through depressurisation of the piping, using a fan, the leakage of radon-bearing soil can be prevented. Testing the indoor radon concentration with both a nonactivated and activated fan provided a method of estimating the effect airﬂows from soil to indoor radon concentration.

3. Results and discussion 3.1. Radon-tight foundation The radon-tight construction practices published in the guidance of the Ministry of Environment were studied in the project [4]. In addition, a new more easy-to-do construction was studied. The construction given in the guidance requires a high level of carefulness. The results of this study suggest that the sealing of the joint between the foundation wall and ﬂoor slab fails often in real building site conditions. A new construction for a radon-tight joint between the foundation wall and ﬂoor slab was developed for houses with slab on ground or with basement. The results were very promising.

Developments in radon-safe building in Finland

621

In the new construction a wide (0.5–1.0 m) strip of bitumen felt is installed above the foundation wall. The other edge of the strip is installed underneath the concrete slab in direct contact with the casting. The key advantage is the exclusion of the need for elastic sealants. The tests made in the laboratory ascertained that the adhesion of the felt to the concrete was good. Indoor radon concentrations in 7 test houses varied in the range of 20–60 Bq m−3 . Five of the seven houses were located in an area where 50% of houses exceed 200 Bq m−3 . In the 8th house where the new method was tested radon concentration was 220 Bq m−3 . However, this was a low value in an area where 90% of houses exceed 200 Bq m−3 (median 590 Bq m−3 ). Lack of the sealing of lead-throughs was considered as the main reason for the remaining radon leakage. In the sealing of the lead-throughs the methods presented in the guidance of the Ministry of Environment were shown to be working. Practical experience showed that special attention should be paid to the development of guidance on sealing the water pipes and electric wiring, which are installed in protective piping beneath the ﬂoor slab. Above-slab installation would be preferable. The results showed that airtight construction is needed in the whole building area. Omitting the sealing work, for example, in storerooms or in rooms for house technology may lead to increases in radon concentrations in living spaces. In these auxiliary spaces radon concentrations were in some cases 5000–10 000 Bq m−3 when the sealing was omitted. 3.2. Air exchange and underpressure The measurements in the 40 low-rise residential houses resulted in air exchange rates of 0.3– 0.7 L h−1 , when the adjustment was in the normal position. However, the possibilities for control of ventilation were in some cases limited. The maximum ventilation rate was in some houses 0.5 L h−1 or the minimum was 0.7 L h−1 . The underpressure in houses with mechanical exhaust ventilation was typically 7–10 Pa, which was remarkably higher than in houses with supply and exhaust ventilation, 1–5 Pa. This difference also affects the indoor radon concentrations. Mechanical supply/exhaust ventilation seemed to be the only safe ventilation solution in order to control the indoor radon concentration in radon prone areas. In one of the houses the pressurisation remarkably distorted the operation of the sub-slab-depressurisation system used for decreasing the indoor radon concentration. High underpressures of 10–40 Pa were measured in apartments [5], which signiﬁcantly increase the ﬂow of radon-bearing air from the soil into indoor spaces. The tests showed that underpressure and indoor radon concentration in the apartments could be reduced by 50% by installing air intakes. Due to the pressure drop in the air-intake, greater reductions in radon concentration cannot normally be achievable. A new type of air-intake was developed in the project. According to the results of the simulation, only a limited pressure drop can be achieved by installing a second air intake into rooms. Higher reductions in radon concentration can be achieved through installation of a new supply and exhaust ventilation system or through other remedial measures. The normal sealing procedures used by the builders in none of the cases provided a complete barrier against the leakage of radon bearing soil air into the apartments.

622

H. Arvela et al.

3.3. Effect of sub-ﬂoor ventilation on ﬂoor slab drying Radon piping installed below the slab can be utilised in improving the drying process of the ﬂoor slab [6]. The study focused on the improved drying process of the ﬂoor slab through the use of sub-slab ventilation. A signiﬁcant difference of several weeks in the drying time was observed as a result of the ventilation. Simultaneously, signiﬁcant amounts of water were transported out from the sub-ﬂoor soil through the piping. In certain cases radon piping can also be utilised in connection with water damage.

4. Conclusions and outcomes • A new construction for an airtight joint between the foundation wall and ﬂoor slab was developed for houses with slab on ground or houses with basements. The construction prevents the ﬂow of radon-bearing air from soil into the house. • The results will be utilised when improving the present guidance material. • The bitumen felt used in the construction works also as a qualiﬁed moisture insulator. • A new guide to be published in the Finnish RT-Building Information File is under preparation. • The practice will be recommended for wide use in the whole country. • The key detail in the construction is the correct application of bitumen felt to the joint between foundation wall and ﬂoor slab. • The companies that participated in the project have renewed their radon-technical practices. • Depressure caused by mechanical air exchange should be taken into account in the planning of both low-rise residential buildings and ﬂats. Mechanical exhaust ventilation may signiﬁcantly increase both depressure and indoor radon concentration. • In ﬂats with no air intakes, signiﬁcant reductions in radon concentration are achievable through installation of new air intakes. • Mechanical supply/exhaust ventilation seemed to be the only safe ventilation solution in order to control the indoor radon concentration in radon prone areas. • The possibilities of controlling the air exchange rate in low-rise residential buildings are sometimes limited. • Radon piping to be installed below the slab can be utilised in reduction of indoor radon concentration and also in reduction of moisture in ﬂoor slab and sub-slab gravel. • In certain cases radon piping can be utilised also in connection with water damage.

References [1] H. Arvela, A.-V. Kettunen, J. Piironen, J. Kurnitski, K. Jokiranta, Review of radon-safe building in Finland, in: Proceedings of Healthy Buildings 2000, vol. 3, Finnish Society for Indoor Air Quality and Climate, Helsinki, 2000, pp. 69–74. [2] H. Arvela, Experiences in radon-safe building in Finland, Sci. Total Environ. 272 (2001) 169–174. [3] H. Arvela, K. Kuuspalo, J. Bergman, R. Yrjölä, J. Kurnitski, K. Jokiranta, M. Matilainen, P. Järvinen, Radonsafe foundation, moisture prevention and air exchange in a healthy building, in: Seminar on Indoor Air, Espoo, Finnish Society for Indoor Air Quality and Climate, 2002, pp. 65–68 (in Finnish).

Developments in radon-safe building in Finland

623

[4] Radon-safe building, foundation construction, Guide 2/1993, Ministry of Environment, 1994 (in Finnish). [5] J. Kurnitski, K. Jokiranta, M. Matilainen. Improvement of mechanical exhaust ventilation. Radon, leakage air and supply air, Report B65, Helsinki University of Technology, Laboratory of Heating, Ventilating and Air Conditioning (in Finnish, abstract in English). [6] A.-V. Kettunen, J. Piironen, Drying the concrete slab on ground and foundation soil using sub-slab suction, in: Proceedings of Healthy Buildings 2000, vol. 3, Finnish Society for Indoor Air Quality and Climate, Helsinki, 2000, pp. 219–224.

624

Stochastic radon lung dosimetry – modeling variability of bronchial cellular doses∗ R. Winkler-Heil, W. Hofmann Institute of Physics and Biophysics, University of Salzburg, Hellbrunner Str. 34, A-5020 Salzburg, Austria

Since the tracheobronchial tree consists of an asymmetrically dividing stochastic branching system and sensitive target cells are nonuniformly distributed in relation to depth in bronchial epithelium, target cell populations may experience signiﬁcant variations in absorbed dose. In the present study, stochastic modeling techniques have been used to simulate the inherent variability of morphological and physiological parameters on particle deposition and clearance and, ﬁnally, on bronchial dose. The stochastic dose calculations indicate that the inherent variability of biological parameters involved in radon lung dosimetry may cause signiﬁcant statistical ﬂuctuations of cellular doses. Computations for deﬁned inhalation conditions demonstrate that epithelial cell doses in the bronchial region can reasonably be approximated by log-normal distributions, with geometric standard deviations as high as 8 for the unattached fraction and up to 3 for the attached fraction.

1. Introduction Calculations of alpha particle doses delivered to sensitive bronchial target cells after inhalation of 222 Rn and its progeny are commonly based on the following assumptions: (i) The bronchial region of the human lung is represented by a sequence of symmetrically bifurcating straight cylindrical tubes, i.e., all airways in a given generation have identical dimensions. As a result of this, particle deposition fractions, particle clearance fractions, and, ﬁnally, alpha-emitting 218 Po and 214 Po surface activities are the same in all airways of a given bronchial generation. (ii) Furthermore, sensitive target cells in bronchial epithelium are assumed to be uniformly distributed throughout given depth intervals. Hence, such deterministic models produce single dose values for basal and secretory cells in a given airway generation. In reality, however, the tracheobronchial tree consists of an asymmetrically dividing stochastic * This research was supported in part by CEC Contract FIGH-CT1999-00005.

RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07076-7

© 2005 Elsevier Ltd. All rights reserved.

Stochastic radon lung dosimetry – modeling variability of bronchial cellular doses

625

branching system [1] and sensitive target cells are nonuniformly distributed in relation to depth in bronchial epithelium [2]. Due to these asymmetries and variabilities, target cell populations may experience signiﬁcant variations in absorbed dose. In the present study, stochastic modeling techniques have been used to simulate the inherent variability of biological parameters on radon progeny deposition, clearance and resulting bronchial dose.

2. Methods An existing stochastic deposition model has recently been extended by a bronchial clearance and cellular dosimetry model to develop a fully stochastic dosimetry model for inhaled radon progeny. This dosimetry model has been used in the present study to explore the effects of biological variability on bronchial doses. The stochastic deposition model [3,4] simulates the random walk of inspired particles through a random airway structure [2]. Upon inhalation, linear airway dimensions are randomly selected by Monte Carlo methods from their distributions and correlations at each airway bifurcation. The probability that a particle selects the major or the minor daughter airway at a given bifurcation is proportional to the splitting of the airﬂow, i.e., proportional to the related cross-sections. Deposition of particles in bronchial airways is computed by analytical equations, i.e., the deposition probabilities of individual particles is given by their average probability. In case of a deposition event, the particle continues its path with decreased statistical weight. Deposition fractions for bronchial airway generations are typically based on 10 000 to 100 000 simulations. The stochastic bronchial clearance model considers both fast and slow bronchial clearance phases [5]. Due to conservation of mass, average mucus velocities in a given airway generation in asymmetrically branching airways are proportional to their respective diameters [6] and mucus velocities in individual bronchial airways are normalized to a tracheal mucus velocity of 5.5 mm min−1 [5]. Mucus delay at carinal ridges of airway bifurcations is considered by a random delay time, which is presently selected from a uniform distribution between 0 and 10 min. The dependence of the magnitude of the slow bronchial clearance fraction fs , with a half-time of 10 days, on geometric particle diameter is modeled by an empirically derived relationship [6]. Half times for transport through epithelium into blood are assumed to be 10 h, for the attached fraction, and 1 h, for the unattached fraction, respectively [7]. Alpha emitting 214 Po and 218 Po surface activities, obtained by dividing the activity retained in a given generation by the total mean surface area of that generation, are subsequently normalized to an exposure of 1 WLM. The dosimetric part of the stochastic model is modeled after an already existing deterministic cellular dose model [8]. Secretory and basal cells are considered to be the critical target cells for cancer induction [5]. The mean depths of the critical cells in the epithelium are based on data by Mercer et al. [2]; their correlation with airway diameter is approximated by a polynomial function. Dose as a function of depth in bronchial epithelium is computed for uniform 218 Po and 214 Po activities on cylindrical bronchial airway surfaces, considering both near and far wall contributions. The distinction between bronchial and bronchiolar airways is based

626

R. Winkler-Heil, W. Hofmann

on a tube diameter of 0.34 cm, i.e., airways with smaller diameters belong to the bronchiolar region and airways with diameters larger than 0.34 cm belong to bronchial airways [9]. Weighted epithelial doses are obtained by weighting the doses to basal and secretory cells in each airway generation by their relative frequencies [2]. For the calculation of committed equivalent doses, an apportionment factor (bronchial: bronchiolar:acinar) of 0.33:0.33:0.33, a tissue weighting factor for the lung of 0.12, and a radiation weighting factor for alpha particles of 20 was used [10]. All frequency distributions presented in this paper are conditional probability density functions, i.e., they are normalized to a total probability of one and all zero values are omitted. 3. Results 3.1. Deposition fractions Deposition has been calculated for a male adult nose breather, applying the standard breathing characteristics for sitting awake, i.e., a tidal volume of 750 ml, an inspiration time of 2.5 s, no breath hold and an expiration time of 2.5 s [5]. The frequency distributions of deposition fractions are plotted in Fig. 1 for 1 and 200 nm unit density particles, representing deposition probabilities of unattached and attached radon

Fig. 1. Probability distributions of the deposited fractions of 1 nm unit density particles in generations 4 (A) and 10 (B), and for 200 nm unit density particles in generations 4 (C) and 10 (D).

Stochastic radon lung dosimetry – modeling variability of bronchial cellular doses

627

Fig. 2. Probability distributions of the surface activity of 218 Po for 1 nm unit density particles in generations 4 (A) and 10 (B), and for 200 nm unit density particles in generations 4 (C) and 10 (D).

progeny, respectively, in airway generations 4 (bronchial region) and 10 (bronchiolar region). While the frequency distributions for the 200 nm particles can reasonably be approximated by log-normal distributions, the corresponding simulations for the 1 nm particles suggest the existence of a bimodal distribution. 3.2. Surface activities The frequency distributions of the steady-state surface activity for 218 Po nuclides in generations 4 and 10 for 1 and 200 nm radon progeny aerosols are shown in Fig. 2 for two selected airway generations. The homogeneously distributed activity on the surface of the airways is normalized to an exposure of 1 WLM. The frequency distributions for the surface activities resemble those for the deposition fractions displayed in Fig. 1, exhibiting again log-normal distributions for 200 nm particles and bimodal distributions for 1 nm particles. 3.3. Dose to critical cells Dose conversion factors for the critical cells, i.e., secretory cells and basal cells, in selected airway generations 4 and 10 are compiled in Table 1 for unattached and attached particle sizes, normalized to a surface activity of 1 Bq cm−2 .

628

R. Winkler-Heil, W. Hofmann

Table 1 Statistical parameters of the dose conversion factors (nGy s−1 per Bq cm−2 ), normalized to unit surface activity for 218 Po and 214 Po nuclides, for airway generations 4 and 10 and for secretory and basal cells (SD: standard deviation; CV: coefﬁcient of variation; N : number of non-zero values).

218 Po

1 nm

200 nm

214 Po

1 nm

200 nm

Secretory, gen 4 Secretory, gen 10 Basal, gen 4 Basal, gen 10 Secretory, gen 4 Secretory, gen 10 Basal, gen 4 Basal, gen 10 Secretory, gen 4 Secretory, gen 10 Basal, gen 4 Basal, gen 10 Secretory, gen 4 Secretory, gen 10 Basal, gen 4 Basal, gen 10

Mean

SD

Max.

Min.

N

CV (%)

85.9 135.8 16.1 106.5 85.9 136.1 17.6 107.0 140.6 174.9 69.6 140.0 140.6 175.2 69.6 140.3

3.3 54.9 29.9 73.1 2.8 55.0 29.6 73.1 2.1 40.6 7.2 56.1 1.7 40.7 7.1 56.2

224.5 294.9 196.3 388.7 149.5 290.1 121.8 358.2 241.5 298.3 219.5 376.3 183.9 294.4 162.9 350.9

85.7 85.0 0.0 0.0 85.7 85.0 0.0 0.0 140.4 139.5 67.7 67.0 140.4 139.5 67.7 67.0

10 000 10 000 422 8221 10 000 10 000 405 8222 10 000 10 000 10 000 10 000 10 000 10 000 10 000 10 000

3.8 40.4 185.8 68.7 3.3 40.4 168.4 68.3 1.5 23.2 10.4 40.1 1.2 23.3 10.2 40.1

Fig. 3. Mean weighted epithelial doses, produced by 218 Po and 214 Po alpha particles, for 1 nm unit density particles in generations 4 (A) and 10 (B), and for 200 nm unit density particles in generations 4 (C) and 10 (D).

Stochastic radon lung dosimetry – modeling variability of bronchial cellular doses

629

Table 2 Median and geometric standard deviation (GSD) of the mean weighted epithelial dose (mGy WLM−1 ), produced by 218 Po and 214 Po alpha particles, as a function of generation number, for 1 and 200 nm unit density particles Generation No.

1 nm Median

GSD

200 nm Median

GSD

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

22.7 36.2 38.5 31.4 27.1 21.7 16.9 11.1 6.87 3.58 2.32 0.96 0.38 0.14 0.059 0.033 0.020 – – –

1.81 2.04 2.14 2.17 1.97 1.93 1.85 1.9 2.43 3.31 4.3 5.4 6.36 6.98 7.21 7.76 5.01 – – –

0.65 1.77 2.08 1.88 1.83 1.82 1.88 1.86 2.16 2.59 3.94 4.54 5.03 5.51 6.28 6.81 7.33 9.67 11.40 12.90

4.50 2.22 2.12 2.25 2.22 2.24 2.36 2.53 2.60 2.67 2.39 2.25 2.24 2.21 2.29 2.41 2.69 2.60 3.09 3.08

3.4. Epithelial dose Figure 3 presents mean weighted epithelial doses, which are obtained by weighing the computed basal and secretory cell doses by the volumetric densities of their cell nuclei in bronchial epithelium [2]. The volumetric density is generally higher for basal cells than for secretory cells in the bronchial airways (0.83 vs. 0.17 for generations 1–5, and 0.85 vs. 0.15 for the remaining bronchial airways). In the bronchiolar airways, no basal cells have been found and therefore only secretory cells are considered as critical cells. The relative contributions of unattached and attached radon progeny to the epithelial dose are based on the following ratios: 218 Po : 214 Pb : 214 Bi/214 Po equals 1.0 : 0.1 : 0.0 for the unattached fraction, and 0.8 : 0.4 : 0.2 for the attached fraction. The distribution of the mean weighted epithelial doses for unattached and attached radon progeny among human bronchial and bronchiolar airways is listed in Table 2.

4. Discussion The main source of the biological variability affecting deposition, clearance and cellular dosimetry for inhaled radon progeny is the asymmetry and variability of linear airway dimensions, such as diameters, lengths, branching and gravity angles, and their correlations,

630

R. Winkler-Heil, W. Hofmann

e.g., cross-section ratios, branching angle to diameter ratio, or the termination probability determining the number bronchial generations along each particle’s paths [1]. The variability of deposition fractions in individual bronchial airways primarily reﬂects the variability of geometrical parameters, affecting deposition probabilities, and the variability of the number of particles transported along a given path. The resulting relative frequencies can be approximated by log-normal distributions, at least for higher generation numbers. The variability of surface activities, i.e., retained fractions of deposited nuclides per unit surface area in a given generation is caused not only by the variability of the geometrical parameters, but also by the variability of the implemented slow clearance mechanisms in bronchial airways. The variability of cellular doses results from the variability of surface activities due to the above variations of particle deposition and clearance, and from variations of the thickness of the epithelium as a function airway diameter, and from the depth distributions of target cells across epithelial tissue. Several dose distributions display bimodal shapes, in which the lower part of the distribution corresponds to the range of dose conversion factors in the bronchial region and the higher part to those in the bronchiolar region. In the present study, all distributions are ﬁtted to log-normal distributions. In conclusion, our stochastic dose calculations indicate that the inherent variability of several biological parameters involved in lung dosimetry may cause signiﬁcant statistical ﬂuctuations of cellular doses. In addition to a more realistic determination of average cellular doses, a stochastic dosimetry model also provides quantitative information about the statistical uncertainty of dose calculations. While bronchial epithelial cell doses in a given generation can reasonably be approximated by log-normal frequency distributions with geometric standard deviations ranging from 2 to 3, cellular dose distributions for the whole bronchial tree, i.e., considering airways in all bronchial airway generations, can have geometric standard deviations as high as 5. The incorporation of systematic individual variations, e.g., speciﬁc lung volumes or breathing patterns, into lung dosimetry may even further increase the width of the dose distributions. Recent developments of biologically based carcinogenesis models have indicated that dose–response relationships are not linear over the whole range of doses or exposures. In case of nonlinearity, the full dose distribution must be used to produce correct risk estimates, instead of average values commonly predicted by deterministic model. The signiﬁcance of the role of dose variability on lung cancer risk has previously been demonstrated through the application of a state-vector model of carcinogenesis to lung cancer risk [11].

References [1] [2] [3] [4] [5] [6] [7]

L. Koblinger, W. Hofmann, Phys. Med. Biol. 30 (1985) 541. R.R. Mercer, M.L. Russell, J.D. Crapo, Health Phys. 61 (1991) 117. L. Koblinger, W. Hofmann, J. Aerosol Sci. 21 (1990) 661. W. Hofmann, L. Koblinger, J. Aerosol Sci. 21 (1991) 675. ICRP Publication 66: Human respiratory tract model for radiological protection, Ann. ICRP 24 (1–3) (1994). R. Sturm, W. Hofmann, G. Scheuch, K. Sommerer, P. Camner, M. Svartengren, Ann. Occup. Hyg., in press. J. Marsh, A. Birchall, W. Hofmann, R. Bergmann, Risk Assessment of Exposure to Radon Decay Products. Work Package 2, Lung Modelling Group. Final Report: CEC Contract No. F14P-CT95-0025, 1999.

Stochastic radon lung dosimetry – modeling variability of bronchial cellular doses

631

[8] R. Winkler-Heil, W. Hofmann, in: W. Burkart, et al. (Eds.), High Levels of Natural Radiation and Radon Areas: Radiation Dose and Health Effects, Elsevier, Amsterdam, 2002, p. 169. [9] R. Bergmann, W. Hofmann, L. Koblinger, J. Aerosol Sci. 28 (Suppl. 1) (1997) S433. [10] ICRP Publication 65: Protection against radon-222 at home and at work, Ann. ICRP 23 (2) (1993). [11] D.J. Crawford-Brown, W. Hofmann, Radiat. Prot. Dosim. 28 (1989) 283.

632

Stochastic state-vector model of radiation carcinogenesis applied to radon-induced lung cancer risk∗ D.J. Crawford-Brown a , W. Hofmann b a Department of Environmental Sciences and Engineering, University of North Carolina at Chapel Hill,

Chapel Hill, NC 27599-1105, USA b Institute of Physics and Biophysics, University of Salzburg, Hellbrunner Str. 34, A-5020 Salzburg, Austria

A biophysical multi-stage state-vector model (SVM) of radiation carcinogenesis has been extended to incorporate stochasticity of transitions and spatial inhomogeneity of cellular doses. Dose-rate dependent cellular transitions between the stages are related to formation of double stranded DNA breaks, repair of breaks, interactions (translocations) between breaks, ﬁxation of breaks, cellular inactivation, stimulated mitosis and promotion through loss of intercellular communication. Each of these transitions from the normal state 0 to the tumour state 7 is simulated stochastically in time through Monte Carlo sampling of the competing events. The stochastic SVM has been applied here to in vitro transformation frequencies by monoenergetic alpha particles and to in vivo lung cancer incidence in uranium miners and laboratory rats exposed to radon progeny. Predictions of the transformation frequency per surviving cell compare favourably with the experimental in vitro data over a wide range of LETs and doses. When incorporating in vivo features of cell differentiation, stimulated cell division and heterogeneity of cellular doses, fair agreement could be obtained between model predictions and lung cancer data from human epidemiological studies as well as from rat inhalation experiments. The model predicts a nonlinear dose–response relationship at low doses, producing a risk, which is approximately a factor 2 smaller at 20 WLM than current risk estimates. The effect of cigarette smoke on the promotion of intermediate cells initiated by alpha radiation produces an increased risk at low exposures, eventually saturating at high exposure levels. 1. Introduction This paper presents a stochastic version of a state-vector model of radiation carcinogenesis, rooted in processes of initiation, promotion and progression modelled previously by * This research was supported in part by EU Contract FIGH-CT1999-00005.

RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07077-9

© 2005 Elsevier Ltd. All rights reserved.

Stochastic state-vector model of radiation carcinogenesis applied to radon-induced lung cancer risk

633

deterministic differential equations [1,2]. Previous formulations of the model, aside from differing from the present with respect to stochasticity, did not fully reﬂect processes related to differentiation of cells and the differences in dose distribution between stem and non-stem cells of the lung tissue. Those previous formulations showed success in predicting correctly in vitro transformation frequency using biologically realistic parameter values, and cancer incidence at low to intermediate doses of radiation in vivo. They did not, however, correctly predict cancer incidence at higher doses typically found in the more highly exposed cohorts of miners [3]. The reasons for this failure might be inappropriate model speciﬁcation (i.e., biological and radiobiological processes are missing or incorrectly speciﬁed in the model); unjustiﬁed extrapolation of some key parameter values from in vitro to in vivo settings; and/or the use of a deterministic rather than stochastic formulation of the model. This paper presents the results of applying a stochastic formulation of the state vector model, adjusted to reﬂect the presence of differentiation and inhomogeneity of doses between stem and non-stem cells, in predicting lung cancer incidence following exposure to radon progeny in humans and rats.

2. Model formulation It is assumed that cells must pass through stages of initiation, promotion and progression to produce a tumor. Initiation refers in the model to cells with a stable, inheritable, alteration in the genome as a result of two double-stranded DNA breaks, followed by translocation of one (an oncogene) onto a region of the DNA under signiﬁcantly less control (the second break). The oncogene-associated break is referred to here as a speciﬁc break, while the other break is referred to as a nonspeciﬁc break. Both breaks are assumed to occur at random in the DNA, with separate ﬁrst-order rate constants (probability per unit dose). If mitosis takes place before this interaction, repair enzymes repair the genome back to its original state with complete efﬁcacy (i.e., mis-repair is not considered here). Repair unrelated to mitosis may take place at any time to repair the speciﬁc break. It is assumed that only the oncogene-associated break is subject to this mode of repair, since it is located in an active gene. The interaction between breaks proceeds with a ﬁxed probability per unit time. Following the interaction, it is necessary for the translocation to undergo ﬁxation, or incorporation into the genome as a permanent feature. The result is a fully initiated cell. Many of the mitotic events, however, result in death of the cell rather than ﬁxation. Only some fraction of the mitoses result in a viable, initiated, cell. Promotion is assumed to occur through the loss of contact inhibition in a cell, and to result in the expansion of an initiated cell into a preneoplastic colony. Both reversible and irreversible promotion is considered here [4–6]. Reversible promotion applies only while the dose-rate of radiation (and/or cigarette smoke in this paper) is non-zero, while irreversible promotion applies even after the dose-rate is removed. Both forms of promotion result from loss of contact inhibition between initiated and surrounding cells. This loss of contact inhibition is induced through inactivation of the cells surrounding an initiated cell (in the case of radiation) or through interference with other forms of signalling (by cigarette smoke). An initiated cell is promoted if at least 4 of the surrounding 6 cells in an epithelial layer lose contact inhibition. The rate of mitosis of that initiated cell then rises above the rate of spontaneous cell death and the colony grows to a preneoplastic lesion.

634

D.J. Crawford-Brown, W. Hofmann

Progression is assumed to result in the growth of the preneoplastic lesion to a fully lethal tumor. There is little evidence of an effect of radiation on this transition, and so it is not considered here to be a function of radiation dose or dose rate. A cell in any state may be inactivated either by radiation or by spontaneous processes, which are assumed here to be governed by ﬁrst-order kinetics for both radiation-induced and background components. The result is removal of the cell from the stream of transitions. Mitosis in a cell containing no chromosomal damage results in two identical cells with no damage. In most cases, mitosis in a cell with one or both forms of chromosomal damage (either speciﬁc or nonspeciﬁc or both) prior to interaction results in repair and the production of two identical normal cells. Mitosis in a cell with both forms of chromosomal damage (speciﬁc and nonspeciﬁc) and following interaction either are inactivated or result in the production of two cells containing the permanent chromosomal alteration (a probability for each of these two pathways is provided). Mitosis in initiated cells results in two identically initiated cells. The resulting model, showing all transitions, is shown in Fig. 1. A cell in any state may be inactivated either by radiation or by spontaneous processes (background rate). The result is removal of the cell from the stream of transitions. Mitosis in a cell containing no chromosomal damage results in two identical cells with no damage. In most cases, mitosis in a cell with one or both forms of chromosomal damage (either speciﬁc or nonspeciﬁc or both) prior to interaction results in repair and the production of two identical normal cells. Mitosis in a cell with both forms of chromosomal damage (speciﬁc and nonspeciﬁc) and following interaction either are inactivated or result in the production of two cells containing the permanent chromosomal alteration (a probability for each of these two pathways is provided). Mitosis in initiated cells results in two identically initiated cells. The resulting model, showing all transitions, is shown in Fig. 1.

Fig. 1. The stochastic model of radiation carcinogenesis used in the present research. Boxes indicate states of a cell and arrows indicate transitions or transfers between states. A full tumor is equivalent to cells in State 7; transformed cells for in vitro studies are equivalent to State 6 colonies.

Stochastic state-vector model of radiation carcinogenesis applied to radon-induced lung cancer risk

635

Ni (t) is the number of cells in State i at any time t during irradiation and/or exposure to cigarette smoke. The state of the cellular community at this time is given by the vector [N0 (t), N1 (t), N2 (t), N3 (t), N4 (t), N5 (t), N6 (t), N7 (t)]. The number of undamaged cells is N0 (t). The number of cells with the speciﬁc break is N1S (t) and the number with only the nonspeciﬁc break is N1NS (t). The number of cells with both the speciﬁc and nonspeciﬁc breaks, but with no interaction or translocation, is N2 (t). The number of cells with both the speciﬁc and nonspeciﬁc breaks, and with interaction but not ﬁxation, is N3 (t). The number of cells with ﬁxation of the breaks is N4 (t); this also is the number of initiated cells in the model. The number of reversibly promoted cells is N5 (t), and the number of irreversibly promoted cells is N6 (t). The number of progressed cells is N7 (t). Cell inactivation by alpha particles in all states is treated as a ﬁrst-order process with rate constant kd . Inactivated cells are removed through lysis on a time scale short compared to the total length of simulation. It is assumed here that each inactivation/removal of a cell in either States 0, 1S , 1NS , 2, 3 or 4 in an in vivo population causes mitosis in another, surviving, cell. This stimulus is not present in in vitro populations until the colony has reached conﬂuence; i.e., until conﬂuence, cell inactivation and lysis does not cause mitosis in surrounding cells. Incorporation of mitosis into the model depends upon cell inactivation and removal (through lysis) of cells in the stem cell population, as well as cells in a differentiated population supported by that stem cell population. The rate constant for inactivation is assumed the same in both populations. The ratio of the number of cells in the differentiated pool over the number of cells in the stem cell pool is assumed constant at Rd/s . In addition, it is assumed that N stages of differentiation occur, with each differentiation during mitosis producing two identical differentiated cells. Therefore, for each division of a stem cell, 2N −1 differentiated cells are produced (each stem cell division produces one stem cell and one cell passing into the ﬁrst stage of differentiation); inversely, each inactivation/removal event in the differentiated pool causes 2−(N−1) mitoses in the stem cell population. Reversibly promoted cells (State 5) are assumed to have an imbalance between the rates of mitosis and inactivation/removal. The difference between the mitotic rate M5 and the rate of inactivation/removal kd in this population is non-zero. Each inactivation results in the eventual removal of the cell from the sequence of states leading to a tumor. Each mitosis results in two State 5 cells. The colony will continue to grow in size until it becomes nutrient limited. At that point in time, assumed to occur when the colony is 1 × 106 cells [1,2], the difference between the mitotic rate and the rate of inactivation/removal becomes zero and the colony ceases to expand. In addition, cells in State 5 may revert back to State 4. This reversion takes place if the irradiation and/or exposure to cigarette smoke are discontinued. For cells remaining in State 5, there is a further ﬁrst-order process causing transitions of individual cells in the colony to State 6, or irreversible promotion. This concept was demonstrated ﬁrst for chemical carcinogenesis, and is assumed here to apply also to radiation-induced and cigarette-induced carcinogenesis [7]. Once a cell is in State 6, the condition that contact inhibition is caused by the presence of surrounding inactivated cells is removed. As a result, State 6 cells do not revert back to either State 5 or State 4 when irradiation ceases. The same imbalance between the mitotic rate and the rate of inactivation/removal α already noted in State 5 cells is also applied to State 6 cells.

636

D.J. Crawford-Brown, W. Hofmann

Finally, it is assumed that State 7 cells result from random mutations of State 6 cells in a colony. It is further assumed that the probability per unit time of an irreversibly promoted colony yielding a full tumor is proportional to the number of State 6 cells in that colony at that time, with a lag time in vivo equal to TL . For these in vivo simulations, the lag time is set equal to 15 years. For cases of in vitro irradiation, the model assumptions are largely the same as those outlined above with two exceptions. First, differentiation is not considered and the cells are not divided into a stem and differentiated population. Second, inactivation and lysis of cells does not stimulate mitotic events in the surviving cells during growth to conﬂuence, since cells during that period already are released from control on growth. Once a colony has reached conﬂuence, however, all of the conditions applying to the in vivo model are in effect with the exception of differentiation (i.e., there is one mitotic event in a surviving cell per inactivation/lysis in surrounding cells). The model structure consists of a series of IF-THEN statements for the various possible events that might occur for a cell in each state. The probability of any event leading to transfer between two states equals the ratio of the probability of that transfer over the sum of the probabilities of all processes leading from that state or leaving the cell in that original state. All cells begin the simulation in State 0. A random number generator was used to generate a random value distributed uniformly in the interval [0, 1]. A given transfer process in a timestep increment was assigned a sub-interval in [0, 1] proportional to the ratio of its probability of occurrence divided by the sum of the probabilities of all processes for that State. If the random number lay in this subinterval, that process was selected for that time step. The timestep increment throughout the simulation was t, taken to be 1 min in the in vitro simulations and 10 minutes in the in vivo simulations. All rate constants used in calculating the transfer probabilities were taken from the previous papers that used the deterministic form of the model [1–3].

3. Results Figure 2 shows the predicted relative number of State 6 colonies following irradiation by alpha particles at a range of dose-rates for an irradiation period of one hour compared to results of

Fig. 2. The results of the model applied to in vitro transformation of C3H 10T1/2 cells as described in the text.

Stochastic state-vector model of radiation carcinogenesis applied to radon-induced lung cancer risk

637

in vitro transformation experiments with C3H 10T1/2 cell [8]. The value of kd was selected to correspond to a D0 value of 1.6 Gy for plated rat tracheal epithelial cells [9]; k67 is set equal to zero since this transition is not present in in vitro cases. Predictions of the transformation frequency per surviving cell compare favourably with the experimental in vitro data over wide range of linear energy transfers (LET) and doses. Figure 3 shows the predicted relative number of State 7 colonies following irradiation by alpha particles at a range of exposures to simulate in vivo exposure of rats to alpha particles via radon progeny, as conducted at Battelle [10]. The exposure lengths were identical in each exposure group to those used in the experiments. The dosimetry used in these calculations has been described elsewhere [3]. A single layer of epithelial cells, with stem and non-stem cells mixed at identical depths, was assumed. Figure 4 shows the predicted relative number of State 7 colonies following irradiation by alpha particles at a range of exposures to simulate in vivo exposure of humans to combinations of alpha particles and cigarette smoke, using the uranium miners as the epidemiological base [11]. The lower curve assumes no exposure to cigarette smoke, and so only the background promotional probability was assumed based on the rat data. The upper curve shows the case in which cigarette smoke is assumed to double the background promotional probability. No attempt has been made here to relate this doubling to a speciﬁc level of smoking. Alpha dose-rates correspond to the mean basal cell dose-rate in lung tissue from exposure to radon progeny by these miners [12]. The numerical value of k67 was established by requiring that the background incidence of tumor, assumed here equal to the background probability of a State 7 colony, equals 0.3. An irradiation period of 4 years was used in all calculations. The value of Rd/s was assumed equal to 2.

Fig. 3. The results of the model applied to in vivo exposure of rats to radon progeny as described in the text. Data were selected to cover a range similar to that in the human studies. The results of the current model are shown as the dashed line. Results of the deterministic model, with equal cellular doses to stem and non-stem cells, and without differentiation, are shown as the solid line. Note the signiﬁcant improvement in ﬁt.

638

D.J. Crawford-Brown, W. Hofmann

Fig. 4. The results of the model applied to in vivo exposure of humans to radon progeny as described in the text. The results of the current model without cigarette smoke are shown as the dashed line. Results of the current model with a doubling of the promotional rate constant due to cigarette smoke are shown as the solid line. Note the two curves converge at high exposures due to the saturation of the promotional effect.

At low doses, the SVM predicts a nonlinear dose-response relationship. If compared with a lifetime lung cancer risk estimate for chronic exposure of 3 × 10−4 per WLM [13] and assuming a linear – no threshold – relationship, the predicted excess tumor probability at 200 WLM is slightly higher than the epidemiologically derived value (0.07 vs. 0.06). At 20 WLM, however, the computed lung cancer risk is only about half of the ICRP value (0.003 vs. 0.006). The apparent nonlinearity of the model predictions at low exposure levels is caused primarily by the nonlinear promotion term and the effects of repair of DNA damage prior to interaction and ﬁxation.

4. Discussion The model presented here represents an improvement over previous formulations based on deterministic principles [3], at least for the case of in vivo exposures. The in vitro exposures appear to be equally well ﬁt by both models. The likely explanation for this ﬁnding is that the improvements in the model ﬁts are due less to the stochastic formulation than to the incorporation of several biological factors in the present model. These are, in particular, the separation of the cell population into stem and non-stem cells (although this is important only for the human population, since these two populations receive the same doses in the rat lungs) and to the incorporation of differentiation in the present model. It is important to note that the quality of these ﬁts is not due simply to a large number of ﬂexible parameters, since all but one parameter (in the in vitro case) and two parameters (in the in vivo case) were constrained by more fundamental experimental data. In the case of in vivo exposures, the two unconstrained parameters were the rate constant from reversibly promoted to irreversibly promoted cells, and the rate constant from irreversibly promoted to progressed cells. In the latter case, the constant was determined by constraining the model to produce the correct background incidence of cancer, and so even this parameter was constrained before ﬁtting to the miner data. The model also shows clearly the effect of cigarette smoking on cancer incidence following irradiation. The effect arises here from an increase in the transition rates between initiated and

Stochastic state-vector model of radiation carcinogenesis applied to radon-induced lung cancer risk

639

both reversibly and irreversibly promoted cells. This is in agreement with the multiplicative, or slightly sub-multiplicative, model of the interaction between radon progeny and smoking found in epidemiological studies.

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13]

D.J. Crawford-Brown, W. Hofmann, Int. J. Radiat. Biol. 57 (1990) 407–423. D.J. Crawford-Brown, W. Hofmann, Math. Biosci. 115 (1993) 123–144. M. Mebust, D.J. Crawford-Brown, W. Hofmann, H. Schoellnberger, Reg. Toxicol. Pharmacol. 35 (2002) 72–79. J. Trosko, C. Chang, B. Madhukar, Radiat. Res. 12 (1990) 241–251. J. Trosko, C. Chang, B. Madhukar, J. Klaunig, Pathobiol. 58 (1990) 265–278. J. Bertram, Radiat. Res. 123 (1990) 252–256. R. Schulte-Hermann, I. Timmermann-Trosiener, G. Bartel, W. Bursch, Cancer Res. 50 (1990) 5127–5135. R. Miller, S. Marino, D. Brenner, S. Martin, M. Richards, G. Randers-Pehrson, E.J. Hall, Radiat. Res. 142 (1995) 54–60. D. Thomassen, F. Seiler, L. Shyr, W. Grifﬁth, Int. J. Radiat. Biol. 57 (1990) 395–405. F. Cross, Radiat. Prot. Dosim. 24 (1988) 463–466. E. Kunz, J. Sevc, V. Placek, J. Horacek, Health Phys. 36 (1979) 699–721. National Research Council, Comparative Dosimetry of Radon in Mines and Homes, National Academy Press, Washington, DC, 1991. ICRP Publication 65: Protection against radon-222 at home and at work, Ann. ICRP 23 (2) (1993).

640

Biophysical mechanisms and radiation doses in radon therapy∗ H. Tempfer a , A. Schober a , W. Hofmann a , H. Lettner a , F. Steger b a Institute of Physics and Biophysics, University of Salzburg, Hellbrunner Str. 34, A-5020 Salzburg, Austria b Division of Health Physics, Department of Radiation Protection, Austrian Research Centers Seibersdorf,

A-2444 Seibersdorf, Austria

In the Gastein valley, Austria, radon-rich thermal water and air have successfully been used for over 100 years for the treatment of various forms of rheumatic diseases. To explore the therapeutic role of radon progeny adsorbed to the skin, their activities on the skin of patients exposed to thermal water (bathtub) and hot vapour (vapour chamber) were measured by alpha spectrometry. Average total alpha activities on the patients’ skin varied from 1.24 to 4.09 Bq cm−2 in the bathtub, and between 1.05 and 2.58 Bq cm−2 in the vapour bath, exhibiting signiﬁcant ﬂuctuations among different parts of the human body and among different patients. While radon progeny adsorption was hardly affected by the degree of blood circulation and greasiness of the skin, a signiﬁcant dependence on the pH-value could be observed, with a distinct maximum deposition around 6 to 7. The penetration of radon progeny into the skin was measured by removing several cell layers, revealing a roughly exponential activity distribution in the upper layers of the skin. The transport of radon through the skin into the blood was determined via radon concentration measurements in the exhaled air of the patients. Calculations of doses to different target organs suggest that radon progeny attachment on skin surfaces may play a major role in the therapeutic response for both treatment schemes, while bronchial doses lie well within the natural ﬂuctuations of the doses received by the population, thus causing no additional radiation hazard to the patients undergoing such treatments.

1. Introduction In radiation protection, the word “radon” is commonly associated with exposure to radon (and its short-lived decay products) in homes and its potential to cause lung cancer at sufﬁciently high levels, e.g., at those found in uranium mines [1]. It is therefore interesting to note that * This research was supported in part by the Forschungsinstitut Gastein–Tauernregion, Projects F PK 109 and 110.

RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07078-0

© 2005 Elsevier Ltd. All rights reserved.

Biophysical mechanisms and radiation doses in radon therapy

641

radon has successfully been used for decades as a therapeutic agent in so-called radon spas, such as Bad Gastein and Bad Hofgastein in Austria, particularly for the treatment of rheumatic diseases [2]. Despite its medically certiﬁed application in radon spas, there is still insufﬁcient information on the relative efﬁciencies of different treatment regimes and the underlying biophysical mechanisms of radon and radon progeny uptake. Two different therapeutic exposure regimes are presently used in Badgastein and Bad Hofgastein: (i) bathing in hot thermal water, and (ii) exposure to hot radon vapour in an exposure chamber or in the thermal gallery, a former mine. In the thermal bath, patients are sitting in a bathtub (37 ◦ C) for a period of 20 min (10 times during the complete treatment) with a radon concentration of about 900 Bq L−1 . Measurements of the radon concentration in the exhaled air indicated that a certain fraction of the radon content in water was taken up by the blood after diffusion through the skin, transported to the lungs (and other organs) via the bloodstream, and then exhaled. In addition, short-lived radon progeny are adsorbed on the skin of the patients. In the vapour bath, patients are exposed to the hot radon vapour (37 ◦ C) in a small chamber for a period of 20 min (10 times during the complete treatment) with a radon concentration ranging from 30 to 200 Bq L−1 (average value: 90 Bq L−1 ). These exposure conditions are similar to those found in the thermal gallery, with the notable exception that the head of the patient remains outside the exposure chamber, thus avoiding inhalation of radon and its progeny (note: the inhalation route is the primary contributor to the organ doses in the thermal gallery). In addition, short-lived radon progeny are adsorbed on the skin of the patients. Two philosophies have been advanced in recent years to explain the observed therapeutic effects [3]: (i) a direct radiation effect in speciﬁc organs of the human body, e.g., regeneration of damaged tissue in the affected organ, through the incorporation of radon (via skin or the lungs), and (ii) an indirect radiation effect, such as induced immune response through irradiation of the deeper layers of the skin by radon progeny activities on the skin. In particular, the role of radon progeny attached to the skin relative to the incorporation of radon and its shortlived decay products requires further examinations. It is, therefore, the goal of the present study to explore the different routes of radon and radon progeny transport from radon-rich water and air to sensitive target cells and tissues for the speciﬁc exposure conditions, as they exist in the radon spas Bad Gastein and Bad Hofgastein, Austria.

2. Measurements and results 2.1. Radon in water and air The radon activities in the thermal water as well as in the ambient air were measured by transferring the radon to a 20 L ionisation chamber. While the Rn-activity in the thermal water was relatively constant at about 980 ± 93 Bq L−1 over 30 measurements, the radon concentration in the vapour bath displayed signiﬁcant temporal variations, ranging from 30 to 200 Bq L−1 . The latter variability is caused by the large temporal variations of radon production in the thermal water wells. The loss of radon during the thermal bath (20 min) by stirring the water is of the order of a few percent.

642

H. Tempfer et al. Table 1 Radon and radon progeny activities in thermal water

Activity (Bq L−1 )

222 Rn

218 Po

214 Pb

214 Bi/214 Po

980 ± 90

912 ± 85

1005 ± 56

950 ± 33

2.2. Radon progeny in water Radon progeny activities in water were measured by exposing 2 cm2 -sized silver plates to 10 L of thermal water. The activities of the decay products 218 Po, 214 Pb and 214 Bi/214 Po on the plates after 20 min of exposure were determined from the spectra and decay curves recorded by an 8-channel alpha spectrometer (Ortec-Octete® ) with 1200 mm2 detectors (Canberra CAM 1200 AM), relative to the activities measured in a 10 L of Ra-standard water, which has been stored in a closed container for more than 3 h to obtain equilibrium between radon and its progeny. The results compiled in Table 1 indicate that equilibrium between radon and its short-lived progeny is reached shortly after ﬁlling the bathtub. Since there is hardly any deposition of radon progeny on the walls of the bathtub, this equilibrium remains during the whole treatment period [4]. 2.3. Transfer of radon from water to exhaled air In order to examine the transfer of Rn from thermal water through the skin and then via the blood into the exhaled air, test persons were bathing in the thermal water for 20 min. Starting 6 min before the bathing phase, samples of exhaled air were taken every 3 min for a period of 50 min. Inhaled air was supplied by a tube from outside the bathing cabin to make sure that no radon was inhaled from the air above the water. Air samples, stored in aluminium-coated bags, were subsequently measured with Lucas-cells and a Pylon® AB 5. The temporal course of the radon concentration in the exhaled air is plotted in Fig. 1 for a selected test person, based on a

Fig. 1. Radon concentrations in the exhaled air of a selected test person based on a radon concentration 850 Bq L−1 in the thermal water.

Biophysical mechanisms and radiation doses in radon therapy

643

radon concentration 850 Bq L−1 in the thermal water. Similar results were obtained for other test persons [5,6]. 2.4. Radon progeny deposition on the skin Patients were exposed to thermal water (thermal bath) or hot radon vapour (vapour bath) for a period of 20 min. Immediately after the patients had left the bathtub or the exposure chamber, the wet skin was dabbed with a towel. Alpha detectors were attached to three different locations of the body (forearms, belly, and lower legs). Alpha spectra and decay curves were measured for 30 min to determine the radon progeny surface activities. To examine potential intra- and inter-individual variations of the radon progeny surface activities, 6 selected patients (3 males, 3 females) were measured three times within one week. The results of these measurements for both exposure conditions are listed in Tables 2 and 3. Signiﬁcant differences among the 6 test persons could be observed. Despite the small sample size, men seem to adsorb radon progeny better than women. Activities measured on artiﬁcial inorganic surfaces (copper and PVC), when exposed to thermal water, were about six times lower than those on human skin. In contrast, radon progeny activities measured on these surfaces, when exposed to vapour, were higher than those measured on the skin. All measurements on human skin as well as on other surfaces like copper, PVC or pig skin showed that 218 Po, 214 Pb and 214 Bi/214 Po are adsorbed from water in a ratio of about 1 : 4 : 3 Table 2 Radon progeny activities (Bq cm−2 ) on various surfaces exposed to thermal water Surface

218 Po

214 Pb

214 Po/214 Bi

Total

Women lower arm (n = 11) Women abdomen (n = 11) Women lower leg (n = 11) Men lower arm (n = 11) Men abdomen (n = 10) Men lower leg (n = 10) Copper (n = 4) PVC (n = 4)

0.20 ± 0.13 0.22 ± 0.10 0.20 ± 0.14 0.31 ± 0.13 0.34 ± 1.18 0.45 ± 0.23 0.18 ± 0.04 0.10 ± 0.02

0.96 ± 0.17 0.57 ± 0.07 0.62 ± 0.20 1.97 ± 0.82 1.12 ± 0.54 1.30 ± 0.44 0.25 ± 0.07 0.33 ± 0.08

0.80 ± 0.11 0.56 ± 0.17 0.42 ± 0.01 1.81 ± 1.02 1.07 ± 0.60 0.90 ± 0.33 0.23 ± 0.08 0.13 ± 0.02

1.96 ± 0.82 1.35 ± 0.19 1.24 ± 0.24 4.09 ± 1.33 3.53 ± 0.82 2.65 ± 1.00 0.66 ± 0.11 0.56 ± 0.08

Table 3 Radon progeny activities (Bq cm−2 ) on various surfaces exposed to radon vapour Surface

218 Po

214 Pb

214 Po/214 Bi

Total

Women lower arm (n = 8) Women abdomen (n = 7) Women lower leg (n = 4) Men lower arm (n = 6) Men abdomen (n = 10) Men lower leg (n = 7) Copper (n = 10) PVC (n = 10)

1.09 ± 0.62 0.77 ± 0.33 0.50 ± 1.19 1.29 ± 0.51 0.86 ± 0.38 0.57 ± 0.22 1.43 ± 0.40 1.63 ± 0.62

0.74 ± 0.37 0.50 ± 0.20 0.41 ± 0.17 0.96 ± 0.44 0.49 ± 0.15 0.40 ± 0.18 1.37 ± 0.57 1.49 ± 0.65

0.32 ± 0.21 0.18 ± 0.15 0.14 ± 0.09 0.33 ± 1.18 0.23 ± 0.13 0.18 ± 0.08 0.58 ± 0.28 0.53 ± 0.18

2.15 ± 0.75 1.45 ± 0.41 1.05 ± 0.27 2.58 ± 0.70 1.58 ± 0.43 1.15 ± 0.30 3.38 ± 1.25 3.65 ± 1.45

644

H. Tempfer et al.

after 20 min. The reason for this behaviour is the fact that daughters of adsorbed nuclides get desorbed from a wet surface by the decay of the mother nuclides. At steady state conditions, i.e., after about two hours, the ratios of the radon progeny activities adsorbed on a wet surface are determined by their half-lives [4]. In order to study the time dependence of decay product adsorption process, test persons were exposed to thermal water and radon vapour, respectively, for 10, 20 (the normal exposure time), 30, 40 and 60 min. Following each exposure, radon progeny decay curves were recorded over a period of 30 min. The time dependence of the absorption process of each nuclide on the patient’s skin is illustrated in Figs. 2 and 3. The time dependence curves illustrate that 218 Po saturates after less than 20 min because of its half-life of 3.05 min. Due to the longer half-lives of the other progeny, no steady state is reached after one hour of exposure.

Fig. 2. Time dependence of radon progeny adsorption on skin in thermal water.

Fig. 3. Time dependence of radon progeny adsorption on skin in radon vapour.

Biophysical mechanisms and radiation doses in radon therapy

645

Table 4 Radon progeny activities on various types of skin, relative to the activities on the untreated lower arm of the respective test person (n = 3): clean pig skin (A), greasy pig skin (B), human skin treated with Finalgon® (C), and human skin after 15 min of jogging (D)

218 Po 214 Pb 214 Bi/Po

A

B

C

D

0.21 ± 0.16 0.14 ± 0.10 0.32 ± 0.36

0.18 ± 0.15 0.13 ± 0.11 0.35 ± 0.34

0.98 ± 0.11 0.91 ± 0.22 1.20 ± 0.33

1.51 ± 1.40 1.02 ± 0.11 0.91 ± 0.63

Several personal factors were investigated which may affect the adsorption process in thermal water and thus explain the apparent inter-subject variations: greasiness of the skin, blood circulation, and transpiration: (i) The inﬂuence of skin greasiness on radon progeny adsorption was examined by comparing two pieces of pig skin; a normal skin and one with the fat removed prior to exposure. (ii) A test person’s arm was treated with a vasculodilating ointment (Finalgon® , BoehringerIngelheim) to enforce blood circulation of the skin. (iii) The inﬂuence of skin transpiration on adsorption was examined by exposing a test person after 15 min of intensive jogging; 5 h later the same person was exposed without prior physical activity. Table 4 shows the comparison of radon progeny activities deposited on the skin after 20 min in thermal water: clean pig skin (A), greasy pig skin (B), human skin treated with Finalgon® (C), and human skin (male test person) after 15 min of jogging (D), based on ﬁve measurements of each surface. The low radon progeny activity on the pig skin suggests that biological mechanisms may play an important role in the adsorption process, while greasiness, blood circulation and transpiration do not seem to be decisive factors. The pH-value of water is likely to inﬂuence adsorption of waterborne ions [7]. Thermal radon water was ﬁlled into 25 L containers and pH was varied between 3 and 9.5 by adding HCl or NaOH. For comparison, one arm of a test person was exposed to different pH-waters, while the second arm was exposed to pH-neutral thermal radon water. The dependence of radon progeny adsorption on the pH-value is plotted in Fig. 4. Radon progeny adsorption seems to be reduced in basic and acidic environments. Reduction in a basic environment can be explained by formation of hydroxides, while the decrease of adsorption in an acidic environment may be explained by a decrease in competition between proton and radon progeny for surface binding sites [8]. 2.5. Radon progeny distribution in upper skin layers To assess radiation doses to the deeper layers of the skin which are relevant to an immune response (Langerhans cells), it is necessary to obtain information about the vertical distribution of radon progeny in the upper layers of the skin. Such information is also necessary for the determination of the efﬁciency of the alpha spectrometer [9].

646

H. Tempfer et al.

Fig. 4. pH dependence of radon progeny adsorption in thermal water.

Fig. 5. Vertical distribution of radon progeny in the upper layers of the skin.

Single cell layers were removed by attaching strips of Tesa-ﬁlm® (area: 16 cm2 ) to the skin and pulling them off after a few seconds [10]. The radon progeny activity at the treated spot was measured and compared with a second detector attached to the untreated skin. These measurements were repeated with 2, 4, 6 and 8 pull-offs. After microscopic analysis of the pulled-off strips, a vertical activity distribution relative to the untreated skin could be plotted (Fig. 5). Since two pull-offs remove approximately one cell layer [10], and the thickness of an epidermic corneocyte swollen in water is about 5 μm [11], it was assumed that two pull-offs correspond to a depth of about 5 μm. The observed quasi-exponential distribution of the radon progeny activity with depth may be caused by diffusion and by transport through crevices or along hair capillary tubes.

Biophysical mechanisms and radiation doses in radon therapy

647

Table 5 Effective target organ doses (μSv) incurred in radon therapy Target organ

Thermal bath

Vapour bath

Skin (epidermis) Kidneys Lungs (TB)

800 40 130

1310 100 1870

3. Radiation doses Together with measurements of ambient radon and radon progeny concentrations in the therapeutic facilities as well as of radon concentrations in blood [5], the above measurements allow a complete radiological description of the various exposure pathways: (i) inhalation of radon and radon progeny in the inhaled air, (ii) incorporation of radon in the bathtub or in the exposure chamber via the skin, and (iii) attachment of decay products on the skin of the patients. Based on these data, effective radiation doses for different organs and tissues in the human body for both treatment schemes are listed in Table 5 [3,12,13]. Two philosophies have been advanced in recent years to explain the observed therapeutic effects: (a) a direct radiation effect in speciﬁc organs of the human body, e.g., regeneration of damaged tissue in the affected organ, through the incorporation of radon (via skin or the lungs); the kidneys are selected here to illustrate typical doses to internal organs; and, (b) an indirect radiation effect, such as induced immune response through irradiation of the deeper layers of the skin (epidermis) by radon progeny activities on the skin. Since lung cancer risk is commonly associated with radon exposure, the doses to the tracheobronchial (TB) region were included for comparison. In contrast to previous dose assessments, which practically neglected skin irradiation [14], the present study indicates that deposition of radon progeny on the skin produces a much higher dose than inhaled radon and progeny to typical internal organs, such as the kidneys, and thus may play an important role in the therapeutic effect. With respect to lung cancer induction, it is important to emphasise that bronchial doses lie well within the natural ﬂuctuations of the doses received by the general population, thus causing no additional radiation hazard to the patients undergoing such treatments. For comparison, the annual effective TB dose to the population living in Salzburg City amounts to 2800 μSv [14].

4. Discussion While our investigations do not attempt to explain the cellular mechanisms leading to an observable therapeutic effect, nor their dependence on dose, they provide the indispensable dosimetric basis for evaluating the relative signiﬁcance of the different exposure pathways. Indeed, our calculations of doses to different target organs suggest that radon progeny attachment on skin surfaces may play a major role in the therapeutic response for both treatment schemes. On the other hand, bronchial doses are comparable to the natural ﬂuctuations of the doses received by the general population, thus causing no additional radiation hazard to the patients undergoing radon therapy.

648

H. Tempfer et al.

References [1] ICRP Publication 65: Protection against radon-222 at home and at work, Ann. ICRP 23 (2) (1993). [2] P. Deetjen, A. Falkenbach (Eds.), Radon und Gesundheit (Radon and Health), Lang, Frankfurt am Main, 1999. [3] W. Hofmann, in: H.G. Pratzel, P. Detjeen (Eds.), Radon in der Kurortmedizin, ISMH, Geretsried, 1997, pp. 57– 67. [4] H. Surbeck, Appl. Radiat. Isot. 53 (2000) 97–100. [5] W. Hofmann, H. Lettner, R. Winkler, W. Foisner, in: Radon und Gesundheit (Radon and Health), Lang, Frankfurt, 1999, pp. 83–91. [6] W.A. Grunewald, H. v. Philipsborn, G. Just, in: Radon und Gesundheit (Radon and Health), Lang, Frankfurt, 1999, pp. 93–101. [7] H. v. Philipsborn, Health Phys. 72 (1997) 277–281. [8] H. Lettner, W. Hofmann, H. Tempfer, A. Schober, in: J.P. McLaughlin, S.E. Simopoulos, F. Steinhäusler (Eds.), The Natural Radiation Environment VII, Proc. VIIth International Symposium on the NRE, Rhodes, Greece, 20–24 May 2002, in: Radioactivity in the Environment, vol. 7, Elsevier, Amsterdam, 2004, this volume. [9] W. Ziechmann, Bodenchemie, Bibliographisches Institut & Brockhaus, Mannheim, 1990. [10] H. Pratzel, Grundlagen des perkutanen Stofftransportes in der Pharmako-Physiko-Therapie und Balneotherapie, Habilitation at Ludwig-Maximilian University Munich, 1985. [11] P.S. Talreja, N. Kleene, W.L. Pickens, T.S. Wang, G.B. Kasting, AAPS Pharm. Sci. 3 (2001), article 13. [12] S.V. Andrejew, B.N. Semjonow, D. Tauchert, Z. Phys. Med. Baln. Med. Klim. 19 (Special issue 2) (1990) 83–89. [13] E. Pohl, J. Pohl-Rueling, Health Phys. 32 (1977) 552–555. [14] E. Pohl, Z. Angew. Bäder-Klimaheilk. 26 (1979) 437–442.

649

Bioaerosols as carriers of radon progeny∗ S. Kagerer a , T. Rettenmoser a , W. Hofmann a , A. Falkensteiner b , F. Steger c a Institute of Physics and Biophysics, University of Salzburg, Hellbrunner Str. 34, A-5020 Salzburg, Austria b Ministry of Environmental Protection of the Government of Salzburg, Ulrich-Schreier-Str. 18,

A-5020 Salzburg, Austria c Division of Health Physics, Department of Radiation Protection, Austrian Research Centers Seibersdorf,

A-2444 Seibersdorf, Austria

Bioaerosols, such as bacteria, pollen and spores, constitute a major fraction of the ambient aerosols, thereby acting as potential carriers for radon decay products. Activity size distributions of the short-lived radon progeny attached to environmental aerosols were measured at four selected sites in urban and rural regions of the Province of Salzburg, Austria, varying in bioaerosol and radon progeny concentrations. Bacteria and spores were sampled with the aid of an Andersen impactor by using selective media, while pollen was collected with a custom-made ﬁlter system. Activities of size-fractionated radon progeny attached to environmental aerosols were determined by in situ gamma spectrometry. Additional measurements comprised mass and particle number size distributions and total number of ambient aerosols. Measured size distributions indicated that a considerable fraction of radon progeny were attached to larger particles, say above 1 μm. At particle sizes above about 5 μm, practically all particles were of biological origin. However, the relative fractions of bioaerosols varied signiﬁcantly with sampling site and local environmental conditions. Based on computed dose– exposure conversion factors, it was estimated that about 20% of the annual effective dose incurred in Badgastein may be attributed to the inhalation of large environmental aerosols, about one third being caused by biological aerosols.

1. Introduction Biological aerosols, such as bacteria, pollen and fungal spores, constitute a major fraction of the ambient aerosol. Unfortunately, there is little information available about background concentration levels of biological agents in outdoor or indoor environments [1]. Such bioaerosols can act as carriers of short-lived radon progeny, thereby affecting the size distribution of * This research was sponsored in part by the Ministry of Social Security and Generations, Project GZ 353.072/1IX/9/01.

RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07079-2

© 2005 Elsevier Ltd. All rights reserved.

650

S. Kagerer et al.

the attached radon progeny fraction. With respect to health effects following inhalation of radon progeny attached to bioaerosols, two aspects must be considered: First, some biological aerosols are in the size range of a few microns, thus increasing deposition in bronchial airways by enhanced inertial impaction, i.e., dose conversion factors are higher than those for submicron particles. Second, inhalation of bioaerosols may cause additional health problems by their very nature, such as allergic reactions to pollen. Thus the goal of this study is to establish a correlation between the occurrence of bioaerosols and short-lived radon decay products at different sites and at different seasons of the year and to investigate their dosimetric implications, in particular their dependence on size, which determines the fate of inhaled particles. Concentration and size measurements were carried out at four selected sites in urban and rural regions of the Province of Salzburg, Austria, varying in anthropogenic and bioorganic aerosols and radon progeny concentrations: Salzburg University: This urban measurement site can be described as a place with low trafﬁc density and a medium density of apartment houses in the neighborhood (altitude above sea level: 425 m). The adjacent botanical garden of the university is an important source of biological aerosols in ambient air. Salzburg City: In contrast to the university, the selected sampling site is directly situated next to main roads in the city (altitude above sea level: 425 m), where the trafﬁc density is relatively high (up to 4000 cars and trucks per hour). In addition, the density of apartment houses and ofﬁce buildings is also quite high. Tamsweg is located in an alpine region of the Hohe Tauern (Northern Alps), Austria (altitude above sea level: 1040 m). It is a small town with a low trafﬁc density and a relatively high density of apartment houses and single-family houses. Tamsweg is a region where agriculture and forestry are very important. Badgastein is another sampling site in a valley of the Hohe Tauern (altitude above sea level: 1000 m). It is a famous radon spa area with low trafﬁc density (about 100–120 cars per hour), surrounded by alpine forests.

2. Measurement techniques Different sampling techniques were used for collecting and analyzing the aerosols. The radon progeny activity size distributions were determined using a custom-made ﬁlter system, consisting of a sequence of six polyester ﬁlters (Osmonic, Poretics) with varying pore sizes (0.1– 10 μm). Air was pumped through this ﬁlter system with a ﬂow rate of 9.5 L m−3 for a period of four hours. This ﬁlter combination was also used to determine the aerosol mass in the different size ranges. The activities of the loaded ﬁlters were then measured by an in situ gamma detector system (Canberra GX3020, XtRa), with an energy range down to 3 keV. Measured activity size concentrations on each ﬁlter were subsequently corrected for pre-ﬁltration in the upstream ﬁlters [2] and normalized to the total activity measured by a high-volume sampler. The total number of particles and the number size distributions were determined by a laserbased portable dust monitor (Grimm Model 1173). Bacteria and fungal spores were collected with the aid of a microbiological air sampler (Andersen ACFM six-stage Viable Particle Sampler), covering the size range between 0.65

Bioaerosols as carriers of radon progeny

651

and 7 μm. For impaction of the bioaerosols, different agar media were used (LB – full medium and selective agar plates for bacteria and fungal spores). After sampling with the Andersen impactor for about 30 min, the agar plates were incubated for 48 hr at 35 ◦ C [2]. Pollen was collected with the above described custom-made ﬁlter system, consisting of 6 ﬁlters with varying pore sizes, and subsequently counted with a Scanning Electron Microscope.

3. Results The present results were obtained in a measurement campaign between May and August of 2001, i.e., they refer to the speciﬁc bioaerosol concentrations and weather conditions during that period of time. To be able to compare results measured at different times during the four month period, measurements were always carried out during similar weather conditions, i.e., when the weather was ﬁne and warm (approximately 20–25 ◦ C). All activity measurements by the custom-made ﬁlter system were corrected for the contribution of unattached radon progeny, which was quite substantial in some cases [3]. The results of the measurements from the different locations shown in Figs. 1–4 represent average values over several measurements. In general, the values of the activity size distribution obtained in this study are comparable to other published data [2,4–6]. In the rural areas of Badgastein and Tamsweg, radon progeny concentrations ranged from 1.8 to 10.6 Bq m−3 . For comparison, Shimo and Saito [4] reported typical outdoor concentrations ranging from 0.27 to 29 Bq m−3 , while the results of Willeke and Baron [2] were between 4 and 15 Bq m−3 . The much higher concentration of radon progeny in outdoor air

Fig. 1. Size distributions of radon progeny activity, bioaerosol and total aerosol concentrations at the Salzburg City measurement site.

652

S. Kagerer et al.

Fig. 2. Size distributions of radon progeny activity, bioaerosol and total aerosol concentrations at the Badgastein measurement site.

Fig. 3. Size distributions of radon progeny activity, bioaerosol and total aerosol concentrations at the Salzburg University measurement site.

Bioaerosols as carriers of radon progeny

653

Fig. 4. Size distributions of radon progeny activity, bioaerosol and total aerosol concentrations at the Tamsweg measurement site.

in Badgastein is primarily caused by the speciﬁc geological composition of the soil and the existence of hot springs with high radon content (Badgastein is a famous radon spa). Concentrations are similar in the fraction below 0.25 μm, but there are signiﬁcant differences in the fractions above 2.0 μm. While the activity size distribution in Badgastein falls off with increasing particle size, it remains approximately level in Tamsweg. The results of the measurements of total particle mass at the different sites are shown in Fig. 5. The results obtained in this study are comparable to the results of Clarke et al. [6]. The highest percentage of biological particles among ambient aerosols was found in Tamsweg (44% in the relevant size fractions), see Table 1. In general, sites with a high concentration of biological aerosols also have higher activities in the corresponding size fractions than sites with a low concentration of bioaerosols. Hence, the percentage of biological aerosols seems to correlate very well with the radon progeny activity size distribution. Results at the Table 1 Percentage of bioaerosols and inorganic aerosols in ambient air at the different measurement sites Measurement site

Bioaerosols

Inorganic aerosols

Bacteria/spores

Pollen

Salzburg City Badgastein Salzburg University Tamsweg

1.5 40 28 44

0.9 0.7 1.1 3.3

0.6 39 27 41

99 34 49 19

All measurement sites

39

1.2

38

61

654

S. Kagerer et al.

Fig. 5. Particle concentrations as a function of particle size.

urban measurement sites support such a correlation, as the lower activities, relative to the rural areas, coincide with a much smaller number of biological aerosols. There are also striking differences between urban and rural measurement sites. In contrast to the rural areas, activity concentrations in urban air are better correlated with the concentration of submicron particles (in the size range of 0.1 to 0.4 μm) than with the larger biological aerosols (ranging from 2 to 10 μm). For example, the activity in the small size fractions (< 0.25 μm) is approximately three times higher in places with a high concentration of small trafﬁc aerosols (various measurement sites in the city of Salzburg), as compared to those with low trafﬁc (University measurement site). Thus radon progeny activities in urban environments seem to be preferentially attached to submicron particles.

4. Dosimetric implications Particle size is the most important factor in determining the sites of inhaled particle deposition in the respiratory tract. In typical indoor environments, radon progeny are commonly attached to submicron particles and the fraction of larger particles, say in the micrometer-size range, can safely be neglected for the computation of dose–exposure conversion factors. Calculations of dose conversion factors in the size range from 1 to 10 μm indicate, however, that the high deposition efﬁciency of large particles in the bronchial region due to the impaction mechanism yields a dose–exposure conversion factor, which is roughly a factor of 2 higher than that for submicron particles [7]. The dose conversion factors displayed in Fig. 6 refer to the sitting awake breathing conditions for an adult male [8]. The concentrations of radon progeny activities attached to biological aerosols were derived from the measured activity and bioaerosol size distributions by calculating radon progeny

Bioaerosols as carriers of radon progeny

655

Fig. 6. Dose conversion factor for average doses to sensitive bronchial epithelial cells as a function of particle size.

Fig. 7. Dose calculations for the different measurement sites.

attachment coefﬁcients as a function of particle size [9]. Resulting bronchial doses at the four measurement sites are shown in Fig. 7. Because of the much higher outdoor radon progeny concentrations, total doses are highest in Badgastein. In addition, the contribution from bioaerosols is also highest there, representing about 30% of the total radon progeny exposure. Thus, in the presence of high bioaerosol concentrations in outdoor environments, the doses received during the time spent outdoors may also play a role for lung cancer risk assessment, despite the lower activity concentrations outdoors.

656

S. Kagerer et al.

5. Discussion Bioaerosols, such as bacteria, pollen and spores, constitute a major fraction of the ambient aerosol. Hence they can act as carriers of the short-lived radon progeny. With respect to health effects following inhalation of radon progeny attached to bioaerosols, two aspects must be considered: First, biological aerosols generally have sizes in the range of a few microns, thus increasing deposition in bronchial airways by enhanced inertial impaction, i.e., dose conversion factors are higher than those for submicron particles in the range of 0.1 to 0.5 μm. Second, inhalation of bioaerosols may cause additional health problems by their very nature, such as allergic reactions to inhaled pollen. In the case of urban aerosols, radon progeny activities were preferentially attached to submicron particle sizes in the size range of 0.1 to 0.4 μm. In rural areas, however, a considerable fraction of the radon progeny activity was found to be associated with large particles, ranging from 2 to 10 μm. At particle sizes above about 5 μm, practically all particles are of biological origin. However, the relative fraction of biological particles and hence their effect on the activity size distributions varied signiﬁcantly with sampling site (availability of bioaerosols) and local environmental conditions (temporal variations of bioaerosol production, local trafﬁc, etc.). In typical indoor environments, radon progeny are commonly attached to submicron particles and the fraction of larger particles can be neglected for the computation of bronchial doses. Based on computed dose–exposure conversion factors for outdoor atmospheres in the size range from 1 to 10 μm, it was estimated that about 20% of the annual effective dose incurred in Badgastein may be attributed to the inhalation of large environmental aerosols, about one third being caused by biological aerosols. Thus, in the presence of high bioaerosol concentrations in outdoor environments, the doses received during the time spent outdoors may also play a role in lung cancer risk assessment, despite the lower activity concentrations outdoors than indoors.

References [1] L. Morawska, M. Hargreaves, in: Conference of the Australian Institute of Occupational Hygienists, Wollong, Australia, 2001. [2] K. Willeke, P.A. Baron, Aerosol Measurement: Principles, Techniques and Applications, Van Nostrand, New York, 1993. [3] Y.S. Cheng, H.C. Yeh, J. Aerosol Sci. 11 (1980) 313–319. [4] M. Shimo, H. Saito, J. Environ. Radioact. 51 (2000) 49–57. [5] J. Porstendörfer, Ch. Zock, A. Reineking, J. Environ. Radioact. 51 (2000) 37–48. [6] A.G. Clarke, G.A. Azadi-Boogar, G.E. Andrews, Sci. Total Environ. 235 (1999) 15–24. [7] R. Weinkler-Heil, W. Hofmann, in: High Levels of Natural Radiation and Radon Areas: Radiation Dose and Health Effects, Elsevier, Amsterdam, 2002, pp. 169–177. [8] ICRP Publication 66: Human respiratory tract model for radiological protection, Ann. ICRP 24 (1–3) (1994). [9] J. Porstendörfer, in: Fifth International Symposium on the Natural Radiation Environment, CEC Report EUR 14411 EN, Commission of the European Communities, Luxembourg, 1993.

657

Radon in Finnish mines – regular monitoring since 1972 M. Annanmäki, E. Oksanen, E. Venelampi, M. Markkanen Radiation and Nuclear Safety Authority – STUK P.O. Box 14, FIN-00881, Helsinki, Finland

Radon measurements in Finnish underground mines were started in 1972. At that time there were 23 operating underground mines in Finland. Some of the mines were old, having a lot of underground abandoned spaces not in active use. Today there are 8 underground mines in operation, most of them being small in size. The ﬁrst radon measurements were made in order to ﬁnd out radon levels in the mines and to estimate the extent of a possible radon problem. Radon measurements were made in work places, as well as, in areas not under work but in which high radon concentrations were expected to occur. Since then regular measurements in all underground mines have been made in order to control radon concentration in work places and to ascertain that the limits set for radon are not exceeded, and in some cases to evaluate the doses to miners. In most cases radon daughter concentration has also been measured together with radon. During the last few years only radon concentrations have been measured. In 1975, a limit for radon was set at 1100 Bq m−3 (30 pCi L−1 ), the value being the radon in equilibrium with its short-lived daughter products. In 1992, an action level of 400 Bq m−3 for radon, an average over the annual working cycle, was adopted. In 1975 the mean effective dose for a miner was 3.5 mSv, in 1985 2.4 mSv, in 1995 1.7 mSv and in 2001 0.9 mSv. The decrease in radon concentration has been partly due to countermeasures made to decrease high radon concentrations but mostly because many of the large, poorly ventilated old mines have now been closed. The modern mines are small in size and it is technically easier to arrange appropriate ventilation in them. The equilibrium factor has varied from 0.2 to 0.9 depending on the mine and the year. The average equilibrium factor when all the mines are taken into account is 0.6. The alpha energy concentration due to thoron 220 Rn (212 Pb) was measured in 8 different mines from a total of 103 ﬁlter samples. The average alpha energy concentration due to 220 Rn was 0.1 μJ m−3 , and the highest measured value was 0.5 μJ m−3 . 1. Introduction The signiﬁcance of radon (222 Rn) as a source of radiation exposure was realised in the 1950s when it was discovered that exposures due to radon (speciﬁcally to its decay products 218 Po, RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07080-9

© 2005 Elsevier Ltd. All rights reserved.

658

M. Annanmäki et al.

Fig. 1. Number of workers in Finnish underground mines. 214 Pb, 214 Bi

and 214 Po) could explain the excess in lung cancers detected among miners. The ﬁrst results indicating a signiﬁcant excess of lung cancer among uranium miners were obtained in the 1960s in the United States. High concentrations of radon exist also in nonuranium mines, in various underground excavation works and other underground workplaces. Today, the epidemiological studies on uranium and non-uranium miners form the main source of information for estimating the risk (the dose) caused by exposures to radon [1]. Radon measurements in the Finnish underground mines were started in 1972. The ﬁrst radon measurements were made in order to ﬁnd out radon levels in the mines and to estimate the extent of a possible radon problem. Radon measurements were made in work places, as well as in areas not in active use but in which high radon concentrations were expected to occur. Since 1972 radon measurements have been made regularly in all underground mines every year or every second year, depending on the radon concentration measured in earlier years. The number of underground mineworkers in Finland is presented in Fig. 1. In the early 1970s there were 23 operating mines and the number of underground workers was about 1300. Today there are only 8 mines, most of them being small in size and employing about 400 workers in total. The minerals mined are mostly copper, zinc, gold and limestone. Earlier iron was also mined. During the years different equipment has been used to measure concentrations of radon and its short-lived daughter products. However, the measurement methods themselves remained the same. Radon was measured with grab sampling using Lucas-type chambers [2]. The shortlived daughter products were measured with the Kusnetz method [3]. In 1983 and 1984 the alpha energy concentration due to 220 Rn (212 Pb) was also measured in most of the mines from the same ﬁlter samples which were used to measure the alpha energy concentration due to 222 Rn.

2. Material and methods 2.1. Regulation of radon in mines The initiative to start radon measurements in 1972 came from the Ministry of Trade and Industry, which is in charge of the safety of mines. First regulations for controlling radon in

Radon in Finnish mines – regular monitoring since 1972

659

underground mines were issued in 1975. At that time, a limit for radon was set at 1100 Bq m−3 (30 pCi L−1 ), deﬁned as the Equivalent Equilibrium Radon concentration (EEC), i.e., radon in equilibrium with its short-lived daughter products. A major revision of the Finnish Radiation Act took effect in 1992 [4]. Since then the Act has covered all work activities, not only mines, involving signiﬁcant occupational and public exposures to natural radiation in accordance with the principles set out in the 1990 ICRP Recommendations. The Act and other provisions issued by virtue of it cover the activities falling under the scope of Title VII of the new BSS Directive [5,6]. In 1992, an action level of 400 Bq m−3 for radon was adopted. The action level is deﬁned as an annual average concentration during working hours. Steps must be taken to reduce the radon concentration if the action level is exceeded. If countermeasures do not work, individual doses of miners must be assessed. In this case, the maximum permissible radon concentration is 3000 Bq m−3 leading to an effective dose of 20 mSv, the dose limit for workers. The individual doses are assessed by measuring the radon concentrations in different working areas and by recording the working hours of the miners in those areas. 2.2. Radon and radon daughter monitors used During the last 30 years several types of equipment have been used to measure radon. However, the measurement method itself, measuring the activity of an air sample (grab sampling) using Lucas-type chambers, has remained the same [2]. Silver activated zinc sulﬁde is used to detect the alpha-activity of radon (222 Rn) and radon daughters (218 Po, 214 Po). In the beginning, there was only one measurement head and the air samples were taken into gas bottles and the air sample was then transferred into the measurement head for its measurement. Later on about 50 detector chambers were used into which air samples were taken directly. The construction of these detectors was our own. There were also several counting systems (a photo multiplier and an electronic counter) into which the chambers could be connected for the measurement. Since 1998, a commercially available monitor with several detector chambers has been used (Pylon AB 5). The detectors were regularly calibrated in a radon chamber maintained at STUK. The reference devices used for calibration purposes were regularly checked against a standard radon concentration produced into a steel drum by emanating radon from a known amount of radium (226 Ra). In 1990, a major study was carried out to check the validity of the calibration system and to recalibrate the radon monitors. As a consequence of the study, the calibration factors of the monitors were changed upwards by about 5% [7]. STUK has participated in several intercomparisons for radon measurements. The ﬁrst was organised with the Swedish National Institute of Radiation Protection (SSI) in 1978. The second, organised in 1982, was made together with the Swedish National Institute of Radiation Protection, the National Institute of Radiation Hygiene (SIS, Denmark) and the National Institute of Radiation Hygiene (NRPA, Norway). In 1987 the third CEC intercomparison of active and passive detectors for the measurement of radon and radon decay products was organised by the National Radiological Protection Board (NRPB, UK) in which STUK also participated [8]. In the 1990s, intercomparisons with the Swedish National Institute of Radiation Protection were also organised. All these intercomparisons showed that the differences between the instrument calibrations of the laboratories used as reference were small.

660

M. Annanmäki et al.

Several instruments have been used to measure radon daughter concentration, but the method used has been the same. A dust sample is collected on a glass ﬁbre or membrane ﬁlter and, after a sampling period of 5 to 10 minutes, the alpha activity collected on the ﬁlter is measured using a silver activated zinc-sulﬁde or a surface barrier detector. The results are expressed in units of μJ m−3 and are calculated using the so-called Kusnetz method [3]. Some of the instruments have been our own construction but since 1986 commercially available equipment has been used (Pylon WL-1000C Working Level Monitor). In the third CEC intercomparison exercise the results obtained with the radon daughter monitors were also compared and STUK’s results were in good agreement with the reference instrument [8]. 2.3. Dose assessment Measurements were made in all operating mines every year or every second year, depending on the concentrations detected in earlier years. These consecutive measurements were made during different seasons so as also to consider, in the long run, possible seasonal changes in ventilation conditions of the mine. The number of radon measurements per mine varied from about 10 to 30 depending on its size and on the number of different working areas. The results were compiled so that an average concentration, considered representative for the whole mine, was derived by calculating an arithmetic mean of concentrations detected in different working areas of the mine. This average concentration was used for comparison with the action level and for assessing the workers doses. The effective dose (E) was calculated using the equation [9]: E = 7.78 × 10−9

Sv ·F ·T ·C Bq h m−3

(1)

where C is the mean radon concentration (Bq m−3 ) in the working places of a mine, F is the equilibrium factor (value 0.64 was used), T is the annual working hours (value 1600 h was used). In all dose assessments presented in this paper, the values F = 0.64 (mean value for all mines during 1972–1992) and T = 1600 h were used. This gives a direct correspondence between the radon concentration and the effective dose: 100 Bq m−3 ↔ 0.8 mSv. In cases where no radon measurements were made in a particular year, the mean radon concentration was calculated as the mean of the previous and following years.

3. Results Regular radon measurements have been made in all underground mines for 30 years in order to ascertain that the action levels set for radon are not exceeded. In most cases also radon daughter concentration has been measured simultaneously. Since 1997 only radon concentrations have been measured. Where action levels were exceeded, the results were used to assess radiation doses received by the workers to ensure that the dose limits were not exceeded. In the 1970s the radon concentrations were less than 400 Bq m−3 in about 50% of all the measured places of work in the mines. Concentrations exceeded 2000 Bq m−3 in about

Radon in Finnish mines – regular monitoring since 1972

661

20% of the measurements. The highest detected concentration in a working area was about 37 000 Bq m−3 . Even higher concentrations were detected in some poorly ventilated unused areas of the mines. Today (results of 2000–2001) the mean radon concentration in all the mines is about 110 Bq m−3 (weighted by the number of workers in each mine) and individual results exceeding 400 Bq m−3 are rather rare. The average individual dose to underground miners was calculated in the following way. First, a mine-speciﬁc individual dose was calculated using equation (1). Then the average of the mine-speciﬁc doses was calculated using the number of workers in each mine as a weighting factor. The results are presented in Fig. 2. The collective dose of Finnish miners was derived by summing up the product of the minespeciﬁc individual dose and the number of workers for all mines operating during the year in question. The results are presented in Fig. 3. Also a mine-speciﬁc mean value for the equilibrium factor F was calculated. Measurements were made annually in 2–15 different mines. Usually about 10 measurements were made in each mine and the average of these results was taken as the mine speciﬁc value. In the calculations only measurements made at working areas were taken into account. The distribution of mine-speciﬁc equilibrium factors is presented in Fig. 4. There were no measurements in 1979.

Fig. 2. Average individual dose received in Finnish underground mines.

Fig. 3. Collective annual dose received in Finnish underground mines.

662

M. Annanmäki et al.

Fig. 4. Equilibrium factor F in Finnish underground mines. The “Minimum” and “Maximum” values are the lowest and the highest mine-speciﬁc values. The “Average” is the average of all the mine-speciﬁc values determined during the year in question.

The alpha energy concentration due to thoron 220 Rn (212 Pb) was measured in 8 different mines from the total of 103 ﬁlter samples. The average alpha energy concentration due to 220 Rn was 0.1 μJ m−3 , and the highest measured value was 0.5 μJ m−3 .

4. Discussion Radon concentrations in mines and doses caused by radon have decreased signiﬁcantly since the early 1970s. The decrease in radon concentration has been partly due to countermeasures made to decrease high radon concentrations but mostly because many of the large, poorly ventilated old mines have now been closed. The modern mines are small in size and it is technically easier to arrange appropriate ventilation in them. Good general ventilation is also required by the use of diesel-powered machinery. In addition, remedial actions have been taken wherever regular control measurements have shown concentrations exceeding the action level. However, sometimes the reduction of radon concentration is technically difﬁcult, especially if the radon originates from radon-rich bedrock-water entering the cavities in large amounts. The most common remedial measures used are increasing the ventilation and preventing the water entering the mine. The very high individual and collective doses calculated for the years 1972–1973 were partially caused by the former lead mine in Kosrnäs whose ore contained signiﬁcantly elevated levels of uranium (238 U). The radon concentrations were in the order of 10 000– 20 000 Bq m−3 . The mine was closed soon after the regular radon monitoring started. There was also one large mine in which the ventilation was not operating correctly resulting in concentrations as high as 10 000–50 000 Bq m−3 . Action was taken to improve the situation resulting in signiﬁcant reduction in radon concentrations. The results of the very early years might also be slightly biased by the fact that during that time many measurements were made in poorly ventilated spaces of old large mines and later on it was difﬁcult to identify whether they really were regular working areas or not. There was a slight increase in both the individual and collective doses in the late 1980s. The statistics show that the increase was caused by two large old mines which were then closed

Radon in Finnish mines – regular monitoring since 1972

663

in 1988 and 1989, respectively. During the last few years of operation these mines exploited some residual ores of the older parts of the mine. It is assumed that ventilation in these parts of the mines was not as good as in the newer parts. In the late 1990s there was once again a slight increase in the average individual dose. This was caused by a mine operating only for some ﬁve years. The concentrations in this particular mine were in the order of 1000–2000 Bq m−3 and temporally concentrations up to about 7000 Bq m−3 were detected in some working areas. The reduction of radon concentration was technically almost impossible, because the radon originated from radon-rich bedrockwater entering the cavities in large amounts. The equilibrium factor F varied from 0.2 to 0.9, depending on the mine and the year. The mean value for the equilibrium factor for all the mines has mostly been between 0.6 and 0.7. The average of all the annual mean values for the years 1972–1992 is 0.64. The equilibrium factors in the mines are somewhat higher than the value 0.4 commonly used for dose assessments [1]. This should be taken into account whenever assessing the miners’ doses. The alpha energy concentrations due to thoron 220 Rn (212 Pb) were so low that it was concluded that thoron does not cause signiﬁcant exposures and regular monitoring was not deemed necessary. It can be concluded that radon in the Finnish mines is not at the moment a major problem. The average concentration in all the mines is about 110 Bq m−3 (results for the years 2000–2001) and very seldom do individual results of different working areas exceed the action level of 400 Bq m−2 . The average in mines is almost the same as the average in Finnish dwellings! Monitoring of normal above ground workplaces has showed, e.g., that in certain areas of Finland the arithmetic mean radon concentration was as high as 500 Bq m−3 [10]. In some of the most radon prone areas in Finland, up to 10–30% of aboveground workplaces and 30–40% of single-family houses may exceed the action level of 400 Bq m−3 [9]. Considering also the number of workers involved, there is no doubt that radon in ordinary above ground workplaces, i.e., not in the underground mines, is the most important source of occupational radon exposure in Finland. In fact, radon in above ground workplaces is by far the most important source of all occupational exposures in Finland [9]. However, the sudden increases in radon concentration detected in some mines in the late 1980s and late 1990s demonstrate clearly that regular monitoring is needed in order to detect circumstances and tendencies which might lead to signiﬁcantly high exposures to the miners as was the case in the early 1970s when the regular monitoring was started.

Acknowledgements During the last 30 years several persons have participated in the radon measurements in mines. Special thanks are due to Mr. Heimo Kahlos who ﬁrst ordered investigation of the possible occurrence of radon in Finnish mines and assessment of whether or not countermeasures were needed. He was the person who was in charge of these measurements in the beginning. Thanks also to Ms. Ritva Paatelainen, who for several years assisted in measurements and recorded all the measurement data into a data ﬁle for later use. We are grateful to Ms. Kyllikki Aakko, who in 1990 made a special study on the calibration of the measurement method used for radon measurements.

664

M. Annanmäki et al.

References [1] ICRP Publication 65: Protection against radon-222 at home and at work, Ann. ICRP 23 (2) (1993). [2] H.F. Lucas, Improved low-level alpha-scintillation counter for radon, Rev. Sci. Instrum. 28 (1957) 680–683. [3] H.L. Kusnetz, Radon daughters in mine atmospheres – a ﬁeld method for determining concentrations, Am. Ind. Hyg. Assoc. Quart. 17 (1956) 85–88. [4] Radiation Act, The Statutes of Finland 592/91, Helsinki, 1991. [5] Council Directive 96/29/Euratom of 13 May 1996 laying down basic safety standards for the protection of the health of workers and the general public against the dangers arising from ionising radiation, Ofﬁcial J. Eur. Commun. Ser. L 159 (29.6.1996). [6] Recommendations for the implementation of Title VII of the European Basic Safety Standards Directive (BSS) concerning signiﬁcant increase in exposure to natural radiation sources, Radiation Protection 88, European Commission, Luxembourg, 1997. [7] K. Aakko, Calibration of ZnS 50 radon monitor of STUK, MSc thesis, University of Jyväskylä, 1990 (in Finnish). [8] J.C.H. Miles, J. Sinnaeve, Results of the third CEC intercomparison of active and passive detectors for the measurement of radon and radon decay products, Report EUR 11882 EN, Commission of the European Communities, Luxembourg, 1988. [9] M. Markkanen, M. Annanmäki, E. Oksanen, Radon in workplaces, Kerntechnik 65 (1) (2000). [10] M. Annanmäki, E. Oksanen, M. Markkanen, Radon at workplaces other than mines and underground excavations, Environ. Int. 22 (Suppl. 1) (1996) S769–S772.

665

Technical guidance on safe installation of a GAC ﬁlter used for removing radon from water M. Markkanen Radiation and Nuclear Safety Authority – STUK, PO Box 14, FIN-00881 Helsinki, Finland

Some ground waters contain radiologically important concentrations of radon (222 Rn). During the past few years, the use of granular activated carbon (GAC) ﬁlters for removing radon from water has rapidly increased in Finland. However, in many cases the risks related to external gamma radiation originating from the ﬁlter during its use are not appropriately considered. In some cases, the ﬁlter has caused signiﬁcant dose rates in the living areas of a home. It was decided to prepare technical guidance on installing a GAC ﬁlter in a safe manner. Radiological criteria for gamma dose were established and used for deriving design parameters regarding appropriate shielding and safety distances. Because of the related risks, however, today it is recommended that GAC ﬁlters should not be installed in homes. The only exception are installations that do not require any speciﬁc considerations concerning external gamma radiation, e.g. cases where the ﬁlter is placed underground into a well.

1. Introduction Some ground waters contain radiologically important concentrations of radon (222 Rn). High concentrations occur especially in drilled wells in granite and granodiorite areas. There are two commonly used techniques to remove radon from water: aeration and granular activated carbon (GAC) ﬁltration. The GAC ﬁltration has many advances, e.g. it is inexpensive, rather maintenance-free and can easily be installed in water systems of existing buildings. The ﬁlter is rather small, typically a cylinder with diameter about 25 cm, height 80–120 cm and volume 40–60 L [1]. However, a signiﬁcant disadvantage of a GAC ﬁlter is the accumulation of radon resulting in a build up of gamma emitting radon daughter products in the ﬁlter. During the past few years, the use of GAC ﬁlters for radon removal has rapidly increased in Finland. In many cases the installers are not aware of the risks related to the external gamma radiation emitted from the ﬁlter during its normal use. In some cases, a ﬁlter has been installed in a room very close to a bedroom or the living room resulting in signiﬁcant gamma dose rate in the living areas of the house. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07081-0

© 2005 Elsevier Ltd. All rights reserved.

666

M. Markkanen

Technical guidance was prepared on installing a GAC ﬁlter in a safe manner. Radiological criteria for the external gamma dose were established and used for deriving design parameters regarding appropriate shielding and safety distances. The purpose of this paper is to discuss the grounds and contents of technical guidance on safe installation of a GAC ﬁlter in private homes. Calculated safety distances and material thickness for two types of shield materials are presented. They are derived from the radiological criteria established for this purpose. 2. Methodology 2.1. Radiological criteria The purpose of installing a GAC ﬁlter is to reduce the dose arising from the use of radonbearing water. However, at the same time the ﬁlter itself becomes a source of external gamma radiation. The remedial action can be considered justiﬁed if the dose averted by removing radon is higher than the dose caused by the increase in gamma radiation. Preferably the averted dose should be as high as possible. One constraint for optimisation is that the gamma dose should not cause an intervention in the future. Therefore, it was considered that the gamma dose should be a small fraction of the dose limit for the members of the public, namely about 0.1 mSv a−1 . The gamma dose rate at the surface of a ﬁlter varies in the range 50–500 μSv h−1 , and 0.5–5 μSv h−1 at 1 meter distance, depending on the radon concentration and average water consumption. If the ﬁlter is installed in an area accessible to the inhabitant, there is always a potential for high unintended exposures if the inhabitant stays for some reason very near to the ﬁlter for considerable times. Therefore, the dose rate in the vicinity of a ﬁlter should be reduced to some moderate level. On these bases, appropriate radiological criteria for installing a GAC ﬁlter in a home are: – the excess external gamma dose shall not exceed 0.1 mSv a−1 , and – the excess external gamma dose rate shall not exceed 1 μSv h−1 at 1 meter from any accessible surface of the ﬁlter. Certain precautions are needed when a ﬁlter is taken out of use or is changed. Regulations on transport of radioactive substances may be applicable if the ﬁlter is transferred immediately after being disconnected from the pipeline. Speciﬁc transport arrangements can be avoided by disconnecting the ﬁlter from the pipeline about 2–3 weeks before the transfer. By this time most of the radon vanishes because of its radioactive decay (half-life 3.8 days). 2.2. Methodology for deriving shield characteristics and safety distances Assuming a removal efﬁciency of 100%, the activity (A) of radon in the ﬁlter is CQ CQ Ψ A= 1 − e−λt −→ = t0 λ λ λ

(1)

where C is radon concentration in water (Bq L−1 ), Q is average water consumption (L d−1 ), t is time from taking the ﬁlter into use (d), λ is radon decay constant, 0.181 d−1 , Ψ is average radon-ﬂow (Bq d−1 ).

Technical guidance on safe installation of a GAC ﬁlter used for removing radon from water

667

When the ﬁlter is taken into use, the amount of radon increases rapidly but reaches a state of equilibrium within about three weeks. The gamma emitting radon daughters have a halflife much shorter than radon and are thus in equilibrium with radon all the time. Therefore, the gamma dose rate is directly proportional to the average radon-ﬂow. On the other hand, the average radon-ﬂow is easily determined from basic measurement results (C and Q). This way it constitutes a useful quantity in relation to which shield characteristics and safety distances can be quantiﬁed. For example, a typical water consumption in a single family house in Finland (3–4 inhabitants) is in the order of 500 L d−1 . If the radon concentration in water is 2000 Bq L−1 , then the radon-ﬂow is 500 L d−1 × 2000 Bq L−1 = 1 MBq d−1 . A method used for calculating external gamma dose rate originating from a solid object is described in [2]. It is based on calculating the attenuation and build-up of photon radiation penetrating layers of materials. Some simplifying assumptions and modelling had to be made in order to use the method for modelling a GAC ﬁlter and its shielding. Therefore, the results are perhaps not scientiﬁcally the best possible but were assumed accurate enough for these practical assessments. The calculation model was tested against some measurements made in the vicinity of an operating GAC ﬁlter. Different types of materials were used as temporary shield between the ﬁlter and a gamma survey meter. The results are presented in Table 1. Considering the possible inaccuracies related to both the calculations and measurements made, the results were relatively encouraging and gave conﬁdence in using the method for assessing shield characteristics and safety distances required to comply with the radiological criteria adopted. In the calculations, two types of living areas of a house were distinguished. Areas such as living rooms and bedrooms were classiﬁed as “high occupancy areas” and others like the bathroom, laundry and garage were “low occupancy areas”. The occupancy time used for these areas was 6000 and 500 h a−1 , respectively. It was assumed that the ﬁlter is installed in a low occupancy area and the average distance of the exposed person to the ﬁlter is 2 m during the stay of 500 h a−1 in this area. In this way, only the most important parameter, i.e. distance to a high occupancy area, needed to be varied in the ﬁnal calculations. Table 1 Validation of the method used for calculating gamma dose rate caused by a GAC ﬁlter Material

2 mm lead 5 mm lead 5 mm aluminium 5 cm concrete 15 cm water 5 cm concrete and 5 mm lead 5 cm concrete and 5 mm aluminium 5 cm concrete and 15 cm water

Excess dose rate (μSv h−1 ) Measured (min–max)∗

Calculated

1.32 (1.11–1.44) 1.10 (0.91–1.13) 1.60 (1.46–1.70) 1.04 (0.96–1.14) 0.73 (0.64–0.78) 0.70 (0.66–0.81) 0.97 (0.94–0.98) 0.44 (0.41–0.47)

1.33 0.96 1.53 1.01 0.85 0.55 0.89 0.41

∗ Measured with a hand-held gamma survey meter. The survey meter was 1 m away from the ﬁlter and the tested material layer near the surface of the ﬁlter. The result is the average of ﬁve readings from which the local background of 0.22 μSv h−1 has been subtracted (the minimum and maximum readings are given in brackets).

668

M. Markkanen

3. Results Based on a large set of calculations it was concluded that the radiological criteria could be complied with in the following way. A shield attenuating gamma radiation shall be placed around the ﬁlter. The shield shall cover all accessible surfaces of the ﬁlter and surfaces facing towards any low or high occupancy areas in the house. The thickness of the shield depends on the average radon-ﬂow and the distance to the nearest high occupancy area in the house. The calculated material thicknesses for two different materials, i.e. iron and concrete, are presented in Figs. 1 and 2. The existing structures of the house (e.g. concrete walls) can be accounted for in sizing the shield. For example, if the calculated thickness is 20 cm of concrete and there is a 10 cm thick concrete wall already in place, only 10 cm of additional concrete shield is needed towards the directions covered by the wall. It was considered that a GAC ﬁlter would probably not be an optimised solution for radon removal for radon-ﬂows exceeding 4 MBq d−1 because the shields become very massive and there is an increasing potential for very high exposures if there are any deﬁciencies in sizing or constructing appropriate shields. Also exposures received during maintenance would become high. Therefore, no results are presented for radon-ﬂows exceeding this value. Lead would be a very effective material for constructing the shield but no values are presented here because

Fig. 1. Calculated material thickness for a shield made of iron.

Fig. 2. Calculated material thickness for a shield made of concrete.

Technical guidance on safe installation of a GAC ﬁlter used for removing radon from water

669

its use in bulk amounts in homes should not be encouraged considering its possible toxic impact on the living environment as well as environmental aspects regarding its disposal in the future.

4. Discussion The European Commission Recommendation on Radon in Drinking Water [3] introduces an action level of 1000 Bq L−1 for consideration of remedial action. According to estimates presented in the Recommendation, the total annual dose caused by water with 1000 Bq L−1 of radon is in the order of 2–4 mSv a−1 . This includes both ingestion via drinking water and inhalation of radon released from the water to indoor air. In this study, an unshielded ﬁlter used for water with 1000 Bq L−1 at an average water consumption of 500 L d−1 will cause an external gamma dose rate of about 0.5, 0.13 and 0.06 μSv h−1 at distances of 1, 2 and 3 m, respectively. This implies that if the ﬁlter is installed in the vicinity (1–3 m) of areas with high occupancy, the annual dose to the inhabitant due to the gamma radiation would be in the order of 0.3–1 mSv a−1 . It should be noted that CAG ﬁlters are also used for removing various other impurities from water. In such a case, the presence of radon should always be examined. Because the amount of accumulated radon is directly proportional not only to the radon concentration but also to the average water ﬂow, special attention should also be paid in cases where the water consumption is exceptionally high, even if the radon concentration is relatively low. A GAC ﬁlter may be operational for years. The continuous radioactive decay of radon and its short-lived daughter products in the ﬁlter will also result in the accumulation of lead (210 Pb) and polonium (210 Po). For example, the activity of these nuclides in a ﬁlter with a radon-ﬂow of 1 MBq d−1 will be about 0.5 MBq after three years of use. The activities of these nuclides need to be considered e.g. when disposing of used ﬁlters. The assessments performed have demonstrated that a GAC ﬁlter could, in principle, be installed in a home safely if appropriate shields against gamma radiation are built and other radiological considerations are taken into account. However, there is a signiﬁcant potential that all conditions necessary to ensure safety would perhaps not be fulﬁlled in practice. Because of these risks, it is recommended that GAC ﬁlters should not be installed in homes. Alternative methods such as aeration should be preferred. Perhaps the only exception could be for installations that do not require any speciﬁc considerations concerning external gamma radiation, e.g. cases where the ﬁlter is placed underground into a well.

References [1] T. Turtiainen, P. Kokkonen, L. Salonen, Removal of radon and other radionuclides from household water with domestic style granular activated carbon ﬁlters, Report STUK-A172, Radiation and Nuclear Safety Authority – STUK, Helsinki, 2002. [2] M. Markkanen, Radiation dose assessments for materials with elevated natural radioactivity, Report STUK-BSTO 35, Radiation and Nuclear Safety Authority – STUK, Helsinki, 1995. [3] Commission Recommendation of 20 December 2001 on the protection of the public against exposure to radon in drinking water supplies, Ofﬁcial J. Eur. Commun. Ser. L 344 (2001).

670

Outdoor radon and thoron in the USA, Canada, Finland and Thailand∗ N.H. Harley a , P. Chittaporn a , M. Heikkinen a , R. Merrill b , R. Medora b a New York University School of Medicine, 550 First Avenue, New York, NY 10016, USA b Fluor Fernald Radiation Control Section, 7400 Wiley Road, Hamilton, OH, USA

A miniature radon (222 Rn) and thoron (220 Rn) passive alpha track detector that can be worn for personal exposure assessment or used as an area monitor was developed through the US Department of Energy, Environmental Management Science Program (EMSP). The detector measures both radon and thoron gas, in duplicate to provide precision. Indoor and outdoor measurements in several countries show the presence of thoron is universal and can bias measurements that permit radon detection without correction for thoron. The grand average of a few measurements made in the USA, Canada, Thailand, Finland is 12 Bq m−3 for both nuclides.

1. Detector development The detector was developed to provide accurate personal exposure assessment at the Fernald Feed Material Production Center, a former USDOE uranium processing facility, where significant onsite stores of radium and thorium were located. There is presently no other detector to measure, in duplicate, radon and thoron gas simultaneously as a personal or area monitor. To assess thoron bronchial dose, the usual measurement is the decay products of thoron 212 Pb and 212 Po [1–4]. Thoron gas itself is rarely measured, because of the difﬁculty in measuring an alpha particle emitting gas with a very short half-life (t1/2 = 55 s). The measurement of the two gases normally requires real time instrumentation with various types of decay chambers to permit a difference in signal with and without the 220 Rn [5,6]. The NYU passive alpha track detector measures the integrated signal from the alpha decay of thoron gas, over any time period, depending upon the exposure assessment needs. The 222 Rn and 220 Rn personal monitor was designed to use alpha track detection ﬁlm in 4 separate chambers using different entry diffusion barriers for signal differentiation between * Support by USDOE (EMSP) Contract DE FG07 97ER62522 is gratefully acknowledged.

RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07082-2

© 2005 Elsevier Ltd. All rights reserved.

Outdoor radon and thoron in the USA, Canada, Finland and Thailand

671

the two gases [7]. A conducting foam, directly beneath the gas entrance holes to the monitor permits only radon or thoron gas passage. The monitor shape is four lobed (similar to a four leaf clover, nickname 4leaf) with a face dimension of 5 cm across chambers and a thickness of 1 cm. It is molded from conducting. ABS (CNi) plastic, i.e., plastic with embedded nickel coated carbon ﬁbers. The nuclear track ﬁlm used for detection of the alpha particles emitted within each chamber is a laser cut 9 × 9 mm square solid-state nuclear track ﬁlm (CR-39). The pristine ﬁlm background is 5 to 15 tracks depending upon the lot. The entire ﬁlm area is counted. The unit displays no charge artifacts and thus the calibration is constant in all situations. The efﬁciency for 222 Rn is (0.009 Tracks per Bq m−3 day) (0.32 Tracks per pCi L−1 day) and the lower limit of detection is 220 Bq m−3 day (6 pCi L−1 day) using a video imaging system. The efﬁciency for thoron is 0.013 Track per Bq m−3 day, (0.48 Track per pCi L−1 day) and the lower limit of detection is 1100 Bq m−3 day (30 pCi L−1 day). The difference in efﬁciency for the 2 gases is due mainly to the change in volume of the detection chamber with the diffusion barrier in place. The 4 separate chambers are versatile and can be used either for 222 Rn measurements alone or a combination of 222 Rn and 220 Rn detection. Each chamber has an O ring seal to prevent leakage around the diffusion barrier, when a barrier is in place to inhibit thoron entry to the radon chamber. Various diffusion barriers were tested for their efﬁciency in permitting radon entry while inhibiting thoron entry. The most appropriate barrier ﬁlm was 3 mil aluminized Mylar, and this is now used in all the 222 Rn measurements. After exposure in our radon and thoron test chambers, or in the ﬁeld, the CR-39 alpha track ﬁlm is etched in 6 N KOH overnight to reveal the alpha particle tracks as shallow pits. The tracks are scored by image analysis with a computerized digitizer. The image analysis scores all tracks on the 9 × 9 mm ﬁlm directly, using an Olympus Zoom microscope, and a Data Translation frame grabber and image analysis program (Global Lab Image/2 ®). Track counting is normally performed using image analysis with about 10 to 20% of samples also scored visually for quality control (QC) after enlarging the 9 × 9 mm area with a microﬁche reader and printing to a standard hard copy paper image (about 23X). Pristine nuclear track ﬁlm and exposed positive controls are etched with each batch of research or ﬁeld samples for QC. This personal or area monitor is based on a miniaturization of our proven passive area 222 Rn alpha track detector [8,9].

2. Results The detectors have been in place on top of the radium containing silos at the USDOE former uranium processing facility in Fernald, Ohio since September 1999. The radium silos contain about 150 TBq 226 Ra (4000 Ci). The measurements from September 1999 to 2002 are shown in Fig. 1. The average 222 Rn concentration was 145 Bq m−3 . Twenty-ﬁve 4leaf detectors have been placed outdoors surrounding the radium containing silos at Fernald since May 2000. The measurements from May 2000 to 2002 for 222 Rn and 220 Rn are shown in Figs. 2 and 3. The 222 Rn concentration west and east of the silos averaged 58 Bq m−3 over the period May 2000 to February 2002. The 220 Rn concentration was not statistically different from zero during

672

N.H. Harley et al.

Fig. 1. Radon and thoron measured on top of radium silos.

Fig. 2. Radon measured west and east of radium silos.

this period. An offsite station at Fernald monitors background and these data are shown in Fig. 4. The 222 Rn at the offsite station averages 26 Bq m−3 , and the 220 Rn was not statistically different from zero.

Outdoor radon and thoron in the USA, Canada, Finland and Thailand

673

Fig. 3. Thoron measured west and east of silos.

Fig. 4. Radon and thoron measured offsite form Fernald.

Measurements for quality control are made at the National Weather Service site in Central Park, Manhattan. The detectors are located at 10 cm and 87 cm above bare earth in two locations within the fenced compound. The 222 Rn and 220 Rn data are shown in Figs. 5 and 6.

674

N.H. Harley et al.

Fig. 5. Radon measured at National Weather Service site, New York City.

Fig. 6. Thoron measured at National Weather Service site, New York City.

In these measurements 220 Rn almost always exceeds the 222 Rn. The average 222 Rn and 220 Rn concentrations were about 15 Bq m−3 over the interval from 1999 to 2002. Measurements are made for quality control, both indoors and outdoors, at a suburban New Jersey home. The outdoor measurements are shown in Fig. 7. Thoron was evident in all sam-

Outdoor radon and thoron in the USA, Canada, Finland and Thailand

675

Fig. 7. Outdoor radon and thoron measured at a suburban New Jersey home.

ples and the 222 Rn and 220 Rn average concentrations were 16 and 17 Bq m−3 , respectively. At this location thoron mostly exceeded the radon concentration. We also measured the thoron decay product 212 Pb in weekly ﬁltered outdoor air samples at this same suburban New Jersey home from 1995 to 1997 or about 2 years (10). The average concentration of 212 Pb was 0.08 ± 0.05 Bq m−3 . The measurements are shown in Fig. 8. The outdoor equilibrium ratio between 220 Rn and the decay product 212 Pb is therefore about 0.08/17 or 0.4%. We had measured 212 Pb in the basement of this home and found the 212 Pb to 220 Rn equilibrium ratio t be 1/50 (or 2% equilibrium). Measurements of 222 Rn and 220 Rn have been made with this detector in a few other locations for special purposes. The outdoor measurements include data from Canada, Finland and Thailand. These are shown in Table 1, along with a summary of the data in Figs. 1 to 8. The measurements in Thailand also showed thoron concentrations in excess of radon. The average outdoor concentration for these 4 countries (excluding the Fernald silos values) is 12 Bq m−3 for both 222 Rn and 220 Rn. It is of signiﬁcance that the detectors measured both radon and thoron simultaneously, at normal background concentrations. Because thoron was in some cases a signiﬁcant fraction, or exceeded the total gas signal, it suggests that historic measurements of radon made either outdoors or in homes with detectors capable of permitting thoron gas entry could have provided an effective 222 Rn concentration that was somewhat higher than actual due to the confounding inﬂuence of 220 Rn. Because the dose from thoron decay products is about a factor

676

N.H. Harley et al.

Fig. 8. Outdoor weekly 212 Pb measured at a suburban New Jersey home [10]. Table 1 Outdoor radon and thoron measurements in four countries Country (location) USA, Central Park Northern New Jersey Fernald offsite Fernald on silos Fernald near silos Canada Near Ft. McMurray Mississauga Thailand (Bangkok) Finland (Rovaniemi)

Date Jul 99–Jun 01 Oct 99–Jun 01 Jun 00–Oct 01 Sep 99–Mar 02 Aug 00–Jan 02 Jul 00–Oct 00

Mar 00–Dec 00 Jan 01–Jan 02 Jan 01–Jan 02 Sep 00–Feb 01 Sep 00–Feb 01 Grand average (omit silos values)

Height (cm) above ground 10 87

222 Rn

S.D.∗

(Bq m−3 ) 17 11 13 26 146 58 7 2 19 10 7 8 10 11 12

220 Rn†

S.D.†

(Bq m−3 ) 7 7 7 9 63 28 6 7 5 2 2 2 2 2 6

15 16 18 14 43 −2 −1.5 10.4 9 8 19 14 N.D.‡ N.D. 12

7 6 7 15 49 28 9 9 6 3 3 3

6

∗ S.D. One-standard deviation of duplicate CR-39 ﬁlms in each detector. † Thoron determined by signal difference. Negative values indicates total (radon plus thoron) minus radon less than

zero. ‡ N.D. Not detectable. Trip background too large to estimate thoron concentration.

Outdoor radon and thoron in the USA, Canada, Finland and Thailand

677

of ﬁve less than that from radon per unit exposure, the health effects from survey data may have been overestimated.

References [1] S. Schery, Health Phys. 49 (1985) 1061. [2] S. Schery, J. Air Waste Manage. 40 (1990) 493. [3] S. Schery, M.J. Zarcony, in: Proceedings of the 18th Midyear Topical Symposium, Health Physics Society, 1985. [4] N.H. Harley, B.S. Pasternack, Health Phys. 24 (1973) 379. [5] H. Israel, J.A.S. Adams, in: W.M. Lowder (Ed.), The Natural Radiation Environment, University of Chicago Press, 1964, p. 313. [6] NCRP, Exposure the population of the United States and Canada to natural background radiation, NCRP Report 94, National Council on Radiation Protection and Measurement, Bethesda, MD, 1987. [7] P. Chittaporn, N.H. Harley, Health Phys. 76 (1999) S163. [8] B. Litt, J. Waldman, N.H. Harley, P. Chittaporn, Health Phys. 61 (1991) 727. [9] N.H. Harley, P. Chittaporn, J. Sylvester, M. Roman, Health Phys. 61 (1991) 737. [10] N.H. Harley, P. Chittaporn, Technology 7 (2000) 407.

678

Early thoron concentration levels in Guarapari: Dosimetric implications A.S. Paschoa a , J. Pohl-Rülling b a Departamento de Física, Pontifícia Universidade Católica do Rio de Janeiro (PUC-Rio), Rio de Janeiro, RJ, Brazil b University of Salzburg, Salzburg, Austria

The three main coastal cities in Brazil with thorium-bearing monazite sands have external gamma radiation levels, in μR h−1 , with the following ranges [a]: 60–130, in Guarapari; 130– 290, in Meaipe; and 50–330, in Cumuruxatiba. One of the main objectives of this paper is to compare the internal doses received by the inhabitants of the thorium rich areas, like Guarapari, with those doses received by populations living in radon rich areas.

1. Introduction There are just a few human populations living permanently exposed to natural radioactivity at levels at least one order of magnitude higher than those averaged for the rest of the world. Guarapari, in Brazil, and some coastal areas in Kerala, India, are examples often mentioned of such populations. In both cases there are high levels of natural 220 Rn (thoron) gas due to 232 Th (thorium) bearing monazite sands. Many years ago, a decision was taken to study the human populations of these two high naturally radioactive (HINAR) areas, taking into consideration a statement made by the Study Group on the Effects of Radiation on Human Heredity of the World Health Organization (WHO) [1]. A study was undertaken in Guarapari which included the following measurements: external levels of environmental radiation on the overall area; external exposures of the general population, using thermoluminescent dosimeters; thoron in breath; chromosomal aberrations; and whole body counting selected workers of the now extinct monazite separation plant “Monazito e Ilmenito do Brasil (MIBRA),” which then processed about 20% of the total amount of monazite sands used annually to produce concentrates. A review of the studies of the natural radioactive environment in Brazil, including those made in Guarapari, was published elsewhere [2]. This paper, however, concerns only previously unpublished data on early thoron concentrations in the air of Guarapari, and some implications related to those levels. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07083-4

© 2005 Elsevier Ltd. All rights reserved.

Early thoron concentration levels in Guarapari: Dosimetric implications

679

2. Thoron in the air and in breath 2.1. Experimental A modiﬁed version of the double ﬁlter technique developed earlier [3–5] was used in Guarapari to measure radon and thoron in air, and in the breath of selected subjects. Figure 1a is a schematic illustration of the cylindrical double ﬁlter of 1.0 meter length, and 15 cm diameter. Two ﬁlters 7 cm in diameter each were placed at the extremities of the cylinder, and the air let through at a rate of about 9 L min−1 . The ﬁlter at the inﬂux extremity removed solid decay products and most airborne particulates, while the other, at the outﬂow end, collected the decay products of radon and thoron gases which entered the cylinder. The outﬂow ﬁlter was then measured for counts resulting from the decay of 220 Rn. The actual counting of the ﬁlter had to take place in a thoron free environment, or a careful background reading had to be done. Figure 1b schematically represents the apparatus used to calibrate the double ﬁlter chamber for measuring thoron (or radon) in the air. In the case of calibrating the instrument to measure thoron, the 228 Th standard solution needs to be at least one month old to achieve equilibrium

(a)

(b) Fig. 1. (a) Modiﬁed version of the double ﬁlter chamber to measure 222 Rn and 220 Rn in air; (b) Apparatus for calibrating the double ﬁlter chamber to measure 222 Rn and 220 Rn in air.

680

A.S. Paschoa, J. Pohl-Rülling

Fig. 2. Double ﬁlter calibration curve for thoron.

with 224 Ra. The ﬂow rate used was 8.7 L min−1 (i.e., 0.145 L s−1 ) under normal atmospheric pressure. The following other factors were taken into proper account: the activity of the 228 Th standard solution; the ﬂow rate of air entering the calibrating apparatus; the thoron losses due to the dead volume (≈ 1.6%)–the 220 Rn concentration in the air entering the inﬂow ﬁlter was 6.7 kBq m−3 (i.e., 181 pCi L−1 ). Considering a “ﬁctive activity”, Ac (at t = 0) – in disintegration per minute (dpm) – the whole efﬁciency of the apparatus (≈ 16%) is given by the expression ε = Ac (at t = 0)/I0 ,

(1)

where I0 is a quantity corresponding to a thoron alpha activity concentration equal to 3.7 kBq m−3 (i.e., 100 pCi L−1 ). Values for I0 were determined for 30, 60 and 120 minute sampling times. Figure 2 is a graph representing the linearity of I0 as a function of sampling time. Calibration factors for different sampling times were estimated based on the graph of Fig. 2. The counting efﬁciency of the ﬁlter was about 30%, so it is reasonable to state that about 50% of the chamber-born decay products of 220 Rn became attached to the walls. The minimum detection activity for a sampling time of 2 hours, and a counting time of 10 hours, allowed one to measure a thoron concentration of about 0.02 kBq m−3 (i.e., 0.5 pCi L−1 ). Thoron activity per unit time in breath, rather than the thoron concentration (in kBq m−3 ), is of interest to infer the 224 Ra body burden of a subject. However, one must take into account that only a few percent of the endogenous thoron are exhaled. The working hypothesis, to infer the 224 Ra body burden from thoron in breath measurements, was that 8% of the thoron produced within the body was exhaled. 2.2. Thoron in the air of MIBRA plant In the mid sixties, when the measurements of thoron concentration in the air of selected sampling places at MIBRA were made, Guarapari was still a small ﬁshing village with less than 10 000 inhabitants. However, during the summer months the overall village population would become more than double that size. The reason for the tourist pilgrimage was the popular faith in the “miraculous” properties of the black sands, to say nothing about the marveleous beaches and ﬁshing locations. Today Guarapari is a modern tourist resort, and ﬁshing is dominated by large motorboats and yachts. The black sands appeared as conspicuous spots of black ilmenite, surrounded by somewhat red to orange sands, which were mostly monazite grains. The mineralogical association of monazite, zirconite, ilmenite, and rutile varied in proportions from deposit to deposit. However, localized spots with monazite concentrations up to 65% have been found. The typical

Early thoron concentration levels in Guarapari: Dosimetric implications

681

monazite sand compositions were: 55 to 65% rare earths; 25 to 35% monazite; 4 to 6% thorium oxide. In 1967 the MIBRA plant had the schematic conﬁguration shown in Fig. 3. Trucks loaded with sands use the site near the entrance shown in Fig. 3 to discharge their load. After the ﬁrst physical separation, the sands were put into sacks for internal transport and subsequent grinding and magnetic separation with the help of vibrating machines. There were a number of further separation steps until the monazite and other concentrates were stored in sacks for transportation. On June 4, 1967 thoron concentration measurements were made at the sites indicated in Fig. 3 by the numbers 1, 2 and 3. Measurements at sites 1 and 2 were made while the plant was operating in the daytime, but at site 3 the measurements were made at night without any plant activity. The results of those measurements are shown in Table 1. Although the thoron concentrations measured at sites 1 and 2 were quite high, at site 3 the value was found to be below the detection limit of the instrument. When those measurements were made, there were about 100 employees working at MIBRA plant in the daytime, and there was not a night shift. The measured air concentrations

Fig. 3. MIBRA plant for the industrial separation of monazite in Guarapari as it was in 1967. Table 1 Thoron concentration in the air of the operating MIBRA plant at selected sampling sites on June 4, 1967 in daytime Sampling site number

Thoron concentration (kBq m−3 )

1 2 3

11 26 < 0.02

682

A.S. Paschoa, J. Pohl-Rülling

of thoron were high during the day at the MIBRA plant, to the extent that the thoron concentration at site 1 was equal to the maximum permissible concentration (MPC) for thoron in air for working places at that time – i.e., 11 kBq m−3 (290 pCi L−1 ) – while at site 2 the measured concentration corresponded to higher than twice that limit. Those levels of thoron concentrations corresponded to about 100 times the indoors thoron concentrations, and more than 300 times higher than the levels found outdoors in Guarapari. The current Brazilian occupational Derived Air Concentration (DAC) may be obtained from guidelines written in the document CNEN-NE-3.01 [6], which follows the recommendations of the ICRP Report 32 [7]. In so being, the equilibrium equivalent DAC for 220 Rn, assuming a mean breathing rate of 1.2 m3 h−1 during a working period of 2.000 h y−1 is 0.330 kBq m−3 (8.9 pCi L−1 ), which is only 3% of the earlier MPC. Thus, for today’s DAC for thoron the MIBRA concentration of thoron in air would reach levels almost 80 times higher than that occupational standard. 2.3. Outdoors and indoors thoron Table 2 presents, for comparison purposes, typical thoron concentrations in air measured outdoors and indoors in Guarapari, and in Rio de Janeiro. The range of outdoors thoron concentration found in Guarapari was comparable to the thorium bearing monazite sands present non-uniformly throughout the village and surroundings, while in Rio de Janeiro the values were below the detection limit of the instrument. Here one must bear in mind that the indoors thoron concentration distribution is quite non-uniform, because of its short half-life (55 seconds). The internal trilateral corners on the ﬂoors of a room might have a local thoron concentration more than one order of magnitude higher than the concentration found in the middle of the room. Whenever one sleeps near such inner trilateral corners near the ﬂoors, the inhalation of thoron may be considerably higher than while standing in a room, or sleeping on a regular bed. Such was the case for many Guarapari inhabitants at that time who slept on low beds near bare ﬂoors. The relatively high indoors thoron concentration, which appears in Table 2, reﬂected the high content of 232 Th and progeny in the building materials used in the construction of the laboratory where the measurement was made. The range of entry rate of 220 Rn in houses due to 232 Th and progeny up to 224 Ra in building materials was reported to be from 0.05 to 5 kBq m−3 h−1 [8]. Here it is worth mentioning that the worldwide mean of the 232 Th natural concentration in soils is 0.03 kBq m−3 , with levels that reach up to 0.4 kBq m−3 [9]. The thorium bearing monazite sands of Guarapari, India, and China had 232 Th concentrations in the order of this latter high value.

Table 2 Thoron concentrations (kBq m−3 ) outdoors and indoors in air (average), and in Guarapari, and in Rio de Janeiro Site

Outdoors

Indoors

Guarapari Rio de Janeiro

0.07–0.2 < 0.02

0.24 0.16

Early thoron concentration levels in Guarapari: Dosimetric implications

683

Table 3 Indoors and outdoors 220 Rn/222 Rn ratios reported by UNSCEAR for selected countries, and in Kerala and Guarapari

Indoors Outdoors

UNSCEAR [9]

Kerala [11]

Guarapari

0.01–0.5 0.01–0.08

23

3.3 × 102 8.3–3.3 × 102

Thus, the entry rate of 220 Rn in houses built over those sands, like the shack houses which were easy to ﬁnd some decades ago in Guarapari, and on the coastal areas of Kerala, had a tendency to be quite high. The MPC of thoron in air used to be 0.37 kBq m−3 (10 pCi L−1 ). Thus, all values presented in Table 2 are below that old limit. Today, however, the correct interpretation of derived limits may become quite complex. Thus, for example, the exempt activity concentration for 220 Rn is 1.2 × 104 kBq m−3 , assuming ρair = 1.2 × 10−3 g cm−3 , while for 222 Rn the same kind of concentration is 12 kBq m−3 [10]. How a factor of 1000 between these two exemption concentrations was determined one might consider to be beyond the scope of this work. However, when one assumes that an inhabitant of an area with thorium rich soil lives in a house built with building materials which are also thorium rich, it is a matter of curiosity to know whether such a situation can, or cannot, be considered “of no regulatory concern.” Here it is interesting to mention that the 220 Rn/222 Rn ratio values reported recently to the European Community Countries (EEC) are quite different from that found in Guarapari, as can be seen in Table 3. One striking aspect that can be noticed in Table 3 is the fact that the 220 Rn/222 Rn ratios are reversed for both indoors and outdoors in selected countries [9], as compared to those ratios found in Kerala [11], and Guarapari. The low concentration of 220 Rn in the air is usually explained through its short half-life, because the activities of 220 Rn and 222 Rn in rocks and soils are considered to be comparable [12]. Thus, only in soils with much higher 232 Th than 238 U contents can the 220 Rn/222 Rn ratio, even without knowledge of the equilibrium-equivalentconcentration (EEC), be reversed, as in the cases of Kerala and Guarapari. 2.4. Thoron in breath measurements In the case of thoron in breath measurements the subject was asked to breath into a rubber sack, which served as a reservoir, while the pump at the other end of the chamber drew the exhaled air through the two ﬁlters of the chamber. Thoron in breath measurements were made initially with three inhabitants of Guarapari who had previously worked in MIBRA for 8 to 17 years. 224 Ra body burdens were inferred taking into account the working hypothesis mentioned earlier that 8% of the thoron produced within the body was supposed to be exhaled. Here it is worth mentioning that 224 Ra is the immediate parent of 220 Rn. The 8% exhalation hypothesis was based on investigations concerning radium (226 Ra and 228 Ra) retention and mobilization in the human skeleton [13–15]. Table 4 shows the inferred 224 Ra body burdens in the three workers measured. Later on, thoron in breath measurements were made with nine other subjects selected at random from Guarapari inhabitants. Figure 4 is a graph representing the results of 224 Ra body

684

A.S. Paschoa, J. Pohl-Rülling

Fig. 4. Graph of the 224 Ra body burdens inferred from thoron in breath measurements in selected Guarapari inhabitants arranged as a function of subject age.

Table 4 224 Ra body burdens inferred from thoron in breath measurements in MIBRA workers Subject

Inferred 224 Ra body burden (Bq)

A B C

10.4 5.9 7.4

burdens inferred from the measurements made with these nine subjects arranged by age. One can notice from Fig. 4 that there was not any correlation between the age of the subject and the inferred 224 Ra body burden. Such lack of correlation seems to show that there was not any contamination of the subjects by long-lived 232 Th series radionuclides. In addition, when one compares the ranges of the 224 Ra body burdens inferred, as presented in Table 4 – from 5.9 to 10.4 Bq – with the equivalent range shown in Fig. 4 – from 2.8 to 9.7 – the conclusion is that there is not any signiﬁcant difference between the two sets of data. This means that the 224 Ra body burdens were likely to have been acquired by inhalation of particulates with 232 Th series progeny attached and ingestion of food contaminated with the same progeny.

Early thoron concentration levels in Guarapari: Dosimetric implications

685

Table 5 Likely annual dose equivalent due to inhalation of 220 Rn in Guarapari Thoron concentration (Bq m−3 )

Annual dose (mSv)

Indoors < 240 Outdoors < 200

55 8

3. Dosimetric implications Considering the bronchial dose equivalent factor used for exposure of 220 Rn and 220 Rn progeny by inhalation [16] – (20–80) × 10−5 mSv per Bq h m−3 – the proposed mean annual dose coefﬁcients for indoor and outdoor exposure of adults to 220 Rn, as recommended by ICRP Publication 50, can be expressed as [17]: HE /c = 0.23 mSv per Bq m−3

for indoors at home,

HE /c = 0.04 mSv per Bq m−3

for outdoors.

and

(2) (3)

For the sake of simplicity, because of lack of information on relative radon daughter concentrations, the factor c is assumed to be equal to unity, so correcting factors can be applied as further information becomes available. Taking into proper account data on indoors and outdoors 220 Rn concentrations in Guarapari, the likely annual dose equivalents are presented in Table 5. As for comparison purposes, one can mention that the annual exposure to external levels in Guarapari, as measured with lithium ﬂuoride dosimeters, were less than 14 mSv [2,18]. Moreover, the average worldwide annual effective doses recently reported by UNSCEAR due to inhalation of 220 Rn and 222 Rn are, respectively, 0.1, and 1.15 mSv. In so being, the doses from inhalation of 220 Rn in thorium rich areas of high natural radioactivity (TRAHINAR) deserve the further attention of investigators of the natural radiation environment. In Guarapari, for example, the annual doses due to thoron inhalation can be more than 500 times higher than the worldwide annual average.

4. Final remarks and conclusions 1. Thoron concentration in the air of a now extinct monazite separation plant in Guarapari, Brazil, reached levels as high as 26 kBq m−3 , while the plant was still in operation, which would represent 80 times the current occupational standard. 2. It is worth mentioning for further reﬂection that the situation of an inhabitant of a HINAR area, with thorium rich soil, living in a house with an indoors concentration in air of 240 Bq 220 Rn m−3 , may be considered “of no regulatory concern.” 3. The 220 Rn/222 Rn ratios in outdoors air in thorium rich HINAR areas, like Kerala and Guarapari, can be higher than 200 and 4000 times, respectively, the equivalent ratios reported recently by UNSCEAR. 4. Body burdens of 224 Ra inferred from thoron in breath measurements in Guarapari inhabitants can reach levels of the order of 10 Bq.

686

A.S. Paschoa, J. Pohl-Rülling

5. Last, but not least, the annual doses due to thoron inhalation in Guarapari may approach a value 500 times higher than the worldwide annual average dose reported by UNSCEAR, thus deserving further investigation.

References [1] World Health Organization, The Study Group on the Effects of Radiation on Human Heredity, WHO, 1957. [2] A.S. Paschoa, More than twenty years of studies of natural radioactivity in Brasil, Technology 7 (2000) 193– 212. [3] D. Blanc, J. Phys. Radiat. 21 (1960) 176–180. [4] W. Jacobi, in: Assessment of Radioactivity in Man, vol. 1, Vienna, 1964, pp. 275–289. [5] J. Fontan, D. Blanc, A. Bouville, Health Phys. 11 (1965) 15–20. [6] Diretrizes Básicas de Radioproteção, CNEN-NE-3.01, Comissão Nacional de Energia Nuclear, 1988, 36 p. [7] ICRP Publication 32: Limits for inhalation of radon daughters by workers, Ann. ICRP 6 (14) (1981). [8] K.H. Folkerts, G. Keller, H. Muth, Radiat. Prot. Dosim. 7 (1984) 41–44. [9] UNSCEAR, Exposures from natural radiation sources, Annex B in:, Sources and Effects of Ionizing Radiation, UNSCEAR 2000 Report to the General Assembly, with Scientiﬁc Annexes, vol. I: Sources, United Nations, New York, 2000. [10] IAEA, International Basic Safety Standards for Protection against Radiation and for the Safety of Radiation Sources, in: IAEA Safety Series, vol. 115, IAEA, Vienna, 1996. [11] A.C. Paul, P.M.B. Pillai, T. Velayudhan, K.C. Pillai, Internal exposure at high background areas, in: K.G. Vohra, U.C. Mishra, K.C. Pillai, S. Sadasivan (Eds.), Natural Radiation Environment, 1982, pp. 50–57. [12] A.H. Nevissi, Detection and measurement of radon and radon decay products, in: S.K. Majumdar, R.F. Schmaltz, Miller E. Willard (Eds.), Environment Radon: Occurrence, Control and Health Hazards, The Pennsylvania Academy of Science, 1990, pp. 110–122. [13] J. Rotblat, G.B. Ward, Phys. Med. Biol. 1 (1956) 57. [14] R.E. Rowlad, J.H. Marshall, Radiat. Res. 11 (1959) 299. [15] R.E. Grillmaier, Thesis, SAAR, Hamburg, 1964, pp. 69–72. [16] NEA, Dosimetric aspects of exposure to radon and thoron daughter products, Report NEA/OECD, Nuclear Energy Agency, Paris, 1983. [17] ICRP Publication 50: Lung cancer risk from indoor exposures to radon daughters, Ann. ICRP 17 (16) (1987). [18] T.L. Cullen, Use of thermoluminescent dosimeters for measurement of external radiation in Guarapari, Brazil, Health Phys. 12 (1966) 970–971.

687

Indoor occupancy and radon exposure in Finland I. Mäkeläinen, S. Moisio, H. Reisbacka, T. Turtiainen STUK, Radiation and Nuclear Safety Authority, PO Box 14, FIN-00881 Helsinki, Finland

The proportion of time spent at home, at work, outdoors and in other locations was analysed by a random sample questionnaire study. In addition, two-month alpha-track radon measurements were performed in the dwellings and at the workplaces of volunteers from the same group. Part of the group carried personal radon dosimeters. The mean proportion of the time Finns spend at home proved to be 0.73, and in private dwelling environments in total, 0.77. The proportion of time spent at work, at school, and in public buildings was 0.14, and that spent outdoors and in vehicles 0.09. The average radon concentration in dwellings was 104 Bq m−3 , in workplaces 30 Bq m−3 , and that recorded by personal dosimeters was 85 Bq m−3 . The concentration calculated from individual occupancy factors and their respective radon concentrations was 88 Bq m−3 . This study showed that the occupancy weighted radon exposure was 15–18% lower than the radon exposure at home.

1. Introduction Radon is the second largest cause of lung cancer after smoking in most countries. Epidemiological studies in residential environments seem to conﬁrm the connection between radon and lung cancer that was found in studies on miners [1]. However, the risk ratios of residential studies have broad conﬁdence intervals and they often lack statistical signiﬁcance [2]. One reason for this is random error in the exposure assessment. This kind of error may even attenuate estimate of risk ratio [3]. To assess radon exposure, the radon concentration in the environment where the participants in the study live and the proportion of time they spend there (occupancy factor) have to be determined. Individuals spend varying times in various human living environments, which have different radon concentrations. Generally, dwellings have the highest average radon concentration. The radon concentration in workplaces and in public buildings is, on average, lower than in dwellings, and it is at its lowest outdoors and in vehicles. In addition, the radon concentration in buildings varies according to the time of day and season, typically being higher at night and in winter [4]. The aim of this study is to give a general overview of exposure to inhaled radon and its variations in Finland. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07084-6

© 2005 Elsevier Ltd. All rights reserved.

688

I. Mäkeläinen et al.

2. Material and methods A questionnaire was sent to 4000 Finns selected randomly from the Central Population Register Centre of Finland. The questionnaire included a general section concerning their time use during the previous year, and a diary section for recording their time use during two speciﬁc days (a weekday and a weekend day or a holiday). The questions concerned the amount of time spent at work, at school and outdoors, and possible time spent at a summer residence (recreational dwelling). To obtain the seasonal differences in occupancy, the number of hours per week spent outdoors was requested separately for winter, summer, spring, and autumn. Details of the location and type of the ofﬁce or workplace were asked for, along with other related questions. The respondents were also asked whether they were willing to carry out an indoor radon measurement at home or at their workplace and to carry a personal dosimeter. The proportion of time spent at home, at work, and outdoors was calculated from the questionnaire by combining the general and diary sections. We crosschecked the occupancy factors outdoors and in vehicles by calculating them from the entries in both sections. Alpha-track dosimeters of STUK (Makrofol DE 1-1 KK) were sent to all those participants who were willing to take at least two types of measurement. The method used for radon measurement has been described earlier [5]. The measurements covered a two-month period, starting from 15th February to 9th March 2001. Dosimeters, to be installed in the living room and bedroom at home, and in the actual place of work at the workplace, were mailed to the respondents. They were asked to carry the personal dosimeters in their pockets or handbags. The assessment of radon concentration during the daily working hours would be biased if only day-and-night-integrating alpha-track dosimeters were used. Therefore, the radon concentration in the workplace during working hours was obtained using the concentration measured by the alpha-track dosimeter multiplied by a correction factor of 0.5 [6]. To calculate the occupancy weighted mean radon concentration an individual is exposed to, the proportions of time spent were used to weight the respective radon concentrations. To generate missing values for workplaces, random values were drawn from the log-normal distribution with geometric mean and standard deviation calculated from the radon concentrations at workplaces in the same province. The values for concentrations in public buildings were drawn from the same distribution as for workplaces. Indoor concentrations in recreational dwellings were calculated using the residential radon concentrations in the same province and, in order to adjust for summer, the main holiday season, multiplied by 0.5. For outdoor radon concentrations we used a value of 10 Bq m−3 [7].

3. Results By the deadline, we had received 833 questionnaires with at least the general section completed, indicating a unit response rate of 21%. As for the diary section, 85% and 88% of those who ﬁlled in the general section responded to the weekday and weekend day sections, respectively. We examined background variables such as age, gender, and the geographical site of the residence and compared them with those of the population of the original sample. The geographical area did not affect the response rate. We also compared the response rates in

Indoor occupancy and radon exposure in Finland

689

Table 1 The mean, coefﬁcient of variation, median and quartiles of the proportion of time spent (occupancy factors) at home, at work, and in other locations calculated from the data obtained by the questionnaire (N = 833) Location

Mean

C.V. (%)

1st quartile

Median

3rd quartile

Indoors at home Indoors at someone else’s home In a recreational dwelling In public buildings At work or at school Outdoors or in vehicles Elsewhere

0.73 0.03 0.02 0.03 0.11 0.09 0.01

26 212 436 141 98 75 429

0.63 0.00 0.00 0.00 0.00 0.05 0.00

0.73 0.00 0.00 0.01 0.11 0.08 0.00

0.89 0.04 0.00 0.04 0.19 0.12 0.00

Fig. 1. Proportion of time spent indoors at home by different age groups according to the questionnaire study (medians). The bars indicate 1st and 3rd quartiles.

municipalities with a high and low proportion of apartment houses and with high- and lowradon areas, but did not ﬁnd any difference. However, the response rate was found to depend on the age of the respondents. Senior citizens responded more willingly than younger adults and people with children. Stratiﬁcation by age group and gender was used to adjust the bias of response rate, though the effect was found to be slight. The mean proportion of time spent at home was 0.73 (Table 1). The mean for persons aged 10 and older was 0.73 and for adults from 24 to 64 was 0.65. The least amount of time spent at home was located in the 24–44 age group (Fig. 1). Womens’ home occupancy factor was 0.74, which was slightly larger than that of men, which was 0.71 on average. This difference was found to be due to men spending more time outdoors and in vehicles. Combined with time spent in someone else’s home or recreational dwelling, the total residential occupancy factor was 0.77. The amount of time spent outdoors and in vehicles seems to be independent of age (Fig. 2). For senior citizens, this accounted for the largest proportion of the time not spent at home. As regards other groups, the workplace or school were, apart from home, the

690

I. Mäkeläinen et al.

Fig. 2. Proportions of time spent at work or at school (medians), at someone else’s home, or outdoors by age group.

main locations at which time was spent. The difference between the genders in the amount of time spent at school or at work was negligible. The number of participants willing to perform all three types of measurement was 211. The number of those willing to perform the measurements at work and at home was 207 and those willing to perform the measurements at home and to carry a personal radon dosimeter numbered 178. The total number of alpha-track dosimeters mailed was 2029, of which 1734 were retrieved, from which it was possible to determine 1680 eligible radon concentrations. The number of participants who returned at least one eligible dwelling dosimeter and either a personal dosimeter or a workplace dosimeter or both was 519. The number of participants from whom we received occupancy data from the questionnaire and the completed radon measurements from their dwelling and from their workplace and/or personal dosimeter was 305. The number of participants with complete data was 123. The indoor radon measurement performed during the spring months is a good estimate of the yearly mean [4]. The concentration in peoples’ homes was estimated to be the mean of the measurements taken, as requested, in the living room and bedroom. In some cases residents did not follow the request concerning the rooms; nevertheless, the mean of the two dosimeters was used, provided the rooms were located in the living area. The concentration in living rooms did not differ signiﬁcantly from that in bedrooms. The slightly higher value for all dwellings than for living rooms and bedrooms is explained by the small number of other room types, some of them located in the basement (Table 2). The radon concentration was about three times higher in dwellings than in workplaces (Table 2). The highest radon concentration in a dwelling was 2256 Bq m−3 , and the same subject also had the highest personal dosimeter reading and calculated occupancy weighted mean concentration, 1436 and 1569 Bq m−3 , respectively. The highest value measured in a workplace was assessed as 474 Bq m−3 . It was found in an ofﬁce, located at ground ﬂoor level in Southern Finland.

Indoor occupancy and radon exposure in Finland

691

Table 2 Basic statistics of the radon concentrations of the study sample measured by two-month alpha-track detectors

Number Arithmetic mean (Bq m−3 ) Standard deviation (Bq m−3 ) Geometric mean (Bq m−3 ) Geometric standard deviation 1st quartile (Bq m−3 ) Median (Bq m−3 ) 3rd quartile (Bq m−3 )

Living room

Bed-room

Dwelling mean

Work-place

Personal dosimeter

Occupancy weighted radon concentration

447 99 151 64 2.4 35 62 112

492 103 165 65 2.4 37 62 109

520 104 159 68 2.4 38 65 110

333 30 40 20 2.6 10 19 35

339 85 122 58 2.2 34 55 93

309 88 118 62 2.2 36 57 95

Fig. 3. Radon concentration measured by personal dosimeter vs. occupancy weighted mean radon concentration calculated using the participant’s occupancy factors from the questionnaire study and the radon concentrations measured at her/his dwelling and workplace (123 observations).

The agreement between the occupancy weighted mean concentration obtained using the calculated concentration, 88 Bq m−3 , was good compared with that obtained using a personal dosimeter, 85 Bq m−3 (Table 2). This can also be seen in Fig. 1. The correlation coefﬁcient between these two variables was 0.97 and the difference was insigniﬁcant on both linear and logarithmic scales (Fig. 3). The geometric standard deviation was, for workplaces, 2.6, a little higher than the ﬁgure of 2.4 for dwellings. For individual exposures assessed using the two methods the geometric standard deviation was 2.2, lower than that of dwellings and workplaces, as might be expected. The correlation coefﬁcients between both calculated or measured concentrations and the indoor radon concentration at home were high, being 0.99 and 0.96, respectively (Fig. 4a). The correlation coefﬁcients between the same concentrations and indoor radon concentration in the workplace were weak but both signiﬁcantly different from zero, at 0.22 (p = 0.017) and 0.21 (p = 0.022), respectively (Fig. 4b). The correlation between the indoor radon concentration at home and in the workplace was also weak, at 0.18 (p = 0.044). The ratios of

692

I. Mäkeläinen et al.

Fig. 4. Occupancy weighted mean radon concentrations calculated using questionnaire data and alpha-track measurements at home and in workplace (solid circles), and by using personal alpha-track dosimeters (open circles), vs. radon concentration measured at home (a) and in the workplace or at school (b) in a group of 123 subjects.

the average radon exposure at home and the occupancy weighted exposure obtained using the occupancy factors and direct measurements were 0.82 and 0.85, respectively. 4. Discussion The results of this study can be compared with the occupancy factor of 0.66, for persons of 10 years and older, obtained in the Time Use Study performed in the years 1987–1988 by Statistics Finland using interviews and diaries of the study subjects [8]. Our results show that the time spent at home by the same age group is 0.73. The higher value obtained in this study was mostly due to the 10–24 year-old age group and, to some extent, the group aged 25– 44 years. The occupancy factors for the two older age groups did not differ between the two studies. The difference may be explained by a real change in the way younger generations use their time, or an error due to the non-response. Unit non-response is a problem in surveys of this kind, especially if the response rate is low. However, we could not identify any sources or error. Mjönes studied a non-random sample of adults in Sweden, obtaining a value of 0.62 for home occupancy, comparable to the corresponding ﬁgure in this study [9]. Brown studied the dose for the UK population, obtaining an occupancy weighted mean value of 0.75 for the population aged over 5 years [10]. This ﬁgure, obtained from a study by the Audience Research Department of the BBC, is also very comparable to this study. The residential indoor radon concentration is the main source of radon exposure in Finland, as expected, because most of the time is spent at home. Furthermore, radon concentration in buildings is highest at night [4]. The lower concentration in workplaces is partly due to the same diurnal variation and partly to the use of efﬁcient mechanical ventilation during working hours. The association of the radon concentration during working hours and the long-term alpha-track measurement needs further studies, as does the outdoor radon concentration. The publication ICRP 65 uses a value of 0.80 as an indoor occupancy factor for calculating doses from residential radon exposure [11]. The ratio of radon concentration at home and the

Indoor occupancy and radon exposure in Finland

693

occupancy weighted mean concentration would be a feasible ﬁgure for an occupancy factor for radon risk projection. It takes into account the lower concentration in workplaces and in public buildings. In this study the occupancy weighted mean radon concentration is, on average, about 15–18% lower than the radon concentration at home, a result in reasonable agreement with the occupancy factor of ICRP.

References [1] Committee on Health Risks of Exposure to Radon, Board on Radiation Effects Research, Commission on Life Sciences, National Research Council, Health Effects of Exposure to Radon: BEIR VI, National Academy Press, Washington, DC, 1999. [2] J.H. Lubin, J.D. Boice Jr., Lung cancer risk from residential radon: meta-analysis of eight epidemiological studies, J. Natl. Cancer Inst. 89 (1997) 49–57. [3] F. Lagarde, G. Pershagen, G. Åkerblom, O. Axelson, U. Bäverstam, L. Damber, A. Enﬂo, M. Svartengren, G.A. Svedjemark, Residential radon and lung cancer in Sweden: Risk analysis accounting for random error in the exposure assessment, Health Phys. 72 (2) (1997) 269–276. [4] H. Arvela, Residential Radon in Finland: Seasonal variation in radon concentrations of 3000 dwellings with model comparison, Radiat. Prot. Dosim. 59 (1) (1995) 33–42. [5] O. Castrén, H. Arvela, I. Mäkeläinen, A. Voutilainen, Indoor radon survey in Finland: methodology and applications, Radiat. Prot. Dosim. 45 (1/4) (1992) 413–418. [6] M. Annanmäki, E. Oksanen, M. Markkanen, Radon at workplaces other than mines and underground excavations, Environ. Int. 22 (Suppl. 1) (1996) S769–S772. [7] UNSCEAR (United Nations Scientiﬁc Committee on the Effects of Atomic Radiation), Sources and Effects of Ionizing Radiation, UNSCEAR 2000 Report to the General Assembly, with Scientiﬁc Annexes, United Nations, New York, 2000. [8] I. Niemi, H. Pääkkönen, V. Rajaniemi, S. Laaksonen, J. Lauri, Annual time use (in Finnish, Vuotuinen ajankäyttö, Ajankäyttötutkimuksen 1987–88 taulukot, Tutkimuksia 183 B), Statistics Finland, 1991. [9] L. Mjönes, Relation between residential gamma radiation and gamma dose to an individual (in Swedish, Samband mellan gammastrålning i bostaden och individuell gammastråldos), SSI rapport 86-22, Statens strålskyddsinstitut, Stockholm, 1986. [10] L. Brown, National radiation survey in the UK: indoor occupancy factors, Radiat. Prot. Dosim. 5 (4) (1983) 203–208. [11] ICRP Publication 65: Protection against radon-222 at home and at work, Ann. ICRP 23 (2) (1993).

694

Indoor radon: controlling factors, deﬁnition of the radon potential and its geographical distribution over Austria∗ P. Bossew, H. Lettner Institute of Physics and Biophysics, University of Salzburg, Hellbrunner Strasse 34, A-5020 Salzburg, Austria

In recent years indoor radon concentrations have been measured in several thousand Austrian buildings in the framework of the Austrian National Radon Project (OENRAP [1]). The measured 222 Rn concentrations not only depend on local conditions related to geology and soil permeability but also on the types of building or room in which the measurement has been performed, like ﬂoor level or window type. Therefore, in order to produce comparable results, a standardised quantity, called the radon potential, must be deﬁned. Furthermore, in order to be able to interpolate the radon potential between measured points and to draw radon maps, it is necessary to quantify its spatial behaviour, like regional tendencies and spatial correlations of the radon potential of locations separated by different distances. The paper discusses the factors which control the indoor radon concentration. Among the main factors are the level of the building in which the room under consideration is located and if the building has a basement, indicating its isolation against soil gas. The paper presents a deﬁnition of the radon potential and investigates its geographical distribution over the area of Austria. It turns out that, in spite of seemingly erratic ﬂuctuations of the radon potential which can often be observed on a local scale, on a regional scale there is a signiﬁcant, systematic spatial behaviour. The resulting radon potential map is presented as well as a radon risk map based on the technique of indicator kriging.

1. Introduction In the past 10 years indoor radon concentrations have been measured at about 7800 locations all over Austria (Fig. 1), in a total of more than 16 000 rooms. Together with the radon result a set of statistical parameters and factors was recorded relating to the locality in which the measurement was performed (geographical coordinates, type of building: basement, principal * Project supported by the Austrian Federal Ministry of Social Security and Generations.

RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07085-8

© 2005 Elsevier Ltd. All rights reserved.

Indoor radon: controlling factors, deﬁnition of potential, and geographical distribution over Austria

695

Fig. 1. Locations of 222 Rn measurement sites in Austria.

construction material, number of ﬂoors, etc.), related to the ﬂat (located on which ﬂoor, heating system, type of windows, number of children living in the place, etc.), and ﬁnally to the type of measurement (time of year, measurement technique). The aim of the analysis was to investigate the controlling factors for the indoor radon concentration and whether the variable shows a regular spatial behaviour, such that it can be interpolated spatially and displayed as a geographical map, or rather ﬂuctuate erratically in space making any mapping impossible.

2. Methods The acquisition of the Radon data was not a subject of this project; details about the methodology can be found in the original reports of the OENRAP project [2]. For the statistical evaluations presented here, standard statistical software like SPSS, Origin, Variowin and Surfer was used.

3. Controlling factors The inﬂuence of the factors, other than site-speciﬁc (geologic etc.), which control the indoor Rn concentration was investigated in the following way. For the inﬂuence of the ﬂoor number in which the radon concentration was measured, the results of the measurements in different ﬂoors of one multi-ﬂoor building were compared, and the results averaged (geometric mean) over all such buildings. Together with the inﬂuence of whether the building has a basement, this is the most important controlling factor. Its signiﬁcance is shown in Fig. 2. In particular, basements show highly increased Rn concentrations, whereas no inﬂuence can be found for rooms located higher than ﬁrst ﬂoor. For the other factors, buildings in the same geographical location (within the accuracy of its determination, i.e. between 100 m and a few km, considered as the nugget region in a geostatistical sense) with different values of a factor, like basement yes/no, window type single/double etc., were compared, and again averaged (geometric mean) over the locations. If in one location several rooms with the same speciﬁcation were found the geometric mean of

696

P. Bossew, H. Lettner

Fig. 2. Dependence of the relative Rn concentration on the ﬂoor in which is measured. (constraint: y(0) = 1).

the associated Rn concentrations was taken. In many cases second level inﬂuences were also investigated (i.e. the inﬂuence of the variation of a second factor while one is kept constant). The main controlling factors are: • • • • • •

the ﬂoor in which the measurement is being made; the total number of ﬂoors; whether the building has a basement; the type of windows; the type of heating system; and the number of children living in the ﬂat (the physical cause has not been investigated: possibly via the increased frequency of opening doors).

Other factors from the list in the dataset related to each measurement, like the year of construction of the building, are also highly signiﬁcant controlling factors, but are strongly correlated to (in this case) the presence of a basement, which can be shown by contingency analysis; therefore they need (and must) not be considered as additional, independent controlling factors. To each controlling factor a value which quantiﬁes its inﬂuence is assigned.

4. The radon potential In order to obtain spatially comparable results for the indoor radon concentration, a quantity is deﬁned post hoc from the measured indoor Rn concentration values, taking into account the independent non-site speciﬁc (geological, etc.) controlling factors, as described above. The other way round, a given Rn potential can serve to estimate the indoor Rn concentration in a building and a room of certain given types. The radon potential is deﬁned as the Rn concentration in a standard building and room, namely:

Indoor radon: controlling factors, deﬁnition of potential, and geographical distribution over Austria

• • • • • •

697

a building without basement; a building with 1 to 3 ﬂoors; windows: double or insulating glass; heating (relevant only for measurements in winter): gas, district heating or oven; one child per ﬂat; measurement in the ground ﬂoor.

Appropriate factors are applied to calculate the Rn potential in non-standard situations (results of the analysis of the controlling factors; for example, the presence of a basement reduces the indoor Rn concentration by a factor 0.665), or vice versa, to estimate the expected indoor Rn concentration in non-standard rooms from the Rn potential (e.g., for a building with basement, the Rn potential value must be multiplied by 0.665). The Rn potential is deﬁned such that, ideally, it depends only on local geological or geophysical factors (like permeability), having eliminated the inﬂuence of factors which are speciﬁc to the building or the measuring circumstances.

5. Radon potential and radon risk maps The spatial correlation of the Rn potential was investigated with the help of the variogram technique. Figure 3 shows one example of the standardised variogram for the logarithms of a set of Rn potential data (logs chosen because of their approximate log-normal distribution). Fitting of a spherical model yields a correlation length (range) of 75 km and a nugget of 63%. (Further substructures of the variograms are recognisable but have not been evaluated nor taken into account.) Similar variograms have been found be Zhu et al. [3] in their radon survey of Belgium: Their main ranges were somewhat smaller with 30–40% though. The result was used as input for the kriging procedure in order to draw a radon potential map (Fig. 4). Using the indicator kriging technique a probability map has been drawn for the radon potential to exceed 400 Bq m−3 (the Austrian intervention limit), see Fig. 5.

Fig. 3. Radon potential: empirical standardised variogram of the logarithms of the data; lag (h) = 1000 m, x-axis unit: m.

698

P. Bossew, H. Lettner

Fig. 4. Radon potential map of Austria. Interpolator: kriging, spherical variogram, range = 75 km, total variance = 65 900 (Bq m−3 )2 , nugget = 63%, no drift and no anisotropy assumed. Grid = 10 km; coordinates: Lambert (m).

Fig. 5. Radon risk map. Grid = 5 km; interpolation: kriging, spherical variogram, nugget = 77%, range = 8.1 km, no anisotropy; coordinate: Lambert (m).

References [1] H. Friedmann, Radon levels in Austria, in: Proc. 5th Conference on High Levels of Natural Radiation and Radon Areas; Radiation Dose and Health Effects, Munich, 4–7 September 2000. [2] H. Friedmann, Ermittlung der Strahlenbelastung der österreichischen Bevölkerung durch Radonexposition und Abschätzung des damit verbundenen Lungenkrebsrisikos – Pilotprojekt. Beiträge, Forschungsberichte des Bundesministeriums für Gesundheit, Sport und Konsumentenschutz, 3/1994 (in German). [3] H.C. Zhu, J.M. Charlet, A. Poffjin, Radon risk mapping in Southern Belgium: an application of geostatistical and GIS techniques, Sci. Total Environ. 272 (2001) 203–210.

699

Simultaneous measurement of radon and thoron exhalation rate from soil and building materials C. Cosma a , O. Cozar a , T. Jurcut b , C. Baciu c , I. Pop d a Babes-Bolyai University, Physics Department, 3400 Cluj-Napoca, Romania b University of Oradea, Sciences Department, 3700 Oradea, Romania c Babes-Bolyai University, Geology Department, 3400 Cluj-Napoca, Romania d Technical University, Physics Department, 3400 Cluj-Napoca, Romania

Our paper presents two methods for simultaneous measurement of radon and thoron exhalation from soil and building materials: (1) charcoal adsorption, (2) Lucas cell methods. Both methods are applicable especially in the case of a soil ﬂux enhanced in thoron gas. In the Lucas cell case, a short accumulation time was used (10–15 minutes). The thoron concentration in the accumulation volume was measured immediately after the sample gas extraction and a regression equation is used for determining equilibrium thoron concentration. In the case of the charcoal method the thoron was measured 4 hours after a special degassing of the sample. Using the LUK-3A device for the building materials, the thoron ﬂux was measured only on a special enhanced thorium + radium sample. 1. Introduction The radon isotopes 220 Rn (thoron) and 222 Rn (radon) were intensely studied in recent times due to their involvement in lung cancer risk [1] and also for geological purposes [2]. The main sources of the indoor radon are the soil radon and radon exhalation from building materials [3]. Active charcoal was often utilized for radon measurement especially for indoor radon [4,5]. The method of soil radon and exhalation measurement from the soil and building materials using adsorption in active charcoal was also used [6–8] with good results but in this case corrections regarding charcoal humidity related to the break point of charcoal are needed [9]. In all these cases the radon adsorbed in charcoal is measured by gamma spectrometry, commonly with NaI (Tl) detectors. For economic and practical reasons, the charcoal canisters must be degassed and re-used many times. Commonly, charcoal cleaning is achieved by degassing the charcoal at 120–140 ◦ C for 12–14 hours, but in the case of high radon content adsorbed in charcoal, for complete desorption of radon, 2–3 steps are necessary [10]. The method for charcoal degassing proposed in this work has the advantage of being more rapid RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07086-X

© 2005 Elsevier Ltd. All rights reserved.

700

C. Cosma et al.

(10–15 minutes) and it produces a much better degassing. On the other hand, the degassing method allows thoron determination from the charcoal even in the case when it is measured by gamma spectrometry using NaI (Tl) detectors. Generally, the measurement methods for thoron determination require either alpha or high resolution gamma spectroscopy. This work presents a method for simultaneous measurement of radon and thoron from soil using charcoal adsorption. Exhalation of radon from some Romanian building materials obtained by charcoal adsorption method was veriﬁed in several cases by the Lucas cell method. In the case of one sample the result was compared with those obtained in the framework of the ERRICCA Intercomparison Exercise from Athens in April 1999 [11]. Finally, both radon and thoron exhalation from a special concrete sample made in our laboratory, artiﬁcially enhanced in 226 Ra and in 224 Ra(Th[NO3 ]2 ) was measured. 2. Experimental methods A brass cylindrical vessel of 0.7 L, Fig. 1, was used for charcoal degassing. In the upper part it is connected to a vacuum pump. After ﬁlling (250 g of charcoal) and closing using six screws and a lead ﬁtting, the vessel is put into another larger vessel with water. The latter is placed on an electric heater. When reaching water boiling temperature, the pump is started for 15 minutes. The charcoal is then removed and can be used for radon-thoron adsorption in the next exposure. By measuring the charcoal radioactivity after this operation, degassing efﬁciency was monitored. After about 3.5 hours from the end of the degassing process, the gamma radioactivity is reduced to the background radioactivity of the charcoal. For radon exhalation gathered from soil the device from Fig. 2 was used, which consists of a frame box of 0.23 m2 (37 × 62 cm) and about 200 g degassed charcoal. After exposure (as a rule, 4–12 hours) the charcoal is introduced into the Marinelli vessel and the gamma radioactivity is measured over the whole spectrum with an NP424 gamma spectrometer equipped with a large NaI (Tl) detector (76 × 45 mm) enclosed in a lead castle of 30 mm wall thickness to reduce the background.

Fig. 1. The brass device for radon degassing.

Fig. 2. The device for radon ﬂux gathering.

Simultaneous measurement of radon and thoron exhalation rate from soil and building materials

701

In the case of thoron being absent from the charcoal, the radon exhalation rate ΦRn can be calculated by following equation: ΦRn = C

ln 2 A0 ·t · e TRn c 2 te (h)S(m )

Bq/ m2 s ,

(1)

where C = 1.412 × 10−3 is the calibration factor, A0 denotes counts per second under the integral of the spectrum (E0 > 60 keV); te is exposure time (usually 4–12 hours); S = 0.23 m2 the collecting area; T = 3.825 days is the half-life of radon and tc the time interval between the middle of the exposure time and the beginning of the measurement. A time interval longer than 3.5 hours must be considered between the end of sampling and the start of measurement. In the case of signiﬁcant thoron exhalation accompanying the radon ﬂux, this equation can be used only if tc is 5–6 times greater than the half time of thoron descendants (10.65 h) that is about two days. In this time the charcoal from the Marinelli vessel must be very well sealed so that the radon remains in this container. The thoron exhalation can be determined if the exposed charcoal is ﬁrstly degassed as above using the device from Fig. 1 and the sample is gamma counted after 4 hours when the radon progeny have decreased by about 500 times. Figure 3 represents such an experiment. Therefore the initial sample must be divided in two: one for radon and the other for thoron measurement. For building material exhalation measurement two types of square (33 × 33 cm2 ) metallic and plastic containers of 8 cm height and 20 cm height respectively were used. These were placed on a shiny surface and air sealed with a special material. On the upper part, there is a small oriﬁce (∅ = 0.3 mm), which must be also sealed by a small quantity of the same material. These oriﬁces allow the extraction, with a special medical syringe, of a known gas volume from the container, which is then analyzed using the Lucas cell method with a LUK 3A device [12]. LUK 3A device produced in the Czech Republic is a portable instrument, programmable for three kinds of measurements: soil radon, soil radon and thoron and radon in water. The samples (concrete, bricks, tills) were enclosed in these containers together with 50 g of recently degassed activated charcoal. The charcoal thickness in this accumulation volume was 1.5–2 mm assuring an efﬁcient adsorbing process both for radon and thoron atoms. After the adsorption time (2–96 h) the charcoal was put into well-sealed standard metallic cans and measured by means of the same NaI (Tl) spectrometer for 500 s. The calibration

Fig. 3. Thoron progeny decreasing four hours after the exposed sample was degassed (T1/2 = tg α = 11.3 h).

702

C. Cosma et al.

for measuring under the whole spectrum (E0 > 60 keV) was made using 226 Ra and 224 Ra (232 Th) sources and the values obtained were c1 = 3.02 Bq count−1 s−1 for radon and c2 = 4.64 Bq count−1 s−1 for thoron, respectively. For energy values higher then 2400 keV, this factor for thoron is 320 Bq count−1 s−1 . The emanated radon quantity A0 (Bq) was adjusted by considering the quantity of radon adsorbed in the charcoal Ac and the quantity Aa present in the real emanation volume V0 − Vs following the equations: A0 = Ac + Aa , Ac = km

Aa , V0 − Vs

(2) (3)

where k = 1.41 m3 kg−1 is the adsorption constant for our charcoal (type CAS-Buzau), m = 50 g being the mass of adsorbing charcoal, and V0 and Vs representing the container volume and the sample volume, respectively. The radon exhalation ﬂux can be calculated as follows: ln 2 A0 (4) · (Bq s−1 ). TRn 1 − e−λT In the case of a long collection time, a correction of about 10–15% is needed due to leakage and back-diffusion processes. In the case of the sample “Specimen 3” measured in the framework of the ERRICCA Intercomparison Exercise [11] another method was also used to determine the radon exhalation from this sample. In this case, a Radim device (Czech Republic) that continuously measures the radon concentration was introduced. The Radim instrument registers the α radiation from radon and its descendants using a Si detector. In Fig. 4 are shown the results of radon activity in the emanation volume as a time function measured with the Radim instrument. From the curve from Fig. 4 it is possible to determine the radon exhalation rate by interpolation of this curve via equation [13]:

C = C0 1 − exp(−λ1 t) . (5) Φ=

In relation (5) λ1 = λRn + λ∗ , λ∗ being the leakage and the back-diffusion constant. The exhalation rate can be found from the relation: Φ = C0 (V0 − Vs ) · λ1 .

Fig. 4. Radon variation in the accumulation container measured with the Radim cell method.

(6)

Simultaneous measurement of radon and thoron exhalation rate from soil and building materials

703

By ﬁtting the experimental points to equation (5) we obtained the following values C0 = 8950 Bq m−3 , λ1 = 0.0105 h−1 and λ∗ = λ1 − λRn = 0.0035 h−1 . From this latter value one can obtain information about the leakage and back-diffusion processes.

3. Results and discussion Table 1 presents the results for soil exhalation using both the charcoal adsorption and Lucas cell methods measured in Cluj-Napoca, where a preliminary high thoron concentration was found in the soil. As one can see from this table, there is a good concordance between the results obtained using the two methods for different sampling locations. The thoron ﬂux is higher than the radon ﬂux in good agreement with other works [14,15]. Table 2 shows the results of radon exhalation from seven different materials. The ﬁrst sample in this table is Specimen 3, which was used in the ERRICCA Intercomparison Exercise from Athens (April 1999) in which many laboratories from Europe participated. Our results obtained using the charcoal method in its two variants and the results obtained with the Lucas cell and Radim methods are in very good agreement with the average value found for this sample by the ERRICCA participants [11], Fig. 5. The samples No. 2 and No. 3 from Table 2 were prepared in our laboratory to verify some models of radon migration in building materials. They are the concrete samples poured following standard technology but enhanced in 226 Ra in two different ways. In preparing sample Table 1 Radon and thoron exhalation (mBq m−2 s−1 ) from soil in the Cluj-Napoca area Place

Charcoal method

A B C

Lucas cell method

Radon ﬂux

Thoron ﬂux

Radon ﬂux

Thoron ﬂux

25.5 ± 2.4 36.1 ± 3.5 30.7 ± 3.5

2180 ± 150 3350 ± 210 2430 ± 180

27.1 ± 1.8 37.4 ± 2.5 32.2 ± 2.1

1980 ± 130 3060 ± 170 2230 ± 140

Table 2 Radon exhalation rates from building material No

1 2 3 4 5 6 7

Sample

Spec 3 ERRICCA Enhanced Ra Enhanced U Concrete BC-20 Bricks Tills BCA

Mass (kg)

19 5.8 4.2 8.5 3.2 5.9 4.5

226 Ra

Radon exhalation (Bq kg−1 h−1 )

(Bq kg−1 )

Charcoal

Lucas cell/Radim

107 ± 5 275 ± 10 272 ± 5 – – – –

0.262 ± 0.022 0.845 ± 0.035 0.067 ± 0.004 0.025 ± 0.002 0.014 ± 0.002 0.016 ± 0.002 0.059 ± 0.003

0.252 ± 0.015 0.810 ± 0.024 0.059 ± 0.004 – – – 0.051 ± 0.003

704

C. Cosma et al.

Fig. 5. The results of the ERRICCA intercomparison exercise [11].

No. 2 a RaCl2 solution was used and for sample No. 3 a ﬁne powder of pitchblende (uranium ore) was melted with the cement used for preparing the sample [16]. The results for two of these samples are quite different because for sample No. 2 the radium atoms are ﬁxed to the surfaces of the mineral grains (the emanation coefﬁcient being high) whereas for sample No. 3 the radium atoms are uniformly distributed in all the small mineral grains of pitchblende ore. The samples Nos. 4–7 are building materials used in Romania at this time. The emanation rates found by us range in the interval mentioned in the literature [3]. To test the possibility of simultaneous measurement of radon and thoron exhalation using the charcoal adsorption method and gamma counting with NaI (Tl) or other detectors with similar resolutions, a special sample of concrete containing 1520 Bq of 226 Ra and 4230 Bq of 224 Ra was prepared. The 226 Ra derived from the same RaCl solution and for 224 Ra 2 g of old 2 Th(NO3 )2 4H2 O in which 224 Ra is in equilibrium with 232 Th was dissolved in the water used to prepare the concrete sample.

Simultaneous measurement of radon and thoron exhalation rate from soil and building materials

705

The thoron exhalation rate was determined by the charcoal adsorption method obtaining 49 Bq kg−1 h−1 and also by the Lucas cell method giving a similar value of 44.5 Bq kg−1 h−1 . The radon exhalation rate was similar to sample No. 2 from Table 2.

References [1] W.W. Nazaroff, A.V. Nero (Eds.), Radon and its Decay Products, Wiley, New York, 1988. [2] G. Akerblom, H. Melander, Geology and radon, in: S.A. Durrani, R. Ilic (Eds.), Radon Measurements by Etched Track Detectors, Word Scientiﬁc, Singapore, 1996. [3] E. Stranden, Building materials as source of indoor radon, in: W.W. Nazaroff, A.V. Nero (Eds.), Radon and its Decay Products in Air, Wiley, New York, 1988. [4] A. George, Health Phys. 46 (1984) 867. [5] H.M. Prichard, K. Marien, Health Phys. 48 (1985) 797. [6] K. Megumi, T. Mamuro, J. Geophys. Res. 79 (1979) 3357. [7] Y. Li, S.D. Schery, B. Turk, Health Phys. 62 (1992) 453. [8] S. Oberstedt, H. Vanmarcke, Radiat. Prot. Dosim. 63 (1996) 69. [9] S.C. Scarpitta, Health Phys. 60 (1996) 673. [10] C. Cosma, A. Van Deynse, A. Pofﬁjn, Radiat. Measur. 31 (1999) 351. [11] N.P. Petropoulos, M.I. Anagnostakis, S.E. Simopoulos, Sci. Total Environ. 272 (2001). [12] C. Cosma, A. Pofﬁjn, D. Ristoiu, Indoor Built Environ. 5 (1996) 236. [13] N. Jonassen, Health Phys. 45 (1983) 369. [14] W.D. Crozier, J. Geophys. Res. 74 (1969) 4199. [15] S.D. Schery, Studies of thoron and thoron progeny: Indications for transport of radioactivity from soil to indoor air, in: Proc. APCA Int. Specialty Conf., Philadelphia, 1986. [16] C. Cosma, F. Dancea, T. Jurcut, D. Ristoiu, Appl. Radiat. Isot. 54 (2000) 467.

706

Exposure to radon in the northern part of the Republic of Kyrgyzstan M. Zhukovsky a , R. Termechikova b a Institute of Industrial Ecology, Sophy Kovalevskoy St., 20A, 620219 Ekaterinburg, Russia b Issyk-Kul State University, K. Tynystanov St., 103, 722360 Karakol, Kyrgyzstan

Indoor radon levels in the Issyk-Kul region of the Republic of Kyrgyzstan are investigated in detail for the ﬁrst time. The average winter and summer radon concentrations were 270 and 180 Bq m−3 , respectively; the winter to summer ratio was calculated to 1.5. There is no statistically signiﬁcant correlation between winter and summer Rn concentration values in dwellings. The inﬂuence of building type (rural or urban) or wall materials on the radon concentration was negligible. Considerable thoron (220 Rn) decay product equivalent equilibrium concentrations (EEC) were observed (average values 5.9 and 15 Bq m−3 for urban and rural dwellings, respectively). An assumption of the dominant role of the diffusion process in radon and thoron entry into dwellings was invoked.

1. Introduction High levels of exposure due to radon, originating in the geological structures, are expected countrywide in the Republic of Kyrgyzstan and especially in the Issyk-Kul region (in the northern mountainous part of the country). The basis for this assumption is the elevated background gamma dose rate in some areas of the Issyk-Kul region (200–300 nGy h−1 ) and increased 226 Ra and 232 Th speciﬁc activity in rocks and building construction materials (up to 100–150 Bq kg−1 for 226 Ra and 100–160 Bq kg−1 for 232 Th). Unfortunately, there exist no data on the radon levels in the dwellings in this region. Therefore, in order to assess the levels of radiation exposure in dwellings, a pilot radon survey for this area was conducted in the period 1999 to 2002. Radon measurements were carried out in 20 settlements situated around Issyk-Kul lake in the Tien-Shan mountains. Twelve settlements were situated to the north and eight to the south of Issyk-Kul lake. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07087-1

© 2005 Elsevier Ltd. All rights reserved.

Exposure to radon in the northern part of the Republic of Kyrgyzstan

707

2. Methods Generally, two techniques were implemented for the radon survey in this region: application of cellulose nitrate (LR-115) alpha track detectors and grab sampling of Rn and Tn decay products by means of a modiﬁed version of the Markov method [1]. The equilibrium equivalent concentrations (EEC) of radon and thoron decay products were determined by measuring the α-activity on ﬁlters after sampling for radioactive aerosols. This method includes a 5min air sampling and 3 cycles of α-activity measurements conducted from the 1st to the 4th minute, from the 7th to the 10th minute and from the 300th to the 330th minute after the end of sampling. The third cycle of measurement is necessary in the case of high thoron decay product activity. The activity concentrations of individual radon decay products, radon and thoron EEC can be calculated by the equations: CPo-218 =

4.37 N1 − N2 + 4.19 · 10−3 N3 , εηυ

(1)

CPb-214 =

1.11 N2 − 7.23 · 10−2 N3 , εηυ

(2)

CBi-214 =

1 2.21N2 − 0.90N1 − 2.62 · 10−3 N3 , εηυ

(3)

EECRn =

1.16 N2 − 7.23 · 10−2 , εηυ

(4)

EECTn =

0.1408 N3 , εηυ

(5)

where N1 , N2 and N3 are numbers of counts during the ﬁrst, second and third time interval of measurement; ε is registration efﬁciency; η is ﬁlter efﬁciency; υ is rate of air sampling (L min−1 ). Using the equilibrium shift between the radon decay products and thoron EEC value the equilibrium factor F between radon gas activity concentration and equivalent equilibrium concentration was assessed [2] by the equation EECTn 0.772 · EECRn 1 − 0.180 · . F= CPo-218 EECRn

(6)

Cellulose nitrate detectors were placed in a special plastic measuring chamber with silicon rubber ﬁlter, which prevented the penetration of 220 Rn and radon decay products into the chamber. The measuring chambers containing the detectors were exposed over 1–3 month periods. The detector material was etched following exposure, in 6N NaOH solution, and track density was determined using a spark counter. The combination of these techniques enables the acquisition of data on the activity concentration of radon gas, the equilibrium factor F and the equivalent equilibrium concentration (EEC) of thoron decay products.

708

M. Zhukovsky, R. Termechikova

3. Results 3.1. Radon concentration distributions Radon concentration measurements were carried out in 113 urban and 630 rural dwellings. In 73 urban and 396 rural dwellings, measurements were carried out in both winter and summer periods. In general, the radon activity concentrations can be described by log-normal distributions (Figs. 1 and 2). Due to the log-normal distribution of radon concentrations, the logarithms of radon concentrations (not the radon concentrations themselves) were further analyzed by standard statistical techniques. Contrary to our results obtained in the north region of Russia [2], in the case of the Issyk-Kul region, there was very little difference between radon concentrations in the urban and rural dwellings. In general, this difference was not statistically signiﬁcant. Only if one separates the sample into “north” and “south” subgroups could a signiﬁcant difference (at p < 0.05) between the urban and rural dwellings in the summer season be detected. The parameters of the distribution of the radon activity concentrations in the Issyk-Kul region are presented in Table 1.

Fig. 1. Distribution of winter radon activity concentration in rural dwellings.

Fig. 2. Distribution of summer radon activity concentration in rural dwellings.

Exposure to radon in the northern part of the Republic of Kyrgyzstan

709

Table 1 Parameters of the distribution of the radon activity concentration Type of dwelling

Winter period Arithmetic mean (Bq m−3 )

Geometric mean (Bq m−3 )

GSD

Summer period Arithmetic mean (Bq m−3 )

Geometric mean (Bq m−3 )

GSD

Urban Rural

235 267

211 200

2.16 2.13

212 181

124 123

3.07 2.51

3.2. Seasonal variations of Rn concentration The seasonal radon concentration dependence CWIN = f (CSUM ), CSUM = f (CWIN ) and the variation in the seasonal distribution of radon concentration were estimated. These parameters can usually be used for the assessment of the annual average radon concentration on the basis of the measurement results from a single measuring period. Unlike the northern regions of Russia [2], no statistically signiﬁcant correlations between CWIN and CSUM were found. There is a statistically signiﬁcant difference CWIN − CSUM , but there are no statistically signiﬁcant correlations between this difference and CWIN or CSUM . The distribution of CWIN − CSUM can be described by a normal distribution and is presented in Figs. 3 and 4 for rural and urban

Fig. 3. Distribution of difference CWIN − CSUM for rural dwellings.

Fig. 4. Distribution of difference CWIN − CSUM for urban dwellings.

710

M. Zhukovsky, R. Termechikova Table 2 Parameters of the distribution of the radon activity concentration seasonal difference CWIN − CSUM Type of dwelling

Arithmetic mean (Bq m−3 )

Standard deviation (Bq m−3 )

Urban Rural

90 86

247 284

Table 3 Parameters of the distribution of the equilibrium factor, F Type of dwelling

Arithmetic mean

Geometric mean

GSD

Urban Rural

0.26 0.28

0.21 0.24

2.20 3.00

dwellings, respectively. The parameters of these distributions are presented in Table 2. There are no statistically signiﬁcant differences in the parameters of these distributions. 3.3. Equilibrium factor F In order to determine the equilibrium factor F we used data on the equilibrium shift between 218 Po and 214 Pb activity concentrations, obtained from the results of grab sampling performed

in the summer period [2]. The distribution of the equilibrium factor F is approximately lognormal. The parameters of this distribution are presented in Table 3. These values are lower than the value of F = 0.4 recommended by the ICRP, and considerably lower than the typical value of F = 0.5 for northern countries. 3.4. Radon concentration dependence on building construction The dependence of indoor radon concentrations on the wall materials, wall cover, age of building and the ﬂoor of measurement (for urban multi-storey houses) were analyzed. The dependence of radon concentrations on the wall materials is presented in Figs. 5 and 6. For urban dwellings, the differences were statistically signiﬁcant only for the summer period. For rural houses, minimal radon concentrations were observed for adobe houses both in winter and summer periods. For the summer period, there is a statistically signiﬁcant decrease of the mean radon concentration for dwellings with walls covered by wallpaper (Figs. 7 and 8). Both these facts can represent circumstantial evidence of the considerable inﬂuence of radon diffusion from building materials, because minimal radon diffusion coefﬁcients are typical and wallpaper can act as a barrier for radon diffusion entry from the adobe blocks. An unusual dependence of radon concentration on the age of buildings was observed (Figs. 9 and 10). For the summer period, the average radon concentrations both for urban and rural dwellings are not dependent on the age of the building. For the winter period, the differences in the average radon concentration in age subgroups are not signiﬁcant for the urban dwellings. But there is a statistically signiﬁcant decrease of the average radon winter concentration for the old rural dwellings.

Exposure to radon in the northern part of the Republic of Kyrgyzstan

Fig. 5. Dependence of winter radon concentrations on wall materials.

Fig. 7. Dependence of winter radon concentrations on wall cover materials.

Fig. 9. Dependence of winter radon concentrations on the age of buildings.

711

Fig. 6. Dependence of summer radon concentrations on wall materials.

Fig. 8. Dependence of summer radon concentrations on wall cover materials.

Fig. 10. Dependence of summer radon concentrations on the age of buildings.

For urban dwellings in the winter period, radon concentrations tend slightly to decrease depending on the ﬂoor measured; this tendency is not statistically signiﬁcant. For the summer period, radon concentration does not depend on the ﬂoor measured (Figs. 11 and 12).

712

M. Zhukovsky, R. Termechikova

Fig. 11. Dependence of winter radon concentrations in urban dwellings on the ﬂoor of measurement.

Fig. 12. Dependence of summer radon concentrations in urban dwellings on the ﬂoor of measurement.

Table 4 Parameters of the distribution of the thoron EEC in the Issyk-Kul region in the summer period Type of dwelling

Number of observations

Arithmetic mean (Bq m−3 )

Geometric mean (Bq m−3 )

GSD

Urban Rural

48 94

5.9 15

3.7 11

3.00 2.41

3.5. Thoron equivalent equilibrium concentration (EECTn ) The preliminary results from the determination of the equivalent equilibrium concentration for thoron (EECTn ) are presented in Table 4. The thoron grab sampling measurements were considerably complicated by the high air temperature (nearly 40 ◦ C), which occurred during the performance of alpha radiometry of aerosol ﬁlters. One of the speciﬁc characteristics of the Issyk-Kul region is the elevated value of the average thoron EECTn . The average thoron EECTn values for this region considerably exceed the average worldwide value [3]. This can be explained by the use of local minerals with high 232 Th speciﬁc activity (100–160 Bq kg−1 ) as building materials.

4. Discussion The results of our research show that there is no adequate model for description of radon seasonal variations. Therefore, the average annual values of radon EEC were calculated by use of direct pair radon activity concentration measurements conducted both in winter and summer periods and measured values for the equilibrium factor F . The equilibrium factor was not determined for the winter periods, so for the cold season (lasting 3 months) we assumed an equilibrium factor of F = 0.5. For the warm season the value of the equilibrium factor was rounded to F = 0.3. The results obtained are presented in Table 5. The differences observed in the mean values or the geometric standard deviation are not statistically signiﬁcant. The

Exposure to radon in the northern part of the Republic of Kyrgyzstan

713

Table 5 Parameter of the distribution of the average annual radon EEC Type of dwelling

Number of observations

Arithmetic mean (Bq m−3 )

Geometric mean (Bq m−3 )

GSD

Urban Rural All dwellings

73 396 472

82 74 75

65 62 62

1.97 1.82 1.84

Table 6 The average annual radon and thoron effective doses Type of dwelling

Rn daughters exposure (mSv)

Tn daughters exposure (mSv)

Urban Rural

3.5 3.2

1.7 4.2

expected occurrence of dwellings with radon EEC above 200 Bq m−3 was 2.8%. The actually observed frequency of dwellings with radon EEC exceeding this action level was 3.0%. The effective doses to population exposed to radon and its decay products were calculated assuming a total of 7000 hours per year spent indoors. For the calculation of the exposure to radon and its decay products, the dose conversion factors recommended by ICRP Publication 65 [4] were used. For exposures to thoron decay products the dose conversion factors recommended by UNSCEAR [3] were used. The average annual effective doses due to exposures to radon and thoron decay products are presented in Table 6. To estimate the dose contribution due to thoron decay products, the doses due to the building materials with 232 Th speciﬁc activity of 100 Bq kg−1 were calculated by the Monte-Carlo method using RESRAD-BUILD 3.1 software. The variable quantities were the radon diffusion coefﬁcient and the thickness of the contaminated layer. The expected mean annual dose due to thoron decay product inhalation was 2.4 mSv with 5 and 95% percentiles of 1.5 and 3.0 mSv, respectively. These results are in agreement with experimental data. Nevertheless, more detailed thoron decay product studies in the Issyk-Kul region are necessary. The non-occurrence of a seasonal radon concentration correlation CWIN = f (CSUM ), CSUM = f (CWIN ) and the absence of a dependence of radon concentration on the type of dwelling can be explained by the diffusion processes being the predominant cause of radon entry into the house (at least during the summer period). During the most extensive warm season, the rooms in houses are either widely open (there is no pressure difference between indoors and outdoors) or tightly closed as a protection from the heat (the air temperature indoors is lower than outdoors). In any case, there is no convective ﬂow of radon-bearing soil gas. The additional arguments supporting this assumption are the correlation of summer radon concentrations with wall materials and covers (Figs. 6 and 8), where the minimal radon concentration values were observed for dwellings with the minimal expected diffusion entry rates.

714

M. Zhukovsky, R. Termechikova

5. Conclusion The existence of a problem of exposure to radon and thoron decay products in the Issyk-Kul region of Republic of Kyrgyzstan has been demonstrated. It is shown that seasonal variations of radon concentrations cannot be ﬁtted by a satisfactory model. Therefore, full year radon measurements are necessary for correct estimation of radon indoor radon exposures. The population radiation exposure due to elevated thoron decay product concentrations is comparable with that for radon exposure. There is a necessity for further detailed investigations of thoron exposure and of mechanisms of radon entry into houses in this region which is characterized by speciﬁc climatic and geological features.

References [1] M.V. Terentyev, Sov. Atom. Energ. 61 (1987) 714 (English translation). [2] M. Zhukovsky, I. Yarmoshenko, Radiat. Prot. Manage. (2) (1998) 34. [3] UNSCEAR (United Nations Scientiﬁc Committee on the Effects of Atomic Radiation), Sources and Effects of Ionizing Radiation, UNSCEAR 2000 Report to the General Assembly, with Scientiﬁc Annexes, United Nations, New York, 2000. [4] ICRP Publication 65: Protection against radon-222 at home and at work, Ann. ICRP 23 (2) (1993).

715

Stratospheric radon measurements in three North American locations I.M. Fisenne a , L. Machta b,1 , N.H. Harley c,2 a US Department of Energy Environmental Measurements Laboratory, 201 Varick Street,

New York, NY 10014-4811, USA b NOAA Air Resources, 1315 East West Highway, Silver Springs, MD 20910, USA c N. H. Harley, New York University School of Medicine, 550 First Avenue, New York, NY 10016, USA

Stratospheric air samples were collected by the US Weather Bureau (now NOAA) in 1962 to explore the possibility of using radon proﬁles as an atmospheric tracer. In the Spring of 1962, WB-57 aircraft collected tropospheric and stratospheric air samples by pressurizing steel spheres with the samples to about 21 MPa. The sampling locations were Alaska, the Southwest USA and the Panama Canal Zone at 8, 32 and 70 degrees North, respectively. The samples were obtained at from 3 to 6 different altitudes ranging from 4 to 20 km, both below and above the tropopause which varied from 8 to 17 km. The 222 Rn concentrations ranged from 1 mBq m−3 at 20 km to 1000 mBq m−3 at 4 km, compared with an average ground level concentration of 15 000 mBq m−3 . The 222 Rn concentration proﬁles for the Canal Zone and the Southwest USA were similar with height, while the Alaskan proﬁles were lower by a factor of 5 to 10. The tropopause appears to be a more effective barrier to 222 Rn transport in the Canal Zone and Alaska than over the Southwest USA. The reason for this effect is not known but may reﬂect the bearing of the jet stream on lower altitude turbulence.

1. Introduction The testing of nuclear weapons in the atmosphere lead to the collection and measurement of the debris not only at the earth’s surface but also at high altitudes using aircraft as the sampling platform. The measurements of fallout radionuclides increased our knowledge of atmospheric circulation, including the discovery of the jet stream and mixing between the northern and southern hemispheres. The emphasis on the measurement of particulate radionuclides also yielded information on attachment rates and particle growth. To give a better description of 1 We regret to report the death of our co-author Lester Machta in August 2001 2 The author gratefully acknowledges the support of USDOE EMSP Contract DE-FG07-97ER62522.

RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07088-3

Published by Elsevier Ltd.

716

I.M. Fisenne et al.

atmospheric circulation, it was thought that the measurement of a gas should be the next step. The naturally occurring noble gas, 222 Rn, with a 3.82 days’ half-life seemed an ideal candidate. Its properties as a noble gas guaranteed that once released from the earth’s surface it would not be subject to chemical changes and could act as a tracer of atmospheric transport processes. There were reports conﬁrming the decrease in radon concentration with height above the earth’s surface [1–3]. This would be expected as the gas is acted upon by tropo-spheric mixing phenomena and is subject to radioactive decay. The ability to collect and measure radon in the stratospheric as well as the tropospheric air became a joint effort between the US Weather Bureau (USWB) and Argonne National Laboratory (ANL) [4]. In May and June 1961, two series of ﬂights to collect whole air samples were conducted over Alaska (70◦ N) and Hawaii (15◦ N) at altitudes of 7.6 to 19.8 km. In general terms these measurements showed the stratospheric radon concentration to be similar for these locations. Estimates were made on vertical diffusion coefﬁcients and it was suggested that rising air rather than turbulent mixing accounted for radon in the stratosphere. The results while interesting were not deﬁnitive. A second series of collections were planned for Spring 1962. ANL withdrew from the program and the USWB approached the US Atomic Energy’s Health and Safety Laboratory (HASL) (now the US Department of Energy Environmental Measurements Laboratory) to provide the measurement capability. The collection and measurement campaign of tropospheric and stratospheric air was completed successfully but was not reported in the open literature due to the press of concerns about fallout radionuclides from atmospheric testing. We have taken this opportunity to present our data to the scientiﬁc community for the ﬁrst time.

2. Materials and methods The ﬂights over Alaska (70◦ N), the Southwest USA (Texas, 32◦ N) and the Panama Canal Zone (8◦ N) were carried out over a 15 day period in April 1962. The air collections were made from the USWB platform, the WB-57. At predetermined locations and altitudes, air was compressed into steel tanks for shipment to the measurement laboratory. Each tank contained approximately 2 m3 of air at standard temperature and pressure. The locations and trajectories of the ﬂights were selected in order to investigate the inﬂuences of tropospheric height, the underlying land mass and thermal gradients on radon concentrations in the atmosphere. The steel tanks were shipped to HASL and radon measurements were performed in 4 days of collection. Prior to the ﬂights many tests were conducted at HASL to improve the radon concentration system, principally by reducing the blank associated with the radon concentration train. The materials of construction, the puriﬁcation train reagents, the charcoal itself contributed to the blank because of the inherent 226 Ra in the materials. At that time commercial absorbents were either coconut charcoal or molecular sieves, both of which showed high and variable (depending on lot) blank values from 226 Ra. Activated charcoal was prepared in-house from peach pits to produce a low level absorbent. The activated peach pits were crushed to a particle size of 3.4 × 4.5 mm (6 × 8 mesh) and 50 to 70 g loaded into copper tubes (1.9 cm inside diameter × 35 cm long) while ﬁt a commercially available tube furnace. The charcoal tubes

Stratospheric radon measurements in three North American locations

717

were ﬁtted with a pressure-vacuum valve at each end. The charcoal trap was immersed in dry ice prior to and during test and actual sample collections. The puriﬁcation train had two Milligan gas washers with a capacity of 270 mL each. The gas entering a washer traversed a spiral path of 90 cm through an absorbent liquid. Two gas washers, in series, were used each containing 100 mL of an absorbent solution. The ﬁrst contained barium hydroxide (50 g Ba(OH)2 ·H2 O L−1 ) for the removal of carbon dioxide gas; the second contained concentrated sulfuric acid for water vapor removal. The assembled train was pressurized with aged air and leak tested under water before use. A fresh puriﬁcation train was prepared and leak tested with each sample. The complete system, that is, the puriﬁcation train and the charcoal trap, were tested for absorption efﬁciency. Aged air was passed through a USDOC National Bureau of Standards (NBS) Standard Reference Material 226 Ra solution, the puriﬁcation train, and the charcoal trap to a wet test meter, which measured the air volume. The ﬂow rate through the system was limited to 6 to 7 L min−1 by the puriﬁcation train. Several trial runs were conducted with air volumes ranging from 0.5 to 2 m−3 at STP. The trial runs indicated 100% collection efﬁciency for the charcoal adsorption with the full puriﬁcation train and over 90% with no carbon dioxide removal. After concentration of the radon on the charcoal trap, the trap was warmed to room temperature, purged with forming gas (85% N2 , 15% H2 ) and placed in a 350 ◦ C tube furnace for one hour. The desorbed radon was slowly ﬂushed with forming gas into a HASL 2L fast pulse ionization chamber (PIC) to an overpressure of 0.24 kPa and the chamber sealed for the measurement period. The efﬁciency of the transfer from the charcoal trap to a single PIC was tested by ﬂushed forming gas through the charcoal trap into a second PIC and measuring for the same period of time as the sample. There were not counts greater than background in the second PIC indicating complete transfer of the sample to the ﬁrst chamber. The HASL PIC calibration and maintenance procedures have been described elsewhere [5,6]. In 1962 there were eight PICs in service. The calibration factor derived from multiple measurements of NBS SRM 226 Ra solutions was 6.35 counts h−1 mBq−1 (225 counts h−1 pCi−1 ) for radon in equilibrium with its short-lived progeny. Statistical testing showed no signiﬁcant difference in the calibration factor among the eight PICs. Six PICs have average background count rates of 6 to 10 counts h−1 and two had background rates of 12 and 18 counts h−1 . For the actual sample measurements, the latter two chambers were used only for the lower altitude collections. In the laboratory, the blanks for the puriﬁcation train and the charcoal trap were minimized by deemanating them immediately prior to a sample run. The radon from the charcoal trap was deemanated by heated in the tube furnace while ﬂowing forming gas through the trap. The puriﬁcation train was deemanated by ﬂowing 20 L of the air sample to be measured through the train and the rubber tubing connections. The rubber tubing had been carefully washed and dried to remove talc, a radium bearing mineral, from the inner surface. The blanks from the puriﬁcation train and the charcoal trap were dependent on the time period required for sample absorption. A 1 m3 air sample was collected over a 2.5 hour period and yielded a total blank of 0.7 to 1.1 mBq (4, 6 or 7 counts h−1 in a PIC), depending on the particular charcoal trap used for the absorption. Thus the overall blank, PIC background count rate, puriﬁcation train and charcoal trap blanks, ranged from 10 to 25 counts h−1 or 1.6 to 4.1 mBq. For the highest altitude air collections careful attention was paid in the selection of the charcoal trap and the

718

I.M. Fisenne et al.

PIC used for the radon measurement. The lower limit of detection [7] at the 95% conﬁdence level could range from 1.6 to 12.4 mBq based on the selected charcoal trap and PIC. The measurement error for the tropospheric samples was in general about 10% and up to 100% for the stratospheric collections where the activity of some samples was below the detection limit. Four of the Alaskan samples, collected at altitudes of 15 to 19 km, and three Southwest USA samples, collected at altitudes between 4.6 and 15 km, were analyzed in duplicate. The duplicates agreed within the error of the measurements.

3. Results and discussion In all 54 samples were measured for radon in April 1962 and only those results greater than the detection limit are included in Figs. 1–3. An average global radon concentration at the earth’s surface of 15 000 mBq m−3 was adopted from the NAS/NRC report summarizing outdoor radon concentrations [8]. All the ﬂight proﬁles at a given location are shown in a single ﬁgure because a breakout by day revealed no particular pattern other than the expected decreasing radon concentration with increasing height above the surface. If the atmospheric concentration of radon was uniform at the surface, one would expect the highest radon concentrations in the Alaska samples and the lowest in the Canal Zone samples because of the relative heights of the tropopause. Likewise one would expect the highest radon concentrations in the Southwest USA samples because of the source term over land. The lowest radon concentration would be expected in the Canal Zone samples because of the decay of radon over the oceans and the tropopause height. These assumptions are not borne out by the measurements. The highest concentrations of radon in the troposphere were measured in the Canal Zone samples. These data are supported by similar tropospheric concentrations reported by Machta and Lucas [4] for air samples collected over Hawaii. The tropospheric radon concentrations in the Southwest USA samples are similar to the Canal Zone proﬁles and are supported by the later measurements of Moore et al. [9]. The lowest tropospheric radon concentrations measured were from Alaska, again supported by the data of Machta and Lucas [4].

Fig. 1. 222 Rn measurements over the Canal Zone.

Stratospheric radon measurements in three North American locations

719

Fig. 2. 222 Rn measurements over Southwest United States.

Fig. 3. 222 Rn measurements over Alaska.

The stratospheric proﬁles reﬂect the expected relationship of geography and source strength. The lowest radon concentrations were measured in the Canal Zone samples and the highest in the Southwest USA samples. The stratospheric radon concentrations measured by Machta and Lucas [4] for samples collected above Hawaii are signiﬁcantly higher than those collected above the Canal Zone. This is also the case for the Alaskan series with signiﬁcantly higher radon concentrations reported by Machta and Lucas [4]. These differences cannot be attributed to seasonality as both series of air collections occurred in the Spring. The measurement results for the three sampling locations are shown in Figs. 1–3 and summarized in Fig. 4. While the radon measurements from the three sampling locations have some differences, their commonality is the decreasing concentration with increasing height. Grouping the data, the mean radon concentration, for each altitude from the three locations into one chart, Fig. 5, clearly shows that an underlying process controls radon in the atmosphere. This is also apparent as the grouped data points might well reﬂect the range expected for a single sampling location over a period of time. Figure 5 reinforces the exponential nature of the declining concentration of radon with height above the earth’s surface. The decrease of a factor of ten in radon concentration from 7.6 to 12.2 km suggests the inﬂuence of the tropopause as a barrier to transport of material from the troposphere to the stratosphere.

720

I.M. Fisenne et al.

Fig. 4. Upper air 222 Rn measurements from the three locations.

Fig. 5. The average 222 Rn concentrations in the troposphere and stratosphere for the three sampling locations.

The effective vertical transport of 222 Rn from the earth’s surface to the various sampling altitudes can be calculated from the mean radon concentrations in Fig. 5, assuming a surface radon concentration of 15 000 mBq m3 . These estimates are shown in Table 1. As would be expected, the “effective” vertical transport of radon increases with increasing height since the density of the atmosphere decreases with increasing height. Although this is an oversimpliﬁcation of the many complex processes occurring in the atmosphere, chemical reactivity to form larger species, as with many aerosols, is eliminated in the case of a noble gas. We are left with two removal processes, radioactive decay and vertical transport. The estimated number of radon half-lives from decay alone explain the measurement difﬁculties encountered in this program. If the mean values in Fig. 5 are used to separate the atmosphere into two compartments, the troposphere at 7.6 km and the stratosphere at 12.2 km, an effective vertical transport may be calculated. Using the mean radon concentration at 12.2 km, the effective vertical transport is shown in Table 2. Again it would be expected that the transport increase with increasing altitude, reducing the number of radon half-lives from the base altitude of 12.2 km.

Stratospheric radon measurements in three North American locations

721

Table 1 Effective vertical transport of 222 Rn in the atmosphere Altitude (km)

222 Rn decay (days)

222 Rn half-lives

Effective vertical transport (cm s−1 )

4.6 7.6 12.2 15.2 18.3 19.8

19.1 19.5 31.8 30.4 40.7 40.8

5.0 5.1 8.3 8.0 10.6 10.7

0.28 0.45 0.44 0.58 0.52 0.56

Table 2 Effective vertical transport of 222 Rn in the stratosphere Altitude (km)

222 Rn decay (days)

222 Rn half-lives

Effective vertical transport (cm s−1 )

18.3 19.8

8.8 9.0

2.3 2.4

0.80 0.98

The 1962 proﬁles collected for radon measurements tend to point to a more homogeneous troposphere than previously thought, a demonstrable decrease in radon concentration with height above the Earth’s surface, increasing rate of vertical transport with increasing height and the need for meticulous measurement quality control and quality assurance.

References [1] [2] [3] [4] [5] [6] [7] [8]

A. Wigand, F. Wenk, Ann. Phys. 86 (1928) 657. M.H. Wilkening, Trans. Am. Geophys. Union 37 (1956) 177. H.A. Miranda Jr., J. Atmos. Terr. Phys. 11 (1957) 272. L. Machta, H.F. Lucas Jr., Science 135 (1962) 296. I.M. Fisenne, H.W. Keller, US Dept. of Energy report EML-437, 1985. I.M. Fisenne, A.C. George, H.W. Keller, J. Res. NIST 95 (1990) 127. B.S. Pasternack, N.H. Harley, Nucl. Instrum. Methods 91 (1971) 533. Committee on Risk Assessment of Exposure to Radon in Drinking Water, Board on Radiation Effects Research, Commission on Life Sciences, National Research Council, Risk Assessment of Radon in Drinking Water, National Academy Press, Washington, DC, 1999. [9] H.E. Moore, S.E. Poet, E.A. Martell, J. Geophys. Res. 78 (1973) 7065.

722

Determination of soil-gas radon concentration in low permeability soils ∗ Martin Neznal, Matej Neznal RADON, v.o.s. corp., Novakovych 6, 180 00 Praha 8, Czech Republic

The collection of soil-gas samples in low permeability soils is often complicated due to a high resistance that must be overcome. An enlargement of the cavity, from which the samples are taken, is then required, but this action may affect the measurement result. A relationship between the measured soil-gas radon concentration and the changing sampling geometry was studied in four reference areas characterized by low permeable soils and/or by a high soil moisture. The results of the survey indicate that measured soil-gas radon concentrations do not depend on the changing sampling geometry if the vertical proﬁle of the uppermost soil layers is homogeneous. A decrease of soil-gas radon concentration with increasing dimensions of the cavity was observed when the soil permeability was higher at shallow depths.

1. Introduction A uniform method for the assessment of the radon potential of soils (the so-called Radon Risk Classiﬁcation of Foundation Soils) has been used in the Czech Republic since 1990. The classiﬁcation is based on the determination of soil-gas radon concentration (222 Rn) and on the permeability classiﬁcation of foundation soils. As for the soil-gas radon concentration measurements, only one sampling and measuring method is commonly used [1]. The equipment for soil-gas sample collection consists of a small-diameter hollow steel probe with a free, sharpened lower end (sharp tip). The probes are pounded into the soil to a depth of 0.8 m. A punch wire is then inserted into the probe. The active area is created in the head of the probe by the extrusion of the tip by means of the punch wire to a distance of several centimeters. Samples of soil-gas are collected using a syringe and introduced into previously evacuated Lucas cells. A similar sampling technique was described by Reimer [2]. The system can be used for soil-gas sampling even in soils with a low permeability. Measurements of soil-gas radon concentration, carried out in the last ten years, indicate a dependence of the determined soil-gas radon concentrations on the depth of sampling, on the * The survey was supported by the Czech Ofﬁce for Nuclear Safety, Praha.

RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07089-5

© 2005 Elsevier Ltd. All rights reserved.

Determination of soil-gas radon concentration in low permeability soils

723

soil permeability, on the dimensions of the cavity from which the soil-gas samples are taken and on the soil-gas sampling technique used [1,3,4]. Low permeability soils, often preventing soil-gas sampling in a standard probe geometry, require the enlargement of the cavity (i.e. the enlargement of the active area from which the soil-gas sample is taken) in the soil.

2. Methods, description of reference areas The internal diameter of the probes that are used for sampling is 8 mm, their external diameter being 12 mm. A typical height of the active area at the lower end of the probe ranges from 2 to 5 cm. In low permeable soil, it is possible to increase the height of the cavity by retracting the probe back towards the soil surface. The relationship between the soil-gas radon concentration and the changing sampling geometry was studied at four reference areas characterized by low permeable soils and/or by a high soil moisture. In each reference area, the measurements were made at nine measuring points. At each measuring point, the soil-gas samples were collected from different sampling depths and using different dimensions of the active area. The four reference areas were called by the names of the nearest villages. A brief description of the geological conditions in the areas is summarized in Table 1. Table 1 Geological conditions Area

Geological conditions

Svetice

Bedrock formed by Ordovician shales (irregular variation of sandy, silty and ﬂinty shale). Quarternary cover represented by eolian to eolic-deluvial sediments (loess and loess loam with a thickness of about 1.5 m). Surface layer made by organic rich clayey loam with thickness of about 0.25 m. Bedrock formed by Cretaceous (Upper Turonian) claystones and marlites. Quarternary cover represented by deluvial and ﬂuvial sediments (sands and clayey sands with thickness more than 1.5 m). Surface layer made by organic rich sandy loam with thickness of about 0.15 m. Bedrock formed by Ordovician clayey shales. Quarternary cover represented by eolic-deluvial sediments (clayey loam and clay) of thickness more than 1.5 m. Surface layer made by organic rich clayey loam with thickness of about 0.50 m. Bedrock formed by porphyric, amphibol-biotitic granodiorite (late Variscan Central Bohemian Pluton, part of the Bohemian Massif). Granodiorite weathering residuum covered by ﬂuvial sediments (sandy clay, clay) with thickness of about 1.0 m. Surface layer made by organic rich loam with thickness of about 0.20 m.

Dubnice

Ptice

Ruzena

The following seven cases were tested: – sampling depth 60–62 cm, height of cavity 2 cm, indicated as “geometry 60 (2) cm”; – sampling depth 80–82 cm, height of cavity 2 cm, indicated as “geometry 80 (2) cm”; – sampling depth 80–85 cm, height of cavity 5 cm, indicated as “geometry 80 (5) cm”; – sampling depth 80–90 cm, height of cavity 10 cm, indicated as “geometry 80 (10) cm”; – sampling depth 70–90 cm, height of cavity 20 cm (the probe was retracted back to the surface), indicated as “geometry 70–90 cm”; – sampling depth 60–90 cm, height of cavity 30 cm, indicated as “geometry 60–90 cm”; – sampling depth 40–90 cm, height of cavity 50 cm, indicated as “geometry 40–90 cm”.

724

M. Neznal, M. Neznal

All ﬁeld measurements were made in 2000, on September 20th in the reference area Svetice, September, 21st at reference area Dubnice, October 26th at reference area Ptice, and November 1st in reference area Ruzena. To avoid any inﬂuence of local inhomogeneities, average soil-gas radon concentrations over the whole reference area not single values were compared.

3. Results As can be seen in Table 2, the results obtained at reference areas Svetice, Dubnice and Ruzena were similar. Values of soil-gas radon concentration were almost the same when “geometry 80 (2) cm”, “geometry 80 (5) cm”, or “geometry 80 (10) cm” was used. Soil-gas radon concentrations in the samples collected using “geometry 70–90 cm” were a little lower but comparable with the previous ones. The results obtained using “geometry 60–90 cm” were lower and similar to the results that were observed when “geometry 60 (2) cm” was used. One example (reference Area Svetice) is presented in Fig. 1. On the other hand, almost no dependence of soil-gas radon concentration on the changing dimensions of the active area was observed at the fourth reference area Ptice, characterized by Table 2 Relative comparison of average soil-gas radon concentrations obtained using different sampling geometry Geometry

Svetice

Dubnice

Ptice

Ruzena

60 (2) cm 80 (2) cm 80 (5) cm 80 (10) cm 70–90 cm 60–90 cm 40–90 cm

0.79 0.99 1.00 1.13 0.82 0.66 0.55

0.82 1.01 1.00 1.02 0.93 0.62 0.39

0.69 0.96 1.00 1.08 1.13 1.13 1.08

0.59 0.90 1.00 0.94 0.95 0.77 0.67

Fig. 1. Reference area Svetice: Soil-gas radon concentration (arithmetic mean ± standard deviation) vs. sampling geometry.

Determination of soil-gas radon concentration in low permeability soils

725

Fig. 2. Reference area Ptice: Soil-gas radon concentration (arithmetic mean ± standard deviation) vs. sampling geometry.

a homogeneous vertical soil proﬁle and by an extremely low soil permeability. The sampling had to be made very slowly and carefully at this reference area. The collection of a 100 ml soil-gas sample very often took more than 10 minutes. As can be seen in Fig. 2, there was no decrease of soil-gas radon concentration, even when the height of the active area was increased to 50 cm (“geometry 40–90 cm”).

4. Conclusions It is possible to conclude that measured soil-gas radon concentrations do not depend on changing sampling geometry if the soil layer is homogeneous and of low permeability. A decrease of soil-gas radon concentration with increasing dimensions of the active area (i.e. using “geometry 70–90 cm”, “geometry 60–90 cm”, or “geometry 40–90 cm”) indicates that the vertical soil proﬁle is not homogeneous and that the soil permeability is higher in shallow depths. In this case, the upper part of the cavity reaches the more permeable soil layers with lower radon concentration. Due to a lower resistance, the soil-gas sample is collected mostly from the upper part of the active area. A good knowledge of the vertical soil proﬁle is thus very important for correct interpretation of measured soil-gas radon concentrations. A perfect sealing of all parts of the equipment is required when soil-gas samples are collected in low permeability soils.

References [1] M. Neznal, M. Neznal, J. Smarda, Environ. Int. 22 (1996) S819. [2] G.M. Reimer, Geophys. Res. Lett. 17 (1990) 809. [3] I. Barnet, M. Neznal (Eds.), Radon Investigations in the Czech Republic V, Czech Geological Survey and Radon corp., Praha, 1994. [4] A.B. Tanner, Nucl. Geophys. 5 (1991) 25.

726

Indoor radon long-term variation assessment I.V. Yarmoshenko a , Z.S. Zunic b , J.P. McLaughlin c , J. Paridaens d , I.A. Kirdin a , K. Kelleher c a Institute of Industrial Ecology, 20A Sophy Kovalevsky Str., GSP-594, 620219 Ekaterinburg, Russia b Institute of Nuclear Sciences “Vinca”, PO Box, 522, 11000 Belgrade, Yugoslavia c University College Dublin, Belﬁeld, Dublin 4, Ireland d Belgian Nuclear Research Centre, SCK/CEN, Boeretang 200, B2400 Mol, Belgium

Investigation of indoor radon long-term variation was conducted basing both on the simulation of the process and results of ﬁeld measurements of indoor radon concentration obtained using retrospective and contemporary techniques. The coefﬁcient of long-term variation (CLTV) of indoor radon is introduced which relates the annual indoor radon concentration with annual radon concentration of the previous or next year. Actual values of the CLTV were estimated by the results of ﬁeld measurements in rural regions of Yugoslavia and Russia. Investigation showed that two main factors govern the process – long-term change of the air exchange rate in living space and the relationship between the contributions of convective and diffusive radon entries to the indoor radon concentration. Recognized and tested parameters of long-term indoor variation can be used for reconstruction of the indoor radon history of houses.

1. Introduction Changes and transformations of construction components (such as basement, construction joints, insulation, interface gaps etc.), underlying soil physical condition and occupants’ living habits may in the course of time result in signiﬁcant changes of the radon entry routes and driving force characteristics and consequently indoor radon entry. A number of tasks in the ﬁeld of radon research require, in particular in residential radon epidemiology, understanding and consideration of indoor radon long-term variations. Retrospective techniques of indoor radon measurements with volume and surface 210 Po traps [1] allow the process to be investigated. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07090-1

© 2005 Elsevier Ltd. All rights reserved.

Indoor radon long-term variation assessment

727

2. Indoor radon long-term variation model To simulate long-term indoor radon variation the model of the process was developed: Ci = Ui + Di + Ai ,

(1)

where Ci is the annual indoor radon concentration in year i, Ui is the contribution of convective entry to indoor radon concentration as an average in year i, Di is the contribution of diffusive entry, Ai is the contribution of outdoor radon. Speciﬁc sub-models of radon entry by different mechanisms were developed as well. The models are quite complicated and include as many as 30 parameters. The set of parameters was reviewed to identify those that might be expected to change in the long-term. The seven parameters considered to be of most importance in this context are now listed: – – – – – – –

moisture content of underlying soil, effective diffusion coefﬁcient of radon in soil, convective transfer factor between soil and basement space, convective transfer factor between basement space and living spaces, air exchange rate in living spaces, air exchange rate in basement, coating thickness of interior wall surfaces.

By the results of simulation the character of the indoor radon long-term variation is close to linear and can be described as the variation of annual radon concentrations on a year-byyear basis. To describe the long-term variation in that way, a coefﬁcient of long-term variation (CLTV, k) is introduced, which relates the annual radon concentration of year i with the annual indoor radon concentration of the previous or next year: Ci−1 = Ci · k.

(2)

To simplify the model, the coefﬁcient k is considered as a constant characteristic of a building occupied by a family. Then if CC is the contemporary indoor radon concentration and Ci is the annual radon concentration i years ago: Ci = CC · k i−1 .

(3)

The CLTV reﬂects the monotonous pattern of long-term indoor radon variation and it is possible to estimate its characteristic value from the results of contemporary and retrospective indoor radon measurements. Such a possibility arises after taking into consideration the age of the volume or surface trap used to measure the retrospective radon concentration and the number of intervals of change (A). If the retrospective indoor radon concentration CR is the average of annual radon concentrations within a period equal to the age of the “trap”, then 1 CC · k i−1 . A A

CR =

(4)

i=1

The solution of the previous equation is CR =

CC k A − 1 · . A k−1

(5)

728

I.V. Yarmoshenko et al.

Fig. 1. Results of CLTV simulation in terms of its dependence on the U/D ratio and the pattern of air exchange rate long-term change; 1: case of air exchange rate long-term increasing, 2: case of air exchange rate long-term decreasing.

Equation (5) cannot be solved symbolically to ﬁnd the functional dependence of k on CC , CR and A; it only allows the speciﬁc solution for k using the observed values of CC , CR and A. Additionally, the model investigation of long-term indoor radon variation has led to the conclusion that two main factors govern the process – long-term change of air exchange rate in indoor living spaces and the relationship between the contributions of convective and diffusive radon entries to indoor radon concentration (U/D). Primary results of the simulation are presented in Fig. 1, which demonstrates the dependence of CLTV on the U/D ratio taking to account two variants of air exchange rate long-term change.

3. Experimental assessment of indoor radon long-term variation Experimental investigation were conducted using data of retrospective measurements (volume and surface traps) obtained in radon surveys performed in typical rural houses of the Gornja Stubla (Kosovo), Kalna and Uzice (Serbia) regions of Yugoslavia [2,3] and Sysert region of the Urals in Russia. The values of the CLTV have been calculated using CC , CR and A values experimentally determined for each house and using Equation (5). Some sets of those values did not return a real positive value of CLTV. Observed CLTV values follow the normal distribution considering the sample as whole and are close to normal in each region. Figure 2 presents the frequency distribution of CLTV in the combined sample and the associated curve of the normal distribution. Average results of the estimation of CLTV for volume (kVT ) and surface (kST ) traps are presented in Table 1. Average values estimated under condition A > 10 are separately presented in Table 1 as well. While the analysis of variances was statistically signiﬁcant, it can be concluded that the mean CLTV across the groups by region are different in magnitude. Figure 3 represents the Box-and-Whisker plot demonstrating the difference of regional mean CLTV which are plotted versus regional mean values of contemporary indoor radon concentration with standard error of estimation and 95% conﬁdence interval.

Indoor radon long-term variation assessment

729

Fig. 2. Frequency distribution of CLTV values. Solid line is ﬁtted normal distribution.

Table 1 Average values of k estimated for houses in four regions Region

Mean A (y)

Mean CR (Bq m−3 )

Mean CC (Bq m−3 )

Mean kST

Mean kVT

Mean kST (A > 10 y)

Mean kVT (A > 10 y)

Gornja Stubla Kalna Uzice Sysert

17 41 17 27

328 128 139 161

756 194 131 88

0.87 0.96 1.03 1.06

1.10 1.03 1.07

0.86 0.96 1.01 1.05

0.95 1.02 1.04

Fig. 3. Regional mean CLTV characteristics vs. contemporary indoor radon concentration.

4. Discussion Simulation and experimental investigation of indoor radon long-term variation show the usefulness of the long-term variation coefﬁcient. The CLTV is considered as a characteristic of the house, which includes both the construction features and living habits of the occupants. By

730

I.V. Yarmoshenko et al.

the results of simulation, the home-speciﬁc CLTV appears to have an approximately constant value. While the retrospective indoor radon measurement techniques are relatively new, there still may be some unrecognized sources of errors, which inﬂuenced the results of indoor radon long-term variation assessment. It is reasonable to suspect that the obtained tendency of regional mean CLTV values to correlate with regional mean indoor radon concentration (Fig. 3) is evidence of such a systematic error. However, such a correlation may reﬂect a real relationship between radon entry and long-term variation processes. Assuming that increasing levels of indoor radon concentration appear to be due to an increasing contribution of convective radon entry, it is valid to suppose that there exists a correlation between the indoor radon concentration and the ratio between the contributions of convective and diffusive radon entries (U/D). Consequently the experimental data presented in Fig. 3 demonstrate the same relationship as is obtained after simulation and is presented in Fig. 1. In addition. the statistical signiﬁcance of analysis of variances supports the viewpoint that the obtained differences between regional mean values of CLTV could not result from the uncertainties in the actual retrospective and contemporary ﬁeld measurements. The obvious and principal difference between k values in subsamples of volume and surface traps (kST < 1 while kVT > 1) can be seen from Table 1. Besides that discrepancy, the obtained average values of k parameters allows the above assumptions to be made on the pattern of indoor radon long-term changes. Recognized parameters of long-term indoor variation can be used for reconstruction of house speciﬁc indoor radon histories. The estimation of CLTV allows us to obtain not only the integral characteristic of indoor radon exposure within the time period equal to “age of trap” but also a year-by-year assessment of annual indoor radon concentration both for years within that period and earlier years as well.

References [1] J.P. McLaughlin, Approaches to the assessment of long term exposure to radon and its progeny, Sci. Total Environ. 272 (2001) 53–60. [2] Z.S. Zunic, J.P. McLaughlin, K. Fujimoto, et al., Field work studies on natural indoor radiation population exposures in Yugoslavia – new results, Proceedings of 3rd International Yugoslav Nuclear Society Conference (YUNSC-2000), Belgrade, Yugoslavia, October 2–5, 2000, Vinca Bull. 6 (1) (2001) 801–806. [3] Z.S. Zunic, J.P. McLaughlin, C. Walsh, et al., The use of SSNTD in the retrospective assessment of radon exposure in high radon rural communities in Yugoslavia, Radiat. Measur. 31 (1999) 343–346.

731

Quality assurance of individual radon measurements T.R. Beck, J. Schwedt, P. Hamel Federal Ofﬁce for Radiation Protection (BfS), Köpenicker Allee 120–130, D-10318 Berlin, Germany

The 96/29/EURATOM guideline and its national transformations demand the application of radiation protection measures if the presence of radon and radon decay products leads to a signiﬁcant increase in exposure of workers. The requirements of the aforementioned EURATOM guideline were met by the amendment to the German Radiation Protection Ordinance in 2001. Presently, a national guideline for the transformation of the legal rules into practice is being implemented to guarantee an effective but appropriate protection against natural radiation. The guideline regulates the responsibilities of the authorities as well as the setting-up of appropriate means and measures to be taken by the employers. Additionally, it will demand from the laboratories carrying out measurement tasks a program for quality assurance and the participation in measurement inter-comparisons. For the evaluation of laboratory inter-comparisons and for the authorization of personnel radon monitors, uniform accuracy criteria have been developed. Around the relevant limit of the measured potential alpha-energy exposure of 14 mJ h m−3 , an over- and under-estimation of the true value by a factor of 1.8 is accepted. Monitors for exposure to radon can be applied at workplaces with equilibrium factors within a range from 0.2 up to 0.7. Measurement uncertainty around the relevant limit of 6000 kBq h m−3 is ±20%.

1. Introduction In 2001 the amendment to the German Radiation Protection Ordinance was published. As a consequence of the transformation of the 96/29/EURATOM guideline, the Radiation Protection Ordinance lays down speciﬁc rules for the protection of workers against exposures due to natural radiation sources for the ﬁrst time. Work ﬁelds where radon and short-lived radon decay products can lead to a signiﬁcant increase in exposure of workers were identiﬁed in the decommissioning of the former uranium industries, in non-uranium mines, in the mining assurance industries, in exhibition caves and mines, in water-supply stations as well as in radon spas. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07091-3

© 2005 Elsevier Ltd. All rights reserved.

732

T.R. Beck et al.

If signiﬁcant exposure of workers exists, appropriate measures to reduce exposure have to be taken by the responsible employers. Personnel monitoring of the employees regarding this radiation exposure could be introduced as well. An important task of the Federal Ofﬁce for Radiation Protection (BfS) is presently to transform the legal requirements into practice by implementation of a national guideline which regulates the responsibilities of the authorities as well as the setting-up of appropriate means and measures to be taken by the employers. Additionally, the guideline will demand from the laboratories carrying out quality assurance measures and the participation in measurement inter-comparisons. This paper describes the current status of the transformation into practice of legal requirements regarding the protection of workers against exposures to radon and radon decay products. It will develop uniform accuracy criteria for personnel radon monitors independent of their physical principle. Results of ﬁrst inter-comparisons of all personnel radon monitors applied in Germany are presented. 2. Protection against natural radiation at workplaces 2.1. Concept and responsibilities According to the new Radiation Protection Ordinance, it is the duty of the employer to carry out an estimation of radon exposure to be expected at his workplaces. Workplace means here work to be performed by one person, even if this work is carried out at different places. The new Radiation Protection Ordinance provides for a step-wise concept for the protection of employees against enhanced radiation exposures due to the inhaling of radon-containing air. In the ﬁrst step, employers have to estimate possible radon exposure at their workplaces. When radon concentrations in the companies cannot be prognosticated, measurements are required. When these estimations result in the possible existence of enhanced radiation exposure, it has to be checked whether remediation measures can possibly be taken in a second step. The involvement of the employees in radiation protection monitoring depends on the success of these measures. In most cases, radon exposure is measured with personnel radon monitors, which have to be worn by the employees outside on the trunk (preferably near the breathing zone) over a certain period of time at work. The employer may give an order to a measuring laboratory to carry out the measurements. The number of radon monitors to be requested is equal to the number of different workplaces plus one reference monitor (“blank value”). That means that for several persons with the same tasks and the same work proﬁle radon exposure may be determined by one person only with the help of a radon monitor. On request of the employers, the monitors are delivered by the measuring laboratories and are evaluated there, too. The employer then gets a test report showing the respective radon exposure. He can use this test report as proof of the fulﬁllment of his duties towards the regulatory authority. 2.2. Interventions If the determined radon exposure is so low that the intervention level of 2000 kBq h m−3 (approximately 6 mSv) is fallen below during the calendar year, the employer does generally

Quality assurance of individual radon measurements

733

not have to take further steps if it is not possible to decrease radon exposure further by taking appropriate measures (requirement of minimization). He has, however, to perform this estimation again if a workplace is essentially modiﬁed which could lead to enhanced radon exposure of the employees. Besides structural changes, essential modiﬁcations can also mean organizational modiﬁcations leading, for example, to the employees’ modiﬁed duration of stay at speciﬁc places. If the result of this estimation is that the intervention level of 2000 kBq h m−3 during the calendar year is exceeded, the employer has to perform remediation measures to decrease radon exposure. Possible measures include, for example, improved ventilation of the affected rooms or sealing of the radon sources. Measurements within the scope of remediation measures do not refer to persons but to rooms. Since, for example, in water-supply stations in most cases only some rooms contribute signiﬁcantly to radon exposure of the employees, these rooms have at ﬁrst to be localized with the help of radon concentration measurements. To elaborate a remediation concept, a private or ofﬁcial measuring institute should be contacted. The success of these remediation measures has then to be documented with further radon exposure measurements. Another possible method to decrease exposure is to reduce the employees’ duration of stay in rooms with enhanced radon concentrations. It will, however, possibly be difﬁcult to monitor this organizational measure. These examples show that if the intervention level of 2000 kBq h m−3 during the calendar year is exceeded, remediation measures are the most important method, since they make further radiation protection monitoring of the personnel unnecessary and simultaneously decrease the risk of health damages for the employees. The employer must guarantee through further regular personal measurements that none of his employees exceeds the radon exposure limit of 6000 kBq h m−3 during the calendar year. Furthermore, these employees have to be examined by a physician in future once per calendar year. Additionally, the employer has, among others, to report regularly the measured radon exposure values of his employees to the responsible authority, so that their names can be entered into the Radiation Protection Register. Measurements for the radiation protection monitoring of the personnel are continuous personal measurements with radon monitors. Corresponding order to perform these measurements has to be given to a measuring laboratory chosen by the authority.

3. Accuracy criteria for individual radon monitors 3.1. Accuracy criteria for occupational radiation monitoring of external radiation Accuracy criteria for occupational radiation monitoring were published by ICRP and IAEA [10,11]. Radiation exposure by radon and radon decay products results from their intake due to inhalation. The measurement principle using a passive personnel detector worn outside on the trunk is, however, similar to the personnel dosimetry of external radiation. For this reason, the personnel radon measurements were treated in the same way. Information concerning the uncertainties that can be expected in making measurements with individual dosimeters at the workplaces is given in ICRP Publication No. 75 [11], which states that:

734

T.R. Beck et al. “The overall uncertainty at the 95% conﬁdence level in the estimation of effective dose around the relevant dose limit may well be a factor of 1.5 in either direction for photons and may be substantially greater for neutrons of uncertain energy and for electrons. Greater uncertainties are also inevitable at low levels of effective dose for all qualities of radiation.”

3.2. Accuracy criteria for individual monitors measuring potential alpha-energy exposure To adopt the aforementioned principles in the ﬁeld of individual measurements of potential alpha-energy exposure, the following is assumed: • In the region near the relevant limit of potential alpha-energy exposure, Pp , a factor k in either direction should be considered acceptable for the overall uncertainty from radon exposure and the equilibrium factor at a 95% conﬁdence interval. • In the region of the recording level, an acceptable uncertainty of ±100% is implied. The introduction of a factor for limitation of accuracy intervals results in asymmetric interval boundaries, from which it follows that an underestimation of the true dose is more restricted than an overestimation. According to ICRP [12], the PP limit of 14 mJ h m−3 corresponds to an annual dose limit of 20 mSv. This represents a radon exposure limit of 6000 kBq h m−3 using an equilibrium factor of 0.4. The recording level PP,0 for radon exposure measurements should be orientated to the natural radon background of 40 Bq m−3 [13]. For a monitoring period assumed to be 3 months, a recording level for the radon exposure PRn,0 of 90 kBq h m−3 results. The corresponding recording level for the potential alpha-energy exposure PP,0 is 0.2 mJ h m−3 . According to IAEA [10], the allowable accuracy interval can be smoothed as a function of the potential alpha-energy exposure level. The upper limit RUL is given by Pp,0 . RUL = k 1 + (1) 2Pp,0 + Pp where Pp is the conventional true potential alpha-energy exposure. The lower limit RLL is given by ⎧ for Pp < Pp,0 , ⎨0 2P RLL = 1 (2) p,0 for Pp Pp,0 . 1− ⎩ k Pp,0 + Pp It should be noted that changes in the recording level would inﬂuence the accuracy interval in the low exposure region. Equations (1) and (2) perform the so-called Trumpet Curve. The factor k was determined from an analysis of the uncertainties and from a practical point of view. It means that the performance of measurement devices available on the market and applied by laboratories with a good laboratory practice can meet the accuracy criteria. The same aspects as mentioned before must be taken into account for measurement devices for radon exposure. In this case the accuracy interval for potential alpha-energy exposure have to be large enough to comprise the uncertainty of the equilibrium factor and the uncertainty of the radon exposure measurements. From the above, the following accuracy criteria for the determination of the potential alpha-energy exposure of short-lived radon decay products are proposed:

Quality assurance of individual radon measurements

735

• In the region near the relevant limit of Pp a factor k = 1.8 in either direction should be considered acceptable for the overall uncertainty from radon exposure and the equilibrium factor at a 95% conﬁdence interval. • For lower Pp the determined value should overestimate the true value not more than a factor k = 2. • In the region of the recording level, an acceptable uncertainty of ±100% is implied. 3.3. Accuracy criteria for individual monitors measuring radon exposure The regulations for the protection against radon exposures laid down in the German Radiation Protection Ordinance convert the recommendations of the ICRP [12]. The causal relationship between the effective dose to radon exposure is given by the effective dose to potential alpha-energy conversion factor and the further calculation by taking the equilibrium factor into account. The uncertainty in determining the potential alpha-energy and therefore the effective dose results from the uncertainty of the equilibrium factor and the uncertainty of the radon exposure measurement. The uncertainty of the equilibrium factor should satisfy the condition that the allowable variability should be large enough to comprise most of the conditions occurring at workplaces. To establish an appropriate uncertainty interval for the equilibrium factor, the knowledge of its local and temporal variability for many workplaces and, in particular, the knowledge of the individual average equilibrium factor during the work should be known. Because of the poor data, it is easy to understand that an exact derivation is not possible. Many measurements in homes, few in mines and recently a few in water-supply stations were published [1–9]. All measurements, however, have the same problem in common: they are only local measurements for a short and in sufﬁcient time interval. Nevertheless, we have constructed a log-normal distribution for the equilibrium factor with a standard deviation σL,F of 0.295 (see Fig. 1). A log-normal distribution obviously reﬂects the expected equilibrium factor. At a 95% conﬁdence interval the lower and upper limit is 0.22 and 0.71, respectively. These limits provide a good compromise of sufﬁcient variation range containing most of the expected equilibrium factors at workplaces. On condition that the sum of the equilibrium factor variance and the variance of the radon exposure measurement shall be equal to the variance of potential alpha-energy, it is now possible to deduce the accuracy interval which is interesting for radon exposure measurements at a 95% conﬁdence level [14]: P ln(RLL ) 2 ln(RUL ) 2 Rn,measured 2 ln 2 . (3) +1.96 − σL,F − σL,F −1.96 1.96 PRn,true 1.96 PRn,true is the conventional true radon exposure and PRn,measured represents the measured value. Measurements of radon exposures with uncertainties within this interval and of an averaged equilibrium factor over the measurement period, which is within a given log-normal interval around the assumed equilibrium factor of 0.4, satisfy the accuracy requirements of the potential alpha-energy exposure resulting from equations (1) and (2). The accuracy interval for the potential alpha-energy exposure according to equations (1) and (2) and the restricted accuracy interval for radon exposure measurements according to equation (3) are shown in Fig. 2. The latter was calculated on condition that the mean equilibrium factor over the measurement

736

T.R. Beck et al.

Fig. 1. Estimation of a log-normal equilibrium factor distribution and a selection of published values for homes and at workplaces.

Fig. 2. Derived accuracy interval for measurements of potential alpha-energy exposure (solid lines) and the restricted interval for radon exposure measurements (dashed lines).

period is between 0.22 and 0.71 (σL,F = 0.295). Around a radon exposure of approximately 6000 kBq h m−3 , which corresponds to a Pp of 14 mJ h m−3 for an assumed equilibrium factor of 0.4 the uncertainty of the relative exposure at a 95% conﬁdence interval is 0.80

PRn,measured 1.18. PRn,true

(4)

Quality assurance of individual radon measurements

737

For lower radon exposures, the span of the interval expands. The restriction of the upper overall limit on factor 2 causes a restriction of the upper exposure limit. For radon exposures of less than 630 kBq h m−3 , the allowable uncertainty of the relative exposure does not exceed a value of 1.47.

4. Results of inter-laboratory comparisons ALGADE personnel monitors for the measurement of the potential alpha-energy exposure of short-lived radon decay products are regularly calibrated and tested at the BfS radon/radon decay product calibration chamber. Test atmospheres generated in the calibration chamber can be adjusted over a wide range of different climatic, aerosol and activity parameters. ALGADE monitor heads sampling air at a constant air ﬂow rate of about 80 L h−1 were used for the calibrations. Figure 3 shows the results of ALGADE monitor readings in relation to the true values depending on the true potential alpha-energy exposures. All results are within the accuracy range allowed for this kind of individual measurement system, the variations of the aerosol particle concentration and therefore the variations of the unattached fraction especially for the measurements at 6 and 7.5 mJ h m−3 with an unattached fraction of approximately 12 and 50%, respectively, and they do not exceed the accuracy limits. BfS carried out inter-laboratory comparisons for German radon measurement services. Six laboratories with different measurement experience and different passive radon measurement systems participated in the comparisons. The summarized results are shown as box plots in Fig. 4. Most of the results are within the accuracy interval of allowable radon exposure. Any results for the lowest (77 and 90 kBq h m−3 ) and the highest (4500 kBq h m−3 ) exposures, however, do not meet the performance criteria. Deviations at high exposures are possibly caused by incorrect calibrations in this range.

Fig. 3. Results of accuracy tests of ALGADE personnel monitors (boundary of the box indicates the 25th and the 75th percentile, a line within the box marks the median, whiskers indicate the 10th and the 90th percentiles, and points indicate the 5th and 95th percentile).

738

T.R. Beck et al.

Fig. 4. Results of accuracy tests of personnel radon monitors issued for German laboratories (see legend of Fig. 3 for box plot characteristics).

5. Conclusions Since the former uranium mines and industries dealt with natural radioactivity radiation, protection against natural sources, especially against radon and radon decay products, have a long tradition in Eastern Germany. Also these days, during the decommissioning of the uranium industries, more than 1500 workers are monitored with on-site or integrating individual monitors, most of them with ALGADE personnel monitors measuring potential alpha-energy exposure. As a result of the transformation of the EURATOM guideline into the German Radiation Protection Ordinance, radiation protection against natural sources has been expanded to workplaces in non-uranium mines, in the mining assurance industries, in exhibition caves and mines, in water-supply stations as well as in radon spas. Presently, many efforts are being undertaken to identify workplaces with exposures to radon exceeding deﬁned limits to introduce measures for remediation and/or for individual monitoring. To determine the exposure to radon or the potential alpha-energy exposure of short-lived radon decay products, integrating individual measurement devices are generally used. On the basis of ICRU and IAEA requirements for the monitoring of workers in case of external radiation, this work has developed physical accuracy criteria and application limits for individual radon and radon decay product monitors. The accuracy criteria describing the upper and lower limits within the measured value can be varied from the true value. Around the relevant limit of the measured potential alpha-energy exposure of 14 mJ h m−3 , an over- and under-estimation of the true value by a factor of 1.8 is accepted. For lower potential alpha-energy exposures, the measured value can overestimate the true value by not more than a factor of 2. Monitors for the exposure to radon can be applied at workplaces with equilibrium factors within a range from 0.2 up to 0.7 to meet the overall accuracy criteria. When the mean equilibrium factor is outside this range, a corrected factor has to be laid down. The measurement uncertainty around the relevant radon exposure limit of 6000 kBq h m−3 is ±20%. For lower radon exposures the measured value shall overestimate the true value by not more than a factor of 1.5. Tests show the applicability of these criteria in practice. In addition to regular performance tests, we recommend the introduction of an appropriate standardized quality assurance system and its accreditation. This ensures the competence

Quality assurance of individual radon measurements

739

of the laboratory services for the measurement task. As a rule, in many countries the competent authorities recognize accreditation as a necessary prerequisite for notiﬁcation as an ofﬁcially accepted measurement service. The BfS calibration laboratory for radon and radon decay products carries out annual inter-laboratory comparisons for integral passive radon measurements as a contribution to quality assurance in this ﬁeld of radiation protection. In these comparisons, the aforementioned accuracy criteria are applied to evaluate the capability of the laboratories and their measurement devices.

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13]

[14]

A. Cavallo, Radiat. Prot. Dosim. 92 (2000) 295. G. Keller, K.H. Folkerts, Health Phys. 47 (1984) 358. A.G. Scott, Health Phys. 45 (1983) 481. E. Straden, L. Berteig, F. Ugletveit, Health Phys. 36 (1979) 413. G.A. Swedjemark, Health Phys. 45 (1983) 453. A. Wicke, J. Porstendörfer, in: Natural Radiation Environment, Wiley Eastern, 1982, p. 481. A.C. George, A.J. Breslin, in: Natural Radiation Environment III, in: DOE Symposium Series, vol. 51, USDOE, Oak Ridge, TN, 1980, p. 1272, Report CONF-780422. J. Schmitz, R.M. Nickels, T. Bünger, Research Report BfS 1/468/00/AS, Salzgitter, 2000. J. Schwedt, W. Ullmann, in: BfS Jahresbericht, Salzgitter, 1994, p. 165. International Atomic Energy Agency, in: Safety Standards Series, No. RS-G-1.3, IAEA, Vienna, 1999. ICRP Publication 75: General principles for the radiation protection of workers, Ann. ICRP 27 (1) (1997). ICRP Publication 65: Protection against radon-222 at home and at work, Ann. ICRP 23 (2) (1993). UNSCEAR (United Nations Scientiﬁc Committee on the Effects of Atomic Radiation), Sources and Effects of Ionizing Radiation, UNSCEAR 2000 Report to the General Assembly, with Scientiﬁc Annexes, United Nations, New York, 2000. T.R. Beck, J. Schwedt, P. Hamel III, in: Proc. 3rd Eurosymposium on Protection against Radon, Liege, Belgium, 2001, p. 199.

740

Sensitivity analysis of an improved method for measuring the radon diffusion coefﬁcient of porous materials W.H. van der Spoel a , M. van der Pal b a Faculty of Civil Engineering and Geosciences, Section Building Engineering, Delft University of Technology,

Stevinweg 1, 2628 CN Delft, The Netherlands b Faculty of Architecture, Building and Planning, Group Building Physics, Eindhoven University of Technology,

PO Box 513, 5600 MB Eindhoven, The Netherlands

Several techniques are available for determining the radon diffusion coefﬁcient of porous materials. The common approach is to position a porous sample between two compartments. In one of the chambers a known high radon activity concentration is introduced while the radon concentration in the other chamber is measured. A steady state or transient analysis may be utilised. In this paper, a transient analysis is outlined for two experimental conditions in a cylindrical geometry, complemented with a rigorous treatment of the counting uncertainty in a continuous radon measurement. The sensitivity and accuracy of a least-squares regression to calculate the diffusion coefﬁcient from artiﬁcially generated data is discussed.

1. Introduction Diffusion of radon in building materials is a major driving force for radon exhalation into buildings with relatively low radon concentrations. The common approach to measure this material parameter in the laboratory is to position a porous sample between two compartments. In one of the chambers a known high radon activity concentration is introduced while the radon concentration in, or ﬂux into, the other (ventilated) chamber is measured. In case radon decay in the sample can be ignored and for steady-state conditions, this technique concerns a direct measurement of the diffusion coefﬁcient. A possible drawback is that one has to wait for equilibrium. A quicker time-dependent method was introduced by Nielson et al. [1] and Zapalac [2]. In the experiments of Nielson et al., a known high radon activity concentration is introduced in one of the chambers while the radon concentration in the ‘detection’ chamber is monitored continuously using a scintillation cell. By ﬁtting an analytical expression, corrected for RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07092-5

© 2005 Elsevier Ltd. All rights reserved.

Sensitivity analysis of an improved method for measuring the radon diffusion coefﬁcient

741

the volume of the detection chamber, for the time-dependent alpha activity in the detection chamber, the diffusion coefﬁcient is inferred. In the technique introduced by Zapalac, the integrated radon ﬂux into the (ventilated) second chamber is measured between contiguous time intervals and, as decay of radon is ignored, may only be applied for thin samples. Several others have used time-dependent methods. Søgaard-Hansen and Damkjær [3] made the assumption of well-mixed chambers, allowing an exact analytical solution including decay in the sample. Tsai and Hsu [4,5] applied a weighted ﬁt to the data obtained with the analytic method of Zapalac. The technique has also been used for radon diffusion through membranes. Wójcik et al. [6] made an analysis that included radon decay in the membrane. In this paper, we present and analyse two different time-dependent methods in a cylindrical geometry, using continuous measurements of the alpha activity in both chambers. The advantage of using a cylindrical sample is that leakage along the sample can be more easily avoided than for most other geometries. We outline a method to estimate the counting uncertainty of such a continuous measurement. These uncertainties are required to perform a weighted leastsquares regression. Finally, the strengths and weaknesses of the two methods are compared based on artiﬁcially generated data.

2. Model description Consider a hollow cylindrical sample with inner radius r1 and outer radius r2 . If we ignore radon generation and advective transport in the sample and assume that the pore–air radon concentration C (Bq m−3 ) only varies in the radial direction, the macroscopic radon transport equation may be written in polar coordinates: ∂C ∂ 2 C De ∂C = De 2 + − λC, ∂t r ∂r ∂r

(1)

where t is time (s), r is the polar coordinate (m), λ is the decay constant (s−1 ) and De is the effective diffusion coefﬁcient (m2 s−1 ). We may write D = βDe ,

(2)

where D is the bulk diffusion coefﬁcient1 and β a coefﬁcient [7,8], also called the “partitioncorrected porosity”, that accounts for partitioning of radon in the air, water and adsorbed phase: β = ε(1 − m + Lm) + ρka ,

(3)

where ε is the material porosity, m the fraction of moisture saturation of the pores, L the Ostwald coefﬁcient for partitioning between air and water, ρ the material density (kg m−3 ) and ka the radon surface adsorption coefﬁcient (m3 kg−1 ). For dry materials and no adsorption, β equals the porosity ε. Two types of experimental conditions are considered. In the ﬁrst type (1), a radon source is placed in the inner compartment with activity S1 (Bq). Assuming a well-mixed leak-tight 1 A reverse deﬁnition is also used.

742

W.H. van der Spoel, M. van der Pal

compartment, the boundary condition at r = r1 is ∂C ∂C (4) = 2πr1 hD − λV1 C + λS1 , r = r1 , t > 0, ∂t ∂r where V1 is the volume of the inner compartment and h the height of the sample. The terms at the right-hand side account for a diffusive ﬂux through the inner surface, radon decay and production, respectively. Since a scintillation cell is attached to the compartment, the volume V1 is larger than the inner volume of the sample (πr12 h). Similarly, the boundary condition at r = r2 is written as ∂C ∂C = −2πr2 hD − λV2 C, r = r2 , t > 0. V2 (5) ∂t ∂r As initial condition we have V1

C = 0,

r1 r r2 , t = 0.

(6)

In the second type (2), the boundary condition describes an (often used) experiment in which the radon concentration in the inner compartment is constant: C = C0 ,

r = r1 , t > 0.

(7)

The other conditions (5) and (6) still apply for this experiment. Note that since C = 0 at t = 0, the radon concentration in the inner compartment is described by a step function. The differential equation (1) with appropriate boundary conditions was solved using Laplace transformation. Basically, the radon concentration as a function of time in each of the compartments can be written as an inﬁnite series: C(t) = b0 +

∞

(8)

bi exp(ai t).

i=1

Note that for the ﬁrst type experiment, C(r, t = 0) = 0, thus type, this only holds for r > r1 , since C(r1 , t > 0) = b0 .

∞

i=1 bi

= −b0 . For the second

3. Experimental premises 3.1. Set-up In an experiment of the ﬁrst type, a radon source is placed in the inner compartment. Before the start of an experiment, the set-up should be ﬂushed for some time (> 3 h) with radonpoor ambient air to remove radon and its progeny. The airﬂow is preferably forced through the porous sample. After stopping the ﬂow, the alpha-activity in both compartments is measured with attached scintillation cells that register the number of pulses due to alpha-decay in contiguous time intervals of typically an hour. In an experiment of the second type, a constant radon concentration in the inner compartment may for example be obtained by placing a radon source in an external closed canister with a large volume ( Vset-up ) for more than about a month to establish equilibrium alphaactivity. The set-up is similarly ﬂushed as in the previous experiment, where after the large canister is connected to the inner compartment and air is continuously circulated between these volumes.

Sensitivity analysis of an improved method for measuring the radon diffusion coefﬁcient

743

3.2. Detection efﬁciencies The count rate measured with the two scintillation ﬂasks depends on the alpha activity in the measuring volume, determined by the concentrations in the scintillation ﬂask of 222 Rn and its alpha-emitting progeny. It therefore depends on the efﬁciencies η1 , η2 and η4 for detecting an alpha particle emitted in the decay of 222 Rn, 218 Po and 214 Po, respectively. The relative detection efﬁciencies η2 /η1 and η4 /η1 are usually > 1. In this paper, we use η2 /η1 = 1.15 and η4 /η1 = 1.33. The efﬁciency η1 for detecting an alpha from radon decay is about 0.5 for a common scintillation cell.

4. Error analysis The analytical expression for the number of counted pulses in the time intervals is ﬁtted to the data using a weighted least-squares method. As a result, the statistical error in the number of counted pulses in each interval is required. Based on the work of Aldenkamp and Stoop [9] and Inkret et al. [10], we have derived the variance of the net number of counted pulses in the case that the radon concentration is described by equation (8). A statistical analysis taking account of correlations in the decay process has also been utilised by Sonoc and Sima [11]. 4.1. Expectation and variance of counted pulses Consider the experiment in which N0 radon atoms (and no progeny) are present at some time t in a scintillation cell and in which the detected pulses due to alpha decay of radon and its progeny are counted from time t1 to t2 (t2 > t1 > t). Let pj (t1 − t, t2 − t) denote the probability of detecting j (j = 0, 1, 2, 3) pulses during the counting interval, originating from a radon atom present at t. Then the expectation value of the number of detected pulses E(X) is, omitting the dependency on t, t1 and t2 for brevity: E(X) = N0 (p1 + 2p2 + 3p3 ) ≡ N0 pE , where we have deﬁned pE ≡ p1 + 2p2 + 3p3 . The variance is given by [10] Var(X) = N0 pE − pE2 + 2p2 + 6p3 ≡ N0 pVar .

(9)

(10)

The alpha-activity of radon and its daughter atoms in the scintillation cell as a function of time are described by the Bateman equations. Integration of the activity of each nuclide over the counting interval from t1 to t2 gives the expected number of transitions in that interval: T1 transitions from 222 Rn take place, T2 from 218 Po to 214 Pb and T4 from 214 Po to 210 Pb. The expected number of detected pulses in the counter interval can be written as E(X) = η1 T1 + η2 T2 + η4 T4 = N0 pE ,

(11)

from which pE may be calculated. To calculate pVar we need the probabilities p2 and p3 . The probability of detecting 2 pulses during the counting interval, originating from a radon atom present at t, consists of 5 terms:

744

W.H. van der Spoel, M. van der Pal

(1) radon atoms present at t1 that decay twice and both decays are recorded; radon atoms present at t1 that decay thrice and (2) the ﬁrst and second decay is recorded and the third is not, (3) the ﬁrst and the third decay is recorded and the second is not, (4) the second and third decay is recorded and the ﬁrst is not; (5) 218 Po atoms present at t1 that decay twice: p2 =

n2 1 T2 − T4 η1 η2 + T4 η1 η2 (1 − η4 ) + T4 η1 η4 (1 − η2 ) + T4 η2 η4 (1 − η1 ) = N0 N0

(12) + T4 η2 η4 ,

where n2 is the number of radon atoms present at t from which two alpha decays are recorded in the counting interval from t1 to t2 , T2 is the number of transitions from 218 Po to 214 Pb originating from radon atoms present at t1 , T4 is the number of transitions from 214 Po to 210 Pb originating from radon atoms present at t , and T is the number of transitions from 1 4 214 Po to 210 Pb originating from 218 Po atoms present at t . The probability of detecting 3 pulses 1 consists of one term: radon atoms present at t1 that decay thrice and all decays are recorded: η1 η2 η4 T4 n3 (13) = . N0 N0 To apply the above model to a continuously changing radon concentration C(t) as described by equation (8), one should consider the rate of change of the number of new radon atoms Nn (t) in a scintillation ﬂask, for which we may write p3 =

dNn (t) dN (t) 1 dAr (t) (14) = + λN(t) = + Ar (t), dt dt λ dt where N(t) is the total number a radon atoms, and Ar (t) is the radon activity in the scintillation cell which equals Vm C(t) with V m the volume of the cell. In the counting interval from t1 to t2 , decays may be recorded from radon atoms (and their progeny) formed in the period between the start of an experiment at t = 0 and t1 , and radon atoms formed in the counting period from t1 to t2 . The expectation of the number of counted pulses is obtained by integration and consists of a contribution from both periods: t1 t2 dNn (t) dNn (t) E X(t1 , t2 ) = (15) pE (t1 − t, t2 − t) dt + pE (0, t2 − t) dt. dt dt 0 t1 For the variance Var(X) we similarly have t1 t2 dNn (t) dNn (t) pVar (t1 − t, t2 − t) dt + pVar (0, t2 − t) dt. (16) Var X(t1 , t2 ) = dt dt 0 t1 Both equations can be written in terms of the radon activity Ar (t), and thus C(t), using equation (14). In the second type diffusion experiment, we have a step-wise concentration change in the inner compartment at t = 0. The expectation of the number of counted pulses and the variance are then given by (15) and (16) added with, respectively: Vm C0 pE (t1 , t2 ) λ

and

Vm C0 pVar (t1 , t2 ) . λ

(17)

Sensitivity analysis of an improved method for measuring the radon diffusion coefﬁcient

745

Fig. 1. Ratio of variance and expectation of number of counted pulses in 1 h intervals as a function of time for different radon detection efﬁciencies η1 .

4.2. Example Consider an experiment with a constant radon source in a closed compartment without radon and progeny initially. The radon concentration is described by C(t) = a(1 − exp(λt)) where a is a constant. For a diffusion experiment the time-dependent radon concentration is slightly different, but the main results are similar. The ratio of the variance and expectation of the number of counted pulses in 1 h contiguous intervals is shown in Fig. 1, where ratios η2 /η1 = 1.15 and η4 /η1 = 1.33 have been used. The time-dependent behaviour of the variance is mainly of concern at the start, for t < 5 h. Thereafter, the equilibrium value is nearly reached which is dependent on the radon detection efﬁciency η1 . In case η1 , η2 and η4 1, or for a counting interval of less than about 30 s, the number of detected pulses approaches a Poisson distribution, i.e. Var(X) ≈ E(X). 5. Sensitivity analysis Artiﬁcial measurement data have been generated for the two experimental conditions based on the models described in Section 2 for an experiment of 7 days and a counting interval of 3 h, giving a total of 112 data points (data from both scintillation cells are ﬁtted simultaneously). A random error conforming to a Gaussian distribution with a calculated variance as outlined in Section 4 was added to each data point. The calculations were performed for a base case where h = 10.4 cm, r1 = 2.55 cm, r2 = 9 cm, V1 = 0.7 L and V2 = 2.8 L (of these, only V2 is varied in the sensitivity analysis). The volume of the scintillation cells is Vm = 0.3 L and the radon detection efﬁciency of both cells is set at η1 = 0.5. Further we assumed no background counts and the parameter β was set at 0.8. The regressions were carried out using the Levenberg–Marquardt method for four values of the effective diffusion coefﬁcient (10−6 , 10−7 , 10−8 and 10−9 m2 s−1 ) and a source strength S1 = 1000 Bq for type-1 conditions. For type-2 conditions, the concentration C0 was set at either 104 Bq m−3 or 105 Bq m−3 . In all cases data points with less than 10 counts were omitted. 5.1. Type-1 experiment First we consider a regression with De and S1 as free parameters and no systematic error in the other quantities. For this regression, the diffusion coefﬁcient is found with a relative

746

W.H. van der Spoel, M. van der Pal

uncertainty between 0.04% (De = 10−8 m2 s−1 ) and 0.6% (De = 10−6 m2 s−1 ). These small uncertainties indicate that taking S1 as a free parameter does not pose any problem. However, if the radon detection efﬁciency of e.g. scintillation cell nr. 2 has a small systematic error of 2%, the ﬁtted diffusion coefﬁcient will contain a systematic error. For De = 10−6 m2 s−1 the systematic error is 20% and the relative uncertainty is 18%. For De = 10−7 m2 s−1 these values are 4% and 2%, respectively. For smaller values of De the errors are less than 1.5%. It is thus rather difﬁcult to accurately determine the diffusion coefﬁcient in this way when De > 10−7 m2 s−1 . The cause of this problem lies in the fact that D e is largely determined by the concentration difference between the inner and outer compartment, which is very sensitive to the relative efﬁciency of the scintillations cells. This unwanted sensitivity would be lower for a larger volume V2 . When V2 = 28 L instead of 2.8 L, the systematic error in De reduces to 5% for De = 10−6 m2 s−1 and even further for larger values of V2 . An optimum is found when V2 is several cubic metres. As an alternative to using a large outer compartment, one may also add the relative efﬁciency as a free parameter in the regression. In doing so, it was found that the effective diffusion coefﬁcient in the range 10−6 –10−9 m2 s−1 is obtained without a systematic error and with a relative uncertainty < 1% when V2 = 2.8 L. The accuracy is better than 0.5% when V2 = 28 L. The accuracy of the method slightly degrades for V2 = 280 L. The above sensitivity analysis ignores the uncertainty in the partition-corrected porosity β, see equation (3). Since β depends on the porosity, water content, density and adsorption coefﬁcient, these quantities should be accurately known to determine the diffusion coefﬁcient. Alternatively, one could take β as a free parameter in the regression. If ηrel is accurately known, a regression with β, De and S1 as free parameter gives both β and De within 1.5% for De 10−7 m2 s−1 , β = 0.8 and V2 = 28 L (slightly worse for V2 = 2.8 L). For De = 10−6 m2 s−1 and V2 = 28 L, the accuracy is about 10%. In case ηrel is also a free parameter (giving a total of 4 free parameters), similar accuracies are obtained. 5.2. Type-2 experiment Similar tests were performed in case the concentration in the inner compartment is constant. For the regression with De and C0 as free parameters, it was found that the calculated diffusion coefﬁcient is less sensitive to systematic errors of the relative efﬁciency. It was also observed that increasing the outer volume has a positive effect on the systematic error in De . From the experimental point of view it is, however, difﬁcult to create a constant concentration in the inner compartment when the outer volume is large and the diffusion length is much larger than the sample thickness. In this respect, the outer volume should be small. The calculations have therefore been restricted to two volumes: V2 = 2.8 L and 28 L. For these values, the relative systematic error in De (10−6 m2 s−1 ) is about 4% when the systematic error in ηrel is 2%. The errors are smaller for lower values of De . Also for this experiment it is worthwhile to perform a regression with the relative efﬁciency as additional free parameter. It was found that in all cases (4 values of De ) with V2 = 2.8 L the calculated diffusion coefﬁcient has no signiﬁcant systematic error and the relative uncertainties are < 0.3%. With V2 = 28 L the results were slightly less accurate, but this is due to a 10 times lower value of C0 . With identical C0 the results are similar.

Sensitivity analysis of an improved method for measuring the radon diffusion coefﬁcient

747

As mentioned, one could also take β as an additional free parameter in the regression. If ηrel is accurately known, a regression with β, De and C0 as free parameters gives both β and De within 1.5% for De 10−7 m2 s−1 , β = 0.8 and V2 = 2.8 L (somewhat worse for V2 = 28 L). In case ηrel is also a free parameter (giving a total of 4 free parameters), similar accuracies are obtained. 5.3. Discussion The above sensitivity analysis ignores any uncertainty in the sample dimensions (h, r1 and r2 ) and other experimental errors such as varying detection efﬁciencies during the experiment due to, e.g, temperature ﬂuctuations, small leaks in the set-up, sample inhomogeneities, etc. In addition, the model assumes well-mixed compartments, which may not reﬂect reality. These effects could be the subject of a further sensitivity analysis. However, most of these errors can be minimised in a carefully prepared and executed experiment.

6. Conclusions Using the regression analysis for the ﬁrst type experiment, a diffusion coefﬁcient in the range 10−6 –10−9 m2 s−1 can accurately be determined using a 6.5 cm thick sample in an experiment of 7 days. For smaller values of De a thinner sample should preferably be used. Further, it was found that the outer compartment should not be too small, i.e., generally V2 > 10V1 , but not so large that the radon concentration becomes too low to be easily detected. Under these circumstances, a regression with S1 , De and the relative detection efﬁciency of the two scintillation cells as free parameters yields accurate results. This means that no information is required on the radon source strength and the relative detection efﬁciency of the scintillation cells to determine the radon diffusion coefﬁcient. A regression with β, De , S1 and ηrel as free parameters gives both β and De within 1.5% for 10−9 De 10−7 m2 s−1 . For larger De , a thicker sample should be used to determine β and De simultaneously. The second type experiment has a slightly better performance and is less sensitive to the volume of the outer compartment. From the experimental point of view, and if β is also a free parameter, a small volume is however preferred. For very small values of De (< 10−9 m2 s−1 ) thinner samples should be used. Again, no information is required on the radon concentration in the inner compartment and the relative detection efﬁciency of the scintillation cells to determine the radon diffusion coefﬁcient. It is concluded that both methods give similarly accurate estimates of De and β. The advantage of the ﬁrst method lies in its experimental simplicity since there is no need to maintain a constant radon concentration in the inner compartment. Also, pressure differentials due to air circulation that may inﬂuence radon transport through the sample are excluded. The regression technique is especially shown to be a useful tool since it does not require data for the radon source strength, or concentration C0 , and the relative detection efﬁciency of the scintillation ﬂasks.

748

W.H. van der Spoel, M. van der Pal

References [1] K.K. Nielson, D.C. Rich, V.C. Rogers, D.R. Kalkwarf, Report NUREG/CR-2875, US Nuclear Regulatory Commission, 1982. [2] G.H. Zapalac, Health Phys. 45 (1983) 377. [3] J. Søgaard-Hansen, A. Damkjær, Health Phys. 53 (1987) 455. [4] S.-C. Tsai, C.-N. Hsu, Geophys. Res. Lett. 20 (24) (1993) 2917. [5] C.-N. Hsu, S.-C. Tsai, S.-M. Liang, Appl. Radiat. Isot. 45 (8) (1994) 845. [6] M. Wójcik, Wlazło, G. Zuzel, G. Heusser, Nucl. Instrum. Methods Phys. Res. A 449 (2000) 158. [7] W.H. van der Spoel, Radon transport in sand, PhD thesis, TU Eindhoven, The Netherlands, 1998. [8] C.E. Andersen, Sci. Total Environ. 272 (2001) 33. [9] F.J. Aldenkamp, P. Stoop, Sources and transport of indoor 226 Radon – measurements and mechanisms, PhD thesis, Rijksuniversiteit Groningen, The Netherlands, 1994. [10] W.C. Inkret, T.B. Borak, D.C. Boes, Radiat. Prot. Dosim. 32 (1990) 44. [11] S. Sonoc, O. Sima, Radiat. Prot. Dosim. 45 (1992) 51.

749

Dose to the fetus from 222Rn in maternal drinking water E.S. Robbins a , N.H. Harley b a Department of Cell Biology Borough of Manhattan Community College,

The City University of New York, NY 10007, USA b Department of Environmental Medicine, New York University School of Medicine,

550 First Avenue, New York, NY 10016, USA

All gases are slightly soluble in water and radon is relatively soluble in groundwater. Concentrations from a few tens to hundreds of bequerels per liter of water can be found in most countries. After maternal ingestion of 222 Rn in raw water, it is transported by the blood and diffuses throughout the body including into the placenta. The short-lived alpha-particle emitting decay products from radon can thus reach an embryo/fetus. In the case of the embryo/fetus, it is of special interest to estimate the radiation dose because there is an early interval of pregnancy when the fetus is at the most risk for severe effects from radiation. In the very tiny early embryos, there are two important factors to consider. The very small size of the embryo should yield a commensurately small target for alpha-particle hits, but conversely, alpha-particle damage to nuclear DNA may have major consequences. For the purposes of dosimetric calculations in the developing embryo and fetus, both the maternal and the fetal placental blood supply at different time points were estimated. Our calculations relied on the available published data for these and for fetal weights as starting points. The 222 Rn accumulating in the various compartments following the ingestion of 100 Bq dissolved in water was estimated and the dose as a function of time is based on the pharmacokinetic model we developed. The clearance half times for the various tissues in the mother are based on some historic published human data for radon. If we assume an average of 0.6 liters of raw tap water at a concentration of 100 Bq per liter are consumed per day, the calculated total dose to the fetus over the term of the pregnancy is about 250 μSv or 25% of the normal background radiation dose. 1. Introduction Radon is a naturally occurring radioactive gas and is present in all environments. Radon is a decay product in the natural uranium decay series and is the direct descendant of radium-226. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07093-7

© 2005 Elsevier Ltd. All rights reserved.

750

E.S. Robbins, N.H. Harley

All persons are exposed not only to the naturally occurring radionuclides in the uranium series, but to other radioactive elements present on Earth because trace amounts are found in food, water and the air we breathe. Therefore the human body contains unavoidable small amounts of radioactive uranium, radium, thorium, carbon, lead, potassium, etc., from daily ingestion and inhalation. Radon is somewhat different from other natural radionuclides to which people are exposed, in that relatively high concentrations can be found in some homes and some drinking water, especially well water. Radon is a gas and a fraction escapes from soil beneath all homes and enters the living space. Depending upon the air exchange rate in the home and the parent radium concentration in the soil, homes can attain concentrations that require radon reduction [1]. All gases are slightly soluble in water and radon is relatively soluble in groundwater. Concentrations from a few hundred to hundreds of thousands of picocuries 222 Rn per liter of water can be found in the USA [2]. Surface water delivered by most water suppliers is low in radon concentration because the 222 Rn can exchange readily with the atmosphere, therefore a possible risk is from raw water ingested directly from a well. The major risk estimated for adults from radon, is lung cancer due to inhalation of the short-lived alpha-emitting decay products [3]. Most 222 Rn comes from the soil beneath a home. However, water can also be a source of human exposure in two ways, liberation into indoor air from water use, and two, ingestion of raw tap water [4,5]. Radon is released into air during water use such as showering and washing where it mixes with the existing 222 Rn from the soil gas source. The best estimate of the radon in home air from water use is that 1/10 000 of the radon concentration in water is found in air in a home. This factor is based on a summary of all the published data [4,5]. Thus, 10 000 pCi L−1 in water would produce 1 pCi L−1 in air in a typical home. The radiation dose to adult body organs, other than the lung, from ingestion of water is small [4,5]. Most solid radioactive substances such as uranium or radium in food do not cross the placental barrier efﬁciently. Radon gas, however, is dissolved in mother’s blood can be transported directly to the placenta. Thus, the fetal dose from the consumption of raw tap water is dependent upon the mother’s blood concentration, and blood supply to the placenta where diffusion to the fetus will occur. In the case of the embryo/fetus, it is of special interest to estimate the radiation dose because there is an early interval of pregnancy when the embryo/fetus is at the most risk for severe effects from radiation. Consideration of the possible dose of 222 Rn which may reach the earliest embryos is of particular difﬁculty because up until the blastocyst stage the embryo is very small and does not yet have a placenta to link it to the maternal blood supply. Its nutrition during this period comes from the yolk sac and allantois not the mother and thus, any 222 Rn reaching the embryo must partition into the maternal ﬂuids within the oviduct and uterus in which the embryo ﬂoats or sits. Radiation damage at this earliest stage could possibly lead to early abortion [6, p. 49] Preimplantation fetal wastage would be almost impossible to detect. The blastocyst embryo at the time of implantation (approximately 7–12 days) should be carefully considered for although the developing embryo (the inner cell mass) is still tiny, 0.1 to 0.2 mm in length [6] and weighing but milligrams, like the earlier cleavage stages, each cell is dividing and can be considered to be a totipotent stem cell [7]. Thus, there are

Dose to the fetus from 222 Rn in maternal drinking water

751

two important factors. The very small size of the embryo should yield a comensurately small target for alpha-particle caused ionizations, but conversely, alpha-particle damage to nuclear DNA may have major consequences, the least of which could yield a smaller than normal embryo. After implantation, the young embryo continues to divide rapidly. It grows from 1.5 mm in length at day 18 to 10 mm at day 39. By 42 days is 13 mm long and weighs 0.2 g. Moore and Reid [6] consider the weeks 4 to 8 to be the most critical period of development during which insults may cause major congenital malformations (p. 90). Alpha-particle damage from 222 Rn to these early stage embryos possibly has the ability to cause smaller than normal fetal size, fetal death, or malformations, as well as possible clones which may become malignant after birth. In A bomb survivors, mental retardation is documented as an effect for fetal irradiation by external radiation in the 8th to 15th weeks of development [8]. After birth, cancer is the major effect seen in follow up studies of irradiated children and adults [9]. It has also been reported that Japanese who suffered A bomb exposure early in utero have several decreased anthropometric measurements when examined at age 18 [10]. In several studies of fetal weights (calculated in various ways from ultrasound data), it has been reported that ﬁrst trimester embryos that are smaller than average risk various perinatal adverse outcomes such as low birth weight and premature delivery [11–13]. At 10 weeks the fetus is 61 mm long [6]. In later fetuses, i.e. after 10 weeks, cell division continues at a phenomenally high rate but the potential for massive or fatal radiation damage may be less because the larger number of cells, and the basic form of the tissues, structures and organs has already been laid down [6]. Although it is beyond the scope of this paper, a complete estimation of possible effects of 222 Rn would include a consideration of the fact that there are different sensitive periods for the causation of different malformations [14,15] and estimations of the fetal weights of many organs have been made [16]. After placentation, maternal blood is the source of 222 Rn to the embryo but quantitative data as to ﬁrst trimester blood ﬂow to the placenta (both maternal and fetal) is non-existent. Blood ﬂow to the uterus has been measured at 20 weeks and at term by angiography of the uterine arteries. At 20 weeks it is reported to be 513 ± 127 mL min−1 and at term to be 970 ± 193 mL min−1 [17]. Of course, only a fraction of the uterine artery blood ﬂow would go to the placenta as it also supplies the endometrium and myometrium. Fetal blood ﬂow to the term human placenta has been measured in the umbilical arteries by Doppler ultrasound and reported as 115 mL kg−1 min−1 [18]. 2. Blood ﬂow and weight data For the purposes of our dosimetric calculations in the developing embryo and fetus, both the maternal and the fetal placental blood supply at different ages was estimated. It was assumed that the weight ratio of blood in the embryo/newborn infant to its total weight is a constant, and that its blood ﬂow to the placenta is constant at 115 mL kg−1 min−1 . These parameters need veriﬁcation and are important quantities, not only for the purpose of these calculations, but for other basic toxicological calculations as well. The body composition of the fetus changes markedly with development. At 12 weeks it is 92.9% water and 0.5% fat while at term it is 72.4% water and 14% fat [19]. These changes will be reﬂected in the uptake and retention time of the fetal body at different times.

752

E.S. Robbins, N.H. Harley

3. Dosimetry The partitioning of radon into the various maternal compartments and fetus following the mother’s ingestion of 222 Rn dissolved in water is shown in the model in Fig. 1. The calculated quantities as a function of time are then based on a pharmacokinetic model which assigned rates and concentrations to different compartments. The clearance half times for the various tissues in the mother are based on a study of Harley et al. [20] where 2 individuals sat in a radon calibration chamber (25.9 kBq m−3 ) for one day and then measured 222 Rn in sequential breath samples. From these data, the clearance half times from lung, blood, intra and extracellular ﬂuid, and adipose tissue for adults were estimated and used in the model. The ability to validate the fetal model, at least for the mother’s total body content from 222 Rn ingested in water, comes from the study of Hursh et al. [21]. Two persons ingested 37 Bq 222 Rn dissolved in 100 ml water in four experiments. The decay product activity remaining in the body was measured using external gamma-ray counting. Radon in water was ingested in four experiments, ﬁrst on an empty stomach, and after a fatty meal. The maternal model ﬁts the Hursh et al. [21] data for ingestion on an empty stomach well (Fig. 2). The fatty meal measurements are not used as less 222 Rn is accessible to the mother’s blood, with subsequent lower dose. The model is therefore conservative, yielding the highest possible dose to the fetus. The half times for 222 Rn in the embryo/fetus are inferred from the estimated numerical blood ﬂow to the fetus, as a fraction of the maternal blood volume shown in Table 1. The dose is calculated based only on the body weight of the fetus, as there are not sufﬁcient data to calculate speciﬁc organ doses to the fetus. Although reported fetal weights for different ethnic groups vary somewhat by report and calculation method [22–25], we have used published American ﬁgures (Table 1). In the model input and clearance half times are assumed to be

Fig. 1. Biokinetic model for 222 Rn transport in maternal drinking water to fetus.

Dose to the fetus from 222 Rn in maternal drinking water

753

Fig. 2. Model of maternal total body 222 Rn and fetal 222 Rn for 100 Bq intake in drinking water. Table 1 Estimates of blood volume, umbilical ﬂow and body weight in the developing embryo/ fetus and the fetal dose from ingestion of 100 Bq of 222 Rn in water Age (weeks)

Blood volume (ml)

Umbilical blood ﬂow (ml min−1 )

Body weight (g)

Equivalent dose (μSv/100 Bq 222 Rn in water)

1 6 9 12 14 16 20 22 25 28 31 33 36 38

0.00007 0.014 0.6 3.5 7 14 35 45 71 106 141 176 212 247

–

0.001 0.2 8 50 100 200 500 630 1000 1500 2000 2500 3000 3500

0 4.1 4.1 3.0 2.1 1.3 0.53 0.42 0.26 0.18 0.13 0.11 0.09 0.08

0.02 0.9 5.7 11.5 23 58 72 115 172 230 288 345 402

Body weights from O’Rahilly and Muller [26] and Moore and Persaud [15]. The radiation weighting factor wr for alpha particles = 20.

equal in the fetus as there is no apparent mechanism for additional removal of dissolved gases from the fetus. The placenta is assumed to have the same 222 Rn concentration as the maternal blood. Again this is conservative because the diffusion of 222 Rn from maternal blood to the embryo/fetus may not completely equilibrate. The dose to the embryo prior to placentation is calculated with the assumption that the small tissue mass is in steady state or equilibrium with the maternal blood. All dose calculations include the alpha activity from build up of the short-lived decay products (218 Po, 214 Po). The decay products formed decay within the fetus with an effective half-life of 30 minutes. Decay

754

E.S. Robbins, N.H. Harley

in the fetus with no clearance is assumed as most are formed within cells. If clearance of the decay products could occur, our model predicts that the dose would be reduced by about a factor of 2. The modeled fetal body content of 222 Rn is shown for a single ingestion of 100 Bq by the mother (Fig. 2). Of course, dose to the fetus would be considerably greater if the mother ingested 222 Rn in her drinking water each day of the pregnancy. The model indicates an increase in dose to the embryo from this single exposure from 9 weeks to about 14 weeks and then a decrease. This is due to the assumed changing blood ﬂow rates. Table 1 shows that the highest calculated equivalent dose occurs between weeks 6 and 16, a key time period of organogenesis [6]. This is due mainly to the small fetal body mass during that interval. The dose at 1 week is zero. This is a consequence of the very small maternal body ﬂuid 222 Rn concentration in which the embryo ﬂoats. Not even one alpha-particle traversal of the embryo is calculated for an initial ingestion of 100 Bq during this earliest embryonic period.

4. Discussion The equivalent dose values in Table 1 can be compared with the dose to every embryo/fetus from natural external gamma-ray and cosmic-ray radiation of approximately 1000 μSv (1 mSv) for the fetal lifetime [27]. Thus, the maximum equivalent dose to the fetus is about 0.4% of the fetal-lifetime external gamma-ray and cosmic-ray dose for each 100 Bq ingested by the mother. If we assume an average of 0.6 liters of raw tap water at a concentration of 100 Bq per liter are consumed per day, the calculated total dose to the fetus over the term of the pregnancy is 250 μSv or 25% of the normal background radiation dose from all sources. The danger posed may be somewhat greater than this % indicates because of the greater DNA damage caused by α particles compared to other types of radiation. Beginning October 1, 1998, the Swedish Government requires 222 Rn in public water supplies greater than 100 Bq per liter to be remediated, and stated that water exceeding a concentration of 1000 Bq per liter is unsafe and cannot be delivered. The US Environmental Protection Agency is required to set a regulation for 222 Rn in drinking water in the United States. The standard they have proposed is 10 Bq per liter but to date no regulation has been set ofﬁcially.

Acknowledgement One author (N.H.H.) acknowledges support from USDOE grant DE-FG07-97ER62522.

References [1] NCRP Evaluation of Occupational and Environmental Radon Risk, National Council on Radiation Protection and Measurements, Bethesda, MD 2002. [2] Committee on Risk Assessment of Exposure to Radon in Drinking Water, Board on Radiation Effects Research, Commission on Life Sciences, National Research Council, Risk Assessment of Radon in Drinking Water, National Academy Press, Washington, DC, 1999.

Dose to the fetus from 222 Rn in maternal drinking water

755

[3] Committee on Health Risks of Exposure to Radon, Board on Radiation Effects Research, Commission on Life Sciences, National Research Council, Health Effects of Exposure to Radon: BEIR VI, National Academy Press, Washington, DC, 1999. [4] National Research Council, Risk Assessment of Radon in Drinking Water, National Academy Press, Washington, DC, 1998. [5] N.H. Harley, E.S. Robbins, A biokinetic model for 222 Rn gas distribution and alpha dose in humans following ingestion, Environ. Int. 20 (1994) 605–610. [6] K.L. Moore, G. Reid, in: The Developing Human, 3rd ed., Saunders, Philadelphia, 1982, pp. 2–6. [7] J.A. Thompson, J. Itskovitz-Elder, S.S. Shapiro, M.A. Waknitz, J.J. Swiergiel, V.S. Marshall, J.M. Jones, Embryonic stem cell lines derived from human blastocysts, Science 282 (1998) 1145–1147. [8] UNSCEAR (United Nations Scientiﬁc Committee on the Effects of Atomic Radiation), Sources and Effects of Ionizing Radiation, UNSCEAR 1993 Report to the General Assembly, with Scientiﬁc Annexes, United Nations, New York, 1993. [9] NCRP, Limitation of exposure to ionizing radiation, NCRP Report 116, National Council on Radiation Protection and Measurement, Bethesda, MD, 1993. [10] E. Nakashima, Relationship of ﬁve anthropometric measurements at age 18 to radiation dose among atomic bomb survivors exposed in utero, Radiat. Res. 138 (1994) 121–126. [11] G.C. Smith, M.F. Smith, M.B. McNay, J.E. Flemming, First trimester growth and the risk of low birth weight, N. Eng. J. Med. 339 (1998) 1817–1822. [12] P. Murru, E. Bertino, A. Coscia, et al., Dimensioni neonatali e differenziazone precoce dell crescita embriofetale, Acta Bi-Medica de L’Ateneo Parmense 71 (Suppl. 1) (2000) 393–396. [13] C.L.D. De Jong, A. Francis, H.P. Van Geijn, J. Gardosi, Customized fetal weight limits for antenatal detection of fetal growth restriction, Ultrasound Obstet. Gynecol. 15 (2000) 36–40. [14] L. Saxen, J. Rapola, in: Congenital Defects, Holt, Rhinehart and Winston, New York, 1989, pp. 112–139. [15] K. Moore, T.V.N. Pershaud, The Developing Human. Clinically Oriented Embryology, 6th ed., Saunders, Philadelphia, 1998. [16] R.H. Luecke, W.D. Wosilait, J.F. Young, Mathematical representation of organ growth in the human embryo/fetus, Int. J. Bio-Med. Comput. 39 (1995) 337–347. [17] J.C. Konje, P. Kaufmann, S.C. Bell, D.J. Taylor, A longitudinal study of quantitative uterine blood ﬂow with the use of color power angiography in appropriate for gestational age pregnancies, Am. J. Obstet. Gynecol. 185 608–613. [18] G.S. Dawes, The fetoplacental circulation, in: C.W.G. Redman, I.I. Sargent, P.M. Starking (Eds.), The Human Placenta, Blackwell, 1993. [19] S.J. Ulijaszek, F.E. Johnson, M.A. Preece (Eds.), The Cambridge Encyclopedia of Human Growth and Development, Cambridge University Press, Cambridge, UK, 1998. [20] J.H. Harley, E.S. Jetter, N. Nelson, Elimination of radon from the body, Environ. Int. 20 (1994) 573–584; Reprinted from USAEC, HASL Report 32, 1958. [21] J.B. Hursh, D.A. Morken, T.P. Davis, A. Lovaas, The fate of ingested 222 Rn ingested by man, Health Phys. 11 (1965) 465–476. [22] R.H. Luecke, W.D. Wosilait, J.F. Young, Mathematical modeling of human embryonic and fetal growth rates, Growth Develop. Ageing 63 (1999) 49–59. [23] J.G. Cecatti, M.R. Machado, F.F. dos Santos, E.F. Marussi, Curve of normal fetal weight values estimated by ultrasound according to gestation age (in Portuguese), Cad Saude Publ. 16 (2000) 1083–1090. [24] J.M. Lim, A.G. Hong, S. Ramam, N. Shyamala, Relationship between fetal femur diaphysis length and neonatal crown-heel length, Ultrasound Obstet. Gynecol. 15 (2000) 131–137. [25] M. Mongelli, A. Biswas, A fetal growth standard derived from multiple modalities, Early Hum. Dev. 60 (2001) 171–177. [26] R. O’Rahilly, F. Muller (Eds.), Human Embryology and Teratology, 2nd ed., Wiley, New York, 1996. [27] NCRP, Exposure the population of the United States and Canada to natural background radiation, NCRP Report 94, National Council on Radiation Protection and Measurement, Bethesda, MD, 1987.

756

Prenatal exposure due to ingestion of natural radionuclides S. Risica, C. Nuccetelli Istituto Superiore di Sanità (National Institute of Health), Physics Laboratory, Viale Regina Elena 299, 00161 Rome, Italy

This paper is devoted to exposure of the embryo and the foetus due to ingestion of natural radionuclides by the mother. Using the dose coefﬁcients of ICRP Recommendation 88, an assessment is made of the activities of the main natural radionuclides and depleted uranium ingested acutely or chronically by the mother or future mother at different stages of pregnancy and before conception, which give an effective dose of 1 mSv to the offspring. Results show that, in particular, chronic intake of some natural radionuclides in the diet could be of concern for the offspring.

1. Introduction In recent years, a revaluation has been made of the possible health effects of prenatal exposure to ionising radiation (see, e.g., the review in [1]). In a previous paper [2], the authors of this paper, in collaboration with other colleagues, analysed the intake of radionuclides by pregnant women as a consequence of both environmental contamination and diagnostic administration of radiopharmaceuticals. Some calculations were carried out estimating the activities of the main radionuclides typical of nuclear fallout and some radionuclides of natural origin, which give the embryo or foetus a 1 mSv equivalent dose. The same calculations were made for depleted uranium (DU). At that time, only one assessment of dose coefﬁcients for embryo/foetus was available, that is, a report of the US Nuclear Regulatory Commission [3], published in the nineties, aimed at developing recommendations on the exposure of pregnant women. The report assesses equivalent dose coefﬁcients for the embryo/foetus at different stages of pregnancy due to inhalation and ingestion by the mother of 30 radioisotopes of occupational, medical and environmental signiﬁcance. Both the internal chronic intake, at different stages of pregnancy, and the pre-existing body burden were considered [3]. Due to the fact that, in almost all ingestion scenarios, 226 Ra and 210 Po dose coefﬁcients are from 1 to 3 orders of magnitude higher than those of I, Cs and Sr (see discussion in [2]), it RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07094-9

© 2005 Elsevier Ltd. All rights reserved.

Prenatal exposure due to ingestion of natural radionuclides

757

became clear that special consideration should be given to possible doses due to ingestion by mothers, or possible mothers, of natural radionuclides in the diet. In recent months, a new ICRP Recommendation [4] has been published on the issue and used for the following elaboration. 2. The estimate of doses to the embryo and foetus by ICRP ICRP Recommendation 88 [4] gives dose coefﬁcients for the offspring following intake by the mother (members of the public as well as workers) of 31 elements for which age-dependent biokinetic models were provided in previous reports. Isotopes of these elements are both artiﬁcial and natural radionuclides. Dose coefﬁcients (doses to the offspring per unit intake by the mother, Sv Bq−1 ) are given for single (acute) intake and for continuous (chronic1 ) intake, either by inhalation or ingestion, occurring both before and during pregnancy. A range of acute intake times was adopted, whereas for continuous exposure a constant intake was assumed for three different periods during pregnancy or before conception (see Table 1). For constant intake before conception, no intake of radionuclides was assumed during the period of pregnancy. For each radionuclide, dose coefﬁcients are given for the effective dose to the offspring up to birth (ein utero ), the committed effective dose from birth to age 70 years (epostnatal ), resulting from activity present at birth, and the total committed effective dose to age 70 years (eoffspring ), the last being the sum of the other two (eoffspring = ein utero + epostnatal ). Since organ weighting factors are not available for the embryo/foetus the ICRP calculation used factors for adults [5]. Moreover, it assessed the equivalent dose up to birth for the tissue receiving the highest equivalent dose, and the equivalent dose to the brain in the period from the end of the 7th to Table 1 Total committed effective dose coefﬁcients (in Sv Bq−1 ) for the offspring of female members of the public from acute and chronic ingestion of main natural radionuclides [4] 210 Pb

210 Po

226 Ra

238 U

232 Th

Acute intake 2.5 y prior to conception 6 months prior to conception at conception week 5 week 10 week 15 week 25 week 35

7.2E–08 9.8E–08 1.1E–07 1.1E–07 1.1E–07 1.1E–07 1.4E–07 2.3E–07

8.6E–15 8.3E–09 2.6E–07 1.9E–07 1.2E–07 1.2E–07 1.2E–07 1.1E–07

1.1E–09 2.0E–09 1.3E–08 2.2E–08 2.1E–07 3.7E–07 5.0E–07 3.6E–07

8.0E–10 9.7E–10 1.2E–08 1.3E–08 1.5E–08 1.5E–08 1.3E–08 1.1E–08

1.1E–09 1.1E–09 3.4E–09 3.3E–09 3.2E–09 4.2E–09 8.8E–09 2.4E–08

Chronic intake for 5 y before conception for 1 y before conception during pregnancy (0–38 weeks)

7.5E–08 1.0E–07 1.4E–07

7.6E–09 3.8E–08 1.3E–07

1.5E–09 3.0E–09 3.2E–07

8.4E–10 1.1E–09 1.3E–08

1.1E–09 1.1E–09 9.4E–09

1 In Publication 88, chronic intake is taken to be constant and continuous over the period with a total intake of 1 Bq.

758

S. Risica, C. Nuccetelli

Table 2 Dose coefﬁcients (in Sv Bq−1 ) due to ingestion of the main natural radionuclides for different age groups of members of the public [6]

3 months 1 year 5 years 10 years 15 years adults

210 Pb

210 Po

226 Ra

238 U

232 Th

8.4E–06 3.6E–06 2.2E–06 1.9E–06 1.9E–06 6.9E–07

2.6E–05 8.8E–06 4.4E–06 2.6E–06 1.6E–06 1.2E–06

4.7E–06 9.6E–07 6.2E–07 8.0E–07 1.5E–06 2.8E–07

3.4E–07 1.2E–07 8.0E–08 6.8E–08 6.7E–08 4.5E–08

4.6E–06 4.5E–07 3.5E–07 2.9E–07 2.5E–07 2.3E–07

the end of the 15th week of gestation, as this is the period in which the brain has the greatest sensitivity to radiation damage. In Table 1, the total committed effective dose coefﬁcients up to age 70 (eoffspring ), due to acute or chronic ingestion by a female member of the public at different stages of pregnancy and before pregnancy, are shown for the main natural radionuclides. First of all, it is interesting to compare the dose coefﬁcients for acute intake with those for different age groups of members of the public [6], reported in Table 2. It can be noted that, unlike the dose coefﬁcients for infants (ﬁrst age group: up to 1 year) and small children (second age group: 1–2 years), which are almost always higher than those for adults, the natural radionuclides offspring dose coefﬁcients considered are generally lower than those for adults, with the exception of 226 Ra (for this nuclide, the age class most at risk is not infants or small children, but 15 year olds due to bone growth). However, it is worthy of note that, even if data are affected by large uncertainties, ICRP Recommendation 60 [5] states that, for most of intrauterine life, the risk of mortality for unit dose can be a few times higher than in the general population. The dose coefﬁcients of Table 1 were used to calculate the radionuclide activities corresponding to 1 mSv dose to the offspring shown in Table 3. The 1 mSv criterion was used because it is the dose limit for members of the public in the case of practices and for an embryo/foetus in the case of the mother’s exposure at work in Euratom Directive 96/29 [7], and the recommended value in the case of medical exposure of the mother in Euratom Directive 97/43 [8]. The same calculations were made for depleted uranium (DU), using the total committed effective dose coefﬁcients to the offspring for different uranium isotopes [4] and their wellknown activity fraction in DU [9] (see Table 4). Possible traces of Pu and other artiﬁcial nuclides were not taken into account. Results are in Table 5. Unfortunately, the results of Tables 3 and 5 cannot be compared to those obtained previously [2] with the NRC method [3] because the ICRP [4] calculates effective dose, whereas the NRC report calculates the equivalent dose to the embryo/foetus, considered as only one organ. Nevertheless, dose to the offspring, particularly for chronic intake of 210 Po and 210 Pb, should probably be considered of concern in both types of assessments. In particular, the consumption of drinking water rich in natural radioactivity could be a signiﬁcant source of exposure for the offspring. Indeed, derived levels (corresponding to 1 mSv) for 210 Pb and 210 Po calculated with ICRP dose coefﬁcients, for 5 years of consumption of drinking water

Prenatal exposure due to ingestion of natural radionuclides

759

Table 3 Activities (in Bq) of some natural radionuclides which, when ingested by the mother, correspond to a 1 mSv total committed effective dose to the offspring 210 Pb

210 Po

226 Ra

238 U

232 Th

Acute intake 2.5 y prior to conception 6 months prior to conception at conception week 5 week 10 week 15 week 25 week 35

1.4E+04 1.0E+04 9.1E+03 9.1E+03 9.1E+03 9.1E+03 7.1E+03 4.3E+03

1.2E+11 1.2E+05 3.8E+03 5.3E+03 8.3E+03 8.3E+03 8.3E+03 9.1E+03

9.1E+05 5.0E+05 7.7E+04 4.5E+04 4.8E+03 2.7E+03 2.0E+03 2.8E+03

1.3E+06 1.0E+06 8.3E+04 7.7E+04 6.7E+04 6.7E+04 7.7E+04 9.1E+04

9.1E+05 9.1E+05 2.9E+05 3.0E+05 3.1E+05 2.4E+05 1.1E+05 4.2E+04

Chronic intake for 5 y before conception for 1 y before conception during pregnancy (0–38 weeks)

1.3E+04 1.0E+04 7.1E+03

1.3E+05 2.6E+04 7.7E+03

6.7E+05 3.3E+05 3.1E+03

1.2E+06 9.1E+05 7.7E+04

9.1E+05 9.1E+05 1.1E+05

Table 4 Depleted uranium composition in terms of mass and speciﬁc activity [9]

234 U 235 U 238 U Total

Mass fraction (%)

Speciﬁc activity (Bq g−1 )

Activity in 1 mg of DU (Bq)

Activity fraction (%)

0.001 0.20 99.8 100

2.31E+08 8.00E+04 1.24E+04

2.31 0.159 12.4 14.9

15.53 1.07 83.35 100

Table 5 Activities and masses of DU which, when ingested by the mother, correspond to a 1 mSv total committed effective dose to the offspring Activities (Bq )

Masses (mg)

Acute intake 2.5 y prior to conception 6 months prior to conception at conception week 5 week 10 week 15 week 25 week 35

1.2E+06 1.0E+06 8.2E+04 7.5E+04 6.5E+04 6.5E+04 7.6E+04 8.8E+04

8.2E+04 6.8E+04 5.5E+03 5.0E+03 4.3E+03 4.4E+03 5.1E+03 5.9E+03

Chronic intake for 5 y before conception for 1 y before conception during pregnancy (0–38 weeks)

1.2E+06 9.0E+05 7.5E+04

7.8E+04 6.0E+04 5.0E+03

760

S. Risica, C. Nuccetelli

– with the 2 L d−1 WHO hypotheses [10], chosen also by the EU for its assessments [11,12] – are 3.6 and 35.6 Bq L−1 , respectively. The former value seems to be exceeded in Finland, for example, as the UNSCEAR Report 2000 [13] gives a range of 0.2 × 10−3 –21 Bq L−1 for 210 Pb concentration in drinking water. It is worth remembering that the recent EU Recommendation on the protection of the public against exposure to radon in drinking water supplies gives reference concentrations of 0.1 Bq L−1 for 210 Po and 0.2 Bq L−1 for 210 Pb “for water supplied as part of a commercial or public activity”, above which “. . . consideration should be given to whether remedial action is needed to protect human health” [14]. In Italy, no systematic nation-wide investigation has ever been carried out on natural radioactivity in drinking water and no data seem to be available in particular for 210 Po and 210 Pb in it. An overview of the main measurements performed on natural radioactivity (222 Rn, 238 U and 226 Ra) in drinking water is provided in Ref. [15]. Starting from these data, it seems that, thermal water excluded, high levels of natural radioactivity in drinking water can probably be found only in some limited areas in Italy. Other foodstuffs, like molluscs and crustaceans, are known to concentrate 210 Po and 210 Pb very efﬁciently. That is why very high concentrations of these radionuclides can be measured in them. However, being a minor dietary component, particularly for pregnant women, they are probably not a signiﬁcant source of exposure, except in singular cases. On the other hand, UNSCEAR Report 2000 [13] mentions signiﬁcant concentrations of 210 Po in meat in UK (0.06–67 Bq kg−1 ). Since meat is a main dietary component, even if its consumption is much lower than drinking water, this concentration should probably be monitored in other countries as well and its origin investigated more deeply. In general, looking at the relevant UNSCEAR Table, it can be said that natural radionuclides are sparsely monitored in various countries and probably require greater attention. As regards DU, no certain information is currently available on possible contamination of drinking water and/or the food chain in the territories of the countries bombed with it, thus making any conclusion on the degree of probability of intake reported in Table 5 impossible. However, some preliminary ﬁndings seem to indicate that the metal may now be present in the food chain and/or drinking water [16]. 3. Conclusion The results of the calculations made by means of dose coefﬁcients taken from the US NRC report [3] and the ICRP Recommendation 88 [4] are hardly comparable. Nevertheless, in both cases, chronic intake by pregnant women, particularly of 210 Po and 210 Pb, could be of concern for their embryo/foetus, in particular situations. On the other hand, as concerns DU, the lack of information about possible contamination of water and food in contaminated areas does not allow for realistic conclusions. Consequently, in the authors’ opinion, natural radionuclides and DU concentration in drinking water and foodstuffs should be investigated more widely, particularly in countries where the natural background is known to be high and in countries that have been bombed. This paper considered only possible exposure of female members of the public due to the ingestion of natural radionuclides. In the authors’ opinion, analysis of possible exposure of pregnant women in workplaces, e.g. in NORM industries, should be carried out to assess the effects of possible inhalation and/or ingestion of natural radionuclides in these situations too.

Prenatal exposure due to ingestion of natural radionuclides

761

Acknowledgements The authors are very grateful to Mr. G. Grisanti and M. Sabatini for their technical assistance.

References [1] P. Fattibene, F. Mazzei, C. Nuccetelli, S. Risica, Acta Paediatr. 88 (1999) 693–702. [2] S. Risica, P. Fattibene, F. Mazzei, C. Nuccetelli, A. Rogani, Radionuclides in pregnancy and breast-feeding, Microchem. J., in press. [3] M.R. Sikov, T.H. Hui, Contribution of maternal radionuclide burdens to prenatal radiation doses, NUREG/CR5631 PNL-7445, Rev. 2, Washington, 1996. [4] ICRP Publication 88: Doses to the embryo and fetus from intakes of radionuclides by the mother, Ann. ICRP 31 (1–3) (2001). [5] ICRP Publication 60: 1990 Recommendations of the International Commission on Radiological Protection, Ann. ICRP 21 (1–3) (1991). [6] ICRP Publication 72: Age-dependent doses to the members of the public from intake of radionuclides, Part 5: Compilation of ingestion and inhalation coefﬁcients, Ann. ICRP 26 (1) (1996). [7] Council Directive 96/29/Euratom of 13 May 1996, laying down basic safety standards for the protection of health of workers and the general public against the danger arising from ionising radiation, Ofﬁcial J. Eur. Commun. Ser. L 159 (29.6.1996). [8] Council Directive 97/43/Euratom of 13 June 1997 on health protection of individuals against the dangers of ionising radiation in relation to medical exposures and repealing directive 84/466/Euratom, Ofﬁcial J. Eur. Commun. Ser. L 180 (9.7.1997). [9] UNEP, Depleted Uranium in Kosovo, Post-conﬂict Environmental Assessment, UNEP Scientiﬁc Mission to Kosovo, 5–19 November 2000. [10] Guidelines for Drinking-Water Quality, 2nd ed., World Health Organisation, Geneva, 1998. [11] Council Directive 98/83/EC of 3 November 1998 on the quality of water intended for human consumption, Ofﬁcial J. Eur. Commun. Ser. L 330/32 (5.12.1998). [12] S. Risica, S. Grande, Council Directive 98/83/EC on the quality of water intended for human consumption: calculation of derived activity concentrations, Rapporto ISTISAN 00/16, 2000, p. 47; also available in the website, http://www.iss.itin PDF format. [13] UNSCEAR (United Nations Scientiﬁc Committee on the Effects of Atomic Radiation), Sources and Effects of Ionizing Radiation, UNSCEAR 2000 Report to the General Assembly, with Scientiﬁc Annexes, United Nations, New York, 2000. [14] Commission Recommendation of 20 December 2001 on the protection of the public against exposure to radon in drinking water supplies, Ofﬁcial J. Eur. Commun. Ser. L 344/85 (28.12.2001). [15] C. Giovani, L. Achilli, G. Agnesod, L. Bellino, M. Bonomi, M. Cappai, G. Cherubini, M. Forte, M. Gravaglia, S. Maggiolo, M. Magnoni, L. Minach, S. Risica, A. Sansone Santamaria, F. Trotti, in: J. Peter, G. Schneider, A. Bayer (Eds.), Proceedings of the 5th International Conference on High Levels of Natural Radiation and Radon Areas: Radiation Dose and Health Effects, Muenchen, Germany, September 4–7, 2000, 2002, p. 35. [16] N.D. Priest, M. Thirwall, Arch. Oncol. 9 (4) (2001) 237.

762

Meta-analysis of twenty radon and lung cancer case–control studies I.V. Yarmoshenko, I.A. Kirdin, M.V. Zhukovsky, S.Y. Astrakhantseva Institute of Industrial Ecology, 20A Sophy Kovalevska Str., GSP-594, 620219 Ekaterinburg, Russia

Epidemiological case–control studies should be considered to be a primary instrument for the investigation of the dose–effect relationship between radon exposure and risk of lung cancer. Eighteen publications on results of radon and lung cancer case–control studies conducted around the world have been found. Additionally, two studies performed in the Ural region of Russia were engaged. Thus, in total, twenty studies which include 12 044 cases and 20 932 controls were involved in this analysis. Two approaches were applied for the combined consideration of the published results. According to the ﬁrst approach, the cases and controls of each study were redistributed by equal ranges of radon concentration with regard to the parameters of a log-normal distribution and odds ratios (ORs) were re-estimated. Then the weighted average of ORs was calculated. The weights were evaluated from comparison of the characteristics of methods and techniques applied for indoor radon exposure assessment in each study and conﬁdence intervals (CIs) estimated for ORs. Obtained weighted ORs (with 95% CIs) were 1, 0.92 (0.86–0.99), 0.94 (0.88–1.02), 1.04 (0.97–1.11), 1.11 (1.01–1.21), 1.31 (1.12–1.53) in the intervals of radon concentration 0–25, 25–50, 50–75, 75–200, 200–400, > 400 Bq m−3 , respectively. When the wider range of radon concentration was considered as reference, the estimated weighted ORs were 1, 1.04 (0.99–1.1), 1.10 (1.03–1.17), 1.17 (1.05–1.3) in the intervals of radon concentration 0–75, 75–150, 150–300, > 300 Bq m−3 , respectively. In the second approach, the published adjusted ORs were used to estimate weighted average values of ORs in the ranges of radon concentration. Obtained OR values are 1, 1.06 (0.99–1.13), 1.15 (1.08–1.22), 1.27 (1.16–1.39), 1.32 (1.32–1.71) in the intervals of radon concentration 0–50, 50–100, 100–200, 200–300, > 300 Bq m−3 , respectively. According to the results of meta-analysis, the volume of case and control groups allows signiﬁcant conclusions on increasing linear dose–response relationships in the range of radon concentration above 75 Bq m−3 . The slope factor of a linear function representing the coefﬁcient of relative risk is 0.0012 (0.007–0.0017) m3 Bq−1 . Estimation in the lower exposure range (below 75 Bq m−3 ) gives some evidence for a U-shaped relationship. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07095-0

© 2005 Elsevier Ltd. All rights reserved.

Meta-analysis of twenty radon and lung cancer case–control studies

763

1. Introduction Radon is a natural radioactive gas entering houses from soil and the material of building construction. Gaseous radon and its progeny are accumulating in all buildings and the variability of concentration is quite high (characteristic geometric standard deviation is up to 2.7). A number of researches conducted around the world have proved that radon is a leading source of radiation exposure. The methods and techniques of indoor radon-concentration measurement are well developed and allow the conduct of large-scale surveys including epidemiological studies. Thus indoor radon exposure is one of the most favorable subjects for investigation of the health effects of low-level radiation exposure on the human population. Up to now, the models of radiation risk due to radon exposure were developed using results of miners cohort studies, which involved exposures to high levels of radon and progeny. Most complete analysis of these and other epidemiological data were undertaken by the BEIR VI Committee and presented in its report [1]. The review of cellular and molecular evidence led the Committee to the selection of a linear non-threshold relation between lung-cancer risk and radon exposure. However, the Committee acknowledged that other relationships, including threshold and curvilinear relationships, could not be excluded with complete conﬁdence, particularly at the lowest levels of exposure. The use of miner-based extrapolations provides uncertain estimates of the size of the relative risk in homes, although the expected relative risk is very small – 1.13 at 150 Bq m−3 with a 95% CI of 1.0–1.2 for a 30-year exposure. For a case–control study, this implies that substantial numbers of subjects are needed to establish a signiﬁcant difference in the distributions and to estimate the effects precisely. The results of meta-analysis involving the eight case–control studies [2] are quite variable with an estimated RR at 150 Bq m−3 of 1.14 (1.0–1.3). Using the results of combined analysis of fourteen case–control studies, the relative risk is 1.06 at 100 Bq m−3 with 95% CI of 1.0–1.1 [3]. Thus, the results of epidemiological case–control studies are generally consistent with the extrapolation based on miner studies. However, in both cases, the range of relative risk estimation at low exposures is still high. Besides that, practically no attempts were undertaken to test the pattern of dose–response relationship and linear dependence was assumed. To date, the results of at least twenty case–control studies of lung cancer and population indoor radon exposure have been published. On our opinion, new combined analysis seems to be essential and beneﬁcial.

2. Method Certain quantitative assessment of health effects of indoor radon exposure and development of a radiation risk model is possible only on a base of large-scale epidemiological studies. Cohort, case–control and ecological studies are three primary types of such study design usually utilized to undertake epidemiological research. In our research, we considered only case– control studies, which are most informative and reliable, to assess radiation risk in the range of exposure arising due to indoor radon. The relationship of effect and risk factor in a case–control study is assessed by estimation of the odds ratio (OR), which is deﬁned as the ratio of probabilities of exposure for members of case and control groups. Considering radon it is important to note that there are no people

764

I.V. Yarmoshenko et al.

unexposed to radon and its progeny and OR is estimated relative to some lowest range of indoor radon exposure. The method of combined analysis used for the published results was meta-analysis. The meta-analysis, as a construction of axiomatic–deductive theory of higher level, allows systematic interpretation and integration. The advantages of meta-analysis are the low cost and operational efﬁciency.

3. Data collection and systematization The copies of eighteen publications on the results of case–control studies of indoor radon and lung cancer incidence were collected. Additionally the results of two of our own studies were involved in the analysis. The total numbers in the case group is 12 044, control group – 20 932. The list of all studies is presented in Table 1. The study titled “USA (B. Cohen)” is quasi-case–control study based on analysis of questionnaire responses about people who had lived in the houses where contemporary radon levels were measured during the radon survey. It is necessary to point out that this study is not in the framework of Cohen’s famous ecological study of US counties’ lung cancer rates and mean indoor radon. The general characteristics of the case and control groups of the studies involved are presented in Table 1 as well. Table 1 General characteristics of the epidemiological case–control studies used in the meta-analysis Study

Year of publication, reference

Case group

Case group

size

smoking status

size

smoking status mixed mixed mixed mixed former and never smokers mixed mixed mixed mixed mixed mixed mixed mixed mixed mixed mixed mixed mixed non-smokers mixed

New Jersey, USA Shenyang, Chine Stockholm, Sweden Winnipeg, Canada Missouri 1, USA

1990 [4] 1990 [5] 1992 [6] 1994 [7] 1994 [8]

480 308 201 738 538

mixed mixed mixed mixed non-smokers

442 356 378 738 1183

Sweden 1 Finland 1 Finland 2 USA (B. Cohen) Southwest England East Germany West Germany Middle Urals, Russia Missouri 2, USA Italy North Urals, Russia Schneeberg, Germany Iowa, USA Sweden 2 Gansu, Chine

1994 [9] 1996 [10] 1996 [11] 1997 [12] 1998 [13] 2000 [14] 2000 [14] 2000 [15] 1999 [16] 2000 [17] 2000 [15] 1999 [18] 2001 [19] 2001 [20] 2002 [21]

1281 164 517 1865 982 1053 1449 195 372 263 127 72 413 258 768

mixed mixed mixed mixed mixed mixed mixed mixed mixed mixed mixed mixed mixed non-smokers mixed

2576 331 517 3034 3185 1667 2297 242 471 265 202 288 614 487 1659

Sex

female female female both sex female both sex male both sex both sex both sex both sex both sex both sex female both sex both sex female female male both sex

Meta-analysis of twenty radon and lung cancer case–control studies

765

Table 2 Methods of indoor radon measurements Study

Measurement technique

Duration of measurements

Mean coverage of exposure estimation (y)

New Jersey, USA Shenyang, Chine Stockholm, Sweden Winnipeg, Canada Missouri 1, USA Sweden 1 Finland 1 Finland 2 USA (B. Cohen) Southwest England East Germany

SSNTD, CC SSNTD SSNTD SSNTD SSNTD SSNTD SSNTD SSNTD CC SSNTD SSNTD, CC

22 24 25 10–25 20 23 25 37

West Germany

SSNTD, CC

Middle Urals, Russia Missouri 2, USA Italy North Urals, Russia Schneeberg, Germany Iowa, USA Sweden 2 Gansu, Chine

SSNTD SSNTD, RT–ST SSNTD SSNTD SSNTD, RT–ST SSNTD, RT–ST SSNTD SSNTD

1 year 1 year 1 year 6 months 1 year 3 months 2 months 1 year 1 week 6 months 1 year (SSNTD), 3 days (CC) 1 year (SSNTD), 3 days (CC) 3 months 1 year 2 half-year periods 3 months 1 year (SSNTD) 1 year 3 months 1 year

30 35 35 17 15–25 35 22 20 20 25 5–30

RT–ST: retrospective technique with surface 210 Po trap; SSNTD: solid state nuclear track detector; CC: charcoal canisters.

Information on the methods of indoor radon measurements, duration of measurements and coverage of radon exposure estimation is presented in Table 2. The radon measurement techniques used to assess indoor radon exposure were: retrospective methods using so-called surface 210 Po traps, solid-state nuclear track detectors (SSNTD), and charcoal canisters. The retrospective methods are most preferable in epidemiological studies as they allow long-term indoor radon variation consideration and indoor radon exposure reconstruction. The SSNTD takes into account medium-term indoor radon variation and provides only measurements of contemporary indoor radon concentration for periods no longer than one year. The most ineffective measurement technique for the reconstruction of indoor radon exposure is that of the charcoal canister. In order to reconstruct the historical radon exposure more precisely, the radon measurement protocol of the majority of the studies included measurements in currently and previously occupied houses. The annual average of indoor radon concentration is used as a surrogate of radon exposure in the majority of the studies. The values of OR in categories of radon exposure are estimated and presented in all publications. Taking into account the methods of indoor radon measurements, duration of measurements and coverage of exposure estimation presented in Table 2, the weights for each study, were evaluated. To evaluate the weights of a study, the scoring scales for each of the above parameters were elaborated. The higher a score corresponded to better techniques and approaches

766

I.V. Yarmoshenko et al.

Table 3 Scores and weights of studies Study

Score by measurement technique

New Jersey, USA Shenyang, Chine Stockholm, Sweden Winnipeg, Canada Missouri 1, USA Sweden 1 Finland 1 Finland 2 USA (B. Cohen) Southwest England East Germany West Germany Middle Urals, Russia Missouri 2, USA Italy North Urals, Russia Schneeberg, Germany Iowa, USA Sweden 2 Gansu, Chine

2 2 2 2 2 2 2 2 1 2 2 2 2 3 2 2 3 3 2 2

Score by duration of measurements 4 4 4 3 4 2 2 4 1 3 4 4 2 4 4 2 4 4 2 4

Score by coverage of exposure estimation

Study weight

2 2 2 1 2 2 2 3 1 3 3 3 1 2 3 1 2 2 2 2

8 8 8 6 8 6 6 9 3 8 9 9 5 9 9 5 9 9 6 8

for radon exposure reconstruction. For example, the retrospective method of exposure reconstruction got the highest score. The scores for the duration of the measurements and the coverage of radon exposure reconstruction were assigned according to duration of time interval. The weight of the study was evaluated by summation of the three scores. The assigned scores and ﬁnal study weights are presented in Table 3. The obtained weights were used to obtain weighted average estimates of the radiation risk characteristics. 4. Results of meta-analysis Two approaches were used for the combined meta-analysis of the twenty studies. According to the ﬁrst approach, the distribution of cases and controls was re-estimated by identical radon exposure categories. By the second approach, the original distributions by radon exposure categories, individual for each study, and the original adjusted OR values were utilized. The estimation by the ﬁrst approach was based on the assumption that distributions of cases and controls by radon exposure are close to log-normal. The published parameters of the log-normal distribution (geometric mean and geometric standard deviation) were used or estimation by indirect parameters was undertaken. With the parameters of the log-normal distribution and under the condition of consistency with the original published frequency distribution of cases and controls, new frequency distributions by identical categories of radon exposure were estimated. New values of the odds ratio (ORi,k ) were calculated for each study (index k) by identical categories of exposure (index i).

Meta-analysis of twenty radon and lung cancer case–control studies

767

Then the weighted average odds ratio (ORw ) was calculated for each category of radon exposure. The weighting was conducted considering both the assigned study weights and the conﬁdence interval estimated for each ORi,k with the equation: ORw (1) (ORi,k wi,k Wk ) wi,k Wk , i = i

i

where index k relates the variable to one of the twenty studies, index i relates the variable to a radon exposure category, wi,k is a weighting coefﬁcient inversely proportional to the 95% conﬁdence interval, Wk is the weight assigned to the k study. The results of the estimation are presented in Table 4 and Fig. 1. The obtained dose–response relationship is close to a curvilinear U-shaped dependence as presented in Fig. 1 by the solid line and ﬁtted by the equation: ORw (CRn ) = e−0.04(CRn −CRn,0 )

0.33

+ 0.0012(CRn − CRn,0 ),

(2)

where CRn is the annual radon concentration (surrogate of radon exposure), CRn,0 is the average radon concentration within the reference category, where OR is set equal to 1. Here CRn,0 = 16 Bq m−3 . In the above case, odds ratio values in the categories of radon concentration higher than 25 Bq m−3 are estimated relative to the lowest range of concentration, where the accepted OR Table 4 Weighted average odds ratio estimated by the ﬁrst approach, assuming ORw = 1 for radon concentration category 0–25 Bq m−3 Radon concentration category (Bq m−3 )

ORw with 95% CI in brackets

0–25 25–50 50–75 75–200 200–400 >400

1 0.92 (0.86–0.99) 0.94 (0.88–1.02) 1.04 (0.97–1.11) 1.11 (1.01–1.21) 1.31 (1.12–1.53)

Fig. 1. Dose–response dependence by the ﬁrst approach, assuming ORw = 1 for radon concentration category 0–25 Bq m−3 .

768

I.V. Yarmoshenko et al. Table 5 Weighted average odds ratio estimated by the ﬁrst approach, assuming ORw = 1 for radon concentration category 0–75 Bq m−3 Radon concentration category (Bq m−3 )

ORw with 95% CI in brackets

0–75 75–150 150–300 >300

1 1.04 (0.99–1.1) 1.10 (1.03–1.17) 1.17 (1.05–1.3)

Fig. 2. Dose–response dependence by the ﬁrst approach, assuming ORw = 1 for radon concentration category 0–75 Bq m−3 .

is 1. After combining the categories of radon concentration up to 75 Bq m−3 , the dependence of OR on radon concentration becomes close to linear. The results of OR estimation in the case of a wider range of reference category are presented in Table 5 and Fig. 2. The obtained linear dependence is well ﬁtted by the equation: ORw (CRn ) = 0.00042(CRn − CRn,0 ) + 1,

where CRn,0 = 38 Bq m−3 .

(3)

Conﬁdence intervals for the slope factor are 0.00011–0.00076 m3 Bq−1 . Published data from the considered case–control studies contain the estimations of the adjusted odds ratio (ORA ). Adjustment of the risk coefﬁcient estimation is a standard procedure in epidemiological studies. Conventionally, the adjustment of OR in case–control studies of the lung cancer and radon exposure relationship is conducted by sex, age and smoking habit of the members of case and control groups. Some researchers make extra adjustments for other factors if necessary. The available values of adjusted OR estimated in all studies with the exception of Cohen’s pseudo case–control study are presented in Fig. 3. The wide range of OR assessments, which prove the weakness of individual studies’ results, can be seen from Fig. 3. For combined analysis by the second approach, the weighted average adjusted odds ratios (ORAW ) were estimated using a modiﬁcation of equation (1): ORA = wi,k Wk . ORAW (4) i,k wi,k Wk i i

i

Meta-analysis of twenty radon and lung cancer case–control studies

769

Fig. 3. The adjusted odds ratios presented in the publications on the results of case–control studies. Table 6 Weighted average odds ratio estimated by the second approach, assuming ORw = 1 for radon concentration category 0–50 Bq m−3 Radon concentration category (Bq m−3 )

ORw with 95% CI in brackets

0–50 50–100 100–200 200–300 >300

1 1.06 (0.99–1.13) 1.15 (1.08–1.22) 1.27 (1.16–1.39) 1.50 (1.32–1.71)

Fig. 4. Dose–response dependence by the second approach, assuming ORw = 1 for radon concentration category 0–50 Bq m−3 .

The results of ORAW estimations are presented in Table 6 and Fig. 4. The dose–response i relationship in this case is well ﬁtted by the linear dependence: ORw (CRn ) = 0.0012(CRn − CRn,0 ) + 1,

where CRn,0 = 26 Bq m−3 .

(5)

770

I.V. Yarmoshenko et al.

The conﬁdence interval for the slope factor of linear function is 0.0007–0.0017 m3 Bq−1 . While conducting all the above combined estimations of radiation risk, an analysis of each study’s inﬂuence on the results was performed by dropping any study which did not change the results signiﬁcantly.

5. Discussion The total sizes of the case and control groups involved in the meta-analysis are sufﬁcient to make quite reliable conclusions on the dose–response relationship in the range of radon concentration above 75 Bq m−3 . To reach the necessary level of signiﬁcance below 75 Bq m−3 , pooled analysis with the original data treatments could be undertaken. It should be mentioned that from the radiation protection point of view, successful radiation risk management below an indoor radon concentration of 75 Bq m−3 is questionable. Above the radon long-term exposure equivalent to an annual radon concentration of 75 Bq m−3 , the dose–response dependence is linearly increasing. Combined analysis of the published adjusted-OR data allows the radiation risk coefﬁcient value per unit radon concentration to be assessed at 0.0012 with a 95% conﬁdence interval of 0.0007–0.0017. Deviation from linearity is observed in the range of radon concentration 0–75 Bq m−3 when comparing the results of the epidemiological study at low categories of exposure. That nonlinearity of dependence may reﬂect a real dependence. However, the non-linearity may appear due to some confounding inﬂuence as well as from the high probability of both systematic and random errors of radiation exposure assessment especially at this low range of indoor radon concentration. A number of countries have adopted indoor radon reference levels in agreement with ICRP recommendations in the range from 200 to 600 Bq m−3 . Comparing the results of the metaanalysis performed here and ICRP recommendations, it may be concluded that meeting those recommendations provides a limitation on indoor radon exposure at such a range of radon concentration where the negative health effects of radon exposure appear to be reliable. The results of meta-analysis of twenty radon and lung cancer case–control studies presented here can be utilized to improve the appropriate radiation risk models.

References [1] Committee on Health Risks of Exposure to Radon, Board on Radiation Effects Research, Commission on Life Sciences, National Research Council, Health Effects of Exposure to Radon: BEIR VI, National Academy Press, Washington, DC, 1999. [2] J.H. Lubin, J.D. Boice Jr., Lung cancer risk from residential radon: meta-analysis of eight epidemiologic studies, J. Natl. Cancer Inst. 89 (1997) 49–57. [3] S. Darby, D. Hill, R. Doll, Radon: a likely carcinogen at all exposures, Ann. Oncol. 12 (2001) 1341–1351. [4] J.B. Schoenberg, J.B. Klotz, H.B. Wilcox, et al., Case–control study of residential radon and lung cancer among New Jersey women, Cancer Res. 50 (1990) 6520–6524. [5] W.J. Blot, Z.Y. Xu, J.D. Boice Jr., et al., Indoor radon and lung cancer in China, J. Natl. Cancer Inst. 82 (1990) 1025–1030. [6] G. Pershagen, Z.H. Liang, Z. Hrubec, et al., Residential radon exposure and lung cancer in Swedish women, Health Phys. 63 (1992) 179–186.

Meta-analysis of twenty radon and lung cancer case–control studies

771

[7] E.G. Letourneau, D. Krewski, N.W. Choi, et al., Case–control study of residential radon and lung cancer in Winnipeg, Manitoba, Canada, Am. J. Epidemiol. 140 (1994) 310–322. [8] M.C. Alavanja, R.C. Brownson, J.H. Lubin, et al., Residential radon exposure and lung cancer among nonsmoking women, J. Natl. Cancer Inst. 86 (1994) 1829–1837. [9] G. Pershagen, G. Akerblom, O. Axelson, et al., Residential radon exposure and lung cancer in Sweden, N. Engl. J. Med. 330 (1994) 159–164. [10] E. Ruosteenoja, I. Makelainen, T. Rytomaa, et al., Radon and lung cancer in Finland, Health Phys. 71 (1996) 185–189. [11] A. Auvinen, I. Makelainen, M. Hakama, et al., Indoor radon exposure and risk of lung cancer: a nested case– control study in Finland, J. Natl. Cancer, Inst. 88 (1996) 966–972. [12] B.L. Cohen, Questionnaire study of the lung cancer risk from radon in homes, Health Phys. 72 (1997) 615–622. [13] S. Darby, E. Whitley, P. Silcocks, B. Thakrar, M. Green, P. Lomas, J. Miles, G. Reeves, T. Fearn, R. Doll, Risk of lung cancer associated with residential radon exposure in South-West England: a case–control study, Br. J. Cancer 78 (1998) 394–408. [14] H.-E. Wichmann, J. Heinrich, M. Gerken, M. Kreuzer, Domestic radon and lung cancer-current status including new evidence from Germany, in: Proceedings of the 5th International Conference on High Levels of Natural Radiation and Radon Areas: Radiation Dose and Health Effects, Munich, September 2000, in: International Congress Series, vol. 1225, 2001. [15] I.A. Kirdin, I.V. Yarmoshenko, V.L. Lezhnin, et al., Radon and lung cancer case–control study in the Middle Urals, in: IV All-Russian Congress on Radiation Researches (Radiobiology, Radioecology, Radiation Protection), Moscow, November 2001 (in Russian). [16] C.R. Michael, P.H. Alavanja Dr., J.H. Lubin, et al., Residential radon exposure and risk of lung cancer in Missouri, Am. J. Public Health 89 (1999) 1042–1047. [17] F. Bochicchio, F. Forastiere, S. Mallone, et al., Case–control study on radon exposure and lung cancer in an Italian region: preliminary results, in: Proceedings of the 10th International Congress of The International Radiation Protection Association “Harmonization of Radiation, Human Life and the Ecosystem”, Hiroshima, Japan, May 2000 (CDRom, P-2a-68). [18] High residential radon health effects in Saxony (Schneeberg Study), Contract N◦ FI4P-CT95-0027 (Coordinator: Dr. Karl Martin), European Commission, DG XII, Nuclear Fission Safety Programme, 1999. [19] W.R. Field, D.J. Steck, B.J. Smith, The Iowa radon lung cancer study – phase I: residential radon gas exposure and lung cancer, Sci. Total Environ. 272 (2001) 67–72. [20] F. Lagarde, G. Axelsson, L. Damber, et al., Residential radon and lung cancer among never-smokers in Sweden, Epidemiology 12 (2001) 0. [21] Z. Wang, J.H. Lubin, L. Wang, et al., Residential radon and lung cancer risk in a high-exposure area of Gansu Province, China, Am. J. Epidemiol. 155 (2002).

772

A new Austrian recommendation guide for radon prevention in the design and construction of new buildings in areas with highly elevated radon levels F.-J. Maringer a,* , P. Schillfahrt b , T. Auer c , R. Pecina d a ARC Seibersdorf Research GmbH, Faradaygasse 3, Arsenal Obj. 214, A-1030 Vienna, Austria b Architecture Ofﬁce, Putzenweg 2, A-6460 Imst, Austria c TA Bauplanungs GmbH & CO KEG, Rosslach 157, A-6441 Umhausen, Austria d Chemical Technical Environmental Agency Tyrol, Wilhelm-Greil-Strasse 17, A-6020 Innsbruck, Austria

The identiﬁcation and localisation of extremely high indoor air radon levels (maximum 274 kBq m−3 ) in the village of Umhausen (Tyrol, Austria) in 1992 led to the urgent need for adequate radon prevention guidelines for the architects of new buildings taking into account the very high radon potential in this speciﬁc area (maximum soil-gas radon concentration about 600 kBq m−3 ). Therefore a team of architects, scientists and engineers created a manual with detailed descriptions and recommendations of adequate radon prevention measures for the design and construction of new buildings in this area on behalf of and ﬁnancially supported by the Government of Tyrol. The scientiﬁc background, the technical methods, and the practical implementation of the recommended prevention measures are shown. The radon prevention concept of the guide is based on three main principles: (1) Air-tight sealing of soil-contact building parts (e.g. basement) and installation penetrations with barriers against radon diffusion and convection. (2) Reduction of vertical convective connections inside the building to limit the vertical temperature and meteorological-driven pressure gradient and decreasing the inﬂow of soil-gas radon into the building. (3) Installation of a gravel layer under the basement for the possibility of the later installation of a sub-ﬂoor depressurisation facility in the case of a decreasing radon barrier effect of the sealing materials. The ﬁrst experiences with the practical implementation of the new guide, the ﬁnancial implications for the builder, and the most important aspects of the general application of the proposed radon prevention principles are discussed. * E-mail address: [email protected] (F.-J. Maringer).

RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07096-2

© 2005 Elsevier Ltd. All rights reserved.

A new Austrian recommendation guide for radon prevention

773

1. Introduction In 1989 a statistically signiﬁcant increase (+40%) of lung cancer death was found in the district of Imst, Tyrol in comparison with the Austrian average value [1,2] during a routine statistical evaluation of causes of death of the Austrian population. After various investigations, unusually high indoor air radon activity concentrations were proved as the reason for the locally elevated lung cancer mortality in the village Umhausen (Fig. 1). The part of Umhausen (2600 inhabitants) with the highest indoor radon concentrations is situated in an alluvial fan of an ancient (∼ 8700 years ago) giant rock slide. The basic soil particles consist of granitic gneiss with median radium-226 activity concentrations of 125 Bq kg−1 (maximum: 17 kBq kg−1 ). Soil gas radon concentrations were found up to 600 kBq m−3 . The medians of the indoor radon concentration were found to be 3750 Bq m−3 (basements) and 1160 Bq m−3 (ground ﬂoor). A maximum indoor radon concentration of about 274 kBq m−3 was found in a basement in summer 1992. Confronted with these results, the responsible governmental authorities urgently needed mitigation and prevention measures to solve this locally restricted real radon problem. Some extremely high radon-burden buildings, were mitigated instantly. To solve the problem concerning new buildings, a team of architects, scientists and engineers created a manual with detailed descriptions and recommendations of adequate radon prevention measures for the

Fig. 1. Location of the radon-prone village of Umhausen, District of Imst, Tyrol, Austria.

774

F.-J. Maringer et al.

design and construction of buildings in this area on behalf of and ﬁnancially supported by the Government of Tyrol. So far, there does not exist a legislative limit for indoor radon activity concentration in Austria. But there is a valid recommendation of the Austrian Radiation Protection Commission [3] with recommended limits (annual average radon-222 concentrations) of 200 Bq m−3 for new buildings (prevention limit) and 400 Bq m−3 for existing buildings (action limit). These values are in agreement with the EU recommendation [4].

2. Scientiﬁc–technical background and methods 2.1. Scope of the guidelines The primary purpose of the developed recommendation guidelines was the long-term reduction of the annual average radon-222 activity concentration in new buildings probably below 400 Bq m−3 . This higher value in opposition to the Austrian recommendation was chosen because of the difﬁcult radon situation in Umhausen to support an economically realistic radon prevention approach. The main focus of the guidelines is placed on single-family houses. However, the principles should be applicable to larger buildings like schools and kindergartens. The recommended technical and construction requirements should be directly applicable by building professionals. All measures should be based on the current state-of-theart construction and building techniques and must fulﬁl economic efﬁciency. 2.2. Applied precautionary principles and recommended building techniques The applied principles and recommendations are based on international experience of radon mitigation and precaution that have been collected during the last 10 years. This international experience has been well published in several papers and guides, e.g. [5–11]. Summing up this review, a three-step strategy is laid down for these guidelines: • installation of radon-barriers in soil-contacting building parts; • reduction of the vertical pressure gradient inside the building; • pre-preparation of a sub-slab depressurisation system. The radon barriers have to be tested and certiﬁed both against radon diffusion and radon convection. Three types of radon barriers in the basement (all soil-contacting parts of the building) are recommended: (a) construction of a waterproof/radonproof concrete basement tub (slab as well as all soiltouching basement walls, including tight installation penetrations); minimum concrete thickness 30 cm (Fig. 2); (b) basement made of concrete and totally sealed with radonproof isolation materials like plastic or bituminous ﬁlms and tight penetrations of installation tubes; minimum concrete thickness 30 cm (Figs. 3–5); (c) combination of basement type (a) and (b) if the soil-touching building parts (basement) are assigned for a usual residential purpose (not only for workshop or storage use).

A new Austrian recommendation guide for radon prevention

Fig. 2. Construction of basement type (a): waterproof/radonproof concrete tub.

Fig. 3. Construction of basement type (b): radonproof isolation of the basement.

Fig. 4. Construction details: tight waste water outlet and installation inlet.

775

776

F.-J. Maringer et al.

Fig. 5. Application of tight installation outlets/inlets in concrete/basement walls.

To avoid or reduce the radon transmission from the soil into the basement, the vertical pressure gradient – caused by temperature differences and wind – should be reduced. The measures taken to reduce the vertical pressure gradient inside the house are: • ﬁlling/closing of all installation perforations, gaps and joints (water, sewage, electricity, heating system, telephone, . . . ) between ﬂoors with elastic sealing compounds; • direct outdoor air support of heating stoves and the chimneys; • installation of automatically closing, airproof doors between soil-contacting and non-soilcontacting parts of the building (e.g. between basement and ground ﬂoor); • avoiding of open staircases inside the building. The third part of the applied radon precautionary strategy is the pre-preparation of a sub-slab depressurisation system. The sub-slab depressurisation system could be installed later on if the radon barrier effect has grown weaker, e.g. due to ageing of materials. This pre-preparation essentially consists of a 15 cm continuous gravel layer (Fig. 6).

A new Austrian recommendation guide for radon prevention

777

Fig. 6. Construction of the basement slab: pre-preparation of a sub-ﬂoor depressurisation system.

3. Implementation experience and ﬁnancial impact of precautionary measures First applications of the radon precautionary concept show, instantly afterwards, radon activity concentration ratios of indoor to soil gas of about 0.5‰. That corresponds to average indoor radon concentrations of about 300 Bq m−3 at soil-gas radon concentrations of 600 kBq m−3 . The most important factor during the practical implementation of the construction using precautionary materials and components is their careful, gap-free build. Only a single small joint or fault in the radon isolation materials is able to eliminate success totally. The additional costs of the application of the recommended radon prevention measures could be estimated for a single-family house (9 m×10 m, basement height 2.5 m) at: (a) radonproof basement tub: EUR 2500–3000; (b) radonproof isolation of basement: EUR 5000–6500; (c) combination of barriers (a) and (b): EUR 6000–7500. The long-term behaviour of the applied concrete and radon barrier materials has to be observed carefully in the concerned buildings. If a decreasing radon isolation effect leads to a signiﬁcantly elevated average indoor radon concentration later on the sub-slab depressurisation system must be installed. Therefore the indoor radon concentration of the concerned houses will be checked by long-term integrating track etch or electret radon detectors (3month measuring period during winter season) periodically (e.g. every 3 years).

4. Prospects: future radon prevention strategy in Austria An evaluation of the application experience and radon reduction success of all the concerned houses is planned in about three years after the publication of the guidelines (2001 → 2004). Based on this planned evaluation, the guidelines will be revised and improved. In parallel, an Austrian standard for radon prevention in new buildings including these guidelines will be published by the Austrian Standardisation Institute OENORM. In this standard, all levels of radon potential areas should be included not only highly elevated radon-prone areas. The publication date could be as soon as the fall of 2002. The application of these guidelines and the new Austrian Standard on new built public buildings in radon-prone areas in Austria will increase experience and could be a valuable means to reducing the radon exposure of the population especially of children and the young generation.

778

F.-J. Maringer et al.

A very important decision of the Austrian scientiﬁc community and governmental authorities in the past was to solve the radon problem in Austria from the high-level side focusing more on new buildings than on existing buildings. This concept supports a cost-effective and collective dose-effective reduction of the radon exposure of the Austrian population.

References [1] O. Ennemoser, et al., High indoor radon concentration in an alpine region of western Tyrol, Health Phys. 67 (2) (1994) 151–154. [2] O. Ennemoser, et al., Unusually high indoor radon concentrations from a giant rock slide, Sci. Total Environ. 151 (1994) 235–240. [3] Richtwerte für die Radonkonzentration in Innenräumen. Österreichische Strahlenschutzkommission, Bundesministerium für Gesundheit, Sport und Konsumentenschutz, Wien, 1992. [4] Commission Recommendation of 21 February 1990 on the protection of the public against indoor exposure to radon, Ofﬁcial J. Eur. Commun. Ser. L 80 (1990). [5] B. Clavensjö, G. Akerblom, The Radon Book – Measures against Radon, The Swedish Council for Building Research, Ljunglöfs Offset AB, Stockholm, 1994. [6] Radon: Technische Dokumentation für Baufachleute, Gemeinden, Kantone und Hauseigentümer, Bundesamt für Gesundheit, Abt. Strahlenschutz, Bern, 2000. [7] Radon-Handbuch Deutschland, Bundesministerium für Umwelt, Naturschutz und Reaktorsicherheit. In Zusammenarbeit mit Bundesamt für Strahlenschutz und Bundesamt für Gesundheit, Schweiz, Bonn, 2001. [8] Assessment Protocols: Durability of Performance of a Home Radon Reduction System. Sub-slab depressurisation systems, US-EPA Publ., no. 625/6-91/032, 1991. [9] Radon Reduction Techniques for Existing Detached Houses, Technical Guidance for Active Soil Depressurization Systems, 3rd ed., US-EPA Publ. No. 625/R-93/011, 1993. [10] T. Gregory, in: Practical Experience in Solving Radon Problems in Public Buildings: a Range of Successful Design Solutions, in: Seminar Series, Radon in Buildings, Institution of Engineers of Ireland, 1992. [11] ÖBV-Richtline, Wasserundurchlässige Betonbauwerke – Weiße Wanne’, Österr. Betonverein, Wien, 1999.

779

The new radon programme in England B.M.R. Green a , L. Davey b a Radon Studies Group, National Radiological Protection Board, Chilton, Didcot, Oxon OX11 0RQ, United Kingdom b Radioactive Substances Division, Department for Environment, Food and Rural Affairs, Ashdown House,

123 Victoria Street, London SW1E 6DE, United Kingdom

Local Authorities in high radon areas of England have been invited to join a partnership with central Government with the joint objectives of identifying and reducing unacceptable high radon levels in homes. Previous radon programmes have successfully identiﬁed more than 40 000 homes with high levels, but the remediation rate has been disappointingly low. The new partnership is based on the experience gained during pilot studies in three separate areas of England. A signiﬁcant increase in the remediation rate was achieved by emphasising the involvement of the Local Councils, informing local health and building professionals and providing local guidance and advice to householders. The new programme is described, early response rates and results reported and discussed and some deﬁnite conclusions drawn.

1. Introduction The ﬁrst comprehensive survey of indoor radon levels throughout the United Kingdom was carried out almost 20 years ago [1] by the National Radiological Protection Board (NRPB) with support from the Commission of the European Communities. Further work over the following years [2–4] culminated in advice to the Government, published in 1990 [5], which detailed a scheme to control exposures to this known carcinogen. The advice provided a framework to identify the areas of the country at most risk of high levels and suggested a reference level (the Action Level, set at 200 Bq m−3 ) at or above which the advice is to reduce radon concentrations in homes. A complete radon map of England was published in 1996 [6], Wales in 1998 [7] and Northern Ireland in 1999 [8]. Central Government fully supported this approach and, in England, the Radioactive Substances Division of the then Department of the Environment, subsequently the Department of the Environment, Transport and the Regions (DETR) and now Department for Environment, Food and Rural Affairs (DEFRA) initiated a programme to identify homes with unacceptable high radon levels and encourage remedial works to reduce the risk [9]. During this period, three-month radon testing using two passive detectors was available to householders, free on RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07097-4

© 2005 Elsevier Ltd. All rights reserved.

780

B.M.R. Green, L. Davey

demand. In addition, in the high-risk areas deﬁned by the radon maps, householders were invited by leaﬂet drops, mail-shots and local publicity to apply for the free test [10]. This programme was successful in persuading between 25% and 35% of householders in high risk areas to accept the offer of a radon test and in identifying a large number of homes – more than 40 000 – with radon concentrations at or above the Action Level, but the remediation rate has been disappointingly low. Research indicated that only 10 to 20% of the householders were taking steps to reduce the radon in their homes [11]. 2. Pilot studies In response to this disappointing statistic, DETR proposed to alter the direction of the radon programme by refocusing at local authority rather than central government level and by restricting offers of a free measurement to householders in the areas of greatest risk, deﬁned as 5% or greater probability of houses exceeding the Action Level. In addition, householders of homes already identiﬁed as having a high radon level would be offered free advice on remediation and a free re-measurement to test the efﬁciency of any works completed. To enable the local authorities to fulﬁl their role, intensive training sessions would be organised by specialist agencies such as NRPB and the Building Research Establishment (BRE). The Communications Directorate of DETR/DEFRA would provide support on public relations and publicity. To evaluate this new approach, pilot studies, commencing in 1998, were carried out in conjunction with three local authorities: Derbyshire Dales District Council in the Peak District; Cherwell District Council in North Oxfordshire; Mendip District Council in Somerset (see Fig. 1). These three Districts are predominantly rural areas with the population split between small market towns, villages and hamlets and outlying farmsteads – characteristics typical of many radon areas in England. High radon levels were identiﬁed in Derbyshire and Somerset in the initial national survey [10] but high levels in parts of Oxfordshire were not identiﬁed until over 10 years later in 1996 [6]. Consequently, Cherwell has little experience of radon compared to the other two districts which have both had extensive, but different initiatives over many years. Cherwell also differs from the other two areas in that it is a relatively afﬂuent area on the edge of the Cotswold Hills and within commuting distance of London by road or rail. A ﬁrm of independent consultants was employed to facilitate the pilot studies, to aid in drawing up Action Plans and to evaluate the success of the programme. The pilot studies came to a successful conclusion in 2000 with the publication of an evaluation report by the consultants [12]. “Response rates were generally high and remediation was successfully speeded up and increased in all three areas, with the total of remediated properties being raised by up to 100% on previously achieved numbers. Key factors in this success included local and pro-active delivery, effective targeting, appropriate timing of publicity, optimum use of technical expertise and sustained support and follow up.” 3. The new programme Based on the success of and the experience gained during the pilot studies, all local authorities at district level in high radon areas of England were invited to join a partnership with

The new radon programme in England

781

Fig. 1. Local authorities working with the Government (axis values denote the National UK Grid).

central Government with the joint objectives of identifying and reducing unacceptable high radon levels in homes. The new partnership also involves the specialist agencies of NRPB and BRE and involves working with local networks of inﬂuential professionals in the building, medical, social care, legal, surveying and house-transaction trades and representatives of large landlords. The whole programme would focus on encouraging the householders and/or landlords of homes identiﬁed with unacceptable high radon levels to take remedial action. In July 2000, the Environment Minister, Michael Meacher announced that over 30 local councils had accepted the opportunity to work in partnership with the Government in a new three-year programme to cut levels of radon in homes [13]. The councils from high radon areas across England (see Figs. 1 and 2) have formed into 12 regional groups (Table 1) to take advantage of the economies of scale possible for such events as information workshops and local publicity. An individual Action Plan is drawn up for each grouping by the independent consultants following discussions with DEFRA, the councils concerned and the specialist

782

B.M.R. Green, L. Davey

Fig. 2. Radon affected areas in England and Wales (the darker the area, the greater the risk).

agencies. These plans provide a framework for the programme: detailing the training seminar by the NRPB and the BRE for council staff, local professionals and others; identifying the number of copies of informative literature, overprinted with the logos of the local councils; setting out an overall timetable for the main events. Central Government is supporting the programmes by making resources and the services of specialist agencies available to the local councils. The agencies help in the drawing up of an Action Plan; providing expertise in the risks associated with exposure to radon and practical advice on ways to reduce levels effectively and efﬁciently in existing dwellings; undertaking the training programme; providing a radon measurement service both to identify the homes at risk and to demonstrate whether remedial methods have been successful as well as providing a backup service to the local authorities. The initial training is carried out by specialist staff from NRPB and BRE with separate sessions for Elected Councillors; Council Ofﬁcers; professionals such as solicitors, building

The new radon programme in England

783

Table 1 Local authority regional grouping Regional group

Local authorities

Regional group

Local authorities

North West

South Lakeland Lancaster Craven Richmondshire S. Somerset N. Dorset N. Kesteven S. Kesteven Alnwick Berwick Kerrier South Hams Torbay West Devon Bath & S. E. Somerset S. Gloucestershire Stroud

West Midlands

Ashﬁeld Derbyshire dales High Peak Staffordshire Moorlands Harborough Melton Rutland Oswestry S. Shropshire Corby Kettering Daventry E. Northants Huntingdon Northampton S. Northants Wellingborough

Somerset Lincolnshire North East Cornwall South West

West

East Midlands

Shropshire N. Northants Northants

Table 2 Phases in a typical radon action plan 1. Initial training 2. Awareness raising 3. Targeting of ﬁrst-time testers

4. Targeting of known high homes 5. Remediation phase 6. Exit strategy

surveyors, architects, estate agents; local building contractors. In addition, a special seminar with a medical bias and involving a senior ofﬁcial from the Department of Health is offered to local health professionals. The seminars, each geared to the particular audience, last about 2 hours including a question and answer session. Tailored information packs are made available to delegates. The plans for each regional group differ in detail, but normally have six main phases, some of which occur in parallel and some consecutively (Table 2). The methods used to increase the general awareness on the dangers of radon vary according to the desires and resources available to the council. DEFRA will arrange for publicity material such as radon bookmarks and drinks coasters to include both the unifying radon logo (Fig. 3) and the logo of the local councils. The same logos are printed onto information leaﬂets, posters and local maps. In some areas, mobile exhibition units visit local villages. News releases and sometimes a photo opportunity with a local dignitary, such as a mayor or local member of parliament, results in local media coverage, backed up by radio or television interviews. A popular publicity piece is sea-side rock (of the mint sweet variety) printed through with the words “NRPB, radon free rock” and used at seminars and public meetings to underline the fact that all soils and rocks, other than the sweets on offer, give off some radon.

784

B.M.R. Green, L. Davey

Fig. 3. The unifying logo for the radon programme.

The householders in high radon areas whose homes have not been measured for radon are referred to as ‘First-time testers’ and those householders living in homes already identiﬁed as having radon levels at or above the Action Level as ‘High-testers’. In both cases, initial targeting is by individual postal packages containing a standard letter from the local authority, a personalised letter from NRPB with a reply slip already partially completed, information leaﬂets and a reply-paid envelope to the local authority. All that the householder needs to do to obtain a free measurement (if a ﬁrst-time tester) or free advice (if a high tester) is to complete the reply slip, seal in the envelope and return it to the Council. Advice on remediation is offered to the high-testers by personal contact with council ofﬁcers. Often, the ofﬁcers will make home visits. Initially, as part of a training process, the ofﬁcer will be accompanied by an expert from BRE, but will soon acquire sufﬁcient experience to deal effectively with the majority of homes. The nominal duration of a regional Action Plan is a calendar year and there is a need for a clear exit strategy for central government commitment at the end of this time. However, each local authority recognises that there will not be a sharp cut-off, but rather a gradual reduction over time during which householders expectations, raised during the programme, will need to be fulﬁlled. The requirement to provide accessible information and expert advice on radon will continue for the foreseeable future.

4. Results to date Following the ministerial announcement in the summer of 2000, the programme got off to a good start with the agreement and implementation of the Action plan for the Northwest group. Over 3300 ﬁrst-time testers, out of over 13 000 invited, took up the offer of a free test; a response rate of 25%. However, following this initial success, the programme slowed for two reasons: the moratorium on Government activity in the run up and during the UK general election campaign of May 2001 and the far-reaching effects of the foot and mouth epidemic in the same year. Virtually all the high radon areas in England are in rural areas; these are the areas most affected by foot and mouth disease in livestock. In the current year (2002) however, the programme is getting back on track with a total of over 32 000 invitations and reminder letters sent to ﬁrst-time testers (Table 3), over 8000 positive replies received and pairs of detectors dispatched and 2000 high-testers offered help from their local councils. For some councils, the Action Plan called for reminder letters to be sent to those householders who did not respond to the initial ﬁrst-time test letter within six weeks. This strategy proved to be cost-effective, increasing the overall response by between 10% and 15% (Table 4).

The new radon programme in England

785

Table 3 Summary of invitations issued to ﬁrst-time testers and the response rates Group

Council

Number of letters sent

Number of positive responses

Positive response (%)

North West

South Lakeland DC Lancaster CC Craven DC Richmondshire DC Bath & NE Somerset Council Sth Gloucestershire Council Stroud DC Sth Shropshire DC Oswestry BC Harborough DC Melton BC Rutland CC

7300 3000 1300 1700 5000 230 1000 2600 47 320 1200 5800

1900 770 190 450 930 66 290 1400 21 120 470 1500

26 25 14 27 19 29 29 55∗ 45 38∗ 39∗ 26

B & NE Somerset, Sth Glos. & Stroud Sth Shropshire & Oswestry Welland Partnership

∗ Includes the response to reminder letters.

Table 4 Summary of response rates when reminder letters are issued Local authority

South Shropshire DC Harborough DC Melton BC

First-time tester letters No. sent Replies

%age

Reminder letters No. sent Replies

%age

Overall response (%)

2551 317 1216

40 27 25

1583 234 908

24 14 18

55 38 39

1000 87 309

400 32 159

Table 5 Results from the Northwest Group Local authority

No. of results

At or above AL

Percentage

South Lakeland Lancaster Craven Richmondshire Totals

1876 853 168 480 3377

265 100 23 88 476

14 12 14 18 14

The results so far from over 3300 ﬁrst-time testers in the Northwest show an encouraging hit rate of 14% with some 470 homes identiﬁed at or above the Action Level out of the results currently available (Table 5). 5. Discussion Typically, the response rate to the initial offer to ﬁrst-time testers was in the range 25% to 30%. Two councils had much lower rates, Craven (14%) and Bath and NE Somerset (19%) and two councils much higher, South Shropshire (40%) and Oswestry.

786

B.M.R. Green, L. Davey

In Craven, householders were not provided with a pre-paid reply envelope, instead they were asked to accept the offer by telephoning the Council Ofﬁces. The obvious conclusion is that householders do not ﬁnd this an easy or convenient option; the response rate in the neighbouring authority, Richmondshire, was almost double at 27%. The substitution of the instruction to respond by telephone rather than providing a pre-paid envelope was the only difference between the offer in these two very similar districts in the Yorkshire Dales. The below average response in Bath and NE Somerset is probably due to the urban nature of the area targeted. This was part of the centre of Bath with a greater proportion of rented accommodation. In addition, some of this accommodation is either occupied by students at nearby colleges or is sheltered housing for the elderly. The higher than average response in Shropshire can be attributed to two factors: a vigorous and effective campaign organised by the local council ofﬁcers with the full support of the elected members and, to a lesser extent, the fact that this area had been targeted on only one occasion in the past unlike many other areas which have been the subject of more than one radon campaign at various times.

6. Conclusions This large programme is still in progress and much work remains to be completed. The next phase is to encourage householders with high levels to carry out effective works to reduce the radon levels in their homes. However, some deﬁnite conclusions can be drawn: • • • • •

Pre-paid envelopes signiﬁcantly increase the response to the offer of a test. The response rate is higher in rural areas than urban areas. An enthusiastic and well planned campaign at local level works. A reminder letter is cost-effective. Radon probability mapping is a useful tool in locating high homes.

References [1] B.M.R. Green, et al., Surveys of natural radiation exposure in UK dwellings with passive and active measurement techniques, Sci. Total Environ. 45 (1985) 459–466. [2] NRPB, Exposure to radon daughters in dwellings, NRPB-ASP10, National Radiological Protection Board, Chilton, 1987. [3] M.C. O’Riordan, A.C. James, B.M.R. Green, A.D. Wrixon, Exposure to radon daughters in dwellings, NRPBGS6, NRPB, Chilton, 1987. [4] A.D. Wrixon, B.M.R. Green, P.R. Lomas, J.C.H. Miles, K.D. Cliff, E. Francis, C.M.H. Driscoll, A.C. James, M.C. O’Riordan, Natural Radiation Exposure in UK Dwellings, NRPB R190, HMSO, London, 1988. [5] M.C. O’Riordan, Human exposure to radon in homes, Docs NRPB 1 (1) (1990) 15–32. [6] Radon affected areas: England, Docs NRPB 7 (2) (1996) 3–9. [7] Radon affected areas: Wales – 1998 review, Docs NRPB 9 (3) (1998) 3–9. [8] Radon affected areas: Northern Ireland – 1999 review, Docs NRPB 10 (4) (1999) 3–8. [9] Department of the Environment, Government accepts NRPB recommendation to reduce the Action Level for radon, Environ. News Release 32 (19 January) (1990). [10] Department of the Environment, Tony Baldry, Junior Environment Minister, launches radon information campaign for Devon and Cornwall, Environ. News Release 134 (11 March) (1991).

The new radon programme in England

787

[11] E.J. Bradley, J.M. Thomas, An analysis of responses to radon remediation advice, NRPB-M707, NRPB, Chilton, 1996. [12] A. Thomas, J. Hobson, Review and evaluation of the radon remediation pilot programme, DETR report no. DETR/RAS/00.004, 2000. [13] Department of the Environment, Transport and the Regions, Environment Minister, Michael Meacher, announces new radon programme, DETR News Release 467 (10 July) (2000).

788

Idea of assessment of the annual dose when non-continuous data exist D. Kluszczy´nski ´ Teresy St., 90-950 Łód´z, Poland Nofer Institute of Occupational Medicine, 8 Sw.

In many settings where the population under exposure is large (e.g. in underground mining), the problem arises of how to estimate the static parameters of dose distribution from ionising radiation (like mean annual dose and the dispersion of the annual doses). The problem results mainly from an insufﬁcient number of measurements and particularly from the lack of extreme values (especially the highest values) in the set of data. Most of the values of static distribution parameters depend on the availability of these extreme values. The paper discusses the inﬂuence of the number of measurements carried out on the probability of recording an extreme value in the result pool. The other problem is how to convert the results of short-term measurements into dose distribution. The paper presents the statistical algorithm that allows conversion of the results obtained during non-continuous (grab) measurements. The concept of a measurement density factor was conceived to describe the correlation between the dose and the dose rate. The observed difference between the theoretical value and the empirical value of the measurement density factor in metal-ore mines is discussed. 1. Introduction In underground mining, the workers are exposed to ionising radiation (especially to radon and its short-time progeny). Due to the large number of exposed miners and the very hard conditions of work, it is not always possible to cover with dosimetry the whole population of underground miners, or the miners may not be covered by continuous measurements. The insufﬁcient number of measurements and the high variability of radon progeny concentration can result in the wrong assessment of the annual dose to miners. In this case, the number of measurements, which should be carried out must base on some statistical formulae. The formulae determine the number of measurements that need to be performed to include the extreme values. Another problem is ﬁnding the correlation between short-term-measurement results and the dose. This paper presents the theoretical and empirical considerations which may be useful in dose assessment for such exposed populations as miners. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07098-6

© 2005 Elsevier Ltd. All rights reserved.

Idea of assessment of the annual dose when non-continuous data exist

789

2. Continuous and non-continuous sets of measurements It is known that the number of measurements which should be carried out is conditioned by the variability of the quantity measured. All the dose distribution parameters depend to a higher or lesser degree on the extreme values, i.e. the extremely low and high values of a given series of measurements. It is not infrequent that the number of measurements performed is a compromise between the technical capacity to collect the necessary data and the expected accuracy of dose assessment. In the case when the data distribution is skewed and the results in the tail of the distribution are relatively rare, the number of measurements should be the most sensitive point in the calculation of the distribution parameters. This kind of distribution describes the radon progeny concentration in a mine, for example. Figure 1 presents the original data from the mine and two distribution curves: based on the original data and when 10% of the highest data are omitted. The answer to the question of what is the probability, P , that at least one result lies outside β percentile (assuming a continuous distribution of quantity and k number of measurements carried out) can be expressed as follows: P = 1 − βk.

(1)

If one considers the case where the distribution of quantity is discrete and only N values may be obtained at random, the probability P =1−

(βN )! (N − k)! , (βN − k)! N !

(2)

where: β deﬁnes the percentile, k is the number of measurements (k N ). In other words, P means the probability that at least one result of k measurements will lie outside β percentile when only N measurements can be performed. The probability curves for β = 0.90 and 0.95 are presented in Figs. 2 and 3, respectively. Let us assume that the underground workforce of a mine contains 1000 miners (N = 1000) and that only one measurement per miner could be carried out. The probability that at least one result of measurement lies outside the 90th percentile (β = 0.90) when 10 (k = 10) measurements were taken equals 0.65, but when 29

Fig. 1. Original data on RnDp concentration and two density curves: plotted for the original data and the 10% cut-off data.

790

D. Kluszczy´nski

Fig. 2. The probability that at least one result of measurement lies outside the 90th percentile (β = 0.90) of N -values’ distribution.

Fig. 3. The probability that at least one result of measurement lies outside the 95th percentile (β = 0.95) of N -values’ distribution.

(k = 29) measurements will be taken, the probability rise to 0.95. Assuming β = 0.95, the probabilities P will equal 0.40 and 0.78, respectively.

3. Annual exposure (dose) assessment It is extremely difﬁcult to determine the physical parameters of an underground mine owing to its intricate internal structure and the mileage of the underground tunnels. This requires modelling of all the values dependent on the ventilation, including annual miners’ exposure to radon progeny. The best way to solve this problem seems to be the application of the stochastic method. To do this, several assumptions have to be made, namely: (a) the mine is homogeneous with regard to the potential radiation hazard as it can be divided into sectors for which this condition is true;

Idea of assessment of the annual dose when non-continuous data exist

791

(b) the average concentration of the alpha potential energy of radon progeny during the measurement time is assumed to be the quantity measured; (c) all of the measurements are independent of one another; (d) the distribution parameters of radon progeny concentration in the mine are available; (e) for each miner, k measurements per year are carried out. The assessment of annual exposure based on a continuous series of measurements requires m consecutive measurements according to the formula: ty (3) , t where ty is the duration of the miner’s work within one year and t is the period of a single measurement. The average concentrations of the potential energy ci (i stands for the ith measurement for a miner) are available and the annual exposure, e, can be calculated from m=

e=

m

(4)

ci t.

i=1

If continuous measurements of exposure were carried out for all the miners, it would be possible to evaluate the distribution of annual miners’ exposure, e. The expected value, E(e), and variance, E 2 (e), can be associated with the following parameters of the measurements: expected value, E(c), and variance, D 2 (c), of the distribution of radon progeny concentration, c: E(e) = ty E(c) and D 2 (e) =

(5)

ty2

D 2 (c). (6) m If for each miner, k measurements were made (k < m), non-continuous measurement conditions would be provided. In this case, the annual exposure for a single miner, ep , can be deﬁned according to the following formulae: def ty

ep =

k

k

ci

(7)

i=1

and next E(ep ) = ty E(c) and D 2 (ep ) =

ty2

D 2 (c). k A comparison of equations (5), (8) and (6), (9) results in the following:

(8) (9)

E(ep ) = E(e) and

(10)

D 2 (e) = δD 2 (ep ),

(11)

792

D. Kluszczy´nski

where def

δ=

k . m

(12)

From these equations the following conclusions can be drawn: (a) the expected value of annual exposure based on non-continuous measurements, ep , equals the expected value of annual exposure based on continuous measurements, e; (b) the variances of distribution of both the exposures (e and ep ) are proportional and the measurement density factor, δ, is the measure of that proportionality.

4. Experimental data The δ factor is dependent on the actual and theoretical time of the measurements, which are or can be carried out for one worker during a year. Let us assume that a single measurement was made for each of the miners. The time of the measurement is 5 min, as for example in grab sampling. In this case δ equals 4.2 × 10−5 (m = 24 000, k = 1). Let 99% of the measurements lie between 0.01 WL and 0.5 WL [1]. It is possible to calculate the distribution of ep exposure with an expected value E(ep ) = 1.12 WLM and variance D 2 (ep ) = 0.96 WLM2 . The distribution of expected annual exposures can be found using the measurement density factor. The expected value, E(e), and variance, D 2 (e), of this distribution would be 1.12 WLM and 3.6 × 10−5 WLM2 , respectively. In the case of a log-normal distribution it would mean that 99% of the measurement results are in the interval between 1.10 and 1.14 WLM, which is inconsistent with the values of the annual exposure distribution. Considering the actual distribution of radon progeny concentration and the annual exposure distribution, the δ factor should amount to 0.03 to 0.05. This range of values differs by a magnitude of one thousand from the theoretical value (4.2 × 10−5 ). The reason for that difference is probably the variation of the concentration during the period of measurement. If the concentrations that are found in the mines are assumed to be constant or nearly constant during a given period of time (e.g. two weeks), reducing the interval between consecutive measurements to less than two weeks does not seem to be effective. It has no inﬂuence on the estimated expected value, E(c), and the variance, D 2 (c). Allowing a two-week period for the possible variation of the concentration, the δ factor assumes a value in the order of 0.03, which is of the same magnitude as the theoretical value for a one-month period of dosimetric measurements (δ = 0.08). The statistical analysis of personal dosimetry measurements carried out in Polish metal-ore mines in 1977–1987 [2,3] reveals considerable changes of δ factor during that period. The dosimeters on miners’ helmets are changed once a month. In this case, the δ factor should reach the value of 0.08, but Fig. 4 demonstrates that δ increases during the years from a value of about 0.08 to a stabilised value of about 0.3 (in 1985–1987 the mean value of δ was 0.32 for zinc–lead ore mines and 0.30 for copper ore mines). The empirical value of δ indicates that there is another period of ﬂuctuations, which is correlated with the seasonal changes.

Idea of assessment of the annual dose when non-continuous data exist

793

Fig. 4. Empirical δ factor calculated for metal-ore mines.

5. Conclusions 1. When the distribution of the observed quantity, as, e.g., radon progeny concentration in an underground mine, is signiﬁcantly skewed, an insufﬁcient number of measurements can determine the recognisable error level of the annual dose assessment. To ensure the probability of 0.95 that among the measurements at least one measurement belongs to the set of the highest available results (β = 0.95) one must carry out at least 57 measurements in a mine with a workforce of 1000 miners. 2. The method discussed in the paper makes it possible to directly calculate the distribution of miners’ annual exposures on the basis of non-continuous measurements or grab sampling results. The connector between these two sets of results is the measurement density factor, δ. 3. Under actual conditions of a given mine, the δ factor equals at least 0.03–0.05. This would indicate that the main contribution to annual exposure value are the two-week ﬂuctuations rather than short-term variations of the concentration of potential energy in the mine’s atmosphere. On the other hand, another type of ﬂuctuation, a seasonal one, may be involved. An analysis of the one-month long measurements points out that the δ factor equals approximately 0.3 instead of the theoretical value of 0.08. If one considers only the seasonal ﬂuctuations, the value of the δ factor should be 0.25 (only for a four-season pattern).

References [1] T. Domanski, D. Kluszczy´nski, J. Olszewski, W. Chruscielewski, Field monitoring versus individual miner dosimetry of radon daughter products in mines, Pol. J. Occup. Med. 2 (2) (1989). [2] W. Chruscielewski, T. Domanski, J. Olszewski, D. Kluszczynski, A. Zorawski, Radiation Exposure of Workers in Coal Mines as Well as in Lead, Copper, Zinc and Chemical Raw Material Mines, in: Stud. Mat. Monogr., vol. 29, 1988 (in Polish). [3] T. Domanski, W. Chruscielewski, D. Kluszczy´nski, J. Olszewski, Radiological classiﬁcation of Polish underground mines and recommendations for surveillance, Pol. J. Occup. Med. 4 (3) (1991).

794

A passive technique to measure radon progeny surface deposition variations in rooms and chambers J.P. McLaughlin, C. Walsh∗ Department of Experimental Physics, University College Dublin, Dublin 4, Ireland

When the alpha track detectors CR-39 and LR-115 are simultaneously exposed to air containing radon and its progeny, the CR-39 will record alpha activity from both radon progeny deposited on its surface and also from radon gas and its progeny in the air above its surface. The LR-115 will only record tracks from alpha particles emitted from radon and its progeny in a restricted volume of air above its surface because of its narrower alpha energy range response compared to that of CR-39. For the same reason, LR-115 will not record tracks from the 6.00 MeV and 7.68 MeV alpha particles emitted by progeny deposited on its surface as these energies are above its registration threshold. This latter characteristic of LR-115 is exploited in the work described here. A series of laboratory exposures were carried out in which it was possible to separately determine the relative responses of both CR-39 and LR-115 to the three alpha emitters 222 Rn, 218 Po and 214 Po in the air. It is then possible to obtain a value for the alpha track density contribution due only to progeny deposited on its surface by using the LR-115 track density and the three relative response ratios to subtract the contributions of the airborne activities from the total track density on the CR-39 detector. This technique was applied in some Dublin dwellings and the results are reported here. With the increased current interest in the use of surface trap measurements of 210 Po on glass as a means of retrospectively assessing radon exposure in epidemiological studies, the availability of the technique presented here may prove to be both timely and useful in improving the accuracy of such retrospective radon exposure assessments.

1. Introduction A detailed understanding of the behaviour of radon decay products in the indoor environment is needed for a number of reasons, not least of which is the need to make improved * Present address: NRPB (National Radiological Protection Board), Chilton, Oxon, OX11 0RQ, UK.

RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07099-8

© 2005 Elsevier Ltd. All rights reserved.

A passive technique to measure radon progeny surface deposition variations in rooms and chambers

795

assessments of human exposure to these species in radon residential epidemiological studies. In particular, a number of such epidemiological studies are presently using the measurement of the alpha recoil implanted surface activities of 210 Po in glass, the so-called surface trap technique, to make retrospective exposure assessments of subjects to radon and its progeny [1–3]. The surface activity of 210 Po on a selected suitable glass object is primarily determined by the deposition ﬂux of short-lived radon progeny onto its surface. Due to local variations in ventilation, airﬂow patterns and aerosol characteristics, etc., such deposition will not be uniform in a typical room. The measured 210 Po surface activity on any selected glass object cannot therefore be taken as representative of the mean radon progeny distribution in the room. This has signiﬁcant implications for the estimation of historical values of radon in a room based on such surface trap measurements. It is therefore necessary to have a simple and accurate means of measuring radon progeny deposition variations in the room to enable a correction to be applied to any individual glass measurement made. In this paper, a simple, reliable and accurate passive technique using two types of detectors is described which will make it possible, by measuring surface deposited radon progeny, to apply such a correction to the retrospective assessment of human exposure to radon and its progeny. The two alpha track detectors used are CR-39 and LR-115 which are mounted side-by-side and simultaneously exposed in an open-face conﬁguration to air containing radon and its progeny. The main problem faced in using open face alpha detectors, of any type, to measure deposited radon progeny is to discriminate the detector signal due to airborne activities from that due to deposited activities on its surface [4]. This problem is overcome in this work by exploiting the quite different track registration characteristics of the two detector types. The alpha energy registration range of CR-39 extends from about 0.1 MeV to energies well above the maximum alpha energy (7.68 MeV from 214 Po) of the 222 Rn series. Consequently, CR-39 records tracks due to (alpha emitting) radon progeny deposited on its surface and those due to radon gas and its (alpha emitting) progeny within alpha range in the airspace above its surface. The LR-115 has a much narrower alpha registration range (approximately from 1.2 to 4.2 MeV) and only records tracks from alpha emitters within an air volume above its surface from which emitted alphas will strike its surface with their reduced energies falling within its alpha registration energy range. Arising from these considerations, it is clear that when CR-39 and LR-115 are simultaneously exposed to air, the CR-39 alpha track density will consist of components due to surface deposited activities and components due to airborne activities while the LR-115 track density will be due only due to airborne activities. The next section describes experimental laboratory investigations by which the relative sensitivities of the two track detectors to the individual airborne alpha emitters 222 Rn, 218 Po and 214 Po were determined. These laboratory results make it possible to determine the value of the CR-39 track density component signal due only to deposited radon progeny.

2. Laboratory calibrations of detectors The three airborne alpha emitters of interest here and their corresponding energies involved are: 222 Rn (5.49 MeV), 218 Po (6.00 MeV) and 214 Po (7.68 MeV). The experimental objective was to separately determine the relative sensitivities (S222 , S218 and S214 ) of the two detector types to each of these radionuclides in the airborne state.

796

J.P. McLaughlin, C. Walsh

2.1. Radon gas sensitivity, S222 The relative sensitivity of the detectors to radon gas was easily obtained by pumping air through a strong radium chloride source and then through an absolute ﬁlter into a cylindrical chamber (radius = 15 cm, length = 30 cm) in which the detectors were mounted. The ﬁltered progeny free air containing radon gas at high concentrations in the range 5.6 to 7 MBq −3 , passed over the detectors which were mounted 10 cm downstream from the ﬁlter. The exposure position chosen in the cylinder ensured that both alpha detectors were out of range of any progeny that might deposit on the walls of the cylinder. At the air speed of 12 cm s−1 the notional transit time of the air from the ﬁlter to the detectors was approximately 1.2 seconds. While theoretically the ingrowth of 218 Po activity in 1.2 seconds is about 0.5% of the radon gas activity the level determined experimentally, by alpha spectrometry of Millipore ﬁlter sampling of chamber air, was found to be ∼ 2% of the radon level. This is somewhat greater than the simple theoretical value and presumably is due to airﬂow mixing patterns in the chamber. The contribution of this concentration of progeny to the dual detector alpha track densities, both directly and due to plate-out on the CR-39, can be shown to be negligible and in any case can be corrected for on the basis of the subsequent sensitivity measurements for 218 Po and 214 Po. Essentially, in this radon sensitivity test procedure the dual detectors were exposed to a pure radon source with minimal interference from progeny. A total of 70 sets of dual detectors were exposed to radon in the manner described and this yielded a mean CR-39/LR-115 relative sensitivity to radon gas of S222 = 3.24. In practical terms therefore if the two detectors are exposed to pure radon gas, for every 100 tracks counted on the LR-115 detector, 324 tracks will be counted on the CR-39. It should be noted that the determined sensitivity ratio of S222 = 3.24 for radon gas, and those given below for 218 Po and 214 Po, are speciﬁc to the etching and track identiﬁcation and counting procedures used. Other detector processing will yield somewhat different ratio values. 2.2.

218 Po

and 214 Po sensitivities, S218 and S214

It is not possible to produce separate airborne concentrations of the two radon progeny 218 Po and 214 Po. In air containing radon, both species will exist with their relative activity concentrations determined by such factors as ventilation, ambient aerosol characteristics, surface deposition and other removal processes, etc. In this work, the sensitivities of CR-39 and LR115 to airborne concentrations of these two alpha emitting radon progeny were separately determined by means of physical simulations of pure 218 Po and pure 214 Po in the airborne state. This was accomplished as follows: A circular stainless steel plate (diameter 11.5 cm, thickness 1.2 cm) was placed inside a 3.5 litre vacuum tight steel chamber into which ﬁltered air containing radon was introduced. The plate was fashioned to have on one side a machined insert which was a removable smaller stainless steel disc (diameter 25 mm, thickness 2.5 mm). The purpose of this removable disc is described below. By the application of a negative potential of 10 kV between the plate and the grounded wall of the chamber an electric ﬁeld of sufﬁcient strength was established to collect the mostly positively charged 218 Po atoms as they were formed from the decay of radon gas. A typical activity collection time of 3 minutes was used at which time the activity on the plate was predominantly 218 Po (T1/2 = 3.05 min). Following the removal of the plate from the

A passive technique to measure radon progeny surface deposition variations in rooms and chambers

797

Fig. 1. 218 Po and 214 Po growth and decay for 3 min collection time.

Fig. 2. Radon progeny exposure apparatus.

collection chamber, according to the laws of radioactive growth and decay, the 218 Po activity decays and the ingrown 214 Po activity increases (see Fig. 1). The collected activity on the plate (after a short collection time such as 3 minutes) can be seen for increasing post-collection times to be essentially an ultra-thin source of 218 Po which changes to a mixed source of 218 Po and 214 Po and ﬁnally to a pure source of 214 Po. To physically simulate pure airborne sources of 218 Po or 214 Po, the plate was mounted in the specially designed apparatus shown in Fig. 2. In this apparatus, the radioactive stainless steel plate is mounted as shown and a piece each of CR-39 and LR-115 are mounted side by side (i.e. in Dual Detector mode) on a moving platform below the plate. This moving platform is driven by a reciprocating stepping motor with a stepping pulse rate of 125−1 . By means of two microswitches the movement of the detectors on the platform is controlled to be a vertical reciprocating movement with a distance range from the stationary radioactive plate from almost in contact (< 0.1 cm) to 8.0 cm which places the detectors just beyond the range of the 7.68 MeV alphas from 214 Po. The cycle time

798

J.P. McLaughlin, C. Walsh

of the motion was 8 seconds, which is small compared to the half-lives of the radionuclides on the plate. Taking the moving detector platform as the frame of reference, this relative motion of the source and the detectors is the close physical equivalent of exposing the detectors to alpha particles from airborne sources of 218 Po and 214 Po. If (as illustrated in Fig. 1) an electrostatic collection time of 3 minutes was used, and by choosing a series of appropriate post-collection exposure times, the stainless steel source may be used in the exposure apparatus as an almost pure source of 218 Po or as a pure source of 214 Po or as a mixture of both. As example, if the detectors are exposed to the source during the post-collection time interval 6–8 minutes then the mean source activities are 93.4% 218 Po and 6.6 % 214 Po. On the other hand, if the exposure takes place in the post-collection time interval 24–26 minutes, the percentages of these radionuclides are almost reversed. The relative percentages of the two alpha emitters present during any detector exposures were determined experimentally. This was achieved by removing the small metal disc insert from the back of the active disc at the end of the collection period. It was then immediately placed in a PIPS (passive implanted planar silicon) alpha spectrometer at the same time as the main activity disc source was being placed in the exposure apparatus. It should be noted that while the total activities on the two discs are different, the relative percentages of activities of 218 Po and 214 Po at any post collection time on the small disc is the same as on the large disc. Because of this, experimental procedures were arranged so that alpha spectrometry of the small disc was synchronised with the exposure period of the CR-39 and LR-115 detectors to the large disc. Thus a real time measurement of the alpha activities being emitted from the small disc took place during the actual time of the alpha exposure of the dual detectors. The two integrated counts of the sharply resolved 6.00 and 7.68 MeV peaks of the alpha spectrum were recorded. These integrated counts are directly proportional to the number of alpha particles emitted by each of 218 Po and 214 Po during the exposures of the dual detectors. A series of exposures were carried out for different post collection time intervals. The resulting CR-39/LR-115 or dual detector track density

Fig. 3. Dual detector ratio values as a function of 218 Po and 214 Po levels.

A passive technique to measure radon progeny surface deposition variations in rooms and chambers

799

ratios were then plotted against the percentages of the alpha particles emitted 218 Po and 214 Po during each exposure as determined on the basis of the alpha spectroscopy of the small disc source. This is shown in Fig. 3 where by extrapolation the best ﬁt straight line to each of its extremities (100% 218 Po and 100% 214 Po) yields the required values of S218 = 4.1 and S214 = 7.2, respectively. 2.3. Results and discussion The results of these laboratory sensitivity measurements are given in Table 1. When the dual detector conﬁguration is exposed to room air, the resulting airborne contributions to the track densities on both track detectors will depend on the actual airborne concentrations of the three alpha emitters in the room air. On the basis of the PAEC (potential alpha energy concentration) of the airborne progeny, it is possible to calculate Ceq , the equilibrium equivalent concentration of radon. Room air with a radon gas concentration C0 can then be characterised by its F or equilibrium factor where F = Ceq /C0 [5]. In typical dwellings, F factors generally range from about 0.3 to 0.6 with a mean value close to 0.4. Using the data in Table 1 for the detector sensitivities to the three separate airborne alpha emitters, it is possible to calculate SF the overall CR-39/LR-115 track density sensitivity ratio as a function of F from 0 to 1. Figure 4 shows the variation of SF over a range of F values from 0.2 to 0.7. On the basis of the determined values of S222 , S218 and S214 , the component of the CR39 track density due to deposited radon progeny can then be determined using the following Table 1 Experimental sensitivity values for 222 Rn, 218 Po and 214 Po Radionuclide

Alpha energy (MeV)

222 Rn

5.49 6.00 7.68

218 Po 214 Po

Sensitivity ratio SI 3.24 4.1 7.2

Fig. 4. Dependence of SF on the equilibrium factor F .

800

J.P. McLaughlin, C. Walsh

relationship: DRnP = CR − SF · LR, where: • • • •

DRnP = alpha track density on CR-39 due to deposited radon progeny, CR = total track density on CR-39 (due to deposited plus airborne activities), LR = track density on LR-115 (due only to airborne activities), SF = overall CR-39/LR-115 sensitivity ratio.

In this work, F = 0.4 is taken as a typical indoor mean value and its corresponding value of SF = 3.8 was used in all determinations of DRnP . Using an indoor F value range from 0.3 to 0.6, the corresponding range in SF values is from 3.6 to 4.1 which correspond respectively to deviations of 5.5 to 8% about the chosen value of SF = 3.8 (F = 0.4). It should be noted that, due to its restricted alpha track registration characteristics and its consequent poor geometrical detection efﬁciency for airborne alpha emitters, the LR-115 track densities found in dual detector exposures in normal dwellings were found to be typically less than 5% that of CR-39. Therefore the calculated DRnP value is not very sensitive to the F values and the aforementioned uncertainties in the chosen SF = 3.8 value translate into small uncertainties in the derived values for DRnP . These are negligible in comparison to other uncertainties such as those arising from alpha track counting statistics which were typically about 5%.

3. Measurements in dwellings To test the dual detector technique, measurements of the variability of radon progeny deposition on room surfaces were carried out in 20 living rooms and bedrooms in 10 normal single family dwellings in Dublin. In each room, 4 dual detectors (CR-39 and LR-115) were mounted; one on each wall. In each case the chosen wall mounting location was close to its geometric centre. The duration of the exposures ranged between 3 and 7 months. Radon gas concentrations, measured by passive SSI type diffusion alpha track radon detectors, during the duration of the dual detector exposures, were found to range from 22 to 164 Bq m−3 . A total of 80 radon progeny wall deposition determinations and 20 radon gas determinations were made [6]. It was convenient for comparison purposes to normalise the deposited surface activities to unit radon exposure (Bq m−3 s−1 ) by expressing them in units of tracks m−2 Bq−1 m3 s−1 . The results are summarised in Fig. 5. Here the walls on which the detectors were placed are coded as follows: WD (wall with door), WW (Wall with window) and W1 , W2 (the other two walls). The deposition of short-lived radon progeny on wall surfaces in the rooms was clearly not uniform. For most of the rooms surveyed, the standard deviation about the mean was within the range 9 to 40%, but for two of the twenty rooms (Nos. 10 and 12) the standard deviations in the measured surface deposition was quite high. In room No. 10, no obvious reason could be found or postulated for enhanced deposition on one wall. In the case of room No. 12 the high standard deviation (94%) was due primarily to the relatively high value of deposited activity on one of the walls. This high deposited activity was found to be due to the high deposition measured at a point on a wall above a concealed 240 V AC electricity cable. It is

A passive technique to measure radon progeny surface deposition variations in rooms and chambers

801

Fig. 5. Measured radon progeny deposition on walls of Dublin dwellings.

considered likely that electric ﬁelds from this cable were responsible as enhanced depositions of radon progeny due to both DC and AC electric ﬁeld gradients are known phenomena [7]. In this context, it should be noted that before ﬁeld use all dual detectors were treated with an anti-static ﬂuid to avoid any electrostatic effects on radon progeny deposition that might occur if the detectors became charged. These results are quite limited and thus should only be considered as indicative of the variability in radon progeny deposition to be expected in normal dwellings. They do, however, suggest that the surface activities of the long-lived radon decay product 210 Pb and its descendant 210 Po which derive from the deposition of the short-lived progeny 218 Po and 214 Po will also display variations of a similar magnitude. The consequential uncertainty introduced in retrospective radon exposure assessment based on measurements of 210 Po is unavoidable but the present work suggests that it is generally likely to be less than 40% if the chosen glass objects are not in obvious locations of either enhanced (near electric ﬁeld gradients) or reduced deposition (in plate-out shadow areas under shelves, etc). On the basis of the work described here, it is to be recommended that open face dual detectors should be used in conjunction with surface trap detectors as a means of quantifying the uncertainty associated with an individual surface trap measurement due to radon progeny deposition variations within a room. When a glass object is chosen in a room for the purpose of making a surface trap measurement, an open face dual detector should also be mounted on the glass object. In addition, a minimum of three other open face dual detectors should be widely dispersed on other room surfaces. The variability of radon progeny deposition measured by these dual detectors can then be used to adjust or correct the measured 210 Po surface activity on the chosen glass object to a value more representative of the mean behaviour of

802

J.P. McLaughlin, C. Walsh

radon progeny in the room. This in turn should help to improve the reliability of the estimated retrospective radon exposure based on the surface trap measurement. It is fully recognised that contemporary radon gas concentrations in a dwelling may differ signiﬁcantly from those in the past but it is not unreasonable to assume that deposition patterns (as distinct from absolute deposition ﬂuxes) of its progeny, in particular in a bedroom, should remain relatively constant over time. Notwithstanding the uncertainties in these considerations and assumptions, the suggested use of dual detectors will be an improvement on the present situation where 210 Po measured on a single chosen glass object in a room is, in effect, taken as a representative surrogate for the mean behaviour of radon progeny in the room. In the case of a residential radon epidemiological study using surface traps, the additional cost of using dual detectors, in the manner outlined above, should cause only a small increase in the cost of the study. It may however, introduce an increase in the level of intrusiveness to study subjects. Nevertheless, it should be considered seriously as a means of helping reduce uncertainty in retrospective radon exposure and risk assessment.

4. Summary and conclusions A new and simple technique to measure the variability of radon progeny deposition onto room surfaces, using a dual detector conﬁguration of CR-39 and LR-115, is described. While this technique may ﬁnd use in basic studies of radon progeny behaviour in enclosed spaces and exposure chambers, its main application may prove to be in the retrospective assessment of radon exposure using surface trap detectors. Here the dual detector technique can be used to determine the variability of radon progeny deposition in rooms so that corrections may be applied to the glass surface trap measurements. In surface trap ﬁeldwork, the biggest cost is that of survey personnel visiting dwellings. The additional cost involved in using the dual detector technique described here will be small in comparison but should considerably improve the quality of the estimated radon exposure data.

Acknowledgements The authors acknowledge the ﬁnancial assistance to this work received from the Commission of the European Communities under research contracts F14P-CT95-0025 and F14P-CT960065.

References [1] [2] [3] [4] [5] [6]

C. Samuelsson, Nature 334 (1988) 338. J.P. McLaughlin, Radiat. Prot. Dosim. 78 (1998) 1. M. Alavanja, J. Lubin, J. Mahaffey, R. Brownson, Am. J. Public Health 89 (7) (1999) 1042. B. Dörschel, E. Piesch, Radiat. Prot. Dosim. 48 (1993) 145. ICRP Publication 65: Protection against radon-222 at home and at work, Ann. ICRP 23 (2) (1993). J.P. McLaughlin, K. Kelleher, K. Scullion, in: Proc 5th Int. Conf. on High Levels of Natural Radiation and Radon Areas, vol. II, Munich, September 2000, in: BfS Schriften, March 2002, p. 444, ISBN 3-89701-808-X. [7] D.L. Henshaw, A. Ross, A.P. Fews, W.A. Preece, Int. J. Radiat. Biol. 69 (1996) 25.

803

Radon on selected underground tourist routes in Poland J. Olszewski, W. Chru´scielewski, J. Kacprzyk, D. Kluszczy´nski, Z. Kami´nski Department of Radiological Protection, Nofer Institute of Occupational Medicine, PO Box 199, 90-950 Lodz, Poland

The underground tourist routes, such as caves, old mines, building constructions under old towns and dungeons are the environments with potential radon concentrations that may exceed several thousand times the atmospheric concentrations. In Poland there are over thirty underground tourist routes, but radon concentrations are measured only in few of them. Among other routes, the following should be singled out: the drift of the former Kowary uranium mine and Niedzwiedzia Cave located near Kletno (the Sudetes). In the former, the mean annual radon concentration is of the order of 500 Bq m−3 , and in the latter it reaches over 2000 Bq m−3 . 1. Introduction Radioactive elements common in nature are responsible for the exposure of the whole population to ionising radiation that originates amongst others from radon. Mean atmospheric concentrations of radon throughout Poland are 4.4 Bq m−3 [1]. The International Agency for Research on Cancer (IARC) has categorised the radioactive gas radon with respect to its potential carcinogenic risk into group 1 as being carcinogenic to humans [2] and, according to the Council Directive 96/29/Euratom, it is covered by the system of radiological protection [3]. Article 23 of the Atomic Law, binding in Poland until 2002, provided that if an occupational activity associated with the presence of natural radiation leads, from the perspective of radiological protection, to increased exposure of workers or the general population, then such an exposure must remain under regular assessment. This paper lists caves and other underground routes as workplaces which require to be assessed for workers’ exposures. Underground tourist routes, such as caves, old mines, building constructions under old towns and dungeons are environments with potential radon concentrations that may exceed several thousand times the atmospheric concentrations. In Poland there are over thirty underground tourist routes, but radon measurements are carried out only in a few of them. Among other routes, the following should be singled out: RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07100-1

© 2005 Elsevier Ltd. All rights reserved.

804

J. Olszewski et al.

the drift of the former Kowary uranium mine and the Niedzwiedzia Cave located near Kletno (the Sudetes). A tourist route of 1200 m in the drift of the former Kowary uranium mine was opened to the public in 2000. The tour along the route takes about one hour. The entrance to Niedzwiedzia Cave was disclosed in 1966, but the Cave opened its door wide for tourists in 1983. In all, the passages of the cave are 2500 m long, including about 400 m of the tourist route; it takes about 45 min to cover it.

2. Methods The surveillance of radon exposure has been carried out by the Nofer Institute of Occupational Medicine, Lodz. The measurements of grab sampling for radon concentrations are taken using RDA 200 scintillation chambers (Pylon Electronic Development Company, Canada). Mean long-term radon concentrations are measured by means of NRPB passive integrating dosimeters with track detector CR-39 (Pershore) or CC-1 dosimeters with detector LR-115 (Kodak Pathe). The dosimeters are placed at the monitoring locations for a period of 1–3 months. The NRPB dosimeters are used to measure the radon exposure of workers serving the “Kowary Drift” tourist route. Calibration of all the methods used for radon concentration measurements are based on the Amersham and Pylon radon standards.

3. Results 3.1. The tourist route in the “Kowary Drift” A thorough assessment of risk connected with radon exposure preceded the opening of the “Kowary Drift” to the public. The radon concentrations were within the range 49.6 to 990 Bq m−3 . Mean annual radon concentrations were 400 ± 250 Bq m−3 . Since the opening of the tourist tour, a constant assessment of the environment has been performed by experimentally measuring radon concentrations on a periodic basis. The tourist service personnel are covered by individual dosimetric surveillance. The latest measurements of radon concentrations were taken at six monitoring points along the tourist route in September 2001. The concentrations varied from 210 to 800 Bq m−3 , and the mean concentration was 380 ± 220 Bq m−3 . The measurements of periodic concentrations are carried out at two monitoring locations on the tourist route. The results of the measurements are summarised in Fig. 1. The results of the study carried out in 2001 revealed that the periodic radon concentrations in the drift ranged from 340 to 690 Bq m−3 , whereas the mean annual concentrations were within the range of 580 ± 110 Bq m−3 . In 2001, 12 employees (guides and tourist service personnel) were under individual exposure surveillance. Because of rotation of personnel, 32 measurements of mean quarterly radon concentrations were taken in 2002. Based on these measurements, the expected annual distribution of individual radon doses was estimated. The distribution is presented in Fig. 2. The mean annual dose was estimated at 0.78 mSv.

Radon on selected underground tourist routes in Poland

805

Fig. 1. The measurements of periodical mean radon concentrations along the “Kowary Drift” tourist route.

Fig. 2. The distribution of radon annual doses received by persons employed on the tourist route.

3.2. The tourist route in Niedzwiedzia Cave In Niedzwiedzia Cave, measurements have been performed since 1995. Mean annual radon concentrations in the cave are higher than those in the drift. In 1995, mean concentrations were 1080 Bq m−3 . During subsequent years, increased radon concentrations were observed, reaching a value of 2000 Bq m−3 in 2001 [4–6]. Fig. 3 presents the results of mean annual

Fig. 3. The results of the mean annual measurements of radon concentrations in Niedzwiedzia Cave.

806

J. Olszewski et al.

measurements of radon concentrations in Niedzwiedzia Cave. In this cave, measurements of the individual radon exposures of the tourist service personnel have not been carried out. Having analysed the available data, the expected dose received by the personnel and tourists can only be estimated. Assuming 500 h of work at a mean concentration of 2500 Bq m−3 , the expected dose to the personnel could be estimated at the level of 3 mSv, and for a tourist who spends about 1 h in the Cave it could amount to 4 μSv.

4. Conclusion The results presented here should be analysed from two perspectives – the tourist service personnel and tourists themselves. Persons employed in the tourist service should be regarded as those occupationally exposed to ionising radiation and as such their exposure has to be controlled. The exposure of employees working in the “Kowary Drift” amounts to about 5% of the recommended annual limit (20 mSv), whereas in Niedzwiedzia Cave this value increases to 15% of the limit. In none of these routes is there a potential to exceed the limit of annual dose. Tourists visiting both routes are exposed to a dose of the order of several μSv, which in comparison with the dose received by every individual during a year (2.4 mSv) is a negligible value, exerting no health effects on tourists. Nevertheless, persons visiting the tourist routes should be informed about the existing ionising radiation originating from the radon present in underground areas and about the dose they may receive during the time spent in the cave or drift.

References [1] Polish Radiological Atlas, Central Laboratory of Radiological Protection, The National Atomic Energy Agency, 1992. [2] International Agency for Research on Cancer (IRAC), Man-made Fibres and Radon, in: Monographs on the Evaluation of Carcinogenic Risk to Humans, vol. 43, IRAC, Lyon, 1988. [3] Council Directive 96/29/Euratom of May 13, 1996 laying down basic safety standards for the protection of health of workers and the general public against the dangers arising from ionising radiation, Ofﬁcial J. Eur. Commun. Ser. L 159 (29.6.1996). [4] T.A. Przylibski, The variability of radon air concentration in Niedzwiedzia Cave, Kletno (the Sudetes), Przeg. Geol. 44 (9) (1996) (in Polish). [5] T.A. Przylibski, Radon concentration changes in the air of two caves in Poland, J. Environ. Radioact. 45 (1999) 81–94. [6] T.A. Przylibski, Personal information (unpublished results).

807

Indoor radon in the karst region of Lithuania L. Pilkyt˙e a , G. Mork¯unas a , G. Åkerblom b a Radiation Protection Centre, Kalvariju˛ 153, LT-2042 Vilnius, Lithuania b Swedish Radiation Protection Authority, SE-17116 Stockholm, Sweden

The highest average indoor radon concentrations in Lithuania are found in the karst region. Radon measurements have been carried out in this region with electrets. Two electrets were used in each dwelling. The duration of the measurements was 3 to 4 weeks. The results are compared with the similar results measured in other regions of Lithuania. The effect of seasonal variations is discussed along with other factors which inﬂuence the indoor radon concentrations, with particular emphasis on indoor radon concentrations in the karst region.

1. Introduction Soil is the main source of radon in Lithuanian houses [1]. Since no elevated Ra-226 concentrations are found in surface soils, soil permeability plays an important role in transfer of radon indoors. Karst phenomena may play an important role in determining the magnitude of indoor radon concentrations [2]. The karst region in Lithuania forms a signiﬁcant proportion of the Lithuanian–Latvian karst region with a total area of 31 000 km2 . It covers up to 20.7% of Lithuanian territory [3]. 17.5% of the Lithuanian population live in this region. However, intensive surface karst phenomena are observed only in a small part of this region – up to 200 km2 . The region is situated in Northern Lithuania, and is the zone of sulphatic karst. The large number of karst pits distinguishes it. In some places, the total area of karst pits is as large as 30% of the territory of the area. The total density of karst pits is up to 64 pits per km2 [3]. The only measurements described here are those which have been carried out in the areas of high indoor radon concentrations. For the sake of brevity, this has been referred to as the karst region, even though this covers only a small part of the whole karst region. A survey of indoor radon concentrations was carried out in Lithuania from 1995 to 1998. Radon concentration measurements were carried out in 400 detached houses, selected at random. Measurements made during the cold season where heating systems in homes were, in general, ‘on’, indicated that the average indoor radon concentration was (55 ± 5) Bq m−3 RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07101-3

© 2005 Elsevier Ltd. All rights reserved.

808

L. Pilkyt˙e et al.

(95% conﬁdence intervals have been adopted) [4]. The parts of the karst region with average indoor radon concentrations exceeding the Lithuanian national average by more than 1.5 times have been identiﬁed. Soil conditions in such parts of the karst region exhibit improved permeability and hence indoor radon concentrations are higher here. The permeability of soil is more important for indoor radon concentrations than 226 Ra concentrations in the surface layer of soil. It has been difﬁcult to detect increased radon concentrations in the karst region during these measurements owing to the relatively small sample size of the measurements – in average only 1.3% of randomly selected houses were in this region. Separate measurements of indoor radon concentrations were carried out in 1995–1996 and in 2001–2002. Areas with karst pits were analysed, though measurements were also made in the areas adjacent to zones with karst pits because karst phenomena cover a much larger area than that with these visual indicators.

2. Materials and methods Standard E-PERMTM electrets (Rad Elec Inc.) were used for measurements [5]. The principle of the measurement techniques is based on the ability of negative ions to discharge an electret, a disk of teﬂonR that has an electrical charge. The electret is placed in the bottom of an ion chamber, made of electrically conductive plastic material. Radon diffuses into the chamber through a ﬁlter that prevents the radon daughters from entering the chamber. When the radon and the radon daughters formed in the chamber decay, the air inside the chamber is ionised. The negatively charged ions are drawn to the positively charged electret and each negative ion discharges the electret by a small amount. The radon concentration in the air, to which the electret is exposed, is determined by measurement of the disk voltage before and after the measurement. The drop in charge of the disk is dependent both on the average concentration of indoor radon and on the duration of the measurement. Two rooms were analysed which were almost permanently occupied and which were closest to the ground. The electret detectors were placed in such rooms in locations where they could not be moved during measurements. They were positioned not closer than 25 cm from the walls, avoiding heating devices, tap water sources, and areas designed for added ventilation. Since basements are not permanently occupied, measurements were not made there. The following information was recorded: name of the owner, address, main construction material, year of construction, number of ﬂoors, ﬂoor on which the measurement took place, destination of the investigated room, date and time of the beginning and end of the measurement, availability of water supply, basement and absorbed dose rate indoors. The duration of any measurement was not shorter than 3 weeks.

3. Results and discussion In 1995–1996, two sets of indoor radon concentration measurements were made. In the month of August, measurements were made in 26 houses and in February, measurements in 35

Indoor radon in the karst region of Lithuania

809

houses. The average concentration in February was (95 ± 20) Bq m−3 , in August (101 ± 27) Bq m−3 . The overall average was (98 ± 16) Bq m−3 . A statistical t-test showed that there was no difference in indoor radon concentrations between summer and winter months (p = 0.38). However, in summer larger differences (by up to 45%) were detected in indoor radon concentrations in separate rooms of the same house. The radon concentrations in the karst region were found to be higher than the national average (p < 0.01). The distribution of concentrations was found to closely resemble the log-normal distribution, which was found for the 400 randomly selected houses throughout Lithuania; however, a second peak in distribution was observed in the houses of the karst region. This has been attributed to what has been described in some parts of the houses of the investigated region. These houses have higher indoor radon concentrations, and this is suggested to be the reason for the second maximum. In July 2001–April 2002, the second set of measurements was carried out in 552 houses in the karst region and areas close to it. The average indoor radon concentration in these houses was (88 ± 7) Bq m−3 . It differs statistically signiﬁcantly (p < 0.01) from the average of concentrations in randomly selected detached houses. The distribution of concentrations received during this set of measurements is given in Fig. 1. A polynomial trend line of the 5th order is also shown in Fig. 1. It indicates the trend for the second maximum at approximately 250 Bq m−3 . As in the ﬁrst trial, it shows that some houses investigated exhibit karst phenomena, i.e. these phenomena have a strong inﬂuence on indoor radon concentrations.

Fig. 1. Distribution of radon concentrations (Bq m−3 ) in houses in the karst region and adjacent areas.

Fig. 2. Distribution of indoor radon concentrations (Bq m−3 ) in the area of the most intensive karst phenomena.

810

L. Pilkyt˙e et al.

The distribution of indoor radon concentrations in the area of the most intensive karst phenomena (distinguished by karst pits in the vicinity of houses) is given in Fig. 2. As seen from Fig. 2, the distribution of indoor radon concentrations in the area of the most intensive karst phenomena is comparable with that of the larger region without karst pits, though situated in the karst region. The distribution of indoor radon concentrations in the area of the most intensive karst phenomena was received from the results of measurements in 302 houses. The average radon concentration in these houses was (109 ± 9) Bq m−3 . It differs statistically signiﬁcantly from the average in the whole karst region (p < 0.01). It indicates that the karst phenomena have an important inﬂuence on indoor radon concentrations. In 149 houses measurements were carried out twice – in the warm and the cold months. The ﬁrst period was July–October (end of measurements), the second – October (start of measurements)–April. The average concentrations in the same houses were (87 ± 9) Bq m−3 in the warm period, (90 ± 13) Bq m−3 in the cold period. According to the t-test, the difference between these averages is statistically insigniﬁcant (p = 0.30). This indicates that the increased ventilation rate in summer does not decrease indoor radon concentrations. However, the differences of indoor radon concentrations in separate rooms are larger in the warm period than in the cold one. This is caused by larger differences in the ventilation of separate rooms in the warm period. Averages of the indoor radon concentrations in houses of different ages are given in Fig. 3. The statistical t-test shows that signiﬁcant differences exist between the radon concentrations in houses of up to 20 years old and those 30–40 years old. The older houses have the highest radon concentrations. Very similar results were received for randomly selected houses all around Lithuania. First of all, this shows that there are no differences between the construction of houses in the karst region and in Lithuania as a whole. On the other hand, it is evident that the improved thermal insulation and consequently decreased ventilation rate in new houses do not increase indoor radon concentrations or at least that an inﬂuence of improved thermal insulation has not been detected yet due to the small number of such houses constructed lately in the region. Such a construction feature as an engineered barrier between the interiors of house and soil is an important reason for the magnitude of indoor radon concentrations. More modern houses have a more signiﬁcant engineered barrier preventing transport of soil gas containing radon. The lower radon concentration in houses built 50 years ago and more may be explained by the fact that about 40 years ago the construction of houses was changed from relatively radon safe crawl spaces to basements. However, the basements constructed 30–40 years ago usually

Fig. 3. Average radon concentrations (Bq m−3 ) in houses of different age (years).

Indoor radon in the karst region of Lithuania

811

do not contain concrete ﬂoors; these began to be constructed 20 years ago. For this reason radon concentrations in older houses with basements are higher than in newer houses with basements with concrete ﬂoors. It should be pointed out that a basement alone is not a reason for lower indoor radon concentrations. The average radon concentration in houses with a basement of any kind is (65 ± 9) Bq m−3 , compared to (103 ± 8) Bq m−3 in houses without any basement. The difference is statistically signiﬁcant (p < 0.01). It should also be pointed out that in some cases the basement is constructed only under a part of the house. In any case, it is evident, that the existence of a barrier that prevents radon from entering into the house exerts a very important inﬂuence on the indoor radon concentrations. This appears to hold true if the barrier is not of any special type or even designed to prevent radon entry into the house. The radon concentrations are higher in houses without water supply, where water is carried indoors using buckets from wells outdoors, than in houses with a water supply (p < 0.05). The concentrations are (96 ± 10) and (84 ± 8) Bq m−3 , respectively. This indicates that water cannot be an important source of radon indoors. The difference can be explained by the fact that houses with a water supply are more recently constructed than the ones without water supply. They were constructed in (1974 ± 1) and (1951 ± 2), respectively. The difference in average ages of houses with and without a water supply is statistically signiﬁcant (p < 0.01). The averages of indoor radon concentrations in houses constructed of different construction materials (materials used for construction of the main part of houses) were also analysed. No statistically signiﬁcant differences were found between radon concentrations in houses constructed of wood, brick, concrete and wood with an outside layer of bricks. This indicates that the construction materials are not an important source of indoor radon in comparison with soil. This conclusion is supported by the fact that there is no correlation between the radon concentration and the ambient dose equivalent rate in separate rooms – the coefﬁcient of correlation is 0.15. This also indicates that the results of dose measurements cannot be used even for screening of indoor radon concentrations.

4. Conclusions The karst region, at least the area of intensive surface karst phenomena, is an important region from the point of view of indoor radon concentrations. In this region, the radon concentrations are twice as high as the national average. The age of a house, i.e., features of its design connected with age, are important with regard to the indoor radon concentrations. Lower indoor radon concentrations are observed in houses constructed less than 25 years ago. Houses older than 45 years also have lower indoor radon concentrations. The thickness of an engineered barrier, and the presence or absence of a basement or crawl space are the causes of large differences in radon concentrations indoors. This feature should be kept in mind when deciding what remedial measure is the optimum option. In the karst region, indoor radon concentrations are not affected directly by ventilation rate. No statistically signiﬁcant difference was found between the indoor radon concentrations in the same houses in the warm and cold periods of year.

812

L. Pilkyt˙e et al.

In general, the relationship between indoor radon concentrations and house construction parameters in the karst region is similar to the relationship in other regions of Lithuania. This indicates that the karst phenomena alone are the main reason for higher indoor radon concentrations. No special relationships have been shown between indoor radon concentrations and house construction characteristics or other parameters in this region.

References [1] G. Mork¯unas, Estimation of the effective dose due to indoor radon in the detached houses, Doctoral thesis, Institute of Physics, Radiation Protection Centre, Vilnius, 2000. [2] UNSCEAR (United Nations Scientiﬁc Committee on the Effects of Atomic Radiation), Sources and Effects of Ionising Radiation, UNSCEAR 2000 Report to the General Assembly, with Scientiﬁc Annexes, vol. 1: Sources, United Nations, New York, 2000. [3] The Northern Lithuanian Karst Region, Geographical Aspects of Nature Use, Institute of Geography, Vilnius, 2000. [4] B. Clavensjö, G. Åkerblom, G. Mork¯unas, Indoor Radon. Remedial Measures, Asveja, Vilnius, 1999. [5] E-PERM system manual, Rad Elec Inc., Frederick, USA, 1991.

813

The behaviour of Rn-222 decay products at the air–glass interface and its implication for retrospective radon exposure estimates B. Roos, C. Samuelsson Department of Radiation Physics, Lund University Hospital, Sweden

Glass sheets investigated regarding their 210 Po signal originating from indoor radon exposure will have a depth distribution down to about 100 nm. The geometrical probability of implantation into glass of glass-attached radon daughter activity in one alpha recoil event is 50%. A particulate layer on the glass sheet (dust, grease or other particles) will stop the alpha recoil nucleus partly or totally, and the alpha-recoiled 210 Po nucleus will be implanted into the glass surface at a lesser depth. Clean glass sheets were pre-exposed to different types of nonradioactive aerosol particles and then exposed to radon-laden air. After a build-up period, the surface activity of 210 Po was determined by alpha spectrometry. The implanted fraction of 210 Po for these sheets of glass, with respect to mass load and type of particulate surface layer, was determined to be between 30 and 80% depending on the mass load (μg cm−2 ).

1. Introduction In the late 1980s, implanted 210 Po in glass surfaces was found to give the historic radon concentration exposure in the environment of the glass [1]. Short-lived airborne 222 Rn decay products contaminate all surfaces in a radon affected dwelling. Subsequent to alpha decay, the recoil energy is sufﬁcient to implant the daughter nucleus in the underlying surface layer (Fig. 1). Alpha recoil implantation of long-lived decay products of 222 Rn into household glass objects has been utilised by many groups as a tool for estimating past radon levels [2–6]. These investigations have provided useful data for the estimation of the absorbed dose from radon daughters. Both [2,7] found a positive correlation between glass-based radon measurements and lung cancer incidence, but a weak or absence of correlation when contemporary radon gas measurements were used as exposure indicator. A common experience from studying the implanted activity of 210 Pb or 210 Po in glass sheets taken from radon-affected dwellings is that the mean residence time of the implanted RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07102-5

© 2005 Elsevier Ltd. All rights reserved.

814

B. Roos, C. Samuelsson

Fig. 1. Decay scheme of radon. Implanting nuclides are 214 Pb and 210 Pb from decaying 218 Po and 214 Po. The activity of 210 Po on the glass surface and that implanted in the glass were determined.

Fig. 2. Authentic household glass exposed to radon-laden air for several years [11, unpublished data]. Loss of 210 Po content in old glass implies an effective half-life of about 10 years. The dashed line is the expected 210 Po content in glass disregarding other losses.

nucleus seems to be much shorter than that expected from radioactive decay alone [8,9]. Similar results [11, unpublished data] have been found from household glass measurements in Sweden (Fig. 2). Careful experiments in radon chambers and dwellings have not been able to pinpoint any removal mechanism besides physical decay (Fig. 3). The absence of a removal mechanism fuels the hypothesis that the observed residence time is due to a decrease in the occurrence of implantation in the glass matrix with exposure time. A decrease in alpha recoil implantation in the glass matrix with time is plausible as the range of the recoiling nucleus is of the same order of or smaller than the linear dimensions of dust particles on surfaces in a dwelling (Fig. 4, a and b).

The behaviour of Rn-222 decay products at the air–glass interface

815

Fig. 3. Radon-exposed glass sheets (5 different glass sheets at different locations in a dwelling) measured at almost equal intervals for up to ten years show no loss of implanted activity.

(a)

(b)

Fig. 4. (a) Particle-attached 218 Po activity deposited on the glass surface and activity deposited on an absorbing layer. Both situations will absorb the energy of the recoiling nucleus, resulting in implantation in the absorbing layer or in shallower implantation of the nucleus in the glass. (b) Implantation paths from surface-deposited 214 Po. 214 Po can exist as surface-deposited activity on the particle (i) or on the glass surface (ii), as implanted activity in the deposited particle (iii) or already implanted in the glass matrix (not indicated in the ﬁgure).

816

B. Roos, C. Samuelsson

It is the aim of this contribution to report on our basic studies of alpha recoil implantation physics and on processes interfering with the implantation of nuclei into the glass matrix itself. In order to facilitate the interpretation of experimental results concerning the implantation probability, the strategy employed was to theoretically and experimentally make a clear distinction between unattached and attached progeny implantation. The experimental exposures were mainly performed in a walk-in radon-aerosol chamber or smaller containers inside this chamber. In experiments to study the inﬂuence of different impurities on the glass surface, authentic glass sheets from dwellings were also investigated. In order to achieve sufﬁcient sensitivity without having to use unrealistically high radon concentrations, large-area samples were used. The samples were analysed by alpha spectrometry in a pulse-ionisation chamber. Complete isolation of the effects of attached progeny implantation in clean glass surfaces is difﬁcult to obtain. In order to eliminate the strong inﬂuence of implantation caused by the decay of unattached 218 Po, the aerosol concentration must be very high, with the consequence that the exposed sample will be contaminated with the stable aerosol during exposure. All our experimental results, however, indicate a signiﬁcantly lower implantation probability for progenies attached to aerosol particles compared with unattached ones [10]. The effect on the implantation process of particulate impurities on the glass surface was quantiﬁed through experiments using glass sheets exposed in an experimental radon chamber. If, during exposure to short-lived decay products, the glass surface is contaminated with particles corresponding to a homogeneous surface layer of unit density and a thickness of 100 nm, the implantation probability of 210 Po atoms decreases roughly by 50%. The inﬂuence of dust on the implantation of the alpha-emitting recoil ions in glass has been investigated by [11], who measured the 210 Po signal in glass with and without a dust layer, but the amount of dust was not quantiﬁed. Later [12] has investigated how aerosol and airborne activity parameters inﬂuenced the alpha recoil implantation process.

2. Materials and methods Glass sheets with an area of 0.022 m2 were exposed to 45 kBq y m−3 (i.e. 2 weeks with 1 MBq m−3 radon-loaded air in an exposure box (1.5 m3 ) placed in the walk-in radon exposure chamber [13]. Normal, commercially available soda lime glass, manufactured by the ﬂoating glass process [14], was used. The clean glass sheets were exposed to different types of surface contamination of particulate matter (dust, smoke, salt, and candle burning particles). The glass sheets were weighed before and after exposure to aerosol particles, and again after exposure to the radon-laden air to quantify any possible particle loss. After 6 months of build-up the analysis of the superﬁcial 210 Po and implanted 210 Po concentration on glass commenced. The glass sheets were ﬁrst analysed with the particle layer intact in order to obtain the total 210 Po activity. Then, after thoroughly cleaning the surface with deionised water, soap and 10% hydrochloric acid, to remove all the particles from the surface and all surface-deposited 210 Po activity, the residual activity was determined. The glass sheets were analysed with an open-ﬂow pulse-ionisation chamber (Fig. 5) similar to the one described by [15].

The behaviour of Rn-222 decay products at the air–glass interface

817

Fig. 5. Pulse ionisation chamber similar to the one described by Johansson et al. [15].

The residual activity thus measured was deﬁned as the implanted fraction, i.e. the measured residual 210 Po activity (Bq m−2 ) divided by the corresponding activity measured before cleaning. In total, 114 glass sheets were investigated.

3. Results Glass sheets exposed to indoor environments will gather particulate matter on the surface of about 10–50 μg cm−2 in 6–12 months. The density of indoor aerosols is usually between 0.9 and 2.5 × 103 kg m−3 , and similar particle loads and densities were used in our experiments. The lower density is due to greasy, candle-burning and tobacco-smoke-related particles. The higher density is due to normal dust particles. In some laboratory experiments the latter particles were replaced by salt particles with a density of 2.2 × 103 kg m−3 . It can be clearly seen from Fig. 6 that the probability of implantation decreases with increasing mass load on the surface. For very clean surfaces, almost all the activity (90–50% for a mass load of 0–5 μg cm−2 ) is implanted. With increasing mass load, the spread of the fraction of implanted compared with the total activity is fairly high and varies by about 50%

Fig. 6. The ratio of 210 Po activity (implanted fraction) before and after cleaning the glass sheet. The implanted activity is the activity remaining after cleaning the glass surface with diluted hydrochloric acid. The total activity is determined on glass sheets without any cleaning. F = corroded (old) glass (and NaCl), × = glass with dust; + = glass with burning candles particulate layer; ∗ = glass exposed to cigarette smoke; ◦ = glass exposed to salt (NaCl).

818

B. Roos, C. Samuelsson

on average. Low-density particles should have less impact on the implantation probability of 210 Po than high-density aerosol particles, but this type of particle (candle-burning particles and tobacco smoke) can be assumed to cover the surface more efﬁciently due to the very high amount of particles of smaller sizes. Low-density particles would therefore not necessarily give rise to a higher implantation ratio and can absorb the recoiling nuclei more efﬁciently than “normal” dust particles. Dust particles can give rise to larger particles of higher weights, but these would cover only a few percent of the glass surface.

4. Discussion The implantation probability of 210 Po is clearly affected by the (intermediate) particulate layer of particles. The particulate layer (measured as the surface weight) absorbs or attenuates some of the recoils, thus decreasing the probability of implantation. On clean glass (0– 5 μg cm−2 ) the implanted fraction should be of the order of 0.75–0.6, taking into account the fact that the geometrical implantation probability after one alpha decay is 50%. In the second alpha decay (214 Po → 210 Pb), about 30% of the already implanted activity escapes from the glass matrix, but an additional amount of implanted activity is created from the decay of 214 Po still residing on the surface. In Fig. 6 the implanted fraction can be seen to vary between 0.6 and 0.9. The spread in values is due to statistical variation plus uncertainties emanating from the practical handling of the glass samples (e.g. the loss of some adsorbed activity). Another reason for variations may be age differences between glass samples. An old glass surface is less smooth and homogeneous than a fresh (untreated) one. With surface particle loads above 10 μg cm−2 the implanted fraction decreases to about 0.5. The ﬂuctuation in this region is about ±25%. In Table 1 it can be seen that 5 μg cm−2 of matter of unit density will cover the glass sheet with a homogeneous layer 50 nm thick. Single-particle layers of particles with different diameters will hypothetically cover 75% of the surface (100 nm particles), 8% (1 μm) or 0.8% (10 μm particles). A few very large particles will of course have a major impact on the particle surface mass load and a large fraction of the surface will remain open. For a mass load of 30 μg cm−2 the single-particle layer hypothesis fails for small aerosol particles as the area covered is 450% (100 nm), 45% (1 μm) and 5% (10 μm). A homogeneous Table 1 The percentage fraction of surface coverage for particles of unit density, and different surface layer mass loads (μg cm−2 ) (values greater than 100% indicate multi-particle layers) Particulate layer (μg cm−2 )

Homogeneous layer (nm)

Diameter of spherical particle (nm) with unit density 50 100 500 1000 10 000 (%) (%) (%) (%) (%)

5 10 20 30 50

50 100 200 300 500

150 300 600 900 1500

75 150 300 450 750

15 30 60 90 150

8 15 30 45 80

0.8 2 3 5 8

The behaviour of Rn-222 decay products at the air–glass interface

819

surface layer of 30 μg cm−2 will prohibit any implantation in the glass proper, but in reality a few large particles dominate this mass load. Consequently, a large fraction of the surface will remain clean and the probability of implantation following two consecutive alpha decays is about 40% (Fig. 6).

5. Conclusions The typical mass load of contaminants on household glass sheets will signiﬁcantly reduce the possibility of alpha recoil implantation in the glass surface. The reduction in implanted 210 Po over decades seen in glass samples, taken from radon affected dwellings, may in part be explained by the effects of these contaminants. If dust were found solely responsible for the observed reduction in 210 Po implantation, a slowly increasing mass load on household glass sheets over time must be assumed.

References [1] C. Samuelsson, Retrospective determination of radon in houses, Nature 334 (1988) 338. [2] M.C.R. Alavanja, J.H. Lubin, J.A. Mahaffey, R.C. Brownson, Residential radon exposure and risk of lung cancer in Missouri, Am. J. Public Health 89 (1999) 1042. [3] A. Birovljev, R. Falk, C. Walsh, F. Bissolo, F. Trotti, J.P. McLaughlin, J. Paridaens, H. Vanmarcke, Retrospective assessment of historic radon concentrations in Norwegian dwellings by measuring glass implanted Po-210 – an international ﬁeld intercomparison, Sci. Total Environ. 272 (2001) 181. [4] R.W. Field, D.J. Steck, M.A. Parkhurst, J.A. Mahaffey, M.C.R. Alavanja, Intercomparison of retrospective radon detectors, Environ. Health Perspect. 107 (1999) 905. [5] J.A. Mahaffey, M.C.R. Alavanja, M.A. Parkhurst, E. Berger, R.C. Brownson, Estimation of radon exposure history for analysis of a residual epidemiological study, Radiat. Prot. Dosim. 83 (1999) 239. [6] Z.S. Zunic, J.P. McLaughlin, C. Walsh, A. Birovljev, S.E. Simopoulos, B. Jakupi, V. Gordanic, M. Demajo, F. Trotti, R. Falk, H. Vanmarcke, J. Paridaens, K. Fujimoto, Integrated natural radiation exposure studies in stable Yugoslav rural communities, Sci. Total Environ. 272 (2001) 253. [7] F. Lagarde, R. Falk, K. Almrén, L. Damber, F. Nyberg, H. Svensson, G. Pershagen, Glass-based radon exposure assessment and lung cancer risk, in preparation, 2001. [8] P. Cauwels, A. Pofﬁjn, An improved model for the reconstruction of past radon exposure, Health Phys. 78 (2000) 528. [9] Fehér, M. Lõrinc, J. Pálfalvi, A new method for measurement of 210 Bi on glass sheets, to estimate retrospective radon exposure, Radiat. Prot. Dosim. 66 (1996) 193. [10] B. Roos, M. Bohgard, C. Samuelsson, Implantation probability of 218 Po in glass. Experiments in a wind tunnel, in preparation, 2002. [11] L. Johansson, M. Wolff, C. Samuelsson, The inﬂuence of dust on the 210 Po signal in retrospective radon measurements, Radiat. Prot. Dosim. 56 (1994) 141. [12] B. Roos, M. Bohgard, H.J. Whitlow, C. Samuelsson, On the implantation of 218 Po in glass, in preparation, 2002. [13] P. Eklund, Design and operation of a walk-in radon environmental chamber, Department of Working Environment, Lund Institute of Technology, Lund University, 1995. [14] S.R. Scholes, Modern Glass Practice, Cahners Books, Massachusetts, 1975. [15] L. Johansson, B. Roos, C. Samuelsson, Alpha-particle spectrometry of large-area samples using an open-ﬂow pulse ionization chamber, Appl. Radiat. Isot. 43 (1992) 119.

820

Radon transfer from ground to houses and prediction of indoor radon in Germany based on geological information J. Kemski a , R. Klingel a , A. Siehl b , R. Stegemann a a Kemski & Partner, Alte Heerstr. 1, D-53121 Bonn, Germany b Institute of Geology, University of Bonn, Nussallee 8, D-53115 Bonn, Germany

Soil gas radon concentrations of distinct geological units in Germany are compared with assigned indoor radon measurements in several test areas with different geological settings. A set of transfer factors for speciﬁc situations depending on geology and dwelling characteristics was derived which enable the prediction of the percentage of houses exceeding reference indoor radon levels. To validate the geology-based area prediction, the soil gas database of a generalised German radon map was matched with about 18 000 long-term indoor radon measurements grouped into administrative units on a district level. For most of the districts, prognosis and measurements are in acceptable agreement, and a clear distinction between areas with enhanced radon both in soil gas and in dwellings versus areas with low radon characteristics can be made. 1. Introduction A geogenic radon potential can be estimated from the mapped distribution of geological units and accompanying measurements of radon activity concentration in soil gas to serve as a regionalised prediction tool for radon levels in houses [1–4]. Using geological and other soilrelated information to deduce a geogenic radon potential for predictions is especially recommended, if – like in Germany – distinct geological conditions are region-speciﬁc in large parts of the country and exhaustive, area-covering information about indoor radon concentrations is not available. Within the framework of a series of research projects launched by the German Federal Ministry of the Environment, Nature Conservation and Nuclear Safety, a standardised in situ measuring procedure for both radon and permeability measurements in 1 m depth was developed to ensure comparability of the collected ﬁeld data. In radon mapping, the utility of permeability measurements is restricted. Gas permeability of soil is varying on a very local scale and strongly depends on soil moisture and its sporadic and seasonal changes. In existing RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07103-7

© 2005 Elsevier Ltd. All rights reserved.

Radon transfer from ground to houses and prediction of indoor radon in Germany

Fig. 1. Radon map of Germany with areas of detailed studies.

821

822

J. Kemski et al.

Fig. 2. Frequency distribution of radon in the soil gas for all grid elements.

houses, the radon ﬂux depends on the construction and the artiﬁcially altered building ground and backﬁll. For new houses to be built, it again will be ruled by these parameters, which cannot be predicted precisely in advance. An a priori determination of permeability at undisturbed ground – and much more exhalation – thus has only a very limited value. Therefore, a generalised radon map of Germany (Fig. 1) has been produced which is based solely on soil gas radon activity measurements at 2213 investigation sites as best estimator of radon potential, assuming uniform highest soil permeability [5]. Regionalisation of the measured values was performed by a distance-weighted interpolation on grid basis (3 × 3 km) within geological units. The interpolation method takes into account geological neighbourhood relations, since geological properties are discontinuously distributed and geological boundaries deﬁne homogeneous subgroups of data. The distribution of the grid data plotted on a log scale (Fig. 2) reﬂects the regional coverage of the different geological units. An extensive area with rather low radon concentrations – mainly quaternary sediments – causes two modi below 15 kBq m−3 , while the third modus, representing the Mesozoic and Paleozoic rocks and their soil cover, is located at about 35 kBq m−3 .

2. Radon in houses related to radon in soil gas Besides the generalised approach, detailed studies were done in six representative test areas (Fig. 1) with about 200 houses each and different geological settings. In 1229 exactly located dwellings, indoor measurements of radon concentration were carried out over a period of one year using CR39 detectors. In every building, two detectors were installed. Following the national and international recommendations of radiation protection agencies, the radon concentration was determined in living rooms in the ground ﬂoor. Since the geological subsoil is the dominant source for indoor radon levels, additional measurements of radon concentrations were conducted in the basement. The occupants reported in a comprehensive questionnaire

Radon transfer from ground to houses and prediction of indoor radon in Germany

823

speciﬁcations of the house characteristics. Geological units attributed to the house locations were carefully classiﬁed according to their radon characteristics and sampled for radon in soil gas at 1303 measuring sites. After interpolation, 1229 raster points were assigned spot-to-spot to the dwellings. Based on the fact that the geological substratum is the dominant source for indoor radon [6–9], the objective of the study was to quantify the fraction of soil gas radon entering dwellings, taking into account geogenic criteria (e.g. radon activity concentration in soil gas, rock type, morphology) as well as relevant properties of houses (e.g. construction type, age, foundation depth, building materials and other site and building speciﬁcations). The test areas differ considerably with respect to the radon characteristics due to the different geological substrata with a large range of variation. The highest radon concentrations are found above the tectonites and igneous rocks of Eastern Bavaria. In all areas, the unconsolidated Quaternary sediments also often show increased radon values, depending on their source rocks. The parameters of the distributions of measured indoor levels in the six research areas are listed in Table 1. The indoor radon activity concentrations obviously reﬂect the distribution of radon activity in soil gas, following approximately log-normal distributions with a total GM for ground ﬂoor rooms of 44 ± 2.3 Bq m−3 and a GM range in the test areas between 30 and 66 Bq m−3 . The maximum measured value in a ground ﬂoor room was 2950 Bq m−3 and in a cellar room 4550 Bq m−3 . Occasionally, a positive correlation between soil gas and indoor radon was doubted [10, 11] and it was questioned whether geologically based and geographically coded information might be safely used as a proxy for human exposure. While a bulk scatter diagram of the measured indoor radon levels and soil gas radon concentration from our study seemingly shows no relationship between the two variables, a signiﬁcant and very meaningful positive correlation with highest indoor values above granitic rocks can be obtained when houses belonging to the same geological unit are considered as homogeneous subgroups and the median values for the 28 geological units are plotted (Fig. 3). The radon measurements and data compiled from the questionnaire were subjected to analysis of variance (Fig. 4). The radon concentration in the ground ﬂoor is most of all inﬂuenced by the existence of a basement. In accordance with other studies [12], geology (here represented by the 28 rock units from all study areas) plays an important role and explains the second highest fraction of variance. The inﬂuence of building materials represents mainly Table 1 Radon concentrations (GM ± GSD) and number of measurements in test areas Test area

Ground ﬂoor (Bq m−3 )

Basement (Bq m−3 )

Soil gas (kBq m−3 )

Bayerischer Wald Franken Sauerland Schleswig–Holstein Schwarzwald Pfalz

66 ± 2.6 (390) 38 ± 2.0 (184) 40 ± 2.1 (168) 36 ± 2.1 (219) 36 ± 2.1 (201) 30 ± 1.7 (67)

106 ± 2.8 (355) 70 ± 2.1 (182) 74 ± 2.8 (176) 49 ± 2.6 (177) 64 ± 2.5 (210) 45 ± 2.2 (66)

37 ± 2.1 (422) 28 ± 2.0 (193) 29 ± 1.9 (184) 20 ± 2.0 (219) 37 ± 2.0 (215) 18 ± 1.9 (70)

Total

44 ± 2.3 (1229)

73 ± 2.7 (1166)

30 ± 1.2 (1303)

824

J. Kemski et al.

Fig. 3. Scatter plot of raw and grouped data.

the use of quarried stones with increased radon emanation. The soil gas radon concentration reﬂects the highly signiﬁcant impact of the geogenic radon potential at the site. Also, insulation and morphological situation (slope position of the house) have signiﬁcance levels above 99.9%. For the basement, the ranking of variables according to their contribution is somewhat different. Geology explains the greatest part of the variance, followed by soil gas radon concentration. Some new variables like the year of construction, the depth of foundation and the wall area which is in direct contact to the soil, are gaining more inﬂuence. Figure 5 points out the interaction of processes affecting indoor radon activity. Emanation from geogenic sources (soil and bedrock) and migration deﬁne the radon availability, while the speciﬁc site and construction characteristics inﬂuence the radon transfer into the house. As a rough estimate to quantify the effect of building-speciﬁc parameters, the relationship of

Radon transfer from ground to houses and prediction of indoor radon in Germany

825

Fig. 4. Analysis of variance for ground ﬂoor and basement.

indoor and soil gas activity called the transfer factor TF [13,14] can be used:

TF[‰] = Rnindoor Bq m−3 /Rnsoil gas Bq m−3 · 1000. It is in fact an apparent transfer factor, implying a distinct portion of uncertainty, because other sources of radon like the building materials come into effect, especially in regions of low soil gas concentration. Processes like meteorological change, ventilation rate and life-style of the inhabitants overprint the inﬂuence of the soil-born primary source, especially when the soil gas radon signal is low.

826

J. Kemski et al.

Fig. 5. Parameters inﬂuencing the transfer factor.

Fig. 6. Variation of transfer factors (ground ﬂoor) with construction.

For a series of construction characteristics, the calculated apparent transfer factors for the ground ﬂoor are listed in Fig. 6. The most common types within our study are massive and half-timbered houses with basement and concrete foundations. We consider this as the ‘normal house situation’. A median TF of 1‰ and a 90 percentile of 2.3‰ in this case means that, above geological units with 100 kBq m−3 radon activity in soil gas, half the houses are expected to exceed an indoor level of 100 Bq m−3 in the ground ﬂoor and 10% of houses an indoor level of 230 Bq m−3 . When quarried stones are used, TF increases considerably to 1.5 and 5.3‰, demonstrating that a direct estimation of indoor radon from soil gas radon without taking into account construction features may lead to errors. Houses without basement show transfer factor values generally twice as high, underlining the effectiveness of a basement in acting as a buffer for the intruding soil gas. The table in Fig. 6 shows that the apparent transfer factor is predominantly a house-type and foundation-dependent variable. This fact is illustrated by its slightly negative correlation with soil gas activity concentration (Fig. 7). Houses in areas of low soil gas concentrations are obviously more inﬂuenced by secondary radon sources yielding relatively high apparent

Radon transfer from ground to houses and prediction of indoor radon in Germany

827

Fig. 7. Dependence of TF from soil gas radon (ground ﬂoor).

transfer factors; while in areas of high potential, especially for newer houses, the conventional insulation proves to be an efﬁcient protection against radon resulting in relatively low TF values.

3. Modelling the prediction of indoor levels Hulka et al. [14] propose in an interesting approach an ‘empirical transfer factor’, derived similarly as the ratio of indoor and soil gas radon concentration and corrected by an assumed outdoor radon concentration of 5 Bq m−3 . They discuss how to use its log-normal distribution for predicting probabilities of exceeding distinct indoor levels. The calculation was based on only a small number of data, and a quite high percentage of houses exceeding 200 Bq m−3 for soil gas concentrations below 10 kBq m−3 is predicted, which seems, at least for Germany, not realistic. One reason for obtaining rather high transfer factors might be the quoted use of the third quartile of 15 measurements located around the measured houses, while in our study a great number of soil gas measurements is afﬁxed to the geological units upon which the houses are built, and at each measuring site the maximum radon value from three boreholes is taken, because during the sampling process soil gas systematically tends to be contaminated by atmospheric air. An improved estimation of the percentage of houses exceeding distinct indoor levels in the ground ﬂoor can be worked out if the characteristics of the different house types and the dependence of the apparent TF on the soil gas concentration are taken into account. When the raw TF data are grouped according to the radon soil gas concentration (Fig. 8), a trend towards lower transfer factors with increasing radon in soil gas is evident (see also Fig. 7). Furthermore, a reduction to a speciﬁc house type (e.g. the mentioned ‘normal house situation’), and a limitation to the prevalent range of soil gas radon activity concentration between 20 and 200 kBq m−3 again lead to a different distribution. Therefore, a generalised prediction of indoor levels with an uncorrected TF distribution may result in an overestimation

828

J. Kemski et al.

Fig. 8. Probability plot TF grouped according to soil gas concentration.

Fig. 9. Probability of not exceeding 400 Bq m−3 in the ground ﬂoor (normal house situation).

of the number of radon affected houses. But it is practicable to make predictions for speciﬁc house types. In case of the ‘normal house situation’ (Fig. 9), a prediction of about 3 percent of houses exceeding 400 Bq m−3 in the ground ﬂoor when radon concentration in soil gas exceeds 100 kBq m−3 is close to reality [15].

Radon transfer from ground to houses and prediction of indoor radon in Germany

829

Accepting indoor radon in the ground ﬂoor as a variable largely dependent on soil gas radon activity concentration, a quantile–quantile plot of both approximately log-normal distributions can provide a concept to infer critical indoor levels from soil gas concentrations (Fig. 10). For the mixed stock of all investigated house types, 80 kBq m−3 Rn in soil gas correspond approximately to 200 Bq m−3 in the ground ﬂoor and 115 kBq m−3 to 400 Bq m−3 . Based

Fig. 10. Deriving critical indoor levels from soil gas concentrations.

Fig. 11. Standardised indoor and soil gas distributions.

830

J. Kemski et al.

Fig. 12. Correspondence of soil based prediction and measured indoor levels.

Radon transfer from ground to houses and prediction of indoor radon in Germany

831

on this gross relation, the two distributions of indoor and soil gas radon are standardised and compared after division by 200 Bq m−3 for indoor and 80 kBq m−3 for soil gas (Fig. 11). At value 1 there is a good correspondence, meaning that 90 to 95% of the measured houses are not expected to exceed 200 Bq m−3 and a similar percentage of the investigated area not to exceed 80 kBq m−3 . The relationship of 2.5‰ holds also for higher activity concentrations both for soil gas and indoors, and we interpret this as the strongly geogenic dominated part of the ground ﬂoor radon distribution. For lower values, there is no linear relationship. 50% of the houses have relatively low indoor values below 40 Bq m−3 (200 Bqm−3 · 0.2), whereas half the area has soil gas concentrations up to 24 kBq m−3 , resulting in a relation of 1.7‰, which can be interpreted by a rather efﬁcient radon protection of a large part of the houses. For a nationwide prediction on the basis of administrative area units, a dataset of approximately 18 000 one-year measurements of indoor radon obtained from the archive of the Federal Ofﬁce for Radiation Protection (BfS) was compared with the interpolated soil gas radon concentration of about 40 000 raster elements derived from the radon map (see Fig. 1). Splitting the indoor radon data according to the 200 Bq m−3 and the soil gas data according to the 80 kBq m−3 level, the regional distribution of coincidence falling below or exceeding the threshold is grouped into district units (Fig. 12). The correspondence of soil-based prediction and measured indoor levels is fairly good. Especially, the distinction between districts with enhanced radon both in soil and indoors and areas with low radon concentrations is clearly visible. The regions of uncertainty point to an insufﬁcient number of measurements (soil gas or indoors) and deserve closer inspection, also taking into account human activities like deep mining and other local speciﬁc features. Prognosis maps of higher resolution are in development, matching the distribution of transfer factors and of radon activity concentration in soil gas for each raster point to predict the probability of exceeding certain indoor levels for speciﬁc house types.

References [1] J. Kemski, R. Klingel, A. Siehl, Classiﬁcation and mapping of radon-affected areas in Germany, Environ. Int. 22 (1996) S789–S798. [2] J. Kemski, A. Siehl, R. Stegemann, M. Valdivia-Manchego, Mapping the geogenic radon potential in Germany, Sci. Total Environ. 272 (2001) 217–230. [3] J. Miksova, I. Barnet, Geological support of the National Radon Programme (Czech Republic), Bull. Czech Geol. Survey 77 (2002) 13–22. [4] H.C. Zhu, J.M. Charlet, A. Pofﬁjn, Radon risk mapping in southern Belgium: an application of geostatistical and GIS techniques, Sci. Total Environ. 272 (2001) 203–210. [5] J. Kemski, R. Klingel, A. Siehl, R. Stegemann, M. Valdivia-Manchego, Transferfunktion für die Radonkonzentration in der Bodenluft und der Wohnraumluft, Schriftenreihe Reaktorsicherheit Strahlenschutz BMU-2002598, 2002, 206 p. [6] G. Akerblom, P. Anderson, B. Clavensjo, Soil gas radon – a source for indoor radon daughters, Radiat. Prot. Dosim. 7 (1984) 49–54. [7] G.M. Reimer, L. Gundersen, A direct correlation among indoor radon, soil gas radon and geology in the Reading Prong near Boyertown, Penn., Health Phys. 57 (1989) 155–160. [8] J. Miles, K. Ball, Mapping radon-prone areas using house radon data and geological boundaries, Environ. Int. 22 (1996) S779–S782. [9] D.K. Talbot, J.D. Appleton, T.K. Ball, M.H. Stunt, A comparison of ﬁeld and laboratory analytical methods for radon site investigations, J. Geochem. Explor. 65 (1998) 79–90.

832

J. Kemski et al.

[10] B.L. Cohen, A national survey of 22 Rn in U.S. homes and correlating factors, Health Phys. 51 (1986) 175–183. [11] L. Friis, N. Carter, O. Nordman, A. Simeonidis, S. Järdö, Validation of a geologically based radon risk map: are the indoor radon concentrations higher in high-risk areas? Health Phys. 77 (1999) 541–544. [12] J.A. Gunby, S.C. Darby, J.C.H. Miles, B.M.R. Green, D.R. Cox, Factors affecting indoor radon concentrations in the United Kingdom, Health Phys. 64 (1993) 2–12. [13] J. Thomas, J. Hulka, L. Tomasek, I. Foitikova, I. Barnet, Determination of radon prone areas by probabilistic analysis of indoor survey results and geological prognosis maps in the Czech Republic, in: International Congress Series, vol. 1225, 2002, pp. 49–54. [14] J. Hulka, L. Tomasek, I. Fojtikova, M. Jiranek, A study of the soil-indoor radon transfer factor – a probabilistic approach, in: BfS Schriften, vol. 24, 2002, pp. 166–168. [15] R. Lehmann, J. Kemski, A. Siehl, R. Stegemann, M. Valdivia-Manchego, The regional distribution of indoor radon concentration in Germany, in: International Congress Series, vol. 1225, 2002, pp. 55–61.

833

Checking the “10 point system” for an evaluation of the soil radon potential D. Bleile, J. Wiegand Essen University, Department 9 – Geology, Universitaetsstrasse, 45141 Essen, Germany

The purpose of this study is to verify the “10 point system” for an evaluation of the soil radon potential, which was proposed by [1]. The “10 point system” was tested at 511 test sites throughout Germany. Devonian to Quaternary sediments, intrusive and extrusive rocks, metamorphic rocks as well as different types of backﬁll were considered. At each test site, the soil radon potential was calculated ﬁrst, and then the actual radon concentration of the soil gas was measured at all test sites at a depth of 0.5 m and at 341 test sites at a depth of 0.8 m, too. The two parameters “trafﬁc vibration” and “soil sealing” were not considered. The results of this study show that the predicted soil radon potential has a highly signiﬁcant positive correlation with the actual measured radon concentration in the soil gas (R 2 of mean values: 0.99).

1. The “10 point system” The “10 point system” is based on geogenic as well as anthropogenic parameters on the soil radon potential and does not include inﬂuences of building characteristics [1]. According to their importance, seven parameters were strung in a ranking system, which can be applied in both rural as well as urban areas: (1) made grounds, (2) geology, (3) relief, (4) vegetation cover, (5) tectonics, (6) soil sealing, (7) trafﬁc vibrations (Table 1). The ﬁrst four parameters control the soil radon potential on a more regional scale, whereas the last three can modify the potential locally. The system is based on a grouping of fundamental parameters of the soil radon potential, like concentration and distribution of 226 Ra in soils, grain size, moisture content and permeability of soils. Because it is not suitable to measure the fundamental parameters at each site, those were assigned to parameters which are already mapped (geology and occurrence of made grounds or faults) or which are noticeable in the ﬁeld (relief, vegetation, soil sealing and trafﬁc vibration). The main advantage of the “10 point system” lies in the possibility to estimate the soil radon potential of any site on a local scale, especially when accompanied by soil radon meaRADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07104-9

© 2005 Elsevier Ltd. All rights reserved.

834

D. Bleile, J. Wiegand

Table 1 The “10 point system” for evaluating the soil radon potential (modiﬁed after [1]) Parameter

P

1. Origin of soil 2. Geology

2.1 variety of rocks

2.2 type of backﬁll 3. Relief

4. Vegetation 5. Local parameters

undisturbed soil or backﬁll < 2 m (go to 2.1) backﬁll > 2 m (go to 2.2) sediment: black shale, phosphate, bauxite magmatic rock: rather silicic (granite, granodiorite, syenite, monzonite, rhyolite, dacite, pumice, pegmatite) and alkaline rocks metamorphic rock: orthogneiss, greisen sediment: till, rocks interbedded with “3 point rocks” magmatic rock: rather intermediate/silicic (trachyte, phonolite) sediment: gravel, pelite, carbonate rock, loess magmatic rock: rather intermediate (diorite, andesite) metamorphic rock: clay schist, mica schist, paragneiss, granulite, marble sediment: sand, sandstone, conglomerate, clay, evaporite magmatic rock: maﬁc, ultramaﬁc (gabbro, basalt, diabase, peridotite) metamorphic rock: quartzite, amphibolite, eclogite, serpentinite slags, ashes, sewage sludge, tailings (ore mining) high 226 Ra conc.: low 226 Ra conc.: sand, gravel, soil aggradation, rubble, tailings (coal mining) upper part of hill lower part of hill plain ﬁeld, meadow or no vegetation forest tectonic elements: fault, mining subsidence soil sealing > 50% strong trafﬁc vibration (trains or trucks) < 10 m distance

2 ⎫0 ⎪ ⎪ ⎬ 3 ⎪ ⎪ ⎭ 2 ⎫ ⎬ 1 ⎭ ⎫ ⎬ 0 ⎭ 3 0 1 0 0 1 0 1 1 1

surements. Common classiﬁcations and rankings, frequently based on small scale and generalising maps (e.g. 1 : 5 Mio), can be implanted into the system [2]. For its application, the ﬁve items should be answered one after another, and the corresponding points are summed up. The resulting score ranges from 0 to 10 points, whereby 0–4 points are classiﬁed as a low soil radon potential; 5–10 are ranked as a high potential, but only very few sites have a higher potential than 7 [1]. Due the ﬁndings of the study, this classiﬁcation into “low” and “high” soil radon potential will be modiﬁed later (see Section 4).

2. Working areas The test sites are located in four different regions of Germany. Most of the urban measuring points (i.e. strongly inﬂuenced by anthropogenic parameters) were located in the region of the urban Ruhr district in NW Germany. The geology of the Ruhr district is build up by Devonian, Carboniferous, Tertiary and Quaternary sediments in the south and Cretaceous and Quaternary sediments in the north. Further test sites are represented by the volcanic “Siebengebirge” SE of the City of Bonn (Tertiary maﬁc, intermediate and silicic extrusive rocks), the crystalline Black Forest (silicic intrusive and metamorphic rocks) and the volcanic “Kaiserstuhl” (Tertiary sediments and maﬁc extrusive rocks) in SW Germany.

Checking the “10 point system” for an evaluation of the soil radon potential

835

Within six weeks of the summer 2001, 852 measurements were conducted at 511 measuring points. At all measuring points, samples of soil gas were taken at a depth of 0.5 m. Among the 511 measuring points, a second sample was taken at a depth of 0.8 m at 341 sites. The soil gas was sampled in evacuated Lucas cells and analysed scintillometrically in the laboratory later. The methods of sampling and analysis are described in [1]. While in urban areas, a sample was taken only at 0.5 m depth (unknown underground cables at depths > 0.6 m), the soil gas was sampled in rural areas at both depths. 222 Rn concentrations at 0.8 m depth are on average approximately 60% higher than values at 0.5 m depth, which follows a normal depth gradient of 222 Rn concentrations in the soil [1].

3. Evaluation of the single parameters 3.1. Geology The parameter “geology” summarises several fundamental parameters of the soil radon potential: the 226 Ra concentration and distribution in the soil, the grain size and the permeability of the soil. In the course of this study, some sediments and intermediate magmatic rocks with increased SiO2 contents turned out to be evaluated wrongly by the initial classiﬁcation of [1] (Table 2). In those cases the soil radon potential was underestimated by 1 point or overrated by 3 points. For this reason, a fourth category with 2 points was added to the parameter “geology” (Table 1). Clastic sediments, which are interbedded with “3 point rocks” (i.e. black shale), tills, as well as trachyte and phonolite are evaluated now with 2 points. Within the highly anthropogenic inﬂuenced Ruhr district, it turned out that not all of the observed backﬁlls were recorded in the geological maps. Therefore, the parameter “origin of soil” as well as “type of backﬁll” was sometimes difﬁcult to evaluate without the use of a drilling stock. Interviews with local residents were an appropriate alternative, since in most cases residents were able to tell about the thickness and the material of the backﬁll (mostly post war rubble). The backﬁll “sewage sludge” (Table 1: 3 points) shows a great variety of radium concentrations. At some sites, this material was signiﬁcantly overrated with 3 points, but at some sites Table 2 Former classiﬁcation of the parameter “geology” (after [1]) 2.1 variety of rocks

sediment: magmatic rock: metamorphic rock: sediment: magmatic rock: metamorphic rock: sediment: magmatic rock: metamorphic rock:

black shale, phosphate, bauxite rather silicic (granite, granodiorite, syenite, monzonite, rhyolite, dacite, pumice, pegmatite) and alkaline rocks orthogneiss, greisen gravel, pelite, carbonate rock, loess, till rather intermediate (diorite, andesite) clay schist, mica schist, paragneiss, granulite, marble sand, sandstone, conglomerate, clay, evaporite maﬁc, ultramaﬁc (gabbro, basalt, diabase, peridotite) quartzite, amphibolite, eclogite, serpentinite

⎫ ⎪ ⎪ ⎬ ⎪ ⎪ ⎭ ⎫ ⎬

3

⎭ ⎫ ⎬

1

⎭

0

836

D. Bleile, J. Wiegand

the sewage sludge had 226 Ra concentration up to 1000 Bq kg−1 , justifying a general “3 point rating” in the sense of safeness. 3.2. Relief The parameter “relief” can be understood as a proxy for the variety of soils. Within the same climate, same geology and same depth of water table, the same soils have been developed under similar topographical situations. The parameter “relief” evaluates the distribution of grain size caused by erosion and soil development, the permeability of soils as well as different soil moisture contents along slopes caused by percolating water and interﬂow. The topographical inﬂuences on the soil radon potential are not the same throughout the year, but change from season to season [3–5]. The important “stack effect” of hill tops (i.e. convective ﬂow of soil gas driven by warm soil gas and cold atmosphere) occurs mainly during the winter [1,4]. In the summer time, diffusive conditions control the soil radon potential [1]. Therefore, during summer, the lower parts of the hills with relatively small grain sizes and relatively high contents of soil moisture are characterised by a high 222 Rn emanation and consequently by highest soil-gas 222 Rn concentrations along a slope [5]. Since the measurements for this study were only conducted in the summer period (June–August), the parameter “relief” was not evaluated according to Table 1 but to Table 3. Nevertheless, to estimate the general soil radon potential throughout the year the classiﬁcation of Table 1 should be used, because winter is the critical season for indoor radon. The evaluation of the parameter “relief” according to Table 3 should only be applied if an estimation of the soil radon potential is used for a correlation with comparable data (like in this study: soil radon potential versus 222 Rn concentrations in soil gas), or in cases of temperate climate zones with smaller differences in temperature during the year. If estimating the soil radon potential in climate zones with distinct seasonal differences in temperature, only Table 1 should be used. The classiﬁcation of the three main types of relief (upper part of hill, lower part of hill and plain) are sketched in Fig. 1. The pictograms show that there is no difference between a coneshaped hill or a ridge (Fig. 1, left side) or a sloping transition between two different altitudes (Fig. 1, right side). The intensity of the effect of the parameter “relief” depends on the slope inclination and therefore on the magnitude of erosion. With a decrease in slope inclination, the erosion power is decreasing too. On very steep slopes, the highly permeable and dry soil-type “virgin soil” (A–C soil) is not limited to the upper parts, but occurs all over the slope. In this case only close to the valley bottom a thicker soil is developed with a more ﬁne grained structure (clay and silt). A division into “upper” and “lower” part of the hill as shown in Fig. 1 is a dynamic division. The borders (dotted lines in Fig. 1) are ﬂexible and should be evaluated according Table 3 Classiﬁcation of the parameter “relief” for the summer (= warmer) period 3. Relief

upper part of hill lower part of hill plain

0 1 0

Checking the “10 point system” for an evaluation of the soil radon potential

837

Fig. 1. Pictogram for the classiﬁcation of the parameter “relief”.

to inclination. Nevertheless, with the exception of very steep and very narrow slopes, most slopes can be subdivided as shown in the ﬁgure. 3.3. Vegetation Subsequent to the parameters “geology” and “relief”, the vegetation is the next most important parameter modifying the soil radon potential. Differences in soil moisture content caused by transpiration and differences in permeability caused by root penetration are evaluated by this parameter. Furthermore, plants are acting as “radon pumps” during transpiration [6]. It was shown, that within the same geological unit, forest sites (low soil moisture content, high soil permeability) have lower 222 Rn concentrations in soil gas than ﬁelds, meadows or areas without vegetation (high soil moisture content, low soil permeability) [6]. These ﬁndings were validated by the results of this study (Fig. 2). Figure 2 demonstrates that the 222 Rn concentrations in soil gas are lower at forest sites than at sites with no vegetation or sites, which are

Fig. 2. 222 Rn concentrations in soil gas (depth of 0.8 m) as a function of the parameters “relief” and “vegetation”.

838

D. Bleile, J. Wiegand Table 4 Classiﬁcation of the parameters “vegetation” and “relief” for a correlation with comparable data (e.g. 222 Rn concentrations in soil gas) 3. Vegetation 4. Relief

ﬁeld, meadow or no vegetation (go to 4. Relief) forest (go to 5. Local parameters) upper part of hill lower part of hill plain

1 0 1 0 0

used as meadows or ﬁelds. This correlation is independent from the geological situation or the relief and visible for all three relief subdivisions. 3.4. Interaction between the parameters “relief” and “vegetation” Another important ﬁnding of this study can be recognised in Fig. 2: obviously, it is not necessary to subdivide the parameter “relief” into “upper” and “lower part of the hill” and “plain” within the forests. The 222 Rn concentrations in the soil gas of forests seem to be very similar at all topographical situations. There are three reasons explaining this interaction between “relief” and “vegetation”. First, the deep root activity of trees causes a homogeneous increase in soil permeability along the whole slope. Second, as a result of tree transpiration and decreased evaporation due to the canopy layer, the water balance of soils in forests is far more equable than it is in soils with low plant mass. Third, the erosion on wooded slopes is much less intensive than it is on slopes used as ﬁelds or meadows. This prevents a grading of soil grain sizes along the slope from coarse grained soils at the hill tops to ﬁne grained soils at the valley bottoms. Table 4 shows the consequences of ﬁnding that forests are counterbalancing the effect of the parameter “relief”. The two parameters “relief” and “vegetation” are changed now against each other, and the parameter “relief” is regarded only if the land is used as a ﬁeld, a meadow or has no vegetation. Within the forests the “relief” is skipped and will be not considered. But comparable to Table 3, this procedure should be applied only if the soil radon potential is used for a correlation with the 222 Rn concentrations in soil gas, for example. Even this was not done within this study, because it turned out that the correlation coefﬁcient was not much improved in doing so. Therefore, the soil radon potential was estimated only by using the classiﬁcations of Table 1. If estimating the soil radon potential in the scope of a radon risk assessment of building land, it is a must to use Table 1, otherwise the soil radon potential would be underrated. 3.5. Local parameters Subsequent to the evaluation of more regional parameters (geology, relief and vegetation), local parameters may occur which can modify the soil radon potential further. Those parameters are: tectonic elements (faults), soil sealing and trafﬁc vibrations. It is well known that tectonic elements, like faults or mining subsidence, can enhance the soil radon potential signiﬁcantly [7,8]. A positive correlation between the occurrence of faults and 222 Rn concentrations in soil gas was observed during this study, too.

Checking the “10 point system” for an evaluation of the soil radon potential

839

The inﬂuences of soil sealing and strong trafﬁc vibrations were proven by [9,10]. Simulations in experiments as well as numerical modelling of the soil sealing effect (sealed area 50%) result in an increase of 222 Rn concentrations in soil gas by a factor of 1.6 [1]. The increase of the concentrations due to strong trafﬁc vibration (trains, trucks) at a distance < 10 m reach a factor of 1.2 [1]. Nevertheless, the net effect of the parameters “soil sealing” and “trafﬁcvibration” are difﬁcult to measure in situ with spot measurements, and were consequently not considered. The main difﬁculty with the parameter “vibration” resulted from the circumstance that during sampling on potential sites, vibrations occurred randomly (i.e. passing trains or trucks), and thus the “pump effect” due to the vibrations was observed only sometimes. Difﬁculties with the parameter “soil sealing” are based on the fact that most measuring points with a degree of soil sealing (> 50%) were placed along roads. In these cases the gravel bed beneath the roads, consisting mainly of highly permeable loose sediments, became apparent. Therefore, the effect of the parameter “soil sealing” was superposed by the effect of the backﬁll.

4. Validation of the “10 point system” The main purpose of this study is to verify the “10 point system”. For all measuring points (n = 511) the soil radon potential was calculated using Table 1, and correlated with the actual measured 222 Rn concentration in the soil gas (Figs. 3 and 4). From a total of 852 measurements, 511 were conducted at a depth of 0.5 m (Fig. 3), and 341 at a depth of 0.8 m at the same sites (Fig. 4). Both ﬁgures show the exponential increase of the arithmetic mean, the median, the 10–90 percentiles and the 25–75 quartiles.

Fig. 3. Correlation of calculated soil radon potential and measured soil 222 Rn concentrations at a depth of 0.5 m (n = 511). The numbers above the 90 percentiles indicate the number of measurements.

840

D. Bleile, J. Wiegand

Fig. 4. Correlation of calculated soil radon potential and measured soil 222 Rn concentrations at a depth of 0.8 m (n = 341). The numbers above the 90 percentiles indicate the number of measurements. Table 5 Grouping of the soil radon potential following the classiﬁcation of the “10 point system”

Low soil radon potential Med. soil radon potential High soil radon potential

Points

Mean

0−3 4−6 7−10

6−21 36−60 107

Median 1−16 25−51 68

25−75%

10−90%

1−27 13−79 51−123

0−49 6−111 13−258

Arithmetic means, medians, 25–75 percentiles and 10–90 quartiles are given for a depth of 0.8 m in Bq L−1 .

The equations in the diagrams indicate the regressions and the correlations (R 2 ). Because the two local parameters “soil sealing” and “trafﬁc vibration” were not rated, a maximum of 8 points could be achieved. Anyway, no site reached this high ranking. The comparison of the two different regressions show that independent of the measuring depth, the “10 point system” is reliable. As expected, the 222 Rn concentrations at 0.8 m depth are on average 60% higher than the values at 0.5 m depth, resulting in two different gradients of the regressions. The highly signiﬁcant correlation of R 2 = 0.99 for both, 0.5 m as well as 0.8 m depth, underlines the power of the “10 point system”. Under the aspect of generalising the 10 degrees of the soil radon potential, the potential can be classiﬁed into three groups: “low”, “medium” and “high” soil radon potential. Especially if one looks at the upper 90 percentile and the 75 quartile borders in Fig. 4, a grouping of the soil radon potential becomes obvious (Table 5). Differences regarding a former classiﬁcation [1] are based on the subdivision into “low”, “medium” and “high” soil radon potentials. The “low soil radon potential” spans 0–3 points, and not 0–4 points as initially suggested by [1]. Especially, Fig. 4 shows clearly the increase of the radon potential at “4 points” for both, the arithmetic means as well as the 75 quartiles and 90 percentiles. A second increase of the 75 quartiles and 90 percentiles is visible between

Checking the “10 point system” for an evaluation of the soil radon potential

841

6 and 7 points, which marks the border between “medium” and “high” soil radon potentials. In the context of risk assessment, it is very satisfying that under a “low soil radon potential” (0–3 points) the upper 90 percentile border stays small (222 Rn concentration at 0.8 m depth < 50 Bq L−1 , Fig. 4).

5. Conclusions Summing up, one has to emphasise the simple and timesaving handling of the “10 point system” on any site without doing measurements. Of course, measurements are always recommended to verify the evaluated soil radon potential. If a soil radon potential of 4 or higher is reached by applying the “10 point system” on building land (i.e. medium or high soil radon potential), the soil radon potential should be checked by measurements or, better, the buildings should be constructed radon-safe right away.

References [1] J. Wiegand, A guideline for the evaluation of the soil radon potential based on geogenic and anthropogenic parameters, Environ. Geol. 40 (2001) 949–963. [2] J. Kemski, A. Siehl, R. Stegemann, M. Valdivia-Manchego, Mapping the geogenic radon potential in Germany, Sci. Total Environ. 272 (1–3) (2001) 217–230. [3] T. Taipale, H. Winqvist, Seasonal variations in soil gas radon concentration, in: B. Bosnjakovic, P.H. van Dijkum, M.C. O’Riorkan, J. Sinnaeve (Eds.), Sci. Total Environ. 45 (1985) 121–127. [4] H. Arvela, A. Voutilainen, T. Honkamaa, A. Rosenberg, High indoor radon variations and the thermal behavior of eskers, Health Phys. 67 (3) (1994) 254–260. [5] J. Wiegand, The topographic situation – an important factor on radon risk mapping, in: I. Barnet, M. Neznal (Eds.), Radon Investigations in the Czech Republic VI and the Third International Workshop on the Geological Aspects of Radon Risk Mapping, Prague, 1996, pp. 62–71. [6] S. Feige, J. Wiegand, Vegetation as an important factor controlling radon potential, in: I. Barnet, M. Neznal (Eds.), Radon Investigations in the Czech Republic VII and The Fourth International Workshop on the Geological Aspects of Radon Risk Mapping, Czech Geological Survey and Radon corp., Prague, 1998, pp. 132–141. [7] A. Mogro-Campero, R.L. Fleischer, Subterrestial ﬂuid convection: a hypothesis for long-distance migration of radon within the earth, Earth Planet. Sci. Lett. 34 (1977) 321–325. [8] J. Kemski, R. Klingel, H. Schneiders, A. Siehl, J. Wiegand, Geological structure and geochemistry controlling radon in soil gas, Radiat. Prot. Dosim. Proced. 45 (1/2) (1992) 235–239. [9] J. Wiegand, B. Schott, The sealing of soils and it’s effect on soil-gas migration, Nuovo Cimento 22C (3–4) (1999) 449–455. [10] S. Schmid, J. Wiegand, The inﬂuence of trafﬁc vibrations on the radon potential, Health Phys. 74 (2) (1998) 231–236.

842

Radon-in-water secondary standard preparation D.J. Karangelos, N.P. Petropoulos, E.P. Hinis, S.E. Simopoulos Nuclear Engineering Section, Mechanical Engineering Department, National Technical University of Athens, 15 780 Athens, Greece

This work presents a method to generate radon-in-water solutions of known concentration that can easily be applied in a laboratory that has access to a radon-in-air calibration facility. The method has been proven to be accurate enough for the solution produced to be usable as a secondary standard, traceable to the calibration of the original radon source. This fact, as well as other attractive features such as low running cost and ease of use, makes the method appropriate for purposes such as quality control, intercalibration of instruments and laboratory intercomparison. 1. Introduction Although radon-in-water standards are necessary for the calibration and quality control of instruments and methods, the short half-life of radon limits their availability. For measurement techniques that do not require the sample to come into physical contact with the detector, such as liquid scintillation, this problem can be addressed by adding 226 Ra to the standard to support 222 Rn. However, when using techniques that require the sample to be introduced in the measuring device and subsequently discarded, such as the method based on gas-transfer membranes described later on, the use of 226 Ra can be a source of potential cross-contamination, as well an added complication in terms of disposal. The method presented in this work was developed to avoid the problems of crosscontamination and disposal while not requiring an additional, dedicated 226 Ra source. It is based on bubbling radon-rich air through water, thus avoiding the addition of 226 Ra, while still not requiring any special encapsulation of the radon source. 2. Description of the method 2.1. The method in principle Consider a closed air–water–radon system, where the volume of the air phase is VA , the volume of the water phase is VW and the total radon activity is R. The concentration of radon in RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07105-0

© 2005 Elsevier Ltd. All rights reserved.

Radon-in-water secondary standard preparation

843

Fig. 1. The Ostwald coefﬁcient.

the air phase (CA ) and the water phase (CW ) when the process of radon diffusion between the two phases has reached equilibrium is determined by the following equations: CA VA + CW VW = R,

(1)

CW (2) = k, CA where k is the dimensionless Ostwald coefﬁcient for radon in water. The Ostwald coefﬁcient determines the solubility of a gas in a liquid and depends on the speciﬁc gas–liquid pair as well as the temperature. As can be seen in Fig. 1, where the Ostwald coefﬁcient for radon, determined from experimental solubility data [1,2], has been plotted against temperature, the variability for ordinary room temperatures is rather limited (∼ 1%/◦ C). We will therefore assume in the following that temperature is controlled to a sufﬁcient degree, so that the value of k can be considered constant in the course of an experiment. If the volumes VA , VW of the system as well as the total radon activity R are known, equations (1) and (2) can be easily solved for the radon concentrations CA and CW : CA =

1 R , VA 1 + kVW /VA

(3)

CW =

k R . VA 1 + kVW /VA

(4)

This is the basis of our method for the generation of radon-in-water standards: A known quantity of radon is introduced in a closed circuit, part of which is a vessel containing water. An air pump is used to drive the air through the water, thus establishing equilibrium. Equation (4) can then be used to calculate the concentration of radon in the water. In particular, if the air volume is much greater than the water volume, VA VW , equation (4) can be simpliﬁed to R CW = k , (5) VA where the water volume VW does not enter the calculation. 2.2. Experimental apparatus The apparatus consists of three separate chambers, connected in series in a closed circuit (Fig. 2):

844

D.J. Karangelos et al.

Fig. 2. Experimental apparatus.

• Chamber A is a radon concentration calibration chamber, in which known amounts of radon can be generated by means of a radon-emanating 226 Ra source. This can be part of an independent radon-in-air calibration facility, as long as inlet and outlet valves are provided. • Chamber B contains the water, into which radon will be dissolved. The air inlet is led to a perforated plastic tube below water level (c), while the outlet is well above the water level. Valves are provided at both the inlet and outlet, which allow this chamber to be removed from the circuit and its water content taken for measurement. • Chamber C is connected between the outlet of chamber B and the inlet of chamber A as a safety vessel, to collect any water droplets that might escape. A gas-tight air pump (a) is used to pump air from chamber A through the water in chamber B, while several measuring instruments – a ﬂow meter (b) and pressure gauge (d) in particular – are also connected. Furthermore, active radon instrumentation can be used to monitor the air in the circuit, either included in chamber A, if space permits it, or connected in parallel with the circuit. 2.3. Standard generation procedure The method proceeds as follows: 1. A known amount of 222 Rn is introduced in chamber A. The secondary radon-in-water standard produced will be traceable to the certiﬁcation of the 226 Ra source used, which in this sense will be the primary standard. Alternatively, a reference instrument can be used to quantify the concentration of 222 Rn in chamber A. In the later case, the radon-in-water standard produced will be traceable to the original calibration of the reference monitor. 2. Chamber B is partially ﬁlled with distilled water and connected to the circuit as in Fig. 2. For the experiments described in this work 4 L of water per run were used. It should however be noted that – depending on the volumes of chambers A and C – accurate determination of the water volume might not be critical, as explained previously 3. All connecting valves are opened and the air pump is operated for a time sufﬁcient for the system to attain equilibrium with respect to the distribution of 222 Rn. Two processes occur simultaneously during this step: air with different radon concentrations is mixed across

Radon-in-water secondary standard preparation

845

Fig. 3. Sub-sampling of the radon-in-water standard.

the circuit, leading to a homogeneous distribution of radon, while radon diffuses through the air–water interface, aided by the increased contact area created by bubbling. It was experimentally determined that, for the particular set-up used in the present work, 20 min of circulation at an airﬂow rate of 10 L min−1 are enough to guarantee equilibrium. 4. The connecting valves are closed and chamber B is removed from the circuit, while remaining closed to the atmosphere. Its water content can now be taken for measurement, either by tilting the chamber and drawing a sample of the water through the outlet valve (Fig. 3), or, if larger volumes are required, by opening the chamber lid, in accordance with the sampling protocol of the method to be used for measurement. 2.4. Uncertainty assessment According to equation (5), the uncertainty of the radon concentration in the water standard produced will be 2 2 δCW δR δV 2 δk (6) = + + . CW k R V The Ostwald coefﬁcient has been calculated from experimental solubility data, with an uncertainty of about 2%, while the systematic uncertainty in the activity of the 226 Ra sources used for the present work is of the order of 4%. The total volume VA of the gas phases of the system is the sum of free volumes in all components of the circuit: VA = VCA + VCB + VCC + VCON where VCA , VCB , VCC are the free volumes of the three chambers and VCON is the free volume of all connecting tubing and associated instrumentation. In the system used for the present work, chamber A is a 1.9 m3 calibration chamber, whose volume has been determined with an uncertainty of about 2%. As the total volume of all connecting tubing is not greater than 5 L, while VCB = 6 L and VC = 4 L, it is reasonable to approximate V ∼ = VCA , within the uncertainty of 2%. Taking these parameters into account according to equation (6), the systematic uncertainty of the method as applied in the present work is seen to be equal to 5%.

846

D.J. Karangelos et al.

Table 1 Method validation results CW,0 (Bq L−1 )

CW (Bq L−1 )

3.1 3.2 4.7 6.6 8.3 13.8 16.9

3.4 ± 0.3 3.5 ± 0.5 4.6 ± 0.3 6.4 ± 0.4 8.5 ± 0.4 16.0 ± 1.1 16.4 ± 1.0

Fig. 4. Method validation results.

3. Validation of the method To validate the method, a set of water samples with radon concentrations ranging from 3 to 17 Bq L−1 were generated and measured using an ionisation chamber coupled to a bubbler. Nominal concentrations (CW,0 ), calculated according to equation (1), and measured values (CW ) for these samples are presented in Table 1. The intercept of the least-squares line for these data points, drawn in Fig. 4, is statistically insigniﬁcant, while the slope is not signiﬁcantly different from unity. The RMS-deviation of measured values from the nominal concentrations is equal to 10%.

4. A calibration example As an example of the usefulness of the method, the quality control of the calibration of a radon-in-water measuring apparatus based on a gas-transfer membrane is presented. The apparatus under control consists of a special microporous membrane tube contained in an air-tight vessel, a magnetic stirrer and an active radon monitor based on a solid-state detector. The principle of operation of the apparatus is as follows: The membrane tube is submerged in the water sample, which is contained in the air-tight vessel and continuously stirred using the magnetic stirrer. Radon is allowed to diffuse through the membrane for a ﬁxed time period and quantiﬁed by the active monitor. The radon quantity that diffuses through the membrane during the above ﬁxed time period is proportional to the initial concentration in the water sample; therefore, a calibration factor f can be determined to calculate the initial concentration in the water sample CW from the concentration in the counting chamber of the instrument, CA . The manufacturer of the apparatus has determined a calibration factor f1 = 4 for 20 min of diffusion, while previous experiments at NES-NTUA, using a different methodology from that presented in this work and based on a theoretical calculation of the time evolution of the radon concentration, led to the value of f2 = 1.7 ± 0.1. A comparison of nominal values

Radon-in-water secondary standard preparation

847

Table 2 Calibration control results CW,0 (Bq L−1 )

CA (kBq m−3 )

3.1 3.2 4.7 6.6 8.3 13.8 16.9

1.4 ± 0.1 1.3 ± 0.2 1.9 ± 0.2 2.8 ± 0.4 3.5 ± 0.1 6.7 ± 0.4 7.5 ± 0.3

CW,1 (Bq L−1 )

CW,2 (Bq L−1 )

5.6 ± 0.5 5.2 ± 0.6 7.6 ± 0.8 11.2 ± 1.6 26.0 ± 0.4 26.8 ± 1.4 30.0 ± 1.3

2.4 ± 0.2 2.2 ± 0.3 3.2 ± 0.3 4.8 ± 0.7 11.1 ± 0.2 11.4 ± 0.6 12.8 ± 0.5

Fig. 5. New calibration curve.

(CW,0 ) for a set of samples generated using the method presented in this work with instrument readings (CA ) and values calculated according to the calibration factors f1 and f2 mentioned above (CW,1 , CW,2 ) is shown in Table 2. It is obvious that neither f1 nor f2 is a valid calibration factor, and a new one needs to be determined. This discrepancy may be attributed to ageing and wear of the membrane over a period of years, which might have an effect on its properties with respect to the diffusion of radon. The instrument reading CA has been plotted against the nominal concentration CW,0 in Fig. 5, along with a least-squares line, which in fact is the new calibration curve. As can be seen, the radon concentration in the water sample is indeed reproducible from the amount that diffuses through the membrane (R2 = 0.99), for the range of concentrations tested. The calibration factor was estimated by the least-squares ﬁt to be equal to f3 = 2.22 ± 0.05, which is signiﬁcantly different from both the value given by the manufacturer (f1 = 4), and that previously estimated at NES-NTUA (f2 = 1.7). 5. Conclusions The method presented in this paper has been successfully applied to the calibration of radonin-water measuring instrumentation. It is easy to apply as well as inexpensive, and is therefore very useful for the periodic control of instruments and methods, as in the example presented in this work.

References [1] H.L. Clever (Ed.), Krypton, Xenon and Radon, IUPAC Solubility Data Series, vol. 2, Pergamon Press, Oxford, 1979. [2] D.R. Lide (Ed.), Handbook of Chemistry and Physics, 74th ed., CRC Press, Boca Raton, 1994.

T . his Page Intentionally Left Blank

5. Cosmic radiation

T . his Page Intentionally Left Blank

851

Sequential measurements of energy spectrum and intensity for cosmic ray neutrons T. Nakamura, T. Nunomiya, N. Hirabayashi, H. Suzuki, Y. Sato, D.A.H. Rasolonjatovo Department of Quantum Science and Energy Engineering, Tohoku University, Aoba, Aramaki, Aoba-ku, Sendai, 980-8579 Japan

The cosmic-ray neutron spectrum and dose rate has been measured with three neutron detectors, Bonner sphere, high-sensitivity rem counter and NE213 organic liquid scintillator continuously from October 2000 when the solar activity reached maximum. The neutron spectrum has three major peaks, thermal energy peak, evaporation peak around 1 MeV and cascade peak around 100 MeV. There can be seen a tendency that a cascade peak increases on the day of higher solar activity having higher neutron ﬂux compared with the other two peaks. The time variation of neutron dose rate shows some correlation with the atmospheric pressure and the solar activity.

1. Introduction The Sun has an 11-year solar cycle, and the solar activity reached a maximum in the period 2000–2002. During the solar maximum, big solar ﬂares sometimes happen and have a large effect on our Earth environment. Dynamic outbreaks of geomagnetic storms and radiation showers to the Earth occur sporadically but with increasing intensity and frequency during the years around maximum. The proton is the main element among the solar event particles that reach the Earth. When the proton comes into, and interacts with, the atmosphere, highenergy secondary neutrons and other electromagnetic waves are generated. About 90% of the secondary particles which can reach the ground are photons and muons, and the rest are neutrons. There are several places in the world where neutron ﬂuxes have been sequentially measured using the moderated-type large BF3 detectors. These detectors, however, only give the neutron counts, but not energy spectra. Today, the silicon semiconductor device is so small and more integrated that soft errors, which are caused by the high-energy cosmic ray neutrons, become a serious problem. In this study, we aimed to sequentially measure the energy spectrum and intensity of cosmic ray neutrons on the ground from 2001 to 2002 with three types RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07106-2

© 2005 Elsevier Ltd. All rights reserved.

852

T. Nakamura et al.

of neutron detectors: Bonner sphere (multi-moderator spectrometer with 3 He counter), highsensitivity rem counter and NE213 organic liquid scintillator.

2. Materials and methods For the long-term sequential measurements of cosmic ray neutrons, we set up an experimental cabin on the ground at the campus of Tohoku University. The cabin is of size 564 cm × 230 cm × 262 cm and is air-conditioned. The measuring position is situated at latitude 38◦ 15 North and longitude 140◦ 52 East and altitude of about 70 m. The geomagnetic latitude in Sendai is about 26◦ and the cut-off rigidity is about 8 GV. The Bonner sphere is composed of a 5.08 cm diameter spherical 3 He counter ﬁlled with 5 atm or 10 atm 3 He gas which is placed at the center of the series of polyethylene moderators of different diameters of 0 (Bare), 8.1, 11.0, 15.0 and 23.0 cm [1]. The high-sensitivity rem counter consists of a 12.9 cm diameter 5 atm 3 He gas-ﬁlled spherical counter with polyethylene moderator and inner absorber [2] for measuring the neutron dose rate. The 12.7 cm diameter by 12.7 cm long NE213 organic liquid scintillator is also used to measure the neutron energy spectrum in the MeV region, while the Bonner sphere gives the energy spectrum over the wide energy region down to thermal energy. The NE213 detector has an adequate sensitivity not only for neutrons but also for gamma rays; thus, the neutron events were separated from gamma-ray events with the pulse-shape discrimination technique which uses the fact that the light output pulses from neutrons are slower than those from gamma rays. From the total counts above the cut-off level measured with the Bonner sphere, the neutron energy spectrum was obtained by unfolding with the SAND-II code [3] and the response functions of ﬁve detectors, which are calculated with the MCNPX Monte Carlo code [4]. In this unfolding, we used the initial guess spectrum which is given as a typical cosmic-ray neutron spectrum by Goldhagen et al. [5]. The neutron dose rate was then estimated by multiplying the neutron energy spectrum with the ﬂux-to-ambient dose conversion factor given by ICRP 74 [6]. The total counts of the rem counter were also converted into ambient dose rate by applying the conversion factor of 21.4 cps/(μSv h−1 ) given in [2]. The neutron energy spectrum in the MeV energy region was obtained by unfolding the pulse height distribution of NE213 with the FORIST code [7] using the response functions in the energy range of 5 MeV to 800 MeV given by Sasaki et al. [8].

3. Results and discussions Figure 1 gives the neutron ambient dose rates measured with the rem counter and the Bonner sphere during the period from March 25, 2001 to January 14, 2002, together with the proton ﬂux of energy above 100 MeV observed by using the satellite GOES-8 launched by NOAA, USA [9]. It is known that the secondary cosmic-ray intensity in the atmosphere surrounding the Earth decreases in some cases during a couple of days after a big solar ﬂare due to a large geomagnetic disturbance, called the Forbush decrease, and increases in some cases just after a big solar ﬂare which emits a large amount of high-energy protons of energy above several hundreds MeV that can reach the Earth, called ground-level enhancement.

Sequential measurements of energy spectrum and intensity for cosmic ray neutrons

853

Fig. 1. Time course of neutron ambient dose rates measured with the Bonner sphere and the rem counter and comparison with the proton ﬂux above 100 MeV observed by GOES-8 Satellite.

Generally, the ambient dose rates obtained with the Bonner sphere and the rem counter show good agreement. But from the middle of April till the middle of May, and from the end of June till the middle of July in 2001, the dose rates with the Bonner sphere are 30 to 50% larger than those with the rem counter, and the reason for this large discrepancy cannot be explained yet. As for the lower values of the rem counter in October and in December, this is discussed later. The dose rates with the rem counter clearly give two distinct peaks on April 3 and 17 which may correspond to the increase of the proton ﬂux of energy above 100 MeV observed with the satellite GOES-8, which is the sign of the occurrence of big solar ﬂares on April 2 and 15. In particular, the dose rate of April 17 is about 3.5 times higher than that of other normal days. The time variation of neutron dose rate shows a correlation with the atmospheric pressure as described later. Figure 2 shows the three neutron energy spectra obtained for 12 hours with the Bonner sphere on April 25, May 20 and June 15, 2001, as typical examples. The neutron spectra have three major peaks, a thermal energy peak, an evaporation peak around a few MeV and a cascade peak around 100 MeV. There can be seen a clear tendency that the cascade peak increases remarkably on the day of relatively higher solar activity having a higher neutron ﬂux on April 25 compared with those on the other two days. On June 15, a day of lower neutron ﬂux, the high-energy peak is much lower, but the evaporation peak does not change so much. This phenomenon may reﬂect the ground-level enhancement to some extent. The thermal neutron peaks keep almost constant values for these three days. Figure 3 shows the three energy spectra obtained for the long accumulation period of pulse counting with the NE213 detector. These three sets of data are for three periods of high count rate, low count rate, and average (medium) count rate from November to December 2001. The NE213 detector could not obtain the neutrons of energy below about 7 MeV because of

854

T. Nakamura et al.

Fig. 2. Neutron energy spectra measured with the Bonner sphere on April 25 (high ﬂux), May 20 (average ﬂux), and June 15 (low ﬂux), 2001.

Fig. 3. Neutron energy spectra measured with NE213 on the days of high, average and low count rates in 2001.

the difﬁculty of gamma-ray discrimination as described earlier, but the unfolded spectra also give two peaks below 10 MeV corresponding to the evaporation neutrons and around 80 MeV to the cascade neutrons. Similarly in Fig. 3, the high energy peak around 80 MeV increases with increasing neutron ﬂux, but the increasing rate is smaller than that in Fig. 2. This may be due to the smoothing-out resulting from the longer pulse accumulation period of about 8 days when using the NE213 than that of only 12 hours using the Bonner sphere.

Sequential measurements of energy spectrum and intensity for cosmic ray neutrons

855

The effect of atmospheric pressure on the neutron dose rate is shown in Fig. 4 for the period March 9 to October 4, 2001. The ambient dose rates measured with the rem counter are normalized to 1 atm (1013 h Pa). The dose rate clearly indicates an inverse effect with the atmospheric pressure and can be ﬁtted by an exponential attenuation curve as follows: Y = 2.2 × exp −6.5 × 10−3 X (1) where Y is the ambient dose rate in μSv h−1 and X is atmospheric pressure in h Pa.

Fig. 4. Relation of ambient dose rates measured with the rem counter and atmospheric pressure.

Fig. 5. Relation of neutron ambient dose rates measured with the rem counter and the proton ﬂux above 30 MeV observed by ACE Satellite. The data are corrected for atmospheric pressure.

856

T. Nakamura et al.

Fig. 6. Relation of neutron ambient dose rates measured with the rem counter and the neutron ﬂux measured at the South Pole. The data are corrected for the atmospheric pressure.

Figures 5 and 6 present a comparison of the ambient dose rates measured with the rem counter to the proton ﬂux above 30 MeV observed in the ACE satellite [10], and to the neutron ﬂux measured with the neutron monitor at the South pole [11] for about one year from December 20, 2000 to January 8, 2002. The dose rates are normalized to the values at 1 atm using equation (1) and the time is adjusted to Japan mean time which is 9 h ahead of GMT. The lower-value data were taken with the rem counter placed in the 2nd ﬂoor of a 6th-ﬂoor concrete building in order to check the attenuation of concrete. The dose rates decrease about 1/3 in the building. There can be seen a change of dose rates after November 26, 2001, when the noise cut-off level suddenly was changed for some reason. There cannot be seen a clear correlation between the observed neutron dose rate on the ground at Sendai, Japan with the ACE proton ﬂux and the neutron ﬂux at the South pole over the whole period, but on April 2 and 15 and the succeeding days in 2001, an increase of dose rates was observed, which corresponds to the increase of ACE proton ﬂux indicating the ground-level enhancement and a decrease of dose rates which corresponds to the decrease of South-pole neutron ﬂux indicating the Forbush decrease.

4. Conclusion Long-term sequential measurements of cosmic ray neutrons have been conducted since October 2000 with three different neutron detectors in an air-conditioned experimental cabin on the grounds of the campus of Tohoku University. The time variations of neutron energy spectra and ambient dose rates have been sequentially observed. This long-term observation clariﬁed some correlation between the measured data and solar activity, and also the inverse effect of

Sequential measurements of energy spectrum and intensity for cosmic ray neutrons

857

atmospheric pressure on neutron ﬂuence. We are continuing the measurements and hope to generate more detailed results over longer time periods. This work is partly ﬁnancially supported by a Grant-in-Aid for Scientiﬁc Research from the Ministry of Education, Culture, Science and Sports in Japan.

References [1] [2] [3] [4]

Y. Uwamino, T. Nakamura, A. Hara, Nucl. Instrum. Methods A 239 (1985) 299. T. Nakamura, A. Hara, T. Suzuki, Nucl. Instrum. Methods A 241 (1985) 554. W.N. McElroy, S. Berg, T. Crockett, R.G. Hawkins, AFWL-TR-67-41, Air Force Weapons Laboratory, 1967. H.G. Hughes, et al., MCNPX for neutron–proton transport, in: Proc. Int. Conf. on Mathematics and Computation, Reactor Physics & Environmental Analysis in Nuclear Applications, Madrid, Spain, American Nuclear Society, September 1999. [5] P. Goldhagen, M. Reginatto, T. Kniss, J.W. Wilson, R.C. Singleterry, I.W. Jones, W. Van Steveninck, Nucl. Instrum. Methods A 476 (2002) 42. [6] ICRP Publication 74: Conversion coefﬁcients for use in radiological protection against external radiation, Ann. ICRP 26 (3) (1996). [7] R.H. Johnson, B.W. Wehring, ORNL/RSIC-40, Oak Ridge National Laboratory, 1976. [8] M. Sasaki, N. Nakao, T. Nakamura, T. Shibata, A. Fukumura, Nucl. Instrum. Methods A 480 (2002) 440. [9] National Oceanic and Atmospheric Administration/Space Environment Center, US Department of Commerce, http://www.sec.noaa.gov. [10] ACE Science Center, California Institute of Technology, http://www.srl.caltech.edu/ACE/index.html. [11] Bartol Research Institute, NSF grant ATM-0000315, http://www.bartol.udel.edu/~neutronm.

858

Cosmic radiation at aircraft altitudes calculated by CARI-6 and its comparison with measurements K. Fujitaka, M. Okano, Y. Uchihori, T. Koi, H. Kitamura International Space Radiation Laboratory, National Institute of Radiological Sciences, 4-9-1 Anagawa, Inage, Chiba 263-8555, Japan

About 200 Japanese domestic ﬂights between Haneda (Tokyo) and Matsuyama were examined to ﬁnd the dose due to cosmic radiation. As the dose ﬂuctuated by roughly 25% around the mean, we may need some room in considering the legal scheme to suppress the dose.

1. Flight data As is well known, cosmic radiation has got very special status in that the exposure is very large in certain circumstances, but it is categorized as natural radiation. When the nature of natural radiation is emphasized, we can ignore the dose. On the contrary, when its large dose is emphasized, we need to take some countermeasures to suppress it. Before we decide whether to support or to ignore a legal limit (there is no legal limit in Japan), we need to examine the actual dose. Our ﬁrst task is to make the dates of ﬂights clear. In total, about 200 measured data covering from February 1993 to March 2001 were collected, for which dates we have also conducted calculations. To make the problem simple, ﬂights were limited to domestic ones connecting two speciﬁc airports. One is the airport of Haneda, which is the old International Tokyo Airport, and the other is Matsuyama Airport. The geomagnetic latitude of Tokyo is 25.6 degrees, while that of Matsuyama is 23.3 degrees, and they are about 800 km apart. As both are located nearly along the same geomagnetic latitude, we can ignore the latitude effect. Measurements and calculations were concentrated on these ﬂights.

2. Calculations For the calculations, a computer code CARI-6 which is most reliable, was used. The CARI-6 requires various inputs before we conduct the calculations. To meet these requirements, we RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07107-4

© 2005 Elsevier Ltd. All rights reserved.

Cosmic radiation at aircraft altitudes calculated by CARI-6 and its comparison with measurements

859

asked air attendants of airline companies to provide information on ﬂight altitudes, and at what time the airplane changed altitude, and so on. All the formats were collected later, and were examined carefully, and questionable data were excluded. If we extend the period of study too long, the heliocentric potential may vary largely. Actually, heliocentric potential goes through a peak, as is seen in Fig. 1. However, it is not the case that this large peak gives a large dose. Figure 2 shows the calculated integrated doses for various ﬂight routes. In these calculations, it was assumed that every ﬂight requires 30 min before reaching 37 000 ft, and another 30 min from that to the ground. Results show that the

Fig. 1. Secular variation of heliocentric potential (MV).

Fig. 2. Cosmic ray dose calculated with CARI-6.

860

K. Fujitaka et al.

Fig. 3. Calculated dose of ﬂights between Tokyo and Matsuyama.

dose for Tokyo to New York decreases with increase in heliocentric potential, while the dose for Tokyo to Matsuyama decreases just a little. Therefore, heliocentric potential becomes inﬂuential only in the long distance international ﬂights. In the present case, it is not the major cause of variation. In the end, data between February 1995 and March 2001 were adopted, which totalled 75 ﬂights. For these, CARI-6 calculations were carried out taking into account real altitude variations. We found that the obtained dose ﬂuctuates around its mean by roughly about 25% (Fig. 3). The purpose of this paper is to reveal the origin of the ﬂuctuation.

3. Measurements Measurements with the 3 ∅ spherical NaI(Tl) scintillation detector [1], which is linked to a 4096-channel spectrometric system, have been repeatedly done in various surface environments which involve cosmic rays. The high-energy component, which exceeds 3 MeV, is summed into a single channel, which in turn is regarded as the cosmic ray contribution. The calibration has been done by simultaneous measurements with this device and an ionization chamber, and sometimes with a He proportional counter as well as a rem counter. We believe that this 3 ∅ NaI(Tl) detector is the most dependable device to estimate the cosmic-ray dose in our environment. As this set is portable, it can be easily employed in measuring aircraft exposures. However, more practical instruments are needed. In general, simplicity cannot be overlooked. Even if we have a very precise and sophisticated device, if it is too delicate, and should be treated with great care, it is impractical to use it on board real airplanes. It is true that “the simpler, the better”. The candidates are pocket dosimeters with detectors made of silicon. However, we are unable to know dose rate with this instrument, and only the integrated dose can be obtained. Therefore, we need another calibration, in which we can convert the

Cosmic radiation at aircraft altitudes calculated by CARI-6 and its comparison with measurements

861

Fig. 4. Dose fraction of neutrons and ionizing components.

indicated dose of this small instrument to a more realistic cosmic-ray dose. Therefore, many simultaneous measurements were taken with the pocket dosimeters and the 3 ∅ NaI(Tl) detector. Such comparison made it clear that 2.7 times the indicated dose, which the pocket dosimeter shows would practically give the total dose. It included neutron dose which was after many ﬂights as seen in the picture of UNSCEAR 2000 (Fig. 4) and others in references [2–4]. The neutron fraction increases at low altitude, and remains stable above. In the present case, it was assumed that about 40% of the total is the neutron dose unless the airplane ﬂies at lower altitude, where the fraction is less, for longer time.

4. Consideration Generally, cosmic-ray dose varies with heliocentric potential, ﬂight course, ﬂight direction, and ﬂight altitude. Heliocentric potential is to be excluded as mentioned before. The ﬂight course is ﬁxed here. Inﬂuences of ﬂight direction, on the other hand, are revealed to be as small as 0.1 μSv. Then, what remains? This should be the “ﬂight altitude”. Airplanes ﬂy up and down according to local weather or atmospheric conditions. Figure 5 shows that calculations with CARI-6 gave systematically higher levels by about 1 μSv than the measured ones. The correlation was about 0.9 (Fig. 6). Both the calculated and the measured data ﬂuctuate to the same degree, and the ﬂuctuation seen in the measurements should not be considered only due to low performance of the instruments. The ionization chamber also gave similar ﬂuctuations. Our conclusion is that altitude variation is the major source of ﬂuctuation. Therefore, we may need some room in considering a legal scheme. Roughly speaking, 25% would be appropriate.

862

K. Fujitaka et al.

Fig. 5. Estimated values by various methods.

Fig. 6. Correlation between calculated with CARI-6 and measured with PD.

References [1] M. Okano, K. Izumo, H. Kumagai, T. Kato, M. Nishida, T. Hamada, M. Kodama, in: Natural Radiation Environment III, vol. 2, CONF-780422, US Department of Energy, Springﬁeld, VA, 1980. [2] Sources and Effects of Ionizing Radiation, UNSCEAR 2000 Report to the General Assembly, with Scientiﬁc Annexes, vol. 1: Sources, United Nations, New York, 2000. [3] T. Nakamura, Y. Uwamino, T. Ohkubo, A. Hara, Health Phys. 55 (5) (1987) 509. [4] M. Okano, K. Fujitaka, K. Izumo, in: Proceedings of IRPA 9th, vol. 2, 1996, pp. 262–264.

863

Factors and processes controlling the concentration of the cosmogenic radionuclide 7Be at high-altitude Alpine stations E. Gerasopoulos a , P. Zanis b , C.S. Zerefos b , C. Papastefanou a , W. Ringer c , H.W. Gäggeler d,e , L. Tobler d , H.J. Kanter f a Nuclear Physics Department, Aristotle University of Thessaloniki, 54006 Thessaloniki, Greece b Laboratory of Atmospheric Physics, Aristotle University of Thessaloniki, 54006 Thessaloniki, Greece c Federal Ofﬁce of Agrobiology, Linz, Austria d Paul Scherrer Institute, Villigen PSI, Switzerland e Departement für Chemie und Biochemie, University of Bern, Switzerland f Fraunhofer Institute, Garmisch-Partenkirchen, Germany

Three years (1996–1998) of activity measurements of the cosmogenic radionuclide 7 Be at three high-altitude Alpine stations – Jungfraujoch, Switzerland; Zugspitze, Germany; and Sonnblick, Austria – were analysed in combination with a set of meteorological and atmospheric parameters, in order to investigate the factors and processes that control the concentration levels and the variability of 7 Be. The 7 Be monthly means indicate an annual cycle with a late-summer maximum, mainly driven by tropopause height as revealed by the high correlation coefﬁcients calculated between monthly means of 7 Be and tropopause height (+0.76, +0.56, +0.60), reﬂecting more vertical transport from upper tropospheric levels into the lower troposphere during the warm season. Cross-correlation analysis performed between 7 Be and the meteorological and atmospheric parameters for the whole period and for each month separately, showed a strong negative correlation with relative humidity and a strong positive correlation with tropopause height, indicating that wet scavenging and downward transport from the upper/middle to lower troposphere within anticyclonic conditions are the main controlling mechanisms throughout the year. Correlation coefﬁcients between 7 Be and tropopause height for different classes of relative humidity showed that the correlation is signiﬁcant mainly for high relative humidity conditions and that during intense downward transport conditions the correlation is partly destroyed, indicating the connection of tropopause folding events, upper-level troughs or cut-off lows with STE events. Superposed-epoch analysis on days with low relative humidity, as is indicative for stratospheric intrusions, reveals an increase of 7 Be to 10–12 mBq m−3 and a synoptic pattern of low tropopause height 2 to 3 days before. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07108-6

© 2005 Elsevier Ltd. All rights reserved.

864

E. Gerasopoulos et al.

1. Introduction 7 Be

is a cosmogenic radionuclide with radioactive decay half-life of 53.3 days and it is a naturally occurring gamma emitter (477.6 keV). The bombardment of atmospheric constituents by cosmic rays (CR) leads to the fragmentation of light atmospheric nuclei, primarily 12 C, 14 N and 16 O, thus giving rise to the production of various cosmogenic isotopes [1]. Neutrons and protons, produced in about equal numbers as a result of the nucleonic cascade initiated near the top of the atmosphere by the most energetic primary component of cosmic radiation, are mainly responsible for nuclear disintegrations in the atmosphere. Low-energy interactions by nucleons with energies of 100–200 MeV have the largest contribution to the isotope production. Since protons experience intense and rapid loss of their energy due to ionisation, the production of the majority of radionuclides is attributed to neutrons [2]. Cosmic rays enter the atmosphere mainly from the geomagnetic poles since they are deﬂected by the magnetic ﬁeld of the Earth. Thus, the cosmogenic nuclide production rate is latitude dependent and as a result 7 Be concentrations are subject to a pronounced north–south gradient especially at stratospheric altitudes [3]. Besides, it has been shown that production rates decrease with atmospheric depth [1], which, in combination with the increase of the atmospheric density and therefore the availability of target nuclei for spallation reactions, leads to the existence of a maximum in the production rate at about 20 km. Only 33% of the 7 Be production takes place in the troposphere (production rate: 2.7 × 10−2 atoms cm−2 s−1 ) and particularly in the upper troposphere, while the rest is produced in the stratosphere [4–6]. Since 7 Be is mainly of stratospheric origin, it has been used in many studies as a tracer of STE [7–12]. The fate of the 7 Be atoms is determined by the mechanisms governing the removal and transport of the aerosols, since they become almost immediately attached to ambient aerosol particles in the accumulation mode 0.4–2 μm [13]. The tropospheric residence time of these aerosols can vary from a few days to a few weeks, depending on meteorological conditions [14–16]. Four processes are mainly responsible for the transport of the 7 Be carrier-aerosols, thus controlling its activity levels, variability and seasonal variation. These are wet scavenging, stratosphere-to-troposphere exchange, downward transfer in the troposphere and horizontal transfer from middle and subtropical latitudes to higher latitudes [17]. The scope of the current study is to investigate the inﬂuence of each process on 7 Be and depict its main climatological characteristics.

2. Methodology of measurements and data presentation On a routine basis, measurements of the 7 Be concentration are being conducted at three elevated Alpine stations located along the mountain chain of the Alps: Jungfraujoch, Switzerland (07◦ 59 E/46◦ 32 N, 3580 m asl), Zugspitze, Germany (10◦ 59 E/47◦ 25 N, 2962 m asl) and Sonnblick, Austria (12◦ 58 E/47◦ 03 N, 3106 m asl). At Jungfraujoch 7 Be measurements are carried out with a time resolution of 48 h, while at both Zugspitze and Sonnblick with a time resolution of 24 h. For the analysis performed in this study, three years of data (1996–1998) were used for each station.

Factors and processes controlling the concentration of the cosmogenic radionuclide 7 Be

865

All stations use the same type of air-samplers with air-ﬂow rates varying from 32 to 46 m3 h−1 . Collection of the 7 Be carrier aerosols is accomplished by the use of air ﬁlters made of either glass ﬁbre or cellulose nitrate, which, after sampling, are folded and pressed into a plastic container. For the acquisition of the spectrum and the calculation of 7 Be activity, high-resolution gamma spectrometry is applied, using coaxial or well type detectors. The total uncertainty due to both sampling procedures and counting statistics is of the order of 10% and the calculated activities are routinely corrected to STP (Standard Temperature Pressure) conditions. In order to identify the role of different parameters on the 7 Be activity levels at Alpine sites and to reveal the main controlling processes in a climatological aspect, time-series of meteorological and other atmospheric data for each station covering the period from 1 January 1996 to 31 December 1998 were used, namely relative humidity, speciﬁc humidity, surface ozone, tropopause height. The tropopause height daily means were extracted from 00 and 12 h UTC measurements available for Payerne (6◦ 57 E/47◦ 48 S). Bi-daily averages were calculated for the various parameters at Jungfraujoch to adjust with the different time resolution of 7 Be measurements.

3. Data analysis and interpretation of results 3.1. Annual cycle of the 7 Be concentrations The mean monthly 7 Be concentrations reveal a distinct seasonality at all three stations, with a late summer maximum in July–August (Fig. 1), while the minimum concentrations are generally met in winter months (December). Both a secondary maximum observed in January– February especially at Sonnblick and a minimum in April (Zugspitze, Sonnblick), can be attributed to the particular meteorological characteristics of these months during the three years of analysis (tropopause height above and below normal, respectively). The January– February maximum is also in coincidence with the secondary maximum in absolute frequency of stratospheric air intrusions found at Zugspitze and Sonnblick [12]. Finally, one can notice that the highest concentrations (9 mBq m−3 ) are found at Jungfraujoch, which is also the most elevated station of the three and the lowest values are met at Zugspitze, also inducing a minimum to the amplitude of the variation due to the seasonal cycle.

Fig. 1. Annual cycle of 7 Be concentrations at the three Alpine stations.

866

E. Gerasopoulos et al.

Correlation coefﬁcients were calculated between the monthly means of 7 Be and the various parameters, in order to investigate the main mechanisms that control the late-summer peak in the 7 Be concentrations. The results showed that the maximum correlation coefﬁcients (+0.76, +0.56, and +0.60 for Jungfraujoch, Zugspitze and Sonnblick, respectively) were between 7 Be concentrations and tropopause height, which is also subject to a summer maximum. Correlation with relative and speciﬁc humidity was generally not signiﬁcant at the 95% conﬁdence level. Thus the enhanced solar heating during the warm season leads to an increase of the tropopause and to more efﬁcient vertical mixing within troposphere having as a result more downward transport from the upper troposphere or the stratosphere to the lower troposphere [17]. In order to eliminate all effects that seasonal variation could pose on the correlation coefﬁcients calculated in the next paragraphs, all analyses were repeated for both initial and after deseasonalisation, time-series. 3.2. Correlation between 7 Be and various parameters and connection with different atmospheric processes Cross-correlation coefﬁcients between 7 Be and the various parameters were calculated in order to clarify the role of each parameter on the 7 Be concentration levels. Table 1 shows the correlation coefﬁcients corresponding to zero lag at which the maximum correlation, in terms of absolute values, was found. The high negative correlation between 7 Be and relative humidity indicates that an important fraction of the 7 Be variance depends on wet scavenging. The role of wet scavenging is reﬂected in the fact that during high relative humidity conditions, the wet scavenging rate of aerosols and thus of 7 Be atoms attached to these aerosols, is increased, since condensation becomes more intense. This anti-correlation between 7 Be and relative humidity cannot be explained alternatively as events of downward transport of dry upper tropospheric or stratospheric air, since in this case one would expect speciﬁc humidity to be also a good predictor of 7 Be concentrations, but as shown in Table 1 the correlation between 7 Be and speciﬁc humidity is not signiﬁcant. From the positive correlation between 7 Be concentrations and tropopause height, which is in turn associated with upper ridges [18], it is concluded that the day-to-day variability of the upper-troposphere synoptic situations is also important since it determines both more efﬁcient downward transport from the upper troposphere during anticyclonic conditions and Table 1 Correlation coefﬁcients between 7 Be and various parameters

Relative humidity Speciﬁc humidity Surface ozone Tropopause height

Jungfraujoch

Zugspitze

Sonnblick

−0.56 −0.03 +0.35 +0.43

−0.51 −0.02 +0.42 +0.36

−0.45 +0.07 +0.34 +0.38

With the exception of speciﬁc humidity, all correlations were found signiﬁcant at the 95% conﬁdence level.

Factors and processes controlling the concentration of the cosmogenic radionuclide 7 Be

867

less wet scavenging during the same conditions. The above conclusion is strengthened by the even stronger positive correlation of 7 Be with the 3-day back-trajectories height [19]. Additional trajectory statistics showed that low 7 Be concentrations typically originate from lower-altitude subtropical ocean areas, while high concentrations arrive from the northwest and high altitudes, as is characteristic for stratospheric intrusions [19]. Surface ozone shows a rather signiﬁcant and consistent positive correlation with 7 Be concentrations, reﬂecting the fact that anticyclonic conditions favour at the same time the more efﬁcient downward mixing of air masses rich in 7 Be and ozone. Also, it should not be disregarded that both 7 Be and ozone have been used in many studies as tracers to identify air masses of stratospheric origin and thus part of the correlation could be attributed to the process of stratosphere-to-troposphere exchange. However, the synoptic conditions mentioned above lead also to more intense photochemical ozone production and thermal convection of ozone precursors from the atmospheric boundary layer and thus there is great difﬁculty when interpreting high surface ozone concentrations. The pattern of the cross-correlation analysis, which is the correlation coefﬁcient when one of the time-series is shifted in time, can be displayed via correlograms. Such a diagram is presented only for Zugspitze (Fig. 2), since at all stations a similar behaviour was revealed. The maximum correlation coefﬁcients mainly arise at zero lag, meaning that each parameter has a temporally direct effect on 7 Be concentration. The sharpness of each pattern can also constitute an indication of the persistence of each correlation. As already mentioned, the same analysis was repeated after removing the annual cycle from the time-series (this was not applicable for relative and speciﬁc humidity). First of all, both the analysis of raw and deseasonalised data produces correlograms of the same pattern. The analysis with the deseasonalised data gives stronger correlations for 7 Be against ozone and speciﬁc humidity (however, speciﬁc humidity still does not explain a great fraction of 7 Be variance) and weaker correlation for 7 Be against tropopause height. The signiﬁcance of the difference between the correlations with initial and deseasonalised data was tested using the F -test. Assuming in general the data set with the lower values as the null model and the other one as the alternative model, an F ratio was calculated which was next compared with the Ft corresponding to the desired signiﬁcance level. The comparison with the Ft corresponding to the 99% signiﬁcance level yielded the conclusion that the alternative model was signiﬁcantly improved relative to the null model. The mechanisms that the correlation of the various parameters with 7 Be expresses were also investigated throughout the year by calculating correlation coefﬁcients for each month

Fig. 2. Correlogram between 7 Be and tropopause height (triangles), relative humidity (white circles), speciﬁc humidity (squares) and ozone (black circles).

868

E. Gerasopoulos et al.

Fig. 3. Seasonal variation of the correlation coefﬁcients between 7 Be and various parameters for the three Alpine stations: Jungfraujoch (squares), Zugspitze (triangles) and Sonnblick (dotted line-circles). Bold dashed lines correspond to 95% conﬁdence level (for Jungfraujoch the 95% conﬁdence level is 0.3 due to the 48 h resolution of data).

(Fig. 3). In general, the negative correlation coefﬁcient between relative humidity and 7 Be remains statistically signiﬁcant during the year with the exception of July–August at Sonnblick for which the correlation is lost (Fig. 3a). Even though different fractions of 7 Be variability may be explained for different months, no seasonal pattern is structured to imply any modiﬁcation of the effect of wet scavenging during the year. On the other hand, the correlation coefﬁcients between speciﬁc humidity and 7 Be are signiﬁcant only during the cold months, from September to February (Fig. 3b). It is rather interesting also to notice that for Sonnblick the correlation becomes positive during summer. Thus, in summer relative humidity is better correlated with 7 Be than speciﬁc humidity is, indicating that wet scavenging is more important than downward transport. In winter, relative and speciﬁc humidity shows similar negative correlation with 7 Be and it is difﬁcult to distinguish if this due to wet scavenging or downward transport. It can be assumed that the higher frequency of deep STE events in winter might also play a role. A rather consistent pattern is found for the positive correlation of 7 Be with ozone throughout the year (Fig. 3c) and ﬁnally, the correlogram between 7 Be and tropopause height also reveals a signiﬁcant positive correlation although for a few particular months no correlation is found (Fig. 3d). 3.3. Prevailing meteorological patterns of high 7 Be concentrations As already discussed in Section 3.2, relative humidity and tropopause height are the parameters that together explain the most signiﬁcant part of 7 Be variance, indicating that wet scavenging and downward transport are important mechanisms. A more thorough investigation of this fact was attempted by calculating correlation coefﬁcients between 7 Be and tropopause height for different classes of relative humidity (Fig. 4). The conclusion was that the correlation is signiﬁcant mainly for high relative humidity conditions and that during intense downward transport conditions the correlation is partly destroyed. One reason for this is that during the

Factors and processes controlling the concentration of the cosmogenic radionuclide 7 Be

Fig. 4. Correlation coefﬁcients between 7 Be and tropopause height for different classes of relative humidity. Gray area corresponds to the 95% conﬁdence level.

869

Fig. 5. Superposed epoch analysis on 7 Be and tropopause height using as key-dates the days with relative humidity lower than 25%. Jungfraujoch (squares), Zugspitze (triangles) and Sonnblick (dashed line-circles).

lower relative humidity conditions, indicative of stratospheric intrusion events, high 7 Be concentrations are found with the presence of tropopause folding events, upper-level troughs or cut-off lows [20], which is in the opposite direction of the positive correlation between 7 Be and tropopause height. Although STE events are usually associated with upper-level troughs or cut-off lows, the stratospheric air typically descends to stations within upper-level ridges (and surface anticyclones) following the troughs. Composite maps of the geopotential height at 500 mbar show that 7 Be concentrations greater than 8 mBq m−3 at Jungfraujoch are associated with an upper ridge over Switzerland [11]. For this reason, superposed epoch analysis was chosen to reveal the patterns of 7 Be and tropopause height, using as key-dates the days with relative humidity lower than 25%. As shown in Fig. 5, dry days (RH ≈ 20%) are inducing an increase of 3–7 mBq m−3 to 7 Be, which reaches the climatologically high levels of 10–12 mBq m−3 . Besides, 2 to 3 days before, the synoptic meteorology reveals a pattern of low tropopause height, possibly associated

870

E. Gerasopoulos et al.

with upper-level troughs or cut-off lows, followed by a substantial increase in the tropopause height.

Acknowledgements One of the authors, E. Gerasopoulos, is kindly supported by the Greek State Scholarship Foundation under contract No. 2955 (32nd program, 1998–1999). This study was carried out within STACCATO (Contract No. EVK2-CT1999-00050), a project funded by the European Commission under the Fifth Framework Programme. Measurements were carried out during VOTALP, a project of the European Commission under the Fourth Framework Programme (Contract No. ENV4-CT1995-0025). We also thank the Swiss Meteorological Institute (SMI) at Payerne for providing the tropopause data.

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20]

D. Lal, B. Peters, in: Handbuch der Physik, vol. 46, 1967, p. 551. D. Lal, P.K. Malhotra, B. Peters, J. Atmos. Terrest. Phys. 12 (1958) 306. A.M. Hillas, in: Cosmic Rays, Pergamon, New York, 1972, p. 299. P.A. Benioff, Phys. Rev. 104 (1956) 1122. D. Lal, B. Peters, in: Progress in Cosmic Ray and Elementary Particle Physics, vol. 6, North-Holland, New York, 1962. K. O’Brien, J. Geophys. Res. 84 (1979) 423. E.F. Danielsen, J. Atmos. Sci. 25 (1968) 502–518. L. Husain, P.E. Coffey, R.E. Meyers, R.T. Cederwall, Geophys. Res. Lett. 4 (1977) 363. H. Elbern, J. Kowol, R. Sladkovic, A. Ebel, Atmos. Environ. 31 (1997) 3207. H.E. Scheel, R. Sladkovic, H.J. Kanter, in: P.M. Borrel, P. Borrel (Eds.), Proceedings of EUROTRAC-2 Symposium 98, WIT Press, Southampton, 1999, p. 260. P. Zanis, E. Schuepbach, H.W. Gäggeler, S. Hübener, L. Tobler, Tellus 51B (1999) 789. A. Stohl, N. Spichtinger-Rakowsky, P. Bonasoni, H. Feldmann, M. Memmesheimer, H.E. Scheel, T. Trickl, S. Hübener, W. Ringer, M. Mandl, Atmos. Environ. 34 (2000) 1323. C. Papastefanou, A. Ioannidou, J. Environ. Radioact. 26 (1995) 273. M.B. Gavini, J.N. Beck, P.K. Kuroda, J. Geophys. Res. 79 (1974) 4447. D.M. Koch, D.J. Jacob, W.C. Graustein, J. Geophys. Res. 101 (1996) 18651. R. Jaenicke, in: G. Fischer (Ed.), Atmospheric Physics and Chemistry in Meteorology, Physical and Chemical Properties of Air, in: Laudelt–Boernstein Series, Group V, vol. 4b, Springer-Verlag, Berlin, 1988, p. 391. H.W. Feely, R.J. Larsen, C.G. Sanderson, J. Environ. Radioact. 9 (1989) 223. G. Vaughan, J.D. Price, Q.J.R. Meteor. Soc. 117 (1991) 1281. E. Gerasopoulos, P. Zanis, A. Stohl, C.S. Zerefos, C. Papastefanou, W. Ringer, L. Tobler, S. Hübener, H.W. Gäggeler, H.J. Kanter, L. Tositti, S. Sandrini, Atmos. Environ. 35/36 (2001) 6347. T.D. Davies, E. Schuepbach, Atmos. Environ. 28 (1994) 53.

871

Aircrew exposure assessment by means of a Si-diode spectrometer F. Spurný a , T. Dachev b a Nuclear Physics Institute of the AS CR, Na Truhláˇrce 39/64, 18086 Praha 8, Czech Republic b Solar Terrestrial Inﬂuence Laboratory of the BAS, Acad. G. Bonchev St. 3, 1113 Soﬁa, Bulgaria

The radiation ﬁelds on-board aircraft are complex since they contain particles with energies up to a few hundreds MeV [1]. We have tested the use of an energy deposition spectrometer based on a Si-detector [2,3] in these ﬁelds. The energy deposited in the detector by a particle is analysed by a 256-channel spectrum analyser. This permits us to distinguish the contribution of different types of radiation to integral dosimetry quantities. Results obtained during onboard aircraft measurements are presented, discussed and analysed. It was found that the equipment could be used to monitor on-board exposure for radiation protection.1

1. Semiconductor spectrometer MDU The Mobile Dosimetry Unit (MDU) [2] can monitor simultaneously the doses and numbers of energy deposition events in a semiconductor Si-detector. The MDU is designed as handy equipment. The amplitude of the pulses is proportional to the energy loss in the detector. Final adjustment of the energy scale is made through the 60 keV photons of 241 Am. The amplitudes are digitised and organised in a 256-channel spectrum. The dose D [Gy] is calculated from the spectrum as D=K · (Ei · Ai )/MD (1) where MD is the mass of the detector in kg; Ei is the energy loss in the channel i; Ai is the number of events in it; and K is a coefﬁcient. The operational time of the instrument depends on the lifetime of the accumulators and on the rate the memory ﬁlls up. In the case of continuous operation the lifetime is about 120 hours with the 1350 mA h accumulators, about 1400 hours with 14 A h batteries. 1 The studies were partially supported through EC Project FI5P-CT00-0068 and the Projects of the Grant Agency

of Czech Republic, Nos. 202/99/0151 and 202/01/0710. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07109-8

© 2005 Elsevier Ltd. All rights reserved.

872

F. Spurný, T. Dachev

2. Results and discussion 2.1. Spectra of deposition events First, the available MDU units were exposed in reference radiation ﬁelds. The results obtained for 137 Cs and 60 Co photons showed that the maximum impulsion height is about 1 MeV. The forms of the measured spectra and those of the spectra calculated by means of the EGS 4transport code were found to be quite similar [4]. The equipment was also exposed to some neutron sources. A rather large difference in the event spectra was observed; the maximum impulsion height reaches up to 10 MeV [3]. This could permit us to distinguish photon and neutron induced events in other radiation ﬁelds. Tests with MDU units were performed on protons and heavy ions as well [2,5] and good agreement was found between the measured and predicted (by the GEANT code simulations) spectra. MDU units have been tested in the CERN-EC high-energy reference ﬁeld behind a concrete shield [6]. It was found that in this case the spectrum is still much larger than that for neutrons, reaching the highest values of the energy deposition higher than 20 MeV. The spectrum becomes harder with increasing beam intensity due to the decreasing importance of the muon background [6], which is contributing only to the signal in the low (below 1 MeV) Edep region. Thus, the signal per monitor unit decreases in that region with increasing intensity. For high Edep events, the signal is independent of the intensity given by the monitor units (Fig. 1). This behaviour is also very important for the interpretation of the data measured on-board aircraft. The main effort was directed to measurements on balloon and on-board aircraft. It was observed that the spectra collected there are similar to those registered at the CERN reference ﬁeld (Fig. 2).

Fig. 1. D(Si) per 1 count of monitor (PIC).

Aircrew exposure assessment by means of a Si-diode spectrometer

873

Fig. 2. Comparison of dose spectral distributions.

2.2. Doses in reference ﬁelds As far as the values of dose calculated from the spectra are concerned, they were in very good agreement with the reference values for 60 Co photons; the value obtained for 137 Cs photons was about 8% lower than the reference. In the CERN reference ﬁeld, the doses calculated for the low Edep region were about 40% lower than the values measured with other standard low LET measuring instruments (RSS 112 chamber, TLDs, etc.). We have observed the same behaviour also for individual electronic dosimeters based on Si-diodes and taken it into account for the interpretation of the MDU detector’s readings. 2.3. Results of on-board aircraft measurements To interpret the data (D(Si)) measured on board aircraft, we used the CERN reference ﬁeld results. The dose in Si measured in the low Edep region was assumed to represent the contribution of low LET radiation and the dose in the high Edep region that of high LET component (neutrons). Taking into account the reference ﬁeld values for these components [6], D(Si) measured on board was recalculated to obtain apparent H ∗ (10) values. Since April 2000, the spectrometer has been used during more than 40 individual and six long-term exposures (about 1400 hours each) with about 600 ﬂights in total. All necessary ﬂight parameters were acquired, which permitted calculation of the effective dose E on-board by means of the CARI 6 code and comparison of the result with the apparent H ∗ (10) values obtained as mentioned above. The results obtained in the case of two CSA aircraft ﬂights are presented in Figs. 3 and 4. One can see there that the values of total apparent H ∗ (10) are in rather good agreement with the E-values calculated by means of the CARI 6 code. Such tendency was observed in most of the other ﬂights treated in the same way. It can also be deduced from Table 1, where integral values, calculated by means of both CARI 6 [8] and EPCARD [7] codes and those deduced from direct readings of MDU semicon-

874

F. Spurný, T. Dachev

Fig. 3. Flight proﬁles New York–Prague.

Fig. 4. Flight proﬁles Bahrajn–Prague. Table 1 Internal exposure levels (in mSv) calculated from codes and deduced from measurements with MDU for all four long-term studies on-board a CSA aircraft during year 2001 Integral exposure level

RUN 1 22/03–07/05

RUN 2 30/05–24/07

RUN 3 29/08–16/10

RUN 4 25/10–10/12

Apparent H ∗ (10) – MDU H ∗ (10) – EPCARD E – CARI 6

2.65 2.65 2.68

3.80 3.60 3.78

2.59 2.50 2.60

1.78 1.72 1.88

Aircrew exposure assessment by means of a Si-diode spectrometer

875

ductor spectrometer for four 2001 year long-term measurements on-board of a CSA aircraft, are compared.

3. Conclusions The semiconductor spectrometer proved its capability to characterise the dosimetric characteristics of complex radiation ﬁelds on-board aircraft. Of course, additional effort is needed to improve its performance. Further calibration of it in the CERN-EC reference ﬁelds is necessary. The accumulation of further on-ﬂight data is also necessary; particularly it would be important to perform on-board measurements in situations when the relative contributions of both components are sufﬁciently different (comparison of routes close to geomagnetic poles and to the equator). Both these approaches are in further progress in our laboratories.

Acknowledgements The authors are much obliged to colleagues from Czech Airlines for their help in organising long-term exposures on-board aircraft. They are much obliged to CERN staff for the help during calibration in reference ﬁelds. Their thanks belong also to colleagues from BAS (B. Tomov, Yu. Matviichuk, and Pl. Dimitrov) for their assistance during equipment preparation and for continuous technical support.

References [1] I. McAulay, et al. (Eds.), Exposure of air crew to cosmic radiation, in: EURADOS Report 1996-01, EURADOS, Luxembourg, 1996, pp. 1–77. [2] Ts. Dachev, F. Spurný, et al., Calibration results obtained with LIULIN-4 type dosimeters, Adv. Space Res. 30 (4) (2002) 917–925. [3] F. Spurný, C. Daˇcev, Aircrew on-board dosimetry with a semiconductor spectrometer, Radiat. Prot. Dosim. 100 (2002) 525–528. [4] F. Spurný, R. Gschwindt, L. Makoviˇcka, C. Daˇcev, Response of a semiconductor spectrometer to photons – Comparison of experiment with MC calculation, Nucl. Energy Safety 11 (49) (2003) 114–117. [5] Y. Uchihori, H. Kitamura, K. Fujitaka, Ts.P. Dachev, B.T. Tomov, P.G. Dimitrov, Y. Matviichuk, Analysis of the calibration results obtained with Liulin-4J spectrometer–dosimeter on protons and heavy ions, Radiat. Measur. 35 (2002) 127–134. [6] A. Mitaroff, M. Silari, The CERN-EU high-energy reference ﬁeld (CERF) facility for dosimetry at commercial ﬂight altitudes and in space, CERN-TIS-2001-006-RP-PP. [7] W. Snyder Friedberg, D.N. Faulkner, Radiation exposure of air carrier crew members II, US FAA Report DOT/FAA/AM-92-2, 1992. [8] S. Roesler, W. Heinrich, H. Schraube, Monte Carlo calculations of the radiation ﬁeld at aircraft altitudes, Radiat. Prot. Dosim. 98 (2002) 367–388.

876

Practical considerations for implementation of the EU Directive in the ﬁeld of cosmic radiation exposure D. Irvine, D.J.C. Flower British Airways Health Services, Waterside, PO BOX 365, Harmondsworth UB7 0GB, England, UK

The derivation of generic ﬂight proﬁles by aircraft type and duration of ﬂight using CARI-6 is described. Results from sensitivity analyses altering ﬂight levels and heliocentric potentials are discussed. Effective dose exposure distributions for the period 1999–2001 by Fleet and occupation following implementation of the Directive are presented. 1. Introduction In 1992 British Airways attended informal discussions with the National Radiological Protection Board (NRPB), Department of Transport (DOT), Civil Aviation Authority (CAA) and the Health and Safety Executive (HSE) examining in-ﬂight occupational exposure to cosmic radiation. This was in response to the recommendations of ICRP 60, the CEC Draft Directive on Basic Safety Standards and the NRPB’s formulation of advice. The forum, which met six monthly thereafter, discussed the political and governmental processes that might be enacted to bring ﬂight and cabin crew exposures under legislative control. Various collaborative dosimetric programmes designed to measure cosmic radiation exposure using active and passive devices were also explored. Davies [1] produced a paper based on active monitoring on Concorde and for the period 1988–1990 concluded that with average annual exposure in the range (3–6 mSv) and maximum annual exposure in the range (6–10 mSv) the new ICRP recommendations would have no serious adverse implications for commercial aviation with no individual doses approaching the new occupational exposure limit of 20 mSv. Following an examination of potential exposure proﬁles for various subgroups of the ﬂight and cabin crew, based on a heuristic model that applied a proportional increase in dose rate per hour as northern latitude increased, it was thought that the most highly exposed group would be some of the Tokyo based cabin crew in Japan, with doses in the region of 6 mSv per year. RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07110-4

© 2005 Elsevier Ltd. All rights reserved.

Practical considerations for implementation of the EU Directive

877

Bagshaw et al. [2] conducted a measurement programme on nine round trips between London and Tokyo in conjunction with the NRPB using batches of thermoluminescence dosimeters and polyallyldiglycol carbonate neutron dosimeters to examine dose levels. Subject to the acknowledged uncertainty with the technique, this conﬁrmed that doses around 6 mSv could be achieved with 900 ﬂight (not block) hours. Recommendations from the Article 31 Group who produced guidance material for Title 7 of the Directive acknowledged the acceptability of computer modelling of the exposure to cosmic radiation provided that validation was carried out on representative routes. On this basis in 1996, following the signing of the BSS Directive (96/29/Euratom) we started discussions with the systems group responsible for maintaining the details of sectors ﬂown. The intention was to develop a system for recording, requiring the minimum of intervention, that would fold ﬂown roster information with sector dose data to provide cumulative dose, rolling yearly dose and previous month dose levels. Rather than limit the recording of dose based on potential exposure in excess of 1 mSv per year, we invoked a system that recorded all occupational exposure irrespective of level.

2. Methods In order to develop a generic ﬂight proﬁle based on aircraft type and duration of ﬂight, a random sample of short-haul and long-haul data traces, that were maintained by the ﬂight data recording group, were examined. Altitude proﬁles were analysed within ﬂight duration categories with the aim of selecting representative proﬁles. Linear regression was initially used to examine the ascent to, and descent from, various cruising altitudes. For long-haul operations where the ﬂight duration was in excess of four hours, altitude plateaux were derived from the average of the levels observed. With a measure of long-term cumulative “effective dose” (dose), and percent deviation from the dose calculated using a standard proﬁle, a sensitivity analysis was performed on the longer duration ﬂights in the route structure. This examined the dependence of dose on ﬂight levels and proportion of time spent at higher altitudes invoking the CARI-6 program supplied by the FAA to calculate doses. The effect of choice of heliocentric potential for a ﬁxed ﬂight proﬁle was also investigated using twenty year and ten-year doses accumulated sequentially throughout the period 1959–1999.

3. Results Figure 1 shows the three monthly moving average cosmic radiation data derived from the Concorde on-board monitors for the period October 1997–July 2000 together with passive data derived by the NRPB on a quarterly basis for the same period, Bartlett et al. [3]. The data follow the changes in the solar cycle and the effective dose data for the passive monitors from the NRPB show good agreement with retrospective calculation of dose using CARI-6, with an overly conservative proﬁle, showing cruise to be at 60 000 ft on the LHR-JFK and JFK-LHR routes for all except the ascent and descent phase (mean sector dose for the period being 35.0 μSv, and 35.4 μSv for the NRPB and CARI-6 data, respectively).

878

D. Irvine, D.J.C. Flower

Fig. 1. Ambient dose equivalent. London–New York sector – 3-month moving average. Table 1 Generic ﬂight proﬁles by duration of ﬂight and aircraft type Duration 0–13

Ascent time 2

Descent time 2

14–26

(T −2) 2

(T −2) 2

27–60

13

13

61–240

20

25

> 240 OTH A/C > 240 777 A/C

30

25

25

30

1st alt (time)

2nd alt (time)

3rd alt (time)

4th alt (time)

20 000 ft (T − 4) 28 000 ft (2) 28 000 ft (T − 26) 35 000 ft (T − 45) 30 000 ft (0.15(T − 55)) 36 000 ft (0.75(T − 55))

34 000 ft (0.20(T − 55)) 39 000 ft (0.25(T − 55))

36 000 ft (0.40(T − 55))

38 000 ft (0.25(T − 55))

T is the ﬂight time in minutes.

Examination of short-haul and long-haul routes determined cut-offs of 13 minutes, 26 minutes, 60 minutes and 240 minutes for changes in ﬂight proﬁle, irrespective of aircraft type. Up to 240 minutes, the ﬂight proﬁle was derived from the sampled data together with a degree of expediency for very short duration ﬂights that were later identiﬁed. Above 240 minutes, the 777 ﬂight proﬁle was differentiated from the remainder, Table 1. Table 2 takes the fourth cruise level model and demonstrates how the mean sector dose across all 373 sectors measured, varies in accordance with altitude level in 1000 ft steps, and alterations in proportion of time at the two highest altitudes using 5% changes. The standard proﬁle is seen to over-estimate the mean sector dose by as much as 6.0% (29.2 μSv versus 27.4 μSv) when compared to the same proﬁle but a 1000 ft lower, and under-estimate the mean sector dose by a maximum of 8.1% when contrasted with a proﬁle 1000 ft higher with 40% of cruise time spent at the highest altitude. A 0.2 μSv increase in mean dose is obtained for each additional 5% increase in the proportion of time spent at the highest level. Table 3 displays similar data for the Boeing 777 model that has two cruise altitudes. The adopted proﬁle of 36 000–39 000 ft in the ratio 0.75 : 0.25 over-estimates dose by 7.4% in comparison to a proﬁle 1000 ft lower with less time at the higher altitude, but under-estimates

Practical considerations for implementation of the EU Directive

879

Table 2 Differences in doses dependent on altitude and time at higher altitudes (not Boeing 777) Altitude proﬁle (ft)

Ratio

Sector mean (μSv)

% difference from (S)

34 000–36 000–38 000 Standard (S) 34 000–36 000–38 000 34 000–36 000–38 000 34 000–36 000–38 000 33 000–35 000–37 000 33 000–35 000–37 000 33 000–35 000–37 000 33 000–35 000–37 000 35 000–37 000–39 000 35 000–37 000–39 000 35 000–37 000–39 000 35 000–37 000–39 000

0.40 : 0.25 0.35 : 0.30 0.30 : 0.35 0.25 : 0.40 0.40 : 0.25 0.35 : 0.30 0.30 : 0.35 0.25 : 0.40 0.40 : 0.25 0.35 : 0.30 0.30 : 0.35 0.25 : 0.40

29.2 29.4 29.6 29.8 27.4 27.6 27.8 28.0 30.9 31.1 31.3 31.5

– −0.65 −1.35 −2.10 5.99 5.33 4.63 3.83 −6.02 −6.65 −7.34 −8.08

Ratio refers to different proportion of ﬂight time spent at the two higher altitudes. The sector mean is derived from 373 sectors with over 240 minutes ﬂight duration.

Table 3 Differences in doses dependent on ﬂight altitude and time at higher altitudes (Boeing 777) Altitude proﬁle (ft)

Ratio

Sector mean (μSv)

% difference from (S)

36 000–39 000 36 000–39 000 Standard (S) 36 000–39 000 36 000–39 000 35 000–38 000 35 000–38 000 35 000–38 000 35 000–38 000 37 000–40 000 37 000–40 000 37 000–40 000 37 000–40 000

0.80 : 0.20 0.75 : 0.25 0.70 : 0.30 0.65 : 0.35 0.80 : 0.20 0.75 : 0.25 0.70 : 0.30 0.65 : 0.35 0.80 : 0.20 0.75 : 0.25 0.70 : 0.30 0.65 : 0.35

32.2 32.5 32.8 33.1 30.1 30.4 30.7 31.0 34.4 34.7 35.0 35.3

0.93 – −0.91 −1.93 7.41 6.46 5.54 4.62 −5.62 −6.56 −7.47 −8.40

Ratio refers to different proportion of ﬂight time spent at these two altitudes. Sector means derived from 373 sectors with over 240 minutes ﬂight duration.

a dose of 35.3 μSv by 8.4% when the levels are increased by 1000 ft and the proportion of time at 40 000 ft is increased to one third (0.35). Increases of 5% in the proportion of time at the higher altitude result in an increase in the mean sector dose of 0.3 μSv. To investigate the dependence of dose on the choice of heliocentric potential, we examined route doses on four long-haul routes over the last 40 years. Recognising that individual doses on a yearly basis will vary dependent on the phase of the solar cycle, we believed it was more pertinent to calculate cumulative lifetime doses that could be used in epidemiological investigations to examine dependencies between dose and health outcome.

880

D. Irvine, D.J.C. Flower

For interpretative purposes, single routes were evaluated under the assumption that lifetime doses would be accumulated on these particular routes. Lifetime doses in reality will involve a complex inter-mix of different sector combinations, but the total dose estimate will lie between the extremes of single pairings evaluated. The CARI-6 package provides heliocentric potential from 1959 to date. The mean dose for a current year was contrasted with estimates based on mean doses calculated using heliocentric potentials from the previous December, November and year, respectively. Twenty one 20-year sequential cumulative dose was derived from 1960–1999 while thirty one 10-year sequential dose was calculated for the same period. Table 4 shows the mean, median and range of differences as percentages of the “actual” value. The clear observation of the very close agreement for lifetime doses based on 20-year Table 4 Difference in 20-year doses for four routes dependent on choice of heliocentric potential Route

Comparison

Mean % difference

Median % difference

Range % difference

LHR–JFK

Actual–Previous Actual–December Actual–November Actual–Previous Actual–December Actual–November Actual–Previous Actual–December Actual–November Actual–Previous Actual–December Actual–November

−0.21 −0.78 0.00 −0.03 −0.25 0.02 −0.13 −0.69 0.00 −0.15 −0.81 −0.01

−0.10 −0.82 0.03 0.04 −0.24 0.01 0.01 −0.69 −0.01 −0.01 −0.83 0.01

−1.06, −1.31, −0.60, −0.37, −0.43, −0.21, −0.91, −1.12, −0.49, −1.05, −1.31, −0.60,

LHR–KUL

LHR–NRT

LHR–SFO

0.48 −0.40 0.36 0.21 −0.07 0.18 0.46 −0.33 0.34 0.54 −0.40 0.38

Data based on twenty one 20-year dose from 1960 to 1980. Table 5 Difference in 10-year doses for four routes dependent on choice of heliocentric potential Route

Comparison

Mean % difference

Median % difference

Range % difference

LHR–JFK

Actual–Previous Actual–December Actual–November Actual–Previous Actual–December Actual–November Actual–Previous Actual–December Actual–November Actual–Previous Actual–December Actual–November

−0.03 −0.94 −0.15 0.04 −0.31 −0.03 0.04 −0.83 −0.14 0.04 −0.97 −0.16

0.03 −0.88 −0.18 0.08 −0.26 −0.03 0.13 −0.77 −0.11 0.11 −0.91 −0.14

−1.94, −2.24, −1.18, −0.70, −0.84, −0.45, −1.70, −2.06, −1.10, −1.95, −2.33, −1.24,

LHR–KUL

LHR–NRT

LHR–SFO

Data based on thirty one 10-year dose from 1960 to 1980.

1.24 0.92 1.53 0.53 0.40 0.63 1.18 0.87 1.36 1.33 0.94 1.57

Practical considerations for implementation of the EU Directive

881

Fig. 2. 1999 ﬂight crew radiation doses by ﬂeet: 744: Boeing 747-400; 747: Boeing 747; 777: Boeing 777; CDE: Concorde.

Fig. 3. 1999 cabin crew radiation doses by ﬂeet: EOG: European Operation Gatwick; NBA: Narrow Bodied Aircraft; WBA: Wide Bodied Aircraft.

exposure is apparent irrespective of the heliocentric potential used. This is to be expected though for the results based on the previous year given that the lifetime cumulative doses will only differ by the doses at the beginning and end of the sequence. With the criterion of acceptability based on the closeness of the mean to the “actual” then use of November’s potential in the calculation shows satisfactory agreement for all four routes (0.00%, 0.02%, 0.00%, −0.01%) with the largest range in the LHR-SFO routing only being a 0.38% under-estimate to a 0.6% over-estimate. In view of the shorter time frame, Table 5 shows a greater range in the difference between the 10-year cumulative dose estimates, but use of the November heliocentric potential is still acceptable with mean over-estimates of (0.15%, 0.03%, 0.14%, 0.16%) and the largest range for the LHR-SFO route only being a 1.24% over-estimate to a 1.57% under-estimate. To provide information on the range and shape of the exposure distributions, Figs. 2 and 3 shows 1999 data by aircraft and ﬂeet for around 2000 ﬂight crew and 15 000 cabin crew using the standard proﬁles and use of the November 1998 heliocentric potential. The maximum value of 5.03 mSv for ﬂight crew was recorded by a pilot on the Concorde ﬂeet while that of 4.67 mSv in the cabin crew population was for a Tokyo based member of staff. Table 6 summarises the mean and maximum value for the years 1999, 2000, and 2001 across the ﬂeets conﬁrming the relatively low levels. The Concorde mean and maximum val-

882

D. Irvine, D.J.C. Flower Table 6 Mean and maximum dose (mSv) by ﬂeet for the period 1999–2001

Fleet

Year 1999 Mean

Max

2000 Mean

Max

2001 Mean

Max

320 FC 737 FC 744 FC 747 FC 757 FC 777 FC 7D4 FC CDE FC EOG CC NBA CC WBA CC

0.89 1.34 2.52 2.81 1.78 2.59 1.46 2.86 1.44 1.23 2.78

2.05 2.22 3.51 4.26 3.13 4.07 2.37 5.03 3.12 3.81 4.67

0.91 1.37 2.51 2.32 1.68 2.70 1.31 1.46 1.31 1.25 2.67

2.08 2.72 3.61 3.55 2.87 3.71 2.34 2.20 2.65 3.79 4.27

1.15 1.20 2.41 1.79 1.59 2.61 1.42 1.03 1.28 0.99 2.45

2.59 1.90 3.31 2.88 2.89 3.58 3.24 2.64 2.67 3.35 4.11

FC: Flight Crew; CC: Cabin Crew; 320: Airbus 320; 737: Boeing 737; 757: Boeing 757; 7D4: European Operations Gatwick.

Fig. 4. Flight and cabin crew cumulative doses 1999–2001.

ues have reduced due the grounding of the aircraft from August 2000 until November 2001. Re-estimation of the maximal dose for the WBA cabin crew member using more extreme altitude proﬁles does not reach 6 mSv, which is the nominal action level for review of ﬂight schedules for staff consistently experiencing doses in this region. Figure 4 provides information on the cumulative dose for staff over the three-year period 1999–2001, with the maximum average yearly dose of 4.25 mSv again being experienced by a Tokyo-based cabin crew member.

Practical considerations for implementation of the EU Directive

883

4. Discussion From an examination of the potential size of errors that might accrue from the use of the wrong proﬁle in the lifetime estimation of cosmic radiation dose, we are satisﬁed that the approach of using a ﬁxed generic proﬁle, in a deterministic sense, based on duration and aircraft type, is sufﬁcient for all practical purposes. In addition, adopting a conservative proﬁle in the estimation of dose for transatlantic Concorde ﬂight shows good agreement with sampled passive data and active dosimetry. The use of the November heliocentric potential, which allows sufﬁcient time at the year end for preparation of the subsequent year’s sector dose estimation is appropriate when dose estimation for epidemiological study is the criterion applied. These conclusions, in relation to the estimation of the health effects of exposure to low levels of cosmic radiation, are based on the mortality experience of our ﬂight deck crew (Irvine [4]), and the likely sensitivity of dose response analyses that will be published later in the year derived from a collaborative exercise across nine European ﬂight deck and cabin crew cohorts. For detailed application of the CARI-6 program, it was necessary for all aircraft types other than Concorde, where city pairings were less than 12 miles apart (a CARI restriction), mainly ﬂights starting from and ending at the same point (typically Engineering test ﬂights, technical faults, Charter ﬂights, etc.) to substitute a nominal sector dose of 5.0 μSv. For Concorde ﬂights, a standard dose was applied based on ﬂight duration data from previous ﬂights and a nominal dose rate at 60 000 ft at the London latitude. The initial development for logging ﬂight and cabin crew doses within the computer systems has been superseded by data warehousing technologies, but the principles are still the same. Within this environment, as part of the duty of care to all our employees we have extended the requirement for data recording beyond ﬂight and cabin crew to all staff travelling as part of their work. These data are extracted from the staff duty travel system. To comply with the requirement for staff to have access to their dose records, the details on cumulative dose, rolling yearly total dose and dose in the previous month are posted on the company’s internal intranet, where the staff member can, under appropriate security validation, access their own record. Alongside, these data is extensive educational information on the fundamentals of cosmic radiation, likely exposure levels and associated risks. This information is constantly available and backs up the initial distribution of leaﬂets to all airline staff covering the same material when the Directive was implemented in May 2000. Operational experience with the data collection system necessitated the estimation of sector dose information for city pairings that had not been in the initial implementation. These data may easily be entered retrospectively into the system. For sectors ﬂown on other “carriers” for which there is no ﬂight duration information, a value of 9 μSv is used representing the median value of the distribution of missing doses derived from a sample of these sectors. To incorporate the retrospective inclusion of radiation dose resulting from a solar particle event (SPE), the sector dose matrix, which is normally used to store dose by year, has the facility to record dose for speciﬁc time periods within the year. This retrospective exercise is most readily carried out if the relevant dose is known within six months of the SPE that gave rise to elevated doses at aviation altitudes, before computer ﬁles are archived. Regarding future developments, we do not believe that the move to more sophisticated ﬂight proﬁles that incorporate way-points is necessary. This judgement is based on the low levels

884

D. Irvine, D.J.C. Flower

of exposure calculated at present, the fact that individual ﬂight schedules will not change, and that epidemiological analysis based on lifetime dose is little affected. We will though keenly follow up comparisons between CARI and EPCARD, recognising the need for harmonisation and standardisation in the implementation of the directive across Member States.

Acknowledgements We acknowledge the contribution of J. Tyerman, analyst programmer, who was key in the initial computer implementation and subsequent extraction of exposure information.

References [1] [2] [3] [4]

D.M. Davies, Radiat. Prot. Dosim. 48 (1) (1993) 121–124. M. Bagshaw, et al., Occup. Environ. Med. 53 (1996) 495–498. D.T. Bartlett, et al., Radiat. Prot. Dosim. 91 (4) (2000) 365–376. D. Irvine, et al., Aviat. Space Environ. Med. 70 (1999) 548–555.

885

Monitoring of the cosmic radiation on IBERIA commercial ﬂights: One year’s experience of in-ﬂight measurements J.C. Saez Vergara a , A.M. Romero Gutiérrez a , R. Rodriguez Jiménez a , R. Dominguez-Mompell Román b , P. Ortiz García b , F. Merelo de Barberá b a CIEMAT, Dosimetría de Radiaciones, Av. Complutense 22, E-28040 Madrid, Spain b IBERIA LAE, Servicio Médico, Zona Industrial Aeropuerto Barajas, E-28042 Madrid, Spain

The results from the ﬁrst year of the experimental measurements of in-board radiation doses received on IBERIA commercial ﬂights are presented. The studied routes cover the most signiﬁcant IBERIA destinations and provide a good estimate of the route doses as required by the new Spanish regulations on air crew radiation protection. The results from the different instruments and the comparisons with the predictions from some route-dose codes (CARI, EPCARD) are discussed. High LET detectors were tested in the CERF facility at CERN, which provides a realistic approach to the radiation ﬁelds at 9–12 km altitudes. In contrast with the data already published, which are mainly focused on North latitudes over parallel 50, many of the data presented in this work have been obtained for routes to Central and South America. These data can be of special interest for the route-dose code developers. The ﬁrst analysis of the obtained data shows that the annual dose received by the IBERIA air crew members ranges from 0.5 to 3.0 mSv, considering 600 hours of effective ﬂying time (from take-off to landing). These values are well below the current dose limits in Spain (roughly 20 mSv y−1 ) but also below the reference of 6 mSv y−1 for radiation workers type A, which should require individual monitoring. However, most of the air crew members receive more than 1 mSv y−1 and therefore a speciﬁc radiation protection programme and annual dose estimates are required by current Spanish regulations.

1. Introduction As a consequence of the ICRP-60 recommendations [1] and the subsequent regulatory documents [2,3], European Union Member States should revise their national legislation on radiation protection. In Spain, the new Law [4] was approved in July 2001 and its text mentions the RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07111-6

© 2005 Elsevier Ltd. All rights reserved.

886

J.C. Saez Vergara et al.

exposure of aircrew during the operation of commercial aircraft. In the case of the annual dose being over 1 mSv, the Spanish Law requires speciﬁc protection measures including dose assessment, working schedule organisation, information for the workers and special protection during pregnancy for the female and the new child to be born. In anticipation of these legal novelties, since May 2000, IBERIA and CIEMAT have been collaborating in a project aimed at determining the radiation doses received during IBERIA commercial ﬂights. The ﬁnal result is to provide recommendations for establishing an adequate radiation protection program for IBERIA aircrews.

2. Instruments and methods 2.1. Selection of instruments Active (powered) dose-rate monitors were selected from commercially available instrumentation. In particular, high-energy neutron monitors recently available were targeted in order to check the real capabilities of such instruments in the presence of high-energy neutrons. Table 1 presents the main characteristics of the selected instruments. The instruments are categorised into those designed to measure the ionising and the neutron components of the cosmic radiation. The ionising component approximately corresponds to the low-LET (< 10 keV μm−1 ) component and the neutron component relates to the neutron and the nuclear interaction of the high-energy proton component of the ﬁeld. The instrument readings are converted to the quantity Ambient dose equivalent H ∗ (10) [5] using appropriate conversion factors when needed. Regular calibration checks are performed in the CIEMAT Secondary Standards Laboratory (ionising component: 137 Cs and 60 Co photon beams) and the CERF facility at CERN (neutron component). Figure 1 shows a linearity check performed with the two neutron counters plus the TEPC. While the TEPC is offering good results (the slope is close to unity), both neutron counters deviate from such behaviour (underestimation of 35% and 55% for SWENDI2 and LINUS, respectively). It is also observed that none of the lines is passing through the origin, which could be related to some high-energy particle contamination (probably muons) coming from the main beam. Table 1 Active instruments employed for dose-rate measurements onboard IBERIA ﬂights Instrument

Detector type

Low LET or ‘ionising’ component Reutes Stokes RS131 Pressure ion chamber Genitron GammaTracer Geiger–Müller tubes Eberline FH-40G Proportional counter High LET or ‘neutron’ component 3 He tube + W moderator Eberline SWENDI2 3 He tube + Pb moderator MAB Linus SNM500X Far West Tech. HAWK TEPC (12.7 cm diameter)

Energy/LET range 70 keV–8 MeV 45 keV–1.3 MeV 45 keV–1.3 MeV Up to 5 GeV > 10 MeV 0.3–1024 keV μm−1

Monitoring of the cosmic radiation on IBERIA commercial ﬂights

887

Fig. 1. Linearity test of neutron component instruments performed at CERF.

2.2. Onboard measurements Due to safety reasons, all the instruments are battery operated during the ﬂight and the autonomy (battery life) ranges from 6 hours (LINUS) to 3 years (GammaTracer). Each instrument records the entire ﬂight (switching on before take-off, switching off after landing) and the integration time is 5 minutes, with the exception of TEPC, which is automatically recording data (including two LET spectra) every minute. Geographic data (latitude, longitude and altitude) on route are obtained automatically from a GPS mounted in the TEPC, which needs a small antenna ﬁtted in a window of the aircraft. The recorded data are stored in each instrument ranging in storage capability from 256 data (more than 21 hours) to practically unlimited (RS131 and TEPC). After each ﬂight, all the data are transferred and processed in a laptop with MS Excel templates to analyse the results. The route data are entered in the two codes employed in the Project: CARI-6 (FAA, USA) [6] and EPCARD 3.1 (GSF, Germany, developed under an EC Contract) [7]. A database with extensive information and analysis results of each ﬂight is maintained at CIEMAT. 2.3. IBERIA destinations Aircrew dosimetry in the Spanish airlines differs from the previous European studies due to the geographic situation of Spain (the most meridional European country) and economic and cultural links with America. In addition, Spanish isle territories (Canary and Balearic Islands) must be served with many ﬂights. As a consequence, an important part of IBERIA ﬂights goes to the Tropical and Equatorial regions, which means lower doses but also probably greater uncertainties (most of the published studies are focused in the North Hemisphere [8–10]). 3. Results The four instruments intended to detect the ionising component seem to be sensitive enough to detect the inﬂuence of ﬂight operation parameters (altitude and latitude) with only a 5 minute

888

J.C. Saez Vergara et al.

integration period. In most ﬂights, the three standard gamma probes agree within ±15% (5% when RS131 is compared with GammaTracer). On the contrary, the TEPC (low LET window) shows up to 20% underestimation in some ﬂights. Results from the neutron monitors are slightly poorer than those from the ionising component monitors, but still they can detect the inﬂuence of altitude and latitude on the dose rate with a 5–10 minute integration period. A study of the ratios LINUS-to-TEPC (high LET) and SWENDI2-to-TEPC (high LET) shows that these ratios are similar to those obtained during the calibration check at CERF. This demonstrates that the three instruments behave in the same way at CERF as when ﬂying in a commercial aircraft. From the basic dose-rate results obtained with each instrument, the following dose quantities are calculated: – route dose: integrated dose from take-off to landing; – mean dose rate during the ﬂight (route dose divided by the ﬂight duration); – annual doses (ambient dose equivalent, effective dose), considering 600 ﬂying hours per year (ﬂying hours are considered from take off to landing). Figure 2 displays two examples of ﬂight dose-rate proﬁles and dose estimates for two relevant destinations in America (New York City and Santiago de Chile). The experimental dose rates for a given altitude can be plotted as a function of Geomagnetic Latitude, which is closely related to the vertical magnetic cut-off rigidity [8]. Figure 3 shows the dose-rate results obtained with the tandem RSS131 + SWENDI2 for two different ﬂight altitudes. The data are in good agreement with the data already published [8–11] and illustrate the well-known behaviour of the dose rate as a function of the geomagnetic latitude and ﬂight altitude. However, all the ﬂights over the Caribbean Sea clearly deviate from this behaviour, yielding higher doses than expected. This fact has been conﬁrmed in all the ﬂights performed from Miami to Central America and some ﬂights from/to Spain, which ﬂy over this geographic area. The reason for this effect is under study but it could be related to some local peculiarity in the Earths’ magnetic ﬁeld that affects the assumptions in calculating the geomagnetic latitude from the geographic coordinates.

4. Discussion The experimental results from the tandem RS131 + SWENDI2 and the TEPC are expressed in terms of the operational quantity ambient dose equivalent H ∗ (10). While they are in good agreement in mean values (0.97 ± 0.09), the differences when considering individual ﬂights can reach ±30%. The code EPCARD 3.1 provides dose estimates in terms of H ∗ (10), which permits a direct comparison with the experimental results (Fig. 4). The mean values of both ratios are close to unity, but again the range of variation for individual ﬂight reaches up to ±30%. Some authors have already reported that for the radiation ﬁelds at aviation altitudes the operational quantity recommended by ICRU and ICRP for area monitoring H ∗ (10) could not be a conservative estimate of the risk-related quantity effective dose E. As the code EPCARD 3.1 computes both quantities for each ﬂight, the ratio of effective dose to ambient dose equivalent estimates was studied for the IBERIA ﬂights in 2001. A mean value of 1.13 ± 0.02 was found

Monitoring of the cosmic radiation on IBERIA commercial ﬂights

Fig. 2. Examples of experimental measurements onboard IBERIA ﬂights. Dose rates were taken from the tandem RSS131 + SWENDI2. 889

890

J.C. Saez Vergara et al.

Fig. 3. Experimental dose rate as a function of geomagnetic latitude for two altitudes. Open dots correspond to ﬂights over the Caribbean Sea.

Fig. 4. Comparison of the experimental results obtained by (RS131 + SWENDI2) and TEPC with the results calculated by EPCARD 3.1. Error bars show the range of ratios.

Monitoring of the cosmic radiation on IBERIA commercial ﬂights

891

Fig. 5. Assessment of annual effective dose derived from the experimental instruments TEPC and RS131 + SWENDI2 and calculated with the computer codes CARI-6 and EPCARD 3.1. Error bars show the range of results.

for the ratio E/H ∗ (10), with a minimum inﬂuence of the geographic area. This value is being employed to calculate effective doses from the experimental results in terms of H ∗ (10). The comparison of effective-dose estimates reported by EPCARD 3.1 and CARI-6 shows that a clear inﬂuence of the ﬂight operation area can be distinguished. While the ratio EPCARD/CARI is clearly lower than unity for Southern destinations (a mean value of 0.9 for the ﬂights from/to South America), Northern destinations show that the ratio is greater than unity, reaching up to 1.2 for the ﬂights from/to North America. The reasons for this discrepancy are probably related to the basic differences between the two codes. The estimate of annual effective dose considers the above-discussed points and is presented in Fig. 5, which shows the good agreement of all the considered dose assessment methods. Furthermore, the differences between distinct geographical areas are reported by all the methods in a similar way. Despite the dose assessment procedure, deviations from mean values up to ±30% can be found for individual ﬂights due to the differences in ﬂight parameters, specially altitude but also geographic track.

5. Conclusions Table 2 summarises the results from the IBERIA ﬂights studied in 2001. Annual effective dose estimates are based on experimental results with the TEPC, considering 600 effective ﬂying hours (from take off to landing) per year. The mean value of the annual effective dose for the 100 studied ﬂights is 1.4 mSv, ranging from 0.4 to 2.7 mSv. As expected, the higher doses are observed when ﬂying to Europe (1.7 mSv) and North America (2.2 mSv). The lower doses (<1.0 mSv) are being observed in domestic ﬂights (excepting Canary Islands) due to the short duration of the ﬂights (0.5 to 1.5 hours) and the low ﬂight levels (around 9 km high). The remaining destinations involve tropical and equatorial latitudes and the observed doses are in the range 1.0–1.6 mSv. From this study and considering the new regulations in Spain, some preliminary suggestions can be made concerning radiation protection of IBERIA aircrew:

892

J.C. Saez Vergara et al. Table 2 Summary of onboard measurements performed on IBERIA ﬂights in 2001 Region

North America Central America South America Africa Europe Spain (excl. Canary Is.) Canary Islands

Number of ﬂights

Flying time (h)

Mean altitude (km)

Annual effective dose (mSv) Mean

Range

6 15 11 6 28

51.4 39.8 114.9 31.9 57.6

10.2 10.6 10.5 10.2 10.4

2.22 1.61 1.43 1.25 1.72

1.78–2.67 1.14–2.09 1.00–1.93 0.81–1.48 0.98–2.46

23 11

18.6 28.1

9.1 10.2

0.70 1.17

0.39–1.44 0.65–1.55

Dose estimates are based on TEPC results and were computed for 600 effective ﬂying hours per year.

– With the exception of domestic ﬂights (excluding Canary Islands), annual effective doses exceed the value of 1 mSv. This fact obliges IBERIA to prepare a Radiation Protection program in the terms required by the current Spanish law. – The higher individual values for annual effective dose are below 3 mSv. This means that the probability of receiving doses over 6 mSv is remote, making it unnecessary to establish individual monitoring for IBERIA aircrew. – An annual dose assessment program based on computer codes for extensive routine calculations and its validation with onboard experimental measurements is being proposed for IBERIA aircrew dosimetry. – Co-operation with other expert groups and airlines is expected in order to complete, maintain and assure the quality of the dosimetry records. It should be noted that these measurements were performed in 2001, when solar activity reached a maximum value. This means that the dose estimates obtained should be considered as minima for the 11-year solar activity cycle. It is estimated that the doses could increase by 5% to 20% depending on the ﬂight parameters (destination, altitude and track). However, the conclusions from this work are still valid even considering those increases in the annual effective dose.

References [1] ICRP Publication 60: 1990 Recommendations of the International Commission on Radiological Protection, Ann. ICRP 21 (1–3) (1991). [2] Council Directive 96/29/Euratom of 13 May 1996 laying down basic safety standards for the protection of the health of workers and the general public against the dangers arising from ionising radiation, Ofﬁcial J. Eur. Commun. Ser. L 159 (1996). [3] Recommendations for the implementation of Title VII of the European Basic Safety Standards Directive (BSS) concerning signiﬁcant increase in exposure due to natural radiation sources, Report Radiation Protection 88, European Commission, Luxembourg 1997. [4] Ministerio de la Presidencia, Reglamento sobre protección sanitaria contra radiaciones ionizantes, Real Decreto 783/2001, BOE 14555, Madrid, 2001 (in Spanish).

Monitoring of the cosmic radiation on IBERIA commercial ﬂights

893

[5] ICRP Publication 75: General principles for the radiation protection of workers, Ann. ICRP 27 (1) (1997). [6] CARI-6 computer program, Developed at the Civil Aerospace Medical Institute, Federal Aviation Authority, USA; Free download from http://www.cami.jccbi.gov/aam-600/610/600Radio.html. [7] H. Schraube, W. Heinrich, G. Leuthold, V. Mares, S. Roesler, Aviation route dose calculation and its numerical basis, in: Proceedings of the Tenth International Congress of the International Radiation Protection Association, Hiroshima, Japan, 2000. [8] B.J. Lewis, M.J. McCall, A.R. Green, L.G.I. Bennett, M. Pierre, U.J. Schrewe, K. O’Brien, E. Felsberger, Aircrew exposure from cosmic radiation on commercial airline routes, Radiat. Prot. Dosim. 93 (4) (2001) 293– 314. [9] I. McAulay, et al. (Eds.), Exposure of air crew to cosmic radiation, EURADOS Report 1996-01, European Commission Report Radiation Protection 85, European Commission, Luxembourg, 1996, pp. 1–77. [10] G. Reitz, K. Schnuer, K.B. Shaw (Eds.), Radiation Exposure of Civil Aircrew, Proc. Workshop, Luxembourg, 1991, Radiat. Prot. Dosim. 48 (1993). [11] M. Kelly, H.-G. Menzel, T. Ryan, K. Schnuer (Eds.), Cosmic Radiation and Aircrew Exposure, Proc. Workshop, Dublin, Ireland, 1998, Radiat. Prot. Dosim. 86 (1999).

894

Health aspects of radiation exposure on a simulated mission to Mars W. Friedberg a , K. Copeland a , F.E. Duke a , K. O’Brien b , E.B. Darden Jr. c,1 a Civil Aerospace Medical Institute, AAM-610, Federal Aviation Administration, PO Box 25082,

Oklahoma City, OK 73125-5066, USA b Department of Physics and Astronomy, Northern Arizona University, Flagstaff, AZ 86011, USA c Oak Ridge Associated Universities, Oak Ridge, TN 37831, USA

In the simulated mission to Mars described in this report, anticipated times of highest solar activity are avoided, and launch times are dates when Earth and Mars are positioned for the shortest travel time consistent with the minimum energy expenditure to travel between the 2 planets. The mission lasts approximately 2.7 years: 536 days in transit plus 439 days on Mars. The estimated effective dose to each of the space travelers is 2.26 Sv, which includes 0.17 Sv received during a solar-proton event. The lifetime risk of fatal cancer from radiation exposure during the mission ranges from 2.4% for men of ages 55–64 (at the time of the mission) to 16.7% for women of ages 25–34. If one of the space travelers conceives a child after the mission is completed, the risk to that child of inheriting a genetic defect caused by the parent’s radiation exposure on the mission would be between 0.7 and 1.1%. The space travelers will not experience acute effects of radiation, but they are likely to incur an increased risk of developing cataracts. Men on the mission may experience temporary reduced fertility.

1. Introduction A major consideration in planning a manned mission in space is exposure of the space travelers to ionizing radiation. Mission planners want to minimize long-term health effects and prevent acute health effects that could jeopardize a mission. At the radiation doses likely to be received during space travel, the major health concern is fatal cancer – the latency period is in years [1,2]. However, space travelers who are inadequately shielded from radiation during a large solar-proton event may receive a dose sufﬁcient to cause serious health effects within hours or up to a couple of months after the event. 1 retired

RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07112-8

Published by Elsevier Ltd.

Health aspects of radiation exposure on a simulated mission to Mars

895

In the simulated mission described here, anticipated times of highest solar activity are avoided, and launch times are dates when Earth and Mars are positioned for the shortest travel time consistent with the minimum energy expenditure to travel between the 2 planets [3]. The length of the mission is approximately 2.7 years. In transit to and from Mars, the space travelers are exposed to charged particles trapped in the Earth’s magnetic ﬁeld, to galactic cosmic radiation (GCR), and to solar energetic particles (SEP) from a solar-proton event. During their stay on Mars, they are exposed to GCR. 2. Sources of ionizing radiation on Mars mission 2.1. Van Allen belts The Earth’s magnetic ﬁeld traps enormous numbers of charged particles in 2 broad overlapping bands that encircle the Earth. The bands are commonly referred to as the Van Allen radiation belts. The belt closest to the Earth, the inner belt, is characterized by the presence of trapped protons with energies in excess of 30 MeV [4]. Electrons are present in both the inner and outer belts and are the main type of particle in the outer belt. Trapped electrons are of negligible importance to manned missions, compared with the trapped protons. The inner belt is closest to the Earth in a region off the Brazilian coast called the South Atlantic Anomaly (SAA), which extends from about 0◦ to 60◦ W and 20◦ to 50◦ S (geographic coordinates) [4]. The SAA is the primary source of radiation exposure for travelers in spacecraft at low altitude and low orbital inclination [4,5]. With the notable exception of the SAA, the International Space Station orbits below the inner Van Allen belt. Trapped particles are almost completely absent at geomagnetic latitudes greater than 70◦ in both the northern and southern hemispheres [6]. 2.2. Galactic cosmic radiation GCR refers to the high-energy subatomic particles that permeate space and consist of 2% electrons and positrons, 85% protons, 12% helium nuclei, and 1% nuclei heavier than helium [5,7]. Essentially all of the GCR effective dose is produced by protons and heavier nuclei. GCR particles enter our Solar System from all directions. Particle energies outside the Earth’s atmosphere extend up to at least 1020 eV [8]. Supernovae are believed to be a main source of GCR. The Sun emits a hot, highly-ionized gas called the solar wind, which carries magnetic ﬁelds. Irregularities in these magnetic ﬁelds scatter GCR particles, thereby reducing the intensity of the GCR [5,9]. When solar activity is high, the solar wind carries more irregularities, resulting in more scattering and a corresponding greater reduction in the GCR intensity. The reduction in GCR intensity is most pronounced for the lower-energy GCR particles, particularly those with energies below 10 GeV [9]. 2.3. Solar energetic particles Often a disturbance in the Sun’s atmosphere results in an explosive emission of huge amounts of matter in the form of low-energy particles. Through a process called shock acceleration,

896

W. Friedberg et al.

this occasionally leads to a signiﬁcant ﬂux of high-energy particles [10]. The most important of these particles, in terms of radiation exposure of space travelers, are high-energy protons. The Space Environment Center of the US National Oceanic and Atmospheric Administration identiﬁes a solar event as a solar-proton event if measurements aboard a GOES satellite indicate that the mean ﬂux of protons with energies greater than 10 MeV equals or exceeds 10 protons/(cm2 srad s) for at least 3 consecutive 5-min periods. Large solar-proton events occur most frequently during about a 4 to 6 year period of highest activity in the Sun’s approximate 11-year solar activity cycle [7,11]. The term solar ﬂare is frequently used to indicate a solar-proton event, but it is also used to indicate other solar-phenomena [12]. Solar-proton events cannot be reliably predicted, nor is it known how high the radiation levels will reach even after an event has begun.

3. Dosimetry While in transit between Earth and Mars, the spacecraft is partially shielded from GCR by planetary bodies. This shielding reduces the mission dose by an extremely small amount, which is not taken into account. Nor is the radiation from charged particles in the Van Allen belts accounted for in the mission dose; its contribution to the mission dose would depend on how long the spacecraft spends in low-Earth orbit. The spacecraft shielding is equivalent to 2 g cm−2 of aluminum plus 4 g cm−2 of polyethylene. The Martian atmosphere is about 95% CO2 and varies in density with altitude [11]. While on Mars, the space travelers are shielded by the Martian atmosphere, modeled as 20 g cm−2 of CO2 , plus the spacecraft shielding. The space travelers’ activities outside the spacecraft while on Mars are not taken into account in the estimated amount of radiation received on the mission. The heliocentric potential used in the calculations to modulate the primary GCR spectrum for the entire mission is 490 MV. The 490 MV is based on the average neutron counting rate for 1997 measured at the Apatity division of the Polar Geophysical Institute in northern Russia. The year 1997 was the time of GCR maximum in the transition from solar activity cycle 22 to 23. Effective doses from primary GCR protons and from particles produced by their interactions with radiation shielding (the Martian atmosphere, when present, and the spacecraft) were calculated using LUIN2000 [13,14] modiﬁed for transport through aluminum, polyethylene, and CO2 . The program LUIN2000 incorporates the ﬂuence-to-effective dose conversion coefﬁcients of Ferrari et al. [15,20]. The transport of GCR heavy nuclei (heavier than protons) through radiation shielding and then to a depth of 5 cm in tissue was carried out with an ad hoc code. Nuclear breakup was handled using a modiﬁed Rudstam CDMD formula [21]. The effective dose from the heavy nuclei was calculated using their contribution to the absorbed dose at a depth of 5 cm in tissue and the radiation weighting factor (wR = 20) recommended by the International Commission on Radiological Protection in their Publication 60 [22]. The effective dose from particles other than heavy nuclei produced by the interactions of GCR heavy nuclei with the radiation shielding was calculated using conversion coefﬁcients of Ferrari et al. [15–20]. Table 1 outlines the 2.7-year simulated mission to Mars. The average daily effective dose to the space travelers from GCR while in transit is 0.0031 Sv. The effective dose of 0.17 Sv

Health aspects of radiation exposure on a simulated mission to Mars

897

Table 1 2.7-year mission to Mars (18 September 2007–20 May 2010) Phase of mission

Time spent (days)

Effective dose (sieverts)

Earth to Mars

280

On Mars

439

0.88 (GCR) 0.17 (SEP) 0.41 (GCR)

Mars to Earth Total

256 975

0.80 (GCR) 2.26

from SEP during the solar-proton event is actually an estimated dose equivalent to the bloodforming organs of astronauts on a simulated space mission during the August 1972 event [23]. (The astronauts were protected by consumables and waste during the event.) The dose from the 1972 event was received over 8 days; in the current study, it is assumed to have been received within 1 day. During a solar-proton event, the intensity of GCR is reduced [24]. Thus, the highest effective dose received by the space travelers within 1 day is about 0.17 Sv.

4. Health effects of radiation exposure 4.1. Early incapacitating effects In considering early incapacitating effects of radiation (Table 2), the main concerns are: (1) injury of blood-forming tissues, leading to blood loss and a weakened ability of the body to respond to infection, and (2) injury of the intestinal tract, leading to diarrhea, loss of body ﬂuids, and bacteria from the intestine gaining access to the blood stream [2]. The effect of radiation on the blood-forming tissues is largely dependent on the amount of damage to the bone marrow [7]. The principal locations of bone marrow are the pelvis, spine, ribs, and proximal ends of the bones of the extremities. The amount of damage to the bone Table 2 Early effects of whole-body exposure to ionizing radiation (based on low-LET radiation exposure received within 1 day, no medical treatment) Effective dose (sieverts)

Effects

<1 1–6

Symptoms, if any, not severe. Survival virtually certain [4,25]. Depending on the dose, varying proportions of exposed individuals will experience anorexia, nausea, vomiting, diarrhea, and easy fatigability, within 1 day after irradiation [1,4,26,27]. Deaths begin occurring after about 2 Sv [26]. Death rate 50% at about 2.5–4.5 Sv [2,26–28]. Fatally irradiated individuals usually die 9–60 days after exposure [29]. Survivors may take up to 2 years to recover [29]. Generally lethal [4,25,26].

>6

898

W. Friedberg et al.

Table 3 Risk of fatal cancer from ionizing radiation (2.26 Sv) received on the mars mission as related to age at exposure and gender Age at exposure (years)

Risk coefﬁcient (10−2 Sv−1 )

% mortality

Men

Women

Men

Women

25–34 35–44 45–54 55–64

4.32 3.13 2.01 1.04

7.38 5.28 3.40 1.82

9.8 7.1 4.5 2.4

16.7 11.9 7.7 4.1

25–64

Men or Women 4.0

Men or Women 9.0

marrow would be reﬂected in peripheral blood counts [7]. Based on the estimated highest dose received within 1 day (0.17 Sv) and the information in Table 2 on health effects from brief exposures, the space travelers will not experience incapacitating effects of radiation. 4.2. Lifetime risk of fatal cancer Two methods were used to estimate the lifetime risk of fatal cancer from ionizing radiation received on the Mars mission (Table 3). One of the methods takes into account the fact that the risk of radiation-induced cancer varies with age at exposure and gender [5]. Women have a higher risk than men of the same age because they live longer, which gives cancer a greater opportunity to develop, and because of their susceptibility to breast cancer and ovarian cancer. Based on the total effective dose of radiation received on the mission, estimates of percent mortality from cancer range from 2.4% for men of ages 55 through 64 (at the time of the mission) to 16.7% for women of ages 25 through 34. In the other method, space travelers are classiﬁed as a single group irrespective of age at exposure or gender, with the risk coefﬁcient for fatal cancer being 4 × 10−2 Sv−1 [22]. Based on this method, the expected cancer mortality from the radiation received on the Mars mission is 9.0%. In the general population of the United States, about 24% of adults (ages 20 years and older) will die of cancer [30]. 4.3. Cataracts Cataracts are an example of a radiation effect of concern to space travelers that would not jeopardize a mission. A cataract is a clouding of the normally transparent lens of the eye that reduces the amount of light reaching the retina and/or distorts the light rays. Usually, the result is impaired vision. The causes include changes in the lens with age, a variety of diseases (e.g. diabetes mellitus), intake of certain toxic chemicals, exposure to infrared and ultraviolet radiation, and exposure to ionizing radiation. Congenital cataracts can result from genetic abnormalities or from maternal disease during pregnancy. Radiation cataractogenesis has been reviewed extensively [5,7,24,27,31]. With low-LET radiation, a minimum dose of about 2 Sv given within about 1 day is required to produce

Health aspects of radiation exposure on a simulated mission to Mars

899

a cataract. For exposures spread over 3 weeks or more, the minimum dose may increase to 4 Sv or more. The time between irradiation and appearance of a cataract (latent period) is about 6 months to 35 years. The latent period becomes shorter with increasing dose. The threshold dose may be a matter of the ability to detect the beginning of cataract formation. A recent small epidemiology study of cataracts in NASA astronauts indicates that exposure to relatively low doses of space radiation increases the risk of cataracts [32]. Astronauts who received an average dose-equivalent to the lens of 0.045 Sv (> 0.008 Sv) had a higher incidence of cataracts than those who received an average lens dose of 0.0036 Sv (< 0.008 Sv). Based on studies with low-LET radiation, the Mars mission dose of 2.26 Sv is below the threshold (4 Sv) for radiation-induced cataracts. However, based on the astronaut study, the travelers on the mission will experience an increased risk of cataracts, presumably because of the greater biological effectiveness of space radiation. The threshold dose for cataracts is lower than that for other ocular lesions [1]. 4.4. Sterility A minimum dose to the testes in a single brief exposure of about 0.15 Sv of low-LET radiation may result in temporary reduced fertility [1,24,27]; 2.5 Sv may produce sterility for about 1 year or longer [31]. For permanent sterility, the minimum single brief exposure is about 3.5– 6.0 Sv [1,27]. Dose fractionation may increase or decrease the effectiveness of the radiation depending on the dose rate [5]. The minimum dose in a single brief exposure of the ovaries to produce permanent sterility has been reported to be 2.5–6 Sv [27]. However, for women over 40 there is some risk of ovarian failure from 1.5 Sv [5]. A single dose of low-LET radiation less than 1 Sv is not likely to have a long-term effect on fertility. Protraction or fractionation of the dose to the ovaries results in less injury than a single brief exposure [5]. Radiation-induced ovarian injury that leads to permanent sterility also results in hormonal changes comparable to those associated with natural menopause [5,27]. The maximum effective dose of 0.17 Sv received by the space travelers in a single day – during the solar-proton event – may result in some of the men experiencing temporary reduced fertility but the dose is unlikely to affect women. There has been no reported loss of libido in men or women from radiation exposures in the range of doses cited above [1]. Note that a single brief exposure of a major portion of the body to 2 Sv, in men or women, is likely to result in serious incapacitating illness and some deaths (Table 2). Thus, if the space travelers received sufﬁcient radiation to have a serious effect on their fertility, that would be the least of their concerns. 4.5. Genetic defects in offspring A child is at risk of inheriting genetic defects because of exposure to ionizing radiation by the parents before the child’s conception. The risk of genetic defects in ﬁrst-generation offspring has been estimated to be between 0.3 and 0.47 × 10−2 Sv−1 [33]. (Estimates based on low-LET radiation effects.) Therefore, if a space traveler on the Mars mission conceived a child some time after the mission is completed, the risk to that child of inheriting a genetic defect from the parent’s radiation exposure (2.26 Sv) on the mission would be between 0.7%

900

W. Friedberg et al.

and 1.1%. If both parents were exposed to radiation before the child was conceived, the risk to the child would be based on the total parental dose. About 6–7% of live-born children in the general population have congenital abnormalities (2–3% if only lethal and severe abnormalities are considered) [33].

5. Concluding remarks The amount of radiation received by the space travelers on the simulated mission to Mars translates into risks of radiation-induced fatal cancer, which, for all but men of ages 55–64, exceed the 3% career limit currently recommended by the National Council on Radiation Protection and Measurements for space crews on low-Earth orbit missions [5]. However, there is considerable uncertainty in the dose estimates and in the radiationinduced health risk coefﬁcients. The risk coefﬁcients are extrapolated from results in studies at higher doses and dose rates, and of lower energy and LET than the galactic cosmic radiation to which space travelers are exposed [5,13,22,34,35]. No data exists for proton- or heavy nuclei-induced cancer in humans [5]. There may be combined effects of radiation with other factors in the space environment.

References [1] NCRP, Guidance on radiation received in space activities, NCRP Report No. 98, National Council on Radiation Protection and Measurements, Bethesda, MD, 1989. [2] E.L. Travis, Primer of Medical Radiobiology, 2nd ed., Mosby, St. Louis, MO, 2000. [3] Planetary positions used to calculate Hohmann transfers, http://nssdc.gsfc.nasa.gov/space/helios/planet.html. [4] J.F. Parker Jr., V.R. West (Eds.), Bioastronautics Data Book, 2nd ed., NASA SP, vol. 3006, National Aeronautics and Space Administration, Science and Technical Information Ofﬁce, Washington, DC, 1973. [5] NCRP, Radiation protection guidance for activities in low-Earth orbit, NCRP Report No. 132, National Council on Radiation Protection and Measurements, Bethesda, MD, 2000. [6] The Illustrated Encyclopedia of Aviation and Space, AFE Press, Los Angeles, CA, 1971. [7] W.H. Langham (Ed.), Radiobiological Factors in Manned Space Flight, National Academy of Sciences, Washington, DC, 1967. [8] D.J. Bird, S.C. Corbato, H.Y. Dai, J.W. Elbert, K.D. Green, M.A. Huang, D.B. Kieda, S. Ko, C.G. Larsen, E.C. Loh, M.Z. Luo, M.H. Salamon, J.D. Smith, P. Sokolsky, P. Sommers, J.K.K. Tang, S.B. Thomas, Astrophys. J. 441 (1995) 144. [9] J.G. Wilson, Cosmic Rays, Wykeham, London, 1976. [10] D.C. Ellison, R. Ramaty, Astrophys. J. 298 (1985) 400. [11] J.W. Wilson, J. Miller, A. Konradi, F.A. Cucinotta (Eds.), Shielding Strategies for Human Space Exploration, NASA CP, vol. 3360, National Aeronautics and Space Administration, Langley Research Center, Hampton, VA, 1997. [12] D.F. Smart, M.A. Shea, Solar radiation, in: G.L. Trigg, E.S. Vera, W. Greulich (Eds.), Encyclopedia of Applied Physics, vol. 18, VCH, New York, 1997, pp. 393–429. [13] K. O’Brien, W. Friedberg, H.H. Sauer, D.F. Smart, Environ. Int. 22 (Suppl. 1) (1996) S9. [14] K. O’Brien, D.F. Smart, M.A. Shea, E. Felsberger, U. Schrewe, W. Friedberg, K. Copeland, Adv. Space Res., in press. [15] A. Ferrari, M. Pelliccioni, M. Pillon, Radiat. Prot. Dosim. 67 (4) (1996) 245. [16] A. Ferrari, M. Pelliccioni, M. Pillon, Radiat. Prot. Dosim. 69 (2) (1997) 97. [17] A. Ferrari, M. Pelliccioni, M. Pillon, Radiat. Prot. Dosim. 71 (2) (1997) 85.

Health aspects of radiation exposure on a simulated mission to Mars [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28]

[29]

[30] [31] [32] [33]

[34] [35]

901

A. Ferrari, M. Pelliccioni, M. Pillon, Radiat. Prot. Dosim. 71 (3) (1997) 165. A. Ferrari, M. Pelliccioni, M. Pillon, Radiat. Prot. Dosim. 74 (4) (1997) 227. A. Ferrari, M. Pelliccioni, M. Pillon, Radiat. Prot. Dosim. 80 (4) (1998) 361. G. Rudstam, Z. Naturforsch. A 21 (7) (1966) 1027. ICRP Publication 60: 1990 Recommendations of the International Commission on Radiological Protection, Ann. ICRP 21 (1–3) (1991). L.C. Simonsen, J.W. Wilson, M.H. Kim, F.A. Cucinotta, Health Phys. 79 (5) (2000) 515. National Research Council, Radiation Hazards to Crews of Interplanetary Missions: Biological Issues and Research Strategies, National Academy Press, Washington, DC, 1996. V.P. Bond, T.M. Fliedner, J.O. Archambeau, Mammalian Radiation Lethality: A Disturbance in Cellular Kinetics, Academic Press, New York, 1965. M.H. Beers, R. Berkow (Eds.), The Merck Manual of Diagnosis and Therapy, 17th ed., Merck Research Laboratories, Whitehouse Station, NJ, 1999. E.J. Hall, Radiobiology for the Radiologist, 5th ed., Lippincott Williams & Wilkins, Philadelphia, PA, 2000. F.A. Mettler Jr., A.K. Guskova, Treatment of acute radiation sickness, in: I.A. Gusev, A.K. Guskova, F.A. Mettler (Eds.), Medical Management of Radiation Accidents, 2nd ed., CRC Press, New York, NY, 2001, pp. 53– 67. A.K. Guskova, A.E. Baranov, I.A. Gusev, Acute radiation sickness: underlying principles and assessment, in: I.A. Gusev, A.K. Guskova, F.A. Mettler Jr. (Eds.), Medical Management of Radiation Accidents, 2nd ed., CRC Press, New York, NY, 2001, pp. 33–51. R.T. Greenlee, M.B. Hill-Harmon, T. Murray, M. Thun, CA Cancer J. Clin. 51 (2001) 15. F.A. Mettler Jr., Direct effects of radiation on speciﬁc tissues, in: I.A. Gusev, A.K. Guskova, F.A. Mettler Jr. (Eds.), Medical Management of Radiation Accidents, 2nd ed., CRC Press, New York, NY, 2001, pp. 69–131. F.A. Cucinotta, F.K. Manuel, J. Jones, G. Iszard, J. Murrey, B. Djojonegro, M. Wear, Radiat. Res. 156 (2001) 460. UNSCEAR (United Nations Scientiﬁc Committee on the Effects of Atomic Radiation), Hereditary Effects of Radiation, UNSCEAR 2001 Report to the General Assembly, with Scientiﬁc Annex, United Nations, New York, NY, 2001. S. Glasstone (Ed.), The Effects of Nuclear Weapons, rev. ed., United States Atomic Energy Commission, Washington, DC, 1964. NCRP, Uncertainties in fatal cancer risk estimates used in radiation protection, NCRP Report No. 126, National Council on Radiation Protection and Measurements, Bethesda, MD, 1997.

902

Observation of dose-rate increases during winter thunderstorms and Monte-Carlo simulation of bremsstrahlung generation T. Torii, M. Takeishi, K. Okubo Tsuruga Head Ofﬁce, Japan Nuclear Cycle Development Institute, 65-20 Kizaki, Tsuruga-shi, 914-8585, Japan

Increases of environmental gamma-ray dose that seem to originate from a thunderstorm were observed around a nuclear facility in Japan. Dose increases measured by thermoluminescent dosimeters (TLDs) exposed during a period including a large lightning ﬂash were up to about 0.1 mGy. Dose rates indicated by environmental radiation monitors around the facility were also increased transiently at that time. From pulse height analysis of a NaI(Tl) scintillation detector, the photon energy spectrum evaluated by spectral unfolding is continuous with energy up to several MeV, consistent with bremsstrahlung emission from energetic electrons. Furthermore, energy spectra of bremsstrahlung photons emitted from high-energy electrons in thunderstorm electric ﬁelds are obtained by Monte-Carlo simulation. These results are discussed in conjunction with observation. 1. Introduction Following C.T.R. Wilson’s suggestion [1] of electron acceleration by the strong electric ﬁelds in thunderclouds, a number of experiments were conducted in order to investigate whether or not energetic electrons and bremsstrahlung photons were generated by thunderstorm electric ﬁelds or lightning discharge processes. In recent years, enhanced radiation inside and above thunderclouds has been detected in experiments using scintillation detectors on a jet, an artiﬁcial satellite, and balloons, demonstrating that radiation is indeed associated with thunderstorm electric ﬁelds [2–5]. In winter, many thunderstorms occur on the west coast of Japan, and it has been suggested that gamma-ray dose may increase occasionally during winter thunderstorms. Recently, gamma-ray dose enhancement, which might be caused by the lightning activity were measured by environmental radiation monitors (ERMs) and TLDs around the site of the fast breeder reactor “Monju”, a nuclear power plant facing the Sea of Japan. In this paper, an occurrence of gamma-ray dose increase during lightning activity is presented, and a discussion RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07113-X

© 2005 Elsevier Ltd. All rights reserved.

Observation of dose-rate increases during winter thunderstorms and Monte-Carlo simulation

903

is provided on electron acceleration and bremsstrahlung generation in strong electric ﬁelds in simulated thunderclouds by a Monte-Carlo calculation.

2. Measurement During the last ﬁve years, dose-rate increases were observed at least four times by the ERMs (2 × 2 NaI(Tl) scintillation detector) around the Monju site. The largest increase occurred at 4:31 JST on January 29, 1997. During that night, there was a thunderstorm that lasted until dawn, and a large lightning ﬂash occurred. At that time, the dose-rate increase was observed by the ERMs; the dose-rate increase duration was several tenths of seconds. Furthermore, the absorbed doses recorded on TLDs (Panasonic UD-200S, CaSO4 :Tm) installed around the site were also greatly enhanced over usual variation ranges at the same points [6]. The monitoring points and the collected data by TLDs are shown in Figs. 1 and 2, respectively. The exposure periods of these particular TLDs were reduced compared to the usual of 3 months, and so the observed ﬂuctuations (average value ±3σ of dose measured for the last 5 ﬁscal years at each point) have been normalized to the reduced exposure period. As shown in Fig. 2, almost all of the dose measured at that time exceeded 3σ from the average. In particular, the doses in the northeastern area of the site were signiﬁcantly high, the dose increases being up to about 0.1 mGy. In Fig. 3 the dose-rate increases indicated by ERMs are also shown. All of these increases exceeded by far the usual ﬂuctuation, and in the case of ERM-1 the dose-rate increase indicated was over 5000 nGy h−1 . As illustrated in Fig. 3, the duration of dose-rate increase indicated by the ERMs was 40 to 60 seconds, and decreased over a period of 20 seconds. Since the time constant of the ERM rate-meters is 0.25 s above about 1000 nGy h−1 , dose rates should return to the background level within about 1 s even if the dose-rate rises instantaneously to several thousand nGy h−1 . Moreover, there is a shift in the initial rise of dose-rate and the peak between the ERMs in different locations. This may suggest that the acceleration of the electrons in the strong electric

Fig. 1. Monitoring points of the ERMs and TLDs around the Monju site.

Fig. 2. The doses (closed circle) are measured by including the time when the lightning occurred near the Monju site. Usual ranges are depicted as open rectangles.

904

T. Torii et al.

Fig. 3. Time variations of indicated dose rate on ERMs. The data displayed are for instantaneous dose rate every 10 seconds.

Fig. 4. The increased pulse height distribution during the lightning activity and the unfolded spectrum.

Fig. 5. Calculated photon spectra on the ground in the case where 5 and 10 MeV electrons are emitted in the sky.

ﬁeld in the thundercloud is a continuous process, which is ceased by a lightning ﬂash. Thus it seems that the dose rates are elevated as the thundercloud moves over the locations of the measuring instruments. At that time, a multi-channel analyzer (MCA) connected to the detector system of the ERM1 stored the pulse-height distribution at one-hour intervals. Figure 4 shows the net pulse-height distribution, which is the distribution including the time that the dose-rate increase occurred; background distribution is subtracted. From that pulse height distribution, we have obtained the energy spectrum of photons by an unfolding calculation using the SAND II code [7]. As depicted in Fig. 4, the unfolded result indicates a continuous spectrum with energy up to several MeV. Furthermore, we calculated the energy spectra of bremsstrahlung photons generated from energetic electrons emitted downwards at heights of 500 and 1000 m. This is the altitude of the base of winter thunderclouds. As shown in Fig. 5, these spectra are consistent with the above unfolded spectrum. However, as a result of the unfolding calculation, the dose increment evaluated from the energy spectrum was about 36 nGy. This result was much different from the dose increment

Observation of dose-rate increases during winter thunderstorms and Monte-Carlo simulation

905

measured by TLD (TL-E1), about 98.8 μGy, in the vicinity of the environmental radiation monitor (ERM-1). 3. Monte-Carlo simulation of bremsstrahlung generation In order to verify the electron acceleration and bremsstrahlung generation in thunderstorms, we have developed a user code of the Monte-Carlo electron and photon transport code system EGS4 [8] simulating electron behavior in air with external electric ﬁelds according to Bielajew’s manner [9]. We have also modeled electric ﬁelds in thunderclouds, and calculated the particle transport using this code [10]. To simplify the calculation, we assumed that the electric ﬁelds were uniform from the ground to an altitude of 1 km and that the ﬁeld strength varied from 100 to 280 kV m−1 . The atmospheric air density was assumed constant and equal to 1.293 kg m−3 . In such ﬁelds, the emission source was a million electrons with initial energy of 5 MeV heading downwards from an altitude of 1 km. Following these assumptions the energy spectra of photons were calculated. As illustrated in Fig. 6, generated photon numbers increased following the increase of the ﬁeld strength; large numbers of photons were emitted at the ﬁeld strength of 280 kV m−1 and over. Furthermore, varying the initial energies of electrons emitted in a constant electric ﬁeld, the avalanche phenomenon of electrons and photons in a strong electric ﬁeld became noticeable with the increase in the energy of incident electrons, as shown in Fig. 7. According to the results of these calculations, it is concluded that the dose enhancement is caused by the acceleration of energetic electrons in a thunderstorm electric ﬁeld and the subsequent generation of a large number of bremsstrahlung photons. 4. Discussion 4.1. Dose increase on TLDs and ERMs A dose increase far exceeding the normal range of ﬂuctuations was observed by ERMs during lightning activity. It is often observed during storms that there is an increase in dose rate due

Fig. 6. Photons spectra on the ground in electric ﬁelds (100–280 kV m−1 ).

906

T. Torii et al.

Fig. 7. The simulated tracks of the bremsstrahlung photons by the Monte-Carlo calculation. In this case, a hundred electrons with energies from 1 to 30 MeV at an altitude of 1 km are emitted downward in an external electric ﬁeld (200 kV m−1 ).

to gamma-ray emissions from radon progeny-ion decay, since the ions are precipitated to the ground by rainfall; however, from past data this never exceeds 200 nGy h−1 . Moreover, if the dose-rate increase were due to radon progeny, gamma-ray peaks of those decay products should be observable in the increment of pulse-height distribution; as indicated in Fig. 4, no such peak is found. Therefore, this increase in dose rate does not seem to be due to radon progeny. Although both the environmental radiation monitors and the respective nearby TLDs showed increased dose values, there was a large quantitative difference between the increments of dose evaluated from the unfolding spectrum and that on a nearby TLD. For example, ERM-1 showed 5.26 μGy h−1 , while the TL-E1 showed 98.8 μGy. At present, it is unknown what produced this difference, but three reasons are conceivable. The ﬁrst is that pulse-counting losses may have suddenly increased due to the dead time of the detector system following the photon burst. The next possible reason is the range of energy measured at the MCA. The measured energy range was from about 150 keV to 5 MeV, and the contribution from photons outside this range was not included in the evaluated dose, almost certainly leading to an underestimation. From the shape of the energy spectrum in Fig. 4, it can be assumed that there were photons with energy under 150 keV and over 5 MeV. The ﬁnal possibility is that the TLDs show different response from that of the scintillation

Observation of dose-rate increases during winter thunderstorms and Monte-Carlo simulation

907

detectors, when exposed to an electron shower occurring near the ground, in addition to the bremsstrahlung photons. 4.2. Elevated dose-rate observation on the ground in winter thunderstorms Dose and dose-rate increases far exceeding the normal range of ﬂuctuations were observed during lightning activity. These increases have been observed only during winter thunderstorms, not thunderstorms in summer. This may be because the cloud base is lower in winter thunderstorms, as low as several hundred meters, with higher electric ﬁeld intensity. If bremsstrahlung photons are emitted downwards at an altitude of several kilometers, as would occur in summertime conditions, dose increases on the ground would be very difﬁcult to detect due to attenuation. 4.3. Electron acceleration and the production of bremsstrahlung photons Here, we assumed a priori the existence of sufﬁciently energetic electrons to produce bremsstrahlung photons in thunderclouds. Although the answer to that question is not clear, secondary cosmic rays or β-rays emitted from airborne radionuclides can be considered as a possible source of the energetic electrons [1,11]. Bremsstrahlung photons at high altitude above the thundercloud may be explained by the theory of runaway breakdown of secondary cosmic rays [12,13], but it is not yet clear if this theory explains the downward emission of bremsstrahlung photons near the ground and with such a large increase of dose as that observed at the Monju site. However, as shown in the above calculation for near ground level, the energetic electrons as the initial seed can be accelerated in a strong electric ﬁeld like that of a thundercloud, and some secondary electrons are produced by the ionization process in air. Some of the produced electrons are decelerated by collision with air molecules, but the rest are accelerated and contribute to the generation of an electromagnetic shower. Since the frictional force of the collision has a minimum at approximately 1 MeV, it seems that the energy of the accelerated electrons ranges from several hundred keV to several tens of MeV. For high-energy electrons above hundreds of MeV, the thunderstorm electric ﬁeld is a minor perturbation. Consequently, as depicted in the above calculation, the electrons with energy of several MeV are mainly affected by the acceleration. In order to conﬁrm the source of such dose increases, we will carry out a further investigation on the transport of energetic electrons and the production of bremsstrahlung photons in thunderstorm electric ﬁelds. Acknowledgements We thank Mr. Sugita of SSL Inc. for his technical assistance, and our colleagues of Tsuruga Head Ofﬁce, JNC for their cooperation. References [1] C.T.R. Wilson, Proc. Cambridge Philos. Soc. 22 (1925) 534.

908

T. Torii et al.

[2] M. McCarthy, G.K. Parks, Geophys. Res. Lett. 12 (1985) 393. [3] G.J. Fishman, P.N. Bhat, R. Mallozzi, J.M. Horack, T. Koshut, C. Kouvelioou, G.N. Pendleton, C.A. Meegan, R.B. Wilson, W.S. Paciesas, S.J. Goodman, H.J. Christian, Science 264 (1994) 1313. [4] K.B. Eack, W.H. Beasley, W.D. Rust, T.C. Marshall, M. Stolzenburg, J. Geophys. Res. 101 (1996) 29637. [5] K.B. Eack, W.H. Beasley, W.D. Rust, T.C. Marshall, M. Stolzenburg, Geophys. Res. Lett. 23 (1996) 2915. [6] T. Torii, M. Takeishi, T. Hosono, J. Geophys. Res. 107 (D17) (2002) 4324. [7] W.N. McElroy, S. Berg, T. Crochett, R.G. Hawkins, AFWL-TR-67-41, Air Force Weapons Laboratory, 1967. [8] W.R. Nelson, H. Hirayama, D.W.O. Rogers, SLAC-265, SLAC, 1985. [9] A.F. Bielajew, in: Monte Carlo Transport of Electrons and Photons, Plenum, New York, 1988. [10] T. Torii, M. Takeishi, T. Hosono, T. Sugita, in: Proc. of the 2nd Int. Workshop on EGS, in: KEK Proceedings, vol. 2000-20, 2000. [11] M.P. McCarthy, G.K. Parks, J. Geophys. Res. 97 (1992) 5857. [12] A.V. Gurevich, G.M. Milikh, R. Roussel-Dupré, Phys. Lett. A 165 (1992) 463. [13] R. Roussel-Dupré, A.V. Gurevich, T. Tunnell, G.M. Milikh, Phys. Rev. E 49 (1994) 2257.

909

Measurements of cosmic radiation doses during air travel by a handy method with an electronic personal dosemeter M. Furukawa Research Center for Radiation Safety, National Institute of Radiological Sciences, 4-9-1 Anagawa, Inage-ku, Chiba 263-8555, Japan

In-ﬂight measurements with an Electronic Personal Dosemeter (EPD) have been conducted on many aviation routes, e.g. transpaciﬁc international ﬂights and domestic ones in Japan, etc., to estimate the individual cosmic radiation doses to aircraft crew and passengers during air travel. In addition, on a domestic aviation route in Japan, the ionizing and neutron components of the cosmic rays were concurrently measured with a NaI(Tl) scintillation spectrometer and a neutron counter, respectively. In order to convert the integrated doses measured with the EPD into realistic ones, tentative coefﬁcients were estimated to be 2.3, 3.1, and 4.6 for aviation routes through low (0–45 degrees), middle (45–60), and high (> 60) geographic latitude areas, respectively, based on the results of the concurrent measurements and reported data. The experimental results, including the converted values, indicate that the EPD is an effective and convenient instrument by which to assess the cosmic radiation doses during air travel.

1. Introduction Exposure due to cosmic rays during air travel is a typical case of technologically or artiﬁcially enhanced natural radiation. The cosmic ray intensity increases with increase in altitude in the atmosphere, so that the cosmic radiation dose rate at aircraft altitude is much higher than that at ground level [1,2]. The intensity of atmospheric cosmic radiation also varies with geomagnetic latitude and solar activity. Although it is extremely rare, the dose rate at subsonicand supersonic-airliner altitudes may increase signiﬁcantly due to enormous solar particle events [1,2]. The integrated cosmic radiation dose during air travel, therefore, depends on the individual ﬂight path and time, as well as on the speed of the aircraft. In particular, the exposure rate at cruising altitude is the main determinant factor of the dose. From the viewpoint of radiation safety, particular attention should be paid to the individual dose from cosmic radiation to aircraft crew, including passengers, because long distance RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07114-1

© 2005 Elsevier Ltd. All rights reserved.

910

M. Furukawa

air travel is increasing. The dose to passengers and crew varies from individual to individual and from path to path. To estimate the realistic dose, therefore, a convenient method for the in-ﬂight measurement should be developed. The purpose of this study is to examine the effectiveness of the handy method using Electronic Personal Dosemeters (EPD) in the assessment of the individual doses from cosmic radiation along various aviation routes.

2. Measurements Individual doses from cosmic rays during air travel have been measured with a pocket-sized EPD (Aloka PDM-101) weight about 60 g. The EPD has sensitivity to the ionizing component of cosmic rays. The value of accumulated dose is displayed on LCD with a minimum dose resolution of 0.01 μSv. Electric power is supplied by a coin-type lithium battery (CR2450), which has a continuous operational life of about one week. In-ﬂight measurements were conducted in the aircraft cabin, and the EPD was kept in a suite pocket during the air travel. The accumulated dose displayed on the LCD and the cruising altitudes were recorded at pre-determined intervals, including the ﬂight time from take off to landing. In order to estimate the realistic total doses during air travel, the ionizing and neutron components of the atmospheric cosmic radiation were concurrently measured with a multi-channel spectrometer with a 3 diameter × 3 cylindrical NaI(Tl) scintillation detector (Aloka JSM102) and a high-efﬁciency neutron dose-equivalent counter (Fuji NSN1), respectively, along a domestic aviation route in Japan. Data analyses on the ionizing component observed at 2 minutes interval were made for the energy spectral region from 3 to 7 MeV, and the cosmic radiation intensity was calculated as the effective dose rate (μSv h−1 ) and the integrated effective dose (μSv) from takeoff to landing. The accumulated dose values displayed on a LCD of the neutron counter with a minimum dose resolution of 0.001 μSv were also recorded at the same interval, and the effective dose rates were also calculated. In addition to these measurements, a comparison of the EPD data with those reported by the other methods has been made in this study.

3. Results and discussions From December 1992 to April 2002, in-ﬂight measurements with the EPD have been performed on about 200 ﬂights on international aviation routes and domestic routes in Japan, China, and Brazil [3–5]. Based on the data from a total of 85 such ﬂights, preliminary results for this experiment are described here. The locations of airports for the measured aviation routes are shown in Fig. 1. The in-ﬂight data obtained with the EPD, including ﬂight time from takeoff to landing, cruising altitude, etc., are summarized in Tables 1 and 2. The maximum value of integrated dose with the EPD, 15.98 μSv, was observed on a ﬂight route of JU14 from New York to Tokyo (Table 1). This agrees with the observations of other studies [2,6]. In the case of much the same aviation route, ﬂight time, and cruising altitude, e.g. Tokyo to Frankfurt (Table 1), Naha to Tokyo (Haneda) in Japan (Table 2), etc., almost similar doses were observed with the EPD. These experimental results strongly suggest that in-ﬂight

Measurements of cosmic radiation doses during air travel by a handy method

911

Fig. 1. Location of airports for the in-ﬂight measurements.

Table 1 Integrated doses due to cosmic rays for the international airlines Flight Route No. Origin

Destination

Japan–Asia JA01 NRT (Tokyo, Japan) JA02 NRT (Tokyo, Japan) JA03 PUS (Pusan, Korea) JA04 SEL (Seoul, Korea) JA05 NRT (Tokyo, Japan) JA06 NRT (Tokyo, Japan) JA07 PEK (Beijing, China) JA08 PEK (Beijing, China) JA09 PEK (Beijing, China) JA10 NRT (Tokyo, Japan) JA11 NRT (Tokyo, Japan) JA12 NRT (Tokyo, Japan) JA13 SHA (Shanghai, China) JA14 SHA (Shanghai, China) JA15 OSK (Osaka, Japan)

PUS (Pusan, Korea) PUS (Pusan, Korea) NRT (Tokyo, Japan) NRT (Tokyo, Japan) PEK (Beijing, China) PEK (Beijing, China) NRT (Tokyo, Japan) NRT (Tokyo, Japan) NRT (Tokyo, Japan) SHA (Shanghai, China) SHA (Shanghai, China) SHA (Shanghai, China) NRT (Tokyo, Japan) NRT (Tokyo, Japan) HKG (Hong Kong, China)

Year

Flight time∗ (min)

Cruising altitude (km)

1996 1998 1996 1998 1998 2001 1997 1998 2001 1993 1997 2001 1993 2001 1992

93 102 99 111 189 180 180 170 161 172 166 162 139 135 236

10.7 8.5 11.3 − 10.8 11.3–12.0 − − 11.4 10.7 9.4 11.2 13.0

Integrated Calculated dose† dose‡ (μSv) [coefﬁcient] (μSv) 0.82 0.64 0.82 1.05 2.06 1.94 1.85 1.60 1.87 1.40 1.50 1.33 1.23 1.19 1.79

1.9 [2.3] 1.5 [2.3] 1.9 [2.3] 2.4 [2.3] 4.7 [2.3] 4.5 [2.3] 4.3 [2.3] 3.7 [2.3] 4.3 [2.3] 3.2 [2.3] 3.5 [2.3] 3.1 [2.3] 2.8 [2.3] 2.7 [2.3] 4.1 [2.3]

(continued on next page)

912

M. Furukawa

Table 1 (continued) Flight Route No. Origin

Year

Flight Cruising time∗ altitude (min) (km)

1992 1997 1996 2001 1993 1996 1997

174 169 382 403 383 367 525

1996 1999 2000 2000 1996 1999 2000 2000 1998 2000 1998 1998 1998 1999 1999

1996 1997 1997 1999 1996 1997 1997 1999 2002 1997 1997

Destination

JA16 HKG (Hon Kong, China) OSK (Osaka, Japan) JA17 HKG (Hong Kong, China) KIX (Kansai, Japan) JA18 NRT (Tokyo, Japan) SIN (Singapore) JA19 NRT (Tokyo, Japan) SIN (Singapore) JA20 SIN (Singapore) NRT (Tokyo, Japan) JA21 SIN (Singapore) NRT (Tokyo, Japan) JA22 KIX (Kansai, Japan) DEL Pelhi, India) Japan–Europe JE01 NRT (Tokyo, Japan) VIE (Vienna, Austria) JE02 NRT (Tokyo, Japan) FRA (Frankfurt, Germany) JE03 NRT (Tokyo, Japan) FRA (Frankfurt, Germany) JE04 NRT (Tokyo, Japan) FRA (Frankfurt, Germany) JE05 FRA (Frankfurt, Germany) NRT (Tokyo, Japan) JE06 FRA (Frankfurt, Germany) NRT (Tokyo, Japan) JE07 FRA (Frankfurt, Germany) NRT (Tokyo, Japan) JE08 FRA (Frankfurt, Germany) NRT (Tokyo, Japan) JE09 NRT (Tokyo, Japan) CDG (Pans, France) JE10 NRT (Tokyo, Japan) CDG (Pans, France) JE11 CDG (Pans, France) NRT (Tokyo, Japan) JE12 NRT (Tokyo, Japan) HEL (Helsinki, Finland) JE13 HEL (Helsinki, Finland) NRT (Tokyo, Japan) JEW NRT (Tokyo, Japan) LHR (London, UK) JE15 AMS (Amsterdam, NRT (Tokyo, Japan) Netherlands) Japan–North America (West Coast) JU01 NRT (Tokyo, Japan) LAX (Los Angeles, USA) JU02 NRT (Tokyo, Japan) LAX (Los Angeles, USA) JU03 NRT (Tokyo, Japan) LAX (Los Angeles, USA) JU04 NRT (Tokyo, Japan) LAX (Los Angeles, USA) JU05 LAX (Los Angeles, USA) NRT (Tokyo, Japan) JU06 LAX (Los Angeles, USA) NRT (Tokyo, Japan) JU07 LAX (Los Angeles, USA) NRT (Tokyo, Japan) JU08 LAX (Los Angeles, USA) NRT (Tokyo, Japan) JU09 LAX (Los Angeles, USA) NGO (Nagoya, Japan) JU10 NRT (Tokyo, Japan) SFO (San Francisco, USA) JU11 SFO (San Francisco, USA) NRT (Tokyo, Japan) Japan–North America (East Coast) JU12 NRT (Tokyo, Japan) JFK (New York, USA) JU13 NRT (Tokyo, Japan) JFK (New York, USA) JU14 JFK (New York, USA) NRT (Tokyo, Japan) North America–South America UB01 LAX (Los Angeles, USA) GRU (Sao Paulo, Brazil) UB02 LAX (Los Angeles, USA) GRU (Sao Paulo, Brazil) UB03 LAX (Los Angeles, USA) GRU (Sao Paulo, Brazil) UB04 LAX (Los Angeles, USA) GRU (Sao Paulo, Brazil) UB05 JFK (New York, USA) GRU (Sao Paulo, Brazil)

Integrated Calculated dose† dose‡ (μSv) [coefﬁcient] (μSv) 1.56 1.64 3.88 3.78 2.94 3.91 5.10

3.6 [2.3] 3.8 [2.3] 8.9 [2.3] 8.7 [2.3] 6.8 [2.3] 9.0 [2.3] 11.7 [2.3]

706 756 685 647 625 630 632 642 679 705 754 633 540 738 581

12.80 13.25 12.58 12.27 15.24 11.87 11.15 12.29 14.96 10.7–11.9 15.35 14.85 10.7 11.10 10.7 9.73 − 14.31 − 11.79

58.9 [4.6] 61.0 [4.6] 57.9 [4.6] 56.4 [4.6] 70.1 [4.6] 54.6 [4.6] 51.3 [4.6] 56.5 [4.6] 68.8 [4.6] 70.6 [4.6] 68.3 [4.6] 51.1 [4.6] 44.8 [4.6] 65.8 [4.6] 54.2 [4.6]

497 513 564 530 669 653 629 683 663 506 616

10.0 − − − 11.0 − − − 10.4–11.3 11.3 11.6

8.98 11.00 9.35 9.99 12.79 13.15 11.77 12.41 12.31 10.22 13.33

27.8 [3.1] 34.1 [3.1] 29.0 [3.1] 31.0 [3.1] 39.6 [3.1] 40.8 [3.1] 36.5 [3.1] 38.5 [3.1] 38.2 [3.1] 31.7 [3.1] 41.3 [3.1]

1999 704 2002 700 1999 821

9.4–11.3 12.92 10.1–11.3 12.43 9.4–11.9 15.98

59.4 [4.6] 57.2 [4.6] 73.5 [4.6]

1996 1997 1997 1999 2002

− − − 10.1 11.3

20.8 [2.3] 23.0 [2.3] 21.3 [2.3] 18.6 [2.3] 16.9 [2.3]

642 658 664 662 522

10.7 11.9 9.3–10.1 11.3 10.8 10.7 − 10.6 10.7 10.1–12.5 11.1 10.1 10.1–12.0

9.04 10.00 9.28 8.10 7.33

(continued on next page)

Measurements of cosmic radiation doses during air travel by a handy method

913

Table 1 (continued) Flight Route No. Origin

Destination

UB06 UB07 UB08 UB09 UB10

LAX (Los Angeles, USA) LAX (Los Angeles, USA) LAX (Los Angeles, USA) LAX (Los Angeles, USA) LAX (Los Angeles, USA)

GRU (Sao Paulo, Brazil) GRU (Sao Paulo, Brazil) GRU (Sao Paulo, Brazil) GRU (Sao Paulo, Brazil) GRU (Sao Paulo, Brazil)

Integrated Calculated dose† dose‡ (μSv) [coefﬁcient] (μSv)

Year

Flight time∗ (min)

Cruising altitude (km)

1996 1997 1997 1999 2002

695 681 673 671 687

− 10.80 − 10.56 − 9.87 − 9.18 10.7–11.9 8.93

24.8 [2.3] 24.3 [2.3] 22.7 [2.3] 21.1 [2.3] 20.5 [2.3]

∗ From takeoff to landing. † Measured with the EPD. ‡ Integrated dose × coefﬁcient.

Table 2 Integrated doses due to cosmic rays for the domestic airlines in Japan Route Origin

Destination

HMD (Haneda) WKN (Wakkanai) HMD (Haneda) HMD (Haneda) HMD (Haneda) HMD (Haneda) AOJ (Aomori) AOJ (Aomori) AOJ (Aomori) AOJ (Aomori) AOJ (Aomori) AOJ (Aomori) HMD (Haneda) HMD (Haneda) FUK (Fukuoka) FUK (Fukuoka) HMD (Haneda) HMD (Haneda) HMD (Haneda) HMD (Haneda) HMD (Haneda) HMD (Haneda) KOJ (Kagoshima) KOJ (Kagoshima) KOJ (Kagoshima) KOJ (Kagoshima) KOJ (Kagoshima)

WKN (Wakkanai) HND (Haneda) AOJ (Aomori) AOJ (Aomori) AOJ (Aomori) AOJ (Aomori) HND (Haneda) HND (Haneda) HND (Haneda) HND (Haneda) HND (Haneda) HND (Haneda) FUK (Fukuoka) FUK (Fukuoka) HND (Haneda) HND (Haneda) KOJ (Kagoshima) KOJ (Kagoshima) KOJ (Kagoshima) KOJ (Kagoshima) KOJ (Kagoshima) KOJ (Kagoshima) HND (Haneda) HND (Haneda) HND (Haneda) HND (Haneda) HND (Haneda)

Year

Flight time∗ (min)

Cruising altitude (km)

Integrated dose† (μSv)

Calculated dose‡ (μSv)

1999 1999 1999 1999 2001 2002 1999 1999 2000 2001 2002 2002 2000 2001 1999 2001 1996 1999 2000 2001 2002 1999 1996 1999 2000 2001 2002

91 98 60 57 51 60 56 63 55 60 60 56 90 87 74 72 78 88 101 98 95 88 88 80 72 72 73

11.0 12.0 7.3 7.0 10.0 11.0 7.3 6.0 8.5 8.0 9.4 10.7 7.9 7.5 11.5 12.5 11.0 10.7 11.9 10.7 11.9 9.2 9.0 11.3 11.3 11.3 11.3

0.87 1.30 0.32 0.25 0.33 0.36 0.29 0.15 0.25 0.32 0.25 0.38 0.44 0.61 0.67 0.62 0.77 0.75 1.14 1.10 0.99 0.65 0.60 0.72 0.52 0.76 0.57

2.0 3.0 0.7 0.6 0.8 0.8 0.7 0.4 0.6 0.7 0.6 0.9 1.0 1.4 1.5 1.4 1.8 1.7 2.6 2.5 2.3 1.5 1.4 1.7 1.2 1.8 1.3

(continued on next page)

914

M. Furukawa

Table 2 (continued) Route Origin

Destination

HMD (Haneda) HMD (Haneda) HMD (Haneda) HMD (Haneda) HMD (Haneda) HMD (Haneda) HMD (Haneda) OKA (Naha) OKA (Naha) OKA (Naha) OKA (Naha) OKA (Naha) OKA (Naha) OKA (Naha) OKA (Naha) OKA (Naha)

OKA (Naha) OKA (Naha) OKA (Naha) OKA (Naha) OKA (Naha) OKA (Naha) OKA (Naha) HND (Haneda) HND (Haneda) HND (Haneda) HND (Haneda) HND (Haneda) HND (Haneda) HND (Haneda) HND (Haneda) HND (Haneda)

Year

Flight time∗ (min)

Cruising altitude (km)

Integrated dose† (μSv)

Calculated dose‡ (μSv)

1994 1999 1999 2001 2001 2002 2002 1994 1996 1999 1999 2000 2001 2001 2001 2002

140 120 190 158 156 147 133 114 107 112 97 121 111 103 105 117

11.9 10.6 12.0 9.4 8.5 11.9 10.7 10.5 11.3 11.5 11.3 11.3 11.3 11.3 11.3 10.0

1.63 1.08 2.18 1.55 0.99 1.74 1.23 1.05 0.95 1.17 0.93 1.12 1.00 0.86 0.99 0.92

3.8 2.5 5.0 3.4 2.3 4.0 2.8 2.4 2.2 2.7 2.1 2.6 2.3 2.0 2.3 2.1

∗ From takeoff to landing. † Measured with the EPD. ‡ Integrated dose × coefﬁcient (2.3).

Table 3 Cosmic radiation doses measured with a neutron counter and a NaI(Tl) scintillation spectrometer Route

Integrated dose (μSv) Ratio (a + b)/c [Average dose rate cruising altitude∗ , μSv h−1 ] a: neutron b: ionizing c: ionizing component component component with EPD

HND (Haneda) → WKN (Wakkanai) 0.49 [0.48] WKN (Wakkanai) → HND (Haneda) 0.79 [0.72]

1.56 [1.58] 2.18 [2.04]

0.87 1.30

2.36 2.28

Coefﬁcient

2.3

∗ Cruising altitudes are shown in Table 2.

measurement with the EPD is a useful method to compare the relative doses for individual ﬂights. For the concurrent measurements with a neutron counter and a NaI(Tl) scintillation detector between Haneda and Wakkanai in Japan (Fig. 1), the dose rates due to neutron and ionizing components at cruising altitude (Table 3) agreed well with the results of other studies conducted in Japan [7,8]. In order to convert the integrated doses measured with the EPD into realistic values, a coefﬁcient was estimated at 2.3 based on the results of the concurrent measurements (Table 3). However, because the cosmic radiation dose rate varies with geomagnetic latitude, the coefﬁcient 2.3 has limited application in the low geographic latitude area (0–45 degrees). Then tentative coefﬁcients for other areas were estimated from the com-

Measurements of cosmic radiation doses during air travel by a handy method

915

Table 4 Integrated doses due to cosmic radiation during air travel Route Origin

Destination

Integrated dose (μSv) a: reported data∗ b: this study

Ratio (a/b)

NRT (Tokyo, JAPAN) LAX (Los Angeles, USA)

LAX (Los Angeles, USA) NRT (Tokyo, JAPAN)

30.0 (24.6–32.2) 38.0 (31.8–40.4)

9.83† 12.53†

3.1 3.0

NRT (Tokyo, JAPAN) JFK (New York, USA)

JFK (New York, USA) NRT (Tokyo, JAPAN)

63.5 (44.3–74.8) 67.1 (48.3–77.7)

12.68† 15.98

5.0 4.2

Coefﬁcient 3.1 4.6

∗ Mean (minimum–maximum) dose due to ionizing and neutron components estimated by model calculation [9]. † Mean dose due to ionizing component measured with the EPD.

parison of the EPD and reported data [9] (Table 4). In this paper, 3.1 and 4.6 were regarded to be coefﬁcients for the aviation routes which pass through middle (45–60) and high (> 60) geographic latitude areas, respectively. The doses converted by these tentative coefﬁcients are shown in Tables 1 and 2. It is considered that the agreement between the converted values for the EPD data and the other reported results by model calculation [6] was within error of less than about 20%. Although the in-ﬂight measurements and the establishment of the handy method are now in progress, the results of this study indicate that the EPD is an effective and convenient tool to assess the cosmic radiation doses during air travel. However, interpreting the data obtained with the EPD in terms of the realistic effective dose requires (1) further measurements with different systems on many ﬂights to improve the method using the EPD and (2) knowledge of the overall properties of the cosmic radiation ﬁelds in the atmosphere, including information on solar activity.

References [1] UNSCEAR, Sources and Effects of Ionising Radiation, UNSCEAR 2000 Report to the General Assembly, with Scientiﬁc Annexes, United Nations, New York, 2000. [2] M. Kelly, et al. (Eds.), Radiat. Prot. Dosim. 86 (1999). [3] M. Furukawa, Radioisotopes 49 (2000) 152. [4] M. Furukawa, Radioisotopes 50 (2001) 282. [5] M. Furukawa, Radioisotopes 50 (2001) 591. [6] M. Kai, et al., J. Health Phys. 27 (1992) 289. [7] T. Nakamura, et al., Health Phys. 53 (1987) 509. [8] M. Okano, et al., in: Proceedings of IRPA 9th, vol. 2, 1996, p. 262. [9] W. Freidberg, et al., Radiat. Prot. Dosim. 86 (1999) 323.

916

Twenty years of TLD measurements on board space vehicles by the Hungarian “Pille” system S. Deme a , I. Fehér a , I. Apáthy a , G. Reitz b , Yu. Akatov c a KFKI Atomic Energy Research Institute, PO Box 49, H-1525 Budapest, Hungary b DLR/FF-ME, Abteilung Strahlenbiologie, Linder Hoehe, D-51147 Koeln, Germany c Institute of Biomedical Problems, 123007 Moscow, Russia

KFKI Atomic Energy Research Institute (KFKI AEKI) has developed and manufactured a series of thermoluminescent dosimeter systems for measuring cosmic radiation doses in the 10 μGy to 10 Gy range, consisting of a set of bulb dosimeters and a small, compact, TLD reader suitable for on-board evaluation of the dosimeters. By means of such a system, highly accurate measurements were carried out on board the Salyut-6, -7 and Mir Space Stations as well as on the Space Shuttle and the International Space Station (ISS). A detailed description of the system is given and the comprehensive results of these measurements are summarized. 1. Introduction One of the many risks of long duration space ﬂights is the dose burden from cosmic radiation, especially during periods of intensive solar activity. At such times, particularly during EVAs, (extravehicular activities) when the wall of the spacecraft does not protect the astronauts, cosmic radiation is a potentially serious health threat. Accurate dose measurement becomes increasingly important during the assembly of large space objects. On-board personal dose measurements are mainly based on thermoluminescent dosimetry. Because of the large dimensions and big mass of the readers, the TLDs used in space activities are generally evaluated only after their return to the ground, in terrestrial laboratories. The disadvantage of on-ground evaluation is that it results in the dose accumulated since the last read-out, i.e. the dose of the whole ﬂight, whereas with the increased duration of space ﬂights periodic and relatively frequent dose measurements would be needed. A small, portable, and space-qualiﬁed TLD reader suitable for reading out the TL dosimeters on board provides the possibility of overcoming the above-mentioned disadvantage. Such a system offers a solution for EVA dosimetry as well. Since the end of the 1970s, KFKI AEKI has developed and manufactured speciﬁcally for spacecraft a series of TLD systems named “Pille” (in English: butterﬂy). Such systems consist RADIOACTIVITY IN THE ENVIRONMENT VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07115-3

© 2005 Elsevier Ltd. All rights reserved.

Twenty years of TLD measurements on board space vehicles by the Hungarian “Pille” system

917

of a set of TL dosimeters and a compact TLD reader suitable for on-board evaluation of the dosimeters. By means of such systems, highly accurate measurements were and are carried out on board the Salyut-6, -7 and MIR Space Stations as well as on the Space Shuttle. The newest implementation of the system has been installed on the International Space Station (ISS).

2. The “Pille” TLD system 2.1. The very ﬁrst version This version of “Pille” was developed for the Hungarian cosmonaut B. Farkas, and he used it during his ﬂight in 1980 on board the Salyut-6 Space Station [1]. Subsequently the same system was repeatedly used for on-board dosimetry by Soviet cosmonauts on Salyut-6 until 1983. The system consisted of a small TLD reader utilizing on-board power, and several bulb dosimeters containing CaSO4 :Tm thermoluminescent material. The measuring dose range of the system was from 10 μGy to 100 mGy and the measured values were displayed and manually recorded. 2.2. Improvement of the ﬁrst version In 1983, the Pille TLD reader was upgraded to achieve a measuring range two orders of magnitude wider (10 μGy–10 Gy) [2]. The upgraded bulb dosimeters were based on CaSO4 :Dy TL material whose supralinearity beyond 1 Gy was less than that of CaSO4 :Tm. The system on board Salyut-6 was replaced by the improved one, and the latter was relocated to Salyut-7 [3] and later to the Mir Space Station, having been used by Soviet cosmonauts. This improved system, including a battery-operated version of the reader, was subsequently used on board a Space Shuttle during the 41-G NASA mission in 1984 by astronaut S. Ride [4]. 2.3. New microprocessor controlled version The development of a brand new version containing a microprocessor-controlled reader with automatic dosimeter identiﬁcation, logging on a memory card, PC-connection, etc., to provide the basis for an up-to-date TLD system for space dose measurements, started in 1993 [5]. The new system appears ﬁrst on board the Mir Space Station as part of ESA’s Euromir’95 mission (1995–1996) [6], while an improved version was used on the same space station in the framework of the NASA4 experiment (1997) [7] measuring – as far as we know for the ﬁrst time in the history of NASA – the dose received by astronauts during an EVA. 2.4. Technical description of the system The thermoluminescent dosimeter (TLD) system [5] consists of numerous bulb TL dosimeters and a lightweight, compact portable TLD reader, suitable for reading out and evaluating the dosimeters at the place of exposure, i.e. on board a spacecraft.

918

S. Deme et al.

Fig. 1. Cross-section of the Pille TL dosimeter.

Figure 1 shows the cross-section of the TL dosimeter. Its essential component is a small vacuum bulb made of glass (a), containing the TL material (b) glued to the surface of a resistive metal plate (c) that is heated electrically. Each TLD bulb is encapsulated in a cylindrical, pen-like metal holder. In the new version, a one-wire-port integrated electronic programmable memory chip (d) mounted inside the holder contains the identiﬁcation code, the individual calibration parameters of the dosimeter, and the date and time of the actual read-out. In the new version, the aperture (e) of the holder is covered by a tube (f) to protect the bulb from light and mechanical effects and to prevent the operator from touching the hot bulb just after read-out. The tube slips backwards automatically when the dosimeter is inserted into the reader. Three gold-plated contacts (g) on one end of the holder provide a lead-in for the heating current and access to the memory chip. A milled-edged head (h) at the other end serves for operation during read-out. Except during read-out, the dosimeter is held within a protective metal case. The newest dosimeters have a diameter of 20 mm, a length of 60 mm, and a mass (together with the carrying case) of 70 g. The TLD reader is based on analogue timing circuits and digital integrated circuits operating in sequential logic in the old generation and on a microprocessor controlling all functions in the new generation. In the old version only the measured dose was displayed. In addition to the measured dose, the new version displays a series of parameters (dosimeter identiﬁer, date and time of the actual and the previous read-out, etc.) and these are stored on a removable ﬂash memory card, which can store data of up to 8000 measurements. The reader is powered through the central power line of the spacecraft; its maximum power consumption (during the period of heating) is less than 10 W, its standby power consumption (new version) is less than 1 W. By virtue of its construction, the reader resists mechanical impacts during launch, and it fulﬁlls space requirements. The mass of the reader is in the range of 1–1.4 kg, depending on the version. 2.5. System operation of the new generation The front panel has an eight-character-wide LED display, memory card slot, hole for the dosimeter, and ﬁve pushbuttons for controlling the reader. The TLD reader has two main modes of operation, manual and automatic. In manual mode, the user can read out the dosimeters, check or set the real time clock, check or set the parameters of the automatic mode, recall the data of any previous read-out, and initiate automatic mode. For manual readings, the dosimeter is inserted into the lightproof compartment of the reader. If one twists the dosimeter, the reading procedure is started automatically. After

Twenty years of TLD measurements on board space vehicles by the Hungarian “Pille” system

919

a measurement, the TL dose in exponential form and the measuring parameters can be indicated one after the other on the display. All these data and the glow curve are stored on the memory card. The reader can operate in automatic data acquisition mode using a dosimeter that remains in the reader. This dosimeter will be automatically read out periodically on the basis of the parameters set in previously. During the time intervals between read-outs, the reader is in sleep mode. In automatic mode, the in-built timer switches on the reader only during read-out. The timer is powered by a small stand-by supply during sleep mode; this requires the very low power consumption of less than 0.1 W. The reader can be connected to a personal computer (PC) via its RS-232 standard serial port. In this way, the parameters can be programmed into the reader and into the dosimeter inserted in the reader, data can be read from the memory card, and service functions can be accomplished. Optionally, the reader can be supplemented by an interface port of any other standard in order to connect it to a local on-board computer network providing remote control and data read-out. 2.6. Dosimetric characteristics of the system For high sensitivity measurements, CaSO4 :Dy is used as the TL material. The measuring dose range of the system with CaSO4 :Dy bulbs is from 3 μGy to 10 Gy at an accuracy level of < 10%; above 10 μGy the accuracy of measurements is < 5%. The read-out precision of the reader is 3 digits + exp.

3. Earlier measurements on board Salyut-6, Salyut-7, and the Space Shuttle 3.1. Salyut-6 The Pille dosimetric system [1] was developed in Hungary for use in international collaborations and was ﬁrst used on the Salyut-6 Space Station in 1980 by the Hungarian cosmonaut B. Farkas and then, by Soviet cosmonauts [3]. The main goal of the dosimetric experiments with the Pille TLD system on Salyut-6 and Salyut-7 was to study the dose distribution inside the space stations and to determine the personal dose of the crew. The results of the ﬁrst experiment on Salyut-6 are presented in Table 1. The next measurement by the Pille system on Salyut-6 was performed 350 days later – on May 18, 1981. The integral dose for this period was found to be 35.0 mGy, i.e. 4.2 μGy h−1 [3]. 3.2. Salyut-7 In the experiments on Salyut-7 the new, upgraded Pille system was used, with a sensitivity of 1 μGy/digit. The complete experimental program carried out in 1983 included the measurement of the dose ﬁeld in the inhabited sections of the station at 13 sites and the determination of the individual doses of the two cosmonauts for various periods of the long-term ﬂight, the results presented in Table 2 [3].

920

S. Deme et al. Table 1 Results of dose ﬁeld measurements on Salyut-6 from May 28 to June 2, 1980 TLD No.

Location

Dose rate (μGy h−1 )

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.

on space suit of Cosmonaut 1 on space suit of Cosmonaut 2 on space suit of Cosmonaut 3 on space suit of Cosmonaut B. Farkas passage, location No. 1 passage, location No. 2 passage, location No. 3 passage, location No. 4 working area, location No. 1 working area, location No. 2 sleeping place No. 1 sleeping place No. 2

3.4 3.6 3.5 3.4 4.5 3.9 4.5 3.7 3.8 2.8 4.5 4.3

Table 2 Results on dose distribution in the inhabited sections of the Salyut-7 measured in 1983 Location

Personal Working area Passage Sleeping place Average

No. of dosimeters

Dose rate in μGy h−1 07.21.–08.26.83

08.26.–09.23.83

09.23.–11.11.83

2 8 3 2 15

5.0–6.0 5.4–6.8 6.3–6.7 5.9–7.3 6.2

6.1–6.7 6.2–8.5 7.5–9.4 7.7–8.8 7.5

4.9–6.0 5.5–7.3 6.5–8.8 6.6–7.7 6.7

Table 3 Total doses in μGy (air) measured at six different dosimeter locations of the Space Shuttle by passive TLDs of the Johnson Space Center (JSC) and by the Pille system Item

Total dose, measured by JSC (μGy) Total dose, measured by Pille (μGy) Difference (%)

Dosimeter location 1

2

3

4

5

6

776 780 0.5

955 967 1.2

839 832 0.8

824 835 1.3

822 829 0.9

871 900 3.3

3.3. Space Shuttle The objective of this experiment was to place the Pille dosimeters in predetermined locations and to read them out periodically during the mission 41G [4]. The mission was ﬂown at an inclination of 57◦ at three altitudes, 352, 274 and 224 km. The mission duration was 197 h. For each of these altitudes two readings of the Pille dosimeters were performed by mission specialist S. Ride. The mean dose rates at 220 km altitude were 85 μGy d−1 , at 270 km 145 μGy d−1 and at 350 km about 240 μGy d−1 . All NASA dosimeters were read out inde-

Twenty years of TLD measurements on board space vehicles by the Hungarian “Pille” system

921

pendently with a standard commercial reader. The single Pille readings for each location were summed up and compared with the results of the NASA total dose measurements for the same locations. Table 3 shows the results and an excellent agreement of both data sets. 4. Russian measurements on Mir In June 1986, the reader of the Pille TLD system previously used on the Salyut-7 Space Station was transported by cosmonauts L. Kizim and V. Soloviev, the very ﬁrst crew of Mir, from the old space station to the new one. A new set of Pille dosimeters arrived to Mir on board the Progress-28 transporting spacecraft in March 1987. During that year, in a period of minimum solar activity, the system was utilized by the second crew for dosimetry mapping the Mir. The results of their measurements are summarized in Table 4. For the ﬁrst time in space history, the Pille system was used to measure the exposure of cosmonauts during their extravehicular activity. Such measurements were carried out twice (on June 12 and 16, 1987) by Y. Romanenko, the commander of the second crew of Mir, during his EVAs. During the EVA, one of the dosimeters was ﬁxed in a pocket on the outer surface of the left leg of his space suit; a second dosimeter was located inside the station for reference measurements. Out of the EVA period both dosimeters were stored at the same location inside the space station and they were read out before leaving and after returning to it. The measured dose rates were 144 and 196 μGy h−1 accordingly. 5. ESA measurements Pille’95, a brand new version of the on-board TLD system of KFKI AEKI, consisting of a microprocessor controlled reader and CaSO4 :Dy bulb dosimeters, was used by ESA astronaut T. Reiter on board the Mir Space Station during the Euromir ’95 mission in 1995–1996 [6]. The measurements were part of ESA’s space dosimetry experiment D-18, led by G. Reitz (DLR, Cologne, Germany) [8]. The dosimeters were exposed at six locations of different shielding around them. Five of them were read out manually once a week; one of them was left in the TLD-Reader to perform frequent measurements (24-hourly and, later, hourly) in automatic mode. Table 4 First dosimetry mapping results on Mir (1987) Location of dosimeters

Cabin of the commander Cabin of the board engineer Lavatory Passage Storage place of photography ﬁlms

Mean dose rate μGy h−1 15.03–23.04

15.05–22.06

02.07–30.07

30.07–10.10

10.11–01.12

Average

14.2 13.3

15.0 11.7

11.7 11.3

10.4 8.8

10.8 7.9

12.4 ± 2.0 10.6 ± 2.2

9.2 12.5 8.3

8.8 14.2 8.3

8.8 12.9 7.5

7.5 10.8 7.1

7.5 9.6 6.7

8.3 ± 0.8 12.0 ± 1.8 7.6 ± 0.8

922

S. Deme et al.

Fig. 2. Dose rate as a function of time, showing SAA crossings, based on hourly dose measurements during the Euromir’95 mission.

Table 5 Results of manual read-out of dosimeters in the Euromir’95 experiment: absorbed dose rates (October 25–December 3, 1995) TLD No.

Location

Mean

1 2 3 4 5 6

Floor of the working area at small diameter Ceiling of the working area at large diameter Wall of the sleeping cabin Ceiling of the working area at large diameter Personal dosimeter Overhead cabin of the board–engineer

12.6 ± 0.3 11.7 ± 0.8 14.0 ± 0.5 12.3 ± 0.5 10.3 ± 0.4 11.6 ± 0.6

Using the hourly measuring period in automatic mode for 170 hours, dose components both of galactic (independent of SAA) and SAA origin were determined. The dose rates measured in automatic mode are shown in Fig. 2. SAA crossings, twice a day, are clearly visible. The three main objectives of the experiment were (1) to measure the dose rate distribution inside the Mir Space Station, (2) to record the personal dose of the astronaut, (3) to fulﬁll measurements in automatic mode with two different (daily and hourly) time periods to allow the measurement of the contribution of the SAA protons to the dose. The results received from manual read-out are given in Table 5. From the data it can be seen that the lowest dose rate was measured on the astronaut’s body; in contrast, the highest dose rate was measured at the usual sleeping places. It was found that the maximum dose due to crossing the SAA was equal to 55 μGy, and that the width of the SAA at 1/10 of the dose maximum was about 60◦ longitude at 30◦ south latitude. Averaging of all the measurements showed that the mean dose rate inside Mir was 12–14 μGy h−1 , and that half of this value was caused by the SAA components.

Twenty years of TLD measurements on board space vehicles by the Hungarian “Pille” system

923

Table 6 Dose rates inside Mir core module during the NASA4 mission February 6– May 6, 1997 Location

Dose rate (μGy h−1 )

Floor of the working area at small diameter Ceiling of the working area at large diameter Wall of the sleeping cabin Personal dosimeter Reference dosimeter stored in the reader

15.1 11.6 18.3 10.0 9.3

6. NASA measurements The advanced TLD system Pille’96 was used during the NASA4 (1997) mission in the FBI’5 (Fundamental Biological Investigation) experiment [7] to monitor the cosmic radiation dose inside the Mir Space Station and to measure the exposure of two of the astronauts during their EVA activities by means of CaSO4 :Dy dosimeters. Internal dose mapping using CaSO4 :Dy dosimeters gave mean dose rates ranging from 9.3 to 18.3 μGy h−1 at locations with different shielding thickness (Table 6). Locations were chosen to provide comparison with other types of dosimeters and to cover a wide range of shielding conditions. The extra dose of the extravehicular activity was studied during an EVA carried out by US astronaut J. Linenger and Russian cosmonaut V. Tsibliev. For EVA dose measurements, CaSO4 :Dy bulb dosimeters were located in specially designed pockets of the ORLAN spacesuits. The measured extra doses due to EVA (5:10 am.–10:08 am. on April 29, 1997–UTC) were compared with the dose measured inside the space station at the same time. During EVAs, on three occasions Mir crossed the SAA. To correct the effect of the higher background inside the space station due to the SAA, our results of the Euromir’95 mission were used. The method of correction was based on our hourly automatic TLD read-outs by the Pille’95 system, producing dose rate values with very good time resolution over a long time period. From these results, each dose during SAA crossings was obtained separately as a function of the crossing latitude. The corrected EVA/inside dose rate ratio was found to be about 3. The mean extra dose rate during EVA was about 55 μGy h−1 . 7. ISS measurements The last version of Pille was and is applied on ISS. The average dose rate measured in the second quarter of 2001 at several locations was about 5.6–8.3 μGy h−1 , which is two times lower than that of, measured on Mir during 1995–1997 (Fig. 3). TLDs are perfect for recording absorbed doses from radiation up to a LET of 20 keV μm−1 , but above this value the efﬁciency decreases rapidly with increasing LET. The response function of different TL materials as a function of LET has already been determined through a series of calibrations, including that for CaSO4 – which is generally used by our system. For the determination of the absorbed dose and especially the dose equivalent, detailed information on the contributions of the high LET components is needed. Therefore, in a ﬁrst step

924

S. Deme et al.

Fig. 3. Mean dose rates measured on board of different space stations.

on board the International Space Station, KFKI AEKI’s TLDs are supplemented by passive plastic nuclear track detectors (PNTDs) provided by Eril Research, Inc. (USA) for measurement of LET spectra 10 keV μm−1 in water. TLDs and PNTDs were exposed together and the LET spectrum measurements from the PNTDs were used to correct the dose measured in TLDs and, using the corrected dose, to determine the dose equivalent. The correction factor is 1.04 only. The measured by PNTD radiation weighting factor is in the range of 2.0–2.4 [9]. 8. Conclusions 8.1. Equipment Various versions of the on-board thermoluminescent dosimeter system Pille developed in KFKI AEKI have proven the system’s capability and reliability in numerous space experiments during the last two decades. Up till now, not a single space qualiﬁed system having similar features is available worldwide. 8.2. Dosimetry mapping Dose rates (12–14 μGy h−1 ) measured during the Euromir’95 mission on Mir [6] were about three times higher than those measured on board the Salyut-6 Space Station during the ﬂight of the Hungarian cosmonaut in 1980 [3]. The main reason for this is the increased ﬂight altitude – by about 100 km. During the NASA4 experiment on Mir [7], three of the Pille dosimeters (1A, 2A and 3A) were located at the same place as three NASA Area Passive Dosimeters (APD, viz. NMA-1, NMA-2 and NMA-3) exposed during the earlier NASA2 mission. The dose rate ratios (Pille TLDs/APD) were found to be 1.11, 0.96 and 1.08, which is an excellent agreement. 8.3. Personal dose During the Euromir’95 experiment, the minimum measured dose inside Mir was that on the body of the astronaut [6]. The same is observed during the NASA4 experiment [7], the dose rate measured with the personal dosimeter of the astronaut (10.0 μGy h−1 ) was signiﬁcantly lower than the mean value of three mapping dosimeters located in the working area and the sleeping places (15.0 μGy h−1 ). This can be explained by the shielding of the personal dosimeter by the body of the astronaut.

Twenty years of TLD measurements on board space vehicles by the Hungarian “Pille” system

925

8.4. EVA dose The dose rate increments measured by the Pille system during the two EVAs of the Soviet cosmonaut member of the second crew of Mir in 1987 were 144 and 196 μGy h−1 , respectively. During the Euromir’95 and the NASA4 missions, at the orbital inclination and altitude of the Mir Space Station, half of the dose inside the station resulted from the short time intervals ﬂying across the SAA. The extra doses of two EVAs during the Euromir’95 and one EVA during the NASA4 experiment were determined and compared with the reference doses measured inside Mir, and the mean ratio without any correction has been found to be about 4.

Acknowledgements We give special thanks to all cosmonauts and astronauts for operating the Pille TLD system during the experiments on different space vehicles. Our gratefulness is extended to all of our scientiﬁc and technical partners who dealt with and were concerned with the various versions of Pille systems throughout the work.

References [1] I. Fehér, S. Deme, B. Szabó, J. Vágvölgyi, P.P. Szabó, A. Cs˝oke, M. Ránky, Yu.A. Akatov, A new thermoluminescent dosimeter system for space research, Adv. Space Res. 1 (1981) 61–66. [2] I. Fehér, B. Szabó, J. Vágvölgyi, S. Deme, P.P. Szabó, A. Cs˝oke, New advanced TLD system for space dosimetry, Report 1983-99 to KFKI, 1983. [3] Yu.A. Akatov, V.V. Arkhangelsky, A.P. Aleksandrov, I. Fehér, S. Deme, B. Szabó, J. Vágvölgyi, P.P. Szabó, A. Cs˝oke, M. Ránky, B. Farkas, Thermoluminescent dose measurements on board Salyut type orbital stations, Adv. Space Res. 4 (1984) 77–81. [4] I. Fehér, R.G. Richmond, S.K. Ride, W. Atwell, Real-time thermoluminescent dosimeter measurements on board the Space Shuttle Orbiter, Unpublished report, 1985. [5] I. Apáthy, S. Deme, I. Fehér, Microprocessor controlled portable TLD system, Radiat. Prot. Dosim. 66 (1996) 441–444. [6] S. Deme, G. Reitz, I. Apáthy, I. Héjja, E. Láng, I. Fehér, Doses due to the South Atlantic Anomaly during the Euromir’95 mission measured by an on-board TLD system, Radiat. Prot. Dosim. 85 (1999) 301–304. [7] S. Deme, I. Apáthy, I. Héjja, E. Láng, I. Fehér, b. Extra dose due to EVA during the NASA4 mission measured by an on-board TLD system, Radiat. Prot. Dosim. 85 (1999) 121–124. [8] G. Reitz, I. Apáthy, R. Beaujean, S. Deme, C. Heilmann, J. Kopp, M. Leicher, K. Strauch, Results of dosimetric measurements in space missions, in: Proceedings Sixth European Symposium on Life Sciences Research in Space, Trondheim, Norway, in: ESA SP, vol. 390, 1996, pp. 183–190. [9] E.R. Benton, E.V. Benton, A.L. Frank, Flight veriﬁcation testing of the ISS passive dosimetry system: Results from expedition 2 DOSMAP experiment, Report No. ERI-011201, 2001.

926

Variation of radiation exposure as a function of altitude between 8 and 30 km B.J. Lewis, L.G.I. Bennett, A.R. Green, A. Butler, M.J. McCall, B. Ellaschuk Department of Chemistry and Chemical Engineering, Royal Military College of Canada, Kingston, Ontario, Canada K7K 7B4

Radiation measurements obtained with a tissue equivalent proportional counter (TEPC) at various ﬂight altitudes (up to 30 km) are presented in this paper. A function that describes the variation of dose equivalent rate with atmospheric depth was developed using 1999 jet-altitude (9–12 km) TEPC data, data obtained in 2001 from a balloon-borne TEPC and a theoretical mass balance analysis of the loss of primary particles and the formation of secondary particles in the atmosphere. This function was veriﬁed using separate TEPC data from (i) measurements made during the 1997 ER-2 AIR campaign, (ii) measurements made on board a British Airways Concorde aircraft at ∼ 16.5 km in 2002, and (iii) measurements obtained on a series of 16 jet-altitude ﬂights conducted during 2001 (close to solar maximum). 1. Introduction The natural radiation level at typical jet aircraft altitudes (∼ 6.1 to 18 km) is much higher than it is at ground level. For the most part, this ﬁeld arises as a result of the interaction of primary galactic cosmic ray (GCR) particles with the Earth’s atmosphere. These primary particles (consisting of ∼ 90% protons, 9% alpha particles and 1% heavy nuclei ranging from carbon to iron) emanate from stellar ﬂares and coronal mass ejections, supernova explosions, pulsar acceleration and explosion of galactic nuclei [1,2]. These particles are accelerated during their interstellar migration to very high energies (typically 100 MeV to 10 GeV, but which may extend up to 1020 eV). The penetrating ability of these ionized particles is directly affected by their magnetic rigidity (i.e., the ratio of their momentum to charge), which is inﬂuenced in an anticoincident manner with the solar cycle due to changing solar modulation. The primary cosmic rays are also affected by the Earth’s magnetic ﬁeld. Those particles that enter near the poles experience little deﬂection while those entering near the equator approach at right angles so that they are deﬂected if their rigidity is below the geomagnetic cutoff rigidity. Hence, the galactic radiation exposure is further dependent on the latitude of the aircraft. RADIOACTIVITY IN THE ENVIRONMENT Crown Copyright © 2005 Published by Elsevier Ltd. VOLUME 7 ISSN 1569-4860/DOI 10.1016/S1569-4860(04)07116-5 All rights reserved.

Variation of radiation exposure as a function of altitude between 8 and 30 km

927

The primary GCR particles that are able to enter into the upper layers of the atmosphere interact with the atmospheric nuclei. In each collision, a proton loses on average ∼ 50% of its energy, which results in secondary particle production of protons, neutrons, and mesons. The buildup of these secondary particles (generated by successive interactions with the primary and/or secondary particles) competes with their reduction through energy loss and further interactions with other atmospheric nuclei. These processes therefore lead to dose rates that vary with altitude, reaching a maximum level at 20 km above sea level, known as the Pfotzer maximum. As a result of 1990 recommendations by the International Commission on Radiological Protection (ICRP) [3], aircrew in many countries (including Canada and countries of the European Union) are being classiﬁed as occupationally exposed to cosmic radiation [4,5]. Speciﬁcally in Canada, an advisory circular by Transport Canada has been issued to recognize the occupational exposure of aircrew and to suggest voluntary action to manage such exposures to a level below 6 mSv y−1 [5]. These regulations present unique challenges to the airline industry, as the conventional dosimetric approach of individually monitoring each exposed member is both expensive and difﬁcult to manage. This leaves the alternative approach of predicting the exposure based on theoretical and experimental knowledge. In a previous study, tissue equivalent proportional counter (TEPC) data obtained from 36 ﬂights in 1999 (at altitudes ranging between 9 and 12 km) were used to develop a correlation between dose equivalent rate, geographical position and altitude (at a heliocentric potential of 650 MV) [6]. From this experimental correlation, a predictive code was then developed to provide the ambient dose equivalent, H ∗ (10), for a given ﬂight. This predictive code incorporated a theoretical analysis to account for the effect of varying solar modulation. As more commercial aircraft are expected to ﬂy at higher altitudes in the future, the capabilities of the predictive code need to be extended to cover these higher altitude ranges (i.e. > 12 km). In the present paper, experimental high-altitude (up to 30 km) TEPC data are presented in order to produce an altitude-scaling factor that can be used in the predictive code to provide H ∗ (10) estimates for higher altitude ﬂights. In addition, experimental data from a series of jet-altitude ﬂights conducted at close to solar maximum are used to improve the solar modulation function used in the predictive code.

2. Experimental procedure Radiation exposure levels were measured on board various ﬂights with a Battelle HAWK tissue equivalent proportional counter (TEPC). The TEPC is capable of detecting both low-LET and high-LET particles. It determines the absorbed dose by using a large (12.7 cm diameter) low-density cavity to simulate a microscopic volume of tissue of equal atomic composition (i.e., a 2-μm diameter tissue site). The Battelle HAWK TEPC was designed for portability and simplicity in airline ﬂight measurements. It consists of a grounded anode type sphericalwalled detector (Far West Technology LET-SW-5) with an attached preampliﬁer and spectrometer box that stores full data spectra each minute on a non-volatile ﬂash memory card (Fig. 1). The instrument in its carry-on case ﬁts into any overhead bin and is powered by 4 standard ‘D’ cells which last for up to ﬁve days of operation.

928

B.J. Lewis et al.

Fig. 1. Arrangement of TEPC in carrying case.

The in-ﬂight operation of this TEPC requires only that the power be switched on prior to the ﬂight and off following the landing. The raw binary data stored on the ﬂash memory card during the ﬂight are then downloaded at a later date to a computer. Battelle proprietary software was used to convert the raw binary data to spectral results that were stored as text ﬁles. The measured lineal energy, y, spectrum can be approximated by the linear energy transfer, LET, of the radiation. Additional Battelle proprietary software was used to relate the absorbed dose spectrum to the operational quantity of dose equivalent, H , using the Q (LET) relationship recommended in ICRP-60 [3]. In the current study, radiation levels were measured on 14 different jet-altitude (8.2– 12.4 km) ﬂights during the period 27 February 01 to 9 June 01, two high-altitude balloon ﬂights conducted by the Italian Space agency (14 July 2001 and 23 July 2001) and two British Airways Concorde engineering test ﬂights (29 January 02 and 8 February 02). Positional data for each ﬂight were obtained either with a handheld GPS (Global Positioning System) or with a GPS incorporated directly into the HAWK TEPC. Where possible, altitude data were based on altimeter readings recorded during the ﬂight. If these readings were not available, the altitude data provided by the built-in GPS were used. 3. Results and data analysis 3.1. Dose equivalent rate dependence upon altitude As mentioned previously, cosmic-ray radiation levels vary with altitude. In our earlier study [6], it was shown that for the jet-altitude range (between 9 and 12 km), the dose equivalent rate data at a given atmospheric depth h (depicted as H˙ (h) in the following analysis) can be subsequently normalized to an altitude of 10.6 km (or atmospheric depth of h0 = 243 g cm−2 ) using a simple exponential scaling law: H˙ (h) = e−ξs (h−h0 ) H˙ 0

(1)

Variation of radiation exposure as a function of altitude between 8 and 30 km

929

where ξs is an effective relaxation length for the atmosphere and the parameter h (in g cm−2 ) in equation (1) is related to the altitude A (in km) with the relation [7]:

1034[1 − 0.0227A]5.26 , A 10.9, h= (2) 230.6 exp −0.1587(A − 10.94) , A > 10.9. In our earlier study, ξs was taken as a constant average value of ∼ 0.0063 cm2 g−1 over the given range of latitudes and atmospheric depths for typical commercial altitudes (9.4 to 11.8 km) [6]. Based on additional data from other research groups, ξs can be further reﬁned as a function of the vertical cutoff rigidity Rc (in GV). For instance, over an altitude range of 8.5 to 11.8 km, the summed ambient dose equivalents from a lead-modiﬁed neutron rem counter (NMX) and an ionization chamber yielded a constant value of ξs ∼ 0.0085 cm2 g−1 when averaged over the North Pole region at vertical cutoff rigidity values of between approximately 0 and 4 GV, whereas a shallower slope of ξ s ∼ 0.0052 cm2 g−1 was obtained over the equatorial region between approximately 11 and 17.6 GV [8]. These values are consistent with other literature values and are comparable to that reported in our earlier study (i.e., our previous value is essentially an average of these two latitude-dependent quantities) [9,10]. Hence, in the current analysis, the following values of ξs are used: ⎧ Rc 4 GV, ⎨ 0.0085 cm2 g−1 , ξs = −4.714 × 10−4 Rc + 0.01039, 4 < Rc < 11 GV, (3) ⎩ Rc 11 GV. 0.0052 cm2 g−1 , Unfortunately, for high-altitude ﬂights, the simple exponential relationship of equation (1) cannot be extrapolated to altitudes near or above ∼ 20 km because of the effect of secondary particle buildup near the Pfotzer maximum (at an atmospheric depth of 60 g cm−2 ) [11]. However, a more general function can be derived from mass balance considerations for the loss of primary particles and the formation of secondary particles in the atmosphere. Incident cosmic radiation (i.e., protons) will be absorbed in the upper layer of the atmosphere. Thus, the rate of change of the primary particle intensity Ip (particle cm−2 s−1 ) with respect to the atmospheric depth h can be described by a simple (ﬁrst-order) absorption law [6]: dIp = −k0 Ip (4) dh where k0 is an effective absorption coefﬁcient for the primary particles. Integrating equation (4) yields, Ip (h) = Ip0 e−k0 h

(5)

where Ip0 is the intensity at the top of the atmosphere at h = 0. Analogous to Chapman layer theory in the formation of the ionosphere, the production of secondary particles can be assumed to be proportional to the rate of absorption of the primary particles, dIp /dz, with an effective proportionality constant β [12]. Thus, accounting for this source, and a ﬁrst-order loss due to absorption in the atmosphere, the conservation statement for the intensity, Is , of the secondary particles is dIs = βk0 Ip − ξs Is dh

(6)

930

B.J. Lewis et al.

where ξs is an effective relaxation length for the secondary particles. Thus, substituting equation (5) into equation (6) and integrating (where it is assumed that there is no secondary particle intensity at the top of the atmosphere, i.e., Is = 0 at h = 0), one obtains: 1 − e−(k0 −ξs )h . Is (h) = βk0 Ip0 e−ξs h (7) k0 − ξs The total intensity is therefore given as

βk0 −ξs h It (h) = Ip (h) + Is (h) = Ip0 e−k0 h + (8) e − e−k0 h . k 0 − ξs Normalizing equation (8) to an atmospheric depth h0 yields the following scaling function (for altitude): −(k0 −ξs )h e−k0 h It (h) k0 − ξs −ξs (h−h0 ) 1 − e fAlt (h) = (9) + =e It (h0 ) βk0 e−ξs h0 − e−k0 h0 1 − e−(k0 −ξs )h0 where the term Ip0 e−k0 h0 has been neglected in the expression for It (h0 ) since the primary ﬂux is negligible at the chosen value of h0 = 243 g cm−2 . As mentioned previously, the parameter k0 accounts for the attenuation of primary particles in the atmosphere. This parameter is ﬁtted to provide a maximum value of the function (equation (9)) at the Pfotzer maximum. Based on FLUKA calculations, the altitude at which the Pfotzer maximum occurs will change with the latitude; for example, near the equator at Rc = 17.6 GV, the Pfotzer maximum is predicted to occur at 16.5 km, but shifts to a slightly higher altitude of 19 km at Rc = 0.7 GV nearer to the poles [13,14]. To account for this effect, the attenuation coefﬁcient for the primary particles is taken as k0 ∼ 0.016 cm2 g−1 . This result is also consistent with the LUIN calculations, which show a greater relaxation length for the primary protons near the top of the atmosphere. The parameter β ∼ 3 and is an effective proportionality constant for the production of secondary particles from primary-particle interactions. This latter parameter has been evaluated with a ﬁtting of equation (9) to the ambient dose equivalent rate data obtained minute-byminute with the TEPC on the balloon-borne ﬂights (Fig. 2). These experiments were conducted at a geographical latitude and longitude of ∼ 38◦ N and ∼ 13◦ E (Rc ∼ 8.3 GV) with

Fig. 2. Comparison of equation (9) to measured TEPC data from balloon-borne experiments.

Variation of radiation exposure as a function of altitude between 8 and 30 km

931

a balloon ascent to a maximum of 32 km. In this ﬁtting, the value of ξs was determined from equation (3) for Rc ∼ 8.3 GV and the measured dose rates were normalized to the given measured value at h0 = 243 g cm−2 . Interestingly, the ﬁrst term in equation (9) has been previously proposed in [6]. The second term in equation (9) acts as a correction for the contribution of primary particles to the dose equivalent, which is only important at high altitudes (i.e., > 20 km). Moreover, for the any jet-altitude ﬂight, equation (9) reduces to the simple exponential law of equation (1). Thus, equation (9) is able to account for the main observed features including a maximum due to secondary-particle buildup and an approximate exponential loss in the lower part of the atmosphere. 3.2. Dose equivalent rate dependence upon solar modulation Cosmic ray intensities vary in a manner, which is anticoincident with solar activity so that galactic radiation is at a maximum during a solar minimum. A previous set of TEPC data from 36 jet-altitude ﬂights was obtained in 1999, a period in the mid-range of solar activity (slightly past the solar minimum) [6]. During the current study, data were obtained from 14 jetaltitude ﬂights in 2001, a more active portion of the solar cycle (Table 1). Together, these data then provide an excellent opportunity to observe experimentally the effect of solar modulation (i.e., the solar cycle) on the dose equivalent rate. The TEPC output provided values of the dose equivalent rate every minute, which were then summed over ﬁve-minute intervals and smoothed using a Savitzky and Golay method to reduce the relative error in the data to approximately 18% [6,15]. These data points were correlated to the given altitude and positional information. The data were then normalized to an altitude of 10.6 km (or atmospheric depth of h0 = 243 g cm−2 ) using equation (9), which, as already noted, reduces to equation (1) at jet-altitudes. Plotting the data versus the vertical cutoff rigidity Rc (based on the International Geomagnetic Reference Field, IGRF1995) yields Fig. 3. (In this ﬁgure, the 1999 data are shown by the open circles and the 2001 data are shown by the grey triangles.)

Fig. 3. Comparison of dose equivalent rate data measured over the solar cycle.

932

B.J. Lewis et al.

Table 1 TEPC measurements on jet-altitude ﬂights in 2001 Flight route∗ Date

Total ﬂight Time of ascent time (min) (min)

Time of descent (min)

Enroute altitude (km)

Time at altitude (min)

TEPC route dose equivalent (μSv)

YTR–BZZ

27-Feb-01

383

16

23

125

13

27

LDZA–YTR 1-Mar-01

500

17

27

YOW–YFB YFB–YRB YRB–YFB YFB–YOW YTR–YBG YBG–YOD

28-Mar-01 28-Mar-01 28-Mar-01 29-Mar-01 24-May-01 24-May-01

173 137 124 153 61 205

21 15 10 17 16 13

22 27 21 22 23 27

YOD–YTR

24-May-01 199

13

19

JFK–MIA

4-Jun-01

155

31

27

MIA–BUE

5-Jun-01

500

37

33

BUE–AKL

6-Jun-01

790

30

25

AKL–LAX

9-Jun-01

690

25

31

90 215 39 12 73 83 373 130 95 93 115 22 22 78 65 156 11 36 63 360 59 19 15 210 75 95 40 70 80 150 70 54 232 100 177

28.0

BZZ–LDZA 28-Feb-01

10.0 10.9 11.2 8.8 11.2 9.4 10.6 10.0 8.2 8.8 10.6 11.2 9.7 10.6 11.2 11.2 10.0 12.4 13.0 11.2 12.4 11.8 9.4 10.0 10.6 10.8 11.2 11.5 11.8 12.1 9.4 10.0 10.6 11.2 11.8

7.22 33.9 9.32 4.30 4.67 8.69 2.67 12.4

15.5 11.0 15.7

55.8

18.7†

∗ Airport codes are: YTR: Trenton, Ontario, Canada; BZZ: Brize Norton Air Base, London, UK; LDZA: Zagreb, Croatia; YOW: Ottawa, Ontario, Canada; YFB: Iqaluit, Nunavut, Canada; YRB: Resolute Bay, Nunavut, Canada; YBG: Bagotville, Quebec, Canada; YOD: Cold Lake, Alberta, Canada; JFK: John F. Kennedy Airport, New York, USA; MIA: Miami, Florida, USA; BUE: Buenos Aries, Argentina, AKL: Auckland, New Zealand; LAX: Los Angeles, California, USA. † Measurement for AKL–LAX ﬂight does not include the ascent of the ﬂight.

Variation of radiation exposure as a function of altitude between 8 and 30 km

933

The data from both curves in Fig. 3 can be best ﬁt to a sigmoid function of the form: fi = ai +

bi

(10)

i 1 + exp Rcd−c i

with Rc given in GV. The sigmoid ﬁtting parameters (i.e., ai , bi , ci and di ) of the data in Fig. 3 are given in Table 2. As expected, the dose rate during solar minimum conditions is greater than that during solar maximum (i.e., f1 > f2 ) due to the effect of solar modulation. Moreover, the geomagnetic knee, which results at high-latitude positions near the poles (i.e., for small values of the cutoff rigidity), is much more pronounced during solar maximum conditions [11]. The effect of the solar cycle can be modeled by correlating f1 and f2 to the given solar modulation. Two different models can be used to determine the solar modulation: (i) the heliocentric potential model of O’Brien [16], which is characterized by a heliocentric potential U (in MV) that is tabulated by the US Federal Aviation Administration (FAA) from daily ground-level neutron monitoring [17]; and (ii) a diffusion–convection model developed by the National Aeronautics and Space Administration (NASA) – Johnson Space Centre (JSC) [18], in which the solar modulation strength is determined by a comparable deceleration potential, Φ (in MV), that depends upon the Climax neutron-monitor count rate. Unfortunately, the two models are not totally consistent with one another since the ratio of Φ/U is not constant. As such, in the current work, both models have been considered. The average values of the two potential parameters over the given measurement period of the two data sets are indicated in Fig. 3. Using these values, the effect of the solar cycle on the (normalized) ambient total dose equivalent rate, H˙ 0 (in μSv h−1 ), can be correlated against either the heliocentric potential (U ) or the deceleration potential (Φ) with the use of linear Lagrange interpolation polynomials such that: f2 − f1 (U − 650) + f1 , H˙ 0U = (11a) 220 f2 − f1 (Φ − 650) + f1 , H˙ 0Φ = (11b) 320 where f1 and f2 are dependent on the cutoff rigidity as detailed in equation (10) (with the given ﬁtting parameters in Table 2). As other data become available over future solar cycles, these data will permit the selection of the better model. Table 2 Fitting parameters for data in Fig. 3 Fitting parameters

Dose rate function f1 (i = 1) (1999 data)

f2 (i = 2) (2001 data)

ai bi ci di

2.0643 4.5105 5.0016 2.7047

1.1744 3.6392 6.4170 2.3073

934

B.J. Lewis et al.

Equation (11) provides the basis for the development of a code to allow for dose equivalent rate prediction for any global position and period in the solar cycle (with an appropriate solar potential model choice). The effect of altitude can be described by (9) so that it can be written generally: H˙ (Rc , h; U, Φ) = H˙ 0U

or Φ

· fAlt .

(12)

The dose equivalent rate in equation (12) can be suitably integrated over a great circle path or between various way points for route–dose prediction for commercial ﬂights at any latitude, altitude or period in the solar cycle.

4. Code development and validation Based on the correlations derived from the experimental data, a Predictive Code for Aircrew Radiation Exposure (PCAIRE) was developed in a Visual C++ platform [6]. This code was written to be user-friendly and requires minimal time for data input, calculation and data storage. The code requires the user to input the date of the ﬂight, the origin and destination airports, the altitudes and times ﬂown at those altitudes. Look-up tables produce the latitude and longitudes of origin and destination, as well as the heliocentric potential or deceleration parameter. A great circle route is produced between the two airports, and the latitude and longitude are calculated for every minute of the ﬂight [19]. The PCAIRE code was validated against TEPC data obtained at jet altitudes during different periods of the solar cycle. Figure 4a illustrates the predictive capabilities of PCAIRE under solar minimum conditions using an independent 26-ﬂight TEPC data set, which was collected in 1999 (i.e., these validation data were not part of the 36-ﬂight data set used for model development in Fig. 4). Figure 4b shows PCAIRE predictions of the route dose equivalent for those ﬂights given in Table 1 (measured close to solar maximum). As shown in the ﬁgures, the PCAIRE predictions of the validation ﬂights are in good agreement with the TEPC measurements. Here the measured TEPC data have a relative error of ∼ 18%, while the code

Fig. 4. Plot of measured TEPC results versus PCAIRE code prediction of route dose equivalent using both solar modulation models for data obtained in (a) 1999 and (b) 2001.

Variation of radiation exposure as a function of altitude between 8 and 30 km

935

has a predictive error of about 20% (which accounts for the uncertainty due to deviations in the ﬂight path from a great circle route as well as uncertainties in the scaling functions for the altitude and solar modulation). The application of PCAIRE up to higher altitudes was tested against two independent sets of TEPC data. The ﬁrst set was data obtained from a Battelle TEPC ﬂown aboard an ER-2 aircraft in June 1997 as part of the Atmospheric Ionizing Radiation (AIR) sponsored by the National Aeronautics and Space Administration (NASA) and the US Department of Energy (DOE) [20]. The ER-2 aircraft was ﬂown at altitudes ranging from 15.2 to 21.3 km, with latitudes spanning 18 to 59◦ N (see Table 3). The second set of data was obtained on board engineering test ﬂights of a British Airways Concorde aircraft in January and February 2002. These ﬂights, at altitudes between 15.2 and 18 km, departed Heathrow International airport, proceeded approximately halfway across the Atlantic Ocean and then returned to Heathrow. As shown in Fig. 5, the new PCAIRE model with the high-altitude function of equation (9), is able to reproduce the measured TEPC data for the various high-altitude ﬂights within Table 3 Flight details and measured route doses for high altitude ﬂights Flight

Date

Total ﬂight time (h)

Flight path

TEPC route dose (μSv)

ER-2 East ER-2 North 1

6/5/97 6/8/97

6.5 7.8

88.7 158

ER-2 South 1 ER-2 North 2

6/11/97 6/13/97

6.5 6.5

ER-2 South 2 Concorde 1 Concorde 2

6/15/97 1/29/02 2/8/02

6.4 3.9 3.7

37◦ N, 122◦ W east to 35◦ N, 100◦ W and return Triangle: 37◦ N, 122◦ W north to 59◦ N, 116◦ W; west for altitude dip; return 37◦ N, 122◦ W south to 18◦ N, 127◦ W and return Triangle: 37◦ N, 122◦ W north to 55◦ N, 117◦ W; west for altitude dip, return Repeat of South 1 51◦ N, 0◦ W west to 48◦ N, 40◦ W and return 51◦ N, 0◦ W west to 4951◦ N, 0◦ W west to 48◦ N, 40◦ W and return ◦ N, 40◦ W and return

46.3 121

22.3 22.3

Fig. 5. Plot of measured TEPC results versus PCAIRE code prediction of route dose equivalent for those high-altitude ﬂights described in Table 3.

936

B.J. Lewis et al.

∼ 23% [21], providing further conﬁdence in the model for high-altitude calculations up to ∼ 20 km. (A correction factor of f = 1/1.15 has been applied to the measured TEPC data where H ∗ (10) = f HTEPC in accordance with calibrations at the PTB [6].) 5. Exposure from solar ﬂares Generally, the exposure resulting from solar ﬂares at typical subsonic altitudes would not contribute signiﬁcantly to the annual aircrew dose, and especially the career dose, as compared to that which arises from the continual galactic cosmic radiation exposure