Volatile Organic Compounds in the Atmosphere

Volatile Organic Compounds in the Atmosphere Volatile Organic Compounds in the Atmosphere Edited by Ralf Koppmann Uni...

Author: Ralf Koppmann

143 downloads 1336 Views 5MB 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

Volatile Organic Compounds in the Atmosphere

Volatile Organic Compounds in the Atmosphere Edited by Ralf Koppmann University of Wuppertal, Germany

Blackwell Publishing

© 2007 by Blackwell Publishing Ltd Blackwell Publishing editorial ofﬁces: Blackwell Publishing Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK Tel: +44 (0)1865 776868 Blackwell Publishing Professional, 2121 State Avenue, Ames, Iowa 50014-8300, USA Tel: +1 515 292 0140 Blackwell Publishing Asia Pty Ltd, 550 Swanston Street, Carlton, Victoria 3053, Australia Tel: +61 (0)3 8359 1011 The right of the Author to be identiﬁed as the Author of this Work has been asserted in accordance with the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs, and Patents Act 1988, without the prior permission of the publisher. First published 2007 by Blackwell Publishing Ltd ISBN: 978-1-4051-3115-5 Library of Congress Cataloging-in-Publication Data Volatile organic compounds in the atmosphere/edited by Ralf Koppmann. – 1st ed. p. cm. Includes bibliographical references and index. ISBN-13: 978-1-4051-3115-5 (hardback : alk. paper) 1. Organic compounds–Environmental aspects. 2. Air quality management. I. Koppmann, Ralf. TD885.5.O74V66 2007 511.51 12–dc22 2006034260 A catalogue record for this title is available from the British Library. Set in 10/12 pt Minion by Newgen Imaging Systems (P) Ltd., Chennai, India Printed and bound in Singapore by Markono Print Media Pte Ltd The publisher’s policy is to use permanent paper from mills that operate a sustainable forestry policy, and which has been manufactured from pulp processed using acid-free and elementary chlorine-free practices. Furthermore, the publisher ensures that the text paper and cover board used have met acceptable environmental accreditation standards. For further information on Blackwell Publishing, visit our website: www.blackwellpublishing.com

Volatile Organic Compounds in the Atmosphere

Contents

Preface

ix

List of Contributors

xi

1

1

2

3

Volatile Organic Compounds in the Atmosphere: An Overview Jonathan Williams and Ralf Koppmann 1.1 Introduction 1.2 Sources 1.3 Sinks 1.4 Atmospheric distribution 1.5 Measurement tools 1.6 Modelling tools 1.7 How organic species affect the atmosphere 1.8 Open questions and future directions References

1 3 5 7 9 10 12 15 19

Anthropogenic VOCs Stefan Reimann and Alastair C. Lewis 2.1 Introduction 2.2 Sources of anthropogenic VOCs 2.3 Atmospheric distribution of VOCs 2.4 Chemical behaviour of VOCs in the atmosphere 2.5 Measurement techniques References

33

Biogenic VOCs Allison H. Steiner and Allen L. Goldstein 3.1 Introduction 3.2 Sources of biogenic VOCs 3.3 Emission inventories of biogenic VOCs 3.4 Global distribution of biogenic VOCs 3.5 Impact on photooxidants and atmospheric chemistry

82

33 33 45 55 60 70

82 83 97 103 107

vi

4

5

6

7

Contents

3.6 Sampling and measurement techniques 3.7 Future directions References

114 116 117

Oxygenated Volatile Organic Compounds Ralf Koppmann and Jürgen Wildt 4.1 Introduction 4.2 Tropospheric mixing ratios and global distribution 4.3 Sources of OVOCs 4.4 Sinks of OVOCs 4.5 Budgets and emission inventories 4.6 Sampling and measurement techniques 4.7 Future directions Acknowledgement References

129

Halogenated Volatile Organic Compounds Simon J. O’Doherty and Lucy J. Carpenter 5.1 Introduction 5.2 Sources of halogenated VOCs 5.3 Atmospheric concentrations: trends and distribution 5.4 Sinks of halogenated VOCs 5.5 Emission inventories 5.6 Sampling techniques 5.7 Measurement techniques References

173

PAN and Related Compounds James M. Roberts 6.1 The chemistry of PANs 6.2 Atmospheric formation 6.3 Measurement and calibration techniques 6.4 Atmospheric measurements 6.5 Modelling and interpretation of ambient measurements 6.6 Conclusions Acknowledgements References

221

Organic Nitrates Paul B. Shepson 7.1 Introduction 7.2 Production mechanism 7.3 Measurement methods 7.4 Atmospheric measurements 7.5 Fate 7.6 Conclusions References

269

129 130 137 149 154 155 160 160 160

173 179 187 192 204 207 210 214

222 229 237 243 249 255 256 256

269 271 274 276 282 285 286

Contents

8

9

vii

High-Molecular-Weight Carbonyls and Carboxylic Acids Paolo Ciccioli and Michela Mannozzi 8.1 Introduction 8.2 Sources 8.3 Atmospheric levels 8.4 Reactivity and impact on the atmosphere 8.5 Sampling and analysis 8.6 Conclusions References

292

Organic Aerosols Thorsten Hoffmann and Jörg Warnke 9.1 Introduction 9.2 Carbonaceous aerosols 9.3 Analysis of organic aerosols Further reading References

342

292 293 309 324 329 333 334

342 345 365 375 375

10 Gas Chromatography-Isotope Ratio Mass Spectrometry Jochen Rudolph 10.1 Introduction 10.2 Fundamentals of stable isotope ratios of VOCs 10.3 Experimental methods 10.4 Kinetic isotope effects 10.5 Stable isotope ratios of atmospheric VOC and their sources 10.6 Conclusions References

388

11 Comprehensive Two-Dimensional Gas Chromatography Jacqueline F. Hamilton and Alastair C. Lewis 11.1 Introduction 11.2 Fundamentals of comprehensive gas chromatography 11.3 Modulators 11.4 Detectors 11.5 Examples of GC × GC use in atmospheric samples 11.6 Conclusions Further reading References

467

Index

489

Color plate appears between pages 268 and 269

388 389 405 420 447 458 460

467 468 471 474 475 482 486 486

Volatile Organic Compounds in the Atmosphere

Preface

Every day, large quantities of volatile organic compounds (VOCs) are emitted into the atmosphere from both anthropogenic and natural sources. They are the ‘fuel’ that keeps atmospheric photochemistry running. Therefore, their sources, sinks and residence times are the subject of current research. In addition to inﬂuencing local, regional and even global photochemistry, several of these compounds have a potential impact on climate, both due to their properties as greenhouse gases and due to their ability to form aerosol particles on oxidation. The formation of gaseous and particulate secondary products caused by the oxidation of VOCs is one of the largest unknowns in the quantitative prediction of the earth’s climate on a regional and global scale, and on the understanding of local air quality. To be able to model and control their impact, it is essential to understand the sources of VOC, their distribution in the atmosphere and the chemical transformations they undergo. Furthermore, organic trace gases can be used as tracer compounds to investigate reactions that are not directly accessible to current measurement techniques or as probes to ‘visualise’ transport processes in the atmosphere or across atmospheric boundaries. In recent years methods and techniques for the analysis of organic compounds in the atmosphere have been developed to increase both the spectrum of detectable compounds as well as the corresponding detection limits. New methods have been introduced to increase the time resolution of those measurements and to resolve more complex mixtures of organic compounds. New technical developments reducing weight and energy requirements made the use of these instruments on various platforms such as balloons or aircraft possible. This book describes the current state of knowledge of the chemistry of VOC as well as the methods and techniques to analyse gaseous and particulate organic compounds in the atmosphere. Chapter 1 is an instructive chapter summarising the variety and the roles of VOC in the atmosphere. Chapters 2 to 9 cover the various compound classes, their distribution in the atmosphere, their chemical transformations and their budgets as well as a survey of currently used measurement techniques. Chapters 10 and 11 describe new methods to measure a large part of the VOC family at a glance and for investigating their stable carbon isotope ratios. In-depth references are provided, enabling each subject to be explored in more detail. The aim is to provide an authoritative review to address the needs of both graduate students and active researchers in the ﬁeld of atmospheric chemistry research. It may also serve as a desktop resource for experienced scientists in the ﬁeld of atmospheric research.

x

Preface

Thanks are due to all the chapter authors for their efforts in the completion of this work. I am grateful to many colleagues for numerous discussions, for their patience, their advice and critical reviews of the chapters. Special thanks are due to Sarahjayne Sierra and Dr Paul Sayer of Blackwell Publishing for their patience in answering all my questions and their persistence in the efforts required for completing this book.

List of Contributors

Dr Lucy J. Carpenter

Department of Chemistry, University of York, York, UK

Dr Paolo Ciccioli

Istituto di Metodologie Chimiche del C.N.R., Monterotondo Scalo, Italy

Professor Allen L. Goldstein

Department of Environmental Science, Policy, and Management, University of California, Berkeley, CA, USA

Dr Jacqueline F. Hamilton

Department of Chemistry, University of York, York, UK

Professor Thorsten Hoffman

Institut für Anorganische und Analytische Chemie, Johannes Gutenberg-Universität Mainz, Mainz, Germany

Professor Ralf Koppmann

Fachbereich Mathematik und Naturwissenschaften, Atmosphärenphysik, Bergische Universität Wuppertal, Wuppertal, Germany

Professor Alastair C. Lewis

Department of Chemistry, University of York, York, UK

Dr Michela Mannozzi

Istituto Centrale per la Ricerca Scientiﬁca e Tecnologica Applicata al Mare, Roma, Italy

Dr Simon J. O’Doherty

School of Chemistry, University of Bristol, Bristol, UK

Dr Stefan Reimann

Eidgenössische Materialprüfungs- und forschungsanstalt, EMPA, Duebendorf, Switzerland

Dr James M. Roberts

NOAA Earth Systems Research Laboratory, Chemical Science Division, Boulder, CO, USA

Professor Jochen Rudolph

Chemistry Department and Centre for Atmospheric Chemistry, York University, Toronto, Ontario, Canada

Professor Paul B. Shepson

Departments of Chemistry, and Earth and Atmospheric Sciences, Purdue University, West Lafayette, IN, USA

Professor Allison H. Steiner

Department of Atmospheric, Oceanic and Space Science, University of Michigan, Ann Arbor, MI, USA

xii

List of Contributors

Dr Jörg Warnke

Institut für Anorganische und Analytische Chemie, Johannes Gutenberg-Universität Mainz, Mainz, Germany

Dr Jürgen Wildt

Institut für Chemie und Dynamik der Geosphäre, Institut 3: Phytosphäre, Forschungszentrum Jülich, Jülich, Germany

Dr Jonathon Williams

Atmospheric Chemistry Department, Max Planck Institute for Chemistry, Mainz, Germany

Volatile Organic Compounds in the Atmosphere Edited by Ralf Koppmann Copyright © 2007 by Blackwell Publishing Ltd

Chapter 1

Volatile Organic Compounds in the Atmosphere: An Overview Jonathan Williams and Ralf Koppmann

1.1

Introduction

The aim of this overview is to highlight the importance of organic trace gases in the atmosphere and to introduce the themes of the chapters to follow. This work is suited to those new to the ﬁeld and to those seeking to place related activities in a broader context. Tens of thousands of organic compounds have been detected in the air we breathe, and the focus here is on the myriad carbon-containing gases present at mixing ratios of some 10 parts per billion (ppbv, 10−9 or nmol/mol) down to some parts per trillion (ppt, 10−12 or pmol/mol). This excludes the three most abundant, but generally less reactive, organic compounds: carbon dioxide (370 parts per million, ppmv, 10−6 or μmol/mol), methane (1.8 ppmv) and carbon monoxide (0.15 ppmv), which have been discussed in detail elsewhere. Unfortunately various terms have been used in the literature to describe the subset of diverse carbon-containing gases under circa 10 ppbv. One of the ﬁrst was nonmethane hydrocarbons (NMHC), which was employed originally to distinguish alkanes, such as ethane, propane and butane, from methane. However, strictly speaking the word ‘hydrocarbon’ indicates a molecule containing only carbon and hydrogen atoms, and therefore this term appears to exclude oxygenated species such as alcohols, carbonyls and acids as well as organic compounds containing other heteroatoms, such as nitrogen or sulphur. In an attempt to embrace all the species relevant to atmospheric chemistry, a further term ‘volatile organic compound’ (VOC) was coined, although there is no general quantitative deﬁnition of what VOCs are. The Environmental Protection Agency (EPA) in the United States has deﬁned VOC as any compound that participates in atmospheric photochemical reactions; however, there have been subsequent attempts to give a more quantiﬁed deﬁnition. The result is that VOCs are considered to be those organic compounds having a vapour pressure greater than 10 Pa at 25◦ C, a boiling point of up to 260◦ C at atmospheric pressure, and 15 or less carbon atoms. The remaining compounds are designated as semivolatile organic compounds (SVOCs). This segregation emphasises the volatile gas phase species from those that partition to the aerosol phase which is reasonable since the later undergo different transport and chemistry (see Sections 1.3 and 1.4). Recently, however, the deﬁnition has become blurred by the use of OVOCs to speciﬁcally identify the oxygenated VOCs. For this chapter we prefer not to draw a divide through the continuum of

Volatile Organic Compounds in the Atmosphere

Stratospheric chemistry

Upper tropospheric chemistry

NO3 HO O3

Aerosol

N

VOC

Oxidation

siti

on

Advection

Remote region chemistry

Wa sh ou t Wet and dry deposition

olo

Bi

(Br, Cl)

Photochemical products Oxidation

uc

De

po

n

tio lea

C o nv e c ti o n

2

al

gic ke

ta

up

CO2 + H2O

Figure 1.1 Sketch of the various processes which determine the fate of VOC in the atmosphere. The individual processes are discussed in the text and the individual classes of VOC in the following chapters of this book.

compounds and instead use the term ‘trace organic compounds’, referring to the dictionary deﬁnitions of organic (designating carbon compounds) and trace (an extremely small amount). Despite being found at extremely low concentrations, trace organic compounds have profound effects in the atmosphere (see Figure 1.1). On the one hand, they are the ‘fuel’ which keeps oxidative atmospheric photochemistry running. Therefore, their sources, sinks and atmospheric residence times are the subject of much current research (see Sections 1.2–1.4). To investigate organic trace gases in the atmosphere it is essential that accurate concentration measurements and careful modelling studies are made (see Section 1.5). In addition to inﬂuencing local, regional and even global photochemistry, several such compounds have a potential impact on climate, both due to their properties as greenhouse gases and due to their ability to form aerosol particles on oxidation (see Section 1.6). Organic trace gases can be used as tracer compounds to investigate reactions which are not directly accessible to current measurement techniques or as probes to ‘visualise’ transport processes in the atmosphere or across atmospheric boundaries. Many open questions remain in this ﬁeld of research, and some of the future challenges in the ﬁeld are summarised in Section 1.7. The intention here is to provide an up-to-date, referenced overview of the ﬁeld emphasising the recent progress made in an exciting and rapidly developing area of research. Recent and review-type references have been preferentially cited along with key older articles so that the interested reader may quickly access more detailed information. The authors would

Volatile Organic Compounds in the Atmosphere: An Overview

3

like to point out that the articles cited here represent a tiny fraction of a vast and widespread literature database. We hope that any omissions of particular works by colleagues will be forgiven in the interests of brevity.

1.2

Sources

Almost everything we do in daily life results in the release of organic species to the atmosphere. Driving a car (Fraser et al. 1998), painting the house (Fortmann et al. 1998), cooking (McDonald et al. 2003), making a ﬁre (Andreae and Merlet 2001), cutting the grass (Fall et al. 1999; Kirstine et al. 1998) and even breathing (Barker et al. 2006; Phillips et al. 1999) – all of these processes result in the emission of organic compounds such as carbonyls, alcohols, alkanes, alkenes, esters, aromatics, ethers and amides. In addition to emissions from human activities, the Earth’s vegetation naturally releases huge amounts of organic gases into the air. As plants assimilate carbon dioxide into biomass through photosynthesis, a fraction of this carbon leaks out in to the atmosphere, predominantly in highly reduced forms such as isoprene and terpenes (Fuentes et al. 2000; Guenther 2002; Kesselmeier et al. 2002). Exactly which compounds are emitted from a particular plant, and how much of each, depends on the age and health of the vegetation as well as the ambient temperature, moisture and light levels (Guenther et al. 1995; Kesselmeier and Staudt 1999). Both plants and invertebrates have been shown to use emission of speciﬁc organic species into the air for signalling (Greene and Gordon 2003; Krieger and Breer 1999). Examples of elaborate chemical mimicry have been found in have been found in insects (e.g. Cremer et al. 2002) and amongst plants to deter attack by herbivores (Kaori et al. 2002; Kessler and Baldwin 2001). While the natural world uses air as a communication medium, man often uses it as a repository for waste products. The emission rates and their associated uncertainties for VOCs from several source categories are summarised in Table 1.1. The anthropogenic contribution to organic emissions in the atmosphere is dominated by the exploitation of fossil fuels (coal, oil and gas). Approximately 100 TgC/year was estimated to be emitted from ‘technological’ sources and 150 TgC/year from all anthropogenic sources including biomass burning (Müller 1992). Coal production mainly leads to methane emission, but minor emissions of ethane and propane are also present. Liquid fossil fuel production, storage and distribution result in a larger variety of organic gas emissions to the atmosphere. Crude oil production platforms are strong point sources of hydrocarbons such as methane, ethane, propane, butanes, pentanes, hexanes, heptanes, octanes and cycloparafﬁns (McInnes 1996). The major sources from processing liquid fossil fuels are catalytic cracking (0.25–0.63 kg/m3 of feed), coking (about 0.4 kg/m3 of feed) and asphalt blowing (about 27 kg of VOC/m3 of asphalt) (Friedrich and Obermeier 1999). Furthermore, so-called fugitive emissions can occur from leaks and evaporation from all types of equipment and installations. Evaporative emissions are estimated to be 2.9 kg/t of fuel at service stations (McInnes 1996) and are familiar to anyone who has ﬁlled a car with gasoline/petrol. Petrochemical products typically contain a limited number of compound classes (e.g. acyclic alkanes, cyclic alkanes, monoaromatics, diaromatics) each consisting of a very large number (tens of thousands) of individual homologues and isomers (Schoenmakers et al. 2000). Major products of the complete combustion of fossil fuels are carbon dioxide and water. However, in practice, combustion leads to CO and organic gas by-products, mainly due to

4

Volatile Organic Compounds in the Atmosphere

Table 1.1 Overview of important sources and global annual emission rates of selected groups of VOC per year Emission rate

Uncertainty range

Fossil fuel use Alkanes Alkenes Aromatic compounds

28 12 20

15–60 5–25 10–30

Biomass burning Alkanes Alkenes Aromatic compounds

15 20 5

7–30 10–30 2–10

Terrestrial plants Isoprene Sum of monoterpenes Sum of other VOC

460 140 580

Oceans Alkanes Alkenes Sum of anthropogenic and oceanic emissions Alkanes Alkenes Aromatic compounds

1 6

200–1 800 50–400 150–2 400 0–2 3–12

44 38 25

Terrestrial plants

1 180

Total

1 287

lack of oxygen, imperfect air/fuel mixing and inappropriate combustion temperatures. Tailpipe emissions from gasoline passenger cars with and without three-way catalyst are estimated, respectively, as 0.68 and 18.92 g of HC/kg of fuel, whereas passenger diesel cars (produced after 1996) emit about 1.32 g of HC/kg of fuel. Similarly, diesel heavyduty vehicle emissions are estimated as 5.4 g of HC/kg of fuel. Exhaust gas emissions from motor vehicles strongly depend on parameters such as vehicle speed, motor load and engine temperature. The predominate emissions for gasoline and diesel combustion engines are −C5 for gasoline cars and methane for diesel engines), C2 − −C5 oleﬁns, ethyne, parafﬁns (C1 − aromatic hydrocarbons (BTEX and C9 aromatics), aldehydes (formaldehyde, acetaldehyde, acrolein, benzaldehyde, tolualdehyde), ketones (acetone) and others (mainly high molecular weight parafﬁns). A smaller emission contribution of anthropogenic gases comes from the solvents industry, and global inventories of these anthropogenic emissions have been compiled (Friedrich and Obermeier 1999; Olivier et al. 1999). These ‘anthropogenic’ emissions are discussed in detail in Chapter 2 of this book. A further strong source of global emissions is from burning of biomass, and giant smoke plumes can nowadays be seen easily on satellite images, especially in the tropics during the dry season (September–October). These emissions are the most difﬁcult to assess as sources, as they are highly dependent on fuel type, humidity and burn rate amongst other

Volatile Organic Compounds in the Atmosphere: An Overview

5

factors (Lobert et al. 1990). Spatial and temporal variability further complicates global budget assessments, and satellite measurements are now being used to monitor the size and location of burning regions (e.g. Duncan et al. 2003). Most burning occurs during human-initiated land clearance but a large component also comes from the domestic use of biomass fuels (Levine 2003). A comprehensive summary of organic gas emissions from biomass burning relative to CO2 has been made recently (Andreae and Merlet 2001). On a global scale, the total amount of reactive biogenic emissions is not well established, although recent estimates indicate that c. 1300 TgC/year are emitted (Guenther 2002). The strongest biogenic emission is thought to be isoprene (C5 H8 ), followed by the less speciﬁed so-called other reactive biogenic compounds that are mainly oxygenated compounds and monoterpenes. Biogenic sources in total are considered to be approximately ten times larger than the sum of anthropogenic emissions including fossil fuel emissions and biomass burning (Guenther 2002; Muller 1992; Olivier et al. 1999). In comparison to terrestrial sources, emissions from the ocean are less well constrained although several important species are known to have a predominately marine source dimethyl sulphide (DMS, see Section 1.6) (Groene 1995) and methyl iodide (Lovelock 1975). A relatively small amount of organics is emitted from the ocean in the form of alkanes and alkenes, c. 5 TgC/year (e.g. Broadgate et al. 1997; Ratte et al. 1993). Recently global oceanic isoprene emissions have been estimated from satellite-derived chlorophyll map and laboratory studies 0.1 TgC/year (Palmer and Shaw 2005), that is, much smaller than the terrestrial source (c. 500 TgC/year). However, the surface ocean has been shown recently to be a massive reservoir for oxygenated organic species (Singh et al. 2003; Williams et al. 2004). Furthermore, a recent study of aerosols at a coastal site in Ireland (O’Dowd et al. 2004) showed that the organic fraction contributes signiﬁcantly (63%) to the sub-micrometer particle mass of aerosols collected over the North Atlantic Ocean during phytoplankton bloom periods. Biogenic emissions in general are discussed in Chapter 3 of this book, while Chapter 4 includes a section concerning the biogenic formation of OVOCs. From a global perspective, geographical location and season determine the relative importance of anthropogenic and biogenic emissions: biogenics are emitted mostly in the tropics whereas most anthropogenic emissions occur in the northern hemisphere between 40◦ N and 50◦ N. All these diverse organic emissions are broken down in the atmosphere into a wider array of partially oxidised species (Atkinson 1994; Atkinson and Arey 2003; Jenkin et al. 1997) and many thousands of gases have been detected in the atmosphere, from the tropics to Antarctica (Ciccioli et al. 1996; Zimmerman et al. 1988).

1.3

Sinks

Since the concentrations of organic trace species do not all simply increase with time there must logically be one or more removal processes (here termed sinks) acting on these compounds. The most important sink for organic trace gases in the atmosphere is chemical oxidation in the gas phase by the hydroxyl radical HO (or to a lesser extent O3 , NO3 and halogen radicals) (Atkinson 1994; Jenkin et al. 2003; Saunders et al. 2003). Certain gas phase organic compounds in the air can absorb sunlight and thereby photolyse to smaller fragments. Some compounds can be efﬁciently removed physically by dry deposition to surfaces such as vegetation (Doskey et al. 2004; Muller 1992) or aerosol (Cousins and

6

Volatile Organic Compounds in the Atmosphere

Mackay 2001); or removed by wet deposition in rain (Fornaro and Gutz 2003; Kieber et al. 2002). The gas phase oxidation of organic compounds in air is mostly initiated by the HO radical, with carbon dioxide and water being the ﬁnal products. In this way atmospheric oxidation is analogous to combustion. Using an everyday example as an analogy, when a cigarette lighter is lit, the hydrocarbon butane burns directly in the ﬂame to form H2 O and CO2 . When the ﬂame is not ignited, then the escaping butane gas is oxidised in the air to the same products, only much more slowly and via many other intermediates. The intermediate oxidation products may have lower vapour pressures, higher polarity or absorb light better than the precursors, making the intermediate products potentially more susceptible to physical removal or photolysis. An alkane must be larger than C20 to be adsorbed onto solid particles (Bidleman 1988), but much smaller multi-functional organic compounds, such as oxalic acid, more readily adsorb and are commonly found on aerosols (Mochida et al. 2003). Further oxidative transformation of these species on the aerosol is also possible (Claeys et al. 2004a; Noziere and Riemer 2003). The overall rate of removal of an organic species from the atmosphere can be derived by summing the reaction rates with radical species, rates of photolysis and the wet and dry deposition rates. From this we may determine the atmospheric lifetime of a species (see spatial distribution section). The rate of reaction of HO with many individual organic compounds under terrestrial conditions is well established from laboratory experiments as a function of temperature and pressure (http://www.iupac-kinetic.ch.cam.ac.uk/ and Mannschreck et al. 2002). Table 1.2 shows the atmospheric lifetime of several commonly measured VOCs with respect to OH, with lifetimes varying from months to minutes. Likewise, global photolysis rates can be calculated for many compounds from laboratory absorption cross section and quantum yield measurements (http://www. iupackinetic.ch.cam.ac.uk/). These rates can be profoundly inﬂuenced by clouds and this in turn can affect trace gas concentrations (Tie et al. 2003). The wet and dry deposition rates for organic compounds are highly variable and are generally empirically determined in the ﬁeld. Generally, organic compounds measured at high and invariable concentrations in the atmosphere are less efﬁciently removed (Junge 1974). Relationships between the variability of organic gas measurements and their rate of removal by HO have been derived (Jobson et al. 1999; Williams et al. 2000) and exploited to derive HO trends. If a long-lived and hence well-distributed organic compound is known to react predominately with HO, and its emission and HO reaction rate are known, then the global HO concentration can be theoretically estimated. Initial attempts based on methyl chloroform indicated large changes in HO concentrations from 1978 to 2003 (Prinn et al. 2001, 2005) with a maximum in 1989 and a minimum in 1998, although other recent evidence suggests that uncertainty in the temporal and spatial emission pattern of methyl chloroform complicates such trend analysis (Krol et al. 2003). Direct biological uptake can also be an effective atmospheric removal process for some organic species (Kesselmeier 2001; Kuhn et al. 2002). The rate of uptake is dependent on the ambient concentration, being strongest when ambient concentrations are high. Compensation points have been deduced for plants, which mark the crossover point between emission and uptake. A surprising recent discovery is that peroxy acetyl nitrate (PAN), an anthropogenic secondary oxidant like ozone, can also be taken up by plants (Sparks et al. 2003). This is an important development for the atmospheric nitrogen cycle as well as the organic species PAN.

Volatile Organic Compounds in the Atmosphere: An Overview

7

Table 1.2 Overview of average tropospheric lifetimes of VOC compound groups and some selected VOCs as examples. Lifetimes are given for an average OH concentration of 6 × 105 per cm3 and an average ozone concentration of 7 × 1011 per cm3 (about 30 ppb)

1.4

Compound

Average lifetime

Alkanes Ethane Propane n-Pentane

Months–days 2.5 months 2.5 weeks 4 days

Alkenes Ethene Propene 1-Butene

Days–hours 1.5 days 11 h 10 h

Cyclic compounds Cyclopentane Methylcyclohexane Cyclohexane

Days–hours 4 days 2 days 3h

Aromatic compounds Benzene Toluene 1,3,5-Trimethylbenzene

Weeks–hours 2 weeks 2 days 7.5 h

Biogenic compounds Isoprene α-Pinene Limonene

Hours–minutes 3h 4h 30 min

Atmospheric distribution

Following emission, volatile organic species are distributed by atmospheric transport processes while undergoing the removal processes described in Section 1.3. The relationship between the atmospheric removal rate (lifetime) of the compound and the average mixing times for the various sections of the atmosphere together determine the extent to which a compound is globally distributed. Atmospheric mixing is physically impeded across the boundary layer temperature inversion (0.5–2 km), the tropopause (10–15 km) and the intertropical conversion zone (ITCZ, 10◦ S–10◦ N), and as a result strong gradients in organic species can develop across these atmospheric interfaces. Typical exchange times are 1–2 days for air to mix vertically out of the boundary layer, 2 weeks to a month for air to be advected zonally around the northern or southern hemisphere, about 1 year for interhemispheric exchange and 4–6 years for an exchange between troposphere and stratosphere. Chemical lifetimes, deﬁned as the time for a chemical concentration to decay to 1/e of its initial value, varies from minutes to hours (terpenes and isoprene), through days to weeks (acetone, methanol, propane), years to decades (methyl chloroform, HCFC 134a), and up to hundreds

8

Volatile Organic Compounds in the Atmosphere

of years for chloroﬂuorocarbons (CFC 11 and CFC 12) whose lifetime is determined by the mixing rate into the stratosphere. Short-lived compounds, such as the biogenic species isoprene (CH2 C(CH3 )CHCH2 ), show strong atmospheric gradients within the boundary layer (0–2 km), whereas longer-lived compounds, such as CFC 113 (lifetime c. 8 years), are better mixed and only show strong gradients between the hemispheres (Boissard et al. 1996; Bonsang and Boissard 1999; Rudolph 1998). Some compounds are more or less uniformly distributed in the troposphere (CFC 12, lifetime 79 years, no remaining sources), only showing concentration gradients in the stratosphere. A large archive of aircraft measurements taken at various locations over the globe is available at http://www-gte.larc.nasa.gov/. There are numerous examples in the literature of regional scale advection where organic pollutants found in remote locations have been linked to distant pollution sources by use of back trajectories (e.g. Blake et al. 1996; Traub et al. 2003). Intercontinental pollution events have been reported (Price et al. 2004; Stohl and Trickl 1999) and trajectories have even been used to track southern hemispheric biomass burning, through the ITCZ to the upper troposphere of the northern hemisphere (Andreae et al. 2001). Secondary photooxidants, such as ozone and PAN, that form en route have also similarly been identiﬁed in plumes emerging from urban centres (e.g. Rappengluck et al. 2003). Interestingly, there is growing evidence to suggest that migrating birds use chemical gradients as an olfactory aid to navigation (Wallraff 2001, 2003). Where the atmosphere is in contact with the Earth, organic species can interact with the various surfaces, such as snow, soil and water (Ballschmiter 1992). Within these media further production or removal mechanisms may exist such as bacterial uptake, enhanced photolysis (Dominé and Shepson 2002; Klán et al. 2003) or biological production. Such processes will affect the lifetime of these species and hence their global distribution. Some higher molecular weight organics with considerably lower vapour pressures tend to partition predominantly to aerosols following release. When such a compound is unreactive, as with persistent organic pollutants (POPs), which are emitted through incomplete combustion or pesticide use, the lifetime of the transporting aerosol will then determine the distribution of this compound. Examples of such compounds include polyaromatic hydrocarbons (PAHs, e.g. Mastral and Callén 2000), Polychlorinated biphenyls (PCBs) and polychlorinated dibenzo-p-dioxins (PCDDs). Whether in gas form or as particles, these compounds can be transported long distances from source regions (Patton et al. 1991). The distribution of the long-lived semi-volatiles is markedly different to that of the volatiles, and with time through repeated volatilisation and adsorption, such compounds tend to concentrate in polar regions (Burkow and Kallenborn 2000) in a manner that could be likened to a global distillation (from the tropics to the poles). Some of these compounds are toxic (Walker 2001a) and can bioaccumulate through the food web (Tanabe et al. 1984), posing a risk to human health and the environment (UNEP 2001). While the boundary layer (0–2 km) tends to be turbulent, the so-called free troposphere above is less well mixed. In addition to the slow process of diffusion, organic gases may be distributed in the atmosphere by meteorological events such as convection (Collins et al. 1999) and via lifting by frontal systems (Purvis et al. 2003). The overall distribution of the organic species varies with latitude and season as a function of the source and sink strengths, as well as prevailing meteorology (Bonsang and Boissard 1999; Singh and Zimmerman 1992). Certain photochemical products, such as organic nitrates (e.g. PAN or alkyl nitrates), have a hemispheric concentration that is maximum in the spring. This has

Volatile Organic Compounds in the Atmosphere: An Overview

9

been explained as the photochemical optimum between the high precursor source and low photochemical sink in wintertime, when PAN precursors are accumulated, and the high photochemical sink in summer (Penkett 1983). In the early years of atmospheric research, it was assumed that after several days to weeks the atmosphere would have effectively removed an organic pollutant, based on the atmospheric lifetime of alkanes. Recently, however, from measurements made between 1 and 13 km over the remote Paciﬁc Ocean, far from source regions, it was shown that volume mixing ratios of oxygenated organic species are some ﬁve times higher than those of the NMHC, alkanes and alkenes (Singh et al. 2001, 2004). Similar high mixing ratios of oxygenates and compound diversity have been reported in other airborne studies (Crutzen et al. 2000), in urban centres (Lewis et al. 2000) and in continental outﬂow from Asia (Jacob et al. 2003; Lelieveld et al. 2001) and Europe (Salisbury et al. 2003). These results concur with earlier theoretical work on the oxidation of organic compounds (Calvert and Madronich 1987; Madronich et al. 1990). Our views about the distribution, sources and role of reactive organic species in the atmosphere are currently being revised rapidly.

1.5

Measurement tools

The human nose is particularly sensitive to several chemical groups (Cain 1979; Firestein 2001). Familiar examples include forests, which emit terpenes (e.g. Geron et al. 2000; Isidorov et al. 2000); oil reﬁneries, which emit aromatic compounds and alkenes (e.g. Doskey et al. 1999); ﬁsh markets, which emit amines (Morita et al. 2003) and freshly cut onions, which emit sulphur compounds (e.g. Ferary and Auger 1996). While human subjects are widely used in odour identiﬁcation studies (Ferreira et al. 2003; Walker 2001b), the nose’s response is inherently subjective (Molhave et al. 1991) and difﬁcult to quantify. Therefore, to investigate the atmosphere quantitatively, researchers have employed a variety of sensitive and speciﬁc sensors, including mass spectrometers, ﬂame-ionisation detectors, electroncapture detectors, optical absorption, chemiluminescence and atomic emission detectors (e.g. Apel et al. 1998; Helmig 1999; Kormann et al. 2002; Sigrist 2003). In many cases such detectors are coupled to pre-separation devices, for example, a gas chromatograph so that individual gases may be isolated prior to detection and a single speciﬁc compound can be measured. Both animal- and plant-type biological detectors have also been deployed for detection of certain molecules. In some studies the amputated sensory antennae of small insects have been connected into measurement devices (e.g. Murlis et al. 2000) and elsewhere the leaves of plants have been analysed for long-term exposure statistics (Hiatt 1999). Ideally for air studies a detectors should also be capable of measuring the huge range of concentrations in the atmosphere. High mixing ratios of several tens of ppbv (nmol/mol) can be found for alkanes and aromatics in polluted urban areas urban areas (Derwent et al. 2000), while halons must be reliably measured at only 0.045 pptv (pmol/mol) (Fraser et al. 1999). The recently reported compound SF5 CF3 was ﬁrst detected at 0.005 pptv (Sturges et al. 2000). This means that if only 200 tonnes of such material would be emitted anywhere in the world it would, in time, be detectable by this instrument. Global networks of detectors are in currently place to routinely monitor changes in greenhouse gases (Prinn et al. 2000). Much of our atmospheric knowledge to date has been driven by what can be reliably measured and how fast (Roscoe and Clemitshaw 1997). Although the ﬁrst atmospheric

10

Volatile Organic Compounds in the Atmosphere

research on organic trace gases (speciﬁcally PAN) was made using infrared spectroscopy (Stephens 1961; Stephens et al. 1956), the following 30–40 years of research on atmospheric organic gases have been dominated by gas chromatography (e.g. Darley et al. 1963; Helmig 1999) coupled to some form of detector. Samples have been either introduced directly into the instrument in the ﬁeld or collected in pressurised steel canisters, absorbent packed cartridges or ﬁlters for later analysis in the laboratory. The alkanes (major components of fossil fuel) were one of the ﬁrst and most widely investigated subset of the reactive organic species (Blomberg et al. 2002). This is because these fully saturated compounds do not interact strongly with most inlet materials or collection vessel surfaces, and the long established technique of gas chromatographic separation with ﬂame ionisation detection has allowed widely available quantitative analysis (Helmig 1999). Many oxidised gases are more difﬁcult to quantify as they may stick to surfaces, thermally decompose or may even be produced in measurement systems (Bates et al. 2000; Helmig et al. 1996; Kelly and Holdren 1995; Lestremau et al. 2001; Li-Jones et al. 2001; Tanner 2003). These techniques are sensitive and speciﬁc, but due to the predetector sample separation, these are limited in sampling frequency. Recently, several important new advances have been made in analytical techniques. These have permitted more organic species to be measured, such as high precision and sensitivity so as to enable δ 13 C isotopic ratios to be determined in organics at mixing ratios below ppbv levels (Rudolph et al. 1997) (see Chapter 10 of this book). Higher frequency measurements have been made possible with chemical ionisation mass spectrometry (CIMS) through use of proton transfer reactions (Hewitt et al. 2003; Lindinger et al. 1998; Williams et al. 2001) (see Chapter 11 of this book), or by other chemical ionisation techniques (Heeb et al. 1999; Leibrock et al. 2003). Several of these high frequency methods have been further developed to measure emission ﬂuxes directly (Bowling et al. 1998; Karl et al. 2004; Warneke et al. 2002). Furthermore, improvements in the ﬁeld of gas chromatography (e.g. multi-dimensional gas chromatography or comprehensive chromatography, Phillips et al. 1985) have delivered considerable improvements in compound separation, identiﬁcation and sensitivity (see Chapter 12 of this book). The physical separation of even enantiomeric monoterpenes, or optical isomers is also now possible from ambient air (Yassaa and Williams 2005). With the arrival of this new generation of measurement systems, more species and timescales are accessible and a new golden age of discovery for ﬁeld measurement has begun. Researchers are now exploiting these latest techniques on planes, ships, balloons and ground sites to establish the global budgets of a wide range of organic species.

1.6

Modelling tools

A variety of numerical models is available today to simulate chemistry and transport in the atmosphere from the level of box models to three-dimensional chemistry and transport models. To simulate atmospheric chemistry in detail, models need to deal with VOCs which play a signiﬁcant role in all reaction cycles in the atmosphere. In this regard the major input data they require are (a) the emission inventories describing the primary emission of VOC including their speciﬁc source compositions and their spatial and temporal variations, (b) the oxidation chemistry of VOC including the kinetics as a function of temperature and pressure and (c) for certain species a consideration of other signiﬁcant loss processes such as

Volatile Organic Compounds in the Atmosphere: An Overview

11

dry and wet deposition. Understandably, due to the huge number of organic compounds in the troposphere, both a complete emission inventory of all possible compounds with the necessary resolution in time and space and a complete coverage of all possible chemical reactions including those of secondary reaction products will never be available. There are only a few comprehensive emission inventories available and these cover the most important compounds relevant for atmospheric chemistry. The global distribution and source strengths of anthropogenic NMHC are usually taken from the Emission Database for Global Atmospheric Research (EDGAR V2.0) database (Olivier et al. 1996, 1999). Detailed information on this database and access to the data is available on http://www.mnp.nl/edgar/. The EDGAR database details sources of fossil-fuel-related activities, biofuel combustion, industrial production and consumption processes (including solvent use) on a per country basis, land-use-related sources, including waste treatment, partially on a grid basis and partially on a per country basis; and natural sources on a grid basis. The database can be used to generate global, regional and national emission data in various formats. For all compounds the reference year is 1990, except for halocarbons, for which 1986 is the reference year. In 2001, version 3.2 was released which comprises emissions by region and source for the period 1990–5. However, organic compounds emissions given in the EDGAR database are known to have large uncertainties in both magnitude and distribution of the emissions. An inventory of biomass burning and natural VOC emissions can be found in the Global Emissions Inventory Activity (GEIA) database. Details of this database can be found on http://www.geiacenter.org/. The biogenic VOC (BVOC) dataset consists of three ﬁles that cover isoprene, terpenes and ‘other’ NVOC data. This database provides BVOC emission measurements and modelling parameters and, in addition, an enclosure database that summarises information from literature and identiﬁes the plant species and the BVOC studied, including enclosure and analytical techniques and other parameters. Presently, about 1 800 plant species from which BVOC emissions have been studied are documented. As will be described in the following chapters, for individual VOC compound classes the chemistry of VOC can be quite complex. Two main approaches are used to simulate the complex chemistry of VOC using numerical models. One approach is to signiﬁcantly reduce the number of organic compounds and hence the complexity of the corresponding reactions. Three methods are used to reduce the number of organic reactions, namely the carbon bond mechanism, the surrogate species method and the lumped species method. In the carbon bond mechanism, individual organic compounds are segregated into one or more bond groups that have a similar chemical reactivity (cf. Gery et al. 1988). With the surrogate species method all VOC of similar reactivity are grouped together. The rate coefﬁcient of each of these compounds is then set equal to that of one particular compound (cf. Atkinson et al. 1982). In the lumped species method, VOC are grouped by their reactivity towards reactions with HO radicals. The rate coefﬁcient is determined by taking the mole fraction weighted average of the reaction coefﬁcient of each compound of the lumped group. Currently the most widely used lumped mechanism is the Regional Acid Deposition Model (RADM) (Stockwell et al. 1990). This mechanism contains 158 chemical reactions and 63 gaseous compounds. Besides primarily emitted inorganic compounds and 16 organic compounds or compound groups, respectively, RADM also includes photochemically produced compounds. This mechanism has meanwhile been updated into a version named

12

Volatile Organic Compounds in the Atmosphere

Regional Atmospheric Chemistry Model (RACM) (Stockwell et al. 1997) and includes reaction schemes for isoprene and the group of monoterpenes. The RACM mechanism again has been improved (Regional Atmospheric Chemistry Model-Mainz Isoprene Mechanism, RACM-MIM) by implementing an explicit chemistry of isoprene (Geiger et al. 2003), the most important BVOC (see Chapter 3). Today RACM is widely used in a variety of mathematical models, ranging from box models to three-dimensional chemistry and transport models. However, the mechanism is also available as a stand-alone model for simulating tropospheric chemistry at typical ambient trace gas concentration levels. The aforementioned methods of condensing VOC reactions have, of course, some disadvantages. The results of a model investigation of a single hydrocarbon oxidation are likely to be erroneous. Investigation of the RACM simulation showed, for example, that for a simulation of the oxidation of some individual VOC, such as branched alkanes, the error can be considerable. Therefore, other approaches have been developed that try to describe the complex chemistry of organics as explicitly as possible, leading to a vast number of chemicals and reactions which have to be taken into account. For example, the widely used Master Chemical Mechanism (MCM) is a detailed chemical mechanism containing 12 600 reactions and 4 500 chemical species (Jenkin et al. 2003). This mechanism includes the complete tropospheric oxidation of 124 VOC. The VOC which are degraded in this mechanism were selected on the basis of available emission data and provide approximately 90% mass coverage of the emissions of uniquely identiﬁable chemical species. The majority of the degradation schemes have been constructed using the methodology described by Jenkin et al. (1997). A review and update of the ideas behind the mechanism as well as recent developments can be found on the corresponding web page (http://mcm.leeds.ac.uk/MCM/).

1.7

How organic species affect the atmosphere

In the 1950s Haagen Smit and co-workers showed that the oxidation of organic species in the presence of NOx and sunlight can form ozone (Haagen Smit 1952). Ozone, which is toxic to humans and plants, has become a major air quality problem in cities and larger areas such as the Mediterranean (Lelieveld et al. 2002) and the south-eastern United States (Solomon et al. 2000). Ozone control strategies adopted in the 1970s were initially unsuccessful due to an underestimation of natural organic emissions in the initial models (Trainer et al. 1987). However, more recent emission controls applied to cars (including NOx and organic species reductions) have reduced regionally produced ozone (Derwent et al. 2003). The capacity of the troposphere to oxidise emissions is also dependent on the amount of organic compounds present. Reaction with the main atmospheric oxidant, the HO radical, is the primary loss mechanism of organics from the atmosphere (see Section 1.3). While the initial reaction is a sink, subsequent oxidation steps may be a source of HOx (HO and HO2 ), making the global effect of organic species complicated. In cities where NOx concentrations are high, increasing concentrations of organics increase the ambient HO. However, in most of the free troposphere HO production is not limited by organics but rather by NOx , and increasing organic concentrations generally decreases ambient HO under these conditions (Wang et al. 1998). In the upper troposphere, where water concentrations are low (<100 ppmv), organic species may provide the main source of HOx radicals. For example, acetone is a source of HOx when photolysed in the dry upper troposphere

Volatile Organic Compounds in the Atmosphere: An Overview

13

(McKeen et al. 1997) and can lead to strong ozone formation. Similar effects are produced by organic peroxides and aldehydes (Jaeglé et al. 2000). At night the organics may also provide an important source of HO radicals (e.g. Platt et al. 2002) through the reactions of alkenes with ozone, and in oceanic regions, organohalogens can provide ozone-depleting halogen radicals such as Br (Platt and Honninger 2003), which can link with inorganic halogen cycles (Sander et al. 2003). The global impact of organics on ozone and the hydroxyl radical have been investigated in models (Houweling et al. 1998; Wang et al. 1998). Sensitivity studies in global models indicate that removing hydrocarbon emissions gives modest decreases in global ozone (<15%) and relatively small increases in global mean HO (<20%). These studies highlight the role of organic nitrates such as PAN in the distribution of NOx and hence HOx . These nitrates form where hydrocarbons are oxidised in the presence of NOx (Roberts 1990). The most abundant nitrate PAN is a lachrymator and largely responsible for the sore eyes experienced in smog. Having a longer lifetime than NOx , organic nitrates may be transported much further from the pollution sources before decomposing to release NOx again. In this way these compounds function as a long-distance transport mechanism for NOx and in so doing they inﬂuence the global oxidation budget (Singh and Salas 1989). The effect of organic species on the global distribution of ozone and HO is a key area of atmospheric research. In the troposphere photooxidation of organic gases in the presence of high NOx (NO and NO2 ) concentrations acts to produce ozone; however in contrast, other trace organic gases can act to destroy ozone in the stratosphere. Prior to man’s proliferation on the planet, long-lived naturally produced organohalogens, such as methyl chloride, represented the most important mechanism for chlorine transport to the stratosphere and hence ozone destruction (Anderson 1990). More recently man-made chlorinated, brominated and ﬂuorinated hydrocarbons have been shown to deliver signiﬁcant additional quantities of chlorine and bromine into the stratosphere (Rowland 1990) and be the cause of the ‘ozone holes’ observed over the polar regions (Molina 1988). As a result of legislation, the overall tropospheric abundance of halogen from halocarbons is now decreasing (Montzka et al. 1996). However, emissions of certain species (e.g. Halons) persist because of a lack of suitable substitutes for critical uses such as ﬁre extinguishants (Butler et al. 1998). In addition to the major organic greenhouse gases CO2 and CH4 , certain other organic species have been implicated in long term or climate effects. Perhaps the best-known example is DMS, a compound which is naturally emitted from the oceans, and that has been proposed as a potential negative feedback to climate warming (Charlson et al. 1987). The hypothesis is that a long-term warming of the oceans would produce more DMS emission, which following oxidation to SO2 and then sulphate would lead to more clouds and hence more reﬂection of incoming sunlight. In polluted regions, where organic species can provide large numbers of condensation nuclei for clouds, then a further radiative consequence of organics emerges. The larger number of nuclei means that the available water in the cloud is more widely distributed, causing the average droplet size to be smaller and the cloud as a result to be more reﬂective (Feingold et al. 2003; Platnick and Twomey 1994). More recently certain oceanic alkyl halides have been linked to cloud formation and hence to possible climatic effects (O’Dowd 2002). While all organic species are infrared active to some extent, their inﬂuence on the global radiative forcing depends on their absorption spectrum and atmospheric abundance (Harries et al. 2001). The CFCs and HCFCs are important in this

14

Volatile Organic Compounds in the Atmosphere

regard. Therefore, organic gases, or the particulate products thereof, can cause a direct effect on climate forcing by reﬂecting or absorbing incoming light, or an indirect effect through modiﬁcation of cloud albedo and lifetime. By inﬂuencing global ozone, organic species may through chemistry also affect the radiation indirectly (Ramanathan 1998; Wigley et al. 2002). Due to the wide range of physical and chemical properties of organic compounds, it is extremely difﬁcult to assess their overall climate feedback effects, especially when the emissions are likely to be also changing as a function of time (Sanderson et al. 2003). When considering the climate effects we may also consider possible roles of organic species on our past climate. Atmospheric organic species have also been implicated in the beginning of life, which has inﬂuenced our climate profoundly. Small micelles and reverse micelles resulting from high molecular weight organics and water coating on aerosol have been proposed as photoreactors within which complex proteins may build up (Tuck 2002; Vaida 2002). On a larger scale, primordial ocean oil slicks of organic species have also been proposed as a means of climate change and important in the formation of the ﬁrst proteins (Nilson 2002). It has also been speculated that other organic species (nitriles) provided a source of nitrogen to photosynthesising marine organisms in the ancient oxygen-free atmosphere of the Earth (Bange and Williams 2000). The trace organic species also have enormous potential for revealing the atmospheric chemical history of the Earth. The analysis of air trapped in ice cores for CH4 and N2 O has provided atmospheric information of the past 120 000 year. Individual organic species can be more speciﬁcally attributed to sources than long-lived species such as CH4 and N2 O. Early work on small cores dating back only c. 40 year have demonstrated this potential (Lee et al. 2003). That gas phase organic compounds and atmospheric aerosols are strongly linked has been established for some time (Went et al. 1967), and organic atmospheric aerosols have recently been reviewed in detail by Seinfeld and Pankow (2003) and Jacobson et al. (2000). The blue hazes and reduced visibility over forests are the result of enhanced scattering of blue light by particles similar in size to the wavelength of light (Watson 2002). These particles can be produced from gaseous organic precursors, which condense from the gas phase to form aerosols (Kavouras et al. 1998). In this respect, oxygenated species with low vapour pressures are favoured; examples include the photochemical products of biogenic emissions such as isoprene (2-methyltetrols), terpenes (pinic and norpinic acid) and sesquiterpenes (Bonn and Moortgat 2003; Claeys et al. 2004b; O’Dowd et al. 2002). Such nuclei may grow in size by coagulation with other particles and later through the condensation of other organic species onto the surfaces (Grifﬁn et al. 1999; Kavouras and Stephanou 2002; Kulmala 2003). Recent evidence has shown that organic compounds absorbed into particles may undergo acidcatalysed reactions: oxidation, hydration, hemiacetal and acetal formation; polymerisation and aldol condensation (Claeys et al. 2004b; Iinuma et al. 2004; Jang et al. 2002; Kalberer et al. 2004; Tolocka et al. 2004) and chemical oxidation of organics such as isoprene can occur in the aerosol liquid phase through acid catalysed reactions with hydrogen peroxide (Claeys et al. 2004a). As the particle grows, the hydrophilic and light scattering properties of the particle can be affected by condensing organic gases or oxidation processes. In pristine conditions over the Amazon, it has been shown that hydrophilic organic species are a large fraction of wet season aerosol mass and they are predicted to signiﬁcantly contribute to particle growth into cloud condensation nuclei (Roberts et al. 2002; Yu 2002). An organic layer has also been recently reported on marine aerosols (Tervahattu et al. 2002). In more

Volatile Organic Compounds in the Atmosphere: An Overview

15

polluted conditions, it has been shown that uptake of organic gases onto soot particles can change particle reﬂectivity (Saathoff et al. 2003), making light absorbing soot particles more reﬂective. The opposite can also be speculated, that organic species can make reﬂective ammonium sulphate particles darker and more light absorbing. The organics, therefore, play a critical role in determining the aerosol albedo. As precursors of cloud condensation nuclei, organic gases can be important in the formation of clouds. Pure water requires extremely high relative humidities (400% or more) in order to produce clusters that provide nuclei for further condensation (Pruppacher and Klett 1997). However, in the presence of aerosols, condensation can occur at conditions more readily found in the atmosphere. Sulphate is a very effective cloud condensation nucleus and hence much research has been focused on DMS and its oxidation products (Pandis et al. 1994). A wide range of anthropogenic and biogenic compounds have been tested for their effect on ice nucleating properties (e.g. Szyrmer and Zawadzki 1997). Amino acids have been shown to be particularly effective in this regard (Milne and Zika 1993), and while long-chain alcohols in monolayers can also promote ice formation (Gavish et al. 1990), some organic acids appear to delay activation (Shantz et al. 2003). There is also some evidence that organic gases can inﬂuence the shape of ice crystals (Hallett and Mason 1958), a parameter that is predicted to have a large effect on snowfall rate (Lohmann et al. 2003). In cities, organic aerosols and soot are also emitted directly, particularly from diesel exhausts. POPs such as PAHs (carcinogenic) and PCBs (toxics derived from burning plastics) are often associated with urban particles. Particles smaller than 10 μm (PM 10) can be effectively inhaled by humans, and correlations have been shown with mortality rates (Brunekreef et al. 1995). In the clean marine environment it has been shown recently that aerosols can be efﬁciently formed from iodine-containing organics, such as diiodomethane (Jimenez et al. 2003). Furthermore, following formation marine aerosol appears to be coated with organic fatty acids. This coating of organic surfactants could have important effects on the physical and chemical properties of the aerosol (Gill et al. 1983). The impact of ship emissions of organics on the atmospheric aerosol is also currently under investigation (von Glasow et al. 2003; Noone et al. 2000). A more detailed discussion of organic aerosols can be found in Chapter 9 of this book.

1.8

Open questions and future directions

The previous sections have outlined how huge emissions of organic gases enter the atmosphere and how oxidation, predominantly initiated by the HO radical, breaks down these gases to CO2 and water. Determining how much reactive carbon is in ambient air and comparing it to the individually measured compounds is an important ongoing task. A similar budgeting process has been successfully performed for atmospheric nitrogen species (Fahey et al. 1986). For the carbon compounds this job is made difﬁcult by the presence of CO, CH4 and CO2 , whose concentrations dwarf those of the more reactive species. Some initial attempts have been made to measure the total carbon, termed Cy by chromatographic separation and then conversion of all reactive species to CH4 (Maris et al. 2003; Roberts et al. 1998). Comparisons to individually measured hydrocarbons are reasonable suggesting that most compounds making up the budget are being measured individually. However, the techniques adopted involve signiﬁcant sample handling, where losses of some

16

Volatile Organic Compounds in the Atmosphere

organic species could occur (Chung et al. 2003). Laboratory experiments and some as yet unpublished ﬁeld studies have shown the potential of HO measurements to determine total reactivity (Jeanneret et al. 2001; Kirchner et al. 2001), but the size and complexity of the instrumentation used has until now precluded widespread use. Recently ground-based measurements made over a forest have shown that a considerable fraction of the measured reactivity could not be accounted for by speciated measurements and that the missing reactivity had a terpene-like emission proﬁle (Di Carlo et al. 2004). Further forest-based ozone ﬂuxes and possible unidentiﬁed photoproducts support this notion (Goldstein et al. 2004; Holzinger et al. 2005). Generally, global models have considerable difﬁculty simulating the chemistry over large forest areas (e.g. the Amazon rainforest) since the high ﬂuxes from known reactive organic compounds, such as isoprene and monoterpenes, tend to consume the available HO radicals in the model. Clearly much more work must be performed to determine what fraction of the ambient organic trace gases are being measured by current techniques. If the total reactivity can be reliably determined, it can simplify photochemical modelling of organic species and aid the development of ozone control strategies. A further open issue is chemical oxidation in the polluted atmosphere. How well we understand the oxidation process can be generally assessed from the agreement between values from a theoretical model and direct measurement. Formaldehyde is considered a good parameter to compare since almost all organic species produce formaldehyde at some stage during gas phase oxidation. While measured and modelled comparisons of formaldehyde show good agreement in clean environments, comparisons in polluted air have consistently shown an underestimate in model values (Fried et al. 2003), the difference being up to a factor of 3–4 in the upper troposphere (Kormann et al. 2003). Assuming this is not a result of transport or source strength errors in the model, this suggests that the organic chemistry in such circumstances is more complex than currently thought. The ozone-forming potential of the organic species is therefore not well understood. Needed are comprehensive ﬁeld measurement datasets of organic compounds including precursor and oxygenated products. These must be compared with explicit chemistry models (e.g. Calvert et al. 2002; Jenkin et al. 1997) to determine for which species real air oxidation chemistry differs from the theoretical oxidation pathways. A complicating factor in this research will be the determination of the gas-to-particle partitioning of the organic species and representation of multi-phase chemistry in models. The next generation of atmospheric chemical transport models will require realistic but tractable mechanisms for organic oxidation, and particulate formation and growth in order to calculate the radiative forcing of climate (Chung and Seinfeld 2002; Grifﬁn et al. 2002). The role of the ocean in the budgets of organic species is surprisingly poorly understood. Many organic species are reported to be emitted from the ocean (including sulphurcontaining gases (Kettle et al. 2001), organohalogens (Carpenter et al. 2003), alkyl nitrates (Chuck et al. 2002)) while many other species are taken up (e.g. acetone, Warneke and de Gouw 2001; methanol, Galbally and Kirstine 2002). In reality, the ocean surface may be a highly variable source or sink for many compounds depending on the latitude, temperature, wind speed and biological composition of the surface water. Ocean emissions may also be dependent on aeolian input of trace elements such as iron and phosphorous. It has been speculated that for some compounds the ocean surface layer represents a giant reservoir of organic species exceeding the amount in the troposphere (e.g. for methanol; Galbally and Kirstine 2002; Singh et al. 2003; Williams et al. 2004). The ocean is able

Volatile Organic Compounds in the Atmosphere: An Overview

17

to assimilate CO2 at rates per unit biomass ﬁve times greater than the largest terrestrial ecosystems, namely, the rainforests, but oceanic organic trace gas production is not well understood. Considering its size and potential importance, the ocean is surprisingly poorly characterised in terms of organic gases, although new measurement programme have recently been initiated (Duce and Liss 2002). Of particular importance to atmospheric chemistry is the characterisation of tropical waters, as emission or uptake in these regions can affect air that is subsequently rapidly convected to the upper atmosphere in the ITCZ. As new instrumentation is deployed in new locations, new compounds are discovered in the air. Important sources and sinks of organic species are continuously being uncovered (e.g. toluene from plants, Heiden et al. 1999; methyl chloride from plants, Yokouchi et al. 2002). Compilation of global emission inventories is then by deﬁnition an ongoing task. Production of an emissions inventory is a long and laborious process, and the ‘current’ version is, therefore, out of date as soon as it is published, leading to large discrepancies between model and measurement even for anthropogenically emitted species (Gros et al. 2003). From the last section it is clear that an understanding of the sources, sinks and chemistry of organic species in the atmosphere is important in predicting future global change. If the Earth warms, as it is predicted to do, then we may expect the distribution of organics to be affected through concomitant increases in temperature, convection and changing vegetation patterns. How biogenic emissions change as a function of the forecasted increases in CO2 will be particularly important, and early indications are that biogenic emissions increase in elevated CO2 conditions (monoterpenes, Constable et al. 1999; methane, Inubushi et al. 2003). However, ozone is also expected to increase, which may also have an effect on these emissions (Mclaughlin and Downing 1995). Predictions based on the Intergovernmental Panel on Climate Change (IPCC) recommended emission scenarios for organic and other species show increases in tropospheric ozone (Zeng and Pyle 2003) and changing radiation budgets (Wigley et al. 2002), which will lead to global air circulation change. These effects may in turn lead to changes in the hydrological and biogeochemical cycles. To offset warming effects of CO2 , a number of proactive methods have been suggested to reduce directly the atmospheric CO2 burden that are relevant to organic species. Methods such as intensive tree planting and fertilising the ocean with iron should be carefully vetted for side effects associated with organic emissions before implementation. For example, plantations of fast-growing trees in polluted areas such as black larch as carbon sinks, can lead to extreme local terpene emissions with important consequences for regional pollution. In a sense this can convert a global problem into a regional one. A similar unexpected outcome can occur in the Amazon, where oil palms are often planted following tree clearance. Since the oil palm is a strong emitter of biogenic reactive compounds, it is conceivable that man’s interference can increase rather than decrease emission rates from a region. Similar effects have been noted in the United States, for example, through proliferation of sweetgum in pine plantations, and it has been suggested that VOC emissions are increasing at 6% per decade in the United States (Purves et al. 2004). Experiments aimed at fertilising the ocean with iron in order to stimulate biological growth and hence CO2 uptake should be also carefully assessed for enhanced organic halogen- or sulphurcontaining emissions which can effect ozone chemistry and aerosol formation (Fuhrman and Capone 1991). In the future, the global population is expected to grow, and megacities will house greater proportions of the increasing population (Cohen 2003). Therefore, it will be increasingly

18

Volatile Organic Compounds in the Atmosphere

important to control the local air quality in order to maintain human health (Beer 2001; Rodwin and Gusmano 2002). This will have to be achieved in the face of a long-term global upward trend in background ozone (Lelieveld et al. 2004; Prather et al. 2003). Considerable effort is being made to reduce the organic emissions of automobile engines through use of alternative fuels such as compressed natural gas (CNG) and liquid petroleum gas (LPG) as well as new engine technologies (Pearson 2001). Use of hydrogen fuel cell technology has been predicted to reduce CO and NOx emissions by up to 50% but its impact on climate forcing is dependent of the technology used (Schultz et al. 2003). The use of ﬁlters in combination with diesel fuel has been predicted to signiﬁcantly improve air quality in the short term (Jacobson et al. 2000). Indoor pollution in cities is also likely to grow in importance. Increased use of terpene-based propellants and increasing background ozone, both combined with the construction of more airtight energy efﬁcient buildings all lead to greater exposure to organic chemicals (Baker 1994; Carslaw 2003). Concerns are already being raised by some nations that other nations situated upwind are responsible for their deteriorating air quality. This echoes the smaller scale situation in the 1980s when cities on the US east coast complained that they could not comply with national air quality standards due to pollution built up in the west. Similarly, in the 1970s acid rain in Scandinavia was attributed to sulphur emissions in the United Kingdom. The issue of international pollution export (UNECE 1991) will almost certainly gain importance in coming years. Legal action is conceivable, where one country demands recompense from another for perceived health effects or tourism decline. Cost estimates for damage attributable to organic compounds have been made already: $1 100 per Mg-VOC (Rabl and Eyre 1998). The monitoring of exported emissions from other countries may also be used as an indirect way of spying on the upwind country. In theory, in the future it could be possible to determine whether published economic ﬁgures are realistic by comparing trends with emissions, or to determine whether certain processes are in operation. Emission trading has already been introduced for CO2 , despite the fact that the Kyoto protocol has only recently been ratiﬁed. Emission trading in other species is under discussion (Solomon 1999). In theory, limits could be imposed to trace organic species but this would require a much better understanding of emissions and would logically entail policing, which considering the short lifetimes and small concentrations of many compounds seems unlikely in the short term. In the near future investigation of such compounds will remain a task for the atmospheric research community. Data from a variety of new satellite platforms (Borrell et al. 2003) is unlikely to contribute many new organic measurements in the near future since most of the organic species do not have unique optical properties and are at very low concentrations. One exception is formaldehyde (CH2 O) and satellite measurements of this compound have been used elegantly to constrain the precursor isoprene emissions (Palmer et al. 2003). Some species, including CFCs-11, -12, -22, CCl4 and ethane, have been measured from a balloon using a Michelson interferometer to 3 km resolution, giving hope for the next generation of satellite measurements. Further candidates for future measurement by satellite are PAN and acetylene. It may also be possible to use typical emission ratios in conjunction with CO measurements to improve estimates in inaccessible regions. Satellite data drive weather forecasts which are increasingly likely to include ozone predictions in the future, as air quality is increasingly recognised as a meteorological hazard (Beer 2001). However, organic trace gas forecasts (e.g. Lawrence et al. 2003) are currently only of academic

Volatile Organic Compounds in the Atmosphere: An Overview

19

interest and available to the public only through a small number of internet sites (e.g. http://dionysos.mpch-mainz.mpg.de/∼lawrence/forecasts.html). Over the past hundred years global anthropogenic emissions of organic species have increased by an estimated factor of seven (van Aardenne et al. 2001), and in the future they can be expected to change still more as land use, vegetation patterns and industrial development alter rapidly (Andreae et al. 2002; Steiner et al. 2002). Since the 1980s the growth in surface emissions in North America and Europe has lessened, although in other areas such as the tropics strongly increasing trends are expected in line with economical development. Emissions from air trafﬁc, which affects cirrus clouds and thereby climate, are also predicted to increase in the future (Seinfeld 1998; Travis et al. 2004). Control of anthropogenic organic emissions is politically difﬁcult to realise both on regional (Grant et al. 1999) and on global scales (e.g. Montreal Protocol http://www.unep.org/ozone/montreal.shtml). To date most of our research on organics in the atmosphere has been passive. We have attempted to quantify emissions into the atmosphere and make deductions from there. On the other hand, deliberate emission of certain species has been proposed and practised. One proposal was the deliberate emission of hitherto unused compounds, of known lifetime, in speciﬁc amounts into the Earth’s atmosphere. Through regular measurements over a suitable time period, it would then be possible to determine accurately the global HO strength (Jöckel et al. 2003). Toxic compounds are also intentionally emitted into the environment as pesticides (Baker et al. 1996). Less than 0.1% of these applied pesticides reach their target raising important ethical questions (Pimentel 1995). It has been suggested that by actively injecting organics (ethane and propane) into the polar stratosphere, ozone depletions may be reduced (Cicerone et al. 1991). Follow-up research suggested, however, that this could have exactly the opposite effect (Elliot et al. 1994), which represents a stark warning against hasty experimentation. Recently it has been proposed to investigate the possibility of injecting sulphur or black carbon aerosol into the stratosphere as a method to attenuate incident radiation and hence engineer an atmospheric cooling to offset global warming effects (Crutzen 2006). We may also be tempted to take active measures if, in the distant future, the planet begins to cool. Humankind may well consider dosing the atmosphere with some of radiatively active organic gases to offset ice ages or to adjust atmospheres of other planets. There is clearly an important role for organic species in the atmosphere long into the future.

References van Aardenne, J.A., Dentener, F.J., Olivier, J.G.J., Goldewijk, C.G.M.K. and Lelieveld, J. (2001) A 1 degrees × 1 degrees resolution data set of historical anthropogenic trace gas emissions for the period 1890–1990. Global Biogeochemical Cycles, 15 (4): 909–28. Anderson, C. (1990) Ozone destruction – methyl chloride gas implicated. Nature, 348 (6300): 377. Andreae, M.O. and Merlet, P. (2001) Emission of trace gases and aerosols from biomass burning. Global Biogeochemical Cycles, 15 (4): 955–66. Andreae, M.O., Artaxo, P., Fischer, H., et al. (2001) Transport of biomass burning smoke to the upper troposphere by deep convection in the equatorial region. Geophysical Research Letters, 28 (6): 951–4.

20

Volatile Organic Compounds in the Atmosphere

Andreae, M.O., Artaxo, P., Brandao, C., et al. (2002) Biogeochemical cycling of carbon, water, energy, trace gases, and aerosols in Amazonia: The LBA-EUSTACH experiments. Journal of Geophysical Research, 107 (D20): 8066. Apel, E.C., Calvert, J.G., Greenberg, J.P., et al. (1998). Generation and validation of oxygenated volatile organic carbon standards for the 1995 Southern Oxidants Study Nashville Intensive. Journal of Geophysical Research, 103 (D17): 22281–94. Atkinson, R. (1994) Gas phase chemistry of organic compounds. Journal of Physics and Chemical Reference Data, R1. Atkinson, R. and Arey, J. (2003) Gas phase tropospheric chemistry of biogenic volatile organic compounds: A review. Atmospheric Environment, 37 (2): S197–219. Atkinson, R., Lloyd, A.C. and Winges, L. (1982) An updated chemical mechanism for hydrocarbons/NOx /SO2 photooxidations suitable for inclusion in atmospheric simulation models. Atmospheric Environment, 16: 1341–55. Baker, E.L. (1994) A Review of recent research on health effects of human occupational exposure to organic solvents – a critical review. Journal of Occupational and Environmental Medicine, 36 (10): 1079–92. Baker, L.W., Fitzell, D.L., Seiber, J.N., et al. (1996) Ambient air concentrations of pesticides in California. Environmental Science and Technology, 30 (4): 1365–8. Ballschmiter, K. (1992) Transport and fate of organic compounds in the global environment. Angewandte Chemie, International Edition, 31 (5): 487–664. Bange, H.W. and Williams, J. (2000) New directions: Acetonitrile in atmospheric and biogeochemical cycles. Atmospheric Environment, 34 (28): 4959–60. Barker, M., Hengst, M., Schmid, J., et al. (2006) Volatile organic compounds in the exhaled breath of young patients with cystic ﬁbrosis, May 2006. European Respiratory Journal, 27 (5): 929–36. Bates, M.S., Gonzalez-Flesca, N., Sokhi, R. and Cocheo, V. (2000) Atmospheric volatile organic compound monitoring. Ozone induced artefact formation. Environmental Monitoring and Assessment, 65 (1–2): 89–97. Beer, T. (2001) Air quality as a meteorological hazard. Natural Hazards, 23 (2–3): 157–69. Bidleman, T.F. (1988) Atmospheric processes – wet and dry deposition of organic-compounds are controlled by their vapor particle partitioning. Environmental Science and Technology, 22: 316–67. Blake, D.R., Blake, N.J., Smith, T.W. Jr., Wingenter, O.W. and Rowland, F.S. (1996) Non-methane hydrocarbon distributions during the Atlantic Stratocumulus transition experiment/marine aerosol and gas exchange (ASTEX-MAGE), June 1992. Journal of Geophysical Research, 101: 4501–14. Blomberg, J., Schoenmakers, P.J. and Brinkman, U.A.Th. (2002) Gas chromatographic methods for oil analysis. Journal of Chromatography A, 972 (4): 137–73. Boissard, C., Bonsang, B., Kanakidou, M. and Lambert, G. (1996) TROPOZ II: Global distributions and budgets of methane and light hydrocarbons. Journal of Atmospheric Chemistry, 25 (2): 115–48. Bonn, B. and Moortgat, G.K. (2003) Origin of atmospheric new particle formation from biogenic hydrocarbons. Geophysical Research Letters, 30 (11): 1585. Bonsang, B. and Boissard, C. (1999) Global distribution of reactive hydrocarbons in the atmosphere. In: N. Hewitt (Ed.) Reactive Hydrocarbons in the Atmosphere. Academic Press, London, pp. 209–65. Borrell, P., Burrows, J.P., Richter, A., Platt, U. and Wagner, T. (2003) New directions: New developments in satellite capabilities for probing the chemistry of the troposphere. Atmospheric Environment, 37 (18): 2567–70. Bowling, D.R., Turnipseed, A.A., Delany, A.C., Baldocchi, D.D., Greenberg, J.P. and Monson, R.K. (1998) The use of relaxed eddy accumulation to measure biosphere–atmosphere exchange of isoprene and of other biological trace gases. Oecologia, 116 (3): 306–15. Broadgate, W.J., Liss, P.S. and Penkett, S.A. (1997) Seasonal emissions of isoprene and other reactive hydrocarbon gases from the ocean. Geophysical Research Letters, 24 (21): 2675–8.

Volatile Organic Compounds in the Atmosphere: An Overview

21

Brunekreef, B., Dockery, D.W. and Krzyzanowski, M. (1995) Epidemiologic studies on short term effects of low levels of major ambient air pollution components. Environmental Health Perspectives, 103 (Suppl. 2): 3–13. Burkow, I.C. and Kallenborn, R. (2000) Sources and transport of persistent pollutants to the Arctic. Toxicological Letters, 112: 87–92. Butler, J.H., Montzka, S.A., Clarke, A.D., Lobert, J.M. and Elkins, J.W. (1998) Growth and distribution of halons in the atmosphere. Journal of Geophysical Research, 103 (D1), 1503–11. Cain, W.S. (1979) Know with the nose – keys to odor identiﬁcation. Science, 203 (4379): 467–70. Calvert, J.G. and Madronich, S. (1987) Theoretical study of the initial products of the atmospheric oxidation of hydrocarbons. Journal of Geophysical Research, 92 (D2): 2211–20. Calvert, J.G., Atkinson, R., Becker, K.H., et al. (2002) The Mechanisms of Atmospheric Oxidation of the Aromatic Hydrocarbons. Oxford University Press, New York. Carpenter, L.J., Liss, P.S. and Penkett, S.A. (2003) Marine organohalogens in the atmosphere over the Atlantic and Southern Oceans. Journal of Geophysical Research, 108 (D9): 4256. Carslaw, N. (2003) New directions. Atmospheric Environment, 37 (39–40): 5645–6. Charlson, R.J., Lovelock, J.E., Andreae, M.O. and Warren, S.G. (1987) Oceanic phytoplankton, atmospheric sulfur, cloud albedo and climate. Nature, 326 (6114): 655–61. Chuck, A.L., Turner, S.M. and Liss, P.S. (2002) Direct evidence for a marine source of C-1 and C-2 alkyl nitrates. Science, 297 (5584): 1151–4. Chung, S.H. and Seinfeld, J.H. (2002) Global distribution and climate forcing of carbonaceous aerosols. Journal of Geophysical Research, 107 (D19): 4407. Chung, M.Y., Maris, C., Krischke, U., Meller, R. and Paulson, S.E. (2003) An investigation of the relationship between total non-methane organic carbon and the sum of speciated hydrocarbons and carbonyls measured by standard GC/FID: Measurements in the Los Angeles air basin. Atmospheric Environment, 37 (Suppl. 2): S159–70. Ciccioli, P., Cecinato, A., Brancaleoni, E., Frattoni, M., Bruner, F. and Maione, M. (1996) Occurrence of oxygenated volatile organic compounds (VOC) in Antarctica. International Journal of Environmental Analytical Chemistry, 62 (3): 245–53. Cicerone, R.J., Elliott, S. and Turco, R.P. (1991) Reduced Antarctic ozone depletions in a model with hydrocarbon injections. Science, 254 (5035): 1191–4. Claeys, M., Wang, W., Ion, A.C., Kourtchev, I., Gelencsér, A. and Maenhaut, W. (2004a) Formation of secondary organic aerosols from isoprene and its gas phase oxidation products through reaction with hydrogen peroxide. Atmospheric Environment, 38: 4093–8. Claeys, M., Graham, B., Gyorgy, V., et al. (2004b) Formation of secondary organic aerosols through photooxidation of isoprene. Science, 303 (5661): 1173–6. Cohen, J.E. (2003) Human population: The next half century. Science, 302 (5648): 1172–5. Collins, W.J., Stevenson, D.S., Johnson, C.E. and Derwent, R.G. (1999) Role of convection in determining the budget of odd hydrogen in the upper troposphere. Journal of Geophysical Research, 104 (D21): 26927–41. Constable, J.V.H., Guenther, A.B., Schimel, D.S. and Monson, R.K. (1999) Modelling changes in VOC emission in response to climate change in the continental United States. Global Change Biology, 5 (7): 791–806. Cousins, I.T. and MacKay, D. (2001) Gas-particle partitioning of organic compounds and its interpretation using relative solubilities. Environmental Science and Technology, 35: 643–7. Cremer, S., Sledge, M.F. and Heinze, J. (2002) Chemical mimicry – male ants disguised by the queen’s bouquet. Nature, 419 (6910): 897. Crutzen, P.J. (2006) Albedo enhancement by stratospheric sulfur injections: A contribution to resolve a policy dilemma? Climatic Change, 77 (3–4): 211–9.

22

Volatile Organic Compounds in the Atmosphere

Crutzen, P.J., Williams, J., Poschl, U., et al. (2000) High spatial and temporal resolution measurements of primary organics and their oxidation products over the tropical forests of Surinam. Atmospheric Environment, 34 (8): 1161–5. Darley, E.F., Stephens, E.R. and Kettner, K.A. (1963) Analysis of peroxyacyl nitrates by gas chromatography with electron capture detection. Analytical Chemistry, 35 (4): 589–91. Derwent, R.G., Davies, T.J., Delaney, M., et al. (2000) Analysis and interpretation of the continuous hourly monitoring data for 26 C-2-C-8 hydrocarbons at 12 United Kingdom sites during 1996. Atmospheric Environment, 34 (2): 297–312. Derwent, R.G., Jenkin, M.E., Saunders, S.M., et al. (2003) Photochemical ozone formation in north west Europe and its control. Atmospheric Environment, 37 (14): 1983–91. Di Carlo, P., Brune, W.H., Martinez, M., et al. (2004) Missing OH reactivity in a forest: Evidence for unknown reactive biogenic VOCs. Science, 304 (5671): 722–5. Dominé, F. and Shepson, P.B. (2002) Air-snow interactions and atmospheric chemistry. Science, 297 (5586): 1506–10. Doskey, P.V., Fukui, Y., Sultan, M., Al Maghraby, A. and Taher, A. (1999) Source proﬁles for nonmethane organic compounds in the atmosphere of Cairo, Egypt. Journal of the Air and Waste Management Association, 49 (7): 814–22. Doskey, P.V., Kotamarthi, V.R., Fukui, Y., Cook, D.R., Breitbeil, F.W. and Wesely, M.L. (2004) Airsurface exchange of peroxyacetyl nitrate at a grassland site. Journal of Geophysical Research, 109 (D10): D10310. Duce, R.A. and Liss, P.S. (2002) The surface ocean – lower atmosphere study (SOLAS). Atmospheric Environment, 36 (32): 5119–20. Duncan, B.N., Martin, R.V., Staudt, A.C., Yevich, R. and Logan, J.A. (2003) Interannual and seasonal variability of biomass burning emissions constrained by satellite observations. Journal of Geophysical Research, 108 (D2): 4100. Elliot, S., Cicerone, R.J., Turco, R.P., Drdla, K. and Tabazdeh, A. (1994) Inﬂuence of the heterogeneous reaction HCL +HOCl on an ozone hole model with hydrocarbon additions. Journal of Geophysical Research, 99 (D2): 3497–508. Fahey, D.W., Hübler, G., Parrish, D.D., et al. (1986) Reactive nitrogen species in the troposphere – measurements of NO, NO2 , HNO3 , particulate nitrate, peroxyacetyl nitrate (PAN), O3 , and total reactive odd nitrogen (NOY) at NIWOT ridge Colorado. Journal of Geophysical Research, 91 (D9): 9781–93. Fall, R., Karl, T., Hansel, A., Jordan, A. and Lindinger, W. (1999) Volatile organic compounds emitted after leaf wounding: On-line analysis by proton-transfer-reaction mass spectrometry. Journal of Geophysical Research, 104 (D13): 15963–74. Feingold, G., Eberhard, W.L., Veron, D.E. and Previdi, M. (2003) First measurements of the Twomey indirect effect using ground-based remote sensors. Geophysical Research Letters, 30 (6): 1287. Ferary, S. and Auger, J. (1996) What is the true odour of cut Allium? Complementarity of various hyphenated methods: Gas chromatography mass spectrometry and high-performance liquid chromatography mass spectrometry with particle beam and atmospheric pressure ionisation interfaces in sulphenic acids rearrangement components discrimination. Journal of Chromatography A, 750 (1–2): 63–74. Ferreira, V., Petka, J., Aznar, M. and Cacho, J. (2003) Quantitative gas chromatography – olfactometry. Analytical characteristics of a panel of judges using a simple quantitative scale as gas chromatography detector. Journal of Chromatography A, 1002 (1–2): 169–78. Firestein, S. (2001) How the olfactory system makes sense of scents? Nature, 413 (6852): 211–18. Fornaro, A. and Gutz, I.G.R. (2003) Wet deposition and related atmospheric chemistry in the Sao Paulo metropolis, Brazil: Part 2 – contribution of formic and acetic acids. Atmospheric Environment, 37 (1): 117–28.

Volatile Organic Compounds in the Atmosphere: An Overview

23

Fortmann, R., Roache, N., Chang, J.C.S. and Guo, Z. (1998) Characterization of emissions of volatile organic compounds from interior alkyd paint. Journal of Air and Waste Management Association, 48 (10): 931–40. Fraser, M.P., Cass, G.R. and Simoneit, B.R.T. (1998) Gas-phase and particle-phase organic compounds emitted from motor vehicle trafﬁc in a Los Angeles roadway tunnel. Environmental Science and Technology, 32 (14): 2051–60. Fraser, P.J., Oram, D.E., Reeves, C.E., Penkett, S.A. and McCulloch, A. (1999) Southern Hemispheric halon trends (1978–1998) and global halon emissions. Journal of Geophysical Research, 104 (D13): 15985–99. Fried, A., Crawford, J., Olson, J., et al. (2003) Airborne tunable diode laser measurements of formaldehyde during TRACE-P: Distributions and box model comparisons. Journal of Geophysical Research, 108 (D20): 8798. Fuhrman, J.A. and Capone, D.G. (1991) Possible biogeochemical consequences of ocean fertilisation. Liminology and Oceanography, 36 (8): 1951–9. Friedrich, R. and Obermeier, A. (1999) Anthropogenic emissions of volatile organic compounds. In: N. Hewitt (Ed) Reactive Hydrocarbons in the Atmosphere. Academic Press, San Diego. Fuentes, J.D., Lerdau, M., Atkinson, R., et al. (2000) Biogenic hydrocarbons in the atmospheric boundary layer: A review. Bulletin of the American Meteorological Society, 81 (7): 1537–75. Galbally, I.E. and Kirstine, W. (2002) The production of methanol by ﬂowering plants and the global cycle of methanol. Journal of Atmospheric Chemistry, 43 (3), 195–229. Gavish, M., Popovitz-Biro, R., Lahav, M. and Leiwerowitz, L. (1990) Ice nucleation by alcohols arranged in monolayers at the surface of water drops. Science, 250: 973–5. Geiger, H., Barnes, I., Bejan, I., Benter, T. and Spittler, M. (2003) The tropospheric degradation of isoprene: An update module for the regional atmospheric chemistry mechanism. Atmospheric Environment, 37: 1503–19. Geron, C., Rasmussen, R., Arnts, R.R. and Guenther, A. (2000) A review and synthesis of monoterpene speciation from forests in the United States. Atmospheric Environment, 34 (11): 1761–81. Gery, M.W., Whitten, G.Z. and Killus, J.P. (1988) Development and testing of the CBM IV for urban and regional modeling. Report EPA-600/3-88-012. U.S. Environmental Protection Agency, Research Triangle Park, NC. Gill, P.S., Graedel, T.E. and Weschler, C.J. (1983) Organic ﬁlms on atmospheric aerosol particles, fog droplets, cloud droplets, raindrops and snowﬂakes. Review of Geophysics, 21 (4): 903–20. von Glasow, R., Lawrence, M.G., Sander, R. and Crutzen, P.J. (2003) Modeling the chemical effects of ship exhaust in the cloud-free marine boundary layer. Atmospheric Chemistry and Physics, 3: 233–50. Goldstein, A.H., McKay, M., Kurpius, M.R., et al. (2004) Forest thinning experiment conﬁrms ozone deposition to forest canopy is dominated by reaction with biogenic VOCs. Geophysical Research Letters, 31 (22): L22106. Grant, W., Perl, A. and Knoepfel, P. (1999) The Politics of Improving Urban Air Quality. Edward Elgar, Cheltenham, UK. Grifﬁn, R.J., Dabdub, D. and Seinfeld, J.H. (2002) Secondary organic aerosol – 1. Atmospheric chemical mechanism for production of molecular constituents. Journal of Geophysical Research, 107 (D17): 4332. Grifﬁn, R.J., Cocker, D.R., Seinfeld, J.H. and Dabdub, D. (1999) Estimate of global atmospheric organic aerosol from oxidation of biogenic hydrocarbons. Geophysical Research Letters, 26 (17): 2721–4. Groene, T. (1995) Biogenic production and consumption of dimethylsulﬁde (DMS) and Dimethylsulfoniopropionate (DMSP) in the marine epipelagic zone – A review. Journal of Marine Systems, 6 (3): 191–209. Greene, M.J. and Gordon, D.M. (2003) Social insects – cuticular hydrocarbons inform task decisions. Nature, 423 (6935): 32.

24

Volatile Organic Compounds in the Atmosphere

Gros, V., Williams, J., van Aardenne, J.A., et al. (2003) Origin of anthropogenic hydrocarbons and halocarbons measured in the summertime European outﬂow (on Crete in 2001). Atmospheric Chemistry and Physics, 3: 1223–35. Guenther, A. (2002) The contribution of reactive carbon emissions from vegetation to the carbon balance of terrestrial ecosystems. Chemosphere, 49 (8): 837–44. Guenther, A., Hewitt, C.N., Erickson, D., et al. (1995) A global-model of natural volatile organic compound emissions. Journal of Geophysical Research, 100 (D5): 8873–92. Haagen-Smit, A.J. (1952) Chemistry and physiology of Los Angeles smog. Industrial Engineering Chemistry, 44 (6): 1342–6. Hallett, J. and Mason, B.J. (1958) Inﬂuence of organic vapours on the crystal habit of ice. Nature, 181 (4607): 467–9. Harries, J.E., Brindley, H.E., Sagoo, P.J. and Bantges, R.J. (2001) Increases in greenhouse forcing inferred from the outgoing longwave radiation spectra of the Earth in 1970 and 1997. Nature, 410 (6826): 355–7. Heeb, N.V., Forss, A.M. and Bach, C. (1999) Fast and quantitative measurement of benzene, toluene and C-2-benzenes in automotive exhaust during transient engine operation with and without catalytic exhaust gas treatment. Atmospheric Environment, 33 (2): 205–15. Heiden, A.C., Kobel, K., Komenda, M., Koppmann, R., Shao, M. and Wildt, J. (1999) Toluene emissions from plants. Geophysical Research Letters, 26, (9): 1283–6. Helmig, D. (1999) Air analysis by gas chromatography. Journal of Chromatography A, 843 (1–2): 129–46. Helmig, D., Pollock, W., Greenberg, J. and Zimmerman, P. (1996) Gas chromatography mass spectrometry analysis of volatile organic trace gases at Mauna Loa Observatory, Hawaii. Journal of Geophysical Research, 101: 14697–710. Hewitt, C.N., Hayward, S. and Tani, A. (2003) The application of proton transfer reaction-mass spectrometry (PTR-MS) to the monitoring and analysis of volatile organic compounds in the atmosphere. Journal of Environmental Monitoring, 5 (1): 1–7. Hiatt, M.H. (1999) Leaves as an indicator of exposure to airborne volatile organic compounds. Environmental Science and Technology, 33: 4126–33. Holzinger, R., Lee, A., Paw, K.T. and Goldstein, A.H. (2005) Observations of oxidation products above a forest imply biogenic emissions of very reactive compounds. Atmospheric Chemistry and Physics, 5: 67–75. Houweling, S., Dentener, F. and Lelieveld, J. (1998) The impact of nonmethane hydrocarbon compounds on tropospheric photochemistry. Journal of Geophysical Research, 103 (D9): 10673–96. Iinuma, Y., Boge, O., Gnauk, T. and Herrmann, H. (2004) Aerosol-chamber study of the alphapinene/O-3 reaction: Inﬂuence of particle acidity on aerosol yields and products. Atmospheric Environment, 38 (5): 761–73. Inubushi, K., Cheng, W.G., Aonuma, S., et al. (2003) Effects of free-air CO2 enrichment (FACE) on CH4 emission from a rice paddy ﬁeld. Global Change Biology, 9 (10): 1458–64. Isidorov, V., Jaroszynska, J., Sacharewicz, T. and Piroznikow, E. (2000) Natural VOC emissions from forests in Poland. Atmospheric Environment, 34: 4739–44. Jacob, D.J., Crawford, J.H., Kleb, M.M., et al. (2003) Transport and Chemical Evolution over the Paciﬁc (TRACE-P) aircraft mission: Design, execution, and ﬁrst results. Journal of Geophysical Research, 108 (D20): 1–19. Jacobson, M.C., Hansson, H.C., Noone, K.J. and Charlson, R.J. (2000) Organic atmospheric aerosols: Review and state of the science. Reviews of Geophysics, 38 (2): 267–94. Jaeglé, L., Jacob, D.J., Brune, W.H., et al. (2000) Photochemistry of HOx in the upper troposphere at northern midlatitudes. Journal of Geophysical Research, 105 (D3): 3877–92. Jang, M.S., Czoschke, N.M., Lee, S. and Kamens, R.M. (2002) Heterogeneous atmospheric aerosol production by acid-catalyzed particle-phase reactions. Science, 298 (5594): 814–17.

Volatile Organic Compounds in the Atmosphere: An Overview

25

Jenkin, M.E., Saunders, S.M., Derwent, R.G. and Pilling, M.J. (1997) Construction and application of a master chemical mechanism (MCM) for modelling tropospheric chemistry. Abstracts of the American Chemical Society, 214: 116-COLL, part 1, 7. Jenkin, M.E., Saunders, S.M., Wagner, V. and Pilling, M.J. (2003) Protocol for the development of the Master Chemical Mechanism, MCM v3 (Part B): tropospheric degradation of aromatic volatile organic compounds. Atmospheric Chemistry and Physics, 3: 181–93. Jeanneret, F., Kirchner, F., Clappier, A., van den Bergh, H. and Calpini, B. (2001) Total VOC reactivity in the planetary boundary layer 1. Estimation by a pump and probe OH experiment. Journal of Geophysical Research, 106 (D3), 3083–93. Jimenez, J.L., Bahreini, R., Cocker, D.R., et al. (2003) New particle formation from photooxidation of diiodomethane (CH2I2). Journal of Geophysical Research, 108 (D10): 4318. Jobson, B.T., McKeen, S.A., Parrish, D.D., et al. (1999) Trace gas mixing ratio variability versus lifetime in the troposphere and stratosphere: Observations. Journal of Geophysical Research, 104 (D13): 16091–113. Jöckel, P., Brenninkmeijer, C.A.M. and Crutzen, P.J. (2003) A discussion on the determination of atmospheric OH and its trends. Atmospheric Chemistry and Physics, 3: 107–18. Junge, C.E. (1974) Residence time and variability of tropospheric trace gases. Tellus, 26 (4): 477–88. Kalberer, M., Paulsen, D., Sax, M., et al. (2004) Identiﬁcation of polymers as major components of atmospheric organic aerosols. Science, 303 (5664): 1659–62. Kaori, S., Maeda, T., Arimura, G., Ozawa, R., Shimoda, T. and Takabayashi, J. (2002) Functions of plant infochemicals in tritrophic interactions between plants, herbivores and carnivorous natural enemies. Japanese Journal of Applied Entomology and Zooology, 46 (3): 117–33. Karl, T., Potosnak, M., Guenther, A., et al. (2004) Exchange processes of volatile organic compounds above a tropical rain forest: Implications for modeling tropospheric chemistry above dense vegetation. Journal of Geophysical Research – Atmospheres, 109 (D18): D18306. Kavouras, I.G. and Stephanou, E.G. (2002) Particle size distribution of organic primary and secondary aerosol constituents in urban, background marine, and forest atmosphere. Journal of Geophysical Research – Atmospheres, 107 (D8): 4069. Kavouras, I.G., Mihalopoulos, N. and Stephanou, E.G. (1998) Formation of atmospheric particles from organic acids produced by forests. Nature, 395 (6703): 683–6. Kelly, T.J. and Holdren, M.W. (1995) Applicability of canisters for sample storage in the determination of hazardous air-pollutants. Atmospheric Environment, 29 (19): 2595–608. Kesselmeier, J. (2001) Exchange of short-chain oxygenated volatile organic compounds (VOCs) between plants and the atmosphere: A compilation of ﬁeld and laboratory studies. Journal of Atmospheric Chemistry, 39 (3): 219–33. Kesselmeier, J. and Staudt, M. (1999) Biogenic volatile organic compounds (VOC): An overview on emission, physiology and ecology. Journal of Atmospheric Chemistry, 33 (1): 23–88. Kesselmeier, J., Ciccioli, P., Kuhn, U., et al. (2002) Volatile organic compound emissions in relation to plant carbon ﬁxation and the terrestrial carbon budget. Global Biogeochemical Cycles, 16 (4): 1126. Kessler, A. and Baldwin, I.T. (2001) Defensive function of herbivore induced plant volatile emissions in nature. Science, 291 (5511): 2141–3144. Kettle, A.J., Rhee, T.S., von Hobe, M., Poulton, A., Aiken, J. and Andreae, M.O. (2001) Assessing the ﬂux of different volatile sulfur gases from the ocean to the atmosphere. Journal of Geophysical Research, 106 (D11): 12193–209. Kieber, R.J., Peake, B., Willey, J.D. and Avery, G.B. (2002) Dissolved organic carbon and organic acids in coastal New Zealand rainwater. Atmospheric Environment, 36 (21): 3557–63. Kirchner, F., Jeanneret, F., Clappier, A., Kruger, B., van den Bergh, H. and Calpini, B. (2001) Total VOC reactivity in the planetary boundary layer 2. A new indicator for determining the sensitivity of the ozone production to VOC and NOx . Journal of Geophysical Research, 106 (D3): 3095–110.

26

Volatile Organic Compounds in the Atmosphere

Kirstine, W., Galbally, I., Ye, Y.R. and Hooper, M. (1998) Emissions of volatile organic compounds (primarily oxygenated species) from pasture. Journal of Geophysical Research, 103 (D9): 10605–19. ˇ Klán, P., Klánová, J., Holoubek, I. and Cupr, P. (2003) Photochemical activity of organic compounds in ice induced by sunlight irradiation: The Svalbard project. Geophysical Research Letters, 30 (6): L016385. Kormann, R., Fischer, H., Gurk, C., et al. (2002) Application of a multi-laser tunable diode laser absorption spectrometer for atmospheric trace gas measurements at sub-ppbv levels. Spectrochemica Acta, A, Molecular and Biomolecular Spectroscopy, 58 (11): 2489–98. Kormann, R., Fischer, H., de Reus, M., et al. (2003) Formaldehyde over the eastern Mediterranean during MINOS: Comparison of airborne in-situ measurements with 3D-model results. Atmospheric Chemistry and Physics, 3: 851–61. Krieger, J. and Breer, H. (1999) Olfactory reception in invertebrates. Science, 286 (5440): 720–3. Krol, M.C., Lelieveld, J., Oram, D.E., et al. (2003) Continuing emissions of methyl chloroform from Europe. Nature, 421 (6919): 131–5. Kuhn, U., Rottenberger, S., Biesenthal, T., et al. (2002) Exchange of short-chain monocarboxylic acids by vegetation at a remote tropical forest site in Amazonia. Journal of Geophysical Research, 107 (D20): 8069. Kulmala, M. (2003) How particles nucleate and grow? Science, (302): 1000–1. Lawrence, M.G., Rasch, P.J., von Kuhlmann, R., et al. (2003) Global chemical weather forecasts for ﬁeld campaign planning: Predictions and observations of large-scale features during MINOS, CONTRACE, and INDOEX. Atmospheric Chemistry and Physics, 3: 267–89. Lee, X.Q., Qin, D.H., Jiang, G.B., Duan, K.Q. and Zhou, H. (2003) Atmospheric pollution of a remote area of Tianshan Mountain: Ice core record. Journal of Geophysical Research, 108 (D14): 4406. Leibrock, E., Huey, L.G., Goldan, P.D., Kuster, W.C., Williams, E. and Fehsenfeld, F.C. (2003) Groundbased intercomparison of two isoprene measurement techniques. Atmospheric Chemistry and Physics, 3: 67–72. Lelieveld, J., Crutzen, P.J., Ramanathan, V., et al. (2001) The Indian Ocean experiment: Widespread air pollution from South and Southeast Asia. Science, 291 (5506): 1031–6. Lelieveld, J., Berresheim, H., Borrmann, S., et al. (2002) Global air pollution crossroads over the Mediterranean. Science, 298 (5594): 794–9. Lelieveld, J., van Aardenne, J., Ficher, H., de Reus, M., Williams, J. and Winkler, P. (2004) Increasing ozone over the Atlantic Ocean. Science, 304 (5676): 1483–7. Lestremau, F., Desauziers, V. and Fanlo, J.L. (2001) Formation of artefacts during air analysis of volatile amines by solid-phase micro extraction. Analyst, 126 (11): 1969–73. Levine, J.S. (2003) Atmospheric chemistry: Burning domestic issues. Nature, 423 (6935): 28–9. Lewis, A.C., Carslaw, N., Marriott, P.J., et al. (2000) A larger pool of ozone-forming carbon compounds in urban atmospheres. Nature, 405 (6788): 778–81. Li-Jones, X., Savoie, D.L. and Prospero, J.M. (2001) HNO3 losses within the cyclone inlet of a diffusiondenuder system under simulated marine environments. Atmospheric Environment, 35 (5): 985–93. Lindinger, W., Hansel, A. and Jordan, A. (1998) Proton-transfer-reaction mass spectrometry (PTR-MS), on-line monitoring of volatile organic compounds at pptv levels. Chemical Society Reviews, 27 (5): 347–54. Lobert, J.M., Scharffe, D.H., Hao, W.M. and Crutzen, P.J. (1990) Importance of biomass burning in the atmospheric budgets of nitrogen-containing gases. Nature, 346 (6284): 552–4. Lohmann, U., Zhang, J. and Pi, J. (2003) Sensitivity studies of the effect of increased aerosol concentrations and snow crystal shape on the snowfall rate in the Arctic. Journal of Geophysical Research, 108 (D11): 4341. Lovelock, J.E. (1975) Natural halocarbons in the air and in the sea. Nature, 256: 193–4.

Volatile Organic Compounds in the Atmosphere: An Overview

27

Maris, C., Chung, M.Y., Lueb, R., et al. (2003) Development of instrumentation for simultaneous analysis of total non-methane organic carbon and volatile organic compounds in ambient air. Atmospheric Environment, 37 (Suppl. 2): S149–58. Madronich, S., Chatﬁeld, R.B., Calvert, J.G., Moortgat, G.K., Veyret, B. and Lesclaux, R. (1990) A photochemical origin of acetic acid in the troposphere. Geophysical Research Letters, 17 (13): 2361–4. Mannschreck, K., Bachmann, K., Barnes, I., et al. (2002) A database for volatile organic compounds. Journal of Atmospheric Chemistry, 42 (1): 281–6. Mastral, A.M. and Callen, M.S. (2000) A review on polycyclic aromatic hydrocarbon (PAH) emissions from energy generation. Environmental Science and Technology, 34 (15): 3051–7. McDonald, J.D., Zielinska, B., Fujita, E.M., Sagebiel, J.C., Chow, J.C. and Watson, J.G. (2003) Emissions from charbroiling and grilling of chicken and beef. Journal of Air and Waste Management Association, 53 (2): 185–94. McKeen, S.A., Gierczak, T., Burkholder, J.B., et al. (1997) The photochemistry of acetone in the upper troposphere: A source of odd-hydrogen radicals. Geophysical Research Letters, 24 (24): 3177–80. Mclaughlin, S.B. and Downing, D.J. (1995) Interactive effects of ambient ozone and climate measured on growth of mature forest trees. Nature, 374 (6519): 252–4. McInnes, G. (1996) Joint EMEP/CORINAIR Atmospheric Emission Inventory Guidebook. European Environmental Agency, Copenhagen, Denmark. Milne, P.J. and Zika, R.G. (1993) Amino acids and nitrogen in atmospheric aerosols. Journal of Atmospheric Chemistry, 16: 361–98. Mochida, M., Umemoto, N., Kawamura, K. and Uematsu, M. (2003) Bimodal size distribution of −C-4 dicarboxylic acids in the marine aerosols. Geophysical Research Letters, 30 (13): 1672. C-2− Molhave, L., Jensen, J.G. and Larsen, S. (1991) Subjective reactions to volatile organic compounds as air pollutants. Atmospheric Environment, 25 (7): 1283–93. Molina, M.J. (1988) The Antarctic ozone hole. Oceanus, 31 (2): 47–52. Montzka, S.A., Butler, J.H., Myers, R.C., et al. (1996) Decline in the tropospheric abundance of halogen from halocarbons: Implications for stratospheric ozone depletion. Science, 272 (5266): 1318–22. Morita, K., Kubota, K. and Aishima, T. (2003) Comparison of aroma characteristics of 16 ﬁsh species by sensory evaluation and gas chromatographic analysis. Journal of the Science of Food and Agriculture, 83 (4): 289–97. Müller, J.F. (1992) Geographical distribution and seasonal variation of surface emissions and deposition velocities of atmospheric trace gases. Journal of Geophysical Research, 97 (D4): 3787–804. Murlis, J., Willis, M.A. and Carde, R.T. (2000) Spatial and temporal structures of pheromone plumes in ﬁelds and forests. Physicological Entomology, 25 (3): 211–22. Nilson, F.P.R. (2002) Possible impact of a primordial oil slick on atmospheric and chemical evolution. Origins of Life and Evolution of the Biosphere, 32: 247–53. Noone, K.J., Johnson, D.W., Taylor, J.P., et al. (2000) A case study of ship track formation in a polluted marine boundary layer. Journal of the Atmospheric Sciences, 57 (16): 2748–64. Noziere, B. and Riemer, D.D. (2003) The chemical processing of gas-phase carbonyl compounds by sulfuric acid aerosols-2,4-pentanedione. Atmospheric Environment, 37 (6): 841–51. O’Dowd, C.D., Aalto, P., Hämeri, K., Kulmala, M. and Hoffmann, T. (2002) Atmospheric particles from organic vapours. Nature, 416: 497–8. O’Dowd, C.D., Facchini, M.C., Cavalli, F., et al. (2004) Biogenically driven organic contribution to marine aerosol. Nature, 431: 676–80. Olivier, J., Bouwman, A., Van der Maas, C., et al. (1996) Description of EDGAR version 2.0: A set of global emissions inventories of greenhouse gases and ozone depleting substances for all anthropogenic and most natural sources on a per country basis and on 1◦ × 1◦ grid, RIVM Rep. 771060 002,

28

Volatile Organic Compounds in the Atmosphere

TNO-MEP Rep. R96/119. National Institute of Public Health and the Environment, Bilthoven, The Netherlands. Olivier, J.G.J., Bloos, J.P.J., Berdowski, J.J.M., Visschedijk, A.J.H. and Bouwman, A.F. (1999) A 1990 global emission inventory of anthropogenic sources of carbon monoxide on 1◦ × 1◦ developed in the framework of EDGAR/GEIA. Chemosphere – Global Change Science, 1 (1–3): 1–17. Palmer, P.I. and Shaw, S.L. (2005) Quantifying global marine isoprene ﬂuxes using MODIS chlorophyll observations. Geophysical Research Letters, 32 (9): L09805. Palmer, P.I., Jacob, D.J., Fiore, A.M., Martin, R.V., Chance, K. and Kurosu, T.P. (2003) Mapping isoprene emissions over North America using formaldehyde column observations from space. Journal of Geophysical Research, 108 (D6): 4180. Pandis, S.N., Russell, L.M. and Seinfeld, J.H. (1994) The relationship between DMS ﬂux and CCN concentration in remote marine regions. Journal of Geophysical Research – Atmospheres, 99 (D8): 16945–57. Patton, G.W., Walla, M.D., Bidleman, T.F. and Barrie, L.A. (1991) Polycyclic aromatic and organochlorine compounds in the atmosphere of Northern Ellesmere Island, Canada. Journal of Geophysical Research, 96 (D6): 10867–77. Pearson, J. (2001) Improving Air Quality – Progress and Challenges for the Auto Industry. SAE International Publishers, Warrendale, PA. Penkett, S.A. (1983) Atmospheric chemistry – PAN in the natural atmosphere. Nature, 302 (5906): 293–4. Phillips, J.B., Luu, D., Pawliszyn, J.B. and Carle, G.C. (1985) Multiplex gas chromatography by thermal modulation of a fused silica capillary column. Analytical Chemistry, 57: 2779–87. Phillips, M., Herrera, J., Krishnan, S., Zain, M., Greenberg, J. and Cataneo, R.N. (1999) Variation in volatile organic compounds in the breath of normal humans. Journal of Chromatography B, 729 (1–2): 75–88. Pimentel, D. (1995) Amounts of pesticides reaching target pests – environmental impacts and ethics. Journal of Agriculture and Environmental Ethicsm, 8 (1): 17–29. Platnick, S. and Twomey, S. (1994) Determining the susceptibility of cloud albedo to changes in droplet concentration with the advanced very high resolution radiometer. Journal of Applied Meteorology, 33 (3): 334–47. Platt, U. and Honninger, G. (2003) The role of halogen species in the troposphere. Chemosphere, 52 (2): 325–38. Platt, U., Alicke, B., Dubois, R., et al. (2002) Free radicals and fast photochemistry during BERLIOZ. Journal of Atmospheric Chemistry, 42 (1): 359–94. Prather, M., Gauss, M., Berntsen, T., et al. (2003) Fresh air in the 21st century? Geophysical Research Letters, 30 (2): 1100. Price, H.U., Jaffe, D.A., Cooper, O.R. and Doskey, P.V. (2004) Photochemistry, ozone production, and dilution during long-range transport episodes from Eurasia to the northwest United States. Journal of Geophysical Research, 109 (D23): D23S13. Prinn, R.G., Weiss, R.F., Fraser, P.J., et al. (2000). A history of chemically and radiatively important gases in air deduced from ALE/GAGE/AGAGE. Journal of Geophysical Research, 105 (D14): 17751–92. Prinn, R.G., Huang, J., Weiss, R.F., et al. (2001) Evidence for substantial variations of atmospheric hydroxyl radicals in the past two decades. Science, 292 (5523): 1882–8. Prinn, R.G., Huang, J., Weiss, R.F., et al. (2005) Evidence for variability of atmospheric hydroxyl radicals over the past quarter century. Geophysical Research Letters, 32 (7): L07809. Pruppacher, H. and Klett, J.D. (1997) Microphysics of Clouds and Precipitation. Kluwer Academic, London. Purvis, R.M., Lewis, A.C., Carney, R.A., et al. (2003) Rapid uplift of nonmethane hydrocarbons in a cold front over central Europe. Journal of Geophysical Research, 108 (D7): 4224.

Volatile Organic Compounds in the Atmosphere: An Overview

29

Purves, D.W., Caspersen, J.P., Moorcroft, P.R., Hurtt, G.C. and Pacala, S.W. (2004) Human-induced changes in U.S. biogenic VOC emissions: Evidence from long-term forest inventory data. Global Change Biology, 10 (10): 1737–1755. doi:10.1111/j.1365–2486.2004.00844.x. Rabl, A. and Eyre, N. (1998) An estimate of regional and global O-3 damage from precursor NOx and VOC emissions. Environment International, 24 (8): 835–50. Ramanathan, V. (1998) Trace-gas greenhouse effect and global warming – underlying principles and outstanding issues – Volvo Environmental Prize Lecture – 1997. AMBIO, 27 (3): 187–97. Rappengluck, B., Melas, D. and Fabian, P. (2003) Evidence of the impact of urban plumes on remote sites in the Eastern Mediterranean. Atmospheric Environment, 37 (13): 1853–64. Ratte, M., Plass-Dulmer, C., Koppmann, R., Rudolph, J. and Denga, J. (1993) Production mechanisms −C4 hydrocarbons in seawater: Field measurements and experiments. Global Biogeochemistry of C2− Cycles, 7: 369–78. Roberts, G.C., Artaxo, P., Zhou, J.C., Swietlicki, E. and Andreae, M.O. (2002) Sensitivity of CCN spectra on chemical and physical properties of aerosol: A case study from the Amazon Basin. Journal of Geophysical Research, 107 (D20): 8070. Roberts, J.M. (1990) The atmospheric chemistry of organic nitrates. Atmospheric Environment, 24 (2): 243–87. Roberts, J.M., Bertman, S.B., Jobson, T., Niki, H. and Tanner, R. (1998) Measurement of total nonmethane organic carbon Cy , development and application at Chebogue Point, Nova Scotia, during the 1993 North Atlantic Regional Experiment campaign. Journal of Geophysical Research, 103 (D11): 13581–92. Rodwin, V.G. and Gusmano, M.K. (2002) The World Cities Project: Rationale, organization, and design for comparison of megacity health systems. Journal of Urban Health–Bulletin of the New York Academy of Medicine, 79 (4): 445–63. Roscoe, H.K. and Clemitshaw, K.C. (1997) Measurement techniques in gas-phase tropospheric chemistry: A selective view of the past, present, and future. Science, 276 (5315): 1065–72. Rowland, F.S. (1990) Stratospheric ozone depletion by chloroﬂuorocarbons. AMBIO, 19: 6–7. Rudolph, J. (1998) Two dimensional distribution of light hydrocarbons – results from the STRATOZ III experiment. Journal of Geophysical Research, 93 (D7): 8367–77. Rudolph, J., Lowe, D.C., Martin, R.J. and Clarkson, T.S. (1997) A novel method for compound speciifc determination of delta C-13 in volatile compounds at pptv levels in ambient air. Geophysical Research Letters, 24 (6): 659–62. Saathoff, H., Naumann, K.H., Schnaiter, M., et al. (2003) Coating of soot and NH4 (2)SO4 particles by ozonolysis products of alpha-pinene. Journal of Aerosol Science, 34 (10): 1297–321. Salisbury, G., Williams, J., Holzinger, R., et al. (2003) Ground-based PTR-MS measurements of reactive organic compounds during the MINOS campaign in Crete, July–August 2001. Atmospheric Chemistry and Physics, 3: 925–40. Sander, R., Keene, W.C., Pszenny, A.A.P., et al. (2003) Inorganic bromine in the marine boundary layer: A critical review. Atmospheric Chemistry and Physics, 3: 1301–36. Sanderson, M.G., Jones, C.D., Collins, W.J., Johnson, C.E. and Derwent, R.G. (2003) Effect of climate change on isoprene emissions and surface ozone levels. Geophysical Research Letters, 30 (18): 1936. Saunders, S.M., Jenkin, M.E., Derwent, R.G. and Pilling, M.J. (2003) Protocol for the development of the Master Chemical Mechanism, MCM v3 Part A, tropospheric degradation of non-aromatic volatile organic compounds. Atmospheric Chemistry and Physics, 3: 161–80. Schoenmakers, P.J., Oomen, M.M., Blomberg, J., Genuit, W. and van Velzen, G. (2000) Comparison of comprehensive two-dimensional gas chromatography and gas chromatography – mass spectrometry for the characterization of complex hydrocarbon mixtures. Journal of Chromatography A, 892: 29–46. Schultz, M.G., Diehl, T., Brasseur, G.P. and Zittel, W. (2003) Air pollution and climate-forcing impacts of a global hydrogen economy. Science, 302 (5645): 624–7.

30

Volatile Organic Compounds in the Atmosphere

Seinfeld, J.H. (1998) Clouds, contrails and climate. Nature, 391 (6670): 837–8. Seinfeld, J.H. and Pankow, J.F. (2003) Organic atmospheric particulate material. Annual review of Physical Chemistry, 54: 121–40. Shantz, N.C., Leaitch, W.R. and Caffrey, P.F. (2003) Effect of organics of low solubility on the growth rate of cloud droplets. Journal of Geophysical Research, 108 (D5): 4168. Sigrist, M.W. (2003) Trace gas monitoring by laser photoacoustic spectroscopy and related techniques (plenary). Review of Scientiﬁc instruments, 74 (1): 486–90, part 2. Singh, H.B. and Salas, L.J. (1989) Measurements of peroxyacetyl nitrate (PAN) and peroxypropionyl nitrate (PPN) at selected, urban, rural and remote sites. Atmospheric Environment, 23 (1): 231–8. Singh, H.B. and Zimmerman, P.R. (1992) Atmospheric distribution and sources of nonmethane hydrocarbons. In: J.O. Nriagu (Ed.) Gaseous Pollutants: Characterization and Cycling, Wiley-Interscience, New York, pp. 177–235. Singh, H., Chen, Y., Staudt, A., et al. (2001) Evidence from the Paciﬁc troposphere for large global sources of oxygenated organic compounds. Nature, 410 (6832): 1078–81. Singh, H.B., Tabazadeh, A., Evans, M.J., et al. (2003) Oxygenated volatile organic chemicals in the oceans: Inferences and implications based on atmospheric observations and air–sea exchange models. Geophysical Research Letters, 30 (16): 1862. Singh, H.B., Salas, L.J., Chatﬁeld, R.B., et al. (2004) Analysis of the atmospheric distribution, sources, and sinks of oxygenated volatile organic chemicals based on measurements over the Paciﬁc during TRACE-P. Journal of Geophysical Research, 109 (D15): D15S07. Solomon, B.D. (1999) New directions in emissions trading: The potential contribution of new institutional economics. Ecological Economics, 30 (3): 371–87. Solomon, P., Cowling, E., Hidy, G. and Furiness, C. (2000) Comparison of scientiﬁc ﬁndings from major ozone ﬁeld studies in North America and Europe. Atmospheric Environment, 34 (12–14): 1885–920. Sparks, J.P., Roberts, J.M. and Monson, R.K. (2003) The uptake of gaseous organic nitrogen by leaves: A signiﬁcant global nitrogen transfer process. Geophysical Research Letters, 30 (23): 2189. Steiner, A., Luo, C., Huang, Y. and Chameides, W.L. (2002) Past and present-day biogenic volatile organic compound emissions in East Asia. Atmospheric Environment, 36 (31): 4895–905. Stephens, E.R. (1961) Long path infrared spectroscopy for air pollution research. Infrared Physics, 1 (3): 187. Stephens, E.R., Hanst, P.L., Doerr, R.C. and Scott, W.E. (1956) Reactions of nitrogen dioxide and organic compounds in air. Industrial and Engineering Chemistry, 48 (9): 1498–504. Stockwell, W.R., Middleton, P., Chang, J.C., and Tang, X. (1990) The second generation Regional Acid Deposition Model chemical mechanism for regional air quality modeling. Journal of Geophysical Research, 95: 16343–67. Stockwell, W.R., Kirchner, F., Kuhn, M. and Seefeld, S. (1997) A new mechanism for regional atmospheric chemistry modeling. Journal of Geophysical Research, 102: 25847–79. Stohl, A. and Trickl, T. (1999) A textbook example of long-range transport: Simultaneous observation of ozone maxima of stratospheric and North American origin in the free troposphere over Europe. Journal of Geophysical Research, 104 (D23): 30445–62. Sturges, W.T., Wallington, T.J., Hurley, M.D., et al. (2000) A potent greenhouse gas identiﬁed in the atmosphere: SF5CF3. Science, 289 (5479): 611–13. Szyrmer, W. and Zawadzski, I. (1997) Biogenic and anthropogenic sources of ice-forming nuclei: A review. Bulletin American Meteorological Society, 78: 209–28. Tanabe, S.,Tanaka, H. and Tatsukawa, R. (1984) Bioaccumulation of PCBs, DDT and HCH isomers in the North Paciﬁc ecosystem. Archive of the Environmental Contamination Toxicology, 13: 731–8. Tanner, R.L. (2003) Measurements in support of air quality improvement – some historical insights. Atmospheric Environment, 37 (9–10): 1271–6.

Volatile Organic Compounds in the Atmosphere: An Overview

31

Tervahattu, H., Hartonen, K., Kerminen, V.M., et al. (2002) New evidence of an organic layer on marine aerosols. Journal of Geophysical Research, 107 (D7): 4053. Tie, X.X., Madronich, S., Walters, S., Zhang, R.Y., Rasch, P. and Collins, W. (2003) Effect of clouds on photolysis and oxidants in the troposphere. Journal of Geophysical Research, 108 (D20): 4642. Tolocka, M.P., Jang, M., Ginter, J.M., Cox, F.J., Kamens, R.M. and Johnston, M.V. (2004) Formation of oligomers in secondary organic aerosol. Environmental Science and Technology, 38: 1428–34. Trainer, M., Williams, E.J., Parrish, D.D., et al. (1987) Models and observations of the impact of natural hydrocarbons on rural ozone. Nature, 329 (6141): 705–7. Travis, D.J., Carleton, A.M. and Lauritsen, R.G. (2004) Regional variations in US diurnal temperature range for the 11–14 September 2001 aircraft groundings: Evidence of jet contrail inﬂuence on climate. Journal of Climate, 17 (5): 1123–34. Traub, M., Fischer, H., de Reus, M., et al. (2003) Chemical characteristics assigned to trajectory clusters during the MINOS campaign. Atmospheric Chemistry and Physics, 3: 459–68. Tuck, A. (2002) The role of atmospheric aerosols in the origin of life. Surveys in Geophysics, 23 (5): 379–409. United Nations Economic Commision for Europe (UNECE) (1991) Protocol to the 1979 convention on long-range transboundary air pollution concerning the control of emissions of volatile organic compounds or their transboundary ﬂuxes. United Nations Economic Commission for Europe, Geneva, Switzerland, ECE/EB.AIR/30. UNEP (2001) Stockholm convention on persistent organic pollutants (Final Draft) http:www.chem.unep.ch/pops/POPs_Inc/dipcon/convtext/disclaimer.htm. Vaida, V. (2002) Atmospheric aerosols as prebiotic chemical reactors. Abstracts of the American Chemical Society, 224: 029-PHYS, part 2. Walker, C.H. (2001a) Organic Pollutants – An Ecotoxicological Perspective. Taylor & Francis, London. Walker, J.C. (2001b) The performance of the human nose in odour measurement. Water Science and Technology, 44 (9): 1–7. Wallraff, H.G. (2001) Navigation by homing pigeons: Updated perspective. Ethology, Ecology and Evolution, 13 (1): 1–48. Wallraff, H.G. (2003) Olfactory navigation by birds. Journal für Ornithologie, 144 (1): 1–32. Wang, Y.H., Jacob, D.J. and Logan, J.A. (1998) Global simulation of tropospheric O3 –NOx – hydrocarbon chemistry. 3. Origin of tropospheric ozone and effects of nonmethane hydrocarbons. Journal of Geophysical Research, 103 (D9): 10757–67. Warneke, C. and de Gouw, J.A. (2001) Organic trace gas composition of the marine boundary layer over the northwest Indian Ocean in April 2000. Atmospheric Environment, 35 (34): 5923–33. Warneke, C., Luxembourg, S.L., de Gouw, J.A., Rinne, H.J.I., Guenther, A.B. and Fall, R. (2002) Disjunct eddy covariance measurements of oxygenated volatile organic compounds ﬂuxes from an alfalfa ﬁeld before and after cutting. Journal of Geophysical Research, 107 (D8): 4067. Watson, J.G. (2002). Visibility: Science and regulation. Journal of the Air and Waste Management Association, 52 (6): 628–713. Went, F.W., Slemmons, D.B. and Mozingo, H.N. (1967) Organic nature of atmospheric condensation nuclei. Science, 156 (3774): 543. Wigley, T.M.L., Smith, S.J. and Prather, M.J. (2002) Radiative forcing due to reactive gas emissions. Journal of Climate, 15 (18): 2690–6. Williams, J., Fischer, H., Harris, G.W., et al. (2000) The variability-lifetime relationship for organic trace gases: A novel aid to compound identiﬁcation and estimation of HO concentrations. Journal of Geophysical Research, 105 (D16): 20473–86. Williams, J., Pöschl, U., Crutzen, P.J., et al. (2001) An atmospheric chemistry interpretation of mass scans obtained from a proton transfer, mass spectrometer ﬂown over the tropical rainforest of Surinam. Journal Atmospheric Chemistry, 38: 133–66.

32

Volatile Organic Compounds in the Atmosphere

Williams, J., Holzinger, R., Gros, V., Xu, X., Atlas, E. and Wallace, D. (2004) Measurements of organic species in air and seawater from the tropical Atlantic. Geophysical Research Letters, 31 (23): L23S06. Yassaa, N. and Williams, J. (2005) Analysis of enantiomeric and non-enantiomeric monoterpenes in plant emissions using portable dynamic air sampling/solid-phase microextraction (PDAS-SPME) and chiral gas chromatography/mass spectrometry. Atmospheric Environment, 39 (27): 4875–84. Yokouchi, Y., Ikeda, M., Inuzuka, Y. and Yukawa, T. (2002) Strong emission of methyl chloride from tropical plants. Nature, 416 (6877): 163–5. Yu, S.C. (2002) Role of organic acids (formic, acetic, pyruvic and oxalic) in the formation of cloud condensation nuclei CCN, a review. Atmospheric Research, 53 (4): 185–217. Zimmerman, P.R., Greenberg, J.P. and Westberg, C.E. (1988) Measurements of atmospheric hydrocarbons and biogenic emission ﬂuxes in the Amazon boundary layer. Journal of Geophysical Research, 93 (D2): 1407–16. Zeng, G. and Pyle, J.A. (2003) Changes in tropospheric ozone between 2000 and 2100 modeled in a chemistry-climate model. Geophysical Research Letters, 30 (7): 1392.

Volatile Organic Compounds in the Atmosphere Edited by Ralf Koppmann Copyright © 2007 by Blackwell Publishing Ltd

Chapter 2

Anthropogenic VOCs Stefan Reimann and Alastair C. Lewis

2.1

Introduction

VOCs were ﬁrst identiﬁed as being involved in photochemical smog formation in the 1950s (Haagen-Smit 1952), and since that time our knowledge of their behaviour in the environment has grown enormously. In this chapter an overview is given on the temporal evolution of anthropogenic VOC emissions in different parts of the world and the relative importance of anthropogenic source groups. This is followed by a discussion about the atmospheric distribution of the anthropogenic VOCs from polluted urban areas to the most remote parts of the globe. The chemical behaviour of, atmospheric VOCs is discussed in relation with the formation of ozone and organic aerosols. The ﬁnal section of the chapter discusses the analytical methods for the detection of anthropogenic VOCs in the atmosphere. The emphasis of this chapter is on non-methane hydrocarbons (NMHCs) as oxygenated and halogenated hydrocarbons will be discussed separately in Chapters 4 and 5, respectively.

2.2

Sources of anthropogenic VOCs

The anthropogenic fraction of atmospheric VOCs is related to the unprecedented usage of fossil fuels for transport, the production of consumer goods and various industrial processes in the past centuries. Mainly three groups of anthropogenic VOCs can be distinguished: NMHCs, oxygenated hydrocarbons (OVOCs) and halogenated hydrocarbons (e.g. chloroﬂuorocarbons [CFCs] and hydroﬂuorocarbons [HFCs]). These substances are either directly emitted to the atmosphere or formed during combustion processes. The global emission of anthropogenic VOCs is estimated to amount to 186 Tg/year (EDGAR 2005). However, the majority of VOCs in the atmosphere are related to biogenic emissions, with estimated totals between 377 TgC/year (i.e. 442 Tg/year) (IPCC 2001) and 1 150 TgC/year (Guenther et al. 1995), which originate nearly exclusively from vegetation and small contributions from oceans and soils. The distinction between biogenic and anthropogenic VOCs in the atmosphere is far from straightforward because many VOC species are produced by both sources. Emissions of alkanes and alkenes, for example, are dominated by anthropogenic sources, but are also produced by soils, wetlands and oceans. On the other hand, OVOCs like acetone and methanol have dominant sources related to emissions from vegetation (Goldstein and Schade 2000;

34

Volatile Organic Compounds in the Atmosphere

Jacob et al. 2002, 2005; Singh et al. 2004), but are also emitted by from combustion of fossil fuel and biofuel. Furthermore, OVOCs are also produced by the oxidation of hydrocarbons during atmospheric transport, which contributes considerably to their abundance at rural and remote sites (Goldstein and Schade 2000; Singh et al. 2004; Solberg et al. 1996). Related to the distinction between biogenic and anthropogenic VOCs in the atmosphere, the newly developed measurements of isotopes of VOCs could become an important analytical tool (Goldstein and Shaw 2003). To make things even more difﬁcult, the deﬁnition of anthropogenic sources itself is delicate, as biomass burning, which could also be regarded as a natural process, has been dramatically enhanced by agricultural practices in the tropics. In this chapter the sources of VOCs will be discussed, using both emission inventories (bottom-up) and emissions derived from modelling of ambient measurements (top down). First, an overview on VOC emissions from different parts of the world and their temporal evolution will be given by using global and regional inventories. Then sources will be distinguished by emission activities and in relation with different groups of VOCs. Finally, methods for the modelling of emissions from measurements will be discussed.

2.2.1

Inventories for anthropogenic VOCs

The EDGAR inventory is a world-wide systematic approach to estimate emissions of anthropogenic greenhouse gases and air pollutants (including VOCs) down to a regional scale. It contributes to the Global Emissions Inventory Activity (GEIA) of IGBP/IGAC and has been used in the IPCC Assessments (IPCC 2001). Within EDGAR, sources are divided into 28 categories, which roughly can be aggregated to biofuel combustion, fossil fuel use and production, industrial processes, agricultural practices, biomass burning and waste treatment. Latest emissions are available for the year 2000 (EDGAR 32FT 2000) on a 1◦ × 1◦ grid (EDGAR 2005) (Figure 2.1). Global anthropogenic VOC emissions in the year 2000 were estimated to be 186 Tg/year, which represents a continuing increase in comparison with 1990 (153 Tg/year) and 1995 (160 Tg/year). The gap of over 20 Tg/year between 1995 and 2000 was predominantly caused by the biomass burning source in 2000 (49 Tg/year) being more than twice as high as in 1995 (22 Tg/year). This is due to a methodological change in the calculation of biomass burning emissions, using observations of burned area instead of burned area statistics, leading to enhanced emissions in Latin America and Africa. In Figure 2.2 the relative contributions of VOC emissions from aggregated activities are shown for the world, East Asia, the OECD (Organisation for Economic Co-operation and Development) and the United States. On a global average, the use of fossil fuels accounts for about 40% of the emitted VOCs. About 25% of the emissions are due to biomass burning and the rest is about equally divided between the use of biofuel and industrial processes. In East Asia (i.e. mainly China), biofuel combustion (e.g. from coal-burning cook stoves) and fossil fuel usage are dominating, with smaller emissions from industrial activities. In the OECD countries and in the United States, the partition between the sources is comparable, with about 50% of the VOC emissions caused by fossil fuel use, followed by a huge contribution from industrial processes and only minor sources from the remaining activities.

Global total: 1.8e + 11 kg (min. = 60.9, max. = 1.1e + 009) Unit: Gg NMVOC =0 0.0–0.1 0.1–1 1–2

2–10 10–50 50–100 100–2000

Figure 2.1 Gridded VOC emissions from anthropogenic sources in the year 2000 from EDGAR 32FT2000 (With permission from EDGAR 2005). This image appears in full colour in the plate section that follows page 268 as Plate 1.

36

Volatile Organic Compounds in the Atmosphere

100%

80%

60%

40%

20%

0%

Global average

USA

OECD

East Asia

Waste management

2701

717

498

215

Biomass burning

48746

1686

163

257

Biofuel combustion

30013

748

288

6093

Industrial processes

26669

6650

5339

2566

Fossil fuel use, production

77428

9129

8040

5642

Figure 2.2 A comparison of aggregated activities related to VOC emissions (Gg/year) in the world (global average), United States, OECD countries and East Asia (EDGAR 2005).

Furthermore, emission inventories on a continental level have been developed in Europe and the United States. For Europe the CORINAIR-EMEP initiative provides yearly emission estimates, either down to a 50 × 50 km2 grid, as national totals, or by activity sectors (EMEP 2005; Vestreng 2003). For the United States the National Atmospheric Emission Inventory (NEI) (NEI 2005) from U.S. EPA estimates emissions on an annual basis since 1970 by deﬁning three classes of sources: point sources (e.g. power plants), area sources (aggregates of small point sources) and mobile sources (vehicles and marine vessels). The ACESS project (Ace-Asia and Trace-P Modelling and Emission Support System; ACESS 2005) is a remarkable attempt to provide accurate emission estimates for the Asian continent. Emissions from this part of the globe are extremely difﬁcult to assess because economical parameters are changing fast and a great variety of poorly deﬁned sources exists (e.g. coal-burning cook stoves, biomass combustion). Emissions are either available for particular VOC species or on a 1◦ ×1◦ grid for the total VOCs. The total Asian anthropogenic emissions were estimated to be 52.2 Tg/year in the year 2000 (ACESS 2005; Streets et al. 2003) with residential combustion of coal and biofuels and transportation having the highest contribution. For China, Klimont et al. (2002) have estimated the VOC emissions in the year 2000 to be 15.6 Tg/year. Due to rapid growth in the use of personal transport and solvents in China, emissions are anticipated to grow fast; however, technical improvements (e.g. catalytic converters) are expected to mediate the increase (Klimont et al. 2002). This behaviour is in contrast to the development in the United States and Europe, where emissions have decreased for many years as a direct result of control strategies such as catalytic converters, better fuel efﬁciencies and an only moderate economic growth (Figure 2.3).

Anthropogenic VOCs

37

40

Tg/year

30 USA EU-15 China

20

10

0 1970

1980

1990

2000

2010

2020

Figure 2.3 Anthropogenic VOC emission estimates and projections from inventories in Europe (EU-15) (EMEP 2005; Vestreng 2003), United States (NEI 2005) and China (Klimont et al. 2002).

2.2.1.1

Regional and urban inventories

Numerous emission inventories have been developed in support of studies related to air quality, atmospheric chemistry and climate. In Europe examples for countries with a sophisticated emission inventory are Great Britain (National Atmospheric Emission Inventory, NAEI) (NAEI 2005) and Germany (Friedrich et al. 2002). In Great Britain emissions have been regularly estimated for many years on a very detailed level in terms of VOC species, spatial resolution and anthropogenic activities (Figure 2.4a). Lately, the German inventory has substantially improved by incorporating measurements at speciﬁc sources (e.g. validating the exhaust source with chassis dynamometer tests) (Schmitz et al. 2000). Serious efforts have been made within urban environments around the world to develop emission inventories for VOCs and other compounds related to tropospheric ozone formation. As an example for an inventory within a developing country, Figure 2.4b shows VOC emissions from the city of Delhi (India) (Gurjar et al. 2004). Figure 2.4 shows the dominant effect of transportation in Delhi in comparison with the moderate fraction related to road transport in Great Britain. Furthermore, total emissions in Great Britain decline since the early 1990s, whereas emissions in Delhi are still increasing at the end of the 1990s, although to a possibly smaller extent in the last years.

2.2.2

Important sources of anthropogenic VOCs

In this section the source activities which lead to emissions of anthropogenic VOCs will be discussed, using grouped activities and quantitative emissions for the year 2000 from EDGAR (2005).

2.2.2.1

Fossil fuel use and production

Largely three source categories can be distinguished within this group: mobile sources, stationary sources and sources due to production, storage and delivery of fossil fuel products. Emissions from this activity are estimated to sum up to a global total of 78 Tg/year,

38

Volatile Organic Compounds in the Atmosphere

3

(a)

Road transport Other transport

Tg/year

2

Extr and Distr fossil fuels Solvents 1

Processes Other

0 1970

1975

1980

1985

1990

1995

2000

(b) 250

Gg/year

200 Industry 150

Domestic Transport

100

Power plants

50 0 1990

1995

2000

Figure 2.4 VOC emissions from main source activities in (a) Great Britain (NAEI 2005) and (b) city of Delhi. Adapted from Gurjar (2004).

predominately due to road transport and oil production. For road transport, areas of highest emissions are located in the United States, Europe, Latin America and South East Asia, whereas highest shares of emissions from oil production can be found in regions with biggest oil drilling activities (i.e. Middle East, Latin America). Mobile sources can be further divided into emissions from the exhaust and fugitive emissions by evaporation. As VOCs from fossil-fuel-driven vehicles globally represent the most important anthropogenic activity, bottom-up inventories using dynamometer test measurements under laboratory conditions have a long history. With those experiments it is not only possible to assess emissions of VOCs from tailpipe exhausts (Heeb et al. 2000a; Schmitz et al. 2000; Siegl et al. 1999) but also to estimate non-tailpipe, fugitive emissions by evaporation (Pierson et al. 1999). According to the valid vehicle legislation, exhaust emissions are measured within driving cycles such as the US FTP-75 and the European driving cycle (EDC). In addition, socalled ‘real-world’ cycles are applied as well on chassis dynamometers. These cycles better represent transient driving at all relevant trafﬁc situations, for example, urban stop-andgo, extended highway driving and driving under cold start conditions. The overall vehicle emissions are then extrapolated by correctly weighting emissions for all relevant driving patterns and vehicle classes. Technical progress like the introduction of catalytic converter systems and more energy-efﬁcient motor systems led to lower emissions per driven distance

Anthropogenic VOCs

39

100% VOCs Benzene 80%

60%

40%

20%

0%

1994

1995

1996

1997

1998

1999

2000

2001

Figure 2.5 Relative decrease of average light-duty car exhaust emissions for VOCs and benzene in summers between 1994 and 2001 in California, with emissions in 1994 corresponding to 100%. Adapted from Kean et al. (2002).

for newer car models (Heeb et al. 2000b; Kean et al. 2002) (Figure 2.5). In addition, fuel composition has been changed over the years for both technical and health reasons. For example, the reduction of carcinogenic benzene in fuel has led to an over-proportional decline of benzene emissions from car exhausts in comparison with the remaining VOCs (Figure 2.5). However, as a noble metal based three-way catalyst also produces benzene from other aromatic compounds (Bruehlmann et al. 2005), the decrease of benzene levels in the fuel is only partially represented in the exhaust emissions (Heeb et al. 2002). Furthermore, time-resolved emission measurements on chassis dynamometers showed that in the wake of emission reductions by catalytic converters ‘cold start emissions’ are becoming more and more important (Baugh et al. 1987; Heeb et al. 2003; Weilenmann et al. 2005). The majority of the VOC emissions of today’s gasoline-fuelled vehicles are released during the ﬁrst few minutes after the engine start. These so-called ‘cold start emissions’ correspond to the time needed to reach the catalyst light-off temperature. Cold start emissions are relevant especially in urban areas, where short-range journeys represent a considerable part of driven distance. Furthermore, chassis dynamometer tests and sealed housings were used to assess the fugitive emissions from passenger cars (Pierson et al. 1999). Fugitive emission rates increase with both, higher ambient temperature and higher vapour pressure of the fuel contents. This behaviour was conﬁrmed by many studies in the warm season, where signiﬁcant fugitive emissions from car tanks were found (Borbon et al. 2003b; Harley et al. 1992; Kourtidis et al. 1999; So and Wang 2004). For assessing the real-world picture of VOC emissions due to road transport, tunnel measurements are a common method, as they give a representative image of the average emissions of the actual car ﬂeet of a certain area. Furthermore, they can be applied to validate emission inventories used for air pollution control models (Staehelin et al. 1998) or to check the effect of changing fuel compositions (Kirchstetter et al. 1999).

40

Volatile Organic Compounds in the Atmosphere

Table 2.1 VOC measurements in tunnels used to assess VOC emissions from mobile sources Year (reference) 1993 (Staehelin et al. 1998) 2003 (Stemmler et al. 2005) 1997 (Schmid et al. 2001) 1998 (Kurtenbach et al. 2002) 1996 (Touaty and Bonsang 2000) 1990 (Haszpra and Szilagyi 1994) 1994/97 (Kirchstetter et al. 1999) 1999 (Kean et al. 2001) 1992 (Gertler et al. 1996) 2000 (McGaughey et al. 2004) 1994 (Duffy and Nelson 1996) 2000 (Hwa et al. 2002) 2000 (Na et al. 2002)

Tunnel location

VOC species

Gubrist, Zurich (Switzerland) Gubrist, Zurich (Switzerland) Tauern (Austria) Wuppertal (Germany) Thiais, Paris (France) Budapest (Hungary) San Francisco Bay (USA) San Francisco Bay (USA) Fort McHenry/Tuscarora (USA) Houston (USA) Sydney Harbour (Australia) Taipei (Taiwan) Seoul (South Korea)

−C10 , OVOCs C2 − −C8 C4 − −C10 , OVOCs C6 − −C10 , OVOCs C2 − −C6 C2 − −C10 C2 − −C9 , OVOCs C2 − OVOCs −C10 C2 − −C8 , MTBE C5 − −C10 C2 − −C9 C2 − −C9 C2 −

In Table 2.1 important tunnel studies from different regions of the world are summarised. Interestingly, some of the tunnels were re-analysed after a certain time lag, which provides a valuable approach to study the success of emission reduction measures. Thus, emission reductions for VOCs in the range of 80–90% were found in Austria and Switzerland, when VOCs were re-measured in the same tunnels (Schmid et al. 2001; Stemmler et al. 2005). Furthermore, emission estimates for the Swiss tunnel were also in compliance with the historical and actual bottom-up inventories for mobile VOC sources for Switzerland. This indicates that the development of the VOC emissions during the last decade is well understood on the basis of ﬂeet composition and dynamometric test measurements. Contrary to these results, calculated emissions from a recent tunnel study in Houston were higher than those from former studies (McGaughey et al. 2004). This atypical result was explained by high ambient temperature during the most recent study, leading to higher fugitive emissions in the tunnel. In some of the tunnel studies also OVOCs were measured. Relative to the NMHC emissions, OVOC were found to be emitted in slightly lower, but still considerable, amounts (Schmid et al. 2001; Staehelin et al. 1998). In a tunnel in Houston (USA), methyl-t butylether (MTBE) was found to be one of the most abundant VOC species in tunnel air, which was related to fugitive emissions (McGaughey et al. 2004). Kean et al. (2001) found formaldehyde to be the most abundant OVOC (excluding MTBE) (accounting for 45% of the OVOCs) followed by acetaldehyde, aromatic aldehydes and acetone. They also detected a decline in emissions for the aldehydes, in the range of 45–70% between 1994 and 1999 from measurements in the same tunnel. Stationary emissions from the use of fossil fuel are due to industrial applications (e.g. reﬁneries and coke ovens), residential heating in the cold season or cooking. Globally, emissions are relatively small compared to mobile sources. However, for certain regions (e.g. China) stationary sources can be a substantial part of the emissions. One of these sources is the extensive use of coal and biofuels in stoves and cookers in the residential

Anthropogenic VOCs

41

sector, which is estimated to contribute to about 30% of total Chinese anthropogenic VOC emissions in the year 2000 (Klimont et al. 2002). A meticulous study of emissions from different types of stoves and biofuel showed that considerable differences exist for emissions of VOCs per produced energy unit (Tsai et al. 2003). Emissions related to production, storage and delivery of fossil fuels predominately occur in those regions where extensive fossil fuel drilling activities exist. However, fugitive emissions can also occur from the transport and distribution of the fuel, such as ships, road tankers and fuel stations. Especially, emissions from fuel stations can be substantial on a local or regional scale. This is not only of concern related to the formation of tropospheric ozone, but also to health issues, as considerable emissions of toxic compounds (i.e. benzene and MTBE) can occur. Natural gas emissions from the delivery infrastructure (pipelines, installations) have been found to be important in urban areas. Their source strength is temperature dependent (Borbon et al. 2003b) and leads to elevated concentrations primarily of ethane and propane (Derwent et al. 1995).

2.2.2.2

Industrial processes

While methods to calculate VOC emissions from fossil fuel use and production are well established, the knowledge of emissions from industrial processes is less certain. The global emissions of 27 Tg/year (EDGAR 2005) are mainly due to the use of VOCs as organic solvents with different physical and chemical properties (i.e. alkanes, ketones, aldehydes, esters and halogenated hydrocarbons). Solvents are used in many consumer goods, such as paints, adhesives and inks, for which the evaporation of the solvent is essential after usage. This leads to elevated emissions at higher temperature during the warm season of the year (Borbon et al. 2003b). Other, more industrial uses encompass the manufacture of pharmaceutics, the cleaning of metal surfaces, the extraction of oil seeds or printing. Nowadays, many of these industries have reduced their emissions in developed countries by either recycling VOCs or thermally destroying them. However, emissions in other parts of the world may still increase. Thus, Chinese VOC emissions from solvents and paints are expected to be nearly ﬁve times higher in 2020, compared to 1990 (Klimont et al. 2002).

2.2.2.3

Biofuel combustion

Biofuel combustion is a primary source of energy in developing countries. The EDGAR inventory for the year 2000 estimates highest regional emissions in Asia (13 Tg/year) and Africa (8 Tg/year) with a global total of 30 Tg/year (EDGAR 2005). The overwhelming part of the emissions derives from heating and cooking in the residential sector. In South America, additionally, a considerable part of VOC emissions from the transport sector is assigned to the biofuel source. This is due to Brazil’s usage of ethanol (derived from plant materials) as vehicle fuel.

2.2.2.4

Biomass burning

Especially in the tropics biomass burning is a common practice in agriculture. In the year 2000 global biomass burning emissions of VOCs was estimated to be 49 Tg/year (EDGAR 2005) with highest contributions from Latin America, Africa and South East Asia.

42

Volatile Organic Compounds in the Atmosphere

Andreae and Merlet (2001) reviewed the literature for gaseous and aerosol emissions of biomass burning. Their global VOC emissions attributed to biomass burning is 60 Tg/year, which is divided into emissions from savannah and grassland (26 Tg/year), tropical forests (22 Tg/year) and extra tropical forests (12 Tg/year). They attributed around 50% of the emissions to oxygenated species like alcohols, ketones, aldehydes and acids.

2.2.2.5

Waste management

VOC emissions due to waste management are low in comparison with other sectors and are estimated to account for about 1–2% of the total anthropogenic VOC sources (EDGAR 2005). Emissions arise from waste incinerators due to incomplete burning and from waste dumping sites, either directly (e.g. by evaporation of deposited solvents) or as organic decay products.

2.2.3 2.2.3.1

Sources of VOC classes Alkanes

Due to its low reactivity and considerable global sources of about 10–13 Tg/year (Gupta et al. 1998; Rudolph 1995), ethane is the most abundant alkane in the atmosphere. The main sources are emissions from exploitation and distribution of natural gas and from biomass burning (Rudolph 1995). Minor sources include emissions from engine exhaust and biogenic emissions from soils and wetlands. Propane as the next higher alkane in the homologous series has main anthropogenic sources related to the distribution of natural gas −C5 -alkanes are primarily emitted from evaporation of and petrochemical industries. C4 − fossil fuel, which is stored in tanks and from the exhaust. C6 -alkanes and higher homologous VOCs are mostly emitted as a consequence of their usage as solvents and from fossil fuel evaporation. However, as the vapour pressure gets lower with the increasing chain length of the VOC, the relative importance of the evaporation becomes smaller for higher molecular VOCs. On the contrary, spillage of fossil fuel becomes more and more signiﬁcant, especially in countries where measures for curbing refuelling losses are not in place.

2.2.3.2

Alkenes

Fossil fuels contain only small amounts of alkenes. Therefore, alkenes are emitted mainly from the vehicle exhaust (due to incomplete combustion of the fuel), from biofuel combustion and biomass burning. Ethene and propene are the most abundant species in polluted air masses. However, their concentrations decline rapidly in more remote areas because of their high reactivity in the atmosphere. Higher molecular alkenes are emitted in lower amounts, but can be important as precursors of tropospheric ozone in urban environments. Furthermore, alkadienes are emitted as a by-product of incomplete combustion processes. Examples are the carcinogenic 1,3-butadiene and isoprene. Interestingly, the latter was long believed to be exclusively of biogenic origin but its correlation with other products from fossil fuel combustion (Borbon et al. 2001; Reimann et al. 2000) suggests a small additional anthropogenic source.

Anthropogenic VOCs

2.2.3.3

43

Alkynes

Ethyne (acetylene) is the dominant alkyne in the atmosphere. In urban areas it is emitted from the incomplete burning of fossil fuel. In other areas (e.g. the tropics) emissions from biomass burning and biofuel combustion are most important (Andreae and Merlet 2001). Furthermore, small amounts of propyne and butyne are emitted by combustion processes and can be measured in polluted environments.

2.2.3.4

Aromatics

Aromatic VOCs are important components of fossil fuels and are predominantly emitted by vehicle exhausts, from fuel evaporation and spillage. In addition, toluene and xylenes also have signiﬁcant industrial emissions because they are used as solvents. The most abundant species in fossil fuels and consequently in urban air masses are benzene, toluene, ethyl benzene, xylenes, styrene and trimethyl benzenes. As benzene has carcinogenic properties it has been replaced in fossil fuels by other aromatics or MTBE. Benzene is also produced during the combustion of biofuel (Barrefors and Petersson 1995), which can be important, for example, in regions with a high ratio of residential wood burning.

2.2.3.5

OVOCs

Oxygenated VOCs are emitted from both anthropogenic and biogenic sources, although on the global scale their predominant sources are emissions from vegetation and biomass burning. A signiﬁcant fraction of OVOCs in the atmosphere are produced by the photochemical conversion of NMHCs to OVOCs (Jacob et al. 2002; Millet et al. 2004; Singh et al. 2000, 2004; Solberg et al. 1996). This secondary formation results in concentrations of OVOCs being normally higher than those of NMHCs in remote areas (Millet et al. 2004; Singh et al. 2004). Direct anthropogenic emissions of OVOCs are related to the incomplete combustion of fossil fuel from mobile sources and the use of OVOCs as solvents. Highest contributions from vehicular sources are from formaldehyde, acetaldehyde, Oxygenated aromatics and acetone (Kean et al. 2001). In addition, the fuel additive MTBE is emitted by evaporation and the exhaust (Kawamoto et al. 2003). However, due to possible detrimental effects on health and environment its use in gasoline has already been phased-out in California and probably will be on a global scale in the near future. After the phase-out, average concentrations of MTBE in California declined from over 1 ppb to 0.15 ppb (CARB 2005).

2.2.4

VOC emission estimations from atmospheric measurements

The aim of many atmospheric VOC measurements has been the assessment of VOC sources and their accurate coverage in emission models. This so-called top-down approach uses the temporal and spatial behaviour of ambient concentrations at given receptors (measurement sites) to infer emissions from a certain region. One of the most straightforward methods is the tracer-ratio technique, where measurements of a substance with unknown source strength are related to concurrent measurements of a substance with known emissions (a priori information). This technique is often

44

Volatile Organic Compounds in the Atmosphere

used in urban studies, where the chemical decay during transport has a negligible inﬂuence on measured concentrations (due to the short distance between the source and the receptor). Examples are the estimation of VOC emissions in Great Britain (Derwent et al. 2000), in Germany (Augsburg (Klemp et al. 2002; Mannschreck et al. 2002; Slemr et al. 2002a), Berlin (Winkler et al. 2002)), in Italy (Lombardy (Dommen et al. 2003)), in China (Hong Kong (Guo et al. 2004a)) and in Mexico (Mexico City (Arriaga-Colina et al. 2004)). Resulting emissions from the top-down approach can differ considerably from those of emission models. Thus, discrepancies between calculated and measured emissions in Augsburg were observed related to the relative contribution of source categories in different seasons (Mollmann-Coers et al. 2002). An illustration for the application in a remote atmosphere is the estimation of global emissions of relatively stable OVOCs from biomass combustion by measurements over the Paciﬁc Ocean (Singh et al. 2004). For the determination of the relative contribution of different sources to measure atmospheric VOCs, the following equation is often taken as the basis: X (n, m) = S(n, p) × Q(p, m), X represents the n × m matrix with a number of n measurements of the concentrations of m species. S is the matrix of the source contributions (n × p matrix) and Q is the matrix of the source proﬁles (p × m matrix, with p sources). There are two principal ways to solve this equation. The ﬁrst is the Chemical Mass Balance (CMB), where it is assumed that the source proﬁles (Q) are known and the equation can be solved with a least-square ﬁt. In the second approach, both S and Q are assumed to be unknown and the equation is solved with multi-variate statistical methods. Both methods have the principal limitation that the mass conservation is implicitly presupposed, which due to atmospheric reactions is only partly fulﬁlled. For the CMB model many studies have been performed in urban environments. Examples for the United States (Watson et al. 2001) are studies in Texas (Fujita 2001), Chicago (Scheff and Wadden 1993) and Michigan (Scheff et al. 1996). In Europe emissions have been estimated, for example, in Birmingham (Great Britain) (Hopkins et al. 2005) and in Helsinki (Finland) (Hellen et al. 2003). Lately, Zhao et al. (2004) developed the CMB method further by introducing measures to account for meteorological factors and human behaviour (i.e. weekdays/weekends). For the multivariate methods the Principal Component Analysis (PCA) is a widely used tool for analysing emission sources with ambient measurements (Buhr et al. 1992). With the help of correlated variables (e.g. concentrations, meteorological data) a smaller number of independent factors is identiﬁed (i.e. principal components, PCs) which can explain the variance in the data. The number of extracted PCs corresponds to the number of sources. This approach has been successfully used in many urban environments, such as London (Great Britain) (Derwent et al. 1995), Lille (France) (Borbon et al. 2003a), Zurich (Switzerland) (Staehelin et al. 2001) and Hong Kong (China) (Guo et al. 2004b). An example for the application of the PCA model in a more remote environment is a study on source contribution at a rural site in China (Guo et al. 2004c). Another multi-variate method is the Positive Matrix Factorisation (PMF), which unlike the PCA, does not rely on the information from the correlation matrix but utilises a

Anthropogenic VOCs

45

weighted least-squares minimisation scheme. Examples of source apportionment studies are from Santiago de Chile (Jorquera and Rappengluck 2004) or Houston (Kim et al. 2005; Zhao et al. 2004). An alternative approach for the detection of different sources, which contribute to ambient concentrations at a given receptor, was developed by Jobson et al. (2004) by adopting earlier work related to VOC reactivities (Jobson et al. 1998). The method is based on the fact that in an urban environment a signiﬁcant part of VOC concentration variability is due to variations in emissions. Applying this method, traces of industrial emissions have been detected at the side of the dominant vehicle exhaust source (Jobson et al. 2004). Furthermore, chemical transport models (CTMs) are used for the estimation of VOC emissions. Thereby, measured concentrations are compared with those resulting from the CTM model, when an emission inventory is used for the initialisation of the model and the chemistry and transport processes are accounted for. CTMs have, for example, been used successfully to validate VOC emissions from Europe by modelling measurements in Scandinavia (Hov et al. 1997) or on a smaller spatial scale to model emissions from urban areas like Paris (France) (Vautard et al. 2003) and Mexico City (Mexico) (West et al. 2004).

2.3

Atmospheric distribution of VOCs

The spatial and temporal distribution of VOCs in the atmosphere depends on the location of their sources, their reactivities and regional and global transport phenomena. In this section, a short historic introduction will be followed by a discussion of observations, ranging from urban centres to global background. Furthermore, general features like diurnal and annual cycles, as well as the spatial and temporal behaviour of VOC concentrations under different conditions will be discussed.

2.3.1

Historic VOC concentrations

The analytic methods for measuring VOCs in ambient air have been developed in parallel with the rapid increase in fossil fuel usage. First measurements were pioneered in the Los Angeles area, where air pollution problems became overwhelming in the 1950s. The most important analytical tool ever since has been gas chromatography (GC) with prior extraction from air onto sampling traps and subsequent analysis using detection by thermionic conductivity (TCD), by ﬂame ionisation (FID) or mass spectrometry (MS). Probably the ﬁrst ambient VOC measurement ever was made in 1956 (Eggertsen and Nelsen 1958), when VOCs from 6.4 l of Los Angeles air were analysed by a GC equipped with a TCD (Figure 2.6). Concentrations were extraordinarily high compared to today’s values (e.g. 180 ppb of ethyne, 150 ppb of propene and 130 ppb of isopentane). In fact, these high concentrations were required for VOCs to be detectable with the chosen method: ‘The threshold of detection is 0.05 p.p.m. or better, which brings the technique well within the concentration range of interest in urban air pollution studies’ (Eggertsen and Nelsen 1958). First measurements of VOCs using GC–MS followed (Quiram and Biller 1958), with an indicative concentration of 110 ppb of total VOCs. Further studies were performed in the same area, with ﬁrst measurements of toluene in the atmosphere (i.e. 60 ppb) in 1958 on a

‘Air’ Propane + N 2O

Peak height (mV)

0.4

0.3

Separating column (dimethylsulfolane on firebrick) Trapping column (dimethylsulfolane on firebrick)

Ethylene

0.2 Ethane

Propylene

0.1

n-Butane Acetylene

Propyne + I-pentene Isopentane

Isobutane

n-Pentane

0.0 45 −183°C

30

3

0

10

20

30

40

Charging 6.4 l 0°C

Figure 2.6 The ﬁrst published chromatogram of ambient VOC measurements showing light hydrocarbons measured in 1956 in Los Angeles (Eggertsen and Nelsen 1958). With permission from the American Chemical Society.

Anthropogenic VOCs

47

day where ‘the smog concentration was ofﬁcially reported as moderate to heavy’ (Farrington et al. 1959). From the very beginning of these measurements vehicular trafﬁc was suggested to be the predominant source of atmospheric VOCs, as concentration patterns were identical to those found in road tunnels (Eggertsen and Nelsen 1958; Farrington et al. 1959). Additional sources were proposed to contribute to the ambient concentrations, such as losses from gasoline evaporation, natural gas distribution, solvents and oil ﬁelds (Altshuller et al. 1971). First diurnal cycles of VOCs were shown by Gordon et al. (1968) and by Altshuller et al. (1971), with a morning maximum related to trafﬁc, followed by a decline during the day hours and an increase in the late afternoon. This was also conﬁrmed by Lonneman and Bufalini (1978), who found that vehicular emissions were dominant on weekday mornings but were normally smaller during the afternoon and on weekends. Simultaneously discussions about the limitations of air pollution policy measures already started and it was speculated that the complete control of automotive emissions, other combustion sources and organic solvents in the Los Angeles basin still is not likely to reduce hydrocarbons to the desired level, if most of these parafﬁnic hydrocarbons were associated with natural gas leakage and petroleum gas leakage (Altshuller et al. 1971). In order to follow the development of anthropogenic VOC emissions in the past 50 years it is illustrative to compare measurements in California starting in the 1950s (Singh and Zimmerman 1992) with the current average yearly concentrations in the same area (CARB 2005) (Figure 2.7). In the beginning, concentrations of benzene and toluene were about

60

50 Benzene Toluene

ppb

40

30

20

10

0 1955

1965

1975

1985

1995

2005

Year

Figure 2.7 Benzene and toluene concentrations in California in the past 50 years (CARB 2005; Farrington et al. 1959; Sexton and Westberg 1984; Singh and Zimmerman 1992).

48

Volatile Organic Compounds in the Atmosphere

40 and 60 ppb, respectively. Since then, levels have declined dramatically and average values are below 1 or 2 ppb for benzene and toluene, respectively. This indicates the ongoing success of the technical solutions such as catalytic converters, which counterbalances the ever increasing vehicular trafﬁc.

2.3.2

Common features of VOCs in polluted areas

In urban atmospheres VOCs exhibit certain recurrent patterns, which are dependent on human behaviour, the geographical location, meteorological parameters and the timedependent contribution of different sources. A typical diurnal cycle for an urban atmosphere is illustrated by measurements from the city centre of Zurich (Figure 2.8). As in many major cities, VOCs in Zurich are emitted predominantly by vehicular trafﬁc. Concentrations in the early morning hours are low, as emissions from the previous day have been diluted or degraded. After 6 am commuter trafﬁc starts and concentrations rise sharply. From about 10 am onwards, concentrations begin to decline steadily until about 4 pm. This behaviour is not only caused by the smaller amount of trafﬁc but also by an enhanced photochemical sink and increasing thermal convection during the daytime hours. This leads to a dilution of the polluted city air with cleaner air masses from above the convective boundary layer. During the evening hours commuter trafﬁc leads again to increased concentrations, which especially on warm, calm days in summer coincides with a stratiﬁcation of the air, caused by the faster cooling of the atmosphere near the surface relatively to the warm air above. Concentrations on a day-to-day basis are additionally inﬂuenced by meteorological phenomena. Generally, the horizontal wind velocity is the most important factor, leading to an efﬁcient dilution of local and regional emissions. As an example, VOCs concentrations from Zurich are shown in Figure 2.9, where stormy conditions (7–9 February 2004) led to a strong decline of measured VOC concentrations. Thereby, the wind not only advects more 2.5

2.0 C4-alkanes C5-alkanes C2-benzenes Benzene Toluene

ppb

1.5

1.0

0.5

0 0

6 AM

12 PM

6 PM

0

Figure 2.8 Average diurnal cycles of C4 -alkanes (n-butane and isobutane), C5 -alkanes (n-pentane and isopentane), C2 -benzenes (ethyl benzene, m-, p-, o-xylene), benzene and toluene in Zurich in the year 2004.

Anthropogenic VOCs

49

9.0

6 Benzene Isopentane Wind velocity

7.5

4

6.0

3

4.5

2

3.0

1

1.5

0

m/s

ppb

5

0.0 4

5

6

7

8

9

10

11

Date in February 2004 Figure 2.9 Concentrations of benzene and isopentane in Zurich (February 2004) showing a negative correlation with the wind velocity.

6 Benzene Toluene C2-benzenes

ppb

4

2

0 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 Year

Figure 2.10 Monthly mean concentrations (1993–2004) of aromatic VOCs at a suburban site (Dubendorf, Switzerland).

pristine air masses to the city but emissions also get transported faster into the surrounding areas. However, this can only conditionally be seen as a general feature, as point sources in the pathway of the advected air masses could also lead to a concentration increase. The seasonal cycle of VOCs is characterised by higher concentrations in winter than in summer. As an example concentrations of aromatic VOCs from Duebendorf (in the surrounding of Zurich, Switzerland) between 1993 and 2004 are shown in Figure 2.10. This recurrent annual cycle is caused by a combination of the generally lower chemical reactivity of the atmosphere in the cold season and higher emissions from fossil-fuel-driven vehicles (‘cold start’, see above). Furthermore, especially in European winter, long-lasting periods

50

Volatile Organic Compounds in the Atmosphere

with constant atmospheric conditions can occur, where emissions are trapped below stable stratiﬁed clouds. In addition, Figure 2.10 illustrates the declining concentrations of aromatic VOCs within the last years in Western Europe. This feature is concurrent with the development in the United States, already discussed for Figure 2.7.

2.3.3

VOCs in urban areas

In this section an overview on important ﬁndings from campaigns in urban environments in different parts of the world is given. Concentrations of anthropogenic VOCs from selected cities show a common pattern with concentrations being lower in Western Europe and Northern America compared to the rest of the world (Table 2.2). In fact, actual Table 2.2 Overview of recent VOC measurements in urban areas from different parts of the world. Houston (Jobson et al. 2004), London (Derwent et al. 2000), Athens (Bakeas et al. 2003; Rappenglück et al. 1998), Chinese cities (Barletta et al. 2005), Hong Kong (Guo et al. 2004b), Porto Allegre (Grosjean et al. 1998), Karachi (Barletta et al. 2002) Houston,a USA 2000

London, GB 1996

Ethane n-Propane n-Butane Isobutane n-Pentane Isopentane n-Hexane

5.90 3.15 1.11 1.21 0.37 0.91–1.23 0.29–0.41

3.4–3.9 2.3–2.6 3.2–4.9 0.8–2.6 0.7–1.0 2.4–4.1 0.2–0.4

Ethene Propene S-butene 1,3-Butadiene S-cis/trans-pentene Isoprene

1.83 0.45 0.14 0.05–0.09 0.02 0.11–0.18

3.0–3.5 1.8–2.1 0.5–0.8 0.2–0.4 0.3–0.4 0.1

2.1–34.8 0.2–8.2 0.2–12.5 0.02–2.5 0.04–14.7 0.04–1.7

0.41

2.9–7.1

2.9–58.3

Benzene Toluene Ethyl benzene m/p-Xylenes o-Xylene

0.32–0.34 0.40–0.42 0.06 0.16 0.06

1.2–1.9 2.7–3.6 0.5–0.7 1.5–2.1 0.7–0.8

Methanol Formaldehyde Acetone Acetaldehyde

3.90 3.90 3.25–3.70 1.94–2.47

Ethyne

Athens, Greece 1994, 2000

1.6–12.4b 1.7–14.2b 3.2–26.3 0.6–4.1

2.5–11.7 6.7–21.2 1.3–4.0 3.2–11.1 1.5–5.5 10.7–17.2 6.5–13.7b 12.3–15.1

Notes: Concentrations are in ppb; S- = sum of. a Concentration compromised by other compounds. b Median of concentrations from different locations.

43 cities, China 2001

3.7–17.0 1.5–20.8 0.6–14.5 0.4–4.6 0.2–7.7 0.3–18.8 0.1–3.6

0.7–10.4 0.4–11.2 0.1–2.7 0.4–15.3 0.1–6.0

Hong Kong, Porto China Allegre, 2001 Brazil 1996

3.6–4.9 5.6–8.8 2.8–3.6 1.5–2.2 3.1–4.7

1.3–2.0 1.8–2.6 0.3–0.4 0.8–1.3 2.1–2.5 13.1–13.5 1.2–1.3 1.6–1.8 0.5–0.6

Karachi, Pakistan 1998/99

2.4 46.1 8.3 4.8 5.9 10.6 6.4

93.0 41.0 19.8 11.0 13.4 12.1 7.5

42.4 13.8 12.8 1.7 5.3 0.2

19.0 5.5 2.8 0.8 0.4 0.8

65.8

18.0

6.5 9.7 2.7 5.4 2.1

5.2 7.1 1.1 2.1 1.0

4.4 2.4

Anthropogenic VOCs

51

concentrations in many cities in developing countries are comparable to those from the ﬁrst VOC measurements in Los Angeles almost 50 years ago. Partly, this is caused by higher evaporative emissions, especially in tropical regions, but mainly by less advanced technical solutions for efﬁcient emission reductions. In North America the early studies in Los Angeles (Section 2.3.1) were followed by campaigns in practically all major cities. A summary of the work of the late 1970s is given by Sexton and Westberg (1984), where data from ambient air monitoring programmes in seven urban centres (Houston, Philadelphia, Baltimore, Washington, Newark, Boston, Milwaukee) and six rural areas are compiled. VOC compositions in cities were found to be more or less identical and the average VOC concentrations in rural areas were at least an order of magnitude below mean urban concentrations. More than a decade later Solomon et al. (2000) provided an overview of ozone-related studies in Northern America and Europe under a wide range of geographical and climatological conditions. Apart from Los Angeles, especially Houston (Texas) has been a major point of interest because of the inﬂuence of its large industrial facilities in the vicinity of the city. After Siddiqi and Worley (1977) ﬁrst addressed the special conditions in Houston, it was recently found that several VOCs like short-chained alkenes and alkanes were systematically inﬂuenced by petrochemical industry emissions (Berkowitz et al. 2004; Kleinman et al. 2002). In Europe atmospheric VOC measurements were lacking some years behind the United States. One of the ﬁrst measurements was performed in 1969 by Grob and Grob (1971), who measured 51 ppb of benzene and 39 ppb of toluene in the air of Zurich. Vehicular emissions are unambiguously detected as the main source, followed by evaporation of fuel (in summer), stationary combustion (in winter) and natural gas emissions. VOC concentrations in North Western European cities are comparable to those in Northern America (Derwent et al. 1995, 2000; Vautard et al. 2003; Winkler et al. 2002). On the other hand, cities in Southern Europe like Athens (Greece) (Rappenglück et al. 1998) and Rome (Italy) (Brocco et al. 1997) endure notorious air pollution problems during the warm summer days. In Athens, for example, pollution is enhanced by a local land-sea breeze, which exports air from the city in the morning and returns it in the afternoon (Flocas et al. 2003; Gusten et al. 1988; Rappenglück et al. 1998). Remarkably, maximum concentrations during stagnant periods reached 61 ppb for benzene and 103 ppb for toluene (Rappenglück et al. 1998) – which, once again, is comparable to values in the 1960s in Los Angeles. Vehicle emissions, fuel evaporation and paint solvents were found to be the main sources in Athens (Moschonas and Glavas 1996). In Italy several campaigns have been performed in the heavily industrialised Po Basin between 1998 and 2003 (Northern Italy) (Dommen et al. 2002; Duane et al. 2002; Neftel et al. 2002; Prevot et al. 1997; Steinbacher et al. 2005a, 2005b). Concentrations were inﬂuenced by both vehicle emissions and industrial solvents. A pronounced difference in the ratio of toluene to benzene was detected during weekends and during main holiday periods, indicating a much reduced amount of industrial emissions during these periods (Steinbacher et al. 2005b). In Japan and Australia measurements have been performed in several major urban agglomerations (e.g. Yamamoto et al. 2000). Resulting concentrations are comparable to those in Northern America, as reducing emissions, for example, by using catalytic converter systems, is a common practice in these countries. Diurnal variations were found to

52

Volatile Organic Compounds in the Atmosphere

be consistent with trafﬁc activities and aromatic VOC concentration was in the lower ppb range. In Africa, South America and Asia the development of so-called megacities creates a big challenge for air pollution control policy. Funding for abatement measures is limited, population and trafﬁc is growing and the inefﬁcient trafﬁc ﬂow leads to a vicious circle with even higher air pollution (Gurjar and Lelieveld 2005). In Africa measurements are restricted to cities in the North of the continent (Abu-Allaban et al. 2002; Cecinato et al. 2002; Doskey et al. 1999). Concentrations are generally higher than in Northern America (e.g. benzene: 8–43 ppb, toluene: 17–111 ppb and n-butane: 71–174 ppb (Abu-Allaban et al. 2002)). In Central America, Mexico City endures enormous air pollution problems because its situation on a high altitude basin leads to a combination of trapped emissions and more efﬁcient ozone formation. Average total VOC concentrations in the morning have been measured to be as high as 1.1–6.7 ppmC (Arriaga-Colina et al. 2004). Liqueﬁed petroleum gas (LPG) represents 16% of the total fuel consumption in Mexico City, which effects in propane and butanes (the main compounds of LPG) to constitute 29% of the atmospheric VOCs (Gasca et al. 2004). In Southern America, Brazil is another focus of interest because ethanol is used as vehicle fuel and accordingly was found to be one of the most abundant VOCs in urban air (Grosjean et al. 1998, 2002). In Asia measurement campaigns have been performed in various urban areas (e.g. Seoul, South Korea (Na et al. 2003); Bombay, India (Rao et al. 1997) and Karachi, Pakistan (Barletta et al. 2002)). In most of the Asian cities indications for considerable solvent emissions have been found in addition to the dominant trafﬁc source (Na et al. 2003; Rao et al. 1997). Concentrations of total VOCs are in the range of several hundreds of ppbs, indicating the need for more efﬁcient emission control. In China the unconstrained growth of the transport sector and the lack of initiatives for measures to reduce emissions are the basis for considerable air pollution problems (Fu 2001). Roadside measurements show the overwhelming trafﬁc source, and concentrations were found to be between tens and hundreds of ppb (Liu et al. 2000). In fact, driving cycle tests of in-use Chinese motor vehicles indicated that VOC emission factors were 5–10 times higher than levels in industrialised countries (Fu 2001). Interestingly, concentrations measured in Hong Kong were comparable to those in other Chinese cities (Guo et al. 2004a; So and Wang 2004). Analysis of recent measurements of VOCs within 43 cities in China (Barletta et al. 2005) results in trafﬁc to be the biggest source. However, in some of the cities also the burning of biofuel and coal as well as natural gas leakages had a considerable inﬂuence.

2.3.4

VOCs in rural and continental background areas

Moving from the urban environment to moderately remote regions, concentrations of anthropogenic VOCs decline depending on the distance of the site to urban areas and the inﬂuence of local sources (Sexton and Westberg 1984). Concentrations are smaller than in urban areas but ratios between the species are generally in good agreement

Anthropogenic VOCs

53

4

ppb

3 C2-benzenes Toluene Benzene 2

1

0 Zurich (centre)

Zurich (suburban)

Rigi (remote)

Jungfraujoch (high altitude)

Figure 2.11 Annual mean concentrations of aromatic VOCs in 2003 at different sites in Switzerland, ranging from the polluted city centre of Zurich to the high Alpine site of Jungfraujoch.

with transported urban pollutants and their photochemical transformation (Borbon et al. 2004). As an example for the effect of dilution and degradation during transport, the development of aromatic VOC concentrations from an urban area to a background site is shown in Figure 2.11. Highest concentrations are measured in the city centre of Zurich. Values are already slightly lower at the suburban site and decline to about a 15–40% of their initial value when they reach the rural station of Rigi. Finally, at the high Alpine background site of Jungfraujoch (3 580 m asl) concentrations decrease to 5–20% of their urban concentration. Thereby, toluene and C2 -benzenes declined more rapidly than benzene because of their shorter atmospheric lifetime. However, short-term meteorological phenomena could mask this general feature. Thus, a strong increase of VOC concentrations can be found at elevated sites during summer afternoons, which is caused by convective transport of air masses from the polluted boundary layer (Helmig et al. 1996; Henne et al. 2006; Li et al. 2005; Prevot et al. 2000a). Concentrations of OVOCs in rural regions tend to grow due to photochemical production during transport and are therefore gaining in relative importance to the other VOCs. Thus, in European summer a considerable part of the measured oxygenated VOCs at rural background sites could be attributed to the secondary production from precursor VOCs (Borbon et al. 2004; Solberg et al. 1996).

2.3.5

VOCs in global background regions

Background concentrations of VOCs exhibit annual cycles, which depend on the hemispheric emissions and on the temporal behaviour of the oxidation capacity of the atmosphere. In addition, concentrations can be enhanced even at the most remote locations due to advection of polluted air masses. For practical reasons it has not been possible to continuously measure VOCs in the global background. Data coverage especially in the

54

Volatile Organic Compounds in the Atmosphere

Southern hemisphere is related to ship cruises or air plane campaigns, mostly performed in Austral summer. In view of the quantity of campaigns at background sites within the last decades it is impossible to give an all-embracing overview of the measured concentrations. However, without being comprehensive, examples of large-scale measurements or their compilation can be found in Rudolph and Johnen (1990), Koppmann et al. (1992), Singh and Zimmerman (1992), Donahue and Prinn (1993), Blake et al. (1997), Gros et al. (1998), Emmons et al. (2000), Singh et al. (2000), Lewis et al. (2001), Boudries et al. (2002) and Warneke and de Gouw (2001). VOCs in background air show an annual cycle with lower concentrations at the end of the summer because of the higher abundance of the oxidising agents (i.e. mainly OH radicals) during the warm season. The amplitude of the seasonal cycles is highest for substances which have an intermediate lifetime of days to weeks, as stable compounds do not signiﬁcantly degrade during transport. Substances which have an atmospheric lifetime in the range of the transport time are partly oxygenated but still reach the remote areas in considerable amounts. On the other hand, very fast-reacting species (e.g. alkenes) are degraded almost totally during transport, which results in very low concentrations in remote areas (Hopkins et al. 2002; Lewis et al. 2001). VOC background concentrations are highest in the mid-latitude of the Northern hemisphere (30◦ –60◦ N) because of the higher amount of fossil fuel usage in the United States and Europe in comparison with other regions of the world and because the VOC lifetime is shorter than the average interhemispheric exchange time (Gupta et al. 1998). On a global level the relatively stable ethane, ethyne and propane are the most important anthropogenic NMHCs, although a considerable fraction of these substances is also emitted by biological processes (Rudolph and Johnen 1990). The abundance of these gases has been relatively stable in the last decades, as Antarctic ice core measurements showed no increase for propane and ethyne and only a minor increase (+1.6 ppt/year) for ethane since 1975 (Kaspers et al. 2004). These values are in good agreement with values of atmospheric campaigns in the Southern Hemisphere (Boissard et al. 1996; Gros et al. 1998; Singh et al. 1988) with the exception of propane being particularly high in the TROPOZ II experiment (Boissard et al. 1996). Concentrations of OVOCs normally dominate over those of NMHCs in remote oceanic air masses (Lewis et al. 2005; Singh et al. 2000), due to the oxidation of primarily emitted VOCs during transport and biological sources.

2.3.5.1

Vertical distribution

Anthropogenic VOCs are mainly emitted within the lowermost layer of the troposphere. Chemical reactions and limited transport to the higher layers of the atmosphere lead to a decline in concentrations at higher altitudes (Blake et al. 2001; Boissard et al. 1996; Rudolph 1988). The rate of decrease is higher for fast-reacting species and only relatively stable compounds can actually reach the uppermost layers of the troposphere or the stratosphere (Blake et al. 1996, 2003b; Scheeren et al. 2003; Singh et al. 1997, 2000). This general feature is, however, at times disturbed by plumes of anthropogenic emissions introduced to higher altitudes by uplifts related to meteorological processes

Anthropogenic VOCs

55

(Fischer et al. 2003; Gros et al. 2004; Jacob et al. 2003) or to topographical barriers, such as mountain chains (Henne et al. 2006; Li et al. 2005; Prevot et al. 2000b; Purvis et al. 2003).

2.3.5.2

Intercontinental transport

In the past years more and more campaigns put the spotlight onto the role of transport of air pollutants from one continent to another (Stohl and Eckhardt 2004). Thus, transport of VOCs from Asia into unpolluted Paciﬁc air masses has been detected, leading to elevated concentrations in the higher troposphere (Blake et al. 2003a; Heald et al. 2003; Jacob et al. 2003). Furthermore, transport from the United States to Europe has been detected (Stohl et al. 2003). This issue is expected to become more and more important with emerging megacities in the tropics and Asia, with China as a special point of concern (Elliott et al. 1997).

2.4

Chemical behaviour of VOCs in the atmosphere

After their release into the atmosphere, VOCs are oxygenated by photochemical processes, which ﬁnally lead to their removal from the atmosphere. For most VOCs the process is initiated by atmospheric radicals like OH, O3 , NO3 and Cl, with the OH radical being by far the most important reactant. Only some OVOCs (e.g. formaldehyde and acetone) are susceptible to destruction by direct photolysis, as they absorb radiation in the wavelength range above 290 nm. The atmospheric lifetime of an individual VOC species is dependent on its chemical structure, the radical concentration and the intensity of solar radiation. When VOCs are degraded in polluted air masses, NO is oxygenated to NO2 , which then gets photolysed and contributes to the formation of tropospheric ozone. As this is a key issue in air pollution control, different concepts have been developed to account for the ozone formation potential of the various VOC species. Globally, the fast-reacting biogenic isoprene and terpenes are more important than anthropogenic VOCs. However, within urban environments anthropogenic VOCs are one of the dominant drivers of air pollution.

2.4.1

VOC degradation pathways

The hydroxyl radical (OH), as the most important substance involved in the degradation of VOCs, has an average global abundance of about 1.1 × 106 radicals/cm3 (Prinn et al. 2005). Its abundance is highest in summer during day hours and, conversely, lowest in winter and during the night. The importance of the Cl radical is still under scrutiny on the global scale, although indications of reactions have been found in at least one urban area (Houston) (Tanaka et al. 2003). For the alkenes also reaction with O3 becomes competitive in polluted areas, whereas the reaction of alkanes and alkenes with NO3 is mainly important during the night (Geyer et al. 2001). Reaction mechanisms and rates coefﬁcients for the degradation of many species have been continuously reviewed by Atkinson and Arey (2003). The determination of rate coefﬁcients for atmospheric reactions has been the result of meticulous laboratory experiments using pulsed and ﬂow experiments. For new or exotic substances, where reaction rate coefﬁcients are not available structure–reactivity relationship methods exist for their estimation.

56

Volatile Organic Compounds in the Atmosphere

2.4.1.1

Degradation of atmospheric alkanes

Alkanes are attacked by OH (or Cl) radicals via hydrogen abstraction, as illustrated for ethane: CH3 CH3 + OH → CH3 CH•2 + H2 O For longer alkanes, with different types of hydrogen–carbon bonds, reaction rates for the OH abstraction are decreasing with the number of hydrogen atoms attached to the same −CH2 → − −CH3 . carbon atom. Therefore, the abstraction is favoured as follows: ==CH → − The resulting alkyl radical is fast reacting with O2 to form an alkyl peroxy radical: CH3 CH•2 + O2 + M → CH3 CH2 O•2 + M In the presence of NO (>10–30 ppt) oxygen is abstracted to form an alkoxy radical: CH3 CH2 O•2 + NO → CH3 CH2 O• + NO2 The produced NO2 molecule promotes the tropospheric ozone formation, whereas the alkoxy radical, depending on the molecule, can react via the following pathways: (a) thermal decomposition, (b) isomerisation and (c) reaction with molecular oxygen (O2 ). (a)

The breaking of the carbon skeleton leads to the formation of a stable aldehyde and a smaller peroxy radical, which again can react as described above. (b) Isomerisation occurs by internal hydrogen abstraction and eventually leads to a hydroxycarbonyl molecule. (c) Reaction with O2 leads to the formation of a stable ketone via the abstraction of a hydrogen atom, which is attached to the same carbon atom as the oxygen radical. Furthermore, alkyl peroxy radicals can also attach an NO or an NO2 molecule. In the case of NO this leads to the formation of relatively stable organic nitrates, whereas with NO2 peroxynitrates are produced: CH3 CH2 O•2 + NO + M → CH3 CH2 ONO2 + M CH3 CH2 O•2 + NO2 + M → CH3 CH2 O2 NO2 + M The yields of the organic nitrates are normally higher for longer chain alkanes. Peroxynitrates (RCH2 OONO2 ) are not be mistaken for peroxyacyl nitrates RC(O)OONO2 (see below). They decompose rapidly at 298K but are more stable in colder temperature. In very clean environments the peroxy radicals can also react with other peroxy radicals (i.e. HO2 or organic peroxy radical). Reaction with HO2 results in the formation of more stable organic peroxides: CH3 CH2 O•2 + HO2 → CH3 CH2 OOH The combination with another peroxy radical is either a sink for the radicals and produces alcohols, ketones and organic acids or leads to alkoxy radicals, which reacts as described above: CH3 CH2 O•2 + CH3 CH2 O•2 → CH3 CH2 OH + CH3 CHO + O2 CH3 CH2 O•2 + CH3 CH2 O•2 → CH3 CH2 O• + CH3 CH2 O• + O2

Anthropogenic VOCs

2.4.1.2

57

Degradation of atmospheric alkenes

In the case of alkenes OH radicals preferably react by addition to the C==C double bond as illustrated for ethene: −CH•2 −CH2 − CH2 ==CH2 + OH → HO− This hydroxyalkyl radical attracts an oxygen molecule to form a hydroxyalkyl peroxy radical: −CH•2 + O2 → HO− −CH2 O•2 −CH2 − −CH2 − HO− which predominately is converted to a hydroxy alkoxy radical via its reaction with NO. −CH2 O•2 + NO → HO− −CH2 O• + NO2 −CH2 − −CH2 − → HO− The hydroxy alkoxy radical can then either react with O2 via H-abstraction to become a glycolaldehyde or by cleavage to form an aldehyde and a hydroxy alkyl radical, which can react again as described above. Furthermore, larger chained molecules have the possibility to undergo isomerisation. The second major pathway for the destruction of alkenes is their reaction via absorption of an ozone molecule (O3 ) over the C==C double bond to form an instable ozonide intermediate, which then decomposes to a carbonyl compound and a Criegee intermediate, as shown for ethene: CH2 ==CH2 + O3 → HCHO + [H2 COO]∗ Criegee intermediates can either stabilise by collision or decompose further. After stabilisation, a possible pathway leads via reaction with water vapour to the formation of organic acids. For further details see, for example, Calvert et al. (2000).

2.4.1.3

Degradation of atmospheric aromatic VOCs

The degradation of aromatic VOCs proceeds by two different pathways: (a) by abstraction of a hydrogen atom from one of the alkyl substitute groups or (b) by OH radical addition to the aromatic ring. In reaction (a) the stable product is an aromatic aldehyde (e.g. toluene leading to benzaldehyde). Reaction (b) is the predominant pathway for the degradation of aromatic VOCs. After the addition of the OH radical to the aromatic ring, molecular oxygen is added to build a cyclic hydroxy peroxy radical. In the following the ring structure is opened and formation of epoxy compounds, different saturated and unsaturated dicarbonyl radicals and ﬁnally methylglyoxal has been observed. Due to a large number of laboratory studies on the atmospheric oxidation of aromatic hydrocarbons in the past years (see the extensive review by Calvert et al. 2000), considerable improvement in the understanding of detailed chemistry has been achieved. However, ring-opening routes (b) are still speculative, whereas ring-retaining routes (a) are comparatively well understood (Hamilton et al. 2003).

2.4.1.4

Degradation of atmospheric OVOCs

OVOCs are either directly emitted by anthropogenic or biogenic sources or are produced in the degradation of VOCs, as described above. For the OVOCs the abstraction of a hydrogen atom from the carbon chain via an OH radical is the predominant reaction. Furthermore, if C==C double bonds are present also the addition of O3 is a possible initial step. The resulting

58

Volatile Organic Compounds in the Atmosphere

peroxy radicals then react as described above. For those OVOCs which have UV-absorbing groups (e.g. aldehydes, ketones, organic peroxides and organic nitrates) also the photolytic degradation can be signiﬁcant. Aldehydes degrade by photolysis and by reaction with OH radicals. Whereas HCHO is degraded predominantly by photolysis, higher aldehydes react mainly with OH radicals. Products of the photolysis ﬁnally lead to the net production of HO2 , which can then be a source for OH radicals, as shown for formaldehyde (HCHO): HCHO + hν → CO + H2 HCHO + hν → HCO + H HCO + O2 → HO2 + CO H + O2 + M → HO2 + M The initial steps of the aldehyde degradation by OH radicals proceed as shown for the example of acetaldehyde: CH3 CHO + OH + O2 → CH3 C(O)O•2 The produced peroxy radical is a precursor for the PAN formation, if enough NO2 is available. The most important peroxyacylnitrate is the peroxyacetyl nitrate (PAN, CH3 C(O)O2 NO2 ), which is well known as eye irritant in emog situations. Pan and organic nitrates (see above) are relatively stable in cold temperatures and are therefore important for long-range transport as they act as reservoirs for reactive nitrogen. In the presence of sufﬁcient NO, however, the peroxy radical of a Cn -aldehyde reacts dominantly with NO, leading to a Cn−1 -aldehyde and CO2 . For the OH radical initiated degradation of ketones the reaction proceeds by H-atom abstraction and subsequent formation of alkoxy radicals. The degradation of alcohols is mainly initiated by OH radicals. If the H-atom is abstrac−H of the CHOH or CH2 OH group, the following reaction of O2 with the ted from C− α-hydroxyradical leads to the formation of a ketone for secondary alcohols or of an aldehyde for primary alcohols. This is shown for 1-butanol and 2-butanol. OHCH2 CH2 CH2 CH3 + OH → OHCH• CH2 CH2 CH3 + O2 → CHOCH2 CH2 CH3 CH3 CH(OH)CH2 CH3 + OH → CH3 C• (OH)CH2 CH3 + O2 → CH3 C(O)CH2 CH3 Noteworthy, no NO is needed for these reactions to form stable products. −H, which leads to reactions Furthermore, OH can abstract H-atoms from other C− analogous to those for alkanes.

2.4.1.5

Lifetimes of VOCs

For the determination of the decline of VOCs in the atmosphere the following equation can be used for the description of its reaction with a radical (as shown for OH), Ct = C0 exp(−k[OH]t ) where Ct represent the concentration at the time t and C0 at the source, k is the reaction rate coefﬁcient and [OH] is the abundance of OH radicals. k can be replaced by the pseudoﬁrst-order rate coefﬁcient k = k × [OH] (i.e. reaction rate) when the [OH] concentration is assumed to be constant.

Anthropogenic VOCs

59

The atmospheric lifetime (T ) of VOCs can then be calculated by the sum of the reciprocal values of the individual rates of the radical reactions (i.e. for OH, O3 , Cl, NO3 ) and the photolysis rate constant (jA ). + jA−1 T = kOH −1 + k −1 + k −1 + k Cl O NO−1 3

3

Thus, T denotes the time that is needed until 1/e of the initial concentration is left. Lifetimes of VOCs in the atmosphere are in the range of days to months for alkanes and alkynes, but can be as short as some hours for the most reactive alkenes, such as terpenes, isoprene and butenes (Atkinson and Arey 2003).

2.4.2

Usage of VOC reactivities for analysis of observations

The speciﬁc atmospheric reactivities of VOCs have been used to address several questions in atmospheric science. The ratio of concentrations of VOCs with different lifetimes has been used for the estimation of the air mass ‘age’. Which means by using the ratio of VOCs (with different lifetime) at the source and at the receptor together with the average abundance of the degrading radicals during transport, it is possible to deduce the travel time of the air parcel. This concept has been used in many studies, ranging from areas downwind of regional sources (Kleinman et al. 2003; McKeen et al. 1996; Roberts et al. 1984) to continental outﬂow (Goldan et al. 2004). However, as pointed out by McKeen and Liu (1993), this idealistic concept is inﬂuenced by non-linear processes, such as sources on the transport path and mixing of different concentrations of VOC in the background air. Vice versa this approach has been used to determine the average OH radical abundance during transport (Kleinman et al. 2003; Kramp and Volz-Thomas 1997; Williams et al. 2000; Wingenter et al. 1999). Furthermore, ratios of VOCs have been used to assess the inﬂuence of Cl radicals to the decline of VOCs in the Arctic boundary layer, as certain species have different rate coefﬁcients for the reaction with OH or Cl (Hopkins et al. 2002; Rudolph et al. 1996). Another approach uses the fact that the spatial and temporal variability of hydrocarbon mixing ratios display an inverse power law relationship to atmospheric lifetime (i.e. variability lifetime relationship). This has been applied to calculate the average OH radical abundance over the Paciﬁc Ocean (Ehhalt et al. 1998) and to assess the inﬂuence from sources to measured concentrations (Hopkins et al. 2005; Jobson et al. 1999).

2.4.3 2.4.3.1

Impacts of atmospheric VOCs Tropospheric ozone formation

Atmospheric VOCs inﬂuence the physical and chemical behaviour of the troposphere in several ways. Probably the most important process in relation to ‘air pollution’ is their contribution to the photochemical ozone formation. The inﬂuence of different VOC species is dependent on the spatial and temporal pattern of emissions, photochemical reaction rates and the molecule-dependent potential for producing ozone. The most widely used approaches to rank VOCs according to their ozone production ability under atmospheric

60

Volatile Organic Compounds in the Atmosphere

conditions are the photochemical ozone creation potential (POCP) (Derwent and Jenkin 1991) and the maximum incremental reactivity (MIR) scale (Carter 1994). The POCP describes the amount of ozone which is built from the introduction of a VOC into the atmosphere. For the calculation of the POCP, reactions from a chemistry model are linked with trajectory models under real-world conditions. For the comparison of the ozone-building potentials of different VOCs, the POCP for a VOCi is calibrated by ratioing its absolute values with those of another VOC (e.g. ethene), which can be expressed in the following equation: POCPi = simulated extra amount of ozone due to VOCi /simulated extra amount of ozone due to ethene × 100 Applications have been the regional scale ozone formation over periods of several days under European conditions using a chemistry transport model and a trajectory model (Andersson-Skold and Holmberg 2000; Derwent et al. 1998). The MIR scale values are determined from irradiations of simpliﬁed model photochemical systems. They have, for example, been used to assess the ozone formation over periods of up to a day under optimum VOC/NOx conditions in urban environments in the United States (Carter 1994; Carter et al. 1995).

2.4.3.2

Formation of SOAs

Secondary organic aerosols (SOAs) are a major component of the total aerosol mass and can therefore contribute to the modiﬁcation of the radiative balance of the atmosphere induced by particulate matters (Kanakidou et al. 2005). Apart from biogenic VOCs (e.g. terpenes) also anthropogenic aromatic VOCs have been found to contribute to the formation of SOAs (Odum et al. 1997). Thereby the atmospheric degradation of gaseous aromatic VOCs leads to the addition of a polar functional group (e.g. carboxylic acid, aldehyde, ketone and nitrate). In comparison with the initial substances oxygenated products exhibit a lower vapour pressure and higher water solubility, which favours their transfer into the particulate matter. Furthermore, oxidation products of aromatic VOCs have been found to form oligomers, which again are incorporated into the aerosol phase (Kalberer et al. 2004), leading to lower volatilities of the SOA and subsequently higher aerosol yields.

2.5

Measurement techniques

The wide range of both short- and long-lived VOC species that are of interest in atmospheric science, coupled to extremely low concentrations and a requirement often for in situ automated analysis, has led to the development of many techniques and methodologies. While some atmospheric species such as organic acids, peroxides and aldehydes are measured using high performance liquid chromatography (HPLC), the majority of organic species are analysed using capillary GC and there is an extensive literature dating back many years (Ioffe et al. 1977). In recent years, on-line MS detectors have found considerable application in atmospheric measurements of VOCs. The diversity of compounds that are of interest has resulted in almost every kind of analytical detector ﬁnding a role within VOC analysis by

Anthropogenic VOCs

61

chromatography. In recent years direct introduction mass spectrometry has also been used and this in many instances has increased the time resolution with which certain species may be measured compared to chromatographic techniques. Chromatography is one of the most common methods of separating complex mixtures into its constituent components, tracing its origins back to the early twentieth century when botanist Mikhail Tswett used a liquid–solid adsorption separation to resolve many coloured components within chlorophyll. This original application gave the technique its name from the Greek for colour or ‘Chroma’ and writing ‘graphein’. From the mid-twentieth century onwards, chromatography expanded rapidly in many directions: 1940, the development of partition chromatography (for which A.J.P. Martin and R.L.M. Synge were awarded the Nobel Prize); 1952, GC, 1970, bonded particles for liquid chromatography and in 1981 the pioneering of electrically driven separations, or electrophoresis (Bartle and Myers 2002). At the heart of all chromatographic techniques, however, is a common set of principles. The mixture of components is solvated in a ﬂuid (referred to as the mobile phase) which is used to transport the analytes through the separation system. The ﬂow of mobile phase passes over or through an immiscible bed of material or stationary phase. Interactions occur between analytes and stationary phase, and this slows the rate at which they pass through the bed. Each analyte may interact (e.g. by adsorption) to a different extent with the stationary phase, resulting in the physical separation of the mixture into its component parts. Chromatography covers a wide range of methods and techniques so it is useful to classify these into a number of key subsets. Perhaps the most basic description that can be made is of the physical state of the mobile phase: gas, liquid or supercritical. A second classiﬁcation describes the way in which the mobile phase interacts with the stationary phase. Where the stationary phase is either packed into or coated onto the walls of a length of tube it is referred to as column chromatography. The physical dimensions can vary considerably; VOC analysis can invoke columns as long as 100 m and as narrow as 50 μm whereas liquid chromatography takes place on columns of relatively modest length but up to several centimetres wide. By applying positive pressure at the start of the column, mobile phase can be driven through the column with a parabolic ﬂow proﬁle. Packed column chromatography involves the ﬂow of mobile phase through a column ﬁlled with a ﬁnely powdered stationary phase, in the order of micrometres in diameter, such as a molecular sieve or silica. In capillary column chromatography, the stationary phase, usually a highly viscous liquid, is coated onto the walls of the column, and the mobile phase ﬂows over this thin ﬁlm. In GC, the mobile phase, a highly diffuse gas, is used to solvate and transport components within a mixture through or over a stationary phase, with separation occurring due to differences in the rates of migration. The mobile phase gas must have a high diffusion coefﬁcient to ensure maximum numbers of gas–stationary phase interactions. The most common gases used in VOC GC are nitrogen, hydrogen and helium. Maximum separation efﬁciency is achieved using the lighter gases and as such nitrogen ﬁnds fewest applications. Highest separation efﬁciencies are achieved using hydrogen gas and this is over a wider linear velocity range than helium. However, the carrier gas of choice is usually helium due to the safety implications associated with hydrogen. High purity gases are used and it is normal to further purify the carrier gas with the use of a combination of oxygen, moisture and hydrocarbon traps.

62

Volatile Organic Compounds in the Atmosphere

The time a compound takes to be eluted from the column is known as the retention time and is used in the identiﬁcation of analytes through standard injections. In a correctly optimised chromatographic separation the analyte band will reach the detector in a narrow plug with a degree of band broadening to produce a Gaussian-shaped peak. The height or area under the peak is representative of the sample concentration giving a quantitative measurement. The overall resolving power of a chromatographic column can be given in terms of peak capacity. This is the maximum number of component peaks that can be theoretically resolved on a given column. In GC peak capacities are generally considerably larger than in liquid chromatography with typical values being 200 and 20, respectively. To completely resolve highly complex atmospheric samples would require extremely high peak capacities and selective detectors Therefore, extraction techniques and derivatisation are often used to simplify the separation required.

2.5.1

Sampling

The initial step of sample acquisition is far from being a trivial task in atmospheric measurements, and deﬁnes a major grouping of techniques into either in situ or post-acquisition analysis. The ability to store an atmospheric sample is a critical factor as to whether analysis may be performed back in the laboratory or on-site immediately following acquisition. Stable species such as methane, CO and CFCs may be successfully stored in certain sample vessels for several weeks without affecting sample integrity, and widespread measurements of these species have been performed. Each analyte however requires careful independent testing for stability, and a single storage material is not appropriate for all species – for example, CO has been shown to be problematic when stored for a longer term in stainless steel vessels. On the other hand, atmospheric organic degradation products, such as PAN (CH3 C(O)OONO2 ), are so unstable that analysis must be performed immediately. For many other important reactive species such as alkenes, monoterpenes and dimethyl sulphide (DMS), the storage of samples has been shown to lead to a degree of analyte losses due to reaction with co-sampled pollutants such as ozone and oxides of nitrogen (Helmig 1997). To overcome these problems of reaction during storage, many forms of scrubber (Helmig 1999; Sin et al. 2001; Vairavamurthy et al. 1992) have been tested to remove such species without affecting sample integrity. In other cases analytical interferences such as CO2 are removed during sampling. A great deal of VOC sampling is performed using stainless steel canisters of 1–10 l in volume (often referred to as whole air samples or ‘grab’ samples), ﬁlled either by vacuum release over a predetermined time via a ﬂow restrictor or by pressurising the sample into the canister using stainless steel bellows or Teﬂon diaphragm pumps. To retain sample integrity there must be minimal interaction with the canister walls and coating methods, such as electropolishing; silica or Teﬂon coating are currently in use. The preparation and cleaning of canisters requires careful attention and high vacuums are often applied along with elevated temperatures and humidiﬁed zero gases for periods of hours or days. Electropolished SUMMA-canisters are often used in campaigns either in air planes or on remote locations. Applications range from very short sampling intervals, for example, in air plane campaigns (Blake et al. 1996) to long-term sampling at remote sites (e.g. EMEP, Solberg et al. 1996).

Anthropogenic VOCs

63

Similar to canister samples is the use of collapsible Teﬂon or Tedlar sample bags. These may be ﬁlled by pump or external vacuum and returned to a laboratory for analysis using almost identical procedures to that of canister samples. While generally of lower unit cost, they often produce a greater degree of sample artefact and analyte losses. Many lower volatility species are unsuitable for collection using canister methods because of problems associated with analyte condensation onto the walls of the container. For this reason a second method of sample acquisition is widely used based on the use of a solid phase adsorbent (which is essentially a very high capacity packed column) as an analyte trap. The adsorbent used in the trap may be chosen to introduce an element of selectivity to the trapping mechanism, although in practice a trap-all approach is commonly used. Allowing air to enter an adsorbent trap purely via diffusion is referred to as passive sampling and is preferred for situations which do not require a very high temporal resolution, for example, for surveying industrial environments or workplace exposures. A huge range of adsorbent supports are commercially available (Camel and Caude 1995), ranging from high surface area (>1 000 m3 /g) carbon material (both charcoals and graphitised) with strong retention characteristics, to lower surface area (<50 m3 /g) polymerics such as Tenax TA. While being relatively low cost compared to sample canisters, care is often required in the cleaning and preparation of sample tubes. Samples may be introduced to the adsorbent tubes either dynamically over short periods of time (typically minutes) or via diffusional sampling over longer periods (typically several days). Carbon-based adsorbents are suitable for a wide range of species ranging from volatile hydrocarbons and CFCs to organic nitrates. Polymeric materials are used mainly for the concentration of lower volatility species such as aromatics and monoterpenes, although compounds as large as 2- and 3-ring polycyclic aromatic species may also be successfully trapped and thermally or solvent desorbed. Recycling of a single adsorbent allows for instrument automation, not possible with canister methods and such recycled traps now form the basis of many national monitoring programmes in the urban environment, as well as automated instruments for research in clean air environments (e.g. see http://www.epa.gov/ttn/amtic/airtoxpg.html). Sulphur species trapping is often performed via chemisorption onto gold wool traps that provide a stable matrix for the sample to be stored for reasonable periods of time (Berresheim et al. 1998). For a review of atmospheric sampling of organic compounds, see Rudolph et al. (1990b).

2.5.2

Removal of water

The inevitable presence of water in atmospheric samples and its removal prior to GC separation is a complex area. In certain circumstances its presence may be both beneﬁcial, for example, with canister samples where it occupies the active sites on the vessel walls, and detrimental, notably through affecting either the detector or reproducibility of the separation. Alumina porous layer open tubular (PLOT) columns are particularly affected by moisture in the sample, and large changes in stationary phase afﬁnity occur when water is introduced. Detectors such as mass spectrometers are also extremely sensitive to water introduced with the sample, and high background noise may result. Even the robust FID can be affected by injection of water, where the ﬂame may be extinguished when water elutes from the column.

64

Volatile Organic Compounds in the Atmosphere

Many forms of selective water removal exist, the simplest of which is the use of condensation traps or stripping coils. Losses of light molecular species are insigniﬁcant although condensation of higher boiling organic material may arise at very low temperatures. Inorganic adsorbents are also commonly used, notably potassium carbonate and magnesium perchlorate. Adsorbents such as these, however, have limited capacity and often require frequent regeneration or replacement. A combination of initial condensation and second stage adsorbent scrubber often provides sufﬁcient capacity to dry a sample stream of air for many hours or days. Continuous drying may be achieved using permeation membranes such as Naﬁon. These forms of dryers operate by generating a steep concentration gradient across a membrane permeable only to highly polar material such as water. A counter-current of dry gas is passed around the outside of the membrane as a sheath gas and carries moisture away to waste. However, this type of drier is unsuitable for samples where the quantitation of polar materials is required.

2.5.3

Sample pre-concentration

Whether an atmospheric sample is collected using adsorbent or canister techniques, several stages of sample preparation are required prior to injection to the analytical column, since the concentration of VOCs, even in polluted environments, is too low to allow for direct loop injection on column. For volatile species collected in canister or whole air samples, analytes are removed from the canister via either internal canister pressure (for pumped in samples) or by vacuum pump (for atmospheric pressure samples) over a pre-concentration trap. A simple method of pre-concentration is to place a sampling loop either directly into or in the headspace of liquid nitrogen (Kohno and Kuwata 1991). Liquid argon has also been used since it may reduce the amount of oxygen retained in the refocusing stage. The pre-concentration zone may consist of a packed tube containing either an absorbent such as Tenax TA, glass beads or may simply be empty stainless steel tubing. Since the pre-concentration stage is at such low temperatures the majority of water vapour in the sample must be removed prior to refocusing in order to stop blockage of lines with ice. Once a sufﬁcient volume has been collected on the refocusing trap, the trap is generally ﬂash heated either electrically or using hot water. This results in a very sharp band of compounds being introduced to the head of the analytical column. Cryogenic methods with careful engineering can be automated successfully and used to measure at low atmospheric concentrations (Rudolph et al. 1990a). With adsorbent tube analysis the collected analytes are generally thermally desorbed either directly onto the analytical column (in the case of programmed temperature vaporisation (PTV) injection), or onto a refocusing cold trap. The desorption temperature is generally deﬁned by the maximum temperature that the adsorbing material can support. For polymeric adsorbents this may be relatively low (<250◦ C) while carbon-based materials may support desorption at temperatures of 500◦ C or higher. If a PTV injector is used, desorption may be sufﬁciently rapid (>16◦ C/s) that initial phase ratio refocusing on-column is sufﬁcient to obtain well-resolved peaks (Lewis et al. 1996). If the desorption from an adsorbent tube is relatively slow then a refocusing step is used, with a similar pre-concentration mechanism to that used with canister samples.

Anthropogenic VOCs

65

Once again, water must be removed from the sample since it may affect the column or the detection system. Automated systems where a single recycled adsorbent trap is used may operate by taking air from either a local manifold or from canister collected samples. The addition of a multiposition valve upstream of the instrument can allow for fully automated canister analysis on systems designed for in situ analysis. Fully automated instruments of this kind using either cryogen or Peltier cooled adsorbent traps are commercially available (e.g. Perkin Elmer, Markes International). While the majority of species are thermally desorbed from adsorbent traps onto the analytical column a few types of compounds require solvent extraction prior to syringe injection. The analysis of some organic nitrates has been described in this way, along with higher molecular weight polycyclic aromatic compounds that may suffer from incomplete or slow release. Derivatisation can be used to selectively convert species from collected atmospheric samples to analytes more suited to GC (Cecinato et al. 2001). An example of post-acquisition derivatisation is the analysis of atmospheric alcohols, particularly of interest in Brazil where ethanol-fuelled cars are used on a large scale (Nguyen et al. 2001). Alcohols are highly polar and can stick to canisters walls or be irreversibly trapped onto absorbents. A novel technique for the analysis of alcohols involved collection of air samples in Pyrex vessels followed by subsequent reaction of alcohols with nitrogen dioxide to form alkyl nitrates. Analysis of the adducts by gas chromatography electron capture detector (GC-ECD) gives detection limits of around 1 ppb for methanol and ethanol.

2.5.4

Separation of VOCs

While a number of speciﬁc applications utilise packed columns (notably in the analysis of methane, CO and N2 O), current methods for the separation of VOC components are performed almost exclusively using capillary column GC. For a general review of the analytical science of GC, see for example Grob and Barry (2004) or chapters within general analytical science textbooks such as Skoog et al. (1998). With the introduction of Al2 O3 PLOT columns, where stationary phase is bonded to the column wall, high resolution analysis of very high volatility species is possible and many applications that historically used packed column GC are now performed using this type of hybrid capillary columns. The wide range of VOC volatilities that are encountered in the atmosphere conﬁnes each analytical system to only a limited range of species that may be completely resolved on a single column. For the most volatile NMHCs, PLOT columns are used widely for species in −C7 (Habram et al. 1998). This type of gas–solid chromatography has the carbon range C1 − a thin porous layer of a ﬁnely divided solid, usually alumina or molecular sieve, deposited on the walls of the tube. Separations on columns of this kind are via adsorption rather than phase partition, the kinetics of which are particularly rapid. As a result theoretical plate numbers in excess of 100 000 are common even for wide bore 0.53 mm internal diameter (i.d.) columns. While in principle suitable for high volatility halogenated compounds hydrogenation and dehydrogenation effects have been reported for such species on PLOT columns. The retention characteristics of PLOT-type columns are unfavourable for oxygenated compounds, where extremely strong retention, often irreversible, can occur. Due

66

Volatile Organic Compounds in the Atmosphere

to the strong retention of polar species, water becomes an important interference and its presence can severely degrade the quality of PLOT separations, manifested in highly variable retention times. The very strong retention of higher boiling point species on PLOT-type columns leads to extensive peak broadening and very lengthy analysis times. Due to this, the analysis of higher molecular weight species including mono aromatics (Hamilton and Lewis 2003), VOCs, CFCs (Frank et al. 1991), HCFCs (hydrogen containing CFC replacements) and terpenoid compounds is generally performed using wall-coated open tubular (WCOT) columns (Ciccioli et al. 1992; Street et al. 1997), where the stationary phase is generally a viscous liquid such as non-polar methylpolysiloxane or slightly polar 5% phenyl-methylpolysiloxane. Typical column speciﬁcations are 0.32 mm (i.d.) × 50 m with ﬁlm thickness of between 1 and 5 μm, and a temperature gradient is used to reduce band broadening in later peaks. Wide bore 0.53 mm (i.d.) columns are also used where direct thermal desorption from a pre-concentration trap to the analytical column is applied. The rate of generation of theoretical plates (and hence peak capacity) on columns of this type are lower than for PLOT types, and as a result to obtain full resolution of some species (e.g. HCFC mixtures in the atmosphere), columns as long as 100 m have been reported. This technique is the most −C12 ) and semi-volatile organic comwidespread method of analysis of mid-range (∼C4 − pounds in the atmosphere. Columns are highly durable and can be used in both ﬁeld and laboratory measurements. To improve the retention and separation of some volatile VOCs (those that fall between standard wall-coated siloxane columns and PLOT columns) without use of sub-ambient cooling, phase thicknesses up to 15 μm have been reported. Band broadening effects through stationary phase diffusion become signiﬁcant with ﬁlms of this thickness and this approach has not been widely adopted. Gas phase species with high molecular weights such as naphthalene, ﬂuorene and anthracene may be separated efﬁciently on non-polar columns with ﬁlm thicknesses of typically 0.25–0.5 μm. Since MS is commonly the detector, very low bleed columns are desirable, although this is difﬁcult to achieve, given the ﬁlm thicknesses typically employed. New stationary phases have been developed to extend the range of compounds amenable to capillary GC analysis. Adding polar functionalities to the methylpolysiloxane backbone can create stationary phases of different polarities and selectivities. For example, 14% cyanopropylphenyl 86% methylpolysiloxane columns are moderately polar and have been used in the analysis of hydrocarbons, oxygenates and CFCs. In addition, a number of speciality phases are available that allow the separation of oxygenated material by capillary GC. Since such species are generally at low concentration in the atmosphere, they often suffer from co-elution with more abundant primary emitted VOCs. While derivatisation followed by HPLC is still a most common method of carbonyl analysis, the use of mixed phase porous layer capillary columns is an emerging technique. An example is the Varian CP-LOWOX column, which uses a multi-layer column coating process to produce an extremely polar column with high stability. Sample pre-concentration can be via standard carbon adsorbent methods and since no derivatisation stages are required, minimum detectable amounts are greatly improved and sampling volumes drastically reduced. Highly polar wall-coated stationary phases have also been developed using polyethylene glycol (WAX) and can be used to separate oxygenated analytes such as alcohols, free acids and esters.

Anthropogenic VOCs

67

GC separations are in general terms performed over timescales of perhaps one sample per hour, but that is not necessarily the limit of performance that can be achieved. Using narrow bore columns and fast temperature ramps (i.e. 40◦ C/min), commonly available on modern GC instrumentation, it is possible to perform much faster GC separations. If sampling, which is often a rate-determining step in the analytical procedure, can be undertaken in parallel, there are no great barriers in reducing cycle time to perhaps 5 or 10 min. This has been used to develop GC-based instruments which can be used from research aircraft (Singh et al. 2004; Whalley et al. 2004).

2.5.5

Detection

As highlighted in the previous section, even the highest resolution capillary column often has insufﬁcient peak capacity to resolve all VOCs in a typical atmospheric sample. Since the introduction of analyte selectivity in the trapping and preparation stages is not always possible, selectivity in detection is a very useful tool for simplifying atmospheric samples. However, it is also often convenient to use a simple detector that gives a universal or groupspeciﬁc response. For an overview of possible detectors for GC, see Hill and McMinn (1992). The FID is in general terms by far the most commonly used detector in GC, since it offers high sensitivity, extremely wide linearity and very good long-term reliability and response. Using well-cleaned fuel gases coupled to low noise electrometer circuitry, it is possible to determine amounts down to as low 1 pg/s of eluting peak. With a typical sample volume of 1 l, detection limits for individual species may therefore be in the low ppt range (10−12 ). Calibration can be performed with relative ease (and in some cases calculated from standard response characteristics) but the complexity of samples can make peak identiﬁcation difﬁcult when co-elutions occur. To overcome this lack of selectivity, analytical methods for alkene and aromatic analysis using a selective response from photoionisation detection (PID) and the reduction gas detector (RGD) have been proposed, although they are not as widespread as FID techniques. For detectors such as the PID, it is the need to regularly calibrate for decreasing bulb/ionisation intensity which accounts for its limited usage. The ECD was speciﬁcally designed for organohalogen measurements in the atmosphere and appears in many published methods in the literature. It offers high sensitivity to electrophilic compounds with almost no response for hydrocarbon species. GC–ECD measurements require careful calibration due to the great variation in response to individual halogenated species, although their high stability allows gas standards to be used over many years. Some halogen-containing species of atmospheric interest (e.g. CH3 Cl, CHF2 Cl, CH2 Cl2 ) have a relatively poor ECD response and the use of the oxygen-doped ECD to enhance their response has been successful. Although MS detection is now the most widespread detection method for these compounds, the ECD still ﬁnds usage (Rivett et al. 2003; Urhahn and Ballschmiter 2000). Selective detection of oxygenated VOCs is difﬁcult, due to low response in both ECD and FID detectors, and low molecular weight fragmentation in MS. Elemental speciﬁc detection such as AED offers some potential in oxygenate analysis although sensitivity is poor at around 100 pg/s. Detectors such as the helium ionisation detector (HID), produce a non-selective, high sensitivity response to these types of compounds, but require full

68

Volatile Organic Compounds in the Atmosphere

chromatographic resolution since they also respond to every other type of VOC (Hunter et al. 1999). Mass spectrometry offers obvious additional analytical resolution, and bench-top GC–MS technology is now at a very advanced stage in terms of reliability, self-calibration, tuning and automation. The eluent from the GC column is transferred to the ion source directly via a gas tight heated transfer line making this a simple and convenient hyphenated technique, without the mobile phase evaporation problems associated with LC-MS. Ions can be produced using a number of different ionisation methods, the most popular being electron (EI) and chemical ionisation (CI). For an overview of MS techniques commonly found in use with atmospheric applications, see for example Hoffmann and Stroobant (2002), although very many other texts are available. In EI, an electron beam, operated at about 70 eV, is ﬁred at molecules as they enter the ion source. The high-energy beam transfers large amounts of energy to molecules, which can break up into smaller ions, releasing excess energy. Fragmentation patterns can give important information about the structure of a molecule, an example being a series of m/z ions that are spaced 14 amu apart, indicating the presence of a hydrocarbon chain. However, the spectral information obtained from GC–MS of many VOC species (in particular hydrocarbons-based compounds) often leads to highly similar fragmentation patterns and assists little in the identiﬁcation of isomeric species. Similarly, identiﬁcation of monoterpene species can only be conﬁrmed through a combination of both spectral information and retention time data. It is often advantageous to obtain the molecular ion to facilitate identiﬁcation. In some cases, EI completely fragments the molecule and the molecular ion is absent from the chromatogram. CI produces ions using an ionised reagent gas (commonly methane, isobutane or ammonia) to transfer protons to (or from) the analyte resulting in the formation of a pseudo-molecular ion. In the positive ion mode the reagent gas donates a proton to the analyte to form [M + H]+ and in the negative mode it abstracts a proton to form [M − H]− . This is a softer form of ionisation and produces less fragmentation in the ion source. Ions are accelerated by the accelerating plate into the mass analyser, which separates ions according to their mass-to-charge ratio (m/z). The quadrupole mass spectrometer has now become one of the most widely used mass spectrometers because of its ease of use, small size and relatively low cost. Mass separation in a quadrupole mass ﬁlter is based on achieving a stable trajectory for ions of speciﬁc m/z values in a hyperbolic electrostatic ﬁeld. Operating currently available quadrupole mass spectrometers in full scan mode has only moderate sensitivity for hydrocarbon-like species perhaps to the 30–50 ppt level. Where GC–MS is particularly strong is in the measurement of halogenated species in the atmosphere (Oram et al. 1996). While hydrocarbon fragmentation is often very similar with little abundant parent ion, many halogenated compounds give highly unique MS fragmentation with abundant large m/z fragments. Long-term measurements of species such as CFCs and their replacements has been performed by GC–MS instruments for many years within the world-wide AGAGE network (Prinn et al. 2000).

2.5.6

Calibration techniques and quality assurance

The quality of data required in any analytical system is of great importance, particularly if any conclusions are to be drawn from the data, such as air quality statistics and model

Anthropogenic VOCs

69

simulations. In GC systems it is normal to use a number of quality control methods, but the most important within this set is the use of calibration standards. A Retention Time Standard, which is a qualitative mix of known components (such as hydrocarbons), is used initially to determine the chromatographic peak retention time and if MS is present, the instrument speciﬁc mass spectra. It can be used during instrument optimisation to ensure maximum peak separation. In addition, it can also be used during routine analysis to ensure there is no retention time shift, which could be due to factors such as changes in water content and ﬂow blockages. A primary calibration standard is used to determine the response of the GC system to speciﬁc analytes and therefore allows quantiﬁcation. This primary calibration standard is generally a gas standard, made up in an inert matrix such as zero air, or nitrogen, containing the compounds of interest at known and typical concentrations. Standard Reference Materials (SRMs) and Certiﬁed Reference Materials (CRMs) are available for some VOCs from speciality gas suppliers and are referenced to a national standard such as the National Institute of Standards and Technology (NIST) in the United States or the National Physical Laboratory (NPL) in the United Kingdom. Compressed gas reference materials such as these are generally prepared via a multiple gas dilutions of a high concentration mother standard, whose VOC content is known accurately from gravimetrical measurement. At each dilution step accuracy may be degraded and there is a tendency for such synthetic standards to be relatively high compared to ambient VOC concentrations. Even though calibration standards with concentrations approaching those of real air samples are preferable, because they reduce any errors associated with dilution during sampling of the standard or non-linearities in instrument response. Gas standards should also ideally be humidiﬁed to ambient levels to ensure there are no water-dependent effects on sampling and response. Given the requirements highlighted above, an ideal working standard is the use of ambient compressed air itself. This works well for stable species such as halocarbons (Prinn et al. 2000), but is more problematic for reactive hydrocarbons. The frequency of calibration is dependent on application and detector. An FID is considerably more stable over time than a quadrupole mass spectrometer, the former requiring a check on calibration perhaps every couple of days, the latter every couple of samples. The intercomparison of GC instruments around the world is also an important aspect of quality control. The AGAGE instruments are calibrated on-site using alternate analysis of ambient air and a 35 l standard tank, with standard tanks lasting up to ten months and then calibrated relative to a central laboratory (currently the Scripps Institution of Oceanography (SIO) California, USA). For a review of calibration methods used at GAW sites, see Plass-Dulmer et al. (2002). A number of international intercomparison exercises have been carried out for the observations of VOCs. The Accurate Measurement of Hydrocarbons in the Atmosphere −C9 (AMOHA) project was developed to evaluate current GC methods of analysis of C2 − NMHCs across Europe (Slemr et al. 2002b). Calibration gas standards were prepared by the UK NPL for all laboratories, to reduce variability associated with the standard, and analysis was carried out in 12–14 laboratories over 4 years. Agreement was generally good for high concentration standard mixtures but disagreement increased as lower concentration, real air samples were analysed. However, some improvements were achieved over the course of the project. Broadly similar ﬁnding arose from the U.S. led NOMHICE intercomparison project (Apel et al. 2003). A number of intercomparisons between different

70

Volatile Organic Compounds in the Atmosphere

analytical measurements including GC-FID/ECD/MS, proton transfer reaction mass spectrometry (PTR-MS) and open path FTIR (OP-FTIR) have been made (Christian et al. 2004). Good agreement between GC and OP-FTIR and between GC and PTR-MS was achieved for a number of compounds. However, differences were found for sticky compounds (e.g. formic acid) and compounds with a low proton afﬁnity (e.g. formaldehyde) with OP-FTIR and PTR-MS having considerable advantages. It is important for real samples to take into account the great difference in time resolution between these methods. In the future, however, it seems certain that validation of techniques via intercomparisons of standards will increase.

References Abu-Allaban, M., Gertler, A.W. and Lowenthal, D.H. (2002) A preliminary apportionment of the sources of ambient PM10, PM2.5, and VOCs in Cairo. Atmospheric Environment, 36: 5549–57. ACESS (2005) Asia and trace-P modelling and emission support system. http://www.cgrer.uiowa.edu/ ACESS/acess_index.htm Altshuller, A.P., Lonneman, W.A., Sutterﬁeld, F.D. and Kopczynski, S.L. (1971) Hydrocarbon composition of atmosphere of Los-Angeles basin – 1967. Environmental Science and Technology, 5: 1009–16. Andersson-Skold, Y. and Holmberg, L. (2000) Photochemical ozone creation potentials (POCP) and replacement of solvents in Europe. Atmospheric Environment, 34: 3159–69. Andreae, M.O. and Merlet, P. (2001) Emission of trace gases and aerosols from biomass burning. Global Biogeochemical Cycles, 15: 955–66. Apel, E.C., Calvert, J.G., Gilpin, T.M., Fehsenfeld, F. and Lonneman, W.A. (15 May 2003) Nonmethane Hydrocarbon Intercomparison Experiment (NOMHICE): Task 4, ambient air. Journal of Geophysical Research – Atmospheres, 108(D9): 4300. Arriaga-Colina, J.L., West, J.J., Sosa, G., Escalona, S.S., Ordunez, R.M. and Cervantes, A.D.M. (2004) Measurements of VOCs in Mexico City (1992–2001) and evaluation of VOCs and CO in the emissions inventory. Atmospheric Environment, 38: 2523–33. Atkinson, R. and Arey, J. (2003) Atmospheric degradation of volatile organic compounds. Chemical Reviews, 103: 4605–38. Bakeas, E.B., Argyris, D.I. and Siskos, P.A. (2003) Carbonyl compounds in the urban environment of Athens, Greece. Chemosphere, 52: 805–13. Barletta, B., Meinardi, S., Simpson, I.J., Khwaja, H.A., Blake, D.R. and Rowland, F.S. (2002) Mixing ratios of volatile organic compounds (VOCs) in the atmosphere of Karachi, Pakistan. Atmospheric Environment, 36: 3429–43. Barletta, B., Meinardi, S., Sherwood Rowland, F., et al. (2005) Volatile organic compounds in 43 Chinese cities. Atmospheric Environment, 39: 5979–90. Barrefors, G. and Petersson, G. (1995) Volatile hydrocarbons from domestic wood burning. Chemosphere, 30: 1551–6. Bartle, K.D. and Myers, P. (2002) History of gas chromatography. Trac-Trends in Analytical Chemistry, 21: 547–57. Baugh, J., Ray, W., Black, F. and Snow, R. (1987) Motor-vehicle emissions under reduced ambienttemperature idle operating-conditions. Atmospheric Environment, 21: 2077–82. Berkowitz, C.M., Jobson, T., Jiang, G.F., Spicer, C.W. and Doskey, P.V. (2004) Chemical and meteorological characteristics associated with rapid increases of O-3 in Houston, Texas. Journal of Geophysical Research – Atmospheres, 109: D10307.

Anthropogenic VOCs

71

Berresheim, H., Huey, J.W., Thorn, R.P., Eisele, F.L., Tanner, D.J. and Jefferson, A. (1998) Measurements of dimethyl sulﬁde, dimethyl sulfoxide, dimethyl sulfone, and aerosol ions at Palmer Station, Antarctica. Journal of Geophysical Research – Atmospheres, 103: 1629–37. Blake, D.R., Chen, T.Y., Smith, T.Y., et al. (1996) Three-dimensional distribution of nonmenthane hydrocarbons and halocarbons over the northwestern Paciﬁc during the 1991 Paciﬁc Exploratory Mission (PEM-West A). Journal of Geophysical Research – Atmospheres, 101: 1763–78. Blake, N.J., Blake, D.R., Chen, T.Y., et al. (1997) Distribution and seasonality of selected hydrocarbons and halocarbons over the western Paciﬁc basin during PEM-West A and PEM-West B. Journal of Geophysical Research – Atmospheres, 102: 28315–31. Blake, N.J., Blake, D.R., Simpson, I.J., et al. (2001) Large-scale latitudinal and vertical distributions of NMHCs and selected halocarbons in the troposphere over the Paciﬁc Ocean during the March–April 1999 Paciﬁc Exploratory Mission (PEM- Tropics B). Journal of Geophysical Research – Atmospheres, 106: 32627–44. Blake, N.J., Blake, D.R., Simpson, I.J., et al. (2003a) NMHCs and halocarbons in Asian continental outﬂow during the Transport and Chemical Evolution over the Paciﬁc (TRACE-P) ﬁeld campaign: Comparison with PEM-West B. Journal of Geophysical Research – Atmospheres, 108(D20): 8806. Blake, N.J., Blake, D.R., Sive, B.C., et al. (2003b) The seasonal evolution of NMHCs and light alkyl nitrates at middle to high northern latitudes during TOPSE. Journal of Geophysical Research – Atmospheres, 108(D4): 8359. Boissard, C., Bonsang, B., Kanakidou, M. and Lambert, G. (1996) TROPOZ II: Global distributions and budgets of methane and light hydrocarbons. Journal of Atmospheric Chemistry, 25: 115–48. Borbon, A., Fontaine, H., Veillerot, M., Locoge, N., Galloo, J.C. and Guillermo, R. (2001) An investigation into the trafﬁc-related fraction of isoprene at an urban location. Atmospheric Environment, 35: 3749–60. Borbon, A., Fontaine, H., Locoge, N., Veillerot, M. and Galloo, J.C. (2003a) Developing receptororiented methods for non-methane hydrocarbon characterisation in urban air – Part I: Source identiﬁcation. Atmospheric Environment, 37: 4051–64. Borbon, A., Fontaine, H., Locoge, N., Veillerot, M. and Galloo, J.C. (2003b) Developing receptororiented methods for non-methane hydrocarbon characterisation in urban air. Part II: Source apportionment. Atmospheric Environment, 37: 4065–76. Borbon, A., Coddeville, P., Locoge, N. and Galloo, J.C. (2004) Characterising sources and sinks of rural VOC in eastern France. Chemosphere, 57: 931–42. Boudries, H., Bottenheim, J.W., Guimbaud, C., et al. (2002) Distribution and trends of oxygenated hydrocarbons in the high Arctic derived from measurements in the atmospheric boundary layer and interstitial snow air during the ALERT2000 ﬁeld campaign. Atmospheric Environment, 36: 2573–83. Brocco, D., Fratarcangeli, R., Lepore, L., Petricca, M. and Ventrone, I. (1997) Determination of aromatic hydrocarbons in urban air of Rome. Atmospheric Environment, 31: 557–66. Bruehlmann, S., Forss, A.M., Steffen, D. and Heeb, N.V. (2005) Benzene: A secondary pollutant formed in the three-way catalyst. Environmental Science and Technology, 39: 331–8. Buhr, M.P., Trainer, M., Parrish, D.D., Sievers, R.E. and Fehsenfeld, F.C. (1992) Assessment of pollutant emission inventories by principal component analysis of ambient air measurements. Geophysical Research Letters, 19: 1009–12. Calvert, J.G., Atkinson, R., Kerr, J.A., et al. (2000) The Mechanisms of Atmospheric Oxidation of the Alkenes. Oxford University Press, New York. Camel, V. and Caude, M. (1995) Trace enrichment methods for the determination of organic pollutants in ambient air. Journal of Chromatography A, 710: 3–19. CARB (2005) http://www.arb.ca.gov/adam/toxics/statesubstance.html Carter, W.P.L. (1994) Development of ozone reactivity scales for volatile organic-compounds. Journal of the Air and Waste Management Association, 44: 881–99.

72

Volatile Organic Compounds in the Atmosphere

Carter, W.P.L., Pierce, J.A., Luo, D.M. and Malkina, I.L. (1995) Environmental chamber study of maximum incremental reactivities of volatile organic-compounds. Atmospheric Environment, 29: 2499–511. Cecinato, A., Di Palo, V., Mabilia, R. and Possanzini, M. (2001) Pentaﬂuorophenylhydrazine as a coating reagent for the HRGC-MS determination of semi volatile carbonyl compounds in air. Chromatographia, 54: 263–9. Cecinato, A., Yassaa, N., Di Palo, V. and Possanzini, M. (2002) Observation of volatile and semi-volatile carbonyls in an Algerian urban environment using dinitrophenylhydrazine/silicaHPLC and pentaﬂuorophenylhydrazine/silica-GC-MS. Journal of Environmental Monitoring, 4: 223–8. Christian, T.J., Kleiss, B., Yokelson, R.J., et al. (2004) Comprehensive laboratory measurements of biomass-burning emissions: 2. First intercomparison of open-path FTIR, PTR-MS, and GC-MS/ FID/ECD. Journal of Geophysical Research – Atmospheres, 109: D02311. Ciccioli, P., Cecinato, A., Brancaleoni, E., Frattoni, M. and Liberti, A. (1992) Use of carbon adsorption traps combined with high-resolution gas-chromatography – mass-spectrometry for the analysis of polar and nonpolar C-4–C-14 hydrocarbons involved in photochemical smog formation. Hrc-Journal of High Resolution Chromatography, 15: 75–84. Derwent, R.G. and Jenkin, M.E. (1991) Hydrocarbons and the long-range transport of ozone and pan across Europe. Atmospheric Environment Part a – General Topics 25: 1661–78. Derwent, R.G., Middleton, D.R., Field, R.A., Goldstone, M.E., Lester, J.N. and Perry, R. (1995) Analysis and interpretation of air-quality data from an urban roadside location in central London over the period from July 1991 to July 1992. Atmospheric Environment, 29: 923–46. Derwent, R.G., Jenkin, M.E., Saunders, S.M. and Pilling, M.J. (1998) Photochemical ozone creation potentials for organic compounds in northwest Europe calculated with a master chemical mechanism. Atmospheric Environment, 32: 2429–41. Derwent, R.G., Davies, T.J., Delaney, M., et al. (2000) Analysis and interpretation of the continuous hourly monitoring data for 26 C-2–C-8 hydrocarbons at 12 United Kingdom sites during 1996. Atmospheric Environment, 34: 297–312. Dommen, J., Prevot, A.S.H., Neininger, B. and Baumle, M. (2002) Characterization of the photooxidant formation in the metropolitan area of Milan from aircraft measurements. Journal of Geophysical Research – Atmospheres, 107(D22): 8197. Dommen, J., Prevot, A.S.H., Baertsch-Ritter, N., et al. (2003) High-resolution emission inventory of the Lombardy region: Development and comparison with measurements. Atmospheric Environment, 37: 4149–61. Donahue, N.M. and Prinn, R.G. (1993) In-situ nonmethane hydrocarbon measurements on Saga-3. Journal of Geophysical Research – Atmospheres, 98: 16915–32. Doskey, P.V., Fukui, Y., Sultan, M., Al Maghraby, A. and Taher, A. (1999) Source proﬁles for nonmethane organic compounds in the atmosphere of Cairo, Egypt. Journal of the Air and Waste Management Association, 49: 814–22. Duane, M., Poma, B., Rembges, D., Astorga, C. and Larsen, B.R. (2002) Isoprene and its degradation products as strong ozone precursors in Insubria, Northern Italy. Atmospheric Environment, 36: 3867–79. Duffy, B.L. and Nelson, P.F. (1996) Non-methane exhaust composition in the Sydney Harbour Tunnel: A focus on benzene and 1,3-butadiene. Atmospheric Environment 30: 2759–68. EDGAR (2005) EDGAR 32FT2000 dataset. http://www.mnp.nl/edgar/model/v32ft2000edgar/ Eggertsen, F.T. and Nelsen, F.M. (1958) Gas chromatographic analysis of engine exhaust and atmosphere – determination of C-2 to C-5 hydrocarbons. Analytical Chemistry, 30: 1040–3. Ehhalt, D.H., Rohrer, F., Wahner, A., Prather, M.J. and Blake, D.R. (1998) On the use of hydrocarbons for the determination of tropospheric OH concentrations. Journal of Geophysical Research – Atmospheres, 103: 18981–97.

Anthropogenic VOCs

73

Elliott, S., Shen, M., Blake, D.R., et al. (1997) Atmospheric effects of the emerging mainland Chinese transportation system at and beyond the regional scale. Journal of Atmospheric Chemistry, 27: 31–70. EMEP (2005) http://webdab.emep.int Emmons, L.K., Hauglustaine, D.A., Muller, J.F., et al. (2000) Data composites of airborne observations of tropospheric ozone and its precursors. Journal of Geophysical Research – Atmospheres, 105: 20497–538. Farrington, P.S., Pecsok, R.L., Meeker, R.L. and Olson, T.J. (1959) Detection of trace constituents by gas chromatography – analysis of polluted atmosphere. Analytical Chemistry, 31: 1512–16. Fischer, H., De Reus, M., Traub, M., et al. (2003) Deep convective injection of boundary layer air into the lowermost stratosphere at midlatitudes. Atmospheric Chemistry and Physics, 3: 739–45. Flocas, H.A., Assimakopoulos, V.D., Helmis, C.G. and Gusten, H. (2003) VOC and O-3 distributions over the densely populated area of greater Athens, Greece. Journal of Applied Meteorology, 42: 1799–810. Frank, H., Frank, W., Neves, H.J.C. and Englert, R. (1991) Automated trace analysis of airborne C1-halocarbons and C2-halocarbons. Fresenius Journal of Analytical Chemistry, 340: 678–83. Friedrich, R., Wickert, B., Blank, P., et al. (2002) Development of emission models and improvement of emission data for Germany. Journal of Atmospheric Chemistry, 42: 179–206. Fu, L. (2001) Assessment of vehicular pollution in China. Journal of the Air and Waste Management Association, 51: 658–68. Fujita, E.M. (2001) Hydrocarbon source apportionment for the 1996 Paso del Norte Ozone Study. Science of the Total Environment, 276: 171–84. Gasca, J., Ortiz, E., Castillo, H., Jaimes, J.L. and Gonzalez, U. (2004) The impact of liqueﬁed petroleum gas usage on air quality in Mexico City. Atmospheric Environment, 38: 3517–27. Gertler, A.W., Fujita, E.M., Pierson, W.R., et al. (1996) Apportionment of NMHC tailpipe vs non-tailpipe emissions in the Fort McHenry and Tuscarora Mountain Tunnels. Atmospheric Environment, 30: 2297–2305. Geyer, A., Alicke, B., Konrad, S., Schmitz, T., Stutz, J. and Platt, U. (2001) Chemistry and oxidation capacity of the nitrate radical in the continental boundary layer near Berlin. Journal of Geophysical Research, 106: 8013–25. Goldan, P.D., Kuster, W.C., Williams, E., Murphy, P.C., Fehsenfeld, F.C. and Meagher, J. (2004) Nonmethane hydrocarbon and oxy hydrocarbon measurements during the 2002 New England Air Quality Study. Journal of Geophysical Research – Atmospheres, 109: D21309. Goldstein, A.H. and Schade, G.W. (2000) Quantifying biogenic and anthropogenic contributions to acetone mixing ratios in a rural environment. Atmospheric Environment, 34: 4997–5006. Goldstein, A.H. and Shaw, S.L. (2003) Isotopes of volatile organic compounds: An emerging approach for studying atmospheric budgets and chemistry. Chemical Reviews, 103: 5025–48. Gordon, R.J., Mayrsohn, H. and Ingels, R.M. (1968) C2 –C5 hydrocarbons in the Los Angeles atmosphere. Environment, Science and Technology, 2: 1117–20. Grob, K. and Grob, G. (1971) Gas–liquid chromatographic-mass spectrometric investigation of C6 –C20 organic compounds in an urban atmosphere – application of ultra trace analysis on capillary columns. Journal of Chromatography, 62: 1–13. Grob, R.L. and Barry, E.F. (2004) Modern Practice of Gas Chromatography (4th edn). John Wiley & Sons, Hoboken, NJ. Gros, V., Martin, D., Poisson, N., et al. (1998) Ozone and C-2–C-5 hydrocarbon observations in the marine boundary layer between 45 degrees S and 77 degrees S. Tellus Series B-Chemical and Physical Meteorology, 50: 430–48. Gros, V., Williams, J., Lawrence, M.G., et al. (2004) Tracing the origin and ages of interlaced atmospheric pollution events over the tropical Atlantic Ocean with in situ measurements,

74

Volatile Organic Compounds in the Atmosphere

satellites, trajectories, emission inventories, and global models. Journal of Geophysical Research – Atmospheres, 109: D22306. Grosjean, D., Grosjean, E. and Moreira, L.F.R. (2002) Speciated ambient carbonyls in Rio de Janeiro, Brazil. Environmental Science and Technology, 36: 1389–95. Grosjean, E., Rasmussen, R.A. and Grosjean, D. (1998) Ambient levels of gas phase pollutants in Porto Alegre, Brazil. Atmospheric Environment, 32: 3371–9. Guenther, A.B., Hewitt, C.N., Erickson, D., et al. (1995) A global model of natural volatile organic compound emissions. Journal of Geophysical Research, 100: 8873–92. Guo, H., Lee, S.C., Louie, P.K.K. and Ho, K.F. (2004a) Characterization of hydrocarbons, halocarbons and carbonyls in the atmosphere of Hong Kong. Chemosphere, 57: 1363–72. Guo, H., Wang, T. and Louie, P.K.K. (2004b) Source apportionment of ambient nonmethane hydrocarbons in Hong Kong: Application of a principal component analysis/absolute principal component scores (PCA/APCS) receptor model. Environmental Pollution, 129: 489–98. Guo, H., Wang, T., Simpson, I.J., et al. (2004c) Source contributions to ambient VOCs and CO at a rural site in Eastern China. Atmospheric Environment, 38: 4551–60. Gupta, M.L., Cicerone, R.J., Blake, D.R., Rowland, F.S. and Isaksen, I.S.A. (1998) Global atmospheric distributions and source strengths of light hydrocarbons and tetrachloroethene. Journal of Geophysical Research – Atmospheres, 103: 28219–35. Gurjar, B.R. and Lelieveld, J. (2005) New directions: Megacities and global change. Atmospheric Environment, 39: 391–3. Gurjar, B.R., Van Aardenne, J.A., Lelieveld, J. and Mohan, M. (2004) Emission estimates and trends (1990–2000) for megacity Delhi and implications. Atmospheric Environment, 38: 5663–81. Gusten, H., Heinrich, G., Cvitas, T., et al. (1988) Photochemical formation and transport of ozone in Athens, Greece. Atmospheric Environment, 22: 1855–61. Haagen-Smit, A.J. (1952) Chemistry and physiology of Los-Angeles smog. Industrial and Engineering Chemistry, 44: 1342–6. Habram, M., Slemr, J. and Welsch, T. (1998) Development of a dual capillary column GC method for the trace determination of C2 –C9 hydrocarbons in ambient air. Hrc-Journal of High Resolution Chromatography, 21: 209–14. Hamilton, J.F. and Lewis, A.C. (2003) Monoaromatic complexity in urban air and gasoline assessed using comprehensive GC and fast GC-TOF/MS. Atmospheric Environment, 37: 589–602. Hamilton, J.F., Lewis, A.C., Bloss, C., et al. (2003) Measurements of photo-oxidation products from the reaction of a series of alkyl-benzenes with hydroxyl radicals during EXACT using comprehensive gas chromatography. Atmospheric Chemistry and Physics, 3: 1999–2014. Harley, R.A., Hannigan, M.P. and Cass, G.R. (1992) Respeciation of organic gas emissions and the detection of excess unburned gasoline in the atmosphere. Environmental Science and Technology, 26: 2395–408. Haszpra, L. and Szilagyi, I. (1994) Nonmethane hydrocarbon composition of car exhaust in Hungary. Atmospheric Environment, 28, 2609–14. Heald, C.L., Jacob, D.J., Fiore, A.M., et al. (2003) Asian outﬂow and trans-Paciﬁc transport of carbon monoxide and ozone pollution: An integrated satellite, aircraft, and model perspective. Journal of Geophysical Research – Atmospheres, 108(D24): 4804. Heeb, N.V., Forss, A.M., Bach, C. and Mattrel, P. (2000a) Velocity-dependent emission factors of benzene, toluene and C-2-benzenes of a passenger car equipped with and without a regulated 3-way catalyst. Atmospheric Environment, 34: 1123–37. Heeb, N.V., Forss, A.M., Bach, C., Reimann, S., Herzog, A. and Jackle, H.W. (2000b) A comparison of benzene, toluene and C-2-benzenes mixing ratios in automotive exhaust and in the suburban atmosphere during the introduction of catalytic converter technology to the Swiss Car Fleet. Atmospheric Environment, 34: 3103–16.

Anthropogenic VOCs

75

Heeb, N.V., Forss, A.M. and Weilenmann, M. (2002) Pre- and post-catalyst-, fuel-, velocity- and acceleration-dependent benzene emission data of gasoline-driven EURO-2 passenger cars and light duty vehicles. Atmospheric Environment, 36: 4745–56. Heeb, N.V., Forss, A.M., Saxer, C.J. and Wilhelm, P. (2003) Methane, benzene and alkyl benzene cold start emission data of gasoline-driven passenger cars representing the vehicle technology of the last two decades. Atmospheric Environment, 37: 5185–95. Hellen, H., Hakola, H. and Laurila, T. (2003) Determination of source contributions of NMHCs in Helsinki (60 degrees N, 25 degrees E) using chemical mass balance and the Unmix multivariate receptor models. Atmospheric Environment, 37: 1413–24. Helmig, D. (1997) Ozone removal techniques in the sampling of atmospheric volatile organic trace gases. Atmospheric Environment, 31: 3635–51. Helmig, D. (1999) Air analysis by gas chromatography. Journal of Chromatography A, 843: 129–46. Helmig, D., Pollock, W., Greenberg, J. and Zimmerman, P. (1996) Gas chromatography mass spectrometry analysis of volatile organic trace gases at Mauna Loa observatory, Hawaii. Journal of Geophysical Research – Atmospheres, 101: 14697–710. Henne, S., Dommen, J., Neininger, B., Reimann, S., Staehelin, J. and Prerot, A.S.H. (2005) Inﬂuence of mountain venting in the Alps on the ozone chemistry of the lower free troposphere and the European pollution export. Journal of Geophysical Research-Atmosphere, 110. Henne, S., Dommen, J., Neininger, B., Reimann, S., Staehelin, J. and Prevot, A.S.H. (2006) Ozone production following Mountain Venting studied with a Lagrangian chemistry-box model. Journal of Geophysical Research, 110: D22307. Hill, H.H. and Mcminn, D. (1992) Detectors for Capillary Chromatography (Chemical Analysis Series). John Wiley & Sons, Hoboken, NJ. Hoffmann, E.D. and Stroobant, V. (2002) Mass Spectrometry Principles and Applications. John Wiley & Sons, Chichester, United Kingdom. Hopkins, J.R., Jones, I.D., Lewis, A.C., Mcquaid, J.B. and Seakins, P.W. (2002) Non-methane hydrocarbons in the Arctic boundary layer. Atmospheric Environment, 36: 3217–29. Hopkins, J.R., Lewis, A.C. and Seakins, P.W. (2005) Analysis and applications of measurements of source dominated hydrocarbon concentrations from the PUMA campaigns in June/July 1999 and January/February 2000 at an urban background site in Birmingham, UK. Atmospheric Environment, 39: 535–48. Hov, O., Sorteberg, A., Schmidbauer, N., et al. (1997) European VOC emission estimates evaluated by measurements and model calculations. Journal of Atmospheric Chemistry, 28: 173–93. Hunter, M.C., Bartle, K.D., Seakins, P.W. and Lewis, A.C. (1999) Direct measurement of atmospheric formaldehyde using gas chromatography pulsed discharge ionisation detection. Analytical Communications, 36: 101–4. Hwa, M.Y., Hsieh, C.C., Wu, T.C., et al. (2002) Real-world vehicle emissions and VOCs proﬁle in the Taipei tunnel located at Taiwan Taipei area. Atmospheric Environments, 36: 1993–2002. Ioffe, B.V., Isidorov, V.A. and Zenkevich, I.G. (1977) Gas chromatographic-mass spectrometric determination of volatile organic-compounds in an urban atmosphere. Journal of Chromatography, 142: 787–95. IPCC (2001) Intergovernmental Panel on Climate Change, Climate Change 2001: The Scientiﬁc Basis. Cambridge University Press, New York. Jacob, D.J., Field, B.D., Jin, E.M., et al. (2002) Atmospheric budget of acetone. Journal of Geophysical Research – Atmospheres, 107(D10): 4100. Jacob, D.J., Crawford, J.H., Kleb, M.M., et al. (2003) Transport and Chemical Evolution over the Paciﬁc (TRACE-P) aircraft mission: Design, execution, and ﬁrst results. Journal of Geophysical Research – Atmospheres, 108: 1–19. Jacob, D.J., Field, B.D., Li, Q.B., et al. (2005) Global budget of methanol: Constraints from atmospheric observations. Journal of Geophysical Research – Atmospheres, 110: D08303.

76

Volatile Organic Compounds in the Atmosphere

Jobson, B.T., Parrish, D.D., Goldan, P., et al. (1998) Spatial and temporal variability of nonmethane hydrocarbon mixing ratios and their relation to photochemical lifetime. Journal of Geophysical Research – Atmospheres, 103: 13557–67. Jobson, B.T., Mckeen, S.A., Parrish, D.D., et al. (1999) Trace gas mixing ratio variability versus lifetime in the troposphere and stratosphere: Observations. Journal of Geophysical Research – Atmospheres, 104: 16091–113. Jobson, B.T., Berkowitz, C.M., Kuster, W.C., et al. (2004) Hydrocarbon source signatures in Houston, Texas: Inﬂuence of the petrochemical industry. Journal of Geophysical Research – Atmospheres, 109: D24305. Jorquera, H. and Rappenglück, B. (2004) Receptor modeling of ambient VOC at Santiago, Chile. Atmospheric Environment, 38: 4243–63. Kalberer, M., Paulsen, D., Sax, M., et al. (2004) Identiﬁcation of polymers as major components of atmospheric organic aerosols. Science, 303: 1659–62. Kanakidou, M., Seinfeld, J.H., Pandis, S.N., et al. (2005) Organic aerosol and global climate modelling: A review. Atmospheric Chemistry and Physics, 5: 1053–123. Kaspers, K.A., Van De Wal, R.S.W., De Gouw, J.A., et al. (2004) Analyses of ﬁrn gas samples from Dronning Maud Land, Antarctica: Study of nonmethane hydrocarbons and methyl chloride. Journal of Geophysical Research – Atmospheres, 109: D02307. Kawamoto, K., Arey, J.S. and Gschwend, P.M. (2003) Emission and fate assessment of methyl tertiary butyl ether in the Boston area airshed using a simple multimedia box model: Comparison with urban air measurements. Journal of the Air and Waste Management Association, 53: 1426–35. Kean, A.J., Grosjean, E., Grosjean, D. and Harley, R.A. (2001) On-road measurement of carbonyls in California light-duty vehicle emissions. Environmental Science and Technology, 35: 4198–204. Kean, A.J., Sawyer, R.F., Harley, R.A. and Kendall, G.R. (2002) Trends in Exhaust Emissions from In-Use California Light-Duty Vehicles, 1994–2001. SAE Technical Paper No. 2002-01-1713. Kim, E., Brown, S.G., Hafner, H.R. and Hopke, P.K. (2005) Characterization of non-methane volatile organic compounds sources in Houston during 2001 using positive matrix factorization. Atmospheric Environment, 39: 5934–46. Kirchstetter, T.W., Singer, B.C., Harley, R.A., Kendall, G.R. and Hesson, J.M. (1999) Impact of California reformulated gasoline on motor vehicle emissions. 2. Volatile organic compound speciation and reactivity. Environmental Science and Technology, 33: 329–36. Kleinman, L.I., Daum, P.H., Imre, D., et al. (2002) Ozone production rate and hydrocarbon reactivity in 5 urban areas: A cause of high ozone concentration in Houston. Geophysical Research Letters, 29: Art. No.-1467. Kleinman, L.I., Daum, P.H., Lee, Y.N., et al. (2003) Photochemical age determinations in the Phoenix metropolitan area. Journal of Geophysical Research – Atmospheres, 108(D3): 4096. Klemp, D., Mannschreck, K., Patz, H.W., Habram, M., Matuska, P. and Slemr, F. (2002) Determination of anthropogenic emission ratios in the Augsburg area from concentration ratios: Results from long-term measurements. Atmospheric Environment, 36: 61–80. Klimont, Z., Streets, D.G., Gupta, S., Cofala, J., Fu, L.X. and Ichikawa, Y. (2002) Anthropogenic emissions of non-methane volatile organic compounds in China. Atmospheric Environment, 36: 1309–22. Kohno, T. and Kuwata, K. (1991) Preconcentration technique for introducing gaseous or volatile compounds into a capillary gas-chromatographic column. Journal of Chromatography, 587: 338–42. Koppmann, R., Bauer, R., Johnen, F.J., Plass, C. and Rudolph, J. (1992) The distribution of light nonmethane hydrocarbons over the mid-Atlantic – Results of the Polarstern Cruise Ant-Vii/1. Journal of Atmospheric Chemistry, 15: 215–34.

Anthropogenic VOCs

77

Kourtidis, K.A., Ziomas, I.C., Rappenglüeck, B., Proyou, A. and Balis, D. (1999) Evaporative trafﬁc hydrocarbon emissions, trafﬁc CO and speciated HC trafﬁc emissions from the city of Athens. Atmospheric Environment, 33: 3831–42. Kramp, F. and Volz-Thomas, A. (1997) On the Budget of OH Radicals and Ozone in an Urban Plume from the Decay of C5 –C8 hydrocarbons and NOx . Journal of Atmospheric Chemistry, 28: 263–82. Kurtenbach, R., Ackermann, R., Becker, K.H., et al. (2002) Veriﬁcation of the contribution of vehicular trafﬁc to the total NMVOC emissions in Germany and the importance of the NO3 chemistry in the city air. Journal of Atmospheric Chemistry 42: 395–411. Lewis, A.C., Bartle, K.D., Mcquaid, J.B., Pilling, M.J., Seakins, P.W. and Ridgeon, P. (1996) Atmospheric monitoring of volatile organic compounds using programmed temperature vaporization injection. Hrc-Journal of High Resolution Chromatography, 19: 686–90. Lewis, A.C., Carpenter, L.J. and Pilling, M.J. (2001) Nonmethane hydrocarbons in Southern Ocean boundary layer air. Journal of Geophysical Research – Atmospheres, 106: 4987–94. Lewis, A.C., Hopkins, J.R., Carpenter, L.J., Stanton, J., Read, K.A. and Pilling, M.J. (2005) Sources and sinks of acetone, methanol, and acetaldehyde in North Atlantic marine air. Atmospheric Chemistry and Physics, 5: 1963–74. Li, Y.S., Campana, M., Reimann, S., et al. (2005) Hydrocarbon concentrations at the Alpine mountain sites Jungfraujoch and Arosa. Atmospheric Environment, 39: 1113–27. Liu, C., Xu, Z., Du, Y. and Guo, H. (2000) Analyses of volatile organic compounds concentrations and variation trends in the air of Changchun, the northeast of China. Atmospheric Environment, 34: 4459–66. Lonneman, W.A. and Bufalini, J.J. (1978) Urban and industrial air-pollution in Houston. Atmospheric Environment, 12: 963–4. Mannschreck, K., Klemp, D., Kley, D., et al. (2002) Evaluation of an emission inventory by comparisons of modelled and measured emission ratios of individual HCs, CO and NOx . Atmospheric Environment, 36: S81–94. Mcgaughey, G.R., Desai, N.R., Allen, D.T., et al. (2004) Analysis of motor vehicle emissions in a Houston tunnel during the Texas Air Quality Study 2000. Atmospheric Environment, 38: 3363–72. Mckeen, S.A. and Liu, S.C. (1993) Hydrocarbon ratios and photochemical history of air masses. Geophysical Research Letters, 20: 2363–6. Mckeen, S.A., Liu, S.C., Hsie, E.Y., et al. (1996) Hydrocarbon ratios during PEM-WEST A: A model perspective. Journal of Geophysical Research – Atmospheres, 101: 2087–109. Millet, D.B., Goldstein, A.H., Allan, J.D., et al. (2004) Volatile organic compound measurements at Trinidad Head, California, during ITCT 2K2: Analysis of sources, atmospheric composition, and aerosol residence times. Journal of Geophysical Research – Atmospheres, 109: D23S16. Mollmann-Coers, M., Klemp, D., Mannschreck, K. and Slemr, F. (2002) Determination of anthropogenic emissions in the Augsburg area by the source-tracer-ratio method. Atmospheric Environment, 36: 95–107. Moschonas, N. and Glavas, S. (1996) C3-C10 hydrocarbons in the atmosphere of Athens, Greece. Atmospheric Environment, 30: 2769–72. −C-9 hydrocarbons concentraNa, K., Kim, Y.P. and Moon, K.C. (2002) Seasonal variation of the C-2− tions and compositions emitted from motor vehicles in a Seoul tunnel. Atmospheric Environment, 36, 1969–78. Na, K., Kim, Y.P. and Moon, K.C. (2003) Diurnal characteristics of volatile organic compounds in the Seoul atmosphere. Atmospheric Environment, 37: 733–42. NAEI (2005) National Atmospheric Emission Inventory, Great Britain. http://www.naei.org.uk/ reports.php?list=AP Neftel, A., Spirig, C., Prevot, A.S.H., et al. (2002) Sensitivity of photooxidant production in the Milan Basin: An overview of results from a EUROTRAC-2 Limitation of Oxidant Production ﬁeld experiment. Journal of Geophysical Research – Atmospheres, 107(D22): 8188.

78

Volatile Organic Compounds in the Atmosphere

NEI (2005) National Emission Inventory USEPA. http://www.epa.gov/ttn/chief/trends/index.html Nguyen, H.T.H., Takenaka, N., Bandow, H., et al. (2001) Atmospheric alcohols and aldehydes concentrations measured in Osaka, Japan and in Sao Paulo, Brazil. Atmospheric Environment, 35: 3075–83. Odum, J.R., Jungkamp, T.P.W., Grifﬁn, R.J., Flagan, R.C. and Seinfeld, J.H. (1997) The atmospheric aerosol-forming potential of whole gasoline vapor. Science, 276: 96–9. Oram, D.E., Reeves, C.E., Sturges, W.T., Penkett, S.A., Fraser, P.J. and Langenfelds, R.L. (1996) Recent tropospheric growth rate and distribution of HFC-134a (CF3CH2F). Geophysical Research Letters, 23: 1949–52. Pierson, W.R., Schorran, D.E., Fujita, E.M., Sagebiel, J.C., Lawson, D.R. and Tanner, R.L. (1999) Assessment of nontailpipe hydrocarbon emissions from motor vehicles. Journal of the Air and Waste Management Association, 49: 498–519. Plass-Dulmer, C., Michl, K., Ruf, R. and Berresheim, H. (2002) C-2-C-8 hydrocarbon measurement and quality control procedures at the Global Atmosphere Watch Observatory Hohenpeissenberg. Journal of Chromatography, A 953: 175–97. Prevot, A.S.H., Dommen, J. and Baumle, M. (2000a) Inﬂuence of road trafﬁc on volatile organic compound concentrations in and above a deep Alpine valley. Atmospheric Environment, 34: 4719–26. Prevot, A.S.H., Dommen, J., Baumle, M. and Furger, M. (2000b) Diurnal variations of volatile organic compounds and local circulation systems in an Alpine valley. Atmospheric Environment, 34: 1413–23. Prevot, A.S.H., Staehelin, J., Kok, G.L., et al. (1997) The Milan photooxidant plume. Journal of Geophysical Research – Atmospheres, 102: 23375–88. Prinn, R.G., Weiss, R.F., Fraser, P.J., et al. (2000) A history of chemically and radiatively important gases in air deduced from ALE/GAGE/AGAGE. Journal of Geophysical Research – Atmospheres, 105: 17751–92. Prinn, R.G., Huang, J., Weiss, R.F., et al. (2005) Evidence for variability of atmospheric hydroxyl radicals over the past quarter century. Geophysical Research Letters, 32 (7): L07809. Purvis, R.M., Lewis, A.C., Carney, R.A., et al. (2003) Rapid uplift of nonmethane hydrocarbons in a cold front over central Europe. Journal of Geophysical Research – Atmospheres, 108(D7): 4224. Quiram, E.R. and Biller, W.F. (1958) Determination of trace quantities of hydrocarbons in the atmosphere. Analytical Chemistry, 30: 1166–71. Rao, A.M.M., Pandit, G.G., Sain, P., Sharma, S., Krishnamoorthy, T.M. and Nambi, K.S.V. (1997) Non-methane hydrocarbons in industrial locations of Bombay. Atmospheric Environment, 31: 1077–85. Rappengluck, B., Fabian, P., Kalabokas, P., Viras, L.G. and Ziomas, I.C. (1998) Quasi-continuous measurements of non-methane hydrocarbons (NMHC) in the greater Athens area during MEDCAPHOT-TRACE. Atmospheric Environment, 32: 2103–21. Reimann, S., Calanca, P. and Hofer, P. (2000) The anthropogenic contribution to isoprene concentrations in a rural atmosphere. Atmospheric Environment, 34: 109–15. Rivett, A.C., Martin, D., Nickless, G., et al. (2003) In situ gas chromatographic measurements of halocarbons in an urban environment. Atmospheric Environment, 37: 2221–35. Roberts, J.M., Fehsenfeld, F.C., Liu, S.C., et al. (1984) Measurements of aromatic hydrocarbon ratios and NOx concentrations in the rural troposphere – Observation of air-mass photochemical aging and NOx removal. Atmospheric Environment, 18: 2421–32. Rudolph, J. (1988) Two-dimensional distribution of light-hydrocarbons – Results from the Stratoz-Iii experiment. Journal of Geophysical Research – Atmospheres, 93: 8367–77. Rudolph, J. (1995) The tropospheric distribution and budget of ethane. Journal of Geophysical Research – Atmospheres, 100: 11369–81.

Anthropogenic VOCs

79

Rudolph, J. and Johnen, F.J. (1990) Measurements of light atmospheric hydrocarbons over the Atlantic in regions of low biological-activity. Journal of Geophysical Research – Atmospheres, 95: 20583–91. Rudolph, J., Johnen, F.J., Khedim, A. and Pilwat, G. (1990a) The use of automated on line gaschromatography for the monitoring of organic trace gases in the atmosphere at low-levels. International Journal of Environmental Analytical Chemistry, 38: 143–55. Rudolph, J., Muller, K.P. and Koppmann, R. (1990b) Sampling of organic volatiles in the atmosphere at moderate and low pollution levels. Analytica Chimica Acta, 236: 197–211. Rudolph, J., Koppmann, R. and Plassdulmer, C. (1996) The budgets of ethane and tetrachloroethene: Is there evidence for an impact of reactions with chlorine atoms in the troposphere? Atmospheric Environment, 30: 1887–94. Scheeren, H.A., Fischer, H., Lelieveld, J., et al. (2003) Reactive organic species in the northern extratropical lowermost stratosphere: Seasonal variability and implications for OH. Journal of Geophysical Research – Atmospheres, 108(D24): 4805. Scheff, P.A. and Wadden, R.A. (1993) Receptor modeling of volatile organic-compounds. 1. Emission inventory and validation. Environmental Science and Technology, 27: 617–25. Scheff, P.A., Wadden, R.A., Kenski, D.M., Chung, J. and Wolff, G. (1996) Receptor model evaluation of the southeast Michigan ozone study ambient NMOC measurements. Journal of the Air and Waste Management Association, 46: 1048–57. Schmid, H., Pucher, E., Ellinger, R., Biebl, P. and Puxbaum, H. (2001) Decadal reductions of trafﬁc emissions on a transit route in Austria – Results of the Tauerntunnel experiment 1997. Atmospheric Environment, 35: 3585–93. Schmitz, T., Hassel, D. and Weber, F.J. (2000) Determination of VOC-components in the exhaust of gasoline and diesel passenger cars. Atmospheric Environment, 34: 4639–47. Sexton, K. and Westberg, H. (1984) Nonmethane hydrocarbon composition of urban and rural atmospheres. Atmospheric Environment, 18: 1125–32. Siddiqi, A.A. and Worley, F.L. (1977) Urban and industrial air-pollution in Houston, Texas. 1. Hydrocarbons. Atmospheric Environment, 11: 131–43. Siegl, W.O., Hammerle, R.H., Herrmann, H.M., Wenclawiak, B.W. and Luers-Jongen, B. (1999) Organic emissions proﬁle for a light-duty diesel vehicle. Atmospheric Environment, 33: 797–805. Sin, D.W.M., Wong, Y.C., Sham, W.C. and Wang, D. (2001) Development of an analytical technique and stability evaluation of 143 C-3-C-12 volatile organic compounds in Summa((R)) canisters by gas chromatography-mass spectrometry. Analyst, 126: 310–21. Singh, H., Chen, Y., Tabazadeh, A., et al. (2000) Distribution and fate of selected oxygenated organic species in the troposphere and lower stratosphere over the Atlantic. Journal of Geophysical Research – Atmospheres, 105: 3795–805. Singh, H.B., Viezee, W. and Salas, L.J. (1988) Measurements of Selected C2–C5 Hydrocarbons in the Troposphere – Latitudinal, Vertical, and Temporal Variations. Journal of Geophysical Research – Atmospheres, 93: 15861–78. Singh, H.B. and Zimmerman, P.B. (1992) Atmospheric distribution and sources of non-methane hydrocarbons. In: J.O. Nriagu (Ed.) Gaseous Pollutants: Characterization and Cycling. Edited by Jerome I. Nriagu, ISBN 0-471-54898-7, Wiley & Sons, Inc., pp. 177–235. Singh, H.B., Chen, Y., Gregory, G.L., et al. (1997) Trace chemical measurements from the northern midlatitude lowermost stratosphere in early spring: Distributions, correlations, and fate. Geophysical Research Letters, 24: 127–30. Singh, H.B., Salas, L.J., Chatﬁeld, R.B., et al. (2004) Analysis of the atmospheric distribution, sources, and sinks of oxygenated volatile organic chemicals based on measurements over the Paciﬁc during TRACE-P. Journal of Geophysical Research – Atmospheres, 109: D15S07. Skoog, D.A., Holler, F.J. and Nieman, T.A. (1998) Gas chromatography. In: Principles of Instrumental Analysis (5th edn). Harcourt Brace College Publishers, Orlando, FL.

80

Volatile Organic Compounds in the Atmosphere

Slemr, F., Baumbach, G., Blank, P., et al. (2002a) Evaluation of modeled spatially and temporarily highly resolved emission inventories of photosmog precursors for the city of Augsburg: The experiment EVA and its major results. Journal of Atmospheric Chemistry, 42: 207–33. Slemr, J., Slemr, F., Partridge, R., D’souza, H. and Schmidbauer, N. (2002b) Accurate Measurements of Hydrocarbons in the Atmosphere (AMOHA): Three European intercomparisons. Journal of Geophysical Research – Atmospheres, 107(D19): 4409. So, K.L. and Wang, T. (2004) C-3-C-12 non-methane hydrocarbons in subtropical Hong Kong: Spatial-temporal variations, source-receptor relationships and photochemical reactivity. Science of the Total Environment, 328: 161–74. Solberg, S., Dye, C., Schmidbauer, N., Herzog, A. and Gehrig, R. (1996) Carbonyls and nonmethane hydrocarbons at rural European sites from the Mediterranean to the Arctic. Journal of Atmospheric Chemistry, 25: 33–66. Solomon, P., Cowling, E., Hidy, G. and Furiness, C. (2000) Comparison of scientiﬁc ﬁndings from major ozone ﬁeld studies in North America and Europe. Atmospheric Environment, 34: 1885–920. Staehelin, J., Keller, C., Stahel, W., Schlapfer, K. and Wunderli, S. (1998) Emission factors from road trafﬁc from a tunnel study (Gubrist tunnel, Switzerland). Part III: Results of organic compounds, SO2 and speciation of organic exhaust emission. Atmospheric Environment, 32: 999–1009. Staehelin, J., Locher, R., Monkeberg, S. and Stahel, W.A. (2001) Contribution of road trafﬁc emissions to ambient air concentrations of hydrocarbons: The interpretation of monitoring measurements in Switzerland by Principal Component Analysis and road tunnel measurements. International Journal of Vehicle Design, 27: 161–72. Steinbacher, M., Dommen, J., Ordonez, C., et al. (2005a) Volatile organic compounds in the Po Basin. Part B: Biogenic VOCs. Journal of Atmospheric Chemistry, 51: 293. Steinbacher, M., Dommen, J., Ordonez, C., et al. (2005b) Volatile organic compounds in the Po Basin. Part A: Anthropogenic VOCs. Journal of Atmospheric Chemistry, 51: 271. Stemmler, K., Bugmann, S., Buchmann, B., Reimann, S. and Staehelin, J. (2005) Large decrease of VOC emissions of Switzerland’s car ﬂeet during the past decade: Results from a highway tunnel study. Atmospheric Environment, 39: 1009–18. Stohl, A. and Eckhardt, S. (2004) Intercontinental transport of air pollution: An introduction. The Handbook of Environmental Chemistry, 4G: 1–12. Stohl, A., Forster, C., Eckhardt, S., et al. (2003) A backward modeling study of intercontinental pollution transport using aircraft measurements. Journal of Geophysical Research – Atmospheres, 108(D12): 4370. Street, R.A., Hewitt, C.N. and Mennicken, S. (1997) Isoprene and monoterpene emissions from a Eucalyptus plantation in Portugal. Journal of Geophysical Research – Atmospheres, 102: 15875–87. Streets, D.G., Bond, T.C., Carmichael, G.R., et al. (2003) An inventory of gaseous and primary aerosol emissions in Asia in the year 2000. Journal of Geophysical Research – Atmospheres, 108(D21): 8809. Tanaka, P.L., Riemer, D.D., Chang, S.H., et al. (2003) Direct evidence for chlorine-enhanced urban ozone formation in Houston, Texas. Atmospheric Environment, 37: 1393–1400. Touaty, M. and Bonsang, B. (2000) Hydrocarbon emissions in a highway tunnel in the Paris area. Atmospheric Environment, 34: 985–96. Tsai, S.M., Zhang, J.F., Smith, K.R., Ma, Y.Q., Rasmussen, R.A. and Khalil, M.A.K. (2003) Characterization of non-methane hydrocarbons emitted from various cookstoves used in China. Environmental Science and Technology, 37: 2869–77. Urhahn, T. and Ballschmiter, K. (2000) Analysis of halogenated C-1/C-2-trace compounds in marine atmosphere. Fresenius Journal of Analytical Chemistry, 366: 365–7. Vairavamurthy, A., Roberts, J.M. and Newman, L. (1992) Methods for determination of lowmolecular-weight carbonyl-compounds in the atmosphere. Atmospheric Environment Part A-General Topics, 26: 1965–93.

Anthropogenic VOCs

81

Vautard, R., Martin, D., Beekmann, M., et al. (2003) Paris emission inventory diagnostics from ESQUIF airborne measurements and a chemistry transport model. Journal of Geophysical Research – Atmospheres, 108(D17): 8564. Vestreng, V. (2003) EMEP/MSC-W Technical report. Review and Revision. Emission data reported to CLRTAP. MSC-W Status Report 2003. EMEP/MSC-W Note 1/2003. http://www.emep.int/mscw/mscw_publications.html Warneke, C. and De Gouw, J.A. (2001) Organic trace gas composition of the marine boundary layer over the northwest Indian Ocean in April 2000. Atmospheric Environment, 35: 5923–33. Watson, J.G., Chow, J.C. and Fujita, E.M. (2001) Review of volatile organic compound source apportionment by chemical mass balance. Atmospheric Environment, 35: 1567–84. Weilenmann, M., Soltic, P., Saxer, C., Forss, A.M. and Heeb, N. (2005) Regulated and nonregulated diesel and gasoline cold start emissions at different temperatures. Atmospheric Environment, 39: 2433–41. West, J.J., Zavala, M.A., Molina, L.T., et al. (2004) Modeling ozone photochemistry and evaluation of hydrocarbon emissions in the Mexico City metropolitan area. Journal of Geophysical Research – Atmospheres, 109: D19312. Whalley, L.K., Lewis, A.C., Mcquaid, J.B., et al. (2004) Two high-speed, portable GC systems designed for the measurement of non-methane hydrocarbons and PAN: Results from the Jungfraujoch High Altitude Observatory. Journal of Environmental Monitoring, 6: 234–41. Williams, J., Fischer, H., Harris, G.W., et al. (2000) Variability-lifetime relationship for organic trace gases: A novel aid to compound identiﬁcation and estimation of HO concentrations. Journal of Geophysical Research – Atmospheres, 105: 20473–86. Wingenter, O.W., Blake, D.R., Blake, N.J., et al. (1999) Tropospheric hydroxyl and atomic chlorine concentrations, and mixing timescales determined from hydrocarbon and halocarbon measurements made over the Southern Ocean. Journal of Geophysical Research – Atmospheres, 104: 21819–28. Winkler, J., Blank, P., Glaser, K., et al. (2002) Ground-based and airborne measurements of nonmethane hydrocarbons in BERLIOZ: Analysis and selected results. Journal of Atmospheric Chemistry, 42: 465–92. Yamamoto, N., Okayasu, H., Murayama, S., Mori, S., Hunahashi, K. and Suzuki, K. (2000) Measurement of volatile organic compounds in the urban atmosphere of Yokohama, Japan, by an automated gas chromatographic system. Atmospheric Environment, 34: 4441. Zhao, W.X., Hopke, P.K. and Karl, T. (2004) Source identiﬁcation of volatile organic compounds in Houston, Texas. Environmental Science and Technology, 38: 1338–47.

Volatile Organic Compounds in the Atmosphere Edited by Ralf Koppmann Copyright © 2007 by Blackwell Publishing Ltd

Chapter 3

Biogenic VOCs Allison H. Steiner and Allen L. Goldstein

3.1

Introduction

Biogenic volatile organic compounds (VOCs) are carbon-containing compounds emitted naturally from the Earth’s surface to the atmosphere. This general categorisation includes a wide range of organic species emitted from vegetation, soils and the oceans, and generally excludes methane, carbon monoxide and carbon dioxide (Kesselmeier and Staudt 1999). This chapter reviews current knowledge on the atmospheric budget of these compounds, from the origin of emission to their fate in the atmosphere. The ﬁrst biogenic VOC emissions speciﬁcally identiﬁed to be important for atmospheric composition were isoprene (Sanadze 1957) and terpenoid compounds (Went 1960). Subsequent studies indicated that a wide variety of organic species are routinely emitted from vegetation (Zimmerman 1979). The discovery that isoprene can be the dominant VOC impacting regional tropospheric photochemistry and ozone production (Chameides et al. 1988; Trainer et al. 1987) dramatically enhanced the amount of research focusing on biogenic VOCs. As research has progressed over the last few decades, the understanding of the types and quantities of these emissions has grown signiﬁcantly. Multiple review articles have been published that can provide a scientiﬁc history of the topic (e.g. Fehsenfeld et al. 1992; Fuentes et al. 2000; Kesselmeier and Staudt 1999; Wiedinmyer et al. 2004). The importance of biogenic VOCs is twofold. First, ecophysiologists investigate biogenic VOCs to elucidate the mechanism of plant photosynthesis and metabolism. Secondly, they are the dominant reactive organic compounds in the troposphere, making them important to atmospheric chemists and climatologists. Many species of biogenic VOCs are rapidly oxidised in the atmosphere, playing an important role in atmospheric chemistry and the formation of ground-level ozone. Some oxygenated biogenic VOCs are photolysed in the upper troposphere, providing a source of HOx radicals and impacting the oxidation capacity of the atmosphere at large scales. Additionally, the oxidation of many biogenic VOCs can yield products that undergo gas-to-particle conversion, leading to the growth of secondary organic aerosols that affect the Earth’s energy balance and thus surface climate. The intersection of these two ﬁelds has resulted in a wide body of research on scales ranging from biological synthesis at the molecular level to atmospheric chemistry at the global level. In this chapter, we will review and discuss the state-of-the-science of biogenic VOC research. We deﬁne the chemical speciation of identiﬁed biogenic VOCs and present their known sources from the molecular level up to the landscape level (Section 3.2), and then present current estimates on the magnitude of these emissions and how they are modelled at the leaf, regional and global scales (Section 3.3). Next, we discuss their atmospheric distributions

Biogenic VOCs

83

(Section 3.4) and their impact on atmospheric chemistry (Section 3.5). Finally, we discuss mixing ratio and ﬂux measurement techniques used to constrain the atmospheric budget (Section 3.6).

3.2 3.2.1

Sources of biogenic VOCs Chemical speciation

Global emissions inventories as recent as the 1990s categorised biogenic VOC emissions into isoprene, monoterpenes and a general grouping of ‘other VOCs’ (Guenther et al. 1995). In early inventories, where little information was available on the chemical species emitted or the speciﬁc sources of the emissions, this term distinguished between ‘other reactive’ VOCs and ‘other VOCs’ with lifetimes greater than or less than 1 day, respectively (Guenther et al. 1995). Here, we re-categorise biogenic VOCs to reﬂect the most recent research by adding two categories: oxygenated VOCs (oxVOCs) and very reactive biogenic VOCs (VR-BVOC) (Holzinger et al. 2004). Table 3.1 shows examples of the chemical formulas and structures of these categories, and they are described in further detail below. The most studied group of biogenic VOCs is the isoprenoid compounds. Isoprenoids are composed of C5 (or hemiterpene) building block structures. The most well-characterised compound of this category in terms of biogenic emissions and atmospheric chemistry is isoprene (2-methyl-1,3-butadiene; C5 H8 with structure shown in Table 3.1), due to its copious emissions, high reactivity and relatively early discovery. Sanadze (1957) and Rasmussen and Went (1965) made the ﬁrst measurements of isoprene emitted from foliage and found that it was released from several types of vegetation species. Since this discovery, isoprene emissions have been measured at the leaf and canopy levels in hundreds of ecosystems. Terpenoids are a large class of compounds containing combinations of the C5 isoprenoid structure. These include C10 compounds known as monoterpenes, C15 compounds or sesquiterpenes and the larger C20 , C25 and C30 molecules. Up to 5 000 different terpenoid structures have been identiﬁed in emissions from vegetation, of which 14 common species have been determined as dominant in biogenic VOC emissions (Geron et al. 2000a). The most abundant of these include the monoterpenes listed and illustrated in Table 3.1. Terpenoid emissions were ﬁrst noted by Went in 1960, with a prescient hypothesis that these biogenic emissions could lead to the haze often seen in rural areas such as the Great Smoky Mountains. Because of their wide range of reactivities and their differences in propensity to form secondary organic aerosols, recent studies have underscored the importance of speciating monoterpenes for understanding their impacts on atmospheric chemistry and climate (Hallquist et al. 1999). Regarding biogenic emissions and their importance in atmospheric chemistry, monoterpenes are the most studied of the terpenoid category. They have a C10 structure that enables the molecule to be quickly oxidised in the atmosphere, yielding lifetimes ranging from several minutes up to a day. Monoterpenes have double bonds that can be contained inside a ring structure (endocyclic) or outside the ring structure (exocyclic). Some monoterpenes are a combination of exocyclic and endocyclic bonds, or they can be open-ended structures. Most importantly, the position of the double bond can affect the reactivity of the molecule, which can greatly inﬂuence atmospheric chemistry. Sesquiterpenes (C15 compounds) have

84

Volatile Organic Compounds in the Atmosphere

Table 3.1 Categories of biogenic volatile organic compounds Category

Chemical formula/deﬁnition

Example compound/structure

Other compounds

Isoprenoids

C5 compounds

Isoprene

Methylbutenol

Exocyclic

C10 H16 Double bonds outside ring structure

β-Pinene

Camphene, sabinene, p-cymene

Endocyclic

C10 H16 Double bonds inside ring structure

α-Pinene

3 -Carene, α-terpinene, γ -terpinene

Combination

C10 H16 Double bonds inside and outside ring structure

Limonene

Terpinolene, β-phellandrene

Open-ended

C10 H16 No ring structure present

Myrcene

Ocimene

Sesquiterpenes

C15 H24

β-Caryophyllene

α-Humulene

Terpenoids

Biogenic VOCs

85

Table 3.1 (Continued) Category

Chemical formula/deﬁnition

Oxygenated

C10 compounds containing oxygen

Methyl chavicol

C and O containing compounds

Acetone

oxVOCs

Example compound/structure

Terpenoid or oxVOCs with lifetimes on the order of minutes to hours

α-Thujene

O

Acetaldehyde, methanol, ethanol, methylbutenol

O

H 3C VR-BVOCs

Other compounds

β-Caryophyllene

CH3 β-Farnesene, α-terpinene

a more elaborate ring structure than monoterpenes and often contain several double bonds, leading to rapid oxidation after they are emitted to the atmosphere. Additionally, some terpenoid compounds can be oxygenated, such as methyl chavicol. Table 3.1 shows some examples of structures and compound names for these different types of terpenoid compounds. Oxygenated biogenic VOCs, or oxVOCs, have received more attention in the past decade as technology to measure their emissions and atmospheric mixing ratios has improved. oxVOCs are deﬁned as carbon-based compounds containing an oxygen atom, including alcohols, ketones, esters and ethers. While oxVOCs have been known for decades to be emitted from plants (Zimmerman 1979), the biogenic contribution to their atmospheric budgets have been investigated only recently (Goldstein and Schade 2000; Heikes et al. 2002; Jacob et al. 2002; Schade and Goldstein 2001; Singh et al. 1994, 2000). Dominant observed oxVOC species include methanol, ethanol, methylbutenol (MBO), formaldehyde, acetaldehyde and acetone. Biogenic emissions are the dominant source of oxygenated

86

Volatile Organic Compounds in the Atmosphere

compounds in the atmosphere, having a signiﬁcant impact on atmospheric chemistry. The importance and emission of oxVOCs are thoroughly reviewed in Fall (2003) and discussed at length in Chapter 4 of this book. In this chapter, the focus is on biogenic oxVOC emissions of MBO, methanol, ethanol, acetone and acetaldehyde, as these are the most studied of the oxVOC biogenic species. In addition to the terpenoid and oxVOC compounds discussed above, a wide range of gases can be emitted from vegetation, including alkanes, alkenes, organic acids, carbonyls, esters and ethers (Kesselmeier and Staudt 1999). Past literature has grouped these compounds together with oxygenated compounds into the ‘other VOC’ category, but recent research provides further speciation and quantiﬁcation of these important emissions. Analyses from several ﬁeld measurement campaigns have recently suggested that a large number of terpenoid compounds are reacting quickly within the forest canopy, leaving them unaccounted for in ﬂuxes measured escaping from the canopy. Ciccioli et al. (1999) discussed the importance of these compounds, particularly the large emissions of β-caryophyllene in a blooming orange grove in Spain. In this study, some monoterpenes were observed in leaf emissions, but could not be detected above the canopy. Holzinger et al. (2004) measured oxidation products in and above a ponderosa pine ecosystem and found evidence of very reactive terpenoid compounds being oxidised before escaping the forest canopy, and termed these emissions as VR-BVOCs. VR-BVOC describes difﬁcult-tomeasure, short-lived compounds that are likely to be oxidised before escaping the forest canopy. A listing of a few of the possible VR-BVOCs is included in Goldstein et al. (2004) and Table 3.1, and includes compounds such as α- and β-farnesene, α- and β-terpinene, α-humulene, β-carophyllene and terpinolene. In addition to direct measurements, there are also several studies that provide indirect evidence of the importance of VR-BVOCs. Kurpius and Goldstein (2003) found that the main process responsible for ozone deposition to a ponderosa pine forest was gasphase chemical reaction, and that this loss scaled with temperature in the same manner as monoterpene emissions. Furthermore, Goldstein et al. (2004) found that the chemical loss term increased signiﬁcantly during forest thinning, a mechanical disturbance that increased monoterpene emissions in the same proportion as ozone loss. In a different ecosystem, Di Carlo et al. (2004) found that the measured biogenic VOCs could not account for all of the OH reactivity in a deciduous forest. They suggest that unmeasured biogenic VOCs were responsible for the missing reactivity because it scaled with temperature in the same relationship as biogenic VOC emissions. These VR-BVOC emissions are important because they act as sources and sinks for atmospheric radicals and provide a source of oxidation products that may persist in the gas phase or partition into secondary organic aerosols. Here, we deﬁne VR-BVOC as compounds with atmospheric lifetimes of minutes or less.

3.2.2 3.2.2.1

Mechanisms for biogenic VOC emissions Isoprene formation

In the past decade, a signiﬁcant amount of progress has been made in understanding the isoprene synthesis process. Two recent papers have thoroughly reviewed the topic of isoprene formation (Sanadze 2004; Sharkey and Yeh 2001), and here we brieﬂy summarise current knowledge of the formation mechanisms. Isoprene is produced via de novo synthesis

Biogenic VOCs

87

Cytoplasm MVA pathway 3-Acetyl-CoA

Chloroplast MEP pathway

HMG-CoA

GA-3P + pyruvate Mevalonate 1-Deoxy-D-xyulose-5-P DMAPP +IPP

IPP

IPP

DMAPP

IPP isoprene

+IPP

+IPP FPP

?

DMAPP

Sesquiterpenes Monoterpenes

Figure 3.1 Isoprene formation pathway, based on Lichtenthaler 1999 and Karl et al. 2002.

(Delwiche and Sharkey 1993; Sanadze et al. 1972), meaning that isoprene emissions occur almost immediately after formation. Initially, the isoprene formation pathway was thought to follow the mevalonic (MVA) pathway, which produces isoprene through acetyl-CoA. While this pathway is responsible for the synthesis of isoprenoids in the cytoplasmic space, recent research shows that a different pathway produces isoprene in the plant plastids (Lichtenthaler et al. 1997; Zeidler et al. 1997). This plastidic pathway is known as the methylerythritol phosphate (MEP) pathway (reviewed in detail in Lichtenthaler, 1999, and shown in Figure 3.1), and produces the main isoprene precursor, dimethylallyl pyrophosphate (DMAPP). DMAPP is then catalysed by isoprene synthase to form isoprene (Silver and Fall 1991, 1995). Even with these recent advances in understanding the formation pathway of isoprene, the biological regulatory controls are still uncertain. It is believed that production is likely regulated by some combination of isoprene synthase activity and DMAPP substrate production (Rosenstiel et al. 2002). Several investigations have attempted to determine the sources of carbon utilised for isoprene formation. Part of the motivation for understanding the synthesis mechanisms and sources of carbon is to illuminate the biological function of isoprene. Recent experiments have used isotopically labelled carbon dioxide to determine the percentage of isoprene’s carbon derived from the photosynthetic cycle. Based on these studies, it is now well established that chloroplastic carbon accounts for about ∼80% of the isoprene carbon under normal conditions with a lesser percentage under stress conditions (Affek and Yakir 2003; Karl et al. 2002; Kreuzwieser et al. 2002; Loreto et al. 2004; Schnitzler et al. 2004). Isoprene synthesis is therefore closely linked to the photosynthetic cycle, but other carbon reserves can be tapped to produce isoprene even when the photosynthetic cycle is operating under drought or temperature stress.

88

Volatile Organic Compounds in the Atmosphere

There is still a wide range of speculation about the biological purpose of isoprene emissions, which often account for a few percent of the carbon photosynthesised by the plant. Possible roles of isoprene include providing thermotolerance for vegetation, acting as an antioxidant, or providing a mechanism to release excess energy and/or carbon. Proponents of the thermotolerance hypothesis suggest that the emission of isoprene increases the stability of protein membranes in the chloroplasts of leaves (Sharkey and Singsaas 1995; Singsaas et al. 1997), although subsequent studies could not reproduce this thermotolerance with statistically signiﬁcant results (Logan and Monson 1999). Others have hypothesised that isoprene can quench ozone to below the toxic level for plants (Loreto et al. 2001b) or protect against damage by singlet oxygen radicals (Affek and Yakir 2002). A third hypothesis is that isoprene could be a mechanism to shunt excess energy (as ATP) or carbon from the leaves (Logan et al. 2000). A new hypothesis by Rosenstiel et al. (2004) suggests that isoprene acts as a metabolic ‘safety valve’ by keeping the carbon substrate and intermediate species (such as DMAPP) at optimal levels. Because of these conﬂicting theories, the biological purpose of isoprene production in plants remains an active area of research.

3.2.2.2

Terpenoid formation

Monoterpenes are formed from the reaction of DMAPP and IPP (see Figure 3.1), and their formation has been likened to the merging of two ﬁve-carbon branched chains (Fuentes et al. 2000). There are two pathways of monoterpene formation that depend on the nature of the emissions. Temperature-dependent terpenes are usually synthesised in leucoplasts or the cytosol by the mevalonate pathway and then stored in secretory organs in the plant such as glandular trichomes and resin ducts (Kesselmeier and Staudt 1999; Loreto et al. 2001a; Niinemets and Reichstein 2002). The terpene formation mechanism was ﬁrst speculated upon by Yokouchi and Ambe (1984) and later conﬁrmed with observations (Kesselmeier et al. 1997; Staudt and Seufert 1995; Steinbrecher 1989). The monoterpene storage structures can be internal to the leaf (such as a resin duct in a pine tree) or external (such as a glandular trichrome on the surface of mint leaves). It is believed that these compounds play a role as a defence mechanism of the plant (hence their storage in specialised structures), and speciﬁc monoterpenes are known to both attract and repel insects. A second formation pathway occurs if the monoterpene emissions are both light and temperature dependent. Like isoprene formation, this production is linked to the photosynthetic cycle and forms monoterpenes in the chloroplasts via the MEP pathway (Loreto et al. 2001a). These light- and temperature-dependent emissions likely perform a similar biological function as isoprene. Higher-order terpenoid emissions, such as sesquiterpenes and diterpenes, are formed from the similar C5 units as isoprenoids (IPP and DMAPP), as shown in Figure 3.1. At this time, it is believed that sesquiterpenes are formed via the mevalonate pathway in the cytosol (Eisenreich et al. 2001). Different terpenoid emissions are likely to have varying functions at varying scales, including protection against cellular damage at the tissue level, acting as antioxidants at the leaf surface level and acting as signal compound for insects and animals at the ecosystem level (Holopainen 2004).

3.2.2.3

oxVOC formation

Here we provide a brief review of the mechanisms of biogenic oxVOC formation; these pathways are discussed in great detail in Fall (2003). In particular, we focus on methanol,

Biogenic VOCs

89

ethanol, MBO, acetaldehyde and acetone as the most information is available about the formation and emission of these compounds. Methanol emissions were ﬁrst noted by MacDonald and Fall (1993) and can be emitted in fairly large amounts from terrestrial vegetation. One known source of methanol is from the reactions of pectin in plant cell walls (Galbally and Kirstine 2002), but other sources from terrestrial vegetation are likely as well (Heikes et al. 2002). Emissions of methanol can also occur due to plant wounding (Warneke et al. 2002) or from decomposing leaf and soil litter (Baker et al. 2001; Schade and Goldstein 2001). Ethanol, another potentially important emission, can be produced in the plant under two possible scenarios (Schade and Goldstein 2002). The ﬁrst is the ‘ﬂooding scenario’, where ethanol is produced via fermentation in submerged roots and then transported to other portions of the plant, later to be emitted via leaf tissue (Kreuzwieser et al. 1999). The other possible mechanism for ethanol emissions is the production in the leaf from fermentation under stress conditions, such as drought or damaging trace gases such as ozone (Schade and Goldstein 2002). MBO is formed in a similar mechanism to that of isoprene, despite the fact that they are emitted from different types of vegetation. As with isoprene, MBO is formed via the MEP pathway with DMAPP as the precursor (Fall 2003). Once DMAPP is formed, a MBO synthase enzyme converts DMAPP to MBO (Fisher et al. 2000). The biological functions of MBO are also unclear, though are expected to play a similar role as isoprene, given their similarities in chemical structure. Acetaldehyde is often emitted during light-to-dark transitions in plants (Holzinger et al. 2000; Karl et al. 2002), and also as a by-product of the ‘ﬂooding scenario’ described above (Kreuzwieser et al. 1999). During light-to-dark transitions, pyruvic acid rises in the leaf and is then catalysed by a safety value to convert excess pyruvate to acetaldehyde, which is often ‘leaked’ from intercellular space. The other possible mechanism for acetaldehyde formation is by the oxidation of root-produced ethanol, which occurs in the leaves and can emit acetaldehyde under certain instances (Kreuzwieser et al. 1999). Acetone, also emitted in large quantities from vegetation, can form by several enzymatic pathways within the leaf. It can be emitted during leaf wounding or from light-dependent or light-independent responses in the leaf (see Fall 2003 for a detailed review of these pathways).

3.2.2.4

Mechanisms of VR-BVOC formation

As the category of VR-BVOCs and speciﬁc compounds therein has been recently delineated, there is little evidence for formation or biological function. Here we categorise VR-BVOCs primarily to include terpenoid compounds based on several measurement studies (Ciccioli et al. 1999; Goldstein et al. 2004; Holzinger et al. 2004). Most C10 terpenes are formed from DMAPP and IPP, and sesquiterpenes are formed from DMAPP and two IPP molecules. Different types of terpenes and sesquiterpenes require different enzymes to form the end products; therefore, it is believed that speciﬁc enzymes are required for different terpenoid formations (Iijima et al. 2004). We would expect the other reactive terpenoid and sesquiterpene compounds to play similar ecological roles as the monoterpene emissions. The presence of these compounds in plants has long been studied by wood products chemists (Zavarin 1968), but investigations on their release into and impact on the atmosphere are preliminary.

90

Volatile Organic Compounds in the Atmosphere

3.2.3

Release of volatiles

Once an organism has overcome its biochemical constraints on production, biogenic VOCs must overcome the physiological constraints and travel from the intercellular air spaces and storage structures into the atmosphere. There are three possible foliar release mechanisms: 1. exit of gases via the stomates; 2. non-stomatal exit via diffusion through the leaf cuticle or woody portions of the plant; 3. release of compounds from the storage pools in plants.

3.2.3.1

Stomatal exit

Compounds that are not stored in specialised structures in the leaf can exit from the intercellular air space to the atmosphere through the stomata. In theory, this exit should be controlled by changes in the stomatal aperture (conductance); however, early studies showed that isoprene emission is not limited by the opening and closing of the stomates (Fall and Monson 1992; Monson 1989; Tingey et al. 1981). This led to speculation that isoprene was unaffected by stomatal conductance because of the large gradient between the intercellular and atmospheric mixing ratios. Fall and Monson (1992) postulated that this gradient was so great that isoprene could overcome the diffusive limitations incurred by the nearly closed stomata. The stomatal ﬂux of gases from the leaf can be described by the following relationship: F = Gs P

(3.1)

where F is the emission ﬂux of the biogenic VOC compound, Gs is the stomatal conductance of that compound, and P is the partial pressure gradient between the atmosphere and the inside of the leaf. In the past, it was assumed that if Gs decreased, the increase in P could overcome any stomatal control restricting release of VOCs. Other compounds, such as α-pinene and light-dependent monoterpenes, have been observed to exhibit a similar lack of stomatal control (Niinemets and Reichstein 2002). However, stomatal closure can reduce the emission of other oxVOCs, including methanol (Nemecek-Marshall et al. 1995), ethanol (Schade and Goldstein 2002), carboxylic acid (Kesselmeier et al. 1997) and other organic acids (Gabriel et al. 1999), yet the mechanistic reason for this control was not understood in these observational studies. Niinemets and Reichstein (2003a, 2003b) hypothesised that varying stomatal controls on biogenic VOC emissions are related to the partitioning between gas and liquid phases in the intercellular air space. If compounds are highly volatile with a high Henry’s law constant (H), then gas-phase mixing ratios build up immediately in the leaf and overcome stomatal resistance. However, if compounds have a low H and are likely to reside in the liquid phase, emissions will not occur until a signiﬁcant gas-phase mixing ratio has accumulated. Niinemets and Reichstein (2003b) use this hypothesis to explain bursts of emissions shortly after stomatal openings.

3.2.3.2

Non-stomatal exit

Early VOC studies investigated the possibility of emissions through the cuticle of the leaf (Fall and Monson 1992; Guenther et al. 1991), though more recent evidence indicates that

Biogenic VOCs

91

this may have resulted from leaf wounding during the experiments (Loreto et al. 1996b). In general, however, non-stomatal emissions are considered to be quite small in comparison with emissions through the stomata. Air diffusion rates through the cuticle are a factor of eight times less than diffusion through the stomata and about four times less for water (Niinemets and Reichstein 2003a). Therefore, most work assumes that the emissions are emitted via the stomata but the possibility remains that small amounts could be emitted through the cuticle.

3.2.3.3

Exit via mechanical disturbance

Another common exit pathway for biogenic VOCs is via mechanical wounding. Emissions have been measured as a result of mechanical disturbance and can last long after the physical disturbance has occurred (Rasmussen and Went 1965; Tingey et al. 1991). This can occur with natural perturbations such as insect damage or herbivore feeding (Litvak et al. 1999) or from anthropogenic inﬂuences such as pasture cutting (Kirstine et al. 1998) and forest thinning (Goldstein et al. 2004; Schade and Goldstein 2003). These emissions can include an increase in existing emissions (as in the case of monoterpene emissions increasing with herbivory feeding or forest thinning), or create emissions of new biogenic VOCs, such as the release of hexanal and hexane family VOCs after plant wounding (Fall et al. 1999). When leaves are broken or disturbed, ducts or trichromes that store biogenic VOCs can be broken open and the compounds are volatilised upon contact with air (Litvak and Monson 1998, Loreto et al. 2000). They can also be volatilised directly from the storage structures, depending on temperatures and the monoterpene pool mixing ratio (Lerdau et al. 1997). Often, this type of release can cause experimental artefacts with leaf-, branchor plant-level sampling methods, where the process of enclosing vegetation for measurements can damage portions of the plant tissue and cause emissions that might not occur otherwise.

3.2.4

Emission sources of biogenic VOCs

The predominant source of biogenic VOC emissions is from the foliage of terrestrial vegetation. This includes natural vegetation such as trees, shrubs, grasses, ferns and mosses, as well as anthropogenically induced vegetation such as crops and urban landscapes. Other minor sources such as oceanic and soil emissions can also contribute to global totals of biogenic VOCs. In general, the following discussion of biogenic VOC emissions will be based on emissions from terrestrial vegetation, as these emissions account for the majority of biogenic VOC emissions.

3.2.4.1

Land vegetation

Trees account for the largest portion of biogenic VOC emissions (Wiedinmyer et al. 2004) and comprise 75% of the total global BVOC (based on the last global estimate of Guenther et al. 1995). The main emission from deciduous trees is isoprene, while terpenoid emissions are dominant from coniferous trees. Certain pine species emit MBO in large quantities equivalent to that of isoprene (Baker et al. 1999; Goldan et al. 1993; Schade and Goldstein 2001). Furthermore, coniferous forests can account for a large fraction of oxVOC emissions,

92

Volatile Organic Compounds in the Atmosphere

including methanol, acetone, ethanol and acetaldehyde (Baker et al. 1999; Goldan et al. 1993; Schade and Goldstein 2001). Both coniferous and deciduous forests can emit VR-BVOC compounds, particularly during the ﬂowering season of deciduous trees (Ciccioli et al. 1999; Holzinger et al. 2004). Other vegetated ecosystems, including shrubs and grasslands, can also be a signiﬁcant source of biogenic VOC emissions. Despite the limited amount of measurements of isoprene emissions from shrubs, the fraction of shrubs that emit isoprene is generally considered to be similar that of trees (Wiedinmyer et al. 2004). Terpenoid compounds can be emitted in high quantities from shrubs, particularly in Mediterranean ecosystems such as those in California and Europe (Hansen et al. 1997; Winer 1983). Undisturbed grasslands were initially considered to be low emitters (Guenther et al. 1999a; Zimmerman 1979), yet other studies have found that they can emit signiﬁcant amounts of isoprene, terpenes and oxVOCs (Fukui and Doskey 2000; Koenig et al. 1995). Cutting and drying of grasses can release a wide range of oxVOCs, including oxygenated species and esters (DeGouw et al. 2000; Kirstine et al. 1998; Warneke et al. 2002). Additionally, other phyla in the plant kingdom can also emit isoprene, including ferns (Evans et al. 1985) and mosses (Hanson et al. 1999). Anthropogenically managed land cover, such as crops or urban landscapes, can also emit biogenic VOCs into the atmosphere. Agricultural crops are a source of a wide variety of biogenic VOCs (Arey et al. 1991), yet there are very few studies identifying and quantifying the emissions for a signiﬁcant range of major crops. Like grasslands, the cutting and harvesting of crops such as hay, alfalfa and corn can also produce signiﬁcant quantities of oxVOC (DeGouw et al. 2000; Karl et al. 2001b). Urban landscapes, including tree and shrub species, also emit a wide range of quantities and species of biogenic VOC compounds. Additionally, the cutting of grasses in urban areas can release a small but signiﬁcant amount of oxVOCs into the atmosphere (Karl et al. 2001b).

3.2.4.2

Oceans and soils

Oceans are a small source of biogenic VOC when compared to terrestrial vegetation (Guenther et al. 1995), but can be the dominant source for some speciﬁc compounds. Several classes of biogenic VOCs are emitted from oceans, including hydrocarbons, oxVOCs, halocarbons and sulphur-containing compounds such as dimethyl sulphide (DMS). The halogenated and sulphur-containing compounds are technically biogenic VOCs, but we will not focus on them in this review as their impacts on atmospheric chemistry are signiﬁcantly different from the terrestrially dominant isoprenoid and oxVOC compounds. Small quantities of isoprene, oxVOCs and other hydrocarbons have been measured as emissions from seawater (Bonsang et al. 1992; Donahue and Prinn 1990, 1993; Plass-Dulmer et al. 1995). Isoprene is emitted as by-products of phytoplankton activity in the ocean surface layer and deeper ocean or from seaweeds (Bonsang et al. 1992; Broadgate et al. 2004). Global oceanic isoprene emissions have been estimated to range from 0.2 to 1.2 Tg C/year, and a recent study by Palmer and Shaw (2005) estimated the marine source of isoprene to be 0.1 Tg C/year. Hydrocarbons, including C2 –C5 compounds, are emitted from seawater and are estimated to be ∼2.1 Tg C/year (Bonsang et al. 1992; Plass-Dulmer et al. 1995). While the magnitudes may be relatively small, oceanic sources may be important for remote atmospheric chemistry, particularly in the case of oxVOC. The importance of oxVOCs and their role in atmospheric chemistry has been noted in studies by

Biogenic VOCs

93

Singh et al. (2000, 2001, 2004) and is discussed in greater detail in Chapter 4. There is conﬂicting evidence as to whether the ocean is a net source or sink for speciﬁc VOCs. Few measurements have been performed to determine the direction of ﬂuxes in the ocean. A recent study by Williams et al. (2004) measured atmospheric and seawater mixing ratios of methanol and acetone and found that the ocean was a sink for methanol and a source for acetone, although the authors caution that these ﬂux directions can vary with location, time of year and the transport of anthropogenic emissions. Inverse modelling studies are often used to determine potential sources and sinks (Heikes et al. 2002; Jacob et al. 2002). Based on several atmospheric measurements and modelling studies, it is believed that the ocean is a source of acetaldehyde (Singh et al. 2003), while acetone is sometimes considered a source (Jacob et al. 2002) and sometimes a sink (Singh et al. 2003). For methanol, Singh et al. (2003) determined that the ocean was in near equilibrium and possibly a net sink. Soils are also considered a minor source of biogenic VOCs or a minor sink, depending on the environmental conditions. However, recent studies have found that oxVOCs can be emitted from bare soil (Schade and Goldstein 2001), with factors such as the amount of litter present, the soil moisture and soil temperature impacting the emission rates. Further research is necessary here to determine the magnitudes of these emissions and their signiﬁcance with respect to terrestrial emissions.

3.2.5

Environmental controls on biogenic VOC

Biogenic VOC emissions are strongly dependent on environmental conditions, making them particularly sensitive to changes in climate. Light and temperature are two important factors controlling the growth and function of plants, and they are the dominant environmental inﬂuences on biogenic VOC emissions as well. In this section, we discuss the impact of immediate (light, temperature, moisture and CO2 -mixing ratios) and long-term (phenology) environmental factors on biogenic VOC emissions.

3.2.5.1

Light dependencies

Since the discovery that plants emit isoprene, the dependence of biogenic VOCs on light has been quantiﬁed (Rasmussen and Went 1965; Sanadze 1957). Isoprene emission is strongly dependent on photosynthetically active radiation (PAR; the portion of the spectrum (400–700 nm) that activates photosynthesis). The response of isoprene to PAR is hyperbolic, and emissions increase with increasing light until they reach a saturation point. The instantaneous effects of light on isoprene emissions have been parameterised in order to model biogenic VOC ﬂuxes (Evans et al. 1985; Guenther et al. 1991, 1993; Pandis et al. 1991), and Figure 3.2a shows the common form of this hyperbolic behaviour. Longer-term effects of light on isoprene emissions have also been noted, particularly that sunlit leaves tend to have higher isoprene emission capacities than shaded leaves (Harley et al. 1996; Lerdau and Throop 1999; Litvak et al. 1996). The current parameterisations of light on isoprene emissions are discussed further in Section 3.3.3. Many studies have found that monoterpene emissions are dominantly a function of temperature and are not impacted by PAR (Dement et al. 1975; Guenther et al. 1993). Yet, Steinbrecher (1989) found that certain species of Mediterranean oak emit monoterpenes as a function of both light and temperature in signiﬁcant quantities that rival isoprene

94

Volatile Organic Compounds in the Atmosphere

(a) 1.2

Light activity factor

1.0 0.8 0.6 0.4 Isoprene and monoterpenes Methylbutenol

0.2 0.0 0

200

400

600

800

1000

1200

1400

PAR (micromoles / m2 h) (b)

Temperature activity factor

2.0

1.5

1.0

0.5 Isoprene Monoterpenes/oxVOC Methylbutenol

0.0 280

290

300

310

Temperature (K) Figure 3.2 (a) Light dependencies of biogenic VOC emissions. (b) Temperature dependencies of biogenic VOC emissions.

production. The mechanisms and functions of these two categories of terpene-emitting plants – temperature-dependent only and temperature- and light-dependent emitters – are quite different. Because the emission of light-dependent monoterpenes is similar to that of isoprene, the isoprene light parameterisation has been used to describe light-dependent emissions of monoterpenes with reasonable results (Ciccioli et al. 1999; Figure 3.2a). Another important light-dependent biogenic VOC is MBO. MBO is an oxVOC emitted from a subset of pine species that grow primarily in the forested regions of the western United States. The presence of MBO was ﬁrst noted by Goldan et al. (1993), who measured high mixing ratios over a lodgepole pine forest in Colorado, and later by Lamanna and Goldstein (2002), over a ponderosa pine forest in the Sierra Nevada. Both studies found that MBO

Biogenic VOCs

95

emissions can rival isoprene in these regions with respect to their impact on regional photochemistry. Harley et al. (1998) measured the leaf-level MBO response to light and found a similar response curve as that for isoprene (as shown in Figure 3.2a). While not speciﬁcally dependent on light, some oxVOCs can be emitted during light to dark transitions. Holzinger et al. (2000) observed bursts of acetaldehyde within minutes of the transition between light and dark environments, although the exact mechanism for this emission is still uncertain. These ﬁndings were also conﬁrmed by Karl et al. (2002) and Graus et al. (2004), who also found that oxVOCs such as hexenal and other C6 VOCs can be emitted from poplar in similar light-to-dark transitions. These emissions represent minor sources to the global budget, but may provide clues about plant metabolism, production and release of the oxVOC species.

3.2.5.2

Temperature dependencies

The temperature dependence of biogenic VOC emissions is noted in early experimental work (Dement et al. 1975) and can inﬂuence all categories of biogenic VOC emissions. Tingey et al. (1981) ﬁrst quantiﬁed how isoprene emissions increased with increasing temperature and isolated this response from the light effects described above. Subsequent studies (e.g. Guenther 1991, 1993) found that isoprene emissions increased with temperature until an optimum of 35–45◦ C (Guenther et al. 1993; Harley et al. 1996), after which emissions decreased (shown in Figure 3.2b). Initially, it was thought that this temperature drop-off was due to destruction of the isoprene synthase enzyme (Monson et al. 1992), yet recent investigations indicate that this temperature decrease may be a regulatory mechanism (Singsaas and Sharkey 2000). Longer-term effects of temperature on isoprene emissions have also been measured. For example, Sharkey et al. (1999) found that the temperature of the previous 2 days can affect the basal emission rate (BER, the emission rate at a speciﬁc reference temperature and PAR) of a particular tree species, and later work by Petron et al. (2001) found that the temperature of the previous several days can alter the temperature optimum as well as the BER. As described in Section 3.2.2.2, non-light-dependent monoterpenes emitters have a different storage and release mechanism than isoprene and light-dependent monoterpene emitters; therefore, their temperature dependencies are quite different. As shown in Figure 3.2b, temperature-dependent monoterpene emitters have emission rates that increase exponentially with increasing temperature. This is because monoterpenes are stored in pools within the plant, and therefore their emission is linked to their volatility and Henry’s law constant. The temperature dependence of oxVOC and some VR-BVOC emissions have been measured for a few plant species and ecosystems. MBO emissions exhibit a similar temperature dependence as that of isoprene (shown in Figure 3.2b). Harley et al. (1998) observed that leaf-level emissions of MBO increase with temperature until they reach an optimum at approximately 35–40◦ C, after which emissions decrease. As with isoprene, longer-term temperature effects have also been noted for emissions of MBO (Gray et al. 2004; Schade and Goldstein 2001). Schade and Goldstein (2001) noted that most of the oxVOCs emitted from a ponderosa pine plantation were exponentially dependent on temperature and they derived temperature response factors for the measured compounds. At this time, oxVOC emissions are typically represented using a terpene-like temperature dependence in global

96

Volatile Organic Compounds in the Atmosphere

and regional emission models, as shown in Figure 3.2b (e.g. Potter et al. 2003; Schade and Goldstein 2001). VR-BVOCs such as sesquiterpenes are expected to also have strong temperature dependence, and analytical techniques to quantify this relationship are currently under development in several laboratories.

3.2.5.3

Moisture dependencies

Monoterpene and some oxVOC emissions have been found to respond to changing moisture conditions. Emissions of monoterpenes generally increase during and following rain events (Helmig et al. 1998a; Schade et al. 1999) or while foliage is wet (Lamb et al. 1985). Other studies have found positive correlations between ambient relative humidity and monoterpene emissions (Dement et al. 1975; Schade et al. 1999). Measurements of acetone ﬂuxes have been shown to increase with increasing humidity, and ethanol varies as a function of the ambient relative humidity, although the observed relationship may have been driven by changes in stomatal conductance as a function of vapour pressure deﬁcit (Schade and Goldstein 2001).

3.2.5.4

CO2 effects

The inﬂuence of atmospheric CO2 mixing ratios on biogenic VOC emissions is an interesting topic in the context of increasing mixing ratios in the atmosphere. However, experimental results have often yielded conﬂicting results, indicating the need for further research. Laboratory measurements found that short-term exposure to elevated CO2 reduced isoprene emissions (Loreto and Sharkey 1990; Monson 1989) or had no effect (Loreto et al. 1996a). However, longer-term exposure to high CO2 conditions has not yielded consistent results. Sharkey et al. (1991) found that elevated CO2 reduced isoprene emissions from aspen trees, while Staudt et al. (2001) found an increase in isoprene emissions in oak trees under high CO2 . Two recent studies performed in the Biosphere II, a large-scale environmentally controlled experimental chamber, also found a negative feedback between isoprene emission and increasing CO2 . Rosenstiel et al. (2003) found isoprene emissions decreased under elevated CO2 , and Pegoraro et al. (2004) conﬁrmed this isoprene reduction when plants were operating at normal conditions. Constable et al. (1999) found that monoterpene emissions were unaffected by increasing CO2 mixing ratios, yet Loreto et al. (2001a) found signiﬁcant reductions in monoterpene emissions under the presence of doubled CO2 due to decreased enzyme activity. In forest measurements performed in the vicinity of a natural CO2 spring, Rapparini et al. (2004) found that leaves growing under ambient, elevated CO2 conditions were not inhibited, though leaves subject to a spike in CO2 mixing ratios were reduced.

3.2.5.5

Seasonal variations (phenology)

In addition to the short-term responses of emissions to light, temperature and moisture, the seasonal cycle of vegetation can also inﬂuence biogenic VOC emissions. The physical, biological and chemical conditions of foliage change over the span of a growing season, and these changes (known as phenological changes) can impact biogenic VOC emissions. Biogenic VOC emissions respond to seasonal factors such as the outbreak, growth, aging and loss of foliage. An obvious example is the ﬂowering of plants (Arey et al. 1991;

Biogenic VOCs

97

Ciccioli et al. 1999) and the related release of a wide variety of biogenic VOCs, which humans can detect without sophisticated instrumentation. Isoprene emissions change on phenological timescales, particularly due to the growth, aging and senescence of leaves. In temperature and boreal deciduous forests, isoprene emissions begin about 2–4 weeks after budbreak, require an additional 2–6 weeks to reach their maximum emission capacity and then decline rapidly during leaf senescence (Fuentes et al. 1999; Goldstein et al. 1998; Monson et al. 1994). This delayed timing of emissions has been attributed to a minimum temperature threshold (Monson et al. 1994) as well as isoprene synthase activity (Lehning et al. 2001). Another important factor with respect to phenology is that of leaf age. Many coniferous trees retain leaves longer than the typical growing season as deﬁned by deciduous forests, and recent experimental work has indicated that leaf age can be important for some monoterpene and oxVOC emissions. While emissions of monoterpenes were unaffected by leaf age, Loreto et al. (2001a) did notice a greater storage of monoterpenes in primary vs secondary needles. MBO emissions are observed to decrease with increasing leaf age (Gray et al. 2004; Harley et al. 1998; Schade et al. 2000). Other studies indicate that other oxVOCs such as methanol may have decreasing emissions with increasing leaf age (Nemecek-Marshall et al. 1995). In a tropical forest in Amazonia, Kuhn et al. (2004) measured isoprene and monoterpene ﬂuxes and found that despite little change in biomass and instantaneous environmental conditions, there were large changes in emissions between the wet and dry seasons. They attribute this change in emissions to possibilities such as leaf maturation, changes in leaf growth environment, variable water and nutrient availability, and oxidative capacity of the canopy air. These observations illustrate that we still have much to learn regarding the phenological factors regulating emissions of biogenic VOCs, particularly in tropical environments.

3.3 3.3.1

Emission inventories of biogenic VOCs History of biogenic VOC emission inventories

Since the earliest reports of isoprene emissions from vegetation, there has been signiﬁcant interest in quantifying the magnitude of biogenic VOC emissions. The ﬁrst published regional-scale emission inventory was by Rasmussen (1970, 1972), who screened plants to determine the range and magnitude of isoprene emitters. Later measurements and regional inventories were performed by Zimmerman (1979) in the Southeastern United States, Winer (1983) in the state of California, and Lamb et al. (1985) in the northeastern United States. These plant enclosure procedures were used to estimate emission rates for various tree species and implemented with foliar biomass data to calculate biogenic emissions over regional scales. These studies laid the groundwork for estimating biogenic emissions over larger spatial scales. Continental-scale estimates were then published for the United States (Lamb et al. 1987), and a detailed model known as BEIS (Biogenic Emissions Inventory System) was developed for use with regional air quality models (Geron et al. 1994; Pierce and Waldruff 1991). Similar types of inventories have been developed for Great Britain (Stewart et al. 2003), Europe (Simpson et al. 1995) and East Asia (Klinger et al. 2002; Steiner et al. 2002). Early

98

Volatile Organic Compounds in the Atmosphere

global estimates were developed in several studies (Allwine 1992; Dignon and Logan 1990; Mueller 1992), but the most comprehensive global model was developed by Guenther et al. (1995). These early biogenic VOC inventories were focused on isoprene and monoterpene emissions, with the addition of a separate ‘other VOC’ category. The Guenther et al. (1995) model is the most widely cited for global emissions estimates. This model estimates biogenic VOC emissions over large spatial scales, using light and temperature parameterisations of emissions based on leaf- and branch-level studies along with plant distribution estimates. Although some modiﬁcations have been made to the initial model, the core structure remains the same and this is described below in Section 3.3.2.1. Recent work has focused on reﬁning these existing algorithms and including chemical species other than isoprene and monoterpenes. For example, Otter et al. (2003) included light-dependent monoterpene emissions in a regional emissions inventory for southern Africa, and other studies have attempted to quantify global oxVOC estimates for acetone (Jacob et al. 2002; Potter et al. 2003), acetaldehyde (Fall 2003) and methanol (Galbally and Kirstine 2002; Heikes et al. 2002). Table 3.2 delineates the various estimates of global biogenic VOC emissions, and Section 3.3.3 discusses how these inventories have changed over time.

3.3.2

Methodology of estimating emission inventories

At present, there are two different general modelling methods used to estimate biogenic VOC emissions. The ﬁrst is an empirical method developed by Guenther et al. (1995), hereinafter referred to as the Guenther 95 (G95) method. This is a widely accepted and utilised model in many studies of regional and global atmospheric chemistry. The emissions estimates correspond fairly well with measurements of emissions, although there are many regions of the Earth for which the model has not been tested against observations. The second method is a physiologically based or mechanistic model that can respond to potential changes in the vegetation environment. While the physiologically based models are still unreliable on larger scales, they may provide a useful tool in the future to investigate changing emissions under different scenarios of global change.

3.3.2.1

The G95 method

The G95 global model is based on the experimental measurements of Guenther et al. (1991, 1993) and regional emission models used in the United States (Geron et al. 1994; Pierce and Waldruff 1991). Using measurements made in various ecosystems throughout the world, the model estimates a global distribution of isoprene, monoterpene and other VOC emissions. Emission ﬂuxes (F ) are calculated as follows: F = BER × D × γ

(3.2)

where BER is the basal emission rate (g VOC emitted/g biomass measured at standard light and temperature conditions), D is the foliar density of the vegetation (g biomass/m2 ground area) and γ is an activity factor that accounts for light, temperature and leaf age effects. We describe each of these factors below in brief detail. Foliar density (D). Biogenic VOC emissions are extremely dependent on the land cover; therefore, the geographic distribution of vegetation is a crucial portion of the inventory.

Biogenic VOCs

99

Table 3.2 A history of global biogenic VOC estimates Category

Global source (Tg C/year)

Year

Reference

432 480 827 491 1 150 312–1 062

1965 1968 1992 1992 1995 2004

Rasmussen and Went 1965 Robinson and Robbins 1968 Allwine 1992 Mueller 1992 Guenther et al. 1995 Wiedinmyer et al. 2004

350 452 450 175 285 420 250 503 559 250–750

1979 1988 1990 1990 1991 1992 1992 1995 2001 2004

Zimmerman 1979 Rasmussen and Khalil 1988 Dignon and Logan 1990 Taylor et al. 1990 Turner et al. 1991 Allwine 1992 Mueller 1992 Guenther et al. 1995 Potter et al. 2001 Wiedinmyer et al. 2004

Monoterpenes

480 143 128 147 127

1979 1990 1992 1992 1995

Zimmerman 1979 Taylor et al. 1990 Allwine 1992 Mueller 1992 Guenther et al. 1995

Total other VOCs

279 94 520

1992 1992 1995

Allwine 1992 Mueller 1992 Guenther et al. 1995

28(19–47) 37.5(14–80)b 100 39–117 48

2000 2002 2002 2002 2004

Singh et al. 2000 Galbally and Kirstine 2002 Heikes et al. 2002 Tie et al. 2003 Singh et al. 2004

Total BVOC

Isoprene

Oxygenated VOCs Methanola

Ethanola

3(2–4)

2004

Singh et al. 2004

Acetonea

6(3–11)c 9(6–12)c 20(15–26)d 33–107e 31(15.5–46.5)c

1994 2000 2002 2003 2004

Singh et al. 1994 Singh et al. 2000 Jacob et al. 2002 Potter et al. 2003 Singh et al. 2004

19(11–27)

2004

Singh et al. 2004

Acetaldehydea

a Values converted to Tg C/year. b Emissions from higher plants only. c Primary biogenic emissions only. d Emissions from terrestrial vegetation (based on the best estimate). e Emissions from live vegetation only.

100

Volatile Organic Compounds in the Atmosphere

Typically, the land cover information required includes (a) the type of land cover present and (b) the density and amount of biomass present. The level of detail available for these estimates can vary from urban-scale high-resolution inventories to global-scale coarse-resolution estimates. For example, urban-scale air quality studies in Los Angeles have used highresolution land cover data to estimate emission inventories at the scale of 1 km (e.g. Scott and Benjamin 2003), while the G95 model used a 0.5 degree resolution to determine biogenic VOC emissions at the global scale. Once a land cover database with appropriate resolution has been selected, the amount of biomass can be estimated from a variety of scale-dependent sources. Urban- and local-scale models can use a detailed forestry dataset that includes speciﬁc information about the types of species and amount of biomass present (e.g. Geron et al. 1994, for the eastern United States, and Solmon et al. 2004, for France). At a global level, more generalised ecosystem data is utilised (Guenther et al. 1995). The approach for estimating foliar density has changed as ground-based and satellite data describing the Earth’s surface becomes more readily available. Early emissions estimates at the regional scale were usually determined from ground-based, forest-speciﬁc data. For example, Lamb et al. (1987) and Geron et al. (1994) used general biomass estimates for D based on the vegetation species type and the class of emitter (i.e. high, low or non-emitter of isoprene). The ﬁrst global model, G95, calculated D from net primary productivity (determined from algorithms based on the temperature and precipitation per model grid cell; see Lieth 1975) and foliar density estimates by Box (1981). Recent regional scales models for California (e.g. Scott and Benjamin 2003) characterised foliar density as a product of two terms: speciﬁc leaf weight (g biomass/m2 leaf area) and the leaf area index (m2 leaf area/m2 ground area). The speciﬁc leaf weight is assumed to remain relatively constant by species and is determined from a database of speciﬁc leaf weights by vegetation species or ecosystem type. Leaf area index can be estimated at high spatial resolutions by satellites and can capture the seasonal changes in biomass. Basal emission rate. The second key piece of information for models is the amount of biogenic VOCs emitted from various types of vegetation. This is scaled from a term known as the emission factor or BER and is usually expressed as the mass of emissions per mass of leaf biomass per time (e.g. μg c g−1 h−1 ). In the G95 model, it is assumed that each plant has a relatively stable BER that can be determined by a series of measurements on individual plant species. BERs are typically measured by branch or leaf enclosures at a standard temperature and light level, and measurement techniques for determining the BER are described in Section 3.6. Over the past several decades, a large database of isoprene and monoterpene BERs has been developed from measurements taken throughout the world, and these are summarised by Wiedinmyer et al. (2004). However, the range of vegetation species for which BERs have been measured represents only a small fraction of the potential emitters. Other biogenic VOC categories are largely uncharacterised elsewhere on the globe. For other oxVOC and VR-BVOC species, very few measurements have been performed, and large-scale estimates using the G95 modelling approach are not possible at this time. Interestingly, recent studies are ﬁnding increasing evidence for the variability of BERs. Early studies noted that longer-term environmental changes can affect the BER on a seasonal scale (Goldstein et al. 1998; Monson et al. 1994) as well as shorter-term changes such as

Biogenic VOCs

101

temperature conditions on the order of several days (Petron et al. 2001; Sharkey et al. 1999). Geron et al. (2000b) found that the BER changed over the course of a growing season in Quercus alba L. (white oak) leaves, as well as changing up to a factor of two over the diurnal cycle. Funk et al. (2003) found that diurnal BER changes varied between species and Staudt et al. (2004) found that there can be signiﬁcant intra- and interspeciﬁc variability in the BER. All of these ﬁndings indicate that the BER can vary over spatial and temporal scales, and this increases uncertainty in the G95 modelling method. Because of the large number of measurements required for determining BERs for species or ecosystems, the possibility of taxonomic relationships to categorise emitters vs nonemitters has been explored as a way to circumvent this problem. In the past, taxonomic categorisation of emission rates has been used (e.g. Guenther 1994; Rasmussen and Khalil 1998) and conﬁrmed vs measurements (He et al. 2004; Karlik and Winer 2001), and this method is currently the only reasonable approach for regions with high numbers of unmeasured species. For example, Harley et al. (2004) screened 125 tree species in Amazonia and separated tree species into ‘isoprene-emitting’ or ‘non-emitting’ categories. They then used a taxonomic method to predict whether an unscreened tree species emits isoprene. However, as noted by Kesselmeier and Staudt (1999), there are limitations to this method. One obvious example is that of the genus Quercus (oak), where some species emit isoprene, some species emit monoterpenes, and others are non-emitters (Harley et al. 1999). Therefore, while this method may be helpful for poorly characterised geographic regions, there is a high potential for error. Activity factors (γ ). The activity factors utilise a standardised BER and correct for instantaneous effects such as light and temperature, as well as a range of longer-term effects such as seasonality, long-term T effects and leaf age. The activity factor is composed of several environmental factors: γ = γL γT γA

(3.3)

where γL is the light correction, γT is the temperature correction, and γA is the leaf age correction (Guenther et al. 2000). The instantaneous effects of light and temperature on emissions are described in detail in Section 3.2.5, and their proﬁles are shown in Figure 3.2. For light-dependent emissions, this curve is parameterised for isoprene, MBO and lightdependent monoterpene emissions as: αCL Q γL = 1 + α2Q 2

(3.4)

where Q is the amount of incoming light as PAR, and α and CL are empirical coefﬁcients (Guenther et al. 1993, 1995), For light- and temperature-dependent emissions such as isoprene and MBO, emissions increase until a temperature optimum and then decrease: γT =

Eopt CT 2 e(CT 1 x) CT 2 − CT 1 (1 − eCT 2 x )

(3.5)

where x = [(1 − Topt ) − (1/T )]/R, Eopt is the maximum normalised emission capacity, T is the leaf temperature, Topt is the temperature at which Eopt occurs,

102

Volatile Organic Compounds in the Atmosphere

and CT 1 and CT 2 are empirical coefﬁcients based on the activation and deactivation energies (Guenther et al. 1993, 1999b). Temperature-only-dependent emissions, such as monoterpenes and other oxVOCs, are parameterised as follows: γT = eβ(T −Ts )

(3.6)

where β is an empirical coefﬁcient, T is the leaf temperature and Ts is the standard temperature of 303 K. The leaf age factor, γA , was added to the algorithm by Guenther et al. (1999b) and deﬁnes Eopt and Topt of equation (1988) as a function of the previous days’ temperature. These factors account for the instantaneous impact of environmental conditions, although it is thought that longer-term effects of light, temperature or leaf age could play an important role. Geron et al. (2000b) and Petron et al. (2001) added an additional factor to Equation (3.5) to account for the longer-term effects of temperature on emissions.

3.3.2.2

Process-based modelling

For estimating the impacts of climate change, it is unclear whether an empirical model such as G95 will be able to accurately assess the effects of a changing environment on biogenic VOC emissions. Several recent studies have attempted to develop a physiologically based method for estimating biogenic VOC emissions. To date, these models have focused on isoprene emissions and the biochemical or biophysical pathway of isoprene production. At present, there are three different process-based models based on recent discoveries of the biochemical pathway of isoprene production. While these models are not widely accepted or applied at the global scale, they warrant mention because they may provide a pathway for future biogenic VOC modelling under different scenarios of climate change. Niinemets et al. (1999) developed an isoprene emission model based on the energy requirements for isoprene synthesis, which links the photosynthetic electron transport rate to isoprene emission rates. A second process-based modelling approach is that of Zimmer et al. (2000), which models the biosynthetic process of isoprene with the MEP pathway (see Section 3.2.2.1 and Figure 3.1). A detailed photosynthesis model is used along with kinetic parameters for each synthesis step, based on Michaelis–Menten enzyme kinetics. Martin et al. (2000) developed the third process-based model that is also grounded in the MEP pathway for isoprene synthesis. The model determines isoprene emission rates based on the three potentially limiting steps of synthesis: the pyruvate (or carbon) supply, the ATP or energy supply for phosphorylation to DMAPP, and the rate of isoprene synthesis from DMAPP.

3.3.3

Changing inventories and the current state of knowledge

In early biogenic VOC inventories, rough estimates were made on global or regional scales in attempts to quantify the magnitudes of biogenic VOC emissions. However, as atmospheric chemistry models have become increasingly sophisticated, the speciation and emission distribution of these compounds has become a more important input parameter. Table 3.2 shows the history of global biogenic VOC emissions inventories and their variations over time (Dignon and Logan 1990; Galbally and Kirstine 2002; Guenther et al. 1995; Heikes et al. 2002; Jacob et al. 2002; Mueller 1992; Palmer et al. 2003; Potter et al. 2001, 2003;

Biogenic VOCs

103

Rasmussen and Khalil 1988; Rasmussen and Went 1965; Robinson and Robbins 1968; Singh et al. 1994, 2000, 2004; Taylor et al. 1990; Tie et al. 2003; Turner et al. 1991; Went 1960; Wiedinmyer et al. 2004; Zimmerman 1979). One of the more notable changes in the past decade is the addition of the biogenic oxVOC species. Work by Singh et al. (2000, 2001) highlights the importance of oxVOC emissions for atmospheric chemistry, and subsequent studies have attempted to constrain atmospheric budgets of acetone, methanol and acetaldehyde. One example is the addition of MBO in the state of California. MBO can be emitted in large quantities and has proven to be important in the modelling of atmospheric chemistry in the regional level (Scott and Benjamin 2003). In response to recent and ongoing research, a change that is likely over the coming decades is the inclusion of VR-BVOC emissions in global inventories. Another new development is the use of satellite-derived products to estimate and validate emissions inventories. Palmer et al. (2003) estimated isoprene emission distributions over North America using satellite-derived formaldehyde columns and compared these emissions with modelled isoprene emissions based on the G95 method. This provides an interesting new method to constrain the global budgets of some biogenic VOC emissions.

3.4 3.4.1

Global distribution of biogenic VOCs Lifetimes

Once emitted to the atmosphere, sinks for biogenic VOCs include dry and/or wet deposition and atmospheric reactions. Biogenic VOCs are quickly oxidised by OH, O3 and/or NO3 (and occasionally chlorine atoms). The presence of double bonds in the chemical structure of biogenic VOCs causes high tropospheric reactivity and makes chemical reactions their dominant loss mechanism. Table 3.3 lists the calculated atmospheric lifetimes for many biogenic VOC species derived from laboratory-measured reaction rate constants and typical mixing ratios of OH, O3 and NO3 (Atkinson and Arey 2003; Atkinson et al. 1990, 1995, 1999; Calvert et al. 2000; Corchnoy and Atkinson 1990; Grosjean and Grosjean 1994; Meyrahn et al. 1986; Papagni et al. 2001; Reissell et al. 2001; Rudich et al. 1996; Smith et al. 1996). Isoprene lifetimes with respect to OH and NO3 are on the scale of hours and with O3 on the scale of days. Monoterpenes generally react more quickly than isoprene, with lifetimes with respect to OH and NO3 on the scale of minutes to hours and with O3 on the scale of minutes to days. Similar to monoterpenes, sesquiterpenes generally have lifetimes of minutes to hours. OxVOCs have much more variable lifetimes, ranging from minutes for compounds such as linalool to days for a compound such as acetone. VR-BVOCs have lifetimes that are estimated to be on the order of minutes or less. Because of their high reactivity, biogenic emissions can be important sinks for radicals in the atmosphere (Fehsenfeld et al. 1992). Isoprene and its oxidation products are generally found to be the dominant sink for OH over forested regions, accounting for up to 15–60% of the daytime removal of OH, depending on location (Carlslaw et al. 2000; Lamanna and Goldstein 2002; Millet et al. 2005). Additionally, recent studies have shown that in areas under urban inﬂuence, nighttime NO3 chemistry can play an important role in ozone formation, and biogenics can act to reduce NO3 mixing ratios as well (Warneke et al. 2002).

104

Volatile Organic Compounds in the Atmosphere

Table 3.3 Calculated atmospheric lifetimes of biogenic volatile organic compounds Lifetimea for reaction with

Biogenic VOC

O3c

OHb

NO3d

Isoprene

1.4 h

1.3 days

1.6 h

Monoterpenes Camphene 2-Carene 3-Carene Limonene Myrcene cis-/trans-Ocimene α-Phellandrene β-Phellandrene α-Pinene β-Pinene Sabinene α-Terpinene γ -Terpinene Terpinolene

2.6 h 1.7 h 1.6 h 49 min 39 min 33 min 27 min 50 min 2.6 h 1.8 h 1.2 h 23 min 47 min 37 min

18 days 1.7 h 11 h 2.0 h 50 min 44 min 8 min 8.4 h 4.6 h 1.1 days 4.8 h 1 min 2.8 h 13 min

1.7 h 4 min 7 min 5 min 6 min 3 min 0.9 min 8 min 11 min 27 min 7 min 0.5 min 2 min 0.7 min

Sesquiterpenes β-Caryophyllene α-Cedrene α-Copaene α-Humulene Longifolene

42 min 2.1 h 1.5 h 28 min 2.9 h

2 min 14 h 2.5 h 2 min >33 days

3 min 8 min 4 min 2 min 1.6 h

61 days (Atkinson et al. 1999) 2.5 days (Reissell et al. 2001) 1.0 days (Corchnoy and Atkinson 1990) 1.3 h (Atkinson et al. 1995) 1.8 h (Atkinson et al. 1995) 52 min (Atkinson et al. 1995) 12 days (Atkinson et al. 1999) 2.4 h (Papagni et al. 2001) 53 min (Smith et al. 1996)

>4.5 yearf

>8 year (Atkinson et al. 1999) >300 days (Reissell et al. 2001) 1.5 year (Corchnoy and Atkinson 1990) 4.1 h (Atkinson et al. 1995) 4.5 h (Atkinson et al. 1995) 6 min (Atkinson et al. 1995) 2.0 year (Atkinson et al. 1999) 7.7 days (Rudich et al. 1996) 9 min (Smith et al. 1996)

Oxygenated VOCs Acetonee Camphor 1,8-Cineole cis-3-hexen-1-ol cis-3-hexenyl acetate Linalool Methanol MBO 6-methyl-5-hepten-2-ol

>235 days (Reissell et al. 2001) >110 days (Atkinson et al. 1990) 6.2 h (Atkinson et al. 1995) 7.3 h (Atkinson et al. 1995) 55 min (Atkinson et al. 1995) >4.5 yearf 1.7 days (Grosjean and Grosjean 1994) 1.0 h (Smith et al. 1996)

Source: Reprinted from Atmospheric Environment, 37: S197–219. Copyright (2003), with permission from Elsevier. a From Calvert et al. (2000) unless noted otherwise. b Assumed OH radical concentration: 2.0 × 106 molecules/cm3 , 12-h daytime average. c Assumed O concentration: 7 × 1011 molecules/cm3 , 24-h average. 3 d Assumed NO radical concentration: 2.5 × 108 molecules/cm3 , 12-h nighttime average. 3 e Photolysis will also occur with a calculated photolysis lifetime of ∼60 days for the lower troposphere, July, 40◦ N Meyrahn et al. (1986). f Estimated.

Biogenic VOCs

105

Similarly, oxVOCs can be important for controlling radical mixing ratios. Lamanna and Goldstein (2002) estimate that MBO accounts for 21% of OH loss over conifers in the Sierra Nevada of California. While biogenic VOCs are normally considered to be sinks for atmospheric radicals, their oxidation in the atmosphere can also provide a new source of radicals. Reactions of ozone with alkenes including terpenes can have high yields of OH radicals (Paulson and Orlando 1996; Paulson et al. 1999). Additionally, photolysis of some oxVOCs such as acetone and acetaldehyde in the upper troposphere can also lead to production of HOx radicals (Wennberg et al. 1998). These impacts are discussed further in Section 3.5.2.

3.4.2

Source distribution

Because of their high reactivity, atmospheric mixing ratios of biogenic VOCs are highly variable and usually concentrated near the emission source. As described in Section 3.3, the source distribution of biogenic VOCs is primarily dependent on the land cover at the Earth’s surface. Emissions are based on the type of vegetation and the amount of biomass present at each location, causing a wide range of temporal and spatial variability in emissions. To give a general idea of the distribution of these source emissions, global isoprene and monoterpene emissions in July (generally considered the global seasonal maximum) are shown in Figures 3.3 and 3.4 as modelled using the G95 algorithm (Guenther et al. 1995). Isoprene is concentrated in regions with large amounts of deciduous forests, such as the southeastern United States, Europe, central Africa, and the tropical regions of Eurasia. Monoterpene emissions, which are considered to be primarily coniferous, are greatest in the western United States, Scandinavia and the forests of Russia. Global distributions of other classes of compounds or VR-BVOCs are still at an early stage of investigation, and thus current global estimates are not available.

3.4.3

Mixing ratio distributions

The mixing ratio distributions in the atmosphere result from the balance between the sources and sinks of biogenic VOC emissions. We have discussed the source distribution of emissions, as well as the reaction with radicals as a major sink for biogenic VOCs. Because of their fast reactivity, transport and deposition processes play minor roles in the global mixing ratio distribution for most biogenic VOCs. Therefore, horizontal surface mixing ratio distributions generally mirror that of the source emissions shown in Figures 3.3 and 3.4. Temporally, different biogenic VOC species have varying diurnal proﬁles. Isoprene mixing ratios have a strong diurnal cycle with daytime values ranging from 0 to 10 ppb (depending on vegetation sources) peaking at midday, and nighttime values on the order of several to hundreds of ppt. Monoterpenes have a very different diurnal cycle primarily when their emissions are independent of light. Mixing ratios at night can range up to 3 ppb near source regions and are generally higher than daytime because they can accumulate in a shallower nocturnal boundary layer (Biesenthal et al. 1998; Hakola et al. 2000; Lamanna and Goldstein 2002). The diurnal behaviour of oxVOC is highly variable depending on the emission species. Light- and temperature-dependent emissions such as MBO have a similar

0

00

–2

50

0

0

00 –2

00 15

20

0

50

25

–1

–1 12

50

00 10

0–

00

10

0

50 0–

25

50

25

10

0–

0–

10

0

Volatile Organic Compounds in the Atmosphere

0

106

00 50 0– 10 10 00 00 –1 25 0

25 0– 5

25 0

10 0–

0–

10 0

Figure 3.3 Global distribution of isoprene ﬂuxes in July, in mg C/m2 , based on the GEIA database (Guenther et al. 1995).

Figure 3.4 Global distribution of monoterpene ﬂuxes in July, based on the GEIA database (Guenther et al. 1995).

Biogenic VOCs

107

diurnal cycle to that of isoprene, with peak daytime values reaching 2–5 ppb at midday (Schade and Goldstein 2001). Other biogenic oxVOCs such as acetone and acetaldehyde often have daytime values of approximately several ppb, with mixing ratios that are slightly higher during the daytime than in the nighttime. Methanol typically has mixing ratios about a factor of ﬁve higher than other oxVOCs in forested regions (Goldan et al. 1995b; Schade and Goldstein 2001). VR-BVOC mixing ratios have been measured although they are often quite low (on the order of several ppt) due to their high reactivity. Recent work by Holzinger et al. (2004) has found that these emissions may in fact be higher than previously assumed, based on the measurements of their oxidation products and their rapid oxidation in the forest canopy. Vertical variations can also affect the spatial mixing ratio distribution, depending on their transport and fate in the atmosphere. Because of their high reactivity, biogenic VOCs generally do not escape the boundary layer. Based on observations of vertical proﬁles of biogenic VOCs at various sites, mixing ratios of isoprene and monoterpenes generally decrease rapidly with increasing heights (Andronache et al. 1994; Helmig et al. 1998b; Warneke et al. 2001). Many vertical proﬁles of oxVOC species have now been performed, yet these compound distributions are more complicated than isoprene and terpenes because of the combination of primary and secondary anthropogenic sources in addition to primary and secondary biogenic sources. Vertical distributions over the Atlantic (Singh et al. 2000) showed decreasing mixing ratios with increasing altitude for acetone and methanol, while in the Paciﬁc (Singh et al. 2001), mixing ratios of methanol (0.9 ppb), acetone (350 ppt) and acetaldehyde (60–100 ppt) were relatively constant with height. Scheeren et al. (2003) measured several oxygenated species in and above the boundary layer over the Eastern Mediterranean, and showed that acetone and methanol (approximately several ppb up to 6 km) decrease with increasing height though much less rapidly than isoprene and monoterpenes, as expected due to their longer lifetimes.

3.5

Impact on photooxidants and atmospheric chemistry

The work of Haagen-Smit in the 1950s determined that photochemical smog was produced by a series of photochemical reactions involving hydrocarbons and nitrogen oxides in the presence of sunlight. In the initial attempts at ground-level ozone control in the 1970s, this led to the control of both nitrogen oxide and anthropogenic hydrocarbon emissions. However, research in the 1980s indicated that in the presence of sufﬁcient NOx , biogenic VOCs can play an important role in the formation of ground-level ozone in both rural and urban regions (Chameides et al. 1988; Trainer et al. 1987), because biogenic VOCs tend to be more reactive than their anthropogenic counterparts (Atkinson and Arey 1998) and their emission rates are relatively high compared to anthropogenic VOCs. As research has progressed, the presence of biogenic VOCs has been shown to strongly inﬂuence tropospheric ozone and NOx chemistry at the regional and global scale (e.g. Horowitz et al. 1998; Pierce et al. 1998; Poisson et al. 2000). In this section, we discuss the role of biogenic VOCs in tropospheric-gas- and aerosolphase chemistry. This topic has been reviewed extensively (e.g. Atkinson and Arey 1998, 2003; Monson and Holland 2001); thus, we provide only a brief discussion here. We discuss some of the major impacts of biogenic VOC in the troposphere, including their effects on

108

Volatile Organic Compounds in the Atmosphere

photooxidants and ground-level ozone, their inﬂuence on the NOx distribution, and the ability of their oxidation products to undergo gas-to-particle conversion, leading to the growth of secondary organic aerosols.

3.5.1

Reaction pathways

Depending on the location and time of day, biogenic VOC emissions generally react with OH or NO3 radicals, O3 or the chlorine atom. For most biogenic VOCs, the reaction with oxidants in the atmosphere occurs by either of the following two main mechanisms: 1. the addition of O3 or OH/NO3 radicals to the double carbon bond in the biogenic VOC; 2. the abstraction of an H atom from the hydrogen–carbon bond by OH or NO3 (Finlayson-Pitts and Pitts 2000). Most biogenic VOCs are likely to follow the addition mechanism rather than the abstraction mechanism, with the exception of double carbon bond aldehydes that tend to react via abstraction (Atkinson and Arey 2003). The following discussion describes the cycle of reactions shown in Figure 3.5. Once the biogenic VOC molecule has been oxidised via the addition or abstraction mechanism, Biogenic VOC

OH

R• O2

ROOH

HO2

RO2•

RO2•

NO2 ROONO2 NO

Carbonyl + Alcohol

RONO2

RO•

Decomposition

O2

Isomerisation

Products

Figure 3.5 General biogenic VOC reaction pathway. Reprinted from Atmospheric Environment, 37: S197– 219. Copyright (2003), with permission from Elsevier.

Biogenic VOCs

109

the primary product is an alkyl radical (R• ). The alkyl radical then immediately reacts with O2 to form an alkyl peroxy (RO•2 ) radical. After the alkyl peroxy radical is formed, it can react with NO, NO2 , HO2 or other alkyl peroxys (RO•2 ). Reactions with NO2 and HO2 terminate the reaction sequence by forming peroxy nitrates (ROONO2 ) and peroxides (ROOH), respectively. However, if the reaction sequence continues via reactions with NO and HO2 , they can yield other important reaction products (Seinfeld and Pandis 1998). The NO pathway is important because it can often lead to the formation of NO2 and, hence, ozone (when NO2 is photolysed forming O, then reacting with O2 to form O3 ), while the HO2 pathway is important because it forms relatively stable products and acts as a sink for radicals. Generally, if NO mixing ratios are low (e.g. in a rural environment), the cycle tends towards the HO2 /RO2 reactions (or peroxy-peroxy reactions) that remove radicals from the system and retard ozone production. Reaction of the alkyl peroxy radical with HO2 forms a peroxide (ROOH), which can then be removed via wet/dry deposition, be photolysed (regenerating OH) or react with OH itself. Reactions of RO•2 can form a variety of products including alcohols, aldehydes and ketones. However, if NO is high such as in polluted regions, then peroxy radicals will tend to react with NO or NO2 . Reactions with NO2 form peroxy nitrates (ROONO2 ), which can also act as a reservoir species for NOx as they can be stable at surface-level temperature and pressure conditions. This stability allows them to be transported for long distances to later dissociate. The formation and fate of peroxy nitrates are discussed in greater detail in Chapter 6 of this volume. Reactions of RO•2 and NO can form either an alkyl nitrate (RONO2 ) or an alkoxy radical (RO• ). The alkyl nitrate is a more stable compound and can remove NOx from the cycle. However, if the reaction forms the alkoxy radical (RO• ), it proceeds to isomerise or decompose into a variety of products. This pathway can result in the formation of NO2 , triggering the production of more ozone. The fate of NO reactions (either forming RONO2 or RO) is determined by the ‘branching ratio’, a ratio representing the likelihood of each NO pathway. An example of biogenic VOC oxidation is the reaction of isoprene and the OH radical. There are six different possible hydroxyalkyl radicals that can form from this reaction, and Figure 3.6 shows one possible reaction pathway of isoprene and OH to completion (Atkinson and Arey 2003). In the pathway shown, OH adds to the primary carbon and then reacts with NO to form an alkoxy radical (assuming sufﬁcient NO is present). OH addition on the ﬁrst or second carbon occurs in approximately 66% of the isoprene–OH reactions and results in the production of methyl vinyl ketone and formaldehyde (Finlayson-Pitts and Pitts 2000). The other third of the reactions involve the OH addition on the third or fourth carbon, producing methacrolein (not shown). The primary products of the isoprene–OH reaction are methyl vinyl ketone, methacrolein and formaldehyde. Another important biogenic VOC oxidation sequence is the α-pinene-ozone reaction, shown in Figure 3.7. This is one of the dominant loss mechanisms for α-pinene and is representative of the type of reactions typically occurring between terpenes and ozone. When ozone adds to the double bond in α-pinene, an ozonide forms that rapidly decomposes to two different Criegee biradicals (Kamens et al. 1999). One of these biradicals (Criegee 1 in Figure 3.7) then reacts to form pinonic acid, norpinonaldehyde and norpinonic acid. The second biradical (Criegee 2) forms pinic acid and several other products. The primary products of this reaction sequence can undergo gas-to-particle conversion and

110

Volatile Organic Compounds in the Atmosphere

OH

CH3

CH3

+

C H 2C

•

CH2

CH2

C

H2COH

CH

CH O2 CH3 C

H2COH

CH

• OO CH2

NO

CH3 H2COH

C

CH

ONO2 CH2 CH3 H2COH

C

CH

•

O

+

NO2

CH2

Decomposition

O •

CH2O

+

O2 HCHO

+

CH2 CH Methyl vinyl ketone H3C

C

HO2

Figure 3.6 Isoprene and OH reaction pathway. (With permission from Finlayson-Pitts and Pitts 2000.)

form secondary organic aerosols, as well as form other important gas-phase species such as carbon monoxide, formaldehyde, and OH and HO2 radicals.

3.5.2

Effects on photooxidants

The oxidation of biogenic VOCs can lead to production of tropospheric ozone via the formation of NO2 . NO2 is formed from the reaction of NO and the peroxy radical, and is quickly photolysed after formation: NO2 + hν → NO + O(3 P) M

O(3 P) + O2 −→ O3

(3.7a) (3.7b)

where M represents a third body (usually an N2 or O2 molecule) required for the termolecular reaction of 3.7b. Because the cycle with biogenic VOC regenerates radicals, the presence

Biogenic VOCs

+

111

O3 O O O

O O

O O O Criegee 1

CHO

Criegee 2

HO2 + O

COOH

Pinonic acid

O

CHO

CO OH H2

Norpinonaldehyde

HOOC

+ Other products COOH

Pinic acid

oxidant

O COOH Norpinonic acid Figure 3.7 α-Pinene and ozone reaction pathway. Reprinted from Environmental Science and Technology, 33: 1430–8. Copyright (1999), with permission from American Chemical Society.

of these molecules in conjunction with sufﬁcient NOx can increase the production of ozone in the boundary layer. This general reaction sequence represents the primary source of ozone in the troposphere. Modelling studies in the late 1980s and early 1990s indicated that biogenic VOCs, in particular isoprene, could be extremely important in estimating regional ozone in both rural and urban locations (Chameides et al. 1988; Council 1991; Trainer et al. 1987). Isoprene can strongly inﬂuence the ozone production efﬁciency (OPE), and was found to account for a 50% increase in OPE during the summertime in the northeastern United States (Hirsch et al. 1996). Regional studies in the western United States and Europe have also found that isoprene can contribute signiﬁcantly to ozone production (Dreyfus et al. 2002; Singh et al. 2004; Tsigaridis and Kanakidou 2002). Global modeling studies have also shown that ozone formation is sensitive to the presence of isoprene. Wang et al. (1998) found that the lack of non-methane hydrocarbons (including

112

Volatile Organic Compounds in the Atmosphere

both anthropogenic and biogenic) in a global photochemistry model reduced ozone mixing ratios by 5–20%, with decreased production near the surface and in the upper atmosphere (above 300 mbar) and increased production in the middle troposphere (300–600 mb), and attribute approximately half of this effect to isoprene. Houweling et al. (1998) found that all non-methane hydrocarbon emissions caused a 17% increase in the global mean tropospheric ozone column and that isoprene from biogenic sources were causing about 70% of this increase. One example of the impacts of oxygenated biogenic VOCs is that of Tie et al. (2003). This study included biogenic methanol emissions in a global tropospheric chemistry model and found that adding these emissions slightly increased O3 (1–2%). Besides the impacts on ozone, the presence of biogenic VOC can also inﬂuence other photooxidants such as OH and HO2 . Near source regions, the presence of biogenic VOCs and their fast reactions with OH can lead to a signiﬁcant sink of OH radicals. The methanol study by Tie et al. (2003) also found that biogenic methanol emissions can slightly increase HO2 mixing ratios (1–2%) and decrease OH (1–3%). This decrease in OH due to the addition of biogenic emissions has also been noted in smaller-scale, canopy-level models (Makar et al. 1999). Biogenic VOCs can also have other impacts on HOx cycling in the troposphere. While isoprene can be a sink for OH, the reaction of ozone with biogenic alkenes can also be a source of OH. This provides a potential OH source at times when mixing ratios would otherwise be negligible (e.g. nighttime) and has been found to be account for 10–15% of the total radical production in the southeastern United States (Paulson and Orlando 1996).

3.5.3

Formation of organic nitrates and sequestering of NOx

Global atmospheric chemistry studies in the late 1980s and 1990s suggested that the export of anthropogenic NOx could inﬂuence ground-level ozone formation in other regions (Jacob et al. 1993; Liu et al. 1987; Mauzerall et al. 1996). One of the methods of transporting NOx is via the formation of peroxy nitrates. As shown in Figure 3.5, peroxy nitrates are formed by the reaction of an alkyl peroxy radical (RO2 ) with NO2 and have a generalised structure of ROONO2 . These relatively stable products can then be transported long distances to be dissociated in other regions, leading to the formation of ozone far from the original sources of anthropogenic pollution. Horowitz et al. (1998) found that peroxy nitrates formed from isoprene are the principle peroxynitrates around the globe. The topic of peroxy nitrates is discussed in further detail in Chapter 6 of this volume.

3.5.4

Formation of secondary organic aerosols

Aerosols, which are liquid or solid particles suspended in the atmosphere, play an important role in the chemistry and climate of the atmosphere. They can scatter and absorb incoming radiation, affecting the radiative budget at the surface of the Earth, and provide a site for heterogeneous reactions in the atmosphere. Biogenic secondary organic aerosols (SOA) can be formed via two possible methods. The ﬁrst is when gas-phase oxidation products condense onto existing particles in the atmosphere. This occurs because some oxidation products have low volatilities, allowing them to condense on pre-existing particles while trying to establish equilibrium between the gas and particle phases. The ability of a biogenic

Biogenic VOCs

113

VOC to form SOA depends on its atmospheric abundance, its chemical reactivity and the volatility of its products (Seinfeld and Pandis 1998). The second formation mechanism is via nucleation, or the formation of new particles, although there is no solid evidence at this time that nucleation is a signiﬁcant source of aerosol to the atmosphere (Kanakidou et al. 2004). Went (1960) ﬁrst discussed the biogenic VOC contribution to haze observed over forested regions. Subsequent studies in the 1980s and 1990s found that the reaction of monoterpenes with photooxidants could produce large amounts of aerosols in smog chamber experiments (Kamens et al. 1981; Pandis et al. 1991). Later studies attributed monoterpenes and sesquiterpenes as the primary contributors to the formation of secondary organic aerosols (Grifﬁn et al. 1999a; Hoffmann et al. 1997). These initial smog chamber studies concluded that gas-to-particle conversion was not occurring for the oxidation products of isoprene. Biogenically formed aerosols can make up a large fraction of the particulate mass of the atmosphere. Pandis et al. (1991) estimated that monoterpenes could account for a range of 15–50% of the particulate matter in Los Angeles. Global inventories by Andreae and Crutzen (1997) estimated that SOA formation from biogenic VOCs was 30–270 Tg/year, while a subsequent study by Grifﬁn et al. (1999b) used the equilibrium principle in conjunction with smog chamber experiments to estimate an annual total of 18.5 Tg of aerosol from biogenic hydrocarbons. In contrast to previous studies, recent investigations have concluded that isoprene can be a precursor of secondary organic aerosols via a different mechanism than that of monoterpenes. Compounds such as carbonyls and dienes (e.g. isoprene) can undergo conversion via heterogeneous reactions in the presence of an acid catalyst such as sulphuric acid (Jang and Kamens 2001; Limbeck et al. 2003). Claeys et al. (2004a) found that biogenic aerosols in Amazonia contained large amounts of low-volatility polyols, which are likely derived from the reaction of isoprene and OH, accounting for ∼2 Tg of biogenic SOA/year. Another possible formation pathway is the reaction of isoprene with hydrogen peroxide under acidic conditions in the liquid phase (Claeys 2004b). These newly hypothesised chemical mechanisms indicate a growing understanding and the need for further research into biogenic SOA. For further information, we refer the reader to a recent review by Kanakidou et al. (2004) and to Chapter 8 of this volume.

3.5.5

Impacts of oxVOCs on atmospheric chemistry

Biogenic oxVOC emissions play a signiﬁcant role in tropospheric chemistry because: (a) they can be carriers of reactive nitrogen and sequester nitrogen in the atmosphere, and (b) they can be photolysed, causing signiﬁcant production of HOx radicals mainly in the upper atmosphere (Jaegle et al. 2001; Singh et al. 2000; Wennberg et al. 1998). Recent measurements in the remote Paciﬁc indicate that the oxVOC species are about twice as abundant as the C2 –C8 hydrocarbons (Singh et al. 2004), indicating their potential magnitude and importance in understanding atmospheric chemistry. Other measurement campaigns over continental regions have found that oxVOC species can comprise a dominant fraction of the total VOC amount and reactivity, and that a large portion of the total oxVOC is likely from biogenic emissions (Goldan et al. 1995a, 1995b; Lamanna and Goldstein 2002; Millet et al. 2005; Riemer et al. 1998).

114

Volatile Organic Compounds in the Atmosphere

Besides oxVOCs from primary sources, the oxidation of other biogenic VOCs can create signiﬁcant amounts of oxygenated products. For example, after biogenic compounds such as isoprene and monoterpenes are oxidised, they can form products such as formaldehyde, acetone and other carbonyl compounds. Globally, the oxidation of isoprene is thought to be the dominant source of formaldehyde. Additionally, secondary formation of acetone by the oxidation of isoprene or MBO can contribute signiﬁcantly to the global acetone budget.

3.6 3.6.1

Sampling and measurement techniques Mixing ratio measurement methods

Mixing ratios of biogenic VOCs can be measured using a variety of sampling techniques, including sampling air into canisters, collecting air samples on solid adsorbents and using automated in-situ instrumentation. These methods are brieﬂy reviewed here, and a thorough review of recently used methods can be found in Helmig (1999). Samples collected into canisters are generally brought back to a central laboratory facility and analysed following cryocondensation of VOCs and subsequent injection into a gas chromatograph equipped with a capillary column and one of several possible detectors. Typical detectors include the ﬂame ionisation detector (FID), quadrupole mass spectrometer, and photoionisation detector (PID). Due to the poor recovery of some biogenic VOCs from canisters, particularly oxVOCs and many terpenoid compounds, techniques have been developed for collecting air samples directly onto carbon-based solid adsorbent cartridges that are later thermally desorbed for injection into a gas chromatography system. Special considerations are required for measurement of many higher-molecular-weight compounds such as sesquiterpenes, and recent progress has been made in creating methods appropriate to these compounds (Helmig et al. 2003). Automated in-situ instrumentation has been developed and deployed to ﬁeld sites in order to create more temporally representative observational data sets of VOCs (e.g. Greenberg et al. 1994) and can be coupled to ﬂux measurement systems to measure pairs of simultaneously collected samples from gradient or relaxed eddy accumulation (REA) systems as described below in the ﬂux measurement section (e.g. Goldstein et al. 1995; Lamanna and Goldstein 2002; Schade and Goldstein 2001). Many recent advances have been made in measuring atmospheric VOCs (including biogenics), which are covered in detail in the last chapters of this book. Chapter 8 discusses gas chromatography-isotope ratio mass spectrometry, which is a technique used to measure the isotopic composition of VOC (recently reviewed by Goldstein and Shaw 2003). Chapter 9 of this volume discusses proton transfer reaction mass spectrometry (PTR-MS), which is a fast-response, high-sensitivity measurement with limited selectivity for individual compounds. This method is now widely used for measuring oxVOCs, biogenic VOC ﬂuxes and biogenic VOC oxidation products. Chapter 10 discusses multidimensional gas chromatography, which has been explored but is not yet widely used for biogenic VOC measurements. This technique is able to separate a complex VOC mixture in atmospheric samples, and could be useful for the speciation of biogenic VOCs and their oxidation products.

Biogenic VOCs

3.6.2

115

Flux measurement methods

Several methods are available to measure the emissions of biogenic VOCs from vegetation. Leaf enclosure and branch enclosure are two ﬂux methods described below to determine the emissions from a small, intact sample of plant matter. The next two ﬂux methods, REA and eddy covariance (EC), are micrometeorological methods used to determine ﬂuxes from ecosystems and generally represent ﬂuxes over an area of 105 m2 or more (Cao and Hewitt 1999). The enclosure methods have the advantage of allowing us to understand the contribution of individual species to biogenic emissions and to manipulate the plant environment to determine response to variables such as light and temperature. The micrometeorological methods provide a broader picture of the hydrocarbon ﬂuxes into and out of a whole undisturbed ecosystem, and can be deployed over longer timescales to observe the response of ﬂuxes to changing environmental conditions.

3.6.3

Leaf and branch enclosure measurements

The most direct way to measure biogenic VOC emissions is with leaf and branch chambers. This technique is commonly used to determine emission factors for biogenic VOC emission modelling purposes. Early emissions measurements were taken via branch enclosures (Lamb et al. 1985; Zimmerman 1979), which envelope an entire branch into a Teﬂon or Tedlar enclosure. Ambient air is then pumped or pulled through the enclosure and an emission rate (ER, mass of emissions per mass of biomass per hour) is determined by the difference in the inlet and outlet air mixing ratios (cin and cout ), the ﬂow rate of the air through the bag, and the biomass within the enclosure: ER =

ﬂow (cout − cin ) mass

(3.8)

ER can also be determined on a per-area basis, by dividing by the leaf area within the enclosure instead of by the mass; however, biogenic VOC emission factors are more typically reported on a per-mass basis (Guenther et al. 1995). Measurements can also be taken at the leaf level, in which a leaf is placed inside a cuvette (chamber). Typically, cuvettes are constructed with the ability to control temperature and light levels to test emissions over a range of conditions and develop emission response curves for modelling purposes. Ambient air is pumped or pulled through the cuvette, and the emission rates are determined in the same manner as Equation (3.8). Past studies have shown that leaf cuvette ER measurements are typically higher than branch enclosure measurements (Guenther et al. 1996), and this has been attributed to light levels and shading within the canopy. When using branch or leaf enclosure techniques, care must be taken not to damage the plant, because that often induces biogenic VOC emissions.

3.6.4

Relaxed eddy accumulation

REA is a micrometeorological method to determine biogenic VOC ﬂuxes in and out of ecosystems or landscapes on a larger scale. This method determines the biosphere–atmosphere ﬂux (F ) of biogenic VOC based on separately collecting and analysing mixing ratios of

116

Volatile Organic Compounds in the Atmosphere

the target species in updrafts (cup ) and downdrafts (cdown ) of air: F = βσw (cup − cdown )

(3.9)

where β is an empirical coefﬁcient that depends on the atmospheric stability, and σw is the standard deviation of the vertical wind speed. A sonic anemometer is used to determine vertical wind speed and direction, and air samples can be collected into canisters or cartridges for later analysis, or can be analysed by automated in-situ instrumentation. For a complete description of this and an example of its application for biogenic VOC ﬂux measurements, see Bowling et al. (1998). This method is particularly effective for measuring ecosystemscale ﬂuxes of compounds that are not amenable to current fast-response measurement methods. For compounds that can be measured with fast time response (better than 1 Hz), the EC method can be used.

3.6.5

Eddy Covariance (EC)

EC is a more direct micrometeorological method to measure biosphere–atmosphere ﬂuxes of biogenic VOCs on the ecosystem scale. Fluxes (F ) are determined from the covariance of the biogenic VOC mixing ratio and the vertical wind speeds using the following: F = Na w c

(3.10) w

is deviation from the mean vertical where Na is the number density of air molecules, wind velocity, and c is the deviation from the mixing ratio of the biogenic VOC. This method requires that the mixing ratio of biogenic VOC be measured as fast or faster than the eddies carry the ﬂux past the sensor or, in other words, faster than the vertical wind is changing directions. Because measurements for the EC method need to be instantaneous, this technique could not be employed for biogenic VOCs until recently. The ﬁrst published ﬂux measurements using the EC method biogenic VOCs utilised a fast chemiluminescence sensor to measure isoprene ﬂux (Guenther and Hill 1998), and more recent studies have used PTR-MS for monoterpene (Lee et al. 2005; Rinne et al. 2002) and oxVOC ﬂuxes (Karl et al. 2001a; Warneke et al. 2002).

3.7

Future directions

Biogenic VOC emissions are a dominant source of reactive organic gases in the atmosphere. In the 1980s and 1990s, the primary research focus was to quantify emissions and understand their role in atmospheric chemistry, with particular attention on contributions to formation of tropospheric ozone and to the oxVOC budget. However, as measurements of these compounds have progressed, categories of extremely reactive compounds are being found to be biogenic in origin (e.g. oxVOCs, VR-BVOCs). These highly reactive compounds form oxidation products in the gas phase and contribute to growth of secondary organic aerosols, which can play an important role in the interactions between chemistry and climate. The interaction between emission of very reactive compounds and their role in secondary aerosol production and growth are now a dominant focus in biogenic VOC research, and one that we believe will remain prominent in the decade to come.

Biogenic VOCs

117

References Affek, H.P. and Yakir, D. (2002) Protection by isoprene against singlet oxygen in leaves. Plant Physiology, 129: 269–77. Affek, H.P. and Yakir, D. (2003) Natural abundance carbon isotope composition of isoprene reﬂects incomplete coupling between isoprene synthesis and photosynthetic carbon ﬂow. Plant Physiology, 131: 1727–36. Allwine, G., et al. (1992) As cited in Fehsenfeld et al. (1992). Andreae, M.O. and Crutzen, P.J. (1997) Atmospheric aerosols: Biogeochemical sources and role in atmospheric chemistry. Science, 276: 1052–8. Andronache, C., Chameides, W.L., Rodgers, M.O., et al. (1994) Vertical distribution of isoprene in the lower boundary layer of rural and urban southern United States. Journal of Geophysical Research, 99 (D8): 16989–99. Arey, J., Winer, A.M., Atkinson, R., et al. (1991) The emission of (z)-3-hexen-1-ol, (z)-3-hexenylacetate and other oxygenated hydrocarbons from agricultural plant species. Atmospheric Environment, 25 (5–6): 1063–75. Atkinson, R. and Arey, J. (1998) Atmospheric chemistry of biogenic organic compounds. Accounts of Chemical Research, 31: 574–83. Atkinson, R. and Arey, J. (2003) Gas-phase tropospheric chemistry of biogenic volatile organic compounds: A review. Atmospheric Environment, 37: S197–219. Atkinson, R., Hasegawa, D. and Aschmann, S.M. (1990) Rate constants for the gas-phase reactions of O3 with a series of terpenes. Journal of Geophysical Research, 97: 6065–73. Atkinson, R., Arey, J., Aschmann, S.M., et al. (1995) Rate constants for the gas-phase reactions of cis3-hexen-1-ol, cis-3-hexenylacetate, trans-2-hexenal, and linalool with OH and NO3 radicals and O3 at 296+/−2K, and OH radical formation yields from the O3 reactions. International Journal of Chemical Kinetics, 27: 941–55. Atkinson, R., Baulch, D.L., Cox, R.A., et al. (1999) Evaluated kinetic and photochemical data for atmospheric chemistry, organic species. Journal of Physical and Chemical Referenece Data, 28: 191–393. Baker, B.M., Guenther, A., Greenberg, J., et al. (1999) Canopy ﬂuxes of 2-methyl-3-buten-2-ol over a ponderosa pine forest by relaxed eddy accumulation: Field data and model comparison. Journal of Geophysical Research, 104 (D21): 26107–44. Baker, B.M., Guenther, A. and Greenberg, J. (2001) Canopy level ﬂuxes of 2-methyl-3-buten-2-ol, acetone, and methanol by a portable relaxed eddy accumulation system. Environmental Science and Technology, 35 (9): 1701–8. Biesenthal, T.A., Bottenheim, J.W., Shepson, P.B., et al. (1998) The chemistry of biogenic hydrocarbons at a rural site in eastern Canada. Journal of Geophysical Research, 103 (D19): 25487–98. Bonsang, B., Poll, C. and Lambert, G. (1992) Evidence of marine production of isoprene. Geophysical Research Letter, 24 (21): 1129–32. Bowling, D.R., Turnipseed, A.A., Delany, A.C., et al. (1998) The use of relaxed eddy accumulation to measure biosphere-atmosphere exchange of isoprene and other biological trace gases. Oecologia, 116: 306–15. Box, E. (1981) Data base of the international biological program and other sources. In: J. Bufalini and R. Arnts (Eds) Atmospheric Biogenic Hydrocarbons. Ann Arbor Science Publishers, Ann Arbor, MI, pp. 121–48. Broadgate, W.J., Malin, G., Kupper, F.C., et al. (2004) Isoprene and other non-methane hydrocarbons from seaweeds: A source of reactive hydrocarbons to the atmosphere. Marine Chemistry, 88 (1–2): 61–73. Calvert, J.G., Atkinson, R., Kerr, J.A., et al. (2000) The Mechanisms of Atmospheric Oxidation of Aromatic Hydrocarbons. New York, NY: Oxford University Press.

118

Volatile Organic Compounds in the Atmosphere

Cao, X.-L. and Hewitt, C.N. (1999) The sampling and analysis of volatile organic compounds in the atmosphere. In: C.N. Hewitt (Ed.) Reactive Hydrocarbons in the Atmosphere. Academic Press, San Diego, CA, pp. 119–57. Carlslaw, N., Bell, N., Lewis, A.C., et al. (2000) A detailed case study of isoprene chemistry during the EASE96 Mace Head campaign. Atmospheric Environment, 34: 2827–36. Chameides, W.L., Lindsay, R.W., Richardson, J., et al. (1988) The role of biogenic hydrocarbons in urban photochemical smog: Atlanta as a case study. Science, 241: 1473–5. Ciccioli, P., Brancaleoni, E., Frattoni, M., et al. (1999) Emission of reactive terpene compounds from orange orchards and their removal by within-canopy processes. Journal of Geophysical Research, 104 (D7): 8077–94. Claeys, M., Graham, B., Vas, G., et al. (2004a) Formation of secondary organic aerosols through photooxidation of isoprene. Science, 303: 1173–6. Claeys, M., Wang, W., Ion, A.C., et al. (2004b) Formation of secondary organic aerosols from isoprene and its gas-phase oxidation products through reaction with hydrogen peroxide. Atmospheric Environment, 38: 4093–8. Constable, J.V.H., Litvak, M.E., Greenberg, J., et al. (1999) Monoterpene emission from coniferous trees in response to elevated CO2 concentrations and climate warming. Global Change Biology, 5: 255–67. Corchnoy, S.B. and Atkinson, R. (1990) Kinetics of the gas-phase reactions of OH and NO3 radicals with 2-carene, 1,8-cineole, p-cymene and terpinolene. Environmental Science and Technology, 24: 1497–1502. Council, N.R. (1991) Rethinking the Ozone Problem in Urban and Regional Air Pollution. National Academy Press, Washington, DC. DeGouw, J.A., Howard, C.J., Custer, T.G., et al. (2000) Proton-transfer chemical-ionization mass spectrometry allows real-time analysis of volatile organic compounds released from cutting and drying of crops. Environmental Science and Technology, 34 (12): 2640–8. Delwiche, C.F. and Sharkey, T.D. (1993) Rapid appearance of 13 C in biogenic isoprene when 13 C is fed to intact leaves. Plant, Cell and Environmental, 16: 587–91. Dement, W.A., Tyson, B.J. and Mooney, H.A. (1975) Mechanism of monoterpene volatilization in salvia mellifera. Phytochemistry, 14: 1555–7. Di Carlo, P., Brune, W.H., Martinez, M., et al. (2004) Missing OH reactivity in a forest: Evidence for unknown reactive biogenic VOCs. Science, 304: 722–5. Dignon, J. and Logan, J.A. (1990) Biogenic emissions of isoprene: A global inventory. EOS Transaction of the American Geophysical Union, 71: 1260. Donahue, N.M. and Prinn, R.G. (1990) Nonmethane hydrocarbon chemistry in the remote marine boundary layer. Journal of Geophysical Research, 95 (D11): 18387–411. Donahue, N.M. and Prinn, R.G. (1993) In-situ nonmethane hydrocarbon measurements on SAGA-3. Journal of Geophysical Research, 98 (D9): 16915–32. Dreyfus, G.B., Schade, G.W. and Goldstein, A.H. (2002) Observational constraints on the production of isoprene oxidation to ozone production on the western slope of the Sierra Nevada, California. Journal of Geophysical Research, 107 (D19): 4365. Eisenreich, W., Rohdich, F. and Bacher, A. (2001) Deoxyxylulose phosphate pathway to terpenoids. Trends in Plant Science, 6 (2): 78–84. Evans, R., Tingey, D. and Gumpertz, M. (1985) Interspecies variation in terpenoid emissions from Englemann and Sitka Spruce seedings. Forest Science, 31: 132–42. Fall, R. (2003) Abundant oxygenates in the atmosphere: A biochemical perspective. Chemical Reviews, 103: 2941–51. Fall, R. and Monson, R.K. (1992) Isoprene emission rate and intercellular isoprene concentration as inﬂuenced by stomatal distribution and conductance. Plant Physiology, 100: 987–992.

Biogenic VOCs

119

Fall, R., Karl, T., Hansel, A., et al. (1999) Volatile organic compounds emitted after leaf wounding: On-line analysis by proton-transfer-reaction mass spectrometry. Journal of Geophysical Research, 104 (D13): 14963–15974. Fehsenfeld, F., Calvert, J., Fall, R., et al. (1992) Emissions of volatile organic compounds from vegetation and the implications for atmospheric chemistry. Global Biogeochemical Cycles, 6 (4): 389–430. Finlayson-Pitts, B.J. and Pitts Jr., J.N. (2000) Chemistry of the Upper and Lower Atmosphere: Theory, Experiments and Applications. Academic Press, San Diego, CA. Fisher, A.J., Baker, B.M., Greenberg, J., et al. (2000) Enzymatic synthesis of methylbutenol from dimethylallyl diphosphate in needles of Pinus sabiniana. Archives of Biochemistry and Biophysics, 383: 128–34. Fuentes, J.D., Wang, D. and Gu, L. (1999) Seasonal variations in isoprene emission from a boreal aspen forest. Journal of Applied Meteorology, 38: 855–69. Fuentes, J.D., Lerdau, M., Atkinson, R., et al. (2000) Biogenic hydrocarbons in the atmospheric boundary layer: A review. Bulletin of the American Meteorological Society, 81 (7): 1537–75. Fukui, Y. and Doskey, P.V. (2000) Identiﬁcation of nonmethane organic compound emissions from grassland vegetation. Atmospheric Environment, 34 (18): 2947–56. Funk, J.L. (2003) Diurnal variation in the basal emission rate of isoprene. Ecological Application, 13 (1): 269–78. Gabriel, R., et al. (1999) Factors controlling the emissions of volatile organic acids from leaves of Quercus ilex L. (holm oak). Atmospheric Environment, 33: 1347–55. Galbally, I.E. and Kirstine, W. (2002) The production of methanol by ﬂowering plants and the global cycle of methanol. Journal of Atmospheric Chemistry, 43: 195–229. Geron, C., Guenther, A.B. and Pierce, T.E. (1994) An improved model for estimating emissions of volatile organic compounds from forests in the eastern United States. Journal of Geophysical Research, 99 (D6): 12773–91. Geron, C., Guenther, A.B., Sharkey, T.D., et al. (2000a) Temporal variability in the basal isoprene emission factor. Tree Physiology, 20: 799–805. Geron, C., Rasmussen, R., Arnts, R., et al. (2000b) A review and synthesis of monoterpene speciation from forests in the United States. Atmospheric Environment, 34: 1761–81. Goldan, P.D., Kuster, W.C., Fehsenfeld, F., et al. (1993) The observation of a C5 alcohol emission in a North American pine forest. Geophysical Research Letter, 20: 1039–42. Goldan, P.D., Kuster, W.C., Fehsenfeld, F., et al. (1995a) Hydrocarbon measurements in the southeastern United States: The Rural Oxidants in the Southern Environment (ROSE) program 1990. Journal of Geophysical Research, 100 (D12): 25945–63. Goldan, P.D., Trainer, M., Kuster, W.C., et al. (1995b) Measurements of hydrocarbons, oxygenated hydrocarbons, carbon monoxide and nitrogen oxides in an urban basin in Colorado: Implications for emissions inventories. Journal of Geophysical Research, 100 (D11): 22771–83. Goldstein, A.H. and Schade, G.W. (2000) Quanitifying biogenic and anthropogenic contributions to acetone mixing ratios in a rural environment. Atmospheric Environment, 34 (29–30): 4997–5006. Goldstein, A.H. and Shaw, S.L. (2003) Isotopes of volatile organic compounds: An emerging approach for studying atmospheric budgets and chemistry. Chemical Reviews, 103: 5025–48. Goldstein, A.H., Daube, B.C., Munger, J.W., et al. (1995) Automated in-situ monitoring of atmospheric non-methane hydrocarbon concentrations and gradients. Journal of Atmospheric Chemistry, 21: 43–59. Goldstein, A.H., Goulden, M.L., Munger, J.W., et al. (1998) Seasonal course of isoprene emissions from a midlatitude deciduous forest. Journal of Geophysical Research, 103 (D23): 31045–56. Goldstein, A.H., McKay, M., Kurpius, M.R., et al. (2004) Forest thinning experiment conﬁrms ozone deposition to forest canopy is dominated by reaction with biogenic VOCs. Geophysical Research Letter, 31 (22): L22106.

120

Volatile Organic Compounds in the Atmosphere

Graus, M., Schnitzler, J.-P., Hansel, A., et al. (2004) Transient release of oxygenated volatile organic compounds during light-dark transitions in grey poplar leaves. Plant Physiology, 135: 1967–75. Gray, D.W., Lerdau, M., Goldstein, A.H. (2004) Inﬂuences of temperature history, water stress, and needle age on methyl butenol emissions. Ecology, 84 (3): 765–76. Greenberg, J., Lee, B., Helmig, D., et al. (1994) Fully-automated gas chromatograph-ﬂame ionization detector system for the in situ determination of atmospheric non-methane hydrocarbons at low parts per trillion concentration. Journal of Chromatography, 676: 389–98. Grifﬁn, R.J., Cocker, D.R.I., Flagan, R.C., et al. (1999a) Organic aerosol formation from the oxidation of biogenic hydrocarbons. Journal of Geophysical Research, 104 (D3): 3555–67. Grifﬁn, R.J., Cocker, D.R.I., Seinfeld, J.H., et al. (1999b) Estimate of global atmospheric organic aerosol from oxidation of biogenic hydrocarbons. Geophysical Research Letter, 26 (17): 2721–4. Grosjean, E. and Grosjean, D. (1994) Rate constants for the gas-phase reactions of ozone with unsaturated aliphatic alcohols. International Journal of Chemical Kinetics, 26: 1185–91. Guenther, A., Zimmerman, P. and Wildermuth, M. (1994) Natural volatile organic-compound emission rate estimates for United States woodland landscapes. Atmospheric Environment, 28 (6): 1197–210. Guenther, A., Geron, C., Pierce, T.E., et al. (2000) Natural emissions of non-methane volatile organic compounds, carbon monoxide and oxides of nitrogen from North America. Atmospheric Environment, 34 (12–14): 2205–30. Guenther, A.B. and Hill, A.J. (1998) Eddy covariance measurement of isoprene ﬂuxes. Journal of Geophysical Research, 103 (D11): 13145–52. Guenther, A.B., Monson, R.K. and Fall, R. (1991) Isoprene and monoterpene emission rate variability: Observations with eucalyptus and emission rate algorithm development. Journal of Geophysical Research, 96: 10799–808. Guenther, A.B., Zimmerman, P., Harley, P.C., et al. (1993) Isoprene and monoterpene emission rate variability: Model evaluations and sensitivity analyses. Journal of Geophysical Research, 98: 12609–17. Guenther, A.B., Hewitt, C.N., Erickson, D., et al. (1995) A global model of natural volatile organic compound emissions. Journal of Geophysical Research, 100 (D5): 8873–92. Guenther, A.B., Zimmerman, P., Klinger, L., et al. (1996) Estimate of regional natural volatile organic compound ﬂuxes from enclosure and ambient measurement. Journal of Geophysical Research, 101 (D1): 1345–9. Guenther, A.B., Archer, S., Greenberg, J., et al. (1999a) Biogenic hydrocarbon emissions and landcover/climate change in a subtropical savanna. Physics, Chemistry and Earth (B), 24 (6): 659–67. Guenther, A., Baugh, B., Brasseur, G., et al. (1999b) Isoprene emission estimates and uncertainties for the Central African EXPRESSO study domain. Journal of Geophysical Research, 104 (D23): 30625–39. Hakola, H., Laurila, T., Rinne, J., et al. (2000) The ambient concentrations of biogenic hydrocarbons at a northern European boreal site. Atmospheric Environment, 34: 4971–82. Hallquist, M., Wangberg, I., Ljungstrom, E., et al. (1999) Aerosol and product yields from NO3 radical-initiated oxidation of selected monoterpenes. Environmental Science and Technology, 33: 553–9. Hansen, U., Van Eijk, J., Bertin, N., et al. (1997) Biogenic emissions and CO2 gas exchange investigated on four Mediterranean shrubs. Atmospheric Environment, 31 (SI): 157–66. Hanson, D.T., Swanson, S., Graham, L.E., et al. (1999) Evolutionary signiﬁcance of isoprene from mosses. Americal Journal of Botony, 86: 634–9. Harley, P.C., Guenther, A. and Zimmerman, P. (1996) Effects of light, temperature and canopy position on net photosynthesis and isoprene emission from sweetgum (Liquidambar styraciﬂua) leaves. Tree Physiology, 16: 25–32.

Biogenic VOCs

121

Harley, P.C., Fridd-Stroud, V., Greenberg, J., et al. (1998) Emission of 2-methyl-3-buten-2-ol by pines: A potentially large natural source of reactive carbon to the atmosphere. Journal of Geophysical Research, 103 (D19): 25479–86. Harley, P.C., Monson, R.K. and Lerdau, M. (1999) Ecological and evolutionary aspects of isoprene emission from plants. Oecologia, 118 (2): 109–23. Harley, P.C., Vasconcellos, P., Vierling, L., et al. (2004) Variation in potential for isoprene emissions among Neotropical forest sites. Global Change Biology, 10: 630–50. He, N.P., Han, X.G., Sun, W., et al. (2004) Biogenic VOCs emission inventory development of temperate grassland vegetation in Xilin River Basin, Inner Mongolia, China. Journal of Environmental Science (China), 16 (6): 1024–32. Heikes, B.G., Chang, W., Pilson, M.E.Q., et al. (2002) Atmospheric methanol budget and ocean implication. Global Biogeochemical Cycles, 16 (4): 1133. Helmig, D. (1999) Review: Gas chromotography air analysis. Journal of Chromatography, 843: 129–46. Helmig, D., Greenberg, J., Guenther, A.B., et al. (1998a) Volatile organic compounds and isoprene oxidation products at a temperate deciduous forest site. Journal of Geophysical Research, 103 (D17): 22397–414. Helmig, D., Balseley, B., Davis, K., et al. (1998b) Vertical proﬁling and determination of landscape ﬂuxes of biogenic nonmethane hydrocarbons with the planetary boundary layer in the Peruvian Amazon. Journal of Geophysical Research, 103 (D19): 25519–32. Helmig, D., Revermann, T., Pollmann, J., et al. (2003) Calibration system and analytical considerations for quantitative sesquiterpene measurements in air. Journal of Chromatography, 1002: 193–211. Hirsch, A.I., Munger, J.W., Jacob, D.J., et al. (1996) Seasonal variation of the ozone production efﬁciency per unit NOx at Harvard Forest, Massachusetts. Journal of Geophysical Research, 101 (D7): 12659–66. Hoffmann, T., Odum, J.R., Bowman, F., et al. (1997) Formation of organic aerosols from the oxidation of biogenic hydrocarbons. Journal of Atmospheric Chemistry, 26 (2): 189–222. Holopainen, J.K. (2004) Multiple functions of inducible plant volatiles. Trends in Plant Science, 9 (11): 529–33. Holzinger, R., Sandoval-Soto, L., Rottenberger, S., et al. (2000) Emissions of volatile organic compounds from Quercus ilex L. measured by proton transfer reaction mass spectrometry under different environmental conditions. Journal of Geophysical Research, 105 (D16): 20573–9. Holzinger, R., Lee, A., Paw U, K.T., et al. (2004) Observations of oxidation products above a forest imply biogenic emissions of very reactive compounds. Atmospheric Chemistry and Physics Discussion, 4: 5345–65. Horowitz, L.W., Liang, J., Gardner, G.M., et al. (1998) Export of reactive nitrogen from North American during summertime: Sensitivity to hydrocarbon chemistry. Journal of Geophysical Research, 103 (D11): 13451–76. Houweling, S., Dentener, F.J. and Lelieveld, J. (1998) The impact of nonmethane hydrocarbon compounds on tropospheric chemistry. Journal of Geophysical Research, 103 (D9): 10673–96. Iijima, Y., Davidovich-Rikanati, R., Fridman, E., et al. (2004) The biochemical and molecular basis for the divergent patterns in the biosynthesis of terpenes and phenylpropenes in the peltate glands of three cultivars of basil. Plant Physiology, 136: 3724–36. Jacob, D.J., Logan, J.A., Gardner, G.M., et al. (1993) Factors regulating ozone over the United States and its export to the global atmosphere. Journal of Geophysical Research, 98 (D8): 14817–26. Jacob, D.J., Field, B.D., Jin, E.M., et al.(2002) Atmospheric budget of acetone. Journal of Geophysical Research, 107 (D10): 4100. Jaegle, L., Jacob, D.J., Brune, W.H., et al. (2001) Chemistry of HOx radicals in the upper troposphere. Atmospheric Environment, 35: 469–89.

122

Volatile Organic Compounds in the Atmosphere

Jang, M. and Kamens, R. (2001) Atmospheric secondary organic aerosol formation by heterogeneous reaction of aldehydes in the presence of a sulfuric acid aerosol catalyst. Environmental Science and Technology, 35: 4758–66. Kamens, R., Jeffries, H.E., Gery, M.W., et al. (1981) The impact of α-pinene on urban smog formation: An outdoor smog chamber study. Atmospheric Environment, 15 (6): 969–81. Kamens, R., Jang, M., Chien, C.-J., et al. (1999) Aerosol formation from the reaction of α-pinene and ozone using a gas-phase kinetics-aerosol partitioning model. Environmental Science and Technology, 33: 1430–8. Kanakidou, M., Seinfeld, J.H., Pandis, S.N., et al. (2004) Organic aerosol and global climate modelling: A review. Atmospheric Chemistry And Physics Discussion, 4: 5855–6024. Karl, T., Guenther, A.B., Jordan, A., et al. (2001a) Eddy covariance measurement of biogenic oxygenated VOC emissions from hay harvesting. Atmospheric Environment, 35 (3): 491–5. Karl, T., Fall, R., Jordan, A., et al. (2001b) On-line analysis of reactive VOCs from urban lawn mowing. Environmental Science and Technology, 35 (14): 2926–31. Karl, T., Curtis, A.J., Rosenstiel, T.N., et al. (2002) Transient releases of acetaldehyde from tree leaves – Products of a pyruvate overﬂow mechanism. Plant, Cell and Environment, 25: 1121–31. Karlik, J.F. and Winer, A.M. (2001) Measured isoprene emission rates of plants in California landscapes: Comparison to estimates from taxonomic relationships. Atmospheric Environment, 36 (6): 1123–31. Kesselmeier, J., Bode, K., Hofmann, U., et al. (1997) Emission of short chained organic acids, aldehydes and monoterpenes from Quercus ilex L. and Pinus pinea L. in relation to physiological activities, carbon budget and emission algorithms. Atmospheric Environment, 31 (S1): 119–33. Kesselmeier, J. and Staudt, M. (1999) Biogenic volatile organic compounds (VOC): An overview on emission, physiology and ecology. Journal of Atmospheric Chemistry, 33: 23–88. Kirstine, W., Galbally, I.E., Ye, Y., et al. (1998) Emissions of volatile organic compounds (primarily oxygenated species) from pasture. Journal of Geophysical Research, 103 (D9): 10605–19. Klinger, L., Li, Q.-J., Guenther, A., et al. (2002) Assessment of volatile organic compound emissions from ecosystems of China. Journal of Geophysical Research, 107 (D21): 4603. Koenig, G., Brunda, M., Puxbau, H., et al. (1995) Relative contribution of oxygenated hydrocarbons to the total biogenic VOC emissions of selected mid-European agricultural and natural plant species. Atmospheric Environment, 29 (8): 861–74. Kreuzwieser, J., Scheerer, U. and Rennenberg, H. (1999) Metabolic origin of acetaldehyde emitted by poplar (Populus tremula × P. alba) trees. Journal of Experimental Botony, 50 (335): 757–65. Kreuzwieser, J., Graus, M., Wisthaler, A., et al. (2002) Xylem-transported glucose as additional carbon source for leaf formation in Quercus robur. New Phytologist, 156: 171–8. Kuhn, U., Rottenberger, S., Biesenthal, T.A., et al. (2004) Seasonal differences in isoprene and light-dependent monoterpene emission by Amazonian tree species. Global Change Biology, 10: 663–82. Kurpius, M.R. and Goldstein, A.H. (2003) Gas-phase chemistry dominates O3 loss to forest, implying a source of aerosols and hydroxyl radicals to the atmosphere. Geophysical Research Letter, 30: 1371. −C10 volatile organic comLamanna, M.S. and Goldstein, A.H. (2002) In situ measurements of C2 − pounds above a Sierra Nevada ponderosa pine plantation. Journal of Geophysical Research, 104 (D7): 21247–62. Lamb, B., Westberg, H., Allwine, G., et al. (1985) Biogenic hydrocarbon emissions from deciduous and coniferous trees in the United States. Journal of Geophysical Research, 90 (D1): 2380–90. Lamb, B., Guenther, A., Gay, D., et al. (1987) A national inventory of biogenic hydrocarbon emissions. Atmospheric Environment, 21 (8): 1695–1705. Lee, A., Schade, G.W., Holzinger, R., et al. (2005) A comparison of new measurements of total monoterpene ﬂux with improved measurements of speciated monoterpene ﬂux. Atmospheric Chemistry and Physics Discussion, 5: 505–13.

Biogenic VOCs

123

Lehning, A., Zimmer, W., Zimmer, I., et al. (2001) Modeling of annual variations in oak (Quercus robur L.) isoprene synthase activity to predict isoprene emission rates. Journal of Geophysical Research, 106 (D3): 3157–66. Lerdau, M., Guenther, A. and Monson, R.K. (1997) Plant production and emission of volatile organic compounds. Bioscience, 47 (6): 373–83. Lerdau, M. and Throop, H.L. (1999) Isoprene emission and photosynthesis in a tropical forest canopy: Implications for model development. Ecological Application, 9 (4): 1109–17. Lichtenthaler, H.K. (1999) The 1-deoxy-D-xylulose-5-phosphate pathway of isoprenoid biosynthesis in plants. Annual Review of Plant Physiology, 50: 47–65. Lichtenthaler, H.K., Schwender, J., Disch, A., et al. (1997) Biosynthesis of isoprenoids in higher plant chloroplasts proceeds via a mevalonate-independent pathway. FEBS Letter, 400: 271–4. Lieth, H. (1975) Modeling the primary productivity of the world. In: H. Lieth and R. Whittaker (Eds) Primary Productivity of the Biosphere. Springer-Verlag: NY, pp. 237–67. Limbeck, A., Kulmala, M. and Puxbaum, H. (2003) Secondary organic aerosol formation in the atmosphere via heterogeneous reaction of gaseous isoprene on acidic particles. Geophysical Research Letter, 30 (19): 1996. Litvak, M.E. and Monson, R.K. (1998) Patterns of induced and constitutive monoterpene production in conifer needles in relation to insect herbivory. Oecologia, 114: 531–40. Litvak, M.E., Loreto, F., Harley, P.C., et al. (1996) The response of isoprene emission rate and photosynthetic rate to photon ﬂux and nitrogen supply in aspen and white oak trees. Plant, Cell and Environment, 19: 549–59. Litvak, M.E., Madronich, S. and Monson, R.K. (1999) Herbivore-induced monoterpene emissions from coniferous forests: Potential impact on local tropospheric chemistry. Ecological Application, 9 (4): 1147–59. Liu, S., Trainer, M., Fehsenfeld, F., et al. (1987) Ozone production in the rural troposphere and the implications for regional and global ozone distributions. Journal of Geophysical Research, 92 (D4): 4191–207. Logan, B.A. and Monson, R.K. (1999) Thermotolerance of leaf discs from four isoprene-emitting species is not enhanced by exposure to exogenous isoprene. Plant Physiology, 120: 821–5. Logan, B.A., Monson, R.K. and Potosnak, M.J. (2000) Biochemistry and physiology of foliar isoprene production. Trends in Plant Science, 5: 477–81. Loreto, F. and Sharkey, T.D. (1990) A gas exchange study of photosynthesis and isoprene emission in red oak (Quercus rubra L.). Planta, 182: 523–31. Loreto, F., Ciccioli, P., Brancaleoni, E., et al. (1996a) Different sources of reduced carbon contribute to form three classes of terpenoid emitted by Quercus ilex L. leaves. Proceedings of the National Academy of Science USA, 93: 9966–9. Loreto, F., Ciccioli, P., Cecinato, A., et al. (1996b) Inﬂuence of environmental factors and air composition on the emission of α-pinene from Quercus ilex leavs. Plant Physiology, 110: 267–75. Loreto, F., Nascetti, P., Graverini, A., et al. (2000) Emission and content of monoterpenes in intact and wounded needles of the Mediterranean pine Pinus pinea. Functional Ecology, 14: 589–95. Loreto, F., Fischbach, R.J., Schnitzler, J.-P., et al. (2001a) Monoterpene emission and monoterpene synthase activities in the Mediterranean evergreen oak Quercus ilex L. grown at elevated CO2 concentrations. Global Change Biology, 7: 709–17. Loreto, F., Mannozzi, M., Maris, C., et al. (2001b) Ozone quenching properties of isoprene and its antioxidant role in leaves. Plant Physiology, 126: 993–1000. Loreto, F., Pinelli, P., Brancaleoni, E., et al. (2004) 13 C labeling reveals chloroplastic and extrachloroplastic pools of dimethylallyl pyrophosphate and their contribution to isoprene formation. Plant Physiology, 135: 1903–7. MacDonald, R. and Fall, R. (1993) Detection of substantial emissions of methanol from plants to the atmosphere. Atmospheric Environment, 27A: 1709–14.

124

Volatile Organic Compounds in the Atmosphere

Makar, P.A., Fuentes, J.D., Wang, D., et al. (1999) Chemical processing of biogenic hydrocarbons within and above a temperature deciduous forest. Journal of Geophysical Research, 104 (D3): 3581–603. Martin, M.J., Stirling, C.M., Humphries, S.W., et al. (2000) A process-based model to predict the effects of climatic change on leaf isoprene emission rates. Ecological Modelling, 131: 161–74. Mauzerall, D.L., Jacob, D.J., Fan, S.M., et al. (1996) Origin of tropospheric ozone at remote high northern latitudes in summer. Journal of Geophysical Research, 101 (D2): 4175–88. Meyrahn, H., Pauly, J., Schneider, W., et al. (1986) Quantum yields for the photodissociation of acetone in air and an estimate for the life time of acetone in the lower troposphere. Journal of Atmospheric Chemistry, 4: 277–91. Millet, D.B., Donahue, N.M., Pandis, S.N., et al. (2005) Atmospheric VOC measurements during the Pittsburgh Air Quality Study: Results, interpretation and quantiﬁcation of primary and secondary contributions. Journal of Geophysical Research, 110: D07S07. Monson, R.K. and Fall, R. (1989) Isoprene emission from aspen leaves. Plant Physiology, 90: 267–74. Monson, R.K. and Holland, E.A. (2001) Biospheric trace gas ﬂuxes and their control over tropospheric chemistry. Annual Review of Ecology Systematics, 32: 547–76. Monson, R.K., Jaeger, C.H., Adams, W.W.I., et al. (1992) Relationships among isoprene emission rate, photosynthesis, and isoprene synthase activity as inﬂuenced by temperature. Plant Physiology, 98: 1175–80. Monson, R.K., Harley, P.C., Litvak, M.E., et al. (1994) Environmental and developmental controls over the seasonal pattern of isoprene emissions from aspen leaves. Oecologia, 99: 260–70. Mueller, J.-F. (1992) Geographical distribution and seasonal variation of surface emissions and deposition velocities of atmospheric trace gases. Journal of Geophysical Research, 97: 3787–804. Nemecek-Marshall, M., MacDonald, R., Franzen, J.J., et al. (1995) Methanol emissions from leaves (enzymatic detection of gas-phase methanol and relation of methanol ﬂuxes to stomatal conductance and leaf development). Plant Physiology, 108: 1359–68. Niinemets, U. and Reichstein, M. (2002) A model analysis of the effects of nonspeciﬁc monoterpenoid storage in leaf tissues on emission kinetics and composition in Mediterranean sclerophyllous Quercus species. Global Biogeochemical Cycles, 16 (4): 1110. Niinemets, U. and Reichstein, M. (2003a) Controls on the emission of plant volatiles through stomata: Differential sensitivity of emission rates to stomatal closure explained. Journal of Geophysical Research, 108 (D7): 4208. Niinemets, U. and Reichstein, M. (2003b) Controls on the emission of plant volatiles through stomata: A sensitivity analysis. Journal of Geophysical Research, 108 (D7): 4211. Niinemets, U., Tenhunen, J.D., Harley, P.C., et al. (1999) A model of isoprene emission based on energetic requirements for isoprene synthesis and leaf photosynthetic properties for Liquidambar and Quercus. Plant, Cell and Environment, 22 (11): 1319–35. Otter, L., Guenther, A., Wiedinmyer, C., et al. (2003) Spatial and temporal variations in biogenic volatile organic compound emissions for Africa south of the equator. Journal of Geophysical Research, 108 (D13): 8505. Palmer, P.I. and Shaw, S.L. (2005) Quantifying global marine isoprene ﬂuxes using MODIS chlorophyll observations. Geophysical Research Letter, 32: L09805, doi:10.1029/2005GL022592. Palmer, P.I., Jacob, D.J., Fiore, A.M., et al. (2003) Mapping isoprene emissions over North American using formaldehyde column observations from space. Journal of Geophysical Research, 108 (D6): 4180. Pandis, S.N., Paulson, S.E., Seinfeld, J.H., et al. (1991) Aerosol formation in the photooxidation of isoprene and β-pinene. Atmospheric Environment, 25A (5/6): 997–1008. Papagni, C., Arey, J. and Atkinson, R. (2001) Rate constants for the gas-phase reactions of OH radicals with a series of unsaturated alcohols. International Journal of Chemical Kinetics, 33: 142–7. Paulson, S.E. and Orlando, J.J. (1996) The reactions of ozone with alkenes: An important source of HOx in the boundary layer. Geophysical Research Letter, 23 (25): 3727–30.

Biogenic VOCs

125

Paulson, S.E., Chung, M.Y. and Hasson, A.S. (1999) OH radical formation from the gas-phase reaction of ozone with terminal alkenes and the relationship between structure and mechanism. Journal of Physical Chemistry A, 103 (41): 8125–38. Pegoraro, E., Rey, A., Bobich, E.G., et al. (2004) Effect of elevated CO2 concentration and vapour pressure deﬁcit on isoprene emission from leaves of Populus deltoides during drought. Functional Plant Biology, 31: 1137–47. Petron, G., Harley, P.C., Greenberg, J., et al. (2001) Seasonal temperature variations inﬂuence isoprene emissions. Geophysical Research Letter, 28: 1707–10. Pierce, T.E. and Waldruff, P. (1991) PC-BEIS: A personal computer version of the biogenic emissions inventory system. Journal of Air and Waste Management Association, 41: 937–41. Pierce, T.E., Geron, C., Bender, L., et al. (1998) Inﬂuence of increased isoprene emissions on regional ozone modeling. Journal of Geophysical Research, 103 (D19): 25611–29. Plass-Dulmer, C., Koppman, R., Ratte, M., et al. (1995) Light nonmethane hydrocarbons in seawater. Global Biogeochemical Cycles, 9: 79–100. Poisson, N., Kanakidou, M. and Crutzen, P.J. (2000) Impact of non-methane hydrocarbons on tropospheric chemistry and the oxidizing power of the global troposphere: 3-Dimensional modeling results. Journal of Atmospheric Chemistry, 36: 157–230. Potter, C.S., Alexander, S.E., Coughlan, J.C., et al. (2001) Modeling biogenic emissions of isoprene: Exploration of model drivers, climate control algorithms, and use of global satellite observations. Atmospheric Environment, 35: 6151–65. Potter, C., Klooster, S., Bubenheim, D., et al. (2003) Modeling terrestrial biogenic sources of oxygenated organic emissions. Earth Interactions, 7 (7): 1–15. Rapparini, F., Baraldi, R., Miglietta, F., et al. (2004) Isoprenoid emission in trees of Quercus pubescens and Quercus ilex with lifetime exposure to naturally high CO2 environment. Plant, Cell and Environment, 27: 381–91. Rasmussen, R. (1970) Isoprene: Identiﬁed as a forest-type emission to the atmosphere. Environmental Science and Technology, 4: 667–71. Rasmussen, R.A. and Khalil, M. (1988) Isoprene over the Amazon basin. Journal of Geophysical Research, 93 (D2): 1417–21. Rasmussen, R.A. (1972) What do hydrocarbons from trees contribute to air pollution? Journal of the Air Pollution Control Association, 22: 537–43. Rasmussen, R.A. and Went, F.W. (1965) Volatile organic material of plant origin in the atmosphere. Proceedings of the National Academy of Science USA, 53: 215–20. Reissell, A., Arey, J. and Atkinson, R. (2001) Atmospheric chemistry of camphor. International Journal of Chemical Kinetics, 33: 56–63. Riemer, D., Pos, W., Milne, P., et al. (1998) Observations of nonmethane hydrocarbons and oxygenated volatile organic compounds at a rural site in the southeastern United States. Journal of Geophysical Research, 103 (D21): 28111–28. Rinne, H.J., Guenther, A., Greenberg, J., et al. (2002) Isoprene and monoterpene ﬂuxes measured above Amazonian rainforest and their dependence on light and temperature. Atmospheric Environment, 36 (14): 2421–6. Robinson, E. and Robbins, R. (1968) Sources, Abundance, and Fate of Gaseous Atmospheric Pollutants. Stanford Research Institute, Menlo Park, CA. Final Report of Project PR-6755. Rosenstiel, T.N., Fisher, A.J., Fall, R., et al. (2002) Differential accumulation of dimethylallyl diphosphate in leaves and needles of isoprene- and methylbutenol-emitting and non-emitting species. Plant Physiology, 129: 1276–84. Rosenstiel, T.N., Potosnak, M.J., Grifﬁn, K.L., et al. (2003) Increased CO2 uncouples growth from isoprene emission in an agriforest ecosystem. Nature, 421: 256–9. Rosenstiel, T.N., Ebbets, A.L., Khatri, W.C., et al. (2004) Induction of a poplar leaf nitrate reductase: A test of extrachloroplastic control of isoprene emission rate. Plant Biology, 6: 12–21.

126

Volatile Organic Compounds in the Atmosphere

Rudich, Y., Talukdar, R.K., Fox, R.W., et al. (1996) Rate coefﬁcients for reactions of NO3 with a few oleﬁns and oxygenated oleﬁns. Journal of Physical Chemistry, 100: 5374–81. Sanadze, G.A. (1957) Emission of organic matters by leaves of Robinia psudaeacia L. Soobsh. Academy of Nauk Global Sustainable Systems Research, 19: 83–6. Sanadze, G.A. (2004) Biogenic isoprene (a review). Russian Journal of Plant Physiology, 51 (6): 810–24. Sanadze, G.A., Dzhaiani, G.I. and Tevzadze, I.M. (1972) Incorporation into the isoprene molecule of carbon from 13 CO2 assimilated during photosynthesis. Soviet Plant Physiology, 19: 17–20. Schade, G.W. and Goldstein, A.H. (2001) Fluxes of oxygenated volatile organic compounds from a ponderosa pine plantation. Journal of Geophysical Research, 106 (D3): 3111–23. Schade, G.W. and Goldstein, A.H. (2002) Plant physiological inﬂuences on the ﬂuxes of oxygenated volatile organic compounds from ponderosa pine trees. Journal of Geophysical Research, 107 (D10): 4082. Schade, G.W. and Goldstein, A.H. (2003) Increase of monoterpene emissions from a pine plantation as a result of mechanical disturbances. Geophysical Research Letter, 30: 1380. Schade, G.W., Goldstein, A.H. and Lamanna, M.S. (1999) Are monoterpene emissions inﬂuenced by humidity? Geophysical Research Letter, 26 (14): 2187–90. Schade, G.W., Goldstein, A.H., Gray, D.W., et al. (2000) Canopy and leaf level 2-methyl-3-buten-2-ol ﬂuxes from a ponderosa pine plantation. Atmospheric Environment, 34 (21): 3535–44. Scheeren, H.A., Lelieveld, J., Roelofs, G.J., et al. (2003) The impact of monsoon outﬂow from India and Southeast Asia in the upper troposphere over the eastern Mediterranean. Atmospheric Chemistry and Physics Discussion, 3: 2285–330. Schnitzler, J.-P., Graus, M., Kreuzwieser, J., et al. (2004) Contribution of different carbon sources for isoprene emitted from poplar leaves. Plant Physiology, 135: 152–60. Scott, K.I. and Benjamin, M.T. (2003) Development of a biogenic volatile organic compound emission inventory for the SCOS97-NARSTO domain. Atmospheric Environment, 37 (2): S39–49. Seinfeld, J.H. and Pandis, S.N. (1998) Atmospheric Chemistry and Physics. John Wiley & Sons, New York. Sharkey, T.D., Loreto, F. and Delwiche, C.F. (1991) The biochemistry of isoprene emission from leaves during photosynthesis. In: T.D. Sharkey, E.A. Holland and H.A. Mooney (Eds) Trace Gas Emissions from Plants. Academic Press, San Diego, CA, pp. 153–84. Sharkey, T.D., Singsaas, E.L., Lerdau, M., et al. (1999) Weather effects on isoprene emission capacity and applications in emissions algorithms. Ecological Application, 9: 1132–7. Sharkey, T.D. and Singsaas, E.L. (1995) Why plants emit isoprene. Nature, 374: 569. Sharkey, T.D. and Yeh, S. (2001) Isoprene emission from plants. Annual Review of Plant Physiology and Plant Molecular Biology, 52: 407–36. Silver, G. and Fall, R. (1991) Enzymatic synthesis of isoprene from dimethylallyl diphosphate in aspen leaf extracts. Plant Physiology, 97: 1588–91. Silver, G.M. and Fall, R. (1995) Characterization of aspen isoprene synthase, an enzyme responsible for leaf isoprene emission to the atmosphere. Journal of Biological Chemistry, 270 (22): 13010–16. Simpson, D., Guenther, A., Hewitt, C.N., et al. (1995) Biogenic emissions in Europe 1. Estimates and uncertainties. Journal of Geophysical Research, 100 (D11): 22875–90. Singh, H.B., O’Hara, D., Herlth, D., et al. (1994) Acetone in the atmosphere: Distribution, sources, and sinks. Journal of Geophysical Research, 99 (D1): 1805–19. Singh, H.B., Chen, Y., Tabazadeh, A., et al. (2000) Distribution and fate of selected oxygenated organic species in the troposphere and lower stratosphere over the Atlantic. Journal of Geophysical Research, 105 (D3): 3795–805.

Biogenic VOCs

127

Singh, H.B., Chen, Y., Staudt, A., et al. (2001) Evidence from the Paciﬁc troposphere for large global sources of oxygenated organic compounds. Nature, 410: 1078–81. Singh, H.B., Tabazadeh, A., Evans, M.J., et al. (2003) Oxygenated volatile organic chemicals in the oceans: Inferences and implications based on atmospheric observations and air-sea exchange models. Geophysical Research Letter, 30 (16): 1862. Singh, H.B., Salas, L.J., Chatﬁeld, R.B., et al. (2004) Analysis of the atmospheric distribution, sources, and sinks of oxygenated volatile organic chemicals based on measurements over the paciﬁc during TRACE-P. Journal of Geophysical Research, 109: D15S07. Singsaas, E.L. and Sharkey, T.D. (2000) The effects of high temperature on isoprene synthesis in oak leaves. Plant, Cell and Environment, 23: 751–7. Singsaas, E.L., Lerdau, M., Winter, K., et al. (1997) Isoprene increases thermotolerance of isopreneemitting species. Plant Physiology, 115: 1413–20. Smith, A.M., Rigler, E., Kwok, E.S.C., et al. (1996) Kinetics and products of the gas-phase reactions of 6-methyl-5-hepten-2-one and trans-cinnamaldehyde with OH and NO3 radicals and O3 at 296+/−2K. Environmental Science and Technology, 30: 1781–5. Solmon, F., Sarrat, C., Serca, D., et al. (2004) Isoprene and monoterpenes biogenic emissions in France: Modeling and impact during a regional pollution episode. Atmospheric Environment, 38 (23): 3853–65. Staudt, M. and Seufert, G. (1995) Light-dependent emissions of monoterpenes by Holm oak (Quercus ilex L.). Naturwissenschaften, 82: 89–92. Staudt, M., Joffre, R., Rambal, S., et al. (2001) Effect of elevated CO2 on monoterpene emission of young Quercus ilex trees and its relation to structural and ecophysiological parameters. Tree Physiology, 21: 437–45. Staudt, M., Mir, C., Joffre, R., et al. (2004) Isoprenoid emissions of Quercus spp. (Q. suber and Q. ilex) in mixed stands contrasting in interspeciﬁc genetic introgression. New Phytologist, 163: 573–84. Steinbrecher, R. (1989) Gehalt und Emission von Monoterpenen in oberirdischen Organen von Picea abes (L.). Universitaet Muenchen, Munich, Germany. Steiner, A.L., Luo, C., Huang, Y., et al. (2002) Past and present-day biogenic volatile organic compound emissions in East Asia. Atmospheric Environment, 36: 4895–905. Stewart, H.E., Hewitt, C.N., Bunce, R.G.H., et al. (2003) A highly spatially and temporally resolved inventory for biogenic isoprene and monoterpene emissions: Model description and application to Great Britain. Journal of Geophysical Research, 108 (D20): 4644. Taylor, J., Zimmerman, P. and Erickson, D. (1990) A 3-D modelling study of the sources and sinks of atmospheric carbon monoxide. Australian National University, Canberra. Tie, X., Guenther, A. and Holland, E.A. (2003) Biogenic methanol and its impacts on tropospheric oxidants. Geophysical Research Letter, 30 (17): 1881. Tingey, D., Evans, R. and Gumpertz, M. (1981) Effects of environmental conditions on isoprene emission from live oak. Planta, 152: 565–70. Tingey, D.T., Turner, D.P. and Weber, J.A. (1991) Factors controlling the emissions of monoterpenes and other volatile organic compounds. In: T.D. Sharkey, E.A. Holland and H.A. Mooney (Eds) Trace Gas Emissions by Plants. Academic Press Inc., San Diego, CA, pp. 93–119. Trainer, M., Williams, E.J., Parrish, D.D., et al. (1987) Models and observations of the impact of natural hydrocarbons on rural ozone. Nature, 329: 705–7. Tsigaridis, K. and Kanakidou, M. (2002) Importance of volatile organic compounds photochemistry over a forested area in central Greece. Atmospheric Environment, 36: 3137–46. Turner, D., Baglio, J., Pross, D., et al. (1991) Climate change and isoprene emissions from vegetation. Chemosphere, 23: 37–56. Wang, Y., Jacob, D.J. and Logan, J.A. (1998) Global simulation of tropospheric O3 -NOx -hydrocarbon chemistry, 3. Origin of tropospheric ozone and effects of nonmethane hydrocarbons. Journal of Geophysical Research, 103 (D9): 10757–67.

128

Volatile Organic Compounds in the Atmosphere

Warneke, C., Holzinger, R., Hansel, A., et al. (2001) Isoprene and its oxidation products methyl vinyl ketone, methacrolein, and isoprene related peroxides measured online over the tropical rain forest of Surinam in March 1998. Journal of Atmospheric Chemistry, 38: 167–85. Warneke, C., Luxembourg, S.L., DeGouw, J.A., et al. (2002) Disjunct eddy covariance measurements of oxygenated volatile organic compounds ﬂuxes from an alfalfa ﬁeld before and after cutting. Journal of Geophysical Research, 107 (D8): 4067. Warneke, C., DeGouw, J.A., Goldan, P., et al. (2004) Comparison of daytime and nighttime oxidation of biogenic and anthropogenic VOCs along the New England coast in summer during New England Air Quality Study 2002. Journal of Geophysical Research, 109: D10309. Wennberg, P.O., Hanisco, T.F., Jaegle, L., et al. (1998) Hydrogen radicals, nitrogen radicals, and the production of O3 in the upper troposphere. Science, 279: 49–53. Went, F.W. (1960) Blue hazes in the atmosphere. Nature, 187: 641–3. Wiedinmyer, C., Guenther, A., Harley, P.C., et al. (2004) Global organic emissions from vegetation. In: C. Granier, P. Artaxo and C. Reeves (Eds) Emissions of Atmospheric Trace Compounds. Kluwer Academic Publishers, Dordrecht, The Netherlands, pp. 115–70. Williams, J., Holzinger, R., Gros, V., et al. (2004) Measurements of organic species in air and seawater from the tropical Atlantic. Geophysical Research Letter, 31: L23S06. Winer, A.M. et al. (1983) Investigation of the role of natural hydrocarbons on photochemical smog formation in California. California Air Resources Board, Statewide Air Pollution Research Center, Riverside, CA. Yokouchi, Y. and Ambe, Y. (1984) Factors affecting the emission of monoterpenes from red pine (Pinus densiﬂora)-long-term effects of light, temperature and humidity. Plant Physiology, 74: 1009–12. Zavarin, E. (1968) Chemotaxonomy of the genus Abies – II. Within tree variation of the terpenes in cortical oleoresin. Phytochemistry, 7: 92–107. Zeidler, J.G., Lichtenthaler, H.K., May, H.U., et al. (1997) Is isoprene emitted by plants synthesized via the novel isopentenyl pyrophosphate pathway? Zeitschrift fuer Naturforschung C, 52: 15–23. Zimmer, W., Brueggeman, N., Emeis, S., et al. (2000) Process-based modelling of isoprene emission by oak leaves. Plant, Cell and Environment, 23: 585–595. Zimmerman, P. (1979) Testing of hydrocarbon emissions from vegetation, leaf litter and aquatic surfaces, and development of a methodology for compiling biogenic emission inventories. U.S. Environmental Protection Agency, Research Triangle Park, NC, EPA-450/4-79-004: 1–113.

Volatile Organic Compounds in the Atmosphere Edited by Ralf Koppmann Copyright © 2007 by Blackwell Publishing Ltd

Chapter 4

Oxygenated Volatile Organic Compounds Ralf Koppmann and Jürgen Wildt

4.1

Introduction

Oxygenated volatile organic compounds (OVOCs) belong to the large family of organic compounds present in the global atmosphere. These compounds include carbonyls, alcohols, ketones, esters, ethers, organic peroxides and organic hydroperoxides. At least the lighter-molecular-weight OVOCs are ubiquitous at relatively high concentrations in the troposphere, and it is widely accepted that they play an important role in atmospheric photochemistry. Despite that recognition, their study has been somewhat neglected, mainly because they are difﬁcult to measure. OVOCs have complex primary and secondary sources. On the one hand they are emitted directly into the atmosphere from a variety of anthropogenic and natural sources. On the other hand, they are products, often the ﬁrst stable ones, in the gas-phase oxidation pathways of organic compounds in the atmosphere. The removal of OVOCs from the atmosphere occurs mainly by the reaction with OH radicals, but also by photolysis and by wet and dry deposition. However, the removal processes are well understood only for a few compounds. As a consequence, the budgets of OVOCs are either poorly understood or not known at all. While the role of formaldehyde (HCHO) as an oxidation product of methane and other VOCs has been studied for more than two decades, the last years have witnessed a growing interest in the role of other OVOCs owing to improved analytic technologies and a deeper understanding of photochemical processes. As an example, acetone was found to be a key atmospheric compound inﬂuencing tropospheric chemistry. It has a signiﬁcant impact on atmospheric chemistry, especially in the upper troposphere, where photolysis leads to the formation of peroxy acetyl nitrate (PAN). In this way, acetone can inﬂuence ozone chemistry by sequestering nitrogen oxides (NOx ) in the form of PAN and by providing free radicals (HOx ) especially in the free troposphere (Jaeglé et al. 2001; McKeen et al. 1997; Singh et al. 1994, 1995; Wennberg et al. 1998). OVOCs may also contribute to organic carbon in aerosols (Jang et al. 2002; Kalberer et al. 2004; Li et al. 2001; Tabazadeh et al. 2004). In addition to their photochemical impact, many of these compounds are harmful to human health especially in urban environments. They affect the respiratory tract and irritate the eyes and some of them are believed to be mutagenic and cancerogenic (Kean et al. 2001; Sin et al. 2001).

130

Volatile Organic Compounds in the Atmosphere

The increasing awareness that these compounds are important players in tropospheric chemistry has lead to signiﬁcant ongoing research although a quantitative determination of oxygenated hydrocarbons is still a challenge.

4.2

Tropospheric mixing ratios and global distribution

The abundance of OVOCs in the atmosphere has been studied on local, regional, and global scales. The investigated compounds cover a variety of aldehydes, ketones, alcohols, and carboxylic acids. Understandably, the largest set of compounds can be found in urban air, in the immediate vicinity of the sources. In these cases the measurements beneﬁt from high mixing ratios and short transport times from the source to the measurement location. In the context of air quality studies, a large number of investigations have been carried out concerning OVOC concentrations on a local scale. These studies dealt with a large set of aldehydes, ketones, and sometimes alcohols in the urban atmosphere. There are a fewer number of investigations at semi-rural or rural sites, and some studies at remote sites or on a global scale. The investigations on global scales are typically limited to formaldehyde, acetone, methanol, and ethanol. This can be attributed to their ubiquitous production or long lifetimes in the atmosphere leading to larger scale distribution and accessible mixing ratios in remote areas and in the free troposphere, respectively. Table 4.1 gives an overview of these studies and lists the concentrations resp. mixing ratios of the three most abundant OVOCs in the urban atmosphere, formaldehyde, acetaldehyde and acetone. In the following we will give an overview of OVOC measurements sorted by compound groups. Table 4.2 summarises the mixing ratios also for highermolecular-weight OVOCs obtained from ﬁeld measurements in rural areas from selected studies. Aldehydes. There are numerous studies on atmospheric mixing ratios of formaldehyde and acetaldehyde. A large part of these investigations has been done in cities and populated areas around the world. Formaldehyde and acetaldehyde are the main OVOCs found in urban environments, together contributing to more than half of the OVOCs by mass (Kean et al. 2001; Sin et al. 2001). Typical concentrations of formaldehyde and acetaldehyde observed in urban air range between 1 and 45 μg/m3 (corresponding to 0.8 and 36 ppb) and 0.7 and 35 μg/m3 (corresponding to 0.5 and 19 ppb), respectively. In rural environments mixing ratios are in the lower ppb range, sometimes less than 1 ppb. At remote locations and in the free troposphere mixing ratios seldom exceed 1 ppb. A few studies report observations of higher-molecular-weight aldehydes in urban air. Higher-molecular-weight aldehydes observed in the rural atmosphere often consist of those formed during lipoxygenase (LOX) activity (C6 aldehydes, see section ‘Biogenic emissions’) −C10 aldehydes such as hexanal, heptanal, octanal, nonanal, and decanal. and saturated C6 − These compounds have been found in mixing ratios from <10 ppt up to more than 1 ppb (Ciccioli et al. 1993; Helmig and Greenberg 1994; Wedel et al. 1998; Yokouchi et al. 1990). Emission from vegetation seems to be one of the possible sources. These emissions can be induced by ozone exposure of plants (Karl et al. 2005). The induction can already occur at ozone concentrations below 10 ppb (Wildt et al. 2003). Ozone is present nearly everywhere

Table 4.1 Overview of studies of OVOCs at urban, rural and remote sites

Urban sites Los Angeles, United States Takasaki, Japan Denver, United States Atlanta, United States Copenhagen, Denmark Leipzig, Germany Rome, Italy Mexico City, Mexico Grenoble, France Kuopio, Finlannd Cairo, Egypt Hongkong Guangzhou, China Rural sites Alabama, United States Boulder, Colorado, United States Schauinsland, Germany Lille Valby, Denmark Birkenes, Sweden Melpitz, Germany Central Ontario, Canada Pabsthum, NE Germany Remote sites Global Pennewitt, NE Germany Georgia, United States SE United States Alert, Canada Free troposphere, United States Paciﬁc Western Europe/Eastern Atlantic Germany North Atlantic

Sampling time

HCHO

CH3 CHO

CH3 COCH3

Autumn 1984 Summer 1983–1986 December 1987–October 1991 July–August 1992 February–June 1994 November 1993 January–March 1995 March–May 1995 May 1995 May 1997–January 1998 March–August 1999 October 1997–September 2000 July–September 2003

2.07–2.81 3.1–14 2.82–4.81 3.3–3.7 3.2 15–35 ppb 13.7 43.5 3.1–22 1.3–2.8 40 3.7–4.9 13.68

2.05–2.84 2.3–12 1.78–2.95 4.7–5.7 1.8 15–23 ppb 8.2 28.6 3.6–18 1.1–3.2 — 1.9–2.4 8.33

— — — — 2.4 30–70 ppb 10.4 — — — — 1.1–1.5 17.76

Kawamura et al. (2000) Satsumabayashi et al. (1989) Anderson et al. (1996) Grosjean et al. (1993) Granby et al. (1997) Müller (1997)2 Possanzini et al. (2000) Báez et al. (1995) Ferrari et al. (1998) Viskari et al. (2000) Khoder et al. (2000) Sin et al. (2001)2 Feng et al. (2005)2

1.3 ppb

4.2 ppb 0.7–5.2 ppb 6.2 4.5 3.1 — 4.3

Goldan et al. (1995a)2 Goldan et al. (1995b)2 Slemr et al. (1996) Christensen et al. (2000) Solberg et al. (1996) Müller (1997) Shepson et al. (1991) Grossman et al. (2003)2

June–July 1990 February 1991 September 1992 May–July 1995 January 1992–December 1995 June 1994/July 1995 July–August 1998 July 1998 January 1991 August 1994 Summer 1991 June 1992 Summer 1995 February–May 2000 July – August 1990 February – March 1994 1994, 1995, 1996 September/October 1991 September 1997

1.2 1.5 1.0 1–8 ppb 2.0

1.3 1.4 0.7 1–3 ppb 0.9 0.1–0.6 ppb

1 ppb–50 ppt 0.6–3.3 ppb

160–40 ppt 0.1–3.8 ppb

11.66 ppb

217 – 410 ppt

0.56 ppb 0.74 ppb 1.56 ppb 134, 202, 166 ppt1

1.8 ppb 1.7 ppb 3.83 ppb 200, 479, 871 ppt1 1.14 ppb 0.2–0.66 ppb 85–2000 ppt 0.1–1 ppb

Reference

Arlander et al. (1995a,1995b) Benning and Wahner (1998) and Wedel et al. (1998)2 Lee et al. (1995) Riemer et al. (1998)2 Boudries et al. (2002)2 Singh et al. (1994) Singh et al. (1995)2 Arnold et al. (1997) Reiner et al. (1999)2 Fried et al. (2002)

Given here are the reported average concentrations or concentration ranges of the three OVOCs (formaldehyde, acetaldehyde and acetone) in μg m−3 unless otherwise noted. In some of these studies other OVOCs have been measured. For these results, we refer to the cited literature. 1 The values refer to measurements in the dark, during the transition period, and during 24-h daylight, respectively. 2 During these studies also other OVOCs have been measured.

132

Volatile Organic Compounds in the Atmosphere

Table 4.2 Atmospheric concentrations of OVOC from ﬁeld measurements in rural areas Compound

Atmospheric concentrations [ppb]

Methanol

6–11

Methanol Methanol Methanol

3.1–22 2–12 ≤7.7

Formaldehyde Formaldehyde Ethanol

480 ng/m3 ≤6.5 0.95–1.2

Ethanol Acetaldehyde Acetaldehyde

0.09–1.32 0.85–1.3 0.4

Acetaldehyde Acetone

360 ng/m3 3.3–4.2

Acetone MBO 1-Penten-3-ol (+methyl-butanals) (Z)-3-Hexenal

1340 ng/m3 2–3 2–6 <0.1–20

Hexenol and Hexanal 1,8 Cineole Hexanal–Decanal Hexanal–Decanal Hexanal–Decanal Methylvinylketone

≤0.02 1.1–8.8 0.7–3.9 ≤0.45 0.4–1.5

Methacrolein

0.5–0.7

Pinonaldehyde

<0.1–7

≤0.15 ppb

Location of measurement

Rural Alabama, Loblolly pine plantation Rural Tennessee Subalpine conifer forest Mixed hardwood forest, northern Michigan Coniferous forest, Hyytiälä Spruce forest, Germany Rural Alabama, Loblolly pine plantation Rural Tenessee Rural Alabama Mixed hardwood forest, northern Michigan Coniferous forest, Hyytiälä Rural Alabama Loblolly pine plantation Coniferous forest, Hyytiälä Sonnblick observatory, Austria Innsbruck, measurements before and after lawn mowing Innsbruck, measurements before and after lawn mowing Maize ﬁeld, Germany Monti Cimini forest, Italy Rome Maize ﬁeld, Germany Rural Alabama, Loblolly pine plantation Rural Alabama, Loblolly pine plantation Rural Germany, Waldstein, Norway spruce forest

Reference

Goldan et al. (1995a) Riemer et al. (1998) Baker et al. (2001) Karl et al. (2003) Hellen et al. (2004) Müller et al. (2006) Goldan et al. (1995) Riemer et al. (1998) Goldan et al. (1995a) Karl et al. (2003) Hellen et al. (2004) Goldan et al. (1995a) Hellen et al. (2004) Goldan et al. (1993) Fall et al. (2001) Karl et al. (2001a) Karl et al. (2001a) Wedel et al. (1998) Ciccioli et al. (1993) Ciccioli et al. (1993) Wedel et al. (1998) Goldan (1995a) Goldan (1995a) Müller et al. (2006)

in the lower troposphere at such low concentrations. Emissions of these aldehydes may therefore be nearly permanent. Seasonal variations in the mixing ratios of formaldehyde and acetaldehyde over European rural sites were reported by Solberg et al. (1996). Again, typical mixing ratios ranged between 1 and 4 ppb for formaldehyde and 0.2–1.5 ppb for acetaldehyde depending on the site and season. Summer maxima were observed at all sites, which is in contrast to the seasonal

Oxygenated Volatile Organic Compounds

133

variations of other VOCs. This may reﬂect on the importance of primary biogenic emissions over secondary formation due to the oxidation of precursor VOC. Latitudinal and seasonal distributions of formaldehyde were measured over the Paciﬁc and Indian Oceans by Arlander et al. (1990). Mixing ratios varied between 0.7 ppb at mid-northern latitudes and 0.2 ppb at mid-southern latitudes. Meridional cross-sections of the global distribution of formaldehyde and acetaldehyde were measured during the TROPOZ II (Tropospheric Ozone) aircraft campaign covering a latitude range between 70◦ N and 60◦ S, and an altitude range between ground level and about 12 km (Arlander et al. 1995a, 1995b). To our knowledge, these are the only available latitudinal cross-sections of these compounds. For formaldehyde, the data showed a vertical decrease with altitude at all latitudes and a broad maximum around the equator. The measured mixing ratios were considerably higher than those expected from the oxidation of methane alone. Using a 2D model simulation, Arlander et al. could show that the formaldehyde distribution could be explained as soon as they included the formaldehyde formation from the oxidation of −C7 hydrocarbons. The CH3 CHO mixing ratios also showed a decrease with altitude, C2 − but a latitudinal decrease from northern to southern latitudes, in contrast to formaldehyde, obviously as a consequence of higher mixing rations of precursor hydrocarbons in northern latitudes. Figures 4.1 and 4.2 show the 2D distributions of formaldehyde and acetaldehyde, as observed by Arlander et al. Vertical proﬁles of formaldehyde over the North Atlantic were reported by Fried et al. (2002). Background formaldehyde mixing ratios were around 230 ppt in the free troposphere and decreased between 35◦ N and 50◦ N from about 500 ppt down to 270 ppt in the marine boundary layer. However, they also report elevated mixing ratios of up to 600 ppt in the 4–8 km altitude range. Singh et al. (2000) measured the vertical proﬁles of formaldehyde, acetaldehyde, and propionaldehyde over the Paciﬁc Ocean. Observed ground-level mixing ratios were about 500 ppt, 400 ppt, and 150 ppt for formaldehyde, acetaldehyde and propionaldehyde, respectively. They report decreasing mixing ratios with altitude (about a factor of four between ground level and 11 km altitude) for all three aldehydes. Except for formaldehyde, GEOSCHEM (Goddard Earth Observing System Chemical)-model-simulated vertical proﬁles showed much lower mixing ratios than measured, indicating that it is not possible to explain the distribution of these compounds with our current knowledge of atmospheric chemistry. New satellite instruments provide the opportunity to measure global column densities of selected OVOCs. As an example, Figure 4.3 shows column densities of formaldehyde derived from measurement of the SCIAMACHY (Scanning Imaging Absorption Spectrometer for Atmospheric Chartography) instrument aboard ESA’s (European Space Agencies) environmental research satellite (ENVISAT) (Wittrock et al. 2006). Observations of higher-molecular-weight aldehydes are very limited. Typically, mixing ratios are in the range of some 100 ppt down to a few ppt. Gas-phase concentrations of these compounds are discussed in detail in Chapter 9. Alcohols. Several studies show that, in rural areas, methanol is the OVOC with the highest concentration. According to current estimations, methanol seems to be the OVOC with the highest global emissions. Since methanol also exhibits a relatively long lifetime, it is the second most abundant volatile organic compound in the atmosphere. Methanol has anthropogenic and natural sources and can be found in polluted as well as in remote areas.

134

Volatile Organic Compounds in the Atmosphere

(a)

10 50

64± 22

55± 21 50± 21

Altitude (km)

8

6

4

220±

214± 35

0

137±

105±

53±

28

31

27

21

55± 21

35±

142±

113±

23

29

39(3) 118± 27

91± 25

151± 39

234±

393±

356± 47

51 314±

690±

116(2)

41

60

195± 33

195± 33

35

200

2

202± 34

13(2)

35 240± 33

54(2)

439± 56

242±

313±

1250±

37

43

154

500

40

400±

20

100

247± 37 191±

135±

33

29

972(2)

1000

374± 44(2)

0

67± 73

20

8 121±

140±

24

25

100

29

141± 29

141±

147±

145± 62(2)

30

193± 33

135± 29

76

118± 36(2)

261±

326±

233±

12

39

44

36

2 394± 71(2)

698± 85

995± 125(2)

713± 95(2)

500

S

229± 21

386± 30

1711±

169±

391±

116±

122(2)

549(4)

146

51

33

0

105± 27

356± 47

534±

1439±

20

77± 26(2)

106± 57(2)

21(3)

100

144± 26

581± 72

1000 300(4)

0 40

211± 35

1031± 696± 545(6) 113(2) 760± 187(2)

53±

117± 28

243± 64(4)

293± 42(4)

51± 21

104± 27

71± 24

376±

527± 86

N

50

207± 34(2)

200

60

60

22

241± 37

187± 33

235± 36

200

40

99± 26

141±

124± 24

6

4

500

56±

64± 25(2)

79± 24

34

335± 66

615± 76

10

199±

140± 30

240± 40

580± 72 1292±

50

Latitude (degrees)

65± 21

Altitude (km)

221±

714± 46

S (b)

19

115±

100

156± 30 215± 41

13±

20

Latitude (degrees)

274± 40

151± 30

171± 66

200

500 40

60 N

Figure 4.1 Meridional cross-sections of the mixing ratios of formaldehyde as observed during the TROPOZ II campaign in January 1991. (a) South-bound ﬂights, (b) North-bound ﬂights. For further details, see Arlander et al. (1995a). With permission from Springer.

Oxygenated Volatile Organic Compounds

135

(a) Altitude (km)

57± 23

4

40± 19

50 93± 28

73± 26

2

63± 45(2)

80

0

60

42± 20

105± 30

80

58± 23

40

75± 26

27± 16 44± 25(2)

50± 22(2)

60± 23

50± 21

20

S

53± 22

60± 23

80

91± 28

41± 22± 10(2) 32(2) 46± 21

0 20 Latitude (degrees)

40

60 N

(b) Altitude (km)

70± 25

4

50± 26

80 2

90± 28

100

0

95± 44(2)

60

40± 26

41± 28

95± 41(2)

112± 114± 46(2) 32(2)

40 S

40± 26

40± 19

100

80

175± 38

123± 131± 73(6) 58(2)

20

131± 43(3)

161± 69(2)

120± 31

106± 30

91± 39(4)

150

177± 39

146± 40

44± 20

104± 49(3)

139± 34

0 20 Latitude (degrees)

40

60 N

Figure 4.2 Meridional cross-sections of the mixing ratios of acetaldehyde as observed during the TROPOZ II campaign in January 1991. (a) South-bound ﬂights, (b) North-bound ﬂights. For further details, see Arlander et al. (1995b). With permission from Springer. 60

HCHO (molecule/cm2)

Latitude

30

2.0 × 1016 1.6 × 1016 1.2 × 1016 8.0 × 1015 4.0 × 1015 0.0

0

−30 −60 −180 −150 −120 −90

−60

−30 0 30 Longitude

60

90

120

150

180

Figure 4.3 The global distribution of formaldehyde derived from SCIAMACHY measurements from August 2004–July 2005 (with permission from F. Wittrock et al. 2006). This image appears in full colour in the plate section that follows page 268 as Plate 2.

In rural areas and over forests, methanol mixing ratios of several ppb are frequently observed. Even values up to 22 ppb have been reported (Riemer et al. 1998). Quite low concentrations of 0.1–1 ppb were measured during a ﬁeld campaign at the Feldberg, Germany, during August and September 2000 (Folkers 2001). Methanol emissions from plants are attributed to growth processes, and it is therefore not surprising that maximum methanol concentrations in rural areas are observable in spring (Karl et al. 2003).

136

Volatile Organic Compounds in the Atmosphere

Compared with methanol, less information exists for ethanol. In rural areas, ethanol mixing ratios range from below 100 ppt to 1.3 ppb (Riemer et al. 1998) and from 100 ppt up to 9.6 ppb (Schade and Goldstein 2001). The main source of ethanol in rural areas is of biogenic origin. Ethanol is emitted from higher plants, particularly under anoxic conditions at the plants roots (Kreuzwieser et al. 1999, 2000). Since these conditions at the roots are extremely variable, the source strength for ethanol and thus atmospheric concentrations will also be variable. The source strength is also highly variable for another group of alcohols. These are the C5 - and C6 -alcohols emitted from plants when membranes are damaged. Examples for this are 1-penten-3-ol or (Z)-3-hexenol. These compounds are synthesised within an enzyme sequence where LOX is a central enzyme. Hence, these compounds are often denominated as LOX products. Since hard stress leading to membrane damage and LOX activity is transient, the emissions are also transient. Such emissions are quite easily detectable during cutting or drying of grass (de Gouw et al. 1999). Atmospheric concentrations therefore occur as pulses with ambient concentrations up to more than 6 ppb for individual LOX products (Karl et al. 2001a). This is quite different for another alcohol, 2-methyl-3-buten-2-ol (MBO). MBO emissions are constitutive and depend on light intensity and temperature in a similar way as isoprene emissions. Emissions are nearly zero during darkness and increase with light intensity. Thus, MBO is observed in concentrations from below measurable up to 7 ppb in the air over strong sources such as Ponderosa pine (Schade and Goldstein 2001). Ketones. The most abundant ketone in the atmosphere is acetone. Except methyl ethyl ketone and methyl vinyl ketone concentrations, which can exceed several ppb, those of all other higher-molecular-weight ketones are lower. For detailed information on these compounds, we refer to Chapter 9 of this book. Local mixing ratios of acetone can be rather high. In polluted areas, values of 10–20 ppb are frequently observed; mixing ratios even up to 70 ppb have been measured. In rural areas, mixing ratios drop to values of several hundred ppt to a few ppb (see Table 4.1). There are a number of studies of acetone concentrations in the free troposphere. Singh et al. (1995) found relatively high concentrations of acetone (0.5 ppb) in the range of 5–10 km at mid-northern latitudes. These concentrations declined to 400 and 200 ppt, respectively, at southern latitudes. Arnold et al. (1997) compiled acetone measurements in the upper troposphere and lower stratosphere, obtained during several aircraft studies at mid-northern latitudes. They report acetone mixing ratios between 200 and 2000 ppt in the upper troposphere (8–13 km) in summer for the different measurement campaigns. During a winter campaign in February 1995, acetone mixing ratios were between 300 and 500 ppt in the upper troposphere, with a steep decrease above the tropopause, which was at about 10 km altitude at that time. Reiner et al. (1999) report vertical proﬁles of acetone mixing ratios over Germany in September and October 1991. They found mixing ratios between 100 and 1 ppb with a tropospheric average of 830 ppt. Above 10 km, acetone mixing ratios decreased with a scale height of 2.2 km, obviously reﬂecting increasing photolysis rates and reduced vertical transport in the lower stratosphere. Measured global distribution, such as the ones shown for formaldehyde and acetaldehyde, is not available for acetone. Singh et al. (1995) conducted model studies based on a total

Oxygenated Volatile Organic Compounds

137

acetone source of 49 Tg/year to derive a height–latitude cross-section over the central Paciﬁc at 175◦ E longitude. Carboxylic acids. The current available data on the abundance of carboxylic acids show that the lower-molecular-weight acids, formic acid and acetic acid, are ubiquitous compounds in the troposphere. Higher-molecular-weight carboxylic acids have also been observed in urban, continental and even in the remote troposphere. However, data on these compounds are sparse. Carboxylic acids are found in fog water, cloud water, rain water, snow and ice, in aerosol particles, and in the gas phase (Chebbi and Carlier 1996, and references therein). The highest values of 13 ppb for formic acid and 16 ppb for acetic acid observed so far were reported for Los Angeles air in 1986 (Grosjean 1989). In other urban and rural areas, formic and acetic acid mixing ratios are in the range of several hundred ppt to a few ppb (cf. Grosjean 1989; Hartmann et al. 1989; Kawamura et al. 1985; Lawrence and Koutrakis 1994; Puxbaum et al. 1988). Similar values were observed over the Amazonian rain forest, with the mixing ratios of formic and acetic acid being four times higher in the dry season than in the wet season (Andreae et al. 1988; Talbot et al. 1990). Even in the remote troposphere, mixing ratios up to 1 ppb were observed with pronounced north-south and vertical gradients, obviously reﬂecting the gradients of the precursor molecules (Arlander et al. 1990; Reiner et al. 1999). The data base for higher-molecular-weight carboxylic acids in the ambient atmosphere is even sparser. There are a few measurements reporting the observation of propionic acid (Kawamura et al. 1985) and pyruvic acid (Andreae et al. 1988), and propionic and n-butyric acid (Satsumabayashi et al. 1989). The mixing ratios were in the range of ten to a few hundred ppt. Dicarboxylic acids are less volatile and are mainly observed in the particulate phase. Kawamura and Kaplan (1990) report that oxalic acid can be observed in the gas phase at elevated temperatures. A compilation of all studies dealing with these compounds can be found in Chebbi and Carlier (1996). In summary, except for some speciﬁc compounds such as light aldehydes, methanol and acetone, it is not yet possible to draw a clear picture of OVOCs’ tropospheric distribution and seasonal variations.

4.3

Sources of OVOCs

OVOCs are emitted into the atmosphere from a variety of sources. Anthropogenic sources are automobile exhaust, industrial processes, solvent evaporation and, to a certain extent, biomass burning. On the other hand, the terrestrial biosphere seems to play an important role in emitting these compounds into the atmosphere. As we will see, plants are able to synthesise these compounds and emit them into the atmosphere, even in a much larger variety than anthropogenic sources. All these sources are primary sources, because the OVOCs are emitted directly into the atmosphere. In addition to the emission from primary sources, OVOCs are formed in the atmosphere in the oxidation pathways of other volatile organic compounds. Even during the oxidation of higher-molecular-weight OVOCs, lower-molecular-weight OVOCs are formed. Thus, we have to deal with what we call secondary sources of these compounds. Under ambient conditions, it is often impossible to distinguish between the different origins of these

138

Volatile Organic Compounds in the Atmosphere

compounds. In the following part, we discuss the different primary and secondary sources of OVOCs as well as their contribution to the global budget, and try to summarise the present state of knowledge.

4.3.1

Anthropogenic emissions

Among the industrial sources that have been identiﬁed to emit OVOCs directly into the atmosphere are reﬁning and petrochemistry, coal chemistry, paint and varnish industry, sewage treatment and even coffee-roasting (cf. the early but still relevant review by Carlier et al. 1986 and references therein). Saturated alcohols have been used in large quantities as industrial solvents. In the context of air quality, anthropogenic OVOC emissions may have a considerable impact on a local or regional scale. For instance, in Brazil, a nationwide alcohol fuel program has been implemented. Methanol, neat ethanol or ethanol–gasoline mixtures are widely used as fuel for vehicles (de Paula Pereira 1999). Furthermore, OVOCs, especially ethers, are frequently used as fuel additives, such as methyl-tert-butyl ether (MTBE), ethyl-tert-butyl ether (ETBE), tert-butyl alcohol (TBA) and tert-amyl-methyl ether (TAME). A variety of anthropogenic sources is discussed for carboxylic acids (cf. Graedel et al. 1986). Among those are animal waste, combustion of synthetic materials, industrial emissions, waste incineration, automobile exhaust and tobacco smoke. Most of these sources are of local importance. There are only a few estimates of the global source strengths from anthropogenic emissions, for example, for methanol and acetone. A global industrial production of methanol of 24.3 Tg/year was reported by Crocco (1997) for 1996. Based on this ﬁgure, a global emission of about 3.7 Tg/year was estimated. This is in agreement with global emission rates of about 4 Tg/year estimated by Singh et al. (1995) based on atmospheric observations. For acetone, Singh et al. (2000) estimate a global emission rate of about 2 Tg/year from anthropogenic sources. These numbers are relatively low compared to other sources. Müller (1992) estimated a global annual emission rate of anthropogenic VOC of about 150 Tg/year. This ﬁgure is the sum of about 100 Tg/year from ‘technological sources’ and about 50 Tg/year from ‘biomass burning’ emissions. Although Müller (1992) does not specify the compounds included in this estimate, he states that this ﬁgure covers all hydrocarbons from all categories of sources. Since OVOCs are included in this estimation, their global annual emission rate can be expected to be somewhere below this ﬁgure. An important contribution to the budget of OVOCs in the atmosphere is due to biomass burning emissions. Biomass burning is associated with deforestation, savannah burning, forest wild ﬁres, agricultural practices and biofuel combustion. Since most of these processes are initiated and controlled by humans, this source is considered as anthropogenic emissions. In the last years, a few studies investigated the emission of OVOCs from this source. Koppmann et al. (1997) reported that the abundance of OVOC in biomass-burning plumes competed with or even exceeded that of other hydrocarbons. Based on this data, they were able to derive emission ratios relative to carbon dioxide for a variety of alcohols, aldehydes, ketones, carboxylic acids and esters. New analytical techniques allowed improving the in-situ measurements of these compounds in biomass burning plumes. For example, Holzinger et al. (1999) used proton

Oxygenated Volatile Organic Compounds

139

transfer reaction mass spectrometry (PTR-MS) to measure oxygenated compounds over Surinam, South America. They report emission ratios relative to carbon monoxide for formaldehyde, acetaldehyde, methanol and acetone. Yokelson et al. (1999) and Goode et al. (2000) identiﬁed formic and acetic acid as well as a number of other oxygenated compounds such as methanol and hydroxyethanal in biomass burning emissions using a Fourier transform infrared spectrometer (FTIR) in laboratory and ﬁeld experiments. Other studies also report signiﬁcant emissions, mainly of formic and acetic acid (Helas et al. 1992; Lacaux et al. 1993). However, to our knowledge, there are no estimates of source strengths for these compounds available up to now. Emission factors and emission ratios including OVOCs from biomass burning were reviewed by Andreae and Merlet (2001). To our knowledge, there is no speciﬁc estimate of the contribution of biomass burning to the global budget of OVOCs available. There are a few estimates of global total VOC emission rates from biomass burning that, in combination with observations, may be used to derive a rough estimate of OVOC emission from this source. The ﬁrst estimation was made by Crutzen et al. (1979). They calculate a global emission rate of nonmethane hydrocarbons of 28 Tg C/year, with an uncertainty of a factor of three. Lobert et al. (1991) calculated a global emission rate of 42 Tg C/year based on laboratory studies of burning of different types of biomass. Based on ﬁeld studies of different types of ﬁres, Koppmann et al. (1997) estimated a global VOC emission rate from biomass burning ranging between 25 and 48 Tg (compounds)/year. These ﬁgures are in the same range as the ﬁgure given by Müller (1992), who based his estimate on the data given by Crutzen et al. (1979), but extended his estimate to other biomass burning contributions. Taking the observation that about 30% of the emitted VOCs are oxygenated compounds (Koppmann et al. 1997), this would add 7.5–15 Tg/year of OVOCs from biomass burning to the total source strength.

4.3.2

Emissions from terrestrial vegetation

Plants are by far the largest natural source of OVOCs in the atmosphere, which warrants a more detailed discussion of the formation and emission mechanisms in a bit more detail here. The compounds emitted are aldehydes, ketones, alcohols and esters. A number of earlier studies have identiﬁed OVOCs in the emissions from forests (Isidorov et al. 1985), from agricultural plants (Arey et al. 1991; Winer et al. 1992) and from a variety of agricultural plants, trees and grassland (König et al. 1995). These studies were reviewed by Puxbaum (1997). Saturated and unsaturated alcohols are directly emitted into the atmosphere by vegetation. The emission of methanol from plants was ﬁrst reported by Fehsenfeld et al. (1992). MacDonald and Kimmerer (1993) observed emissions of methanol from a variety of trees in a similar order of magnitude as isoprene emissions. Since 1993, it is known that 2-methyl-3-buten-2-ol (MBO) is emitted from plants. In a pine forest in Colorado, Goldan et al. (1993) and Harley et al. (1998) observed MBO mixing ratios in the range of several ppb. MBO showed a diurnal cycle as did isoprene; the mixing ratio, however, exceeded that of isoprene by a factor of 5–8. Schade et al. (2000) measured canopy scale ﬂuxes of MBO over a ponderosa pine plantation in California of up to 2 mg C/m2 /h in July and August. On this local scale, this value corresponds to about 0.5% of the gross carbon uptake. MBO thus makes up 20% of the total OH reactivity (Lamanna and Goldstein 1999) and about 50% of the acetone yield (Ferronato et al. 1998).

140

Volatile Organic Compounds in the Atmosphere

A number of compounds, which are commonly referred to as leaf alcohols and leaf esters, have been observed to be present in ambient air. Examples of such compounds are (Z)-3-hexen-1-ol and (Z)-3-hexenylacetate. Additionally, higher-molecular-weight leaf aldehydes such as (Z)-3-hexenal or (E)-2-hexenal are co-emitted with the leaf alcohols. Other alcohols have been observed under various conditions, for example, linalool during the blossoming period of orange trees (Arey et al. 1991). −C10 -OVOCs. Ciccioli et al. Another group of OVOC are saturated aldehydes, mainly C6 − (1993) found these compounds in the atmosphere, and from vertical distributions of the OVOC concentrations they concluded that these OVOCs are emitted from vegetation. Owen et al. (1997) observed direct emissions of hexanal, nonanal and decanal in measurements with plants using a plant enclosure technique. Emissions of these saturated aldehydes were also found from plants under insect attack and from plants exposed to ozone (Karl et al. 2005; Wildt et al. 2003). The main contribution to biogenic emissions occurs in tropical regions, which have a high leaf area index, almost no seasons and more or less constant high temperatures (Guenther et al. 1995). Compared to (see Chapter 3), little is known about the biosynthesis and emission mechanisms of OVOCs. However, direct emissions from plants are known for a couple of OVOCs. Table 4.3 gives an overview of observed emissions for the most abundant OVOCs obtained from ﬁeld measurements. Table 4.4 summarises results from enclosure and laboratory studies.

4.3.2.1

Biosynthesis of OVOCs

Plants synthesise and emit a large number of volatile organic compounds (Gerzhenzon and Croteau 1991; Knudsen et al. 1993; Puxbaum 1997). These biogenic VOCs are the products of different physiological processes, which take place in different parts of the plants. The biosynthesis of the most abundant biogenically emitted VOCs, isoprene and monoterpenes, has been thoroughly investigated and is well understood (Fall 1999; Kesselmeier and Staudt 1999; Kreuzwieser et al. 1999; Lichtenthaler 1999). Despite a growing number of studies, our knowledge concerning the biosynthesis of OVOCs is still incomplete. In the following, we give an overview of biosynthetic pathways of OVOC formation in the order of increasing numbers of carbon atoms. C1 -OVOCs (methanol, formaldehyde and formic acid). C1 compounds are believed to be synthesised during developmental processes of plants (Kreuzwieser et al. 1999). Meanwhile, it is widely accepted that methanol is synthesised in plants, although – like ethanol – it is toxic in higher concentrations (Joseph and Kelsey 2000). The most probable way of methanol formation seems to be pectin demethylation. Details of this process have been investigated in various studies (Levy and Staehelin 1992; Obendorf et al. 1990; O’Neill et al. 1990; Ricard et al. 1986). Other processes, such as mending of proteins (Mudgett and Clarke 1993) or degradation of lignin (Lewis and Yamamot 1990) seem to be less likely. Methanol emissions have been observed during cell expansion of roots, leaves, and fruits (Fall and Benson 1996), growth and maturation of seeds (Obendorf et al. 1990), cell degradation (Fall and Benson 1996; Levy and Staehelin 1992; Nemecek-Marshall et al. 1995), leaf abscission and senescence of plant tissues (Harrimen et al. 1991) and after exposing tobacco plants to very high ozone concentrations (Beauchamp et al. 2005). Methanol emissions thus seem to be related to growth processes, and also to cell repair. These processes occur both in

Oxygenated Volatile Organic Compounds

141

Table 4.3 Flux densities obtained from ﬁeld measurements Compound

Estimated ﬂux densities

Location of measurement/plant species

Reference

Baker et al. (2001) Karl et al. (2001b) Schade and Goldstein (2001) Warneke et al. (2002)

Karl et al. (2001b)

Methanol Methanol Methanol Methanol

1 mg/m2 /h 1–8.4 mg/m2 /h 2.9 mg (C)/m2 /h 0.5–1.5 mg/m2 /h

Methanol

0.5 (0.9) mg/m2 /h

Methanol Ethanol Ethanol Acetone Acetone Acetone

0–0.2 mg (C)/m2 /h 0.64 mg (C)/m2 /h 0.17 μg (C)/g(dw)/h 0.37 mg (C)/m2 /h 0.1–1.5 mg/m2 /h 0.5 (0.3) mg/m2 /h

Acetone Acetaldehyde Acetaldehyde Acetaldehyde

0.23 μg (C)/g(dw)/h 0.2 mg (C)/m2 /h 0.5–3 mg/m2 /h Mean 0.3 (0.2) mg/m2 /h

Acetaldehyde Acetaldehyde C5 -LOX-Products + 2-methyl-butanal C6 -LOX-Products

0.5–100 nmol/m2 /min 0.19 μg (C)/g(dw)/h 0.1–0.9 mg/m2 /h

Subalpine conifer forest Hayﬁeld, Austria Ponderosa pine Alfalfa ﬁeld before and after harvesting Mixed hardwood forest, Michigan Field in Germany Ponderosa pine Norway spruce Ponderosa pine Hayﬁeld, Austria Mixed hardwood forest, Michigan Norway spruce Ponderosa pine Hayﬁeld, Austria Mixed hardwood forest, Michigan Spruce forest, Germany Norway spruce Hayﬁeld, Austria

0.1–1.5 mg/m2 /h

Hayﬁeld, Austria

Karl et al. (2003) Schade and Custer (2004) Schade and Goldstein (2001) Grabmer et al. (2006) Schade and Goldstein (2001) Karl et al. (2001b) Karl et al. (2003) Grabmer et al. (2006) Schade and Goldstein (2001) Karl et al. (2001b) Karl et al. (2003) Cojocariu et al. (2004) Grabmer et al. (2006) Karl et al. (2001b)

aboveground and belowground parts of plants. Due to its high water solubility, methanol can diffuse out of cells and enter the transpiration stream. Hence, methanol produced in roots can be transported to the leaves by the xylem ﬂow and subsequently emitted. Formaldehyde and formic acid are most probably produced in an oxidation process from methanol (Fall 1999). Kesselmeier et al. (1998) report a light-dependent emission either directly caused by a light-dependent process or by the light dependence of stomatal conductance. It is assumed that formic acid is produced from methanol by the enzyme methanol oxidase. The oxidation processes may be dependent on photosynthetic active radiation (PAR). C2 -OVOCs (ethanol, acetaldehyde and acetic acid). Ethanol is synthesised in the plant’s roots due to alcoholic fermentation. With the xylem transportation stream, it is transported to the leaves where oxidation by alcohol dehydrogenase forms acetaldehyde in a ﬁrst step and acetic acid in a second step (e.g. Kreuzwieser et al. 1999). Possibly, acetic acid is also formed from acetyl coenzyme A (Kesselmeier and Staudt 1999). Biosynthesis and emission of these compounds seem to appear mainly during stress situations, particularly during anaerobic

142

Volatile Organic Compounds in the Atmosphere

Table 4.4 Flux densities obtained from enclosure or laboratory measurements Compound

Fluxes

Methanol

1.5–45 μg/g (dw)/h

Methanol Methanol Formaldehyde Formic acid Ethanol Acetaldehyde

6.8–34 μg/g (dw)/h 0.03–7.6 mg (C)/m2 /h 382–589 ng/g (dw)/h 221–434 ng/g (dw)/h 0.001–0.67 mg (C)/m2 /h 0.3–0.8 mg/m2 /h

Acetaldehyde Acetic acid Acetone Acetone

691–1034 ng/g (dw)/h 36.6–289 ng/g (dw)/h 7.4 ± 1.5 μg/g (fr wt)/h 0.1–0.2 mg/m2 /h

Acetone Butanone Butanone

7.5 × 10−11 mol/m2 /s = 0.06 μg/g (dw)/h 0.1–10.3 mg (C)/m2 /h 11–80 μg/g (dw)

LOX-products

0–10−9 mol/m2 /s

LOX-products (Z)-3-Hexenol

≤23 000 μg (C)/m2 /h ≤1.3 μg/h/g (dw)

(Z)-3-Hexenal + (E)-2-Hexenal Nonanal

30–240 μg/g (dw) ≤0.014 μg/g (dw)/h

Species Eleven different species Russian olive, Aspen Clover, Grass Holm oak and pine Holm oak and pine Clover, Grass Bluegrass, orchard grass Holm oak and pine Holm oak and pine Abies concolor Bluegrass, orchard grass Scots pine Clover Bluegrass, orchard grass Tobacco, sunﬂower, tomato, Clover and grass Eighteen agricultural and two natural plant species Bluegrass and Orchard grass Sunﬂower, Scots pine

Reference MacDonald and Fall (1993b) Nemececk-Marshall et al. (1995) Kirstine et al. (1998) Kesselmeier et al. (1997) Kesselmeier et al. (1997) Kirstine et al. (1998) Karl et al. (2001b) Kesselmeier et al. (1997) Kesselmeier et al. (1997) MacDonald and Fall (1993a) Karl et al. (2001b) Shao et al. (2001) Kirstine et al. (1998) Karl et al. (2001b) Heiden et al. (2003) Kirstine et al. (1998) Arey et al. (1991)

Karl et al. (2001a) Wildt et al. (2003)

conditions in the plants’ roots. For example, young poplar trees emit high amounts of acetaldehyde upon ﬂooding of the roots (Kreuzwieser et al. 1999). It is known that the production of ethanol and acetaldehyde also increases as soon as the plant encounters drought stress, extreme cold and high concentrations of ozone or SO2 . Stress-induced emissions are already described by Kimmerer and Kozlowski (1982). Strong increases of acetaldehyde emissions are observed during abrupt light–dark transitions. However, the change of PAR in the environment is much more moderate and insufﬁcient to make this effect important under natural conditions (Graus et al. 2004). C3 -OVOCs. The most important C3 -OVOC known to be emitted from plants is acetone (Fall 1999). The biosynthetic pathways of acetone production are still not known. Acetone is speculated to be synthesised by decarboxylation from acetoacetic acid as an intermediate (Fall 1999; Kesselmeier and Staudt 1999). Another C3 -compound that seems to be directly

Oxygenated Volatile Organic Compounds

143

emitted by plants is propanol. Kirstine et al. (1998) report emissions of propanol from pastures, even though these emissions were low and seemed not to be of large signiﬁcance. C4 -OVOCs. 2-Butanone is the only a C4 -OVOC that was found to be directly emitted from higher plants (Karl et al. 2001b, Kirstine et al. 1998). Again, not much is known about the biosynthetic pathways leading to the formation of 2-butanone and its emission mechanisms. C5 -OVOCs. Zeidler and Lichtenthaler (2001) have shown that MBO is synthesised from dimethyl-allyl-diphosphate (DMAPP) as is isoprene. Goldan et al. (1993) found very high MBO emissions from conifers in Colorado. They observed mixing ratios of 2–3 ppb in ambient air and diurnal variations with maxima around noon and minima during darkness. The emissions were found to be parallel to those of isoprene, which may be a consequence of similar internal plant formation processes. Since DMAPP is also the precursor of isoprene, it is not surprising that the emissions of MBO are determined by PAR and temperature in a similar way as those of isoprene (Harley et al. 1999). Other C5 -OVOCs emitted from vegetation are 1-penten-3-ol, 1-penten-3-one or 3-pentanone. These C5 -OVOCs are synthesised via the octadecanoid pathway (Fisher et al. 2003). Fall et al. (2001) measured high atmospheric concentrations of C5 -OVOCs after a hard freeze event. These compounds showed concentrations in the range of several ppb over a time period of 5 days and diurnal variations with maximum concentrations during darkness. These compounds belong to the family of LOX products, which will be addressed in the following section. C6 -OVOCs. The octadecanoid pathway is also the origin of C6 -OVOCs, often termed as green leaf volatiles, wound-VOC or LOX products. C6 -OVOCs are well known to most of us since these compounds are emitted upon cell wounding, giving the typical odour during the mowing of grass. The biochemical formation of these compounds is well understood (e.g. Croft et al. 1993; Gardner 1995; Hatanaka 1993). They are formed from the common lipid constituents of plant membranes in a sequence of enzyme reactions. Substrates for these OVOCs are free fatty acids consisting of 18 carbon atoms. These fatty acids are oxidised by LOX enzymes. The hydroperoxides formed in this oxidation process are decomposed by lyase enzymes, and the C6 -OVOCs are formed. Depending on the position where the oxygen is added to the substrate, C9 -OVOCs can also be produced and emitted. Emissions of LOX products are typically observed after wounding (Buttery and Ling 1993; Fall et al. 1999), during drying (de Gouw et al. 1999), after pathogen attack (Croft et al. 1993) and after exposure to ozone at elevated levels (Heiden et al. 1999). The main volatiles produced within this pathway during acute stress situations are the C6 -OVOCs. Emissions of C5 - and C9 -OVOCs are lower than those of the C6 -OVOCs. However, as shown at the example of pathogen-infested tobacco plants, emissions of C5 -LOX products may last longer than those of C6 -LOX products (Heiden et al. 2003). For a given plant species, the emission pattern was found to be independent of the elicitor (Heiden et al. 2003). This behaviour might be understood from the basic processes. Once the free fatty acids are produced, the same processes lead to the emission of these OVOCs, independent of the exact elicitor leading to the production of the free fatty acids. However, except for ozone uptake with very high ozone ﬂuxes (Beauchamp et al. 2005), a deﬁnite reason for the emission strength of these OVOC has not yet been identiﬁed.

144

Volatile Organic Compounds in the Atmosphere

C10 -OVOCs. Oxygenated monoterpenes can be termed as C10 -OVOCs. Examples are verbenone, myrtenal (C10 H14 O), 1,8-cineole (also known as eucalyptol) or linalool (C10 H18 O). Like MBO, these compounds are also synthesised via DMAPP, implying that they are emitted constitutively. This also implies that the emission strengths may be described with phenomenological algorithms, considering temperature and light intensity as variables determining the emissions. The emission rates of oxygenated monoterpenes seem to be lower than those of non-oxygenated ones. −C10 Saturated aldehydes. Another group of OVOCs are saturated aldehydes (mainly C6 − OVOCs). Owen et al. (1997) showed direct emissions of hexanal, nonanal and decanal in measurements with plants using a plant enclosure technique. These compounds have also been observed in the ambient rural atmosphere (Wedel et al. 1998). Emissions of these OVOCs are also found from plants exposed to ozone or from plants under insect attack (Wildt et al. 2003). It is still unclear whether these aldehydes are synthesised by enzymatic processes or by nonenzymatic oxidative processes inside the plants. Aromatic compounds. Aromatic OVOCs are also emitted by plants. Some of these compounds are synthesised within the phenyl propanoid pathway. Methyl salicylate (C8 H8 O3 ) is the predominant OVOC emitted from this pathway (e.g. Heiden et al. 1999; Seskar et al. 1998). Methyl salicylate is known as a signal molecule for plants (Shulaev et al. 1997); however, there is only one observation of such compounds in the atmosphere known to us (Th. Hoffmann, unpublished results).

4.3.2.2

Emission mechanisms and algorithms

When we discuss OVOC emissions from vegetation, we can distinguish between two classes of OVOCs – OVOCs emitted constitutively and those emitted particularly during stress conditions. Emissions are termed as constitutive emissions if they appear permanently with more or less the same emission strengths, that is the emissions are similar at the same external conditions. Constitutive emissions can be described by phenomenological algorithms. Such algorithms describe the dependence of emission rates on temperature and PAR. The absolute emission rates are considered by standard emission rates often termed as basal emissions or emission factors at ﬁxed temperature and PAR. Phenomenological algorithms are successfully used to describe these dependencies for VOC originating from DMAPP, such as isoprene or monoterpenes (Guenther et al. 1993, 1995; Schuh et al. 1997; Tingey et al. 1991). These algorithms should also be applicable to describe the emission rates of OVOCs originating from DMAPP. This is the case for MBO. To our knowledge, no algorithms have been reported for oxygen-containing monoterpenes. Emissions of oxygenated monoterpenes seem to be lower than those of nonoxygenated ones. For emissions from Quercus ilex and Pinus pinea, Kesselmeier et al. (1997) report that oxygenated monoterpenes contribute between 10 and 26% to the total monoterpene emissions. More data regarding the applicability of phenomenological algorithms to describe the emission rates of other OVOC are sparse. Acetone emissions from pine (Pinus sylvestris L.) depend on temperature as well as on PAR (Shao and Wildt 2002; Shao et al. 2001). These dependencies are also described by a phenomenological algorithm. Kesselmeier et al. (1997) found emissions of formaldehyde, formic acid, acetic acid and acetaldehyde from

Oxygenated Volatile Organic Compounds

Methanol

2.0×10−9

50 1.6×10−9

40

1.2×10−9

30 20

8.0×10−10

10

Temperature (°C)

Φ(Methanol) (mol /m2/s1)

145

4.0×10−10 0 12:00

12:00

12:00 12:00 12:00 Daytime (hh:mm)

12:00

12:00

Figure 4.4 Diurnal course of methanol emissions from beech (open squares, left scale). The dotted line indicates the period of illumination. During illumination was 480 μmol/m/s. At day 3, the temperature (black line, right scale) was increased from 20◦ C to 30◦ C, leading to an additional increase in methanol emissions.

Quercus ilex. The diurnal variations of these emissions implied that the emissions are driven by temperature as well as by PAR. The emission rates could be described with the algorithm of Guenther et al. (1993, 1995). However, emissions of formaldehyde and acetaldehyde from pine seemed to be independent of PAR. Also, methanol emissions appear to be permanent and may thus be considered as constitutive emissions. As a typical example, Figure 4.4 shows the diurnal variation of methanol emission from beech measured in a plant chamber under deﬁned conditions. The emission rates clearly show light and temperature dependence. Methanol emissions are related to growth processes. Growth processes may be triggered by temperature or PAR, but they follow plant internal diurnal cycles (Walter and Schurr 2005). Dependencies of methanol emissions on temperature or PAR, therefore, are superimposed by diurnal growth cycles and variations of stomatal aperture. The latter impacts the equilibrium of methanol concentrations between the gas phase in the substomatal cavity and the apoplastic water (Niinemets and Reichstein 2003). From the high water solubility of methanol, it is evident that methanol emissions cannot be described by phenomenological algorithms as they are used today. Emissions of many OVOCs are induced by elicitors such as stress. The impacts of these elicitors vary with time. Hence, such induced emissions are not continuous and cannot be described by phenomenological algorithms developed so far. Only in case of ozone as −C10 -aldehydes (Wildt et al. the elicitor, continuous emissions are observed for the C6 − 2003). The emission rates correlate quite well to each other (except of hexanal in some cases because hexanal emissions may also be due to LOX activity). Also, correlations between their concentrations are observed in the atmosphere (Ciccioli et al. 1993; Wedel et al. 1998). These ﬁndings imply similar formation processes and similar emission mechanisms for these OVOCs. The continuity of these emissions allows investigating their dependence on external parameters and, thus, to describe the emissions in dependence of those variables affecting

146

Volatile Organic Compounds in the Atmosphere

the emissions. In case of ozone exposure as an elicitor, temperature and ozone ﬂux dens−C10 -aldehyde emissions. Despite their ity had the most pronounced impacts on the C6 − unknown origin, the emissions of these aldehydes can be described using a phenomenological algorithm. In case of insect or pathogen attack, it is still unknown what measured category can be used as a reference. Thus, in case of these elicitors, an applicable algorithm is missing. The emissions of many other OVOCs are transient. This transient character nearly rules out a quantitative description of stress-induced OVOC emissions.

4.3.3

Natural sources other than vegetation

Besides terrestrial vegetation, other sources also emit OVOCs in non-negligible amounts. Schade and Custer (2004) showed that methanol is emitted from bare, ploughed soil. They determine emission rates up to 0.2 mg (C)/m2 /h, which is about an order of magnitude less than methanol emissions reported for vegetation. An OVOC source with signiﬁcant strength and thus with signiﬁcant impact on the troposphere is decay of dead plant material. Warneke et al. (1999) found that wetting of dried leaves, grass and conifer needles leads to emissions of many OVOCs. Especially methanol and acetone are formed. This source could account for global annual emissions of 6–8 Tg of acetone and 18–40 Tg of methanol. Also, human breath is a source of OVOCs. Since more than two decades, selected VOCs are measured as diagnostic tool for certain diseases. Among these, VOC acetone, acetaldehyde and ethanol play a key role. With mixing ratios of several hundred ppb, acetone concentrations in exhaled human breath can be quite high (Pleil and Lindstrom 1995). Barker et al. (2006) analysed exhaled breath samples to investigate the patterns of speciﬁc VOC as a noninvasive diagnostic tool for cystic ﬁbrosis. As a general result of their study, they observed that, besides isoprene, acetone is the most abundant compound in exhaled human breath, with typical mixing ratios of >400 ppb. In addition to acetone, methanol, ethanol and 2-propanol proved to be the most abundant OVOCs in exhaled human breath, with mixing ratios of about 300 ppb, 200 ppb and 10 ppb, respectively. Most probably, acetone and other OVOCs can also be found in the breath of many animals. However, there are no reliable data and estimations of the importance of this source for the budgets of these compounds in the troposphere. Bacterial metabolism is discussed to be a source for a direct production of carboxylic acids. This process can occur in cloud or rain drops (Herlihy et al. 1987; Kawamura and Kaplan 1990) and in soil (Enders et al. 1992; Sanhueza and Andreae 1991; Talbot et al. 1990). Estimates are not available of the emission rates of any of these sources. Graedel and Eisner (1988) report emissions of carboxylic acids from formicine ants, although this does not seem to be an important source. Other insects too can produce aldehydes and ketones (Morgan and Tyler 1977). Recently, the oceans are being discussed to be a source for speciﬁc OVOCs, and a sink for others. To our knowledge, there are no reliable data on seawater concentrations available so far. However, some estimates have been published based on model calculations, taking into account atmospheric concentrations. Singh et al. (2003) evaluated airborne OVOC measurements obtained during TRACE-P over the Paciﬁc Ocean. Using a global 3D model and an air–sea exchange model, they estimated that the mixed layer of the ocean may be a

Oxygenated Volatile Organic Compounds

147

considerable reservoir for OVOCs. They conclude that the ocean is a source for aldehydes (125 Tg/year for acetaldehyde, 45 Tg/year for propanal) and a sink for acetone (−8 Tg/year) and methanol (−14 Tg/year). Marandino et al. (2005) estimated a global ocean uptake of acetone of 48 Tg/year, based on air/sea ﬂux measurements over the North Paciﬁc Ocean. This is a factor of three higher than previously estimated, and considering this change alone would have signiﬁcant consequences for the atmospheric acetone budget (see below). Assuming depths of 50 m for the well-mixed layer, they estimated an oceanic reservoir of 4, 1, 50 and 10 Tg for acetaldehyde, propanal, methanol and acetone, respectively. This accounts for 2–25 times their atmospheric burden. This is consistent with other model studies, which also come to the result that the oceans must be a considerable reservoir for acetone (de Laat et al. 2001) and methanol (Galbally and Kirstine 2002). In the same way, sunlight-initiated reactions decompose dissolved organic matter, leading to the formation of alkenes in the surface water of the oceans (cf. Ratte et al. 1993); a variety of oxygenated compounds seem to be formed in these reactions (Zhou and Mopper 1997). Altogether, the role of the oceans in the global budget of OVOCs remains unexplored up to now.

4.3.4

Photochemical OVOC formation

Besides the primary emission from anthropogenic and natural sources, OVOCs are formed in the atmosphere in the oxidation pathways of other volatile organic compounds (including OVOCs themselves). These secondary carbonyl compounds are of outstanding importance due to their role as radical promoters in the atmosphere. Carlier et al. (1986) have compiled the carbonyl compounds identiﬁed to be formed in the reaction pathways of nonmethane hydrocarbons derived from various experiments in simulation chambers under simulated atmospheric conditions. The most abundant secondary compounds are formaldehyde, acetaldehyde, benzaldehyde, acetone, glyoxal and some higher-molecular-weight aldehydes and ketones. The largest source of formaldehyde in the atmosphere is the oxidation of VOC due to the reaction with OH radicals. Besides formaldehyde, higher-molecular-weight aldehydes are formed in the oxidation of higher-molecular-weight VOC (cf. Atkinson 2000). Acetone is known to be formed during the oxidation of hydrocarbons with isostructure and monoterpenes (Alvarado et al. 1999; Chatﬁeld et al. 1987; Orlando et al. 2000; Reissel et al. 1999; Singh and Hanst 1981; Singh et al. 1994). Acetone was also observed as a product of the MBO oxidation (Alvarado et al. 1999). To a lesser extent, higher-molecular-weight VOC such as isoalkanes and isoalkenes with more than ﬁve carbon atoms also produce acetone (Singh and Zimmerman 1992). Goldstein and Schade (2000) have shown that 45% of the acetone mixing ratio in the Sierra Nevada (California, United States) can be attributed to biogenic sources, 14% to anthropogenic sources and 41% to the regional background. Of biogenic acetone, 35% was direct emissions from vegetation into the atmosphere, while 63% was formed in the oxidation of biogenic MBO and 2% in the oxidation of monoterpenes. Figure 4.5 gives a simpliﬁed reaction pathway leading to the formation of formaldehyde, acetaldehyde and acetone from the oxidation of methane, ethane and propane. Figure 4.6 is an example of the complex reaction chains following the oxidation or photolysis (see below) of a Cn -aldehyde leading to the formation of a Cn−1 -aldehyde. As can be seen, the different

148

Volatile Organic Compounds in the Atmosphere

CH4 + •OH

→

•CH 3

+ O2

→

CH3O2•

CH3O2• + NO

→

CH3O• + NO2

CH3O• + O2

→

HCHO + HO2•

•CH 3

+ H2O

CH3CH3 + •OH

→

CH3•CH2 + H2O

CH3•CH2 + O2

→

CH3CH2O2•

CH3CH2O2• + NO

→

CH3CH2O• + NO2

CH3CH2O• + O2

→

CH3CHO• + HO2•

CH3CH2CH3 + •OH

→

CH3•CHCH3 + H2O

CH3•CHCH3 + O2

→

CH3CHOO•CH3

CH3CHOO•CH3 + NO

→

CH3CHO•CH3 + NO2

CH3CHO•CH3 + O2

→

CH3COCH3 + HO2•

Formation of formaldehyde (a)

Formation of acetaldehyde (b)

Formation of ketones

(c)

Figure 4.5 Examples for the formation of OVOCs from the oxidation of VOC in the troposphere. (a) Formation of formaldehyde from the oxidation of methane, (b) formation of acetaldehyde from the oxidation of ethane and (c) formation of acetone from the oxidation of propane. For further details and the formation of these and other OVOCs from the oxidation of higher-molecular-weight VOCs, we refer to textbooks on atmospheric chemistry (e.g. Finlayson-Pitts and Pitts 2000; Seinfeld and Pandis 2000).

reaction pathways further lead to the formation of carboxylic acids, hydroperoxides and organic nitrates. Oxidation of isoprene provides another important OVOC source. Reactions of isoprene with OH radicals, ozone and NO3 radicals yield signiﬁcant amounts of methyl vinyl ketone, methacrolein and 3-methyl furan. The ozonolysis of unsaturated hydrocarbons is the main source of the formation of carboxylic acids in the troposphere. The reaction proceeds by addition of ozone to the double bond to form a primary ozonide, which rapidly decomposes into carbonyl and initially energy-rich Criegee biradicals. Criegee biradicals may collisionally stabilise and isomerise to yield the corresponding carboxylic acid. Although there are a number of kinetic studies on ozonolysis of alkenes in the gas phase, there are still large uncertainties concerning the yield of carboxylic acids from these reactions. Chebbi and Carlier (1996) reviewed the studies up to that date and summarised the known yields of different alkene–ozone reactions. The amount and distribution of OVOCs formed in the oxidation of organic compounds can be used as a measure of photochemical processing of polluted air masses. Meanwhile, remote sensing from satellites allows drawing a global picture of the formation of secondary

Oxygenated Volatile Organic Compounds

R

hν

h OH, + O2 NO3

PER-Rn +1-ACID

OH HO2

h

R

RnCO3

HO2′ RO2

Rn +1-ACID

149

M

NO3′ RO2′ NO

OH

P-Rn +1-N

NO2

RnO2

NO

OH HO2 RnOOH

RnNO3 RO2

NO3` RO2` NO

h

OH

h RnOH

RnO

OH

OH OH

O2

RO2

OH

R h , OH, + O2 NO3

−C4 aldehydes following the reaction with OH or Figure 4.6 Simpliﬁed scheme of the oxidation of C2 − nitrate radicals and/or photolysis according to the MCM mechanism (with permission from Bossmeyer 2006).

OVOCs, at least for some selected compounds. Again, SCIAMACHY results are an excellent example for the investigation of photochemical processing of VOC. Figure 4.7 shows the global distribution of glyoxal, which is formed during the oxidation of VOC emitted from terrestrial vegetation, biomass burning and anthropogenic activities (Wittrock et al. 2006).

4.4

Sinks of OVOCs

OVOCs are removed from the atmosphere by chemical or photolytic degradation, by dry and by wet deposition. Chemical degradation of OVOCs in the gas phase occurs via the reaction with radicals (OH, NO3 , in some special situations also Cl) and ozone. Of these processes, the most important one is the oxidation initiated by the reaction with OH radicals. The initial step of the oxidative removal of saturated OVOCs is the abstraction of a hydrogen atom, in the case of unsaturated OVOCs the addition of OH to the double bond. Unsaturated OVOCs are also oxidised by ozone and, mainly during nighttime, by the nitrate radical (NO3 ). Another important sink of OVOCs, for some compounds even the most important, is photolysis. Figure 4.8 shows the reaction channels initiated by the photolysis of formaldehyde, acetaldehyde and acetone. The corresponding radical pathways

150

Volatile Organic Compounds in the Atmosphere

60

Glyoxal (molecule/cm2)

Latitude

30

1.2 × 1015 1.0 × 1015 8.0 × 1014 6.0 × 1014 4.0 × 1014 2.0 × 1014

0

−30 −60 −180 −150 −120 −90

−60

−30 0 30 Longitude

60

90

120

150

180

Figure 4.7 The global distribution of glyoxal (CHOCHO) for the year 2005 derived from SCIAMACHY measurements. Glyoxal is an oxidation product of VOCs and thus an indicator for photochemical processing of polluted air masses (with permission from Wittrock et al. 2006). This image appears in full colour in the plate section that follows page 268 as Plate 3.

Reaction HCHO + hν CH3CHO + hν

CH3COCH3 + hν

HOx yield →

H• + HCO• H2 + CO

2 HO2

→ → → →

CH3• + HCO• CH4 + CO CH3CO• + H

2 HO2

→ →

CH3• + CH3CO• 2 CH3• + CO

1 HO2 2 HO2

1 HO2

Figure 4.8 Examples for the photolysis of OVOCs (formaldehyde, acetaldehyde and acetone) and the yield of HOx radicals in the troposphere. For further details on the photolysis of higher-molecular-weight OVOCs, we refer to textbooks on atmospheric chemistry (e.g. Finlayson–Pitts and Pitts 2000: Seinfeld and Pandis 2000).

lead to the formation of HOx radicals and thus have a signiﬁcant impact on atmospheric photochemistry. Since OVOCs are polar compounds, they can easily interact with aerosol particles or solve in rain or fog droplets. OVOCs are thus removed from the atmosphere by dry and wet deposition (Matsunaga et al. 2003; Seinfeld and Pandis 2000). The interaction with aerosol particles and the formation of secondary organic aerosols is described in Chapter 9 and will not be discussed here. For some OVOCs, it is known that there is an uptake by plants as long as their atmospheric concentration is higher than a certain compensation point (Kesselmeier 2001).

4.4.1

Chemical loss and photolysis of OVOCs

The typical lifetimes of OVOCs in the troposphere are dependent on the rate coefﬁcients of the reactions with OH, O3 and NO3 , respectively, and of the concentration of

Oxygenated Volatile Organic Compounds

151

Table 4.5 Mean atmospheric lifetimes for some selected OVOCs OVOC

Methanol Ethanol Formaldehyde Acetaldehyde Acetone Propanal MBO (Z)-3-Hexenol MACR MVK

kOH /10−12 cm3 s

τOH

kO3 /10−18 cm3 s

τO 3

kNO3 /10−17 cm3 s

τNO3

0.944a 3.27a 9.2a 16.0a 0.2a 20.0a 28.0b 108.0c 33.0d 19.0d

8.6 day 2.4 day 20.1 h 11.6 h 38.6 day 9.3 h 6.6 h 1.7 h 5.6 h 9.8 h

— — <0.01e <0.01e slowb slowb 10.0b 64.0c 1.1d 5.8d

— — >4.3 year >4.3 year — — 1.6 day 2.5 day 2.7 day 19.8 h

24.0a 200.0a 58.0a 270.0a <3.0a — — 2.7 × 104 c — —

5.4 year 235.2 day 2.2 year 1.7 day >43.0 year — — 1.7 day — —

Note: Mean lifetimes are estimated using OH concentrations of 1.5×106 cm−3 , O3 concentrations of 7 10 cm−3 (30 ppb) and NO3 concentrations of 2.4 107 cm−3 (1 ppt). a Atkinson (1997). b Grosjean and Grosjean (1995). c Atkinson et al. (1995). d Grosjean et al. (1993). e Kotzias et al. (1997).

the corresponding oxidant. Table 4.5 summarises the lifetimes of some speciﬁc OVOCs. Owing to their chemical structure and reactivity towards chemical processes in the atmosphere, OVOCs are often divided into two groups, highly reactive OVOCs (acetaldehyde, 2-methyl-3-butene-2-ol, C6 -alcohols, C6 -aldehydes, etc.) and less reactive OVOCs (methanol, ethanol, formic acid, acetic acid, acetone, etc.). Aldehydes and Ketones. The major transformation processes for aliphatic aldehydes, ketones and some dicarbonyls in the troposphere are photolysis and the reaction with OH radicals (cf. Atkinson 2000 and references therein). Reactions with ozone or with NO3 radicals are not important in the troposphere. Photolysis seems to be the dominant loss process for formaldehyde and the α-dicarbonyls. Photolysis and reaction with OH radicals compete in the loss of acetone in the lower troposphere where sufﬁcient OH radicals are available. In the upper troposphere, photolysis accounts for 90% of the loss of acetone. For all other compounds, the reaction with OH radicals is the dominant loss process in the troposphere. As regards the photolysis of these compounds, absorption cross-sections and quantum yields are available for formaldehyde, acetaldehyde and acetone. Also, for many other compounds, these data are reliably known (Atkinson 2000, and references therein). The same holds true for the rate coefﬁcients for the reactions with OH radicals. For the lessermolecular-weight compounds, the rates are available; for the others, these data are fairly well known. The photolysis of acetone and methyl hydroperoxide is a direct source of OH radicals in the upper troposphere. Acetone photolysis dominates over the primary formation of OH radicals from ozone photolysis and subsequent reaction with water vapour for (H2 O) <100 ppm. Under these conditions, the yield of HOx radicals from the photolysis of acetone varies between 2 and 3 in the upper troposphere (Jaeglé et al. 2000).

152

Volatile Organic Compounds in the Atmosphere

Organic hydroperoxides are typically formed in the gas phase following the reaction of peroxy radicals with HO2 . The information on reactions of hydroperoxides is sparse. The atmospheric chemistry of organic hydroperoxides is brieﬂy reviewed by Gunz and Hoffmann (1990). Alcohol. The dominant loss process of alcohols in the atmosphere is the reaction with OH −H bond. Abstraction of an H-atom from the O− −H radicals via H-atom abstraction at a C− bond is negligible because of the higher bond strength (104 vs 94 kcal/mol). The resulting −H hydroxyalkyl radicals react with oxygen generally by H-atom abstraction from the O− bond leading to a carbonyl and an HO2 radical. The important aspect to note is that the new radicals formed in these reactions again lead to the formation of aldehydes, namely formaldehyde, acetaldehyde or glycoladehyde. The chemistry of a variety of saturated alcohols has been reviewed by Grosjean and Grosjean (1997). Details of the product formation and kinetic data can be found in Atkinson (2000) and references therein. Photolysis, reactions with ozone or nitrate radicals are of negligible importance for saturated alcohols. Carboxylic acids. Carboxylic acids are not removed in signiﬁcant amounts by gas-phase reactions. The main loss reaction with OH radicals is slow. The corresponding rate coefﬁcients are on the order of 10−12 cm3 /molecule/s leading to average tropospheric residence times of several days up to a week. The main removal process of these compounds is dry and wet deposition (see below), which thus controls the actual residence times. Depending on the season, time of day, humidity and local circumstances, the residence time may be much shorter than given by chemical degradation.

4.4.2

Dry and wet deposition

Dry deposition refers to the loss of a compound on a surface. The compound is transported through the atmosphere to the surface by turbulence. It penetrates through a laminar layer between the turbulent atmosphere and the surface, and it is thereafter lost either at the surface itself or inside the corpus surrounded by the surface. Losses can appear on aerosols, on the sea surface, on soil or on plants. The loss rates depend on the turbulence in the atmosphere, on the gas-phase concentrations and properties of the respective compound as well as on the properties of the surface. If one of these speciﬁed steps is inefﬁcient, the whole process is inefﬁcient as a loss process. To assess the efﬁciency of dry deposition as a loss process, lifetimes with respect to dry deposition have to be compared to lifetimes with respect to other loss processes such as atmospheric oxidation or photolysis. A compound for which deposition is not negligible is methanol. Its lifetime with respect to dry deposition was estimated to be in the range of 24 days, and its lifetime with respect to atmospheric oxidation is 18 days (Heikes et al. 2002 and references therein). Both processes are of similar timescale and therefore lead to similar loss rates. In case of methanol, dry deposition cannot be neglected when balancing its budget. Other OVOCs for which dry deposition has been measured are acetic acid, formic acid, formaldehyde and acetaldehyde. Loss processes of acetic acid and formic acid on plants have been observed by Kesselmeier et al. (1998). This was attributed either to losses in water

Oxygenated Volatile Organic Compounds

153

droplets on the leaves or increases in stomatal opening. However, both acids are generally emitted from plants (Kesselmeier et al. 1998). For carboxylic acids, dry and wet deposition are the main sinks. Khan and Brimblecombe −C6 carboxylic acids. Their results show (1992) have measured Henry law constants for C2 − that the partitioning of carboxylic acids between liquid and gas phase is dependent on temperature, the amount of water present and its pH. Grosjean (1989) estimated that less than 10% of the lesser-molecular-weight carboxylic acids are removed from the troposphere by wet deposition. Dry deposition velocities have only been estimated up to now. The values range between 1 cm/s (Grosjean 1989; Jacob and Wofsy 1988; Keeler et al. 1990; Talbot et al. 1990) and 1.3 cm/s (Helas et al. 1992). From gradient measurements over a mixed deciduous forest, Krinke (1999) reported deposition velocities for formaldehyde in the range of 0.5–1.45 cm/s. From these data, a lifetime with respect to dry deposition of about 26 hours was estimated. This lifetime is much longer than that with respect to oxidation by OH (≈19 hours, Atkinson 2000) or with respect to photolysis (≈8 hours, e.g. Atkinson 2000). Thus, globally dry deposition is not important as a loss process for formaldehyde. For acetaldehyde, deposition velocities between 0.16 and 0.21 cm/s have been reported (Rottenberger et al. 2004). On the one hand, these deposition velocities are slower than those reported for formaldehyde. On the other hand, oxidation and photolysis are of similar efﬁciency as for formaldehyde (e.g. Atkinson 2000). Also, for acetaldehyde, dry deposition does not seem important as a loss process.

4.4.2.1

Uptake by plants

As we have discussed above, vegetation has been shown to be an important source for OVOCs in the atmosphere. The role of vegetation as a possible sink for OVOC is less investigated. If a compensation point exists for a given OVOC, its atmospheric concentration determines whether mainly loss by uptake occurs or emission dominates (e.g. Kesselmeier 2001). Rottenberger et al. (2004) report a compensation point for acetaldehyde of about 0.6 ppb during the transition of the dry-to-wet season in the Amazonian. Cojocariu et al. (2004) measured a compensation point of about 6 ppb for spruce seedlings and Karl et al. (2005) found between 1.2 and 8.5 ppb over a loblolly pine plantation. However, as obvious from the production pathway of acetaldehyde (e.g. Kreuzwieser et al. 2000), the compensation point for its uptake by plants will strongly depend on the oxygen concentration in the soil. Mechanistic measurements, with respect to the impact of formaldehyde uptake by plants and with respect to the fate of formaldehyde in plants, have been conducted by Giese et al. (1994), who measured the capacity of cell-suspension cultures to metabolise formaldehyde. This capacity is by far higher than that necessary to metabolise all formaldehyde capable to diffuse through the plant’s stomata. It may thus be assumed that formaldehyde destruction inside plants is not the limiting step for its deposition on plants. Also, for acetic and formic acid, bidirectional ﬂuxes are known. Uptake of both acids is observed and was attributed either to losses in water droplets on the leaves or increases in stomatal opening (Kesselmeier et al. 1998). Other research concerning the uptake of organic compounds by plants has been conducted in the scope of food quality since toxic VOCs may accumulate in plant material. It is

154

Volatile Organic Compounds in the Atmosphere

known that VOCs and OVOCs are absorbed in plant cuticles and may thus enter the food chain. The storage capacity of the cuticle may be very large (e.g. Merk and Riederer 1997; Riederer et al. 2002; Welke et al. 1998); however, it is still unclear whether or not storage of VOCs or OVOCs in the cuticle acts as an efﬁcient sink or as a buffer. In summary, dry deposition of OVOCs may be important from different points of view. As a loss process, dry deposition is important in case other loss processes are slow. Deposition may also play a signiﬁcant role if the sources of the OVOCs are near to the plants. In this case, dense vegetation can weaken the source strengths for OVOCs, also for those that are emitted from vegetation. Possible uptake of these VOCs may impact the role of vegetation as a VOC source. Knowledge about the uptake mechanisms may furthermore help assess differences between the amount of directly emitted carbon and the amount of carbon transported out of a canopy.

4.5

Budgets and emission inventories

To our knowledge, a detailed emission inventory for anthrop/ogenic emissions of OVOCs is presently not available. The Emission Database for Global Atmospheric Research (EDGAR V2.0) provides data on the global distribution and source strengths of anthropogenic NMHC (see Chapter 1). However, the database for OVOCs is rather limited. Speciﬁc data are only available for formaldehyde. Other aldehydes, ketones and alcohols are lumped into groups. As mentioned before, there are only a few estimates of the global source strengths of individual OVOCs from anthropogenic emissions such as methanol and based on very limited data sets. For all other OVOCs, little is known about anthropogenic sources. A number of investigations dealt with automobile emissions (see Chapter 2), but again the data are limited and difﬁcult to extrapolate to a global scale. Much more seems to be known about biogenic emissions. However, the great variety of OVOCs emitted from vegetation and the lack of knowledge of emission behaviour as well as of emission rates makes it difﬁcult to estimate global source strengths. Reported emission rates are in the range of 0.5–5 mg/h/kg dry leaf. Guenther et al. (1995) took a geometric mean value of 1.5 mg/h/kg dry leaf to estimate a global emission rate of 520 Tg C/year. This ﬁgure does not include methanol and the major organic acids emitted from plants, formic acid and acetic acid. Estimations of source strengths for naturally emitted OVOCs are based on the knowledge of the processes leading to the emissions and on a reference that can be used for extrapolations. They are based on atmospheric data, model calculations of biogenic and anthropogenic emissions, and removal processes, as well as 3D chemistry and transport models. Atmospheric budgets are only available for methanol (Galbally and Kirstine 2002) and acetone (Jacob et al. 2002; Singh et al. 1994, 1995). Methanol. Galbally and Kirstine (2002) estimated a global source of methanol of about 150 Tg/year with a range of 83–260 Tg/year. They estimate that terrestrial plants, as the largest single source, contribute 100 Tg/year. Secondary formation in the atmosphere contributes 19 Tg/year, biomass burning and decay of plant material 13 Tg/year each, and direct anthropogenic emission 4 Tg/year. The removal of methanol from the atmosphere is

Oxygenated Volatile Organic Compounds

155

estimated to be about 150 Tg/year with a range of 82–273 Tg/year. The reaction with OH radicals was the dominant removal process, followed by dry and wet deposition. The model estimated an average concentration of about 500 ppt in the southern hemisphere and about 800 ppt in the northern hemisphere. Both mixing ratios are slightly higher than the average values derived from observations. Acetone. Singh et al. (1994) estimated a global acetone source of 40 Tg/year with a range of 30–46 Tg/year. Based on more observational data, this best estimate was updated to 56 Tg/year, with a range of 37–80 Tg/year (Singh et al. 2000). Secondary formation from the atmospheric oxidation of precursor hydrocarbons such as propane, isobutane and isobutene were believed to be the largest source, with a contribution of about 50%, followed by biomass burning emissions and direct biogenic emissions. Only 3% were attributed to direct anthropogenic emissions. Photolysis, reaction with OH radicals and deposition are the removal processes for acetone with a contribution of 64%, 24% and 2%, respectively. The average lifetime of acetone was estimated to be 16 days (Singh et al. 1995). Jacob et al. (2002) did a comprehensive investigation of the global budget of acetone using a 3D global model simulation based on a priori estimates of sources and sinks. They used observations from 14 surface sites and 5 aircraft campaigns to improve their estimates through an inversion analysis. They give a best estimate of the global source of acetone of 95 ± 15 Tg/year, with a speciﬁc attribution to individual sources. Recently, Marandino et al. (2005) presented an updated acetone budget, taking into account their estimated global air-to-sea ﬂux of acetone of 48 Tg/year. This ﬁgure would signiﬁcantly imbalance the current atmospheric acetone budget. However, they also updated the estimated loss due to photolysis using new quantum yields measured by Blitz et al. (2004). Based on these data, the estimated loss is reduced by a factor of two from 46 to 24 Tg/year. Combining the updated loss rates and taking into account the uncertainties of the input parameters, the total loss changes from 96 Tg/year given by Jacob et al. (2002) to 101 Tg/year.

4.6

Sampling and measurement techniques

Compared to other hydrocarbons, OVOCs are more reactive, sticky, and water soluble, and are thus much more difﬁcult to measure with sufﬁcient accuracy in ambient air. A variety of techniques have evolved, most of them focussing on individual compounds. A review of measurement techniques is given by Vairavamurthy et al. (1992), and a summary is given in Table 4.6 taken from Benning (1997). Here, we give a brief overview of the different methods. In doing so, we discuss separately the techniques for the measurement of formaldehyde and other OVOCs. Formaldehyde. There are a number of different measurement techniques for formaldehyde in the atmosphere. A thorough compilation of measurement techniques and intercomparison experiments has recently been published by Hak et al. (2005). Here, we only give an overview of the methods and discuss brieﬂy the advantages and disadvantages of the different techniques. In general, we can distinguish between three different approaches to

Table 4.6 Compilation of measurement techniques Measurement of formaldehyde Colorimetric methods

Measurement of carbonyl compounds

Fluorescence methods

In-situ techniques

HPLC with derivatisation

Chromotropic acid UV/Vis Detector

2,5-diacetyl-1,4dihydrolutidine Fluorescence detector

Differential optical absorption spectroscopy long path, point measurement (White cell)

3-Methyl-2benzothiazolonhydrazone UV/Vis Detector

Reduction gas analyzer Mass spectrometric detector

o-Pentaﬂuorobenzylhydroxylamine with GC–MS, GC-ECD, GC-nitrogen phosphorous detector

Pararosanilin UV/Vis Detector

Enzyme catalised derivatisation Fluorescence detector

Fourier transform infrared spectroscopy long path, point measurement (White cell) Laser induced ﬂuorescence TOLAS

Anthrone Fluorescence detector 2-Diphenylacetylindan-1,3dionhydrazine Fluorescence detector 5,5-Dimethyl-1,3cyclohexadione Fluorescence detector 1,3-Cyclohexadione Fluorescence detector Dansylhydrazine Fluorescence detector 2,4-Dinitrophenyhydrazine UV/Vis Detector

Electron capture detector Flame ionisation detector 2-dimensional GC with FID

n-BenzylethanolamIne with GC–MS, GC-NPD Pentaﬂuorophenylhydrazine with GC-ECD

Coil gas liquid scrubber diffusion scrubber

Continuous measurements

Coil gas liquid scrubber Adsorbents

Adsorbents

Adsorbents

Impinger

Cryogenic sampling

Impinger

Other methods Source: Adapted from Benning (1997).

Impinger

GC without derivatisation

GC and derivatisation

Discontinuous measurement

Oxygenated Volatile Organic Compounds

157

measure formaldehyde in ambient air: 1

optical methods such as DOAS (Differential Optical Absorption Spectroscopy) or FTIR white cell systems, LIF (Laser-induced Fluorescence), and tuneable diode laser absorption spectroscopy (TDLAS). 2 in-situ wet chemical methods such as the Hantzsch reaction. 3 derivatisation with chromatographic separation techniques such as dimethylphenylhydrazine (DNPH)/high performance liquid chromatography (HPLC). DOAS systems typically operate with optical path lengths between 200 and 2 000 m in a wavelength range between 300 and 360 nm. The detection limit is around 1 ppb with accuracy typically better than 10% at a time resolution of a couple of minutes. FTIR systems operate at path lengths of up to 1 000 m, covering a spectral region between 1 800 and 3 500 cm−1 . Typically, the doublet at 2 779 and 2 781.50 cm−1 is used to detect formaldehyde. The detection limit is at about 0.5 ppb, with a precision of 10–30%. The time resolution is again a couple of minutes. Another optical method is TDLAS (Fried et al. 2002; Harris et al. 1989). These systems have detection limits between 40 ppt and 1 ppb, depending on the system, and time resolutions of a few minutes. Today, a commonly used method is a ﬂourometric method using the Hantzsch reaction. To apply this method, formaldehyde has to be transferred from the gas phase into the liquid phase using a stripping coil. The method applies the Hantzsch reaction ﬁrst used by Nash (1953) to determine formaldehyde. The continuous monitor for the detection of gaseous formaldehyde is described in detail by Kelly and Fortune (1994). The detection of formaldehyde is based on a reaction of aqueous formaldehyde with 2,4-pentanedione and ammonia, leading to 3,5-diacetyl-1,4-dihydrolutidine. The excitation wavelength is 412 nm; the ﬂuorescence is detected at 510 nm. The detection limit is about 40 ppt at a precision of 5%. Due to the response time of the system, the time resolution is on the order of 5 min. The chromatographic method applies the reaction of formaldehyde with 2,4-DNPH. By this reaction, formaldehyde and also other carbonyls are transferred to 2,4dinitrophenylhydrazone derivatives, which are separated by HPLC and detected by ultraviolet (UV) absorption spectroscopy at 360 nm. The typical detection limit is between 5 and 20 ng of formaldehyde, corresponding to ambient mixing ratios of about 0.5 ppb. The precision is on the order of 10%, but the time resolution is determined by the sampling procedure and is typically on the order of 1 h or more. Up to date, there have been about a dozen intercomparison experiments concerning formaldehyde measurements in ambient air. These experiments have been done on a laboratory level, in simulation chambers, but also during ﬁeld campaigns. For a detailed description of the individual exercises and their results, we refer to Hak et al. (2005) and the references therein. To brieﬂy assess the results of these experiments, the measurement of formaldehyde in ambient air is still a challenge and a matter of continuous discussion. Especially the wet chemical methods and the methods applying a derivatisation seem to be sensible to interferences and cross sensitivities. The discrepancies between these methods and optical methods are sometimes quite high, in some cases up to a factor of two. Compared to optical methods, especially measurements by the Hantzsch method show large, often unsystematic deviations. In some cases, the results obtained with these techniques show a good

158

Volatile Organic Compounds in the Atmosphere

agreement with about 15% of the measured values. However, in other cases, the Hantzsch method yields mixing ratios up to a factor of two higher or lower than those determined by optical methods. In most cases, these disagreements cannot be certainly attributed to speciﬁc problems such as sampling problems, cross interferences, or calibration problems. In cases where long-path DOAS or FTIR measurements have been compared with point measurements, the discrepancies may be attributed to the different air masses that have been probed. The determination of HCHO mixing ratios from DOAS measurements is based on absorption cross-sections measured in laboratory experiments. The recommended crosssections published by Cantrell et al. (1990) and Meller and Moortgat (2000) differ by 11.4%. However, cross-sections reported by other studies are lower (Rogers 1990) or higher (Pope et al. 2005) than the recommended values. Since for some studies the applied cross-section is not even speciﬁed, DOAS measurements imply a potential error of up to about 30%. Other OVOCs. For higher-molecular-weight oxygenated compounds, gas chromatography is still the most appropriate method to measure at least a number of OVOCs such as alcohols, ketones and >C1 -aldehydes (e.g. Apel et al. 2002; Goldan et al. 1995a; Riemer et al. 1998). For a recent compilation of GC techniques, we refer to Helmig et al. (1999a; 1999b). Because of their low atmospheric concentrations, they have to be preconcentrated from a larger volume of air before analysis. Thus, the measurement procedure typically consists of three steps: preconcentration, separation and detection. For a preconcentration, three methods are used: cryogenic preconcentration on porous glass beads at liquid nitrogen temperature, selective sampling on solid adsorbents (Ciccioli et al. 1993; Winer et al. 1992) or derivatisation. Separation of a mixture of preconcentrated VOCs can be done using gas chromatography, HPLC and chemical ionisation mass spectrometry (CIMS). A recently developed method is PTR-MS, which does not need any preconcentration. Following the preconcentration step, the compound mix is thermally desorbed and then transferred to the gas chromatograph. The most commonly used detectors for OVOC measurements are ﬂame ionisation detectors or mass spectrometry. Examples can be found in Goldan et al. (1995a) and Schade et al. (2000). In the past few years, Gas chromatography– mass spectrometry (GC–MS) has become increasingly popular for OVOC measurements (e.g. Apel et al. 2002; Biesenthal et al. 1998; Leibrock and Slemr 1997). Other hyphenated techniques have also been applied, such as GC–AED (Gas chromatography–atomic emission detector) (Apel et al. 1998a,b), GC–RGD (Gas chromatography–reduction gas detector) (O’Hara and Singh 1988), PTR-MS (Hansel et al. 1998) and FTIR (Yokelson et al. 1999). Recently, helium ionisation detectors are also applied for this purpose. A general problem with all GC techniques is the interference with water. For chromatographic separation using capillary columns, it is essential to remove water from the sample. This is typically done before preconcentration. However, removal of water from a sample may affect its OVOC composition. Various methods are described to remove water from the sample: selective preconcentration (Ciccioli et al. 1992; Heiden et al. 1999; Sturges and Elkins 1993), chemical drying (Ciccioli et al. 1992; Knoeppel et al. 1980; Löfgren et al. 1991; Staehelin, et al. 1991; Sturges and Elkins 1993), semipermeable membranes (Ciccioli et al. 1992; Foulger and Simmonds 1979; Hofmann et al. 1992; McClenny et al. 1984), freezing of water (Cooper and Saltzman 1991; Goldan et al. 1993; Mannschreck 2001;

Oxygenated Volatile Organic Compounds

159

Riemer et al. 1998), condensation (Rosen and Pankow 1991), selective desorption (Greenberg et al. 1994) and chromatographic techniques (Folkers 2001; Leibrock and Slemr 1997; Riemer et al. 1998). Chemical removal of water seems to be the most problematic removal procedure. Chemicals used for this purpose are K2 CO3 , Na2 CO3 , CaCO3 , NaOH, Na2 SO4 , Mg(ClO4 )2 and P2 O5 . In all cases, reactions of OVOCs with these chemicals and thus a deterioration of the sample composition cannot be excluded. Therefore, these methods cannot be recommended for an application in OVOC analysis. The most common semipermeable membrane to remove water from a sample is a copolymer from Teﬂon and perﬂuorosulfonic acid (NAFION). Because this material is chemically inert and temperature and pressure resistant, it is used in many commercial analytical systems. However, the fast and very effective removal of water from a sample is accompanied by high blank values, especially for hydrocarbons with double bonds, and a sometimes complete loss of polar OVOC (Kelly et al. 1993). More reactive compounds degrade at the surface of the membrane (Burns et al. 1983). While this method is suitable for the analysis of unpolar aliphatic hydrocarbons and aromatic compounds, it cannot be recommended for an analysis of OVOCs. Selective adsorption allows, in some special cases, a preconcentration of OVOC without trapping the water. Several adsorbents are available for the sampling of OVOC. For a number of compounds, TENAX TA provides the possibility to avoid the adsorption of water and CO2 . Sometimes other adsorbents are used to trap OVOC (various Carbotraps, Carbosieve S-III and Carboxen 569). All these adsorbents capable of trapping OVOCs seem to have speciﬁc problems reaching from artefacts during thermal desorption, degradation of some compounds on the adsorbent and memory effects. These materials also have the disadvantage that they can adsorb up to 400 mg of water per gram adsorbent. Using these materials makes the removal of water either prior to sampling or prior to injection necessary. Selective desorption of water using a reversed ﬂow in the absorption tube after trapping the OVOCs is also frequently used. This method is applicable, although it shows compoundspeciﬁc problems. However, a large volume of dry gas (ultra-pure nitrogen or helium) is necessary to remove water to a substantial extent. Great care has to be taken that the volume of purge gas is less than the breakthrough volume of the OVOC. Several authors report volumes of up to 4 L to remove the trapped water (Kelly et al. 1993; Koch et al. 1997). Another disadvantage is the time necessary for purging the trap. At typical ﬂow rates of some 100 mL/min, the cycling time for one analysis is considerably increased, which then would not be acceptable. Another method frequently used is a cold trap made of Teﬂon or SilcoSteel tubing kept at temperatures of 263–223 K. Riemer (1998) used PTFE (poly tretra ﬂuoro ethene) tubing at 263 K without observing any signiﬁcant losses of OVOC. Other groups use various tubing material at lower temperatures. However, the reported results vary considerably. MacDonald and Kimmerer (1993) do not report any losses of OVOC, while Ciccioli et al. (1992) report losses of more than 50% for alcohols and higher-molecular-weight aldehydes. Although not easy, a promising method is the chromatographic separation of water from the sample. Folkers (2001) describes a 2D chromatography using a packed column (Sorbitol) to remove water from a sample and subsequently a CP Wax 52 CB capillary column to separate OVOC. He reports minor problems with methanol due to insufﬁcient separation of the water and methanol peaks, but for all other OVOCs this method worked well. There have been only a few instrument intercomparisons so far with GC–MS and cartridge-based methods for the measurement of selected OVOCs. Differences greater than

160

Volatile Organic Compounds in the Atmosphere

a factor of two were common. A CIMS technique and a GC simultaneously measured acetone on board an aircraft in the free troposphere and lower stratosphere over the Atlantic (Wohlfrom et al. 1999). Very recently, an intercomparison experiment has been carried out in the frame of the European Atmospheric Composition Change – A Network of Excellence (ACCENT) at the atmosphere simulation chamber SAPHIR in Jülich, Germany. Twelve groups with a total of 16 different instruments participated in this exercise. In a formal blind intercomparison, 16 different compounds were intercompared at different, but almost typical ambient concentrations. Furthermore, the conditions were changed between dry and humid air, as well as zero and ambient ozone concentrations. On average, most instruments and compounds agreed within a factor of two with the calculated values from the chamber volume and the injected amount of OVOCs. Some instruments matched better than this, and some compounds appeared to be easier to measure than others.

4.7

Future directions

For the lesser-molecular-weight compounds representing the family of aldehydes, ketones and alcohols: Formaldehyde, acetaldehyde, acetone and methanol, our knowledge has signiﬁcantly increased over the past few years. Although by far not complete, their sources and sinks as well as their tropospheric budgets are reasonably well understood. But, in the view of the presently quite incomplete knowledge of all higher-molecular-weight compounds both with respect to their tropospheric distribution, their sources and sinks, but also with respect to the kinetic parameters describing their fate in the troposphere, the frequently used statement ‘more measurements are needed’ essentially holds true. Further investigations are certainly needed concerning atmospheric measurements, but also concerning laboratory data. For the latter, atmosphere simulation chambers that allow measurements at ambient conditions seem to be the most promising tool to investigate the chemical and photolytical loss processes of OVOCs as well as their secondary formation in the oxidation pathways of other volatile organic compounds. The increasing speed in the development of more sophisticated and sensitive measurement techniques will allow us to better investigate these difﬁcult-to-measure compounds even at low concentration levels. The results will help to better understand the role of these compounds and thereby the role of volatile organic compounds in the troposphere.

Acknowledgement The authors thank Prof. Dieter H. Ehhalt for a critical review of this chapter.

References Alvarado, A., Tuazon, E.C., Aschmann, S.M., Arey, J. and Atkinson, R. (1999) Products and mechanisms of the gas-phase reactions of OH radicals and O3 with 2-methyl-3-buten-2-ol. Atmospheric Environment, 33 (18): 2893–905.

Oxygenated Volatile Organic Compounds

161

Anderson, L.G., Lanning, J.A., Barrell, R., Miyagishima, J., et al. (1996) Sources and sinks of formaldehyde and acetaldehyde: An analysis of Denver’s ambient concentration data. Atmospheric Environment, 30: 2113–23. Andreae, M.O. and Merlet, P. (2001) Emission of trace gases and aerosols from biomass burning. Global Biogeochemical Cycles, 15: 955–66. Andreae, M.O., Talbot, R.W., Andreae, T.W. and Harriss, R.C. (1988) Formic and acetic acid over the central Amazon region, Brazil 1. Dry season. Journal of Geophysical Research, 93 (D2): 1616–24. Apel, E.C., Calvert, J.G., Riemer, D., et al. (1998a) Measurements comparison of oxygenated volatile organic compounds at a rural site during the 1995 SOS Nashville Intensive. Journal of Geophysical Research–Atmospheres, 103 (D17): 22295–316. Apel, E.C., Calvert, J.G., Greenberg, J.P., et al. (1998b) Generation and validation of oxygenated volatile organic carbon standards for the 1995 Southern Oxidants Study Nashville Intensive. Journal of Geophysical Research–Atmospheres, 103 (D17): 22281–94. Apel, E.C., Riemer, D.D., Hills, A., et al. (2002) Measurement and interpretation of isoprene ﬂuxes and isoprene, methacrolein, and methyl vinyl ketone mixing ratios at the PROPHET site during the 1998 Intensive. Journal of Geophysical Research, 107 (D3): ACH 7-1 to ACH 7-15. Arey, J., Winer, A.M., Atkinson, R., Aschmann, S.M., Long, W.D. and Morrison, C.L. (1991) The emission of (Z)-3-hexen-1-ol, (Z)-3-hexenyl acetate and other oxygenated hydrocarbons from agricultural plant species. Atmospheric Environment, 24A (5/6): 1063–75. Arlander, D., Croun, D.R., Farmer, J.C., Menzia, F.A. and Westberg, H.H. (1990) Gaseous oxygenated hydrocarbons in the remote marine troposphere. Journal of Geophysical Research, 95: 16.391–16.403. Arlander, D.W., Brüning, D., Schmidt, U. and Ehhalt, D.H. (1995a) The distribution of acetaldehyde in the lower troposphere during TROPOZ II. Journal of Atmospheric Chemistry, 22: 243–9. Arlander, D.W., Brüning, D., Schmidt, U. and Ehhalt, D.H. (1995b) The tropospheric distribution of formaldehyde during TROPOZ II. Journal of Atmospheric Chemistry, 22: 251–68. Arnold, F., Bürger, V., Droste-Franke, B., et al. (1997) Acetone in the upper troposphere and lower stratosphere: Impact on trace gases and aerosols. Geophysical Research Letter, 24: 3017–20. Atkinson, R. (1997) Gas-phase tropospheric chemistry of volatile organic compounds: 1. Alkanes and alkenes. Journal of Physical and Chemical Reference Data, 26: 215–90. Atkinson, R. (2000) Atmospheric chemistry of VOCs and NOx . Atmospheric Environment, 34: 2063–101. Atkinson, R., Arey, J., Aschmann, S.M., Corchnoy, S.B. and Shu, Y. (1995) Rate constants for the gas phase reactions of cis-3-hexen-1-ol, cis-3-hexenylacetate, trans-2-hexenal, and linalool with OH and NO3 radicals and O3 at 296 ± 2 K, and OH radical formation yields from the O3 reactions. International Journal of Chemical Kinetics, 27: 941–55. Báez, A.P., Belmont, R. and Padilla, H. (1995) Measurements of formaldehyde and acetaldehyde in the atmosphere of Mexico City. Environmental Pollution, 89: 163–7. Baker, B., Guenther, A., Greenberg,J. and Fall, R. (2001) Canopy level ﬂuxes of 2-methyl-3-buten-2-ol, acetone and methanol by a portable relaxed eddy accumulation system. Environmental Science and Technology, 35: 1701–8. Barker, M., Hengst, M., Schmid, J., et al. (2006) Volatile organic compounds in the exhaled breath of young patients with cystic ﬁbrosis. European Respiratory Journal, 27: 929–36. Beauchamp, J., Hansel, A., Kleist, E., et al. (2005) Ozone induced emissions of biogenic VOC from tobacco: Relationships between ozone uptake and emission of LOX products. Plant, Cell and Environment, 28 (10): 1334–43. −C2 -Aldehyd-Konzentrationen während der Feldmesskampagne Benning, L. (1997) Messung von C1 − POPCORN: Ihr Einﬂuss auf die Radikalchemie. PhD Thesis, Univeristy of Cologne, Germany.

162

Volatile Organic Compounds in the Atmosphere

Benning, L. and Wahner, A. (1998) Measurements of atmospheric formaldehyde (HCHO) and acetaldehyde (CH3 CHO) during POPCORN 194 using 2.4-DNPH coated silica cartridges. Journal of Atmospheric Chemistry, 31: 105–17. Biesenthal, T.A., Bottenheim, J.W., Shepson, P.B. and Brickell, P.C. (1998) The chemistry of biogenic hydrocarbons at a rural site in eastern Canada. Journal of Geophysical Research–Atmospheres, 103 (D19): 25487–98. Blitz, M.A., Heard, D.E. and Piling, M.J. (2004) Pressure and temperature-dependent quantum yields for the photodissociation of acetone between 279 and 327.5 nm. Geophysical Research Letter, 31: L06111. Bossmeyer, J. (2006) Studies of aldehydes in an atmosphere simulation chamber. Report of the Research Center Jülich, Germany, JÜL-4218. Boudries, H., Bottenheim, J.W., Guimbaud, C., et al. (2002) Distribution and trends of oxygenated hydrocarbons in the high Arctic derived from measurements in the atmospheric boundary layer and interstitial snow air during the ALERT2000 ﬁeld campaign. Atmospheric Environment, 36: 2573–83. Burns, W.F., Tingey, D.T., Evans, R.C. and Bates, E.H. (1983) Problems with a naﬁon membrane dryer for drying chromatographic samples. Journal of Chromatography, 269 (1): 1–9. Buttery, R.G. and Ling, L.C. (1993) Volatile components of tomato fruit and plant parts: relationship and biogenesis. In R. Teranishi, R.G. Buttery and H. Sugisawa (Eds) Bioactive Volatile Compounds from Plants. American Chemical Society, Washington, DC, 23–34. Cantrell, C.A., Davidson, J.A., McDaniel, A.H., Shetter, R.E. and Calvert, J.G. (1990) Temperaturedependent formaldehyde cross-sections in the near-ultraviolet spectral region. Journal of Physical Chemistry, 94: 3902–8. Carlier, P., Hannachi, H. and Mouvier, G. (1986) The chemistry of carbonyl compounds in the atmosphere – A review. Atmospheric Environment, 20: 2079–99. Chatﬁeld, R.B., Gardner, E.P. and Calvert, J.G. (1987) Sources and sinks of acetone in the troposhere: behavior of reactive hydrocarbons and a stable product. Journal of Geophysical Research, 92: 4208–16. Chebbi, A. and Carlier, P. (1996) Carboxylic acids in the troposphere, occurrence, sources, and sinks: A review. Atmospheric Environment, 30 (24): 4233–49. Christensen, C.S., Skov, H., Nielsen, T. and Lohse, C. (2000) Temporal variation of carbonyl compound concentrations at a semi-rural site in Denmark. Atmospheric Environment, 34: 287–96. Ciccioli, P., Cecinato, A., Brancaleoni, E. and Frattoni, M. (1992) Use of carbon adsorption traps combined with high-resolution gas chromatography-mass spectrometry for the analysis of polar −C14 hydrocarbons involved in photochemical smog formation. Journal of High and non-polar C4 − Resolution Chromatography Chromatography Communications, 15: 75–84. Ciccioli, P., Brancaleoni, E., Frattoni, M., Cucinato, A. and Brachetti, A. (1993) Ubiquitous occurrence of semi-volatile carbonyl compounds in tropospheric samples and their possible sources. Atmospheric Environment, 27A: 1891–901. Cojocariu, C., Kreuzwieser, J. and Rennenberg, H. (2004) Correlation of short chained carbonyls emitted from Picea abies with physiological and environmental parameters. New Phytologist, 162: 717–27. Cooper, D.J. and Saltzman, E.S. (1991) Measurements of atmospheric dimethyl sulﬁde and carbon disulﬁde in the western Atlantic boundary layer. Journal of Atmospheric Chemistry, 12: 153–68. Crocco, J.R. (1997) World methanol demand forecast to grow 2.7% over ﬁve years. Hydrocarbon Process, 76: 25. Croft, K.P.C., Jüttner, F. and Slusarenko, A.J. (1993) Volatile products of the lipoxygenase pathway evolved from (Phaseolus vulgaris L.) leaves inoculated with Pseudomonas syringae pv. Phaseolicola. Plant Physiology, Bd. 101 (1): S13–24.

Oxygenated Volatile Organic Compounds

163

Crutzen, P.J., Heidt, L.E., Krasnec, J.P., Pollock, W.H. and Seiler, W. (1979) Biomass burning as a source of atmospheric gases CO, H2 , N2 O, NO, CH3 Cl, and COS. Nature, 282: 253–7. Enders, G., Dlugi, R., Steinbrecher, R., et al. (1992) Biosphere atmosphere interactions integrated research in a European coniferous forest ecosystem. Atmospheric Environment, 26A (1): 171–90. Fall, R. (1999), Biogenic emissions of volatile organic compounds from higher plants. In: C.N. Hewitt (Eds) Reactive Hydrocarbons in the Atmosphere. San Diego: Academic Press. 43–96. Fall, R. and Benson, A. (1996) Leaf methanol – The simplest natural product form plants. Trends in Plant Science, 1 (9): 296–301. Fall, R., Karl, T., Hansel, A., Jordan, A. and Lindinger, W. (1999) Volatile organic compounds emitted after leaf wounding. On-line analysis by proton-transfer-reaction mass spectrometry. Journal of Geophysical Research, 104: 15963–74. Fall, R., Karl, T., Jordan, A. and Lindinger, W. (2001) Biogenic C5 VOCs: release from leaves after freezethaw wounding and occurrence in air at a high mountain observatory. Atmospheric Environment, 35: 3905–16. Fehsenfeld, F., Calvert, J., Fall, R., et al. (1992) Emissions of volatile organic compounds from vegetation and the implications for atmospheric chemistry. Global Biogeochemical Cycles, Bd. 6: S.389–430. Feng, Y., Wen, S., Chen, Y., et al. (2005) Ambient levels of carbonyl compounds and their sources in Guangzhou, China. Atmospheric Environment, 39: 1789–800. Ferrari, C.P., Kaluzny, P., Roche, A., Jacob, V. and Foster, P. (1998) Aromatic hydrocarbons and aldehydes in the atmosphere of Grenoble. Chemosphere, 37 (8): 1587–601. Ferronato, C., Orlando, J.J. and Tyndall, G.S. (1998) Rate and mechanism of the reactions of OH and Cl with 2-methyl-3-buten-2-ol (atmospheric chemistry). Journal of Geophysical Research, 103: (D19), 25579–86. Finlayson-Pitts, B.J. and Pitts, J.N. (2000) Chemistry of the Upper and Lower Atmosphere. Academic Press, San Diego, CA. Fisher, A.J., Grimes, H.D. and Fall, R. (2003) The biochemical origin of pentenol emissions from wounded leaves. Phytochemistry, 62: 159–63. Folkers, A. (2001) Sauerstoffhaltige ﬂüchtige organische Verbindungen in der Troposphäre: Entwicklung und Anwendung einer gaschromatographischen Nachweismethode. PhD Thesis, University of Cologne, Germany. Foulger, B.E. and Simmonds, P.G. (1979) Drier for ﬁeld use in the determination of trace atmospheric gases. Analytical Chemistry, 51 (7): 1089–90. Fried A., Lee, Y.-N., Frost, G., et al. (2002) Airborne CH2 O measurements over the North Atlantic during the 1997 NARE campaign: Instrument comparisons and distributions. Journal of Geophysical Research, 107 (D4): 4039. Galbally, I. and Kirstine, W. (2002) The production of methanol by ﬂowering plants and the global cycle of methanol. Journal of Atmospheric Chemistry, 43: 195–229. Gardner, H.W. (1995) Biological roles and biochemistry of the lipoxygenase pathway. Hort Science, Bd. 30 (2): S.197–2. Gerzhenzon, J. and Croteau, R. (1991) Terpenoids. In G.A. Rosenthal and M.R. Berenbaurn (Eds), Herbivores. Their interaction with secondary plant metabolites, Vol. 1: The chemical participants. Academic Press. Giese, M., Bauer-Doranth, U., Langebartels, C. and Sandermann H. Jr, (1994) Detoxiﬁcation of formaldehyde by spider plant (Chlorophytum comosum L.) and by soybean (Glycine max L.) cellsuspension cultures. Plant Physiology, 104: 1301–9. Goldan, P.D., Kuster, W.C., Fehsenfeld, F.C. and Montzka, S.A. (1993) The observation of a C5 alcohol emission in a north American pine forest. Geophysical Research Letters, 20 (11): 1039–42.

164

Volatile Organic Compounds in the Atmosphere

Goldan, P.D., Kuster, W.C., Fehsenfeld, F. and Montzka, S.A. (1995a) Hydrocarbon measurements in the southeastern United States: The Rural Oxidants in the Southern Environment (ROSE) Program 1990. Journal of Geophysical Research, 100 (D12): 25945–63. Goldan, P.D., Trainer, M., Kuster, W.C., et al. (1995b) Measurements of hydrocarbons, oxygenated hydrocarbons, carbon monoxide, and nitrogen oxides in an urban basin in Colorado: Implications for emission inventories. Journal of Geophysical Research, 100 (D11): S22771–83. Goldstein, A.H. and Schade, G.W. (2000) Quantifying biogenic and anthropogenic contributions to acetone mixing ratios in a rural environment. Atmospheric Environment, 34: 4997–5006. Goode, J.G., Yokelson, R.J., Ward, D.E, et al. (2000) Measurement of excess O3 , CO2 , CO, CH4 , C2 H4 , C2 H2 , HCN, NO, NH3 , HCOOH, CH3 COOH, HCHO, and CH3 OH in 1997 Alaskan biomass burning plumes by airborne Fourier transform infrared spectroscopy (FTIR). Journal of Geophysical Research, 89 (D17): 22147–66. de Gouw, J.A., Howard, C.J., Custer, T.G. and Fall, R. (1999) Emissions of volatile organic compounds from cut grass and clover are enhanced during the drying process. Geophysical Research Letters, 26: 811–14. Grabmer, W., Kreuzwieser, J., Wisthaler, A., et al. (2006) VOC emissions from Norway spruce (Picea abies L. [Karst]) twigs in the ﬁeld – Results of a dynamic enclosure study. Atmospheric Environment, 40: 128–137. Graedel, T.E. and Eisner, T. (1988) Atmospheric formic acid from formicine ants: A preliminary assessment. Tellus, 40B: 335–9. Graedel, T.E., Hawkins, D.T. and Claxton, L.D. (1986) Atmospheric Chemical Compounds: Sources, Occurrence, and Bioessay. Academic Press, Orlando, FL. Granby, K., Christensen, C.S. and Lohse, C. (1997) Urban and semirural observations of carboxylic acids and carbonyls. Atmospheric Environment, 31: 1403–15. Graus, M., Schnitzler, J.P., Hansel, A., et al. (2004) Transient release of oxygenated volatile organic compounds during light-dark transitions in grey poplar leaves. Plant Physiology, 135: 1967–75. Greenberg, J.P., Lee, B., Helmig, D. and Zimmerman, P.R. (1994) Fully automated gas chromatographﬂame ioniyation detector system for the in situ determination of atmosperic nonmethane hydrocarbons at low parts per trillion concentration. Journal of Chromatography A, 676: 389–98. Grosjean, D. (1989) Organic acids in southern California Air: Ambient concentrations, mobile source emissions, in situ formation and removal processes. Environmental Science and Technology, 23: 1506–14. Grosjean, D., Williams, E.L. and Grosjean, E. (1993) Atmospheric chemistry of isoprene and of its carbonyl products. Environmental Science and Technology, 27: 830–40. Grosjean, E. and Grosjean, D. (1995) Carbonyl products of the ozone unsaturated alcohol reaction. Journal of Geophysical Research, 100: 22815–20. Grosjean, E. and Grosjean, D. (1997) The gas phase reaction of unsaturated oxygenates with ozone carbonyl products and comparison with the alkene-ozone reaction. Journal of Atmospheric Chemistry, 27: 271–89. Grossman, D., Moortgat, G.K., Kibler, M., et al. (2003) Hydrogen peroxide, organic peroxides, carbonyl compounds, and organic acids measured at Pabstthum during BERLIOZ. Journal of Geophysical Research, 108 (D4): 8250. Guenther, A., Hewitt, C.N., Erickson, D., et al. (1995) A global model of natural volatile organic compound emissions. Journal of Geophysical Research, Bd. 100 (D5): S8873–92. Guenther, A.B., Zimmerman, P.R., Harley, P.C., Monson, R.K. and Fall, R. (1993) Isoprene and monoterpene emission rate variability: Model evaluation and sensitivity analysis. Journal of Geophysical Research, Bd. 98 (D7): 12609–17. Gunz, D.W. and Hoffmann, M.R. (1990) Atmospheric chemistry of peroxides. A review. Atmospheric Environment, 24A: 1601–33.

Oxygenated Volatile Organic Compounds

165

Hak, C., Pundt, I., Trick, S., et al. (2005) Intercomparison of four different in-situ techniques for ambient formaldehyde measurements in urban air. Atmospheric Chemistry and Physics, 5: 2881–900. Hansel, A., Jordan, A., Warneke, C., Holzinger, R. and Lindinger, W. (1998) Improved detection limit of the proton-transfer reaction mass spectrometer: On-line monitoring of volatile organic compounds at mixing ratios of a few pptv. Rapid Communications in Mass Spectrometry, Bd. 12 (13): S871–5. Harley, P., Fridd-Stroud, V., Greenberg, J., Guenther, A. and Vasconcellos, P. (1998) Emission of 2-methyl-3-buten-2-ol by pines: A potentially large natural source of reactive carbon to the atmosphere. Journal of Geophysical Research, 103 (D19): 25479–86. Harley, P., Fridd-Stroud, V., Greenberg, J., Guenther, A. and Vasconcellos, P. (1999) Emission of 2-methyl-3-buten-2-ol by pines: A potentially large natural source of reactive carbon to the atmosphere. Journal of Geophysical Research, Bd. 103 (D19): S25475–86. Harriman, R.W., Tieman, D.M. and Hanada, A.K. (1991) Molecular cloning of tomato pectin methylesterase gene and its expression in Rutgers, ripening inhibitor, non-ripening, and never ripe tomato fruits. Plant Physiology, 97: 80–87. Harris, G.W., Mackay, G.I., Iguchi, T., Mayne, L.K. and Schiff, H.I. (1989) Measurements of formaldehyde in the troposphere by tunable diode laser absorption spectroscopy. Journal of Atmospheric Chemistry, 8: 119–37. Hatanaka, A. (1993) The biogeneration of green odour by green leaves. Phytochemistry, Bd. 34 (5): S1201–18. Hartmann, W.K., Andreae, M.O. and Helas, G. (1989) Measurements of organic acids over central Germany. Atmospheric Environment, 23: 1531–33. Heiden, A.C., Hoffmann, T., Kahl, J., et al. (1999) Emission of volatile organic compounds from ozone-exposed plants. Ecological Application, 94: 1160–7. Heiden, A.C., Kobel, K., Langebartels, C., Schuh-Thomas, G. and Wildt, J. (2003) Emissions of oxygenated volatile organic compounds from plants, part I: Emissions from lipoxygenase activity. Journal of Atmospheric Chemistry, 45: 143–72. Heikes, B.G., Chang, W., Pilson, M.E.Q., et al. (2002) Atmospheric methanol budget and ocean implication. Global Biogeochemical Cycles, 16 (4): 1133. Helas, U.H., Kessel, M., Kesselmeier, J., et al. (1992) Biosphere atmosphere interactions integrated research in a European coniferous forest ecosystem. Atmospheric Environment, 26A (1): 1992. Hellen, H., Hakola, H., Reissell, A. and Ruuskanen, T.M. (2004) Carbonyl compounds in boreal coniferous forest air in Hyytiälä, Southern Finland. Atmospheric Chemistry and Physics Discussions, 4: 2991–3011. Helmig, D. and Greenberg, J.P. (1994) Automated in situ gas chromatographic-mass spectrometric analysis of ppt level volatile organic trace gases using multistage solid-adsorbent trapping. Journal of Chromatography A, 677 (1): 123–32. Helmig, D., Klinger, L.F., Guenther, A., Vierling, L., Geron, C. and Zimmerman, P. (1999a) Biogenic volatile organic compound emissions (BVOCs) I: Identiﬁcations from three continental sites in the U.S. Chemosphere, 38 (9): 2163–2187. Helmig, D., Klinger, L.F., Guenther, A., Vierling, L., Geron, C. and Zimmerman, P. (1999b) Biogenic volatile organic compound emissions (BVOCs) II: Landscape ﬂux potentials from three continental sites in the U.S. Chemosphere, 38 (9): 2189–204. Herlihy, L.J., Galloway, J.N. and Mills, A.L. (1987) Bacterial utilization of formic and acetic acid in rainwater. Atmospheric Environment, 21: 2397–402. Hofmann, U., Hofmann, R. and Kesselmeier, J. (1992) Cryogenic trapping of reduced sulfurcompounds using a naﬁon drier and cotton wadding as an oxidant scavenger. Atmospheric Environment A, 26 (13): 2445–9.

166

Volatile Organic Compounds in the Atmosphere

Holzinger, R., Warneke, C., Hansel, A., et al. (1999) Biomass burning as a source of formaldehyde, acetaldehyde, methanol, acetone, acetonitrile, and hydrogen cyanide. Geophysical Research Letters, 26: 1161–4. Isidorov, V.A., Zenkevich, I.G. and Ioffe, B.V. (1985) Volatile organic compounds in the atmosphere of forests. Atmospheric Environment, 19: 1–8. Jacob, D.J. and Wofsy, S.C. (1988) Photochemistry of biogenic emissions over the Amazon forest. Journal of Geophysical Research, 93: 1477–86. Jacob, D.J., Field, B.D., Jin, E.M., et al. (2002) Atmospheric budget of acetone. Journal of Geophysical Research, 107, D10: 4100. Jaeglé, L., Jacob, D.J., Brune, W.H. et al. (2000) Photochemistry of HOX in the upper troposphere at northern midlatitudes, Journal of Geophysical Research, 105: 3877–92. Jaeglé, L., Jacob, D.J., Brune, W.H. and Wennberg, P.O. (2001) Chemistry of HOx radicals in the upper troposphere.Atmospheric Environment, 35: 469–89. Jang, M., Czoschke, N.M., Lee, S. and Kamens, R.M. (2002) Heterogeneous atmospheric aerosol production by acid-catalyzed particle-phase reactions. Science, 298 (5594): 814–17. Joseph, G. and Kelsey, R.G. (2000) Physiology of growth of duglar-ﬁr seedlings treated with ethanol solutions. Plant Science, Bd. 150 (2): S191–9. Kalberer, M., Paulsen, D., Sax, M., et al. (2004) Identiﬁcation of polymers as major components of atmospheric organic aerosols. Science, 303: 1659–62. Karl, T., Fall, R., Jordan, A. and Lindinger, W. (2001a) On-line analysis of reactive VOCs from urban lawn mowing. Environmental Science and Technology, Bd. 35 (14): S2926–31. Karl, T., Guenther, A., Jordan, A., Fall, R. and Lindinger, W. (2001b) Eddy covariance measurement of biogenic oxygenated VOC emissions from hay harvesting. Atmospheric Environment, 35: 491–5. Karl, T., Guenther, A., Spirig, C., Hansel, A. and Fall, R. (2003) Seasonal variation of biogenic VOC emissions above a mixed hardwood forest in northern Michigan. Geophysical Research Letters, 30: 2186. Karl, T., Harley, P., Baker, B., Jardine, K., Nemitz, E. and Guenther, A. (2005) The bi-directional exchange of oxygenated VOCs between a loblolly pine (Pinus taeda) plantation and the atmosphere. Atmospheric Chemistry and Physics Discussion, 5: 5875–907. Kawamura, K. and Kaplan, I.R. (1990) Stabilities of carboxylic acids and phenols in Los Angeles rainwaters during storage. Water Resource, 24 (II): 1419–23. −Cl10 ) in the Kawamura, K., Ng, L.-L. and Kapian, I.R. (1985) Determination of organic acids (C1 − atmosphere, motor exhausts, and engine oils. Environmental Science and Technology, 19: 1082–6. −C10 monocarboxylic Kawamura, K., Steinberg, S. and Kaplan, I.R. (2000) Homologous series of C1 − −C6 carbonyls in Los Angeles air and motor vehicle exhaust. Atmospheric acids and C1 − Environment, 34: 4175–91. Kean, A.J., Grosjean, E., Grosjean, D. and Harley, R.A. (2001) On-road measurement of carbonyls in California light-duty vehicle emissions. Environmental Science and Technology, 35: 4198–4204. Keeler, G.J., Japar, S.M., Brachaczek, W.W., Gorse, R.A. Jr, Norbeck, J.M. and Pierson, W.R. (1990) The sources of aerosol elemental carbon at Allegheny Mountain. Atmospheric Environment, 24A: 2795–2805. Kelly, T.J. and Fortune, C.R. (1994) Continuous monitoring of gaseous formaldehyde using an improved ﬂuorescence approach. International Journal of Environmental Analytical Chemistry, 54: 249–63. Kelly, T.J., Callahan, P.J., Pleil, J. and Evans, G.F. (1993) Method development and ﬁeld measurements for polar volatile organic compounds in ambient air. Environmental Science and Technology, 27 (6): 1146–53. Kesselmeier, J. (2001) Exchange of short chain oxygenated volatile organic compounds (VOCs) between plants and the atmosphere: A compilation of ﬁeld and laboratory studies. Journal of Atmospheric Chemistry, 39: 219–33.

Oxygenated Volatile Organic Compounds

167

Kesselmeier, J. and Staudt, M. (1999) Biogenic volatile organic compounds (VOC): An overview on emission, physiology and ecology. Journal of Atmospheric Chemistry, 33 (1): 23–88. Kesselmeier, J., Bode, K., Gerlach, C. and Jork, E.M. (1998) Exchange of atmospheric formic and acetic acid with trees and crop plants under controlled chamber and puriﬁed air conditions. Atmospheric Environment, 32 (10): 1765–75. Kesselmeier, J., Bode, K., Hofmann, U., et al. (1997) Emission of short chained organic acids, aldehydes and monoterpenes from Quercus ilex L. and Pinus Pinea L. in relation to physiological activities, carbon budget and emission algorithms. Atmospheric Environment, 31: 119–33. Khan, I. and Brimblecombe, P. (1992) Henry’s law constants of low molecular weight (<130) organic acids. Journal of Aerosol Science, 23 (Suppl. 1): 897–900. Khoder, M.I., Shakour, A.A., Farag, S.A. and Abdel Hameed, A.A. (2000) Indoor and outdoor formaldehyde concentrations in homes in residential areas in Greater Cairo. Journal of Environmental Monitoring, 2 (2): 123–6. Kimmerer, T.W. and Kozlowski, T.T. (1982) Ethylene, ethane, acetaldehyde and ethanol production by plants under stress. Plant Physiology, Bd. 69 (4): S840–7. Kirstine, W., Galbally, I., Ye, Y. and Hooper, M. (1998) Emissions of volatile organic compounds (primarily oxygenated species) from pasture. Journal of Geophysical Research, Bd. 103 (D9): S10605–19. Knoeppel, H., Versiono, B., Schlitt, H., Peil, A., Schauenburg, H. and Vissers, H. (1980) Proceedings of the ﬁrst European Symposium on Physical–Chemical Behaviour of Atmospheric Pollutants. Ispra, 16–18 October 1979, Commission of the European Communities. Knudsen, J.T., Tollsten, L. and Bergström, L.B. (1993) Floral scents – A checklist of volatile compounds isolated by head-space technique. Phytochemistry, Bd. 33 (2): S253–80. Koch, E., Wózniak, P., Gawlowski, J. and Niedzielski, J. (1997) Removal of water from multibed carbon traps with adsorbed polar volatile organic compounds. Chemia Analityczna, 42 (1): 49–56. König, G., Brunda, M., Puxbaum, H., Hewitt, C.N., Duckham, S.C. and Rudolph, J. (1995) Relative contribution of oxygenated hydrocarbons to the total biogenic VOC emissions of selected midEuropean agricultural and natural plant species. Atmospheric Environment, 29 (8): 861–74. Koppmann, R., Khedim, A., Rudolph, J., et al. (1997) Emissions of organic trace gases from savanna ﬁres in southern Africa during the 1992 southern African ﬁre atmosphere research initiative and their impact on the formation of tropospheric ozone. Journal of Geophysical Research, 102: 18879–88. Kotzias, D., Konidari, C. and Sparta, C. (1997) Volatile carbonyl compounds of biogenic origin: Emission and concentration in the atmosphere. In: G. Helas, J. Slanina and R. Steinbrecher (Eds.) Biogenic Volatile Organic Compounds in the Atmosphere. The Hague, The Netherlands SPB Academic Publishing. Kreuzwieser, J., Scheerer, U. and Rennenberg, H. (1999) Metabolic origin of acetaldehyde emitted by poplar (Populus tremula x P. alba) trees. Journal of Experimental Botany, Bd. 50 (335): S757–65. Kreuzwieser, J., Kühnemann, F., Martis, A., Rennenberg, H. and Urban, W. (2000) Diurnal pattern of acetaldehyde emission by ﬂooded poplar trees. Physiologia Plantarum, Bd. 108 (1): S79–86. Krinke, S. (1999) Experimentelle Bestimmung der Depositionsgeschwindigkeit von Formaldehyd und Ozon über einem Laubwaldbestand. PhD thesis, University of Cologne, Germany. de Laat, A.T.J., de Gouw, J.A., Lelieveld, J. and Hansel, A. (2001) Model analysis of trace gas measurements and pollution impact during INDOEX. Journal of Geophysical Research, 106 (28): 469–28, 480. Lacaux, J.-P., Cachier, H. and Delmas, R. (1993) Biomass burning in Africa: An overview of its impact on atmospheric chemistry. In: P.J. Crutzen and J.G. Goldammer (Eds) Fire in the Environment: The Ecological, Atmospheric, and Climatic Importance of Vegetation Fires. pp. 159–91, John Wiley & Sons, Chichester, England.

168

Volatile Organic Compounds in the Atmosphere

−C10 volatile organic Lamanna, M.S. and Goldstein, A.H. (1999) In situ measurements of C2 − compounds above a Sierra Nevada ponderosa pine plantation. Journal of Geophysical Research, 104 (D17): 21247–62. Lawrence, J.E. and Koutrakis, P. (1994) Measurement of atmospheric formic and acetic acids: Methods evaluation and results from ﬁeld studies. Environmental Science and Technology, 28: 957–64. Lee, Y.-N., Zhou, X. and Hallok, K. (1995) Atmospheric carbonyl compounds at a rural site in the southeastern United States, Journal of Geophysical Research, 100: 25933–44. Leibrock, E. and Slemr, J. (1997) Method for measurement of volatile oxygenated hydrocarbons in ambient air. Atmospheric Environment, 31 (20): 3329–39. Levy, S. and Staehelin, L.A. (1992) Synthesis, assembly and function of plant cell wall macromolecules. Current Opinions in Cell Biology, 4: 856. Lewis, N.B. and Yamamot, E. (1990) Lignin: occurrence, biogenesis and biodegradation. Annual Review of Plant Physiology and Plant Molecular Biology, Bd. 41: S455–96. Li, Y., Wang, J., Chen, Z.R. and Zhou, X.T. (2001) Artiﬁcial neural network for the quantitative analysis of air toxic VOCs. Analytical Letters, 34 (12): 2203–19. Lichtenthaler, H.K. (1999) The 1-deoxy-D-xylulose-5-phosphate pathway of isoprenoids biosynthesis in plants. Annual Review of Plant Physiology and Plant Molecular Biology, 50: 47–65. Lobert, J.M., Scharffe, D.H., Hao, W.M. et al. (1991) Experimental evaluation of biomass burning emissions: nitrogen and carbon containing compounds. In: J. Levine (Ed.) Global Biomass Burning: Atmospheric, Climatic, and Biospheric Implications. MIT Press, Cambridge, MA and London, GB, pp. 290–304. Löfgren, L., Berglund, P.M. Nordlinder, R., Pertersson, G. and Ramnäs, O. (1991) Selective assessment −C6 alkenes in air by adsorption sampling and gas-chromatography. International Journal of C2 − of Environmental Analytical Chemistry, Bd. 45 (1): S39–44. MacDonald, R.C. and Fall, R. (1993a) Acetone emissions from conifer buds. Phytochemistry, 34: 991–3. MacDonald, R.C. and Fall, R. (1993b) Detection of substantial emissions of methanol from plants to the atmosphere. Atmospheric Environment, 27A: 1709–13. MacDonald, R.C. and Kimmerer, T.W. (1993) Metabolism of transpired ethanol by eastern cottonwood (Populus deltoides Bartr.). Plant Physiology, Bd. 102 (1): S173–9. Mannschreck, K. (2001) Experimentelle Bestimmung von städtischen Emissionen anhand von Konzentrationsmessungen im Lee einer Stadt, Untersuchungen zum Beitrag verschiedener Quelltypen und Vergleich mit einem Emissionsberechnungsmodell. Berichte des Forschungszentrums Jülich (Jül.): 3846. Marandino, C.A., de Bruyn, W.J., Miller, S.D., Prather, M.J. and Saltzmann, M.E. (2005) Oceanic uptake and the global atmospheric acetone budget. Geophysical Research Letter, 32: L15806. Matsunaga, S., Mochida, M. and Kawamura, K. (2003) Growth of organic aerosols by biogenic semi-volatile carbonyls in the forestal atmosphere. Atmospheric Environment, 37: 2045–50. McClenny, W.A., Pleil, J.D. Holdren, M.W. and Smith, R.N. (1984) Automated cryogenic preconcentration and gas-chromatographic determination of volatile organic-compounds in air. Analytical Chemistry, 56 (14): 2947–51. McKeen, S.A., Gierczak, T., Burkholder, J.B. et al. (1997) The photochemistry of acetone in the upper troposphere: A source of odd-hydrogen radicals. Geophysical Research Letter, 24: 3177–80. Meller, R. and Moortgat, G.K. (2000) Temperature dependence of the absorption cross-sections of formaldehyde between 223 and 323 K in the wavelength range 225–375 nm. Journal of Geophysical Research, 105 (D6): 7089–101. Merk, S. and Riederer, M. (1997) Sorption of volatile C1 to C6 alkanols in plant cuticles. Journal of Experimental Botany, 48: 1095–104. Morgan, E.D. and Tyler, R.C. (1977) Microchemical methods for the identiﬁcation of volatile pheromones. Journal of Chromatography, 134: 174–7.

Oxygenated Volatile Organic Compounds

169

Mudgett, M.B. and Clarke, S. (1993) Characterization of plant L-isoaspartyl methyltransferases that may be involved in seed survival: Puriﬁcation, cloning and sequence analysis of the wheat germ enzyme. Biochemistry, Bd. 32 (41): S11100–11. Müller, J.-F. (1992) Geographical distribution and seasonal variation of surface emissions and deposition velocities of atmospheric trace gases. Journal of Geophysical Research, Bd. 97 (D4): S3787–804. Müller, K. (1997) Determination of aldehydes and ketones in the atmosphere: A comparative long time study at an urban and a rural site in Eastern Germany. Chemosphere, 35: 2093–106. Müller, K., Haferkorn, S., Grabmer, W. et al. (2006) Biogenic carbonyl compounds within and above a coniferous forest in Germany. Atmospheric Environment, 40: 81–91. Nash, T. (1953) The colorimetric estimation of formaldehyde by means of the Hantzsch reaction. Journal of Biochemistry, 55: 416–21. Nemecek-Marshall, M., MacDonald, R.C., Franzen, F.J., Wojciechowski, C.L. and Fall, R. (1995) Methanol emission from leaves. Plant Physiology, Bd. 108 (4): S1359–68. Niinemets, Ü. and Reichstein, M. (2003) Controls of the emission of plant volatiles through stomata: Differential sensitivity of emission rates to stomatal closure explained. Journal of Geophysical Research, 108 (D7): 4208.. Obendorf, R.L., Koch, J.L., Gorecki, R.J., Amable, R.A. and Aveni, M.T. (1990) Methanol accumulation in maturing seeds. Journal of Experimental Botany, Bd. 41 (225): S489–95. O’Hara, D. and Singh, H.B. (1988) Sensitive gas chromatographic detection of acetaldehyde and acetone using a reduction gas detector. Atmospheric Environment, 22: 2613–15. O’Neill, M., Albersheim, P. and Darvill, A. (1990) The pectic polysaccharides of primary cell walls. In: Methods in Plant Biochemistry, Bd. 2, London: Academic Press, 415–41. Orlando, J.J., Noziére, B., Tyndall, G.S. et al. (2000) Product studies of the OH- and ozoneinitiated oxidation of some monoterpenes. Journal of Geophysical Research, 105 (D9): 11561–72. Owen, S., Boissard, C., Street, R.A., Duckham, C., Csiky, O. and Hewitt, C.N. (1997) Screening of 18 Mediterranean plant species for volatile organic compound emissions. Atmospheric Environment, 31 (SI): 101–18. de Paula Pereira, P.A., Santos, E.T.S., De Freitas Verreira, T. and De Andrade, J.B. (1999) Determination of methanol and ethanol by gas chromatography following air sampling onto ﬂorosil cartridges and their concentrations at urban sites in the three largest cities in Brazil. Talanta, 49 (2): 245–52. Pleil, J.D. and Lindstrom, A.B. (1995) Measurement of volatile organic compounds in exhaled breath as collected in evacuated electropolished canisters. Journal of Chromatography B: Biomedical Application, 665 (2): 271–9. Pope, F.D., Smith, C.A., Ashfold, M.N.R. and Orr-Ewing, J. (2005) High resolution absorption cross sections of formaldehyde at wavelengths from 313 to 320 nm. Physical Chemistry Chemical Physics, 7: 79–84. Possanzini, M., Di Palo, V., Brancaleoni, E., Frattoni, M. and Ciccioli, P. (2000) A train of DNPHcoated cartridges for the determination of carbonyls from C1 to C12 in air and emission samples. Atmospheric Environment, 34: 5311–18. Puxbaum, H. (1997) Biogenic emission of alcohols, ester, ether and higher aldehydes. In: G. Helas, J. Slanina and R. Steinbrecher (Eds) Biogenic Volatile Organic Carbon Compounds in the Atmosphere, Amsterdam, The Netherlands: SPA Academic Publishing. Puxbaum, H., Simeonov, V. and Kalina, M.F. (1998) Ten years trends (1984–1993) in the precipitation chemistry in central Austria. Atmospheric Environment, 32 (2): 193–202. −C4 Ratte, M., Plass-Dülmer, C., Koppmann, R. and Rudolph, J. (1993) Production mechanism of C2 − hydrocarbons in sea water: Field measurements and experiments. Global Biogeochemical Cycles, 7: 369–378.

170

Volatile Organic Compounds in the Atmosphere

Reiner, T., Möhler, O. and Arnold, F. (1999) Measurements of acetone, acetic acid, and formic acid in the northern midlatitude upper troposphere and lower stratosphere. Journal of Geophysical Research, 104 (D11): 13943–52. Reissell, A., Harry C., Aschmann, S.M. (1999) Formation of acetone from the OH radicaland O3 -initiated reactions of a series of monoterpenes. Journal of Geophysical Research, 104: 13869–79. Ricard, J., Noat, G., Diamantidis, G., Woudstra, M. and Ricard, J. (1986) Electrostatic effects and the dynamics of enzyme reactions at the surface of plant cells. A theory of the ionic control of a complex multienzyme system. European Journal of Biochemistry, Bd. 155 (1): S199–202. Riederer, M., Daiss, A., Gilbert, N. and Kohle, H. (2002). Semi-volatile organic compounds at the leaf/atmosphere interface: Numerical simulation of dispersal and foliar uptake. Journal of Experimental Botany, 53: 1815—23. Riemer, D., Pos, W., Milne, P., et al. (1998) Observations of nonmethane hydrocarbons and oxygenated volatile organic compounds at a rural site in the southeastern United States. Journal of Geophysical Research, 103: 28111–28. Rogers, J.D. (1990) Ultraviolet absorption cross-section and atmospheric photodissociation rate constants of formaldehyde. Journal of Physical Chemistry, Bd. 94: S4011–15. Rosen, M.E. and Pankow, J.F. (1991) Adsorption/thermal desorption for the determination of volatile organic compounds in water. Journal of Chromatography, 537: 321–8. Rottenberger, S., Kuhn, U., Wolf, A., et al. (2004) Exchange of short-chain aldehydes between Amazonian vegetation and the atmosphere. Ecological Applications, 14 (4): 247–62. Sanhueza, E. and Andreae, M.O. (1991) Emission of formic and acetic acids from tropical savanna soils., Geophysical Research Letter, 18: 1707–10. Satsumabayashi, H., Kurita, H., Yokouchi, Y. and Ueda, H. (1989) Mono-and di-carboxylic acids under long-range transport of air pollution in central Japan. Tellus, 41B: 219–29. Schade, G.W. and Custer, T.G. (2004) OVOC emission from agricultural soil in northern Germany during the 2003 European heat wave. Atmospheric Environment, 38: 6105–14. Schade, G.W. and Goldstein, A.H. (2001) Fluxes of oxygenated volatile organic compounds from a ponderosa pine plantation. Journal of Geophysical Research, 106 (D3): 3111–24. Schade, G.W., Goldstein, A.H., Gray, D.W. and Lerdau, M.T. (2000) Canopy and leaf level 2methyl-3-buten-2-ol ﬂuxes from a ponderosa pine plantation. Atmospheric Environment, 34: 3535–44. Schuh, G., Heiden, A.C., Hoffmann, Th., et al. (1997) Emissions of volatile organic compounds from sunﬂower and beech: dependence on temperature and light intensity. Journal of Atmospheric Chemistry, 27: 291–318. Seinfeld, J.H. and Pandis, S.N. (2000) Atmospheric Chemistry and Physics. Wiley, New York. Seskar, M. Shulaev, V. and Raskin, I. (1998) Endogenous methyl salicylate in pathogen-inoculated tobacco plants. Plant Physiology, 116: 387–92. Shao, M. and Wildt, J. (2002) Quantiﬁcation of acetone emissions from pine plants. Science in China (B), 45: 532–40. Shao, M., von Czapiewski, K.C., Heiden, A.C., et al. (2001) Volatile organic compound emissions from scots pine: Mechanisms and description by algorithms. Journal of Geophysical Research, Bd. 106 (D17): S20483–91. Shepson, P.B., Hastie, D.R. Schiff, H.I. et al. (1991) Atmospheric concentrations and temporal −C3 carbonyl compounds at two rural sites in central Ontario. Atmospheric variations of C1 − Environment, 25A: 2001–15. Shulaev, V. Silverman, P. and Raskin, I. (1997) Airborne signalling by methyl salicylate in plant pathogen resistance. Nature, 385: 718–21. Sin, D.W.M., Wong, S.C. and Louie, P.K.K. (2001) Trends of ambient carbonyl compounds in the urban environment of Hong Kong. Atmospheric Environment, 35: 5961.

Oxygenated Volatile Organic Compounds

171

Singh, H., Chen, Y., Tabazadeh, A., et al. (2000) Distribution and fate of selected oxygenated organic species in the troposphere and lower stratosphere over the Atlantic. Journal of Geophysical Research, 105: 3795–805. Singh, H.B., Tabazadeh, A., Evans, M.J., et al. (2003) Oxygenated volatile organic chemicals in the oceans: Interferences and implications based on atmospheric observations and air-sea exchange models. Geophysical Research Letter, 30 (16): 1862. Singh, H.B., Kanakidou, M., Crutzen, P.J. and Jacob, D.J. (1995) High concentrations and photochemical fate of oxygenated hydrocarbons in the global troposphere. Nature, 378: 50–4. Singh, H.B., O’Hara, D., Herlth, K., et al. (1994) Acetone in the atmosphere: Distribution, sources, and sinks. Journal of Geophysical Research, 99: 1805–19. Singh, H.B. and Zimmerman, P.R. (1992) Atmospheric distribution and sources of nonmethane hydrocarbons. In: J.O. Nriagu (Ed.) Gaseous Pollutants: Characterization and Cycling. WileyInterscience. Wiley, New York. Singh, H.B. and Hanst, P.L. (1981) Peroxyacetyl nitrate/PAN/in the unpolluted atmosphere – An important reservoir for nitrogen oxides. Geophysical Research Letter, 8: 941–4. Slemr, J. and Junkermann, W. (1996) Temporal variations in formaldehyde, acetaldehyde, and acetone and budget of formaldehyde at a rural site in Southern Germany. Atmospheric Environment, 30: 3667–76. Solberg, S., Schmidtbauer, N., Semb, A., Stordal, F. and Hov, O. (1996) Boundary-layer ozone depletion as seen in the Norwegian Arctic in spring. Journal of Atmospheric Chemistry, 23: 301–32. Staehelin, J., Graber, N. and Widmer, H.M. (1991) Automated-analysis of light-hydrocarbons in ambient air using the commercially available system AAR-KA 5890A – Technical adaptations and experience from ﬁeld-measurements. International Journal of Environmental Analytical Chemistry, 43 (4): 197–208. Sturges, W.T. and Elkins, J.W. (1993) Use of adsorbents to collect selected halocarbons and hydorhalocarbons of environmental interest from large air volumes. Journal of Chromatography, 642 (1–2): 123–34. Tabazadeh, A., Yokelson, R.J., Singh, H.B., Hobbs, P.V., Crawford, J.H. and Iraci, L.T. (2004) Heterogeneous chemistry involving methanol in tropospheric clouds. Geophysical Research Letter, 31: L06114. Talbot, R.W., Andreae, M.O., Berresheim, H., Jacob, D.J. and Beecher, K.M. (1990) Sources and sinks of formic, acetic, and pyruvic acids over central Amazonia. II – Wet season. Journal of Geophysical Research, 95: 16799–811. Tingey, D.T., Turner, D.P. and Weber, J.A. (1991) Factors controlling the emission of monoterpenes and other volatile organic compounds. In: T.D Sharkey, E.A. Holland and H.A. Mooney (Eds) Trace gas emissions by plants, Academic Press, San Diego. Vairavamurthy, A., Roberts, J.M. and Newman, L. (1992) Methods for determination of low molecular weight carbonyl compounds in the atmosphere, a review. Atmospheric Environment, 26A: 1965–1993. Viskari, E.L., Vartiainen, M. and Pasanen, P. (2000) Seasonal and diurnal variation in formaldehyde and acetaldehyde concentrations along a highway in Eastern Finland. Atmospheric Environment, 34: 917–23. Walter, A. and Schurr, U. (2005) Dynamics of leaf and root growth: Endogenous control versus environmental impact. Annals of Botany, 95, 891–900. http://dx.doi.org/10.1093/aob/mci103 Warneke, C., Karl, T., Judmaier, H., et al. (1999) Acetone, methanol, and other partially oxidized volatile organic emissions from dead plant matter by abiological processes: Signiﬁcance for atmospheric HOx chemistry. Global Biogeochemical Cycles, Bd. 13 (1): S9–17. Warneke, C., Luxembourg, S., de Gouw, J., Rinne, J., Guenther, A. and Fall, R. (2002) Disjunct eddy covariance measurements of oxygenated VOC ﬂuxes from an alfalfa ﬁeld before and after cutting. Journal of Geophysical Research, 107: ACH6-1 to ACH6-11.

172

Volatile Organic Compounds in the Atmosphere

Wedel, A., Müller, K.P., Ratte, M. and Rudolph, J. (1998) Measurements of volatile organic compounds (VOC) during POPCORN 1994: Applying a new on-line GC–MS-technique. Journal of Atmospheric Chemistry, 31: 73–103. Welke, B., Ettlinger, K. and Riederer, M. (1998) Sorption of volatile organic chemicals in plant surfaces. Environmental Science and Technology, 32: 1099–1104. Wennberg, P.O., Hanisco, T.F., Jaeglé, L., et al. (1998) Hydrogen radicals, nitrogen radicals, and the production of O3 in the upper troposphere. Science, Bd. 279 (5347): S49–53. Wildt, J., Kobel, K., Schuh-Thomas, G. and Heiden, A.C. (2003) Emissions of oxygenated volatile organic compounds from plants, part II: Emissions of saturated aldehydes. Journal of Atmospheric Chemistry, 45: 173–96. Winer, A.M., Arey, J., Atkinson, R., et al. (1992) Emission rates of organics from vegetation in California’s Central Valley. Atmospheric Environment, 26A: 2647–59. Wittrock, F., Richter, A., Oejten, H., et al. (2006) Simultaneous global observations of glyoxal and formaldehyde from space. Geophysical Research Letter, 33: L16804. Wohlfrom, K.-H., Hauler, T., Arnold, F. and Singh, H. (1999) Acetone in the free troposphere and lower stratosphere: aircraft based CIMS and GC measurements over the North Atlantic and a ﬁrst comparison. Geophysical Research Letters, 26: 2849–52. Yokelson, R.J., Goode, J.G., Ward, D.E., et al. (1999) Emissions of formaldehyde, acetic acid, methanol, and other trace gases from biomass ﬁres in North Carolina measured by airborne fourier transform infrared spectroscopy. Journal of Geophysical Research, 104: 30109–25. Yokouchi, Y., Mukai, H., Nakajima, K. and Ambe, Y. (1990) Semi-volatile aldehydes as predominant organic gases in remote areas. Atmospheric Environment, 24A: 439–42. Zeidler, J. and Lichtenthaler, H.K. (2001) Biosynthesis of 2-methyl-3-buten-2-ol emitted from needles of pinus ponderosa via the non-mevalonat DOXP/MEP pathway of isoprenoid formation. Planta, Bd. 213 (2): S323–6. Zhou, X. and Mopper, K. (1997) Photochemical production of low-molecular-weight carbonyl compounds in seawater and surface microlayer and their air-sea exchange. Marine Chemistry, 56:, 201–13.

Volatile Organic Compounds in the Atmosphere Edited by Ralf Koppmann Copyright © 2007 by Blackwell Publishing Ltd

Chapter 5

Halogenated Volatile Organic Compounds Simon J. O’Doherty and Lucy J. Carpenter

5.1

Introduction

The importance of halogenated volatile organic compounds (VOCs) in the atmosphere became apparent in the early 1970s as a result of the use of new measurement tools and the proposal that chloroﬂuorocarbons (CFCs) released into the atmosphere could result in chlorine chemistry in the stratosphere, which might ultimately lead to catalytic ozone destruction. Lovelock developed the electron capture detector (ECD) in the late 1950s (Lovelock 1962, 1963; Simmonds et al. 1967). This novel detector was used by Lovelock in 1969 to make the ﬁrst measurements of CFCl3 (CFC-11) near his home in Bowerchalke, United Kingdom, and Adrigole, Ireland. Soon, it became apparent from these early measurements that the CFCs were accumulating in the Earth’s atmosphere without any means for removal. Early vertical proﬁles carried out by Lovelock also indicated that the CFCs exhibited steady levels in the troposphere and a decline in concentration in the stratosphere (Lovelock 1974). Tropospheric inertness of the CFCs was thus conﬁrmed, and lifetimes of up to hundreds of years were indicated. These long lifetimes, coupled with their strong infrared absorptions, also make the CFCs signiﬁcant greenhouse gases. Around the same time, a catalytic cycle of ozone destruction in the stratosphere involving chlorine released from CFCs was proposed by Molina and Rowland (1974a) and Stolarsk and Cicerone (1974). These ﬁndings were conﬁrmed by scientists of the British Antarctic Survey (BAS), who made regular ozone concentration measurements from their base at Halley Bay at 76◦ S. They detected a decline in springtime ozone concentration of the order of 30%; in addition, the depletion was increasing year by year (Farman et al. 1985). The combination of these ﬁndings sparked the CFC chemistry and ozone depletion debate, one of the most important environmental debates of the century, and propelled halogenated VOCs to the forefront of atmospheric chemistry. The Montreal Protocol on substances that deplete the ozone layer was brought into existence in 1987 in an effort to control and reduce the production and use of halocarbons; this protocol and its Amendments have caused dramatic reductions in industrial halocarbon emissions with an overall downturn in total tropospheric chlorine. A summary of the Montreal Protocol is described in Box 5.1. When considering the halogen budget in the atmosphere, many other halogenated compounds must also be taken into account. Biogenic (natural) sources of halocarbons, contribute considerably to the total halogen budget, CH3 Cl is the most abundant halocarbon in the atmosphere, and is largely derived from natural sources. Bromine and iodine compounds are minor constituents in the Earth’s atmosphere; however, they

174

Box 5.1

Volatile Organic Compounds in the Atmosphere

Montreal Protocol

In 1985, the Vienna Convention established mechanisms for international co-operation in research into the ozone layer and the effects of ozone-depleting chemicals (ODCs). The ﬁrst discovery of the Antarctic ozone hole also occurred in 1985. On the basis of the Vienna Convention, the Montreal Protocol on Substances that Deplete the Ozone Layer was negotiated in Montreal and signed by 24 countries and by the European Economic Community in September 1987. The Protocol became effective in 1989 and called for the parties to phase down the use of CFCs, halons and other man-made ODCs. The Montreal Protocol stipulated that the production and consumption of compounds that deplete ozone in the stratosphere (CFCs, halons, carbon tetrachloride (CCl4 ) and methylchloroform) were to be phased out by 2000 (2005 for methylchloroform). The Montreal Protocol on Substances that Deplete the Ozone Layer was one of the ﬁrst international environmental agreements that included trade sanctions to achieve the stated goals of a treaty. It also offered major incentives for non-signatory nations to sign the agreement. The treaty negotiators justiﬁed the sanctions because depletion of the ozone layer is an environmental problem most effectively addressed on the global level. Furthermore, without the trade sanctions, there would be economic incentives for non-signatories to increase production, damaging the competitiveness of the industries in the signatory nations as well as decreasing the search for less damaging CFC alternatives. At meetings in London (1990), Copenhagen (1992), Vienna (1995), Montreal (1997) and Beijing (1999), amendments were adopted that were designed to speed up the phasing out of ozone-depleting substances. Table 5.1 summarises the phase-out timetable for selected compounds.

have a much higher efﬁciency of ozone destruction compared to chlorine. Our understanding of natural halocarbon source and sinks is still poor. Anthropogenic (man-made) halocarbons such as the hydrochloroﬂuorocarbons (HCFCs) and hydroﬂuorocarbons (HFCs) are currently being developed and used as replacements for the CFCs. Perﬂuorocarbons (PFCs) are employed in a variety of applications; these compounds can have atmospheric lifetimes as long as 50 000 years. The halons, fully halogenated compounds containing bromine have been used as ﬁre suppressants. To confuse matters even further, many halogenated VOCs have mixed anthropogenic and biogenic inputs. In this chapter, we will describe the current understanding of the sources and sinks, trends, distribution and chemistry of halogenated VOCs in the atmosphere. A well-deﬁned knowledge of these factors is required to understand the present-day distribution of stratospheric ozone and to predict its future behaviour in response to natural and anthropogenic forcings. All of these gases play a role in the forcing of climate change by altering the infrared radiative budget of the atmosphere (IPCC 2001; WMO 1998; WMO 2003). The importance of these gases in the chemistry of both the marine boundary layer and free troposphere is also of interest because they are involved in many reaction cycles that can change the oxidation power of the atmosphere indirectly by inﬂuencing the main oxidants O3 and its photolysis product OH, and directly by reaction of the Cl radical with hydrocarbons

Halogenated Volatile Organic Compounds

175

Table 5.1 The Montreal Protocol phase-out schedule for selected compounds Ozone-depleting substances

Developed countries

Developing countries

CFCs

Phased out at end of 1995a

Halons

Phased out at end of 1993

Total phase out by 2010

Carbon tetrachloride

Phased out at end of 1995a

Total phase out by 2010

Methylchloroform

Phased out at end of 1995a

Total phase out by 2015

HCFCs

Freeze from beginning of 1996b

Freeze in 2016 at 2015 base level Total phase out by 2040

35% reduction by 2004 65% reduction by 2010 90% reduction by 2015 Total phase out by 2020c Methyl bromide

Freeze in 1995 at 1991 base leveld 25% reduction by 1999 50% reduction by 2001 50% reduction by 2000 Total phase out by 2005f

Total phase out by 2010

Freeze in 2002 at average 1995–1998 base level 20% reduction by 2005e Total phase out by 2015

a Developed nations can continue to produce CFCs up to 15% of their 1986 baseline to help developing countries meet their domestic needs and for essential uses such as medical devices. b Based on 1989 HCFC consumption with an extra allowance (ODP weighted) equal to 2.8% of 1989 CFC consumption. c Up to 0.5% of base level consumption can be used until 2030 for servicing existing equipment. d All reductions include an exemption for pre-shipment and quarantine uses. e Review in 2003 to decide on interim further reductions beyond 2005. f Except for applications granted Critical Use Exemption (CUE).

(e.g. CH4 ). We will consider the huge volume of research that has been carried out (at different research locations, using different measurement techniques associated with various sampling platforms) in an attempt to understand the role of halogenated VOCs in the atmosphere. The main factors determining the atmospheric concentration of an individual halocarbon are its sources and sinks. Production and subsequent releases into the atmosphere deﬁne the sources of individual halocarbons. These sources can be anthropogenic or biogenic, and the magnitude of the sources varies spatially and temporally. The residence time (or atmospheric lifetime) of a halocarbon is deﬁned by its rate of removal or sink. The average atmospheric lifetime not only controls the build-up of the halocarbon in the atmosphere, but also impacts on the global distribution of halocarbons and the role in evaluating the impact of individual species on the stratospheric and tropospheric ozone depletion, climate change, regional pollution and on future scenarios, and ﬁnally determines the timescale for the atmosphere to recover after emissions cease. Two key indices used to assess the environmental effect of halocarbons in the atmosphere are described below.

176

Volatile Organic Compounds in the Atmosphere

5.1.1

Ozone depletion potential

A method of expressing the effect of a source gas in depleting stratospheric ozone is in terms of ozone depletion potential (ODP). This term relates the amount of ozone destroyed by a source gas relative to that of CFC-11 (CFCl3 ) on a kilogram-for-kilogram basis. ODPi =

Global O3 loss due to unit mass emission of i Global O3 loss due to unit mass emission of CFC-11

ODPs are calculated using a variety of models and semi-empirical approaches. Original ODP calculations presupposed gases were at steady state; however, more recently the concept of time-dependent ODP provides information on the near term effects of a short-lived species relative to the near term effects of CFC-11. Compounds such as HCFCs will release their chlorine much earlier than CFC-11, because they have shorter atmospheric lifetimes; hence the time-dependent ODPs calculated for the HCFCs are much larger than their steady state values. This is discussed in much greater detail in WMO (2003).

5.1.2

Global warming potential

The climate impact of a given mass of a halocarbon emitted into the atmosphere can be estimated from its radiative properties and atmospheric lifetime. The two can be combined to provide the global warming potential (GWP) that is a proxy for the climate effect of a gas. GWP is the ratio of the warming caused by a substance to the warming caused by a similar mass of carbon dioxide. Thus, the GWP of CO2 is deﬁned to be 1.0. Mathematically, the GWP of component x is given by TH ax × [x(t )]dt GWPx (TH ) = 0TH ar × [x(t )]dt 0 where TH is the time horizon over which the calculation is performed, ax is the radiative forcing per unit mass due to an increase in the atmospheric abundance of the species, x(t ) is the time-dependent decay in abundance of species x, ar and r(t ) are the corresponding quantities for the reference gas, CO2 . This is a convenient index to use for the purposes of regulation; however, there are many fundamental problems with this index. Uncertainties include the use of CO2 as a reference gas because its lifetime is long and very uncertain. The general deﬁnition is suitable for long-lived gases that are well-mixed in the atmosphere and do absorb the thermal infrared radiation. The index is much less useful for short-lived species because these species are not uniformly distributed in the troposphere. Their distributions depend on where and when (time of year) they are emitted. Thus, it is not possible to assign a single steady-state change in burden in the troposphere per unit mass emission. These index parameters are illustrated in Table 5.2. GWP values are subject to change as better estimates become available for either the radiative forcing associated with the gas, or the lifetime associated with the gas or the reference (CO2 ). Relative to the IPCC (1995)-reported values, the recommended GWPs (over 100 years) have been modiﬁed from −16% to +51%, depending on the gas, with an average of −15%.

SF6 CF4 C2 F6 C3 F8 C4 F10 C6 F14 c-C4 F8 CCl3 F CCl2 F2 CCl2 FCClF2 CClF2 CClF2 CClF2 CF3 CBrF3 CBrClF2 CBrF2 CBrF2 CCl4 CH3 Br CH2 BrCl CH3 CCl3 CH2 Cl2 CH3 Cl

PFC-14 PFC-116 PFC-218 PFC-31-10 PFC-51-14 PFC-318

CFCs CFC-11 CFC-12 CFC-113 CFC-114 CFC-115

Halons Halon-1301 Halon-1211 Halon-2402

Chlorocarbons Carbon tetrachloride Methyl bromide Bromochloromethane Methylchloroform Methylene chloride Methyl chloride

PFCs

5b 0.46 1.3

26b 0.7

65 16b 20b

45 100 85 300 1 700

3 200 50 000 10 000 2 600 2 600 3 200 3 200

Lifetimea (years)

0.06 0.03 0.01

0.13 0.01

0.32 0.3 0.33b

0.25 0.32 0.3 0.31 0.18

0.52 0.08 0.26 0.26 0.33 0.49 0.32

Radiative Forcinga (Wm−2 /ppb)

476 35a 55a

2 540 16

7 970 4 460 3 460

6 300 10 200 6 150 7 560 4 990

15 290 3 920 8 110 5 940 5 950 6 230 6 870

TH-20b

144 10a 16a

1 380 5

7 030 1 860 1 620

4 600 10 600 6 030 9 880 7 250

22 450 5 820 12 010 8 690 8 710 9 140 10 090

TH-100 b

GWP

0.12 <0.001f 0.02f

0.73 0.38

12.0 6.0 <8.6

1 1.0 1.0 0.9f 0.4f

ODPe

Table 5.2 Lifetimes, Radiative Forcings, GWPs and ODPs for the ozone-depleting substances (ODSs) and their replacements

CHF3 CH2 F2 CHF2 CF3 CH2 FCF3 CH3 CF3 CH3 CHF2 CF3 CHFCF3 CF3 CH2 CF3 CHF2 CH2 CF3 CH3 CF2 CH2 CF3

HFCs HFC-23 HFC-32 HFC-125 HFC-134a HFC-143a HFC-152a HFC-227ea HFC-236fa HFC-245fa HFC-365mfc

Notes: a From IPCC (2001). b Updated in WMO (2002). c Updated from two averaged model results. d Scaled to new RE. e From WMO (2002). f From Singh and Fabian (1999).

CHClF2 CHCl2 CF3 CHClFCF3 CH3 CCl2 F CH3 CClF2 CHCl2 CF2 CF3 CHClFCF2 CClF2

HCFCs HCFC-22 HCFC-123 HCFC-124 HCFC-141b HCFC-142b HCFC-225ca HCFC-225cb

Table 5.2 (Continued)

270b 4.9b 29 14b 52 1.4 34.2b 240b 7.6b 8.6b

12b 1.3b 5.8b 9.3 17.9b 1.9b 5.8b

Lifetimea (years)

0.187c 0.111c 0.23 0.161c 0.13 0.09 0.257c 0.28 0.28 0.21

0.2 0.14b 0.22 0.14 0.2 0.2b 0.32

Radiative Forcinga (Wm−2 /ppb)

11 100d 2 220d 5 970 3 600d 5 540 411 4 930d 7 620 3 180 2 370

4 850 257 1 950 2 120 5 170 404 1 910

TH-20b

14 300d 670d 3 450 1 410d 4 400 122 3 140d 9 500 1 020 782

1 780 76 599 713 2 270 120 586

TH-100b

GWP

<3 × 10−5 <1.5 × 10−5

<4 × 10−4

0.05 0.02 0.02 0.12 0.07 0.02 0.03

ODPe

Halogenated Volatile Organic Compounds

179

Section 5.2 will explain and categorise the nature and location of sources by compound type; in Section 5.3, the most recent distributions of halocarbons will be presented; and in Section 5.4, the processes that lead to the transformation and removal of each class of halocarbon from the atmosphere will be deﬁned. Section 5.5 deals with methods of deriving global emissions of halogenated VOCs, Section 5.6 discusses sampling techniques and, ﬁnally, Section 5.7 outlines some of the key instrumentation used to make the atmospheric measurements detailed in this chapter.

5.2

Sources of halogenated VOCs

The atmosphere contains large amounts of halocarbons, primarily in the form of chlorine, with lesser amounts of ﬂuorine, bromine and iodine. Halogens can be present in their inorganic (e.g. HCl) or organic forms (e.g. CH3 Cl). The source of these compounds is related to several natural and anthropogenic processes at the Earth’s surface and their atmospheric lifetimes range from minutes to centuries. Chlorine containing compounds provide by far the largest ﬂux of halocarbons to the atmosphere. Tropospheric sources of inorganic chlorine are dominated by emissions of HCl from volatilisation of sea salt aerosol, emissions also occur from volcanoes (Bobrowski et al. 2003), and a variety of anthropogenic sources such as combustion of coal (McCulloch et al. 1999a). The ﬂux is large (possibly 100 Tg Cl/year, Tg = 1012 g) but very uncertain, moreover, sea salt and HCl are efﬁciently returned to the Earth’s water and land surfaces, minimising their long-range transport through the atmosphere and conﬁning them to the troposphere. Keene et al. (1999) have compiled existing data on the concentration and ﬂuxes of HCl. Although organic halogen compounds can be sub-divided into many classiﬁcations such as, halogen type, anthropogenic or biogenic sources, the WMO report 1998 chose long-lived and short-lived species, this division was useful as it implies the location of their chemical destruction. Long-lived halocarbons are destroyed almost exclusively by photochemical processes in the stratosphere and have lifetimes in the order of 16–50 000 years. These halocarbons are important because after their release at the Earth’s surface they mix rapidly in the troposphere, and are advected into the stratosphere where photolysis leads to the formation of inorganic halogen species. It is these inorganic species that participate in the catalytic destruction of ozone. Short-lived source gases react predominantly with tropospheric hydroxyl radicals (OH) and have lifetimes on the order of minutes to several years. Whatever classiﬁcation is chosen, there is no deﬁnitive division between species because many classes of gases share common factors in both their life cycles and effects. In this chapter we will classify halocarbons by halogen type and attempt to review the current understanding of each halogen type in terms of sources, sinks, trends and distribution.

5.2.1

Mixed chlorine and ﬂuorine compounds

The system for naming compounds described in this section is as follows: CFC-01234a is: 0 = Number of double bonds (omitted if zero) 1 = Carbon atoms −1 (omitted if zero)

180

Volatile Organic Compounds in the Atmosphere

2 = Hydrogen atoms +1 3 = Fluorine atoms 4 = Replaced by Bromine (‘B’ preﬁx added) a = Letter added to identify isomers, the ‘normal’ isomer in any number has the smallest mass difference on each carbon, and a, b or c are added as the masses diverge from normal.

5.2.1.1

Chloroﬂuorocarbons (CFCs)

Following their introduction in the 1930s as safe refrigerants, production and emissions of CFCs 11 and 12 (CCl3 F and CCl2 F2 ), the original compounds in the group, remained comparatively low through to the 1950s. They then increased rapidly as applications of CFCs grew to encompass use as aerosol propellants, in air conditioning, as blowing agents for energy-efﬁcient foam insulation and packaging, and as cleaning solvents. As non-ﬂammable chemicals with low toxicity and reactivity, CFCs were acceptable from the standpoint of worker and consumer safety. Annual production grew steadily, reaching a peak total of over 800 000 metric tonnes of CFC-11 and CFC-12 by 1974, the year in which their potential role in ozone depletion was ﬁrst postulated (McCulloch 1999; Midgley 1997). In 2003, the comparable ﬁgure was less than 12 800 metric tonnes (http://www.afeas.org/production_and_sales.html). CFC-113 was widely used in solvent and dry-cleaning applications. Worldwide production of CFC-113 (CClF2 CCl2 F) dropped from 250 000 to 599 metric tonnes between 1989 and 2003. A number of less abundant CFCs have are emitted into the atmosphere from a variety of uses such as specialist refrigerants (CFC-114, CClF2 CClF2 ; CFC-115, CClF2 CF3 ; CFC-13, CClF3 ), aerosol propellants (CFC-114 and CFC-115) and foam blowing agents (CFC-114). Small amounts of CFC-13 are also emitted during aluminium production and during the manufacture of CFC-12. Calculated emissions of CFC-114 (2005) are less than 10% of the peak year (1989). Production of the CFCs, have been phased out in developed countries by the 1987 Montreal Protocol report and subsequent amendments on regulation of substances that deplete the ozone layer. Although emissions of the CFCs have declined rapidly, concentrations have not declined as fast; this is due to the long lifetimes of these compounds; CFC-11 (45 year), CFC-12 (100 year), CFC-13 (640 year), CFC-115 (1 700 year), CFC-114 (300 year).

5.2.1.2

Hydrochloroﬂuorocarbons

HCFCs are organic compounds that consist of carbon, hydrogen, chlorine and ﬂuorine. They are much more reactive than CFCs, so a large fraction of the HCFCs emitted break down in the troposphere by reaction with the hydroxyl radical, and hence are removed before they have a chance to affect the ozone layer. These compounds do have the capacity to trap heat in the atmosphere and as such have a relatively large GWP. Global atmospheric levels of the principal HCFCs have been increasing rapidly in the atmosphere since their introduction (in the early 1970s, 1980s and 1990s for HCFC-22 (CHClF2 ), HCFC-142b (CH3 CCl2 F) and HCFC-141b (CH3 CClF2 ), respectively). This is due to sustained emissions associated with their use as replacements for stratospheric ozone depleting CFCs. HCFC-142b and HCFC-141b are mainly used as foam blowing agents, largely replacing CFC-11, although relatively small amounts of HCFC-141b are also used as a solvent, replacing CFC-113 in high-precision cleaning. HCFC-22 is commonly used in self-contained air conditioners, but

Halogenated Volatile Organic Compounds

181

the major uses are in commercial scale chillers and cold storage refrigeration. HCFC-22 and HCFC-142b are also used as chemical feedstocks, and as such a fraction of their production is captivity converted and not emitted to the atmosphere. The HCFCs are controlled under the Montreal Protocol as ‘transitional substances’ with a 2 030 phase-out in developed countries and a 2 040 phase-out in developing countries. The production and sales of HCFCs and HFCs for emissive uses are surveyed annually by the Alternative Fluorocarbons Environmental Acceptability Study (AFEAS). AFEAS participants include all of the world’s major producers, located in Europe, North America and Japan, and their subsidiaries, giving virtually complete global coverage for HFCs and for most HCFCs. Industrial sales of HCFC-22 from companies reporting to AFEAS increased steadily from about 56 × 106 kg/year in 1970 to 245 × 106 kg/year in 1992. In the most recent years for which data exist (1993–2003), sales into dispersive uses have remained fairly constant, with approximately 95% of sales in the Northern Hemisphere (NH) (http://www.afeas.org/). Industrial production of HCFC-141b and HCFC-142b increased markedly in the late 1980s and early 1990s as these HCFCs found use as replacements for CFCs and chlorinated solvents. Although production of HCFC-142b (initially mainly for feedstock use) began earlier than that of HCFC-141b, by 1995 nearly three times as much HCFC-141b was sold into dispersive uses. Sales ﬁgures for that year were 113 × 106 kg for HCFC-141b and 39 × 106 kg for HCFC-142b, and indicate that more than 99% of sales of these compounds were in the NH (http://www.afeas.org/). Sales ﬁgures for both compounds peaked in 2000 and have now reduced in response to the Montreal Protocol, with sales in 2003 being 74 × 106 kg for HCFC-141b and 20 × 106 kg for HCFC-142b.

5.2.2 5.2.2.1

Fluorine compounds Hydroﬂuorocarbons

HFCs are organic compounds that contain hydrogen, carbon and ﬂuorine. HFCs, which do not contain chlorine, do not have any potential for the destruction of ozone, and so are thought to be suitable replacements for CFCs. HFC-134a (CH2 FCF3 ) is the preferred substitute for CFC-12, primarily in refrigeration and air-conditioning applications. Industrial sales of this ‘chlorine-free’ refrigerant increased from 0 in 1990 to 167 × 106 kg in 2003 (http://www.afeas.org/). Demand for other HFCs such as HFC-125 (CHF2 CF3 ), HFC-143a (CH3 CF3 ), HFC-152a (CH3 CHF2 ), HFC-227ea (CF3 CHFCF3 ) and HFC-365mfc (CH3 CF2 CH2 CF3 ) have also increased in recent years. HFC-23 (CHF3 ) is emitted into the atmosphere as a by-product during the manufacture of HCFC-22 (via overﬂuorination of CHCl3 ). The major use of HFC-32 (CH2 F2 ) and 143a is in refrigerant blends with other HFCs. Emissions of these HFCs are expected to arise only from anthropogenic sources, and this has been conﬁrmed from ﬁrn air (unconsolidated snow) measurements that show a clear absence of any natural sources (Butler et al. 1999; Sturrock et al. 2002).

5.2.2.2

Perﬂuorocarbons

This group of fully ﬂuorinated compounds have extremely high GWP values as a result of their ability to absorb and trap heat in the atmosphere coupled with their very long

182

Volatile Organic Compounds in the Atmosphere

atmospheric lifetimes. The major emission source for CF4 is aluminium electrolysis, speciﬁcally when a fault condition occurs in the electrolysis process (a condition known as an Anode Effect). Weston (1996) reviewed measurements of CF4 concentrations in the atmosphere and combined aluminium production rates to provide an estimate of 1.3–3.6 kg of CF4 emitted per ton of aluminium produced. Natural sources of CF4 outgassing from natural ﬂuorites (CaF2 ) and igneous and metamorphic rocks has also been reported (Harnisch and Eisenhauer 1998; Harnisch et al. 2000). C2 F6 is also emitted from an Anode Effect during aluminium smelting; CF4 and C2 F6 should be emitted in an average reported ratio of 10 : 1. The semiconductor industry uses C2 F6 in applications such as silicon etching and CVD (Chemical Vapour Deposition) chamber cleaning.

5.2.3 5.2.3.1

Chlorine compounds Methyl chloride

Methyl chloride (CH3 Cl) is by far the most abundant chlorocarbon in the atmosphere with a reported 5 Tg in the earth’s atmosphere and with a reported mixing ratio in the range of 530–560 ppt (Montzka and Fraser 2003; Simmonds et al. 2004). The reported sources of CH3 Cl in order of magnitude are biomass burning (Andreae and Merlet 2001; Lobert et al. 1999), oceans (Cox et al. 2003; Singh et al. 1983), tropical plants (Yokouchi et al. 2000, 2002), wood-rotting fungi (Moore et al. 2005; Watling and Harper 1998), salt marshes (Rhew et al. 2002), coal combustion, wetlands, waste incineration, industry and rice paddies (Redeker et al. 2000). Although largely natural in origin, CH3 Cl is responsible for about 16% of chlorine-catalysed ozone destruction in the stratosphere (Montzka and Fraser 2003). The global budget for CH3 Cl is far from certain with known sinks signiﬁcantly outweighing sources by about 2.2 Tg/year (tera = 1012 ). The ocean source has been questioned by Tokarczyk et al. (2003), who suggest that direct measurements of CH3 Cl degradation rates in coastal seawater (Bedford Basin, Nova Scotia), using a stable isotope incubation technique, indicate rapid loss attributed to microbial activity and suggested that this could make the oceans a major sink for CH3 Cl and lower the overall atmospheric lifetime of CH3 Cl from the 2003 estimate of 1.3 to about 1.0 years. Measurements of trace gases in air trapped in polar ﬁrn have demonstrated that the pre-twentieth-century burden of CH3 Cl was close to that at present (Butler et al. 1999). This reinforces the view that CH3 Cl is predominantly natural in origin, and measurements of its latitudinal gradient suggest that terrestrial emissions are located in the tropics. In an attempt to reduce the uncertainties surrounding the budget of CH3 Cl, Keppler et al. (2005) have used a stable isotope ratio (13 C/12 C) technique to suggest that abiotic methylation of chloride in plants and soil organic matter accounts for 1.8–2.5 Tg/year and that the microbial soils sink is much larger than previously assumed (>1 Tg/year).

5.2.3.2

Chloroform

The emissions of chloroform (CHCl3 ) to the atmosphere are primarily (90%) from natural sources (McCulloch 2003). These sources are poorly characterised; however, it has been reported that ocean sources predominate at 360 ± 90 Gg/year (giga = 109 ), followed by soils 220 ± 100 Gg/year; in contrast, industrial sources such as pulp/paper

Halogenated Volatile Organic Compounds

183

manufacture and water treatment total less than 70 Gg/year. The open ocean source estimates (Khalil et al. 1983, 1999) have been questioned by O’Doherty et al. (2001), who used modelling of global CHCl3 data to infer much larger emissions from soils and suggested that an ocean source could only account for 20 ± 20% of total emissions. Reconstructed atmospheric levels of CHCl3 using Law Dome, Antarctic ﬁrn air suggest that ocean emissions are less than estimated by Khalil et al. (1999), but anthropogenic sources for CHCl3 are greater than previously thought (Trudinger et al. 2004). However, the industrial estimates from (McCulloch 2003) are believed to be well constrained, leading to the possibility that human inﬂuence through changes in agriculture have increased soil sources. The global average mixing ratio of CHCl3 ranges from 10 to 20 ppt, and NH mixing ratios are about a factor of 2 higher than those in the SH (Khalil et al. 1983; O’Doherty et al. 2001).

5.2.3.3

Methylene chloride

The major sources of methylene chloride (CH2 Cl2 ) are anthropogenic (about 70%) from industrial and commercial uses. Estimates suggest emissions of 160 Gg/year in 1960, increasing up to a peak of 650 Gg/year in the mid-1980s (Trudinger et al. 2004). Estimated industrial emissions from audited sales data were about 580 Gg in 1990, and have steadily declined since then at about 12 Gg/year (McCulloch et al. 1999b). The nature and magnitude of natural CH2 Cl2 sources are very uncertain. There are natural emissions from oceans of 160 Gg/year (Khalil et al. 1999) and biomass burning 50 Gg/year (Lobert et al. 1999), although the ocean source is reported to be ‘poorly constrained’ (Baker et al. 2000). The major removal process for CH2 Cl2 in the troposphere is destruction by OH with an atmospheric lifetime of 5–6 months (Ko and Poulet 2003). Graedel and Keene (1996) estimate that about 2% of CH2 Cl2 emissions reach the stratosphere. Measurements near the tropical tropopause in 1992 averaged 14.9 ± 1.1 ppt (Schaufﬂer et al. 1993, 2003). The atmospheric history of CH2 Cl2 reconstructed from the analysis of Antarctic ﬁrn air shows a steep increase in concentration from about 1960, peaking in 1990, and then a decrease. Furthermore, the ﬁrn record suggests signiﬁcantly lower emissions due to the oceans than the estimates by (Khalil et al. 1999), and an enhanced terrestrial source (Trudinger et al. 2004). Southern Hemisphere (SH) CH2 Cl2 levels in the pre-1940s were 1–2 ppt, 15–20% of peak levels in the 1980s–1990s, suggesting that natural emissions are less than 20% of peak emissions. Tropospheric mixing ratios of 40–50 ppt in the NH and 15–20 ppt in the SH were obtained during ocean cruises (Koppmann et al. 1993). Similar results have been reported from several NH and SH ground-based sites by (Elkins 1997). The average background concentration observed at Cape Grim, Tasmania during 1998–2000 was 8.9 ± 0.2 ppt (Cox et al. 2003).

5.2.3.4

Carbon tetrachloride

The source of this compound is purely anthropogenic. Prior to the Montreal Protocol taking effect, CCl4 was primarily used as a chemical intermediate in the production of CFC-11 and CFC-12. There are large discrepancies between the observed burdens and trends of this gas and reported consumption data. Suggested emissions of 41 Gg from a total production of 203 Gg in 1996 only account for half the emissions required to explain the observed concentrations in the atmosphere (WMO 2003). The likely explanation is that not all production and usage has been reported and/or estimates of CCl4 lifetime are incorrect.

184

5.2.3.5

Volatile Organic Compounds in the Atmosphere

Methylchloroform

Methylchloroform (CH3 CCl3 ) was used primarily as a cleaning solvent. It is also employed as a chemical feedstock, but that fraction of its production is not emitted into the atmosphere. Production has dropped from 270 Gg/year in the 1980s to about 126 Gg/year in 1993 (Midgley and McCulloch 1995). In 1996, there was no new production for sales into emissive uses in the developed world. Total consumption of CH3 CCl3 in developing countries operating under Article 5 of the Montreal Protocol was approximately 30 × 106 kg in 1994 (United Nations Environment Programme (UNEP) 1997). It has been estimated that these requirements were mostly met by material exported from developed countries and that production capacity in developing countries lies well below this level (Midgley and McCulloch 1995). Consumption in developing countries was phased out in 2005.

5.2.3.6

Trichloroethene

The major source of trichloroethene (C2 HCl3 ) is from industrial usage as a degreasing agent. McCulloch and Midgley (1996) report global emissions estimated from audited sales for 1988 and 1995 of 260 Gg and 240 Gg, respectively. All but about 2% of the sales are in the NH. The main removal process is with OH. It has an atmospheric lifetime of about 4–7 days. C2 HCl3 can also undergo reductive de-chlorination to 1,2-dichloroethylene through the activity of soil microbes (Kleopfer et al. 1985). Global emissions have remained relatively constant from 1988 to 1996 at about 240 Gg/year. Natural sources of C2 HCl3 are from the oceans and seawater algae (Abrahamsson et al. 1995; Khalil et al. 1999). Recently, it has been reported that salt lakes are also a natural source of C2 HCl3 due to the microbial activity of halobacteria (Weissﬂog et al. 2005). Estimated tropospheric mixing values are 1–5 ppt for the NH and 0.01–0.1 ppt for the SH, contributing to an atmospheric burden of 3.1–5.3 Gg (Ko and Poulet 2003). From weekly samples collected at Alert, Canada, between early 1992 and mid-1994, Yokouchi et al. (1996) reported winter/summer NH concentration ratios of approximately 60, with 24 winter peak concentrations of 6–8 ppt. This illustrates the very short atmospheric lifetime of C2 HCl3 and also why C2 HCl3 itself is unlikely to impact stratospheric chlorine levels.

5.2.3.7

Tetrachloroethene

Perchloroethene (C2 Cl4 ) is mainly used for dry cleaning and as a metal degreasing solvent. Small but signiﬁcant quantities of C2 Cl4 are emitted in the ﬂue gas from coal-ﬁred power plants (Garcia et al. 1992). Emissions of 580 Gg estimated for 1984 by Class and Ballschmiter (1987) are signiﬁcantly greater than global emissions derived from audited production ﬁgures for the years 1993–1996 of 261 Gg, 258 Gg, 278 Gg and 289 Gg, respectively (McCulloch and Midgley 1996). McCulloch and Midgley (1996) reported global emissions during 1990 of 366 Gg, which have declined subsequently at about 23 Gg/year. Emissions from coal-ﬁred power plants are estimated to be only 19 Gg/year. The atmospheric lifetime of C2 Cl4 is 3–4 months, and its primary atmospheric sink is reaction with OH (Olaguer 2002).

5.2.4

Bromine compounds

The current understanding of stratospheric ozone depletion chemistry resulting from halogen compounds indicates that even though the concentration of bromine in the stratosphere

Halogenated Volatile Organic Compounds

185

is much smaller than that of chlorine, bromine accounts for a substantial degree of ozone depletion because of its much greater ODP. Bromine atoms in the troposphere have also been implicated in ozone depletion events (ODEs) in Polar Regions, mainly the Arctic. Sources of the main bromine containing compounds are highlighted below.

5.2.4.1

Methyl bromide

Methyl bromide (CH3 Br) is an ozone-depleting compound and has signiﬁcant natural and anthropogenic sources. CH3 Br is the single largest carrier of bromine to the stratosphere where, on a per-atom basis, it is estimated to be 50–60 times more effective than chlorine in depleting ozone (Kurylo and Rodriguez 1999). CH3 Br has many anthropogenic sources, it is a highly effective fumigant used to control insects, nematodes, weeds, pathogens and rodents in more than 100 crops, in forests and ornamental nurseries, and in wood products. The primary uses are for soil fumigation, post harvest protection and quarantine treatments. CH3 Br is also released during anthropogenically inﬂuenced biomass burning (Blake et al. 1996; Mano and Andreae 1994), and from automobiles using leaded petrol (Thomas et al. 1997). The use of CH3 Br was phased out in 2005 under the Montreal Protocol. CH3 Br also originates in large part from natural sources such as oceans, which are estimated to be the largest identiﬁed source for atmospheric CH3 Br and the second largest sink, overall the ocean is thought to be a net sink (Baker et al. 1999; King et al. 2000). Other natural sources include biomass burning, coastal salt marshes and other terrestrial ecosystems. Like CH3 Cl, the calculated budget for CH3 Br is far from certain, with estimated sinks outweighing estimated sources. The disparity is slowly being narrowed as new ﬁndings on ecosystem net ﬂuxes are being discovered. Total emissions have been estimated to be 159 Gg/year with a wide possible range of 74–284 Gg/year (WMO 2002). Yokouchi et al. (2000) estimated the anthropogenic contribution to be 14–52% in the NH and 6–41% in the SH for a possible atmospheric lifetime (τa ) range of 0.3–1.4 year.

5.2.4.2

Halons

The halon nomenclature system consists of a four-digit number used to identify each compound. The ﬁrst digit represents the number of carbon atoms in the compound; the second digit, the number of ﬂuorine atoms; the third digit, the chlorine atoms; and the fourth the number of bromine atoms. Halon-1211 (CBrClF2 ) is used almost exclusively in portable equipment (hand-held ﬁre extinguishers and, to a limited extent, in larger capacity wheeled units mainly used at civil and military airports). Halon-1301 (CBrF3 ) is employed primarily in automatic ﬁxed systems (also known as total-ﬂooding systems). These ﬁre protection systems are designed to provide a ﬁre extinguishing concentration of the agent. Applications include: computer rooms, telecommunications facilities, control rooms, shipboard machinery spaces, aircraft engines and cargo bays, and other similar high-value ﬁre risks. The production of halons peaked in 1998 and despite limits imposed on halon production in developed nations; concentrations in the atmosphere continue to rise. This is thought to be due to emissions from halons that have been banked; these reserves of halons are large compared to the relative 2005 emission rates. The reserves of Halon-1211 have been estimated to be between 65 and 132 Gg, and Halon-1301 between 45 and 65 Gg (WMO 2003). Two lesser-abundant halons are also present in the atmosphere, halon-2402 (CBrF2 CBrF2 ) used a ﬁre extinguisher

186

Volatile Organic Compounds in the Atmosphere

primarily in the former Soviet Union, and halon-1202 (CBr2 F2 ) used by the military in a few minor applications or produced by over-bromination during halon-1211 production. All of these compounds have high ODP values and have been reported to contribute about one third of total bromine entering the stratosphere (Wamsley et al. 1998).

5.2.4.3

Bromoform

Bromoform (CHBr3 ) is the second most abundant ‘reactive’ organobromine gas in the background troposphere (Sturges et al. 2000). Atmospheric concentrations range from 1.8 to 20 ppt. It is a potentially signiﬁcant contributor to reactive bromine in the upper atmosphere/lower stratosphere because it carries three bromine atoms. The major sources of CHBr3 are thought to be marine in origin, namely macroalgae (Gschwend et al. 1985; Sturges et al. 1992), ocean phytoplankton (Carpenter et al. 2003) and water chlorination (Gschwend et al. 1985). However, research into terrestrial sources such as spruce forests (Haselmann et al. 2000), soil (Hoekstra et al. 1998) and coastal soil/peatlands are showing that these sites might provide a large seasonal input (as much as 10% of the global source). The total global source of CHB3 is currently reported to be 300 Gg/year (Dvortsov et al. 1999).

5.2.5 5.2.5.1

Iodine compounds Methyl iodide

Methyl iodide (CH3 I) has several terrestrial sources that together are believed to comprise up to 30% of the total budget (Bell et al. 2002). Estimates of annual terrestrial emissions of CH3 I are between 1.4 × 108 mol and 4.1 × 108 mol (Redeker et al. 2000) from rice paddies, 5×107 mol from natural wetland (Dimmer et al. 2001) and 6×107 mol from biomass burning (Bell et al. 2002). Keppler et al. (2000, 2001, 2005) proposed an abiotic route for alkyl halide production in soils and sediments from halide ion alkylation during the oxidation of organic matter by an electron acceptor such as Fe(III). The authors proposed that production of C1 –C4 alkyl iodides from soils containing Fe(III) and iodide could be signiﬁcant globally, although their data did not allow for an estimate of emission. Terrestrial sources for the di- and tri-iodated compounds have not been identiﬁed. Manley and delaCuesta (1997) calculated a mean CH3 I production rate from 15 species of marine phytoplankton of 8×106 mol/year, similar to estimates from seaweed release. Neither source is signiﬁcant compared to the total estimated global CH3 I emission strength of 0.9–2.5×109 mol/year (Moore and Groszko 1999), suggesting missing sources. It should be noted, however, that laboratory incubations may miss potentially important ocean interactions, leading to underestimation of trace gas release rates. For example, work by Hughes (PhD Thesis, Biogenic iodocarbon production in the sea, 2004) has found that pelagic iodinated VHOC (Volatile Halogenated Organic Compound) production is signiﬁcantly enhanced in the presence of communities of microorganisms associated with particles formed during the plankton death and decay process such as marine snow, diatom mucilage aggregations and phytodetritus. Laboratory evidence suggests that a wide variety of both marine and terrestrial bacteria are capable of CH3 I production and of methylating environmental levels of I− ; whether these provide important global sources in their own right is yet to be determined. A further marine source of CH3 I is via sunlit irradiation of seawater (Moore and Zaﬁriou 1994). The suggested mechanism is photochemical generation of methyl and iodine radicals from,

Halogenated Volatile Organic Compounds

187

respectively, dissolved organic matter (DOM) and iodide present in the surface waters (Moore and Zaﬁriou 1994). Most DOM contains components that do not absorb light in the solar region; thus, they are unable to participate in the photochemical reactions. Components that absorb in the solar region include fatty acids, lipids, pigment systems and amino acids. Recent measurements in the tropical Atlantic have conﬁrmed that the photochemical source of CH3 I is abiotic, and proposed that it can support at least half of the average sea-to-air ﬂux of 23 nmol/m2 day (Richter and Wallace 2004).

5.3

Atmospheric concentrations: trends and distribution

The previous section described in detail the processes that emit halogenated VOCs into the atmosphere. Once emitted, these compounds mix throughout the atmosphere. In general, it takes approximately 1 month for a gas to mix throughout a hemisphere and 1 year to mix between hemispheres. Vertical mixing throughout the planetary boundary layer (surface to 1 km) takes about 1 h, and 1 month throughout the troposphere (10 km), and if the atmospheric lifetime of a compound is long enough it takes an estimated 5 years for tropospheric/stratospheric (10–30 km) transport to occur (Engel et al. 2002). Whilst the halogenated VOCs are mixing they also under go a myriad of complex transformations and chemical reaction in what has been described as an ‘uncontrolled and moving reaction vessel’ (Graedel and Crutzen 1993). In order to try and understand the processes that govern the accumulation and distribution of these compounds in the atmosphere, it is necessary to make physical measurements at various points around the globe. Around the world, there exist a number of global networks that make measurements of halogenated VOCs in surface air. The principal groups involved in this monitoring effort are: 1.

Atmospheric Lifetime Experiment/Global Atmospheric Gases Experiment/Advanced GAGE – ALE/GAGE/AGAGE 2. National Oceanic and Atmospheric Administration/Global Monitoring Division – NOAA/GMD 3. System for Observation of Halogenated Greenhouse Gases in Europe – (SOGE) 4. University of California at Irvine – (UCI); and as atmospheric columns 5. Network for Detection of Stratospheric Change – (NDSC). The scientiﬁc objectives of these measurements are several in number and of considerable importance in furthering our understanding of a number of important global chemical and climatic phenomena. These objectives can be summarised as follows: (a)

to determine optimally from observations, the rates of emission and/or chemical destruction (i.e. atmospheric lifetimes) of halogenated VOCs; (b) to document accurately the global distributions and temporal behaviours over the globe; (c) to determine optimally the average concentrations and trends of OH radicals in the troposphere by determining the rate of destruction of atmospheric CH3 CCl3 , HFCs and HCFCs from their continuous measurements of their concentration together with industrial estimates of their emissions (dealt with later in the chapter);

188

(d)

(e)

Volatile Organic Compounds in the Atmosphere

to determine optimally, using high frequency trace gas data (and theoretical estimates of their rates of destruction), the magnitude and distribution by region of the surface sources of these gases; to provide an accurate data base on the rates of accumulation of gases over the globe which can be used to test synoptic-, regional-, and global-scale circulation predicted by three-dimensional (3-D) models.

Tropospheric concentrations and their growth rates are presented in WMO (2003) (Montzka and Fraser 2003); a detailed discussion of the observed trends can also be found in the WMO report. Here, the same data are listed in Table 5.3, for year 2000. In addition, Figure 5.1 presents concentrations for selected species from the period 1996–2003 based on NOAA-GMD and AGAGE data (Montzka 2003; Simmonds et al. 2004, private communications). Many of the species for which data are tabulated here and in WMO (2003) are regulated in the Montreal Protocol. Two of the most abundant CFCs, CFC-11 and CFC113, have been decreasing in the atmosphere since around 1996. Although their emissions have decreased very substantially in response to the Montreal Protocol, their long lifetimes of around 45 and 85 years, respectively, mean that their sinks can reduce their levels only about 2% and 1%/year, respectively. The other major CFC, CFC-12 has ﬁnally reached a plateau in its atmospheric levels and is beginning to decline. The concentrations of the less abundant CFC-114 and CFC-115 (with lifetimes of 300 and 1 700 year, respectively) are relatively stable. Table 5.3 Mixing ratios, growth rates and atmospheric lifetimes for selected CFCs, Halons, HCHFs, HFCs and PFCs. Data are taken from WMO (2003). Included are also global emissions derived from the given mixing ratios and trends (see text for explanation). Data are given for year 2000, with a few exceptions Species

Tropospheric abundance (ppt)

Growth rate (ppt/year)

Notes

Year

CFCs CFC-12

CCl2 F2

542.9 534.5 534.0 535.7

2.3 1.9 1.8 1.9

AGAGE, in situ GMD, in situ GMD, ﬂasks UCI, ﬂasks

2000 2000 2000 2000

CFC-11

CCl3 F

260.5 263.2 262.6 261.0

−1.1 −2 −1.5 −1

AGAGE, in situ GMD, in situ GMD, ﬂasks UCI, ﬂasks

2000 2000 2000 2000

CFC-113

CCl2 FCClF2

82.0 82.1 81.1

−0.35 −0.32 −0.49

AGAGE, in situ GMD, ﬂasks UCI, ﬂasks

2000 2000 2000

CFC-114

CClF2 CClF2

16.5 17.2

0 −0.1

UEA, SH, ﬂasks AGAGE, in situ

1996 2000

CFC-115

CClF2 CF3

UEA, SH, ﬂasks AGAGE, in situ

1996 2000

7.5 8.1

0.16

Halogenated Volatile Organic Compounds

189

Table 5.3 (Continued) Species

Tropospheric abundance (ppt)

Growth rate (ppt/year)

Notes

Year

Halons Halon 1211

CBrClF2

4.1 4 3.9 4.4

0.13 0.10 0.12 0.13

AGAGE, in situ GMD, ﬂasks UCI, ﬂasks UEA, SH, ﬂasks

2000 2000 2000 2000

Halon 1301

CBrF3

3 2.6 2.3

0.06 0.06 0.05

AGAGE, in situ GMD, ﬂasks UEA, SH, ﬂasks

2000 2000 2000

Chlorocarbons Carbon tetrachloride

CCl4

96.1 99.6 99.2

−0.94 −0.95 −1.03

AGAGE, in situ GMD, in situ UCI, ﬂasks

2000 2000 2000

Methylchloroform

CH3 CCl3

45.4 46.4 45.7 47.6

−8.7 −10.2 −9.1 −13

AGAGE, in situ GMD, in situ GMD, ﬂasks UCI, ﬂasks

2000 2000 2000 2000

HCFCs HCFC-22

CHClF2

143.2 141.9

5.4 5.1

AGAGE, in situ GMD, ﬂasks

2000 2000

HCFC-141b

CH3 CCl2 F

13.0 12.7

1.8 1.7

AGAGE, in situ GMD, ﬂasks

2000 2000

HCFC-142b

CH3 CClF2

12.5 11.7

1.1 1

AGAGE, in situ GMD, ﬂasks

2000 2000

HCFC-123

CHCl2 CF3

0.03

0

UEA, SH, ﬂasks

1996

HCFC-124

CHClFCF3

1.34

0.35

AGAGE, in situ

2000

HFCs HFC-23 HFC-125 HFC-134a

CHF3 CHF2 CF3 CH2 FCF3

15.5 2.3 25.7 25.5

0.9 0.45 4.2 3.2

UEA, SH, ﬂasks AGAGE, in situ AGAGE, in situ GMD, ﬂasks

2000 2000 2000 2000

HFC-152a

CH3 CHF2

2.4

0.4

AGAGE, in situ

2000

PFCs CF4 C2 F6 C3 F8

FC-14 FC-116 FC-218

— 0.1 — 0.02

MPAE, NH, ﬂasks UEA, SH, ﬂasks Culbertson et al. (2004) UEA, SH, ﬂasks

1998 1996 1997 2000

UEA = University of East Anglia. MPAE = Max Plank Institute for Aeronomy.

76 2.9 0.2 0.22

Volatile Organic Compounds in the Atmosphere

CCl4

105

100

95

90 1990 1992 1994 1996 1998 2000 2002 2004 180

HCFC-22

160 140 120 100 80 1990 1992 1994 1996 1998 2000 2002 20

HFC-23

15

10

5 1990 1992 1994 1996 1998 2000 2002 3

Tropospheric mole fraction (ppt)

110

HFC-152a

2.5 2 1.5 1 0.5 0 1990 1992 1994 1996 1998 2000 2002 Year

500 480 460 1990 1992 1994 1996 1998 2000 2002 5

Halon-1211

4.5

4

3.5

3 1990 1992 1994 1996 1998 2000 2002 20

10

5

0 1990 1992 1994 1996 1998 2000 2002

75 70 65 60 1990 1992 1994 1996 1998 2000 2002 2004 Halon-1301 3.0 2.8 2.6 2.4 2.2 1990 1992 1994 1996 1998 2000 2002 20

3 2 1 0 1990 1992 1994 1996 1998 2000 2002 SF6

5.5 5 4.5 4 3.5 1990 1992 1994 1996 1998 2000 2002 Year

HCFC-142b

15

10

5

0 1990 1992 1994 1996 1998 2000 2002

HFC-125

4

6

80

HCFC-141b

15

5

Tropospheric mole fraction (ppt)

250 1990 1992 1994 1996 1998 2000 2002

520

CFC-113

85

Tropospheric mole fraction (ppt)

255

540

90

Tropospheric mole fraction (ppt)

260

CFC-12

30

Tropospheric mole fraction (ppt)

265

560

Tropospheric mole fraction (ppt)

270

Tropospheric mole fraction (ppt)

275

Tropospheric mole fraction (ppt)

CFC-11

Tropospheric mole fraction (ppt)

280

Tropospheric mole fraction (ppt)

Tropospheric mole fraction (ppt)

Tropospheric mole fraction (ppt)

Tropospheric mole fraction (ppt)

Tropospheric mole fraction (ppt)

Tropospheric mole fraction (ppt)

190

HFC-134a

25 20 15 10 5 0 1990 1992 1994 1996 1998 2000 2002 140

MCF

120 100

CFC = chlorofluorocarbon

HFC = hydrofluorocarbon

HCFC = hydrochlorofluorocarbon

MCF = methyl chloroform

80 60 40 20 0 1990 1992 1994 1996 1998 2000 2002 Year

Figure 5.1 Global annually averaged tropospheric mole fractions. In situ abundances from AGAGE and GMD as well as ﬂask samples from GMD are included. The AGAGE abundances are global lower troposphere averages processed through their 12-box model (Prinn et al. 2000; updates provided by D. Cunnold). The GMD abundances are area weighted global means for the lower troposphere (from http://www.gmd.noaa.gov/hats/index.html), estimated as 12-month averages centred around 1 January

Halogenated Volatile Organic Compounds

191

The two most abundant halons, halon-1211 and halon-1301, are still increasing in 2005, but at a reduced rate; this is despite a ban on their production and sales in developed nations in 1994. Halon scenarios from past WMO Scientiﬁc Assessments of Ozone Depletion reports are roughly consistent with current observed mixing ratios and trends. However, large uncertainties regarding emission from banked (stored) halons make accurate projections difﬁcult. The concentration of CCl4 reached a maximum of 104 ppt in 1989–1990, and has declined since then at a rate of 1 ppt/year. Large reductions in emissions of CH3 CCl3 , and its relatively short lifetime (5 years), have lead to a rapid decline in its atmospheric abundance. Global emissions have decreased from about 720 to 20 Gg and the atmospheric concentration has declined from 140 to 150 ppt in the 1990s to 25 ppt in 2003 (Reimann et al. 2005). Observation of CH3 CCl3 has proved to be an important measurement to make, because it has been used historically to indirectly determine OH concentrations. The indirect measurement is important because OH is the dominant oxidising chemical in the atmosphere, destroying most air pollutants and many gases involved in ozone depletion and global warming. Global-scale direct measurements of OH are not possible because of its extremely short lifetime (∼1 s). Long-term high-frequency measurements of CH3 CCl3 are combined with accurate estimates of emission and the knowledge that reaction with OH is the major removal mechanism for CH3 CCl3 (minor removal mechanisms such as photochemical loss and loss to oceans are accounted for) to determine regional- to global-scale concentrations and trends of OH. The three most abundant HCFCs, HCFC-22, HCFC-141b and HCFC-142b, are still increasing signiﬁcantly in 2005 (between 4 and 8%/year), but the rates of increase are slowing down somewhat. HFC concentrations are currently signiﬁcantly lower than for the most abundant CFCs, but are increasing rapidly. HFC-134a and HFC-23 are the most abundant components, the former growing most rapidly, with a growth rate of about 20%/year due to its use as a replacement for some CFC refrigerants. Observations of PFCs are sparser and have a relatively short emission history; as such growth rates are not reported. CF4 is by far the most abundant species in this group. Further, the atmospheric histories of a group of halocarbons have been reconstructed from analyses of air trapped in ﬁrn (snow above glaciers) (Butler et al. 1999; Sturges et al. 2001; Sturrock et al. 2002). The results show that concentrations at the beginning of the twentieth century of CFCs, halons and HCFCs were generally less than 2% of the current concentrations. The compounds mentioned so far are anthropogenic in origin and have relatively long atmospheric lifetimes and as such are well mixed in the atmosphere. It is much more difﬁcult to quantify the effect in global terms of short-lived species in the atmosphere especially ones

Figure 5.1 (contd.) each year (e.g. 2 002.0 = 1 January 2002). GMD data for 2003 are preliminary and will be subject to recalibration. The UCI data are based on samples at latitudes between 71ºN and 47ºS. The UEA data (HFC-23) are from Cape Grim, Tasmania, and are represented here by a ﬁt to a series of Legendre polynomials (up to third degree). Reproduced with permission from IPCC/TEAP, IPCC/TEAP Special report on safeguarding the ozone layer and the global climate system: Issues related to hydroﬂuorocarbons and perﬂuorocarbons. Prepared by Working Groups I and III of the Intergovernmental Panel on Climate Change, and the Technical and Economic Assessment Panel. Cambridge University Press, UK and New York, 2005.

192

Volatile Organic Compounds in the Atmosphere

that have biogenic or mixed sources. Short-lived gases are not transported far from their sources and hence concentrations will tend to follow the regions of sources and sinks. Even though the atmospheric lifetimes of some short lived gases vary considerably, measurements of these compounds (CH3 Br, CH3 Cl, CH3 I, etc.) at speciﬁc sites tend to provide atmospheric information on processes and transformation local to the site of measurement. For example, transformation of bromine and iodine species in Polar Regions can result in complete tropospheric ozone depletion (see next section for more detail). Halocarbons with large sources from industrialised countries have higher concentrations in the middle and high latitudes of the NH and lowest concentrations in the SH. This combined with a 1-year transport time from northern to SHs results in an interhemispheric gradient. Changes in atmospheric bromine (Br) indicate that the global tropospheric burden of Br from the sum of halons and short-lived species such as CH3 Br peaked in 1998 and has since declined by nearly 5%. These changes are driven primarily by a decrease of CH3 Br since 1998 which is about a factor of two larger than expected given reported declines in industrial production, a result that may suggest revisions to our understanding of the global atmospheric budget for this gas (Montzka et al. 2003). CH3 Cl has predominantly natural sources, and as such is not restricted by the Montreal Protocol. However, it is one of the most abundant chlorinecontaining gases in the remote atmosphere and trends for this gas could affect the trend in atmospheric chlorine. Measurements suggest no systematic trend between 1998 and 2005; however, mixing ratios of this gas show signiﬁcant inter-annual variability.

5.4

Sinks of halogenated VOCs

The atmospheric burden, or the total amount of halogenated VOCs in the atmosphere is determined by the competing processes of emission (rate of release into the atmosphere), and rate of removal from the atmosphere. While emissions (discussed in the previous section) are primarily a result of human activities, the rates of removal are largely determined by complex natural atmospheric processes that vary both spatially and temporally, and may be quite different depending on the gas of interest. The rate of removal of a gas is directly linked to its atmospheric lifetime. The concept of atmospheric lifetime of halogenated VOCs in the atmosphere is key to understanding the destruction patterns in the atmosphere and hence their atmospheric build-up and distribution. For example, rapid removal implies a short atmospheric lifetime, while a longer lifetime (slow removal) means that a given emission rate will result in a larger atmospheric burden. It must be remembered that the atmospheric lifetime of a compound is not generally constant. This is because emission and loss processes are not uniform in space or time. Atmospheric sinks for halogenated VOCs include 1.

2. 3. 4. 5.

photochemical losses within the troposphere and stratosphere, predominantly reaction with OH in the troposphere and photolysis by ultraviolet (UV) radiation in the stratosphere; reaction with electronically excited atomic oxygen (O(1 D)) and atomic chlorine (Cl); uptake in oceanic surface waters through chemical and biological degradation processes; biological degradation in soils; surface reactions on minerals.

Halogenated Volatile Organic Compounds

PFCs SFs

F

Degradation products

OH O'D

UV

193

Mesosphere Cl

Br

CFCs Halons

Stratosphere light, H2O

O3 light

TFA

NO

NO2

HCl

Troposphere

O'D Degradation Products

HO2 H2O OH

CFCs Halons

Degradation products

Degradation products

OH

HCFCs HFCs

PFCs

HFEs

SFs

CH3CCl2

CO2

HCs

NH3 Aerosols

Removal

Land

Ocean

Land

CFC = chlorofluorocarbon PFC = perfluorocarbon HCFC = hydrochlorofluorocarbon HFE = hydrofluoroether HFC = hydrofluorocarbon

HC = hydrocarbon

Figure 5.2 Schematic of the degradation pathways of CFCs, HCFCs, HFCs, PFCs, HFEs, HCs and other replacements species in the various atmospheric reservoirs. HFE = Hydroﬂuoroether. Reproduced with permission from IPCC/TEAP, IPCC/TEAP Special report on safeguarding the ozone layer and the global climate system: Issues related to hydroﬂuorocarbons and perﬂuorocarbons. Prepared by Working Groups I and III of the Intergovernmental Panel on Climate Change, and the Technical and Economic Assessment Panel. Cambridge University Press, UK and New York, 2005.

A schematic of the various removal processes affecting the concentration of the considered species once emitted in the atmosphere is provided in Figure 5.2. Individual removalprocesses have various impacts on different gases in different regions of the atmosphere. The dominant process will control the local concentration of a gas. For example, reaction with OH is the dominant removal process for many hydrogenated halocarbons in both the troposphere and stratosphere. For those same gases, reactions with O(1 D) and Cl play a large role only in the stratosphere. The main degradation pathways of a number of halogenated VOCs are shown in Figure 5.2. The determination of the spatial distribution of trace gas sink strengths are of interest because it is relevant to both the atmospheric lifetimes and the impact on air quality of these substances. The lifetimes of many halogenated VOCs have been reassessed by the IPCC (Prather et al. 2001) and by the WMO Scientiﬁc Assessment of Ozone Depletion (Montzka and Fraser 2003).

5.4.1

Loss processes for the CFCs and halons

The loss processes for these compounds are deﬁned by their chemical inertness. Their long atmospheric lifetimes result in their transport to the stratosphere where photochemical

Volatile Organic Compounds in the Atmosphere

Altitude (km)

194

40

40

30

30

20

20

10

10 CF2Cl2 0 0.1

CFCl3 1

Normal mixing ratio

0 0.001

0.01

0.1

1

Normal mixing ratio

Figure 5.3 Vertical distributions of CFC-12, and CFC-11, depicting various dynamical disturbances observed on 27 March 1987 (open circle), and 16 April 1994 (ﬁlled square) over Hyderabad. These proﬁles belong to the QBO-east-phase above about 40 hPa. The MPIC model simulated proﬁle (continuous line) and the observed proﬁle on 9 April 1990 over Hyderabad (open triangle, QBO-west-phase above 40 hPa) are shown for a comparison. The stratospheric proﬁles of CFC-12 are known to be controlled by both chemistry and dynamics, and that of CFC-11 are mainly controlled by chemistry (attains photochemical equilibrium). QBO = Quasi-biennial zonal wind oscillation, MPIC = Max Plank Institute for Chemistry. Adapted with permission from Patra, P.K., Lal, S., et al. (2000) Chlorine partitioning in the stratosphere based on in situ measurements. Tellus Series B-Chemical and Physical Meteorology, 52 (3): 934–46. Blackwell Publishing, 2000.

destruction leads to the release of free halogen atoms (Cl or Br). Figure 5.3 illustrates data for a CFC-11 and CFC-12, how the concentration remains virtually constant throughout the troposphere, but drops rapidly once past the tropopause as UV radiation causes photolysis and release of active chlorine (Patra et al. 2000). The chemical composition of these compounds will determine the altitude at which photolysis occurs. More Cl or Br atoms in the molecule cause a shift in the UV absorption spectrum to longer wavelengths, allowing photolysis to occur at lower altitudes where their absolute impact is greater (Wayne 1995). The catalytic activity of Br relative to that of Cl is calculated to be of the order of 100 at an altitude of 20 km (WMO 2003). In-depth coverage of the processes involved in stratospheric ozone depletion, especially ozone depletion in the polar stratosphere are beyond the scope of this chapter; however, such topics are widely covered in the literature (Brasseur et al. 1999; Cicerone et al. 1974; Farman and Gardiner 1987; Molina and Rowland 1974b; Wayne 1995). A brief explanation of the major loss processes follows: CFCl3 + hν → CFCl2 + Cl CFCl2 + O2 + M → CFCl2 O2 + M CFCl2 O2 + NO → CFCl2 O + NO2 CFCl2 O + M → COFCl + Cl + M COFCl + hν → FCO + Cl In a similar fashion, halons are photolysed to form free Br atoms; however, because the presence of bromine in the molecule shifts the absorption spectra to longer wavelengths, photolysis of bromine compounds tends to occur faster than chlorine compounds and at

Halogenated Volatile Organic Compounds

195

lower altitudes: CF3 Br + hν → CF3 + Br Followed by reaction of Cl and Br atoms with O3 , leading to the formation of ClO and BrO: Cl + O3 → ClO + O2

Br + O3 → BrO + O2

ClO + O → Cl + O2

BrO + O → Br + O2

O + O3 → 2O2

O + O3 → 2O2

ClO and BrO can then undergo a number of null cycles to regenerate Cl and O3 . This type of regeneration reaction has no effect on the ozone layer. However, other catalytic cycles exist that lead to the loss of odd oxygen and do have a profound effect on the ozone layer. The major ozone depleting Cl cycle (if ClO alone is present as catalyst): Cycle 1: Cl + O3 → ClO + O2 ClO + O → Cl + O2 O + O3 → 2O2 Reactions of ClO with other HOx , NOx and ClOx species also lead to O3 destruction: Cycle 2: Cl + O3 → ClO + O2 ClO + NO → Cl + NO2 NO2 + O → NO + O2 O + O3 → 2O2 Cycle 3: Cl + O3 → ClO + O2 OH + O3 → HO2 + O2 ClO + HO2 → HOCl + O2 HOCl + hν → OH + Cl 2O3 → 3O2 The magnitude of ozone depletion is reduced by conversion of Cl and ClO into reservoir species, the main ones are: ClONO2 and HCl. Once converted active Cl and ClO are not involved in ozone depletion until the reservoir species reacts with OH or by photolysis to return the active species. The atmospheric lifetime of reservoir species is again dependent upon altitude. For example the lifetime of ClONO2 is approximately 6 h in the midlatitude

196

Volatile Organic Compounds in the Atmosphere

lower stratosphere (below 30 km) and decreases to about 1 h at 40 km as the intensity of UV light increases. Cycle 4: Cl + O3 → ClO + O2 NO + O3 → NO2 + O2 ClO + NO2 + M → ClONO2 + M ClONO2 + hν → Cl + NO3 NO3 + hν → NO + O2 Cycle 5: Cl + CH4 → CH3 + HCl Cl + HO2 → HCl + O2 Cl + H2 → HCl + H Cl + H2 O2 → HCl + HO2 The loss rates for these cycles are altitude dependent, above 20 km Cycle 1 is dominant, at lower altitudes Cycles 2–5 become dominant. The chemistry of bromine containing compounds in the stratosphere is interesting because of a number of reasons. Firstly, they have a higher potential to destroy ozone (by a factor of 45–60) than the equivalent amount of chlorine due to the fact that the reservoir species HBr and BrONO2 are rapidly photolysed. Secondly, because of a chlorine-bromine cycle involving a coupling reaction between BrO and ClO: Cycle 6: BrO + ClO → Br + Cl + O2 Br + O3 → BrO + O2 Cl + O3 → ClO + O2 2O3 → 3O2 In addition to the above reaction there are two other reactions that can occur to regenerate halogen atoms and lead to catalytic destruction of odd oxygen: BrO + ClO → Br + OClO OClO + M → Cl + O2 BrO + ClO → BrCl + O2 BrCl + hν → Br + Cl

Halogenated Volatile Organic Compounds

197

CX3CXYH (year) H2O

OH

CX3CXY (μs) O2

CX3C(O)X (Y=H) HO2

+

NO2 CX3CXYO2 (min)

CX3CXYOOH (days)

CX3CXYOONO2 (h) Δ

OH NO2

NO h

CX3CXYO (μs) Δ

Δ O2

CX3 (μs) + C(O)XY O2 CX3O2

CX3C(O)X + Cl

CX3C(O)X (weeks)

(Y= Cl)

+ HO2 (Y=H)

Figure 5.4 Generalised scheme for the atmospheric oxidation of a halogenated organic compound, CX3CXYH (X, Y = H, Cl, or F). Transient radical intermediates are enclosed in ellipses; products with less transitory existence are given in the boxes.

5.4.2

Loss processes for the HCFCs and HFCs

These compounds contain at least one carbon–hydrogen bond, and react in the troposphere with the hydroxyl radical (OH) by hydrogen abstraction, to produce haloalkyl radicals. For example, taking the generalised structure CX3 CXYH, where X, Y are H, Cl or F: OH + CX 3 CXYH → H2 O + CX 3 CXY The haloalkyl radicals form the corresponding peroxyl radicals (CX3 CXYOO) and the subsequent reactions are illustrated in Figure 5.4. The oxidation occurs predominantly in the troposphere where scavenging of water-soluble, oxidation products prevents most of the reactive halogen carried by these trace gases from being released in the stratosphere. Lifetimes with respect to oxidation by OH for most HCFCs are longer than several years (see Table 5.2), thereby permitting small but signiﬁcant amounts of these compounds to reach the stratosphere (Lee et al. 1995; Zander et al. 1996). Once in the stratosphere, the propensity for a compound to degrade via photodissociation or chemical oxidation at stratospheric temperatures determines its contribution to the reactive halogen burden of the stratosphere. Oxidation by OH in the troposphere is the predominant loss process for most HCFCs, and much less of the chlorine they contain is released to the stratosphere as reactive halogen atoms when compared to similar emissions of the fully halogenated CFCs.

198

Volatile Organic Compounds in the Atmosphere

Stratosphere CFCs

inorganic X free troposphere

CH2Br + OH → ... → BrO

CFCs

organic and inorganic X

O3 destruction via XO

locally in polluted environments: O3 production

all particles including cloud droplets S(IV) oxidation by HOX Cl, Br, I

boundary layer

h Cl + HC → HCI + R

new particles

IO2

I, (Br, Cl)

DMSO uptake of I

catalytic O2 depletion events BrO + BrO organic halogens DMS ice

XO + ...

Cl

recycling

ice

polar regions

Cl2 HOCI HCI

CH2X

release of Br, Cl

HX

sea sat aerosol cooling towers

ocean

OH

swimming industry biomass pools burning

volcanoes

continents

Figure 5.5 Schematic depiction of the most important halogen-related processes in the troposphere.

5.4.3

Loss processes for chlorine and bromine compounds

These classes of compounds will be considered together because they are both partially oxidised and undergo reaction in the troposphere via reaction with OH. However, whereas chlorine chemistry is the key to understanding stratospheric ozone depletion, bromine chemistry is the key to tropospheric ozone depletion and has been implicated in the process of ‘bromine explosions’ leading to ODEs mainly over the Arctic Ocean during polar sunrise. The phenomenon of springtime tropospheric ozone depletion is not of global signiﬁcance but it is puzzling and to date not clearly understood (Singh and Fabian 1999). The most important Br + Cl reaction cycles are depicted schematically in Figure 5.5. Reaction of CH3 Cl with OH results in a number of partially oxidised species: Cycle 7: OH + CH3 Cl → CH2 Cl + H2 O CH2 Cl + O2 + M → CH2 ClO2 + M CH2 ClO2 + NO → CH2 ClO + NO2 CH2 ClO + O2 → HCOCl + HO2 These partially oxidised species are then either removed by rainout, or undergo further oxidation leading to free Cl-atom production: HCOCl + hν → HCO + Cl

Halogenated Volatile Organic Compounds

199

The main sink for chlorine radicals is the reaction with hydrocarbons and especially with CH4 : Cl + RH → HCl + R As previously mentioned, it is the self-reaction of BrO when present at high levels (e.g. in the polar regions) that dominates tropospheric halocarbon reactions. It was discovered that remarkably low levels (even down to zero) of surface ozone measured in the Arctic spring were anti-correlated with high levels of ‘ﬁlterable’ bromine (Barrie et al. 1988): Cycle 8: 2(Br + O3 → BrO + O2 ) BrO + BrO → 2Br + O2 BrO + BrO → Br2 + O2 Br2 + hν → 2Br 2O3 → 3O2 Although it is clear that the BrO+BrO catalytic reaction cycle is very efﬁcient, it is not yet fully understood where the source is of the reactive bromine species that leads to what had been termed ‘bromine explosions’ (Platt and Honninger 2003). Many reaction paths have been proposed involving chemical reactions taking place in or on airborne sea salt, the snow pack, or more likely on the Arctic ocean sea ice to produce photolabile halogen compounds, especially BrO radicals, which rapidly destroy surface ozone. A great deal of research effort has been expended in this area over the past decade, it is clear, however, that there are still many unknowns including possible climate change feedbacks on source strength of sea salt aerosol or biogenic halogen precursors.

5.4.4

Loss processes for iodine compounds

Over the last few years, there has been increasing evidence that iodine has an important inﬂuence on tropospheric chemistry, ﬁrstly by participating in catalytic ozone destruction cycles and associated perturbations of HOx and NOx speciation (Bloss et al. 2005; McFiggans et al. 2000), and secondly by forming new aerosol particles in the coastal boundary layer, which potentially act as cloud condensation nuclei (CCN) (O’Dowd et al. 2002). These processes are unique to the marine boundary layer. Until recently, CH3 I, which was ﬁrst detected in 1973 (Lovelock and Maggs 1973), was believed to be the major source of atmospheric iodine. However, due to their rapid photolysis rates, it has been suggested that CH2 I2 , CH2 ClI and CH2 BrI may be more important organic iodine atom precursors (Carpenter et al. 1999). Observations have shown that molecular iodine (I2 ) is a signiﬁcant source of coastal atomic iodine (Saiz-Lopez and Plane 2004). Currently, the iodocarbons are still considered to be the dominant vectors for the sea-to-air transfer of iodine, although the importance of I2 in this transfer is under review. Figure 5.6 uses a simple schematic diagram to illustrate iodine chemistry in the marine boundary layer. Photolysis of organoiodines occurs rapidly, with lifetimes ranging from several days (CH3 I, C2 H5 I, C3 H7 I: Rattigan et al. 1997; Roehl et al. 1997), several hours (CH2 ICl:

200

Volatile Organic Compounds in the Atmosphere

NO2, M

INO2 HOI

CH3I

HO2

CH2I2

O3

CH2CII I2

I

NO2, M

HOI +

IONO2

NO OH

Ocean

IO

OH

HO2

HI

H – – Cl , Br , l

Aerosol

–

– IO3

HOCI, HOBr, HOI

IO, BrO

IX

I–

OIO I2O2

Atmosphere Figure 5.6 Simpliﬁed schematic of iodine chemistry in the marine boundary layer.

Rattigan et al. 1997; Roehl et al. 1997), an hour or less (CH2 IBr: Mossinger et al. 1998) to about 5 min at midday (CH2 I2 : Mossinger et al. 1998; Roehl et al. 1997). Although the lifetimes of the polyiodomethanes are controlled almost entirely by photodissociation, OHand Cl-initiated attack could account for 10–20% of the removal of CH3 I and compete with photolysis for removal of the propyl iodides (Cotter et al. 2001). The iodine atoms released react with ozone to yield iodine oxide (IO), which reacts with HO2 and NO2 to yield the intermediate reservoir species HOI and IONO2 . One fate for these intermediates is photolysis, which regenerates I atoms without concomitant O atom formation and can therefore lead to catalytic O3 loss: I + O3 → IO + O2 IO + HO2 → HOI + O2 HOI + hν → OH + I O3 + HO2 → OH + 2O2 The HOI cycle has been suggested as the dominant iodine-initiated O3 loss cycle at NOx levels below about 500 pptv (Stutz et al. 1999). For typical conditions at Mace Head, Ireland, Stutz et al. calculated 0.3 ppb/h O3 loss at 100 pptv of NOx from the HOI cycle. Another major fate of the HOI and IONO2 intermediates is uptake onto marine aerosols, yielding aqueous HOI, which will react with available halide ions in the presence of H+ to form the interhalogen species IBr and ICl (Vogt et al. 1999). These are released into the gas phase and produce two halogen atoms on their photolysis. This, and equivalent reactions of HOBr yielding BrCl, Br2 and Cl2 , leads to the universal depletion of Br− and frequently of Cl− from sea salt aerosols relative to seawater, and to the release of reactive gaseous halogens from the extensive halide reservoir provided by sea salt aerosol (Ayers et al. 1999; Sander

Halogenated Volatile Organic Compounds

201

et al. 2003). The comparatively trace levels of iodide (∼0.1–0.3 μg (Wong 1991)) present in seawater, compared to signiﬁcant chloride and bromide concentrations, and the relative aqueous reaction rates of HOX (where X = I, Br, Cl) with Br− and Cl− , results in sea salt aerosols being a signiﬁcant source of reactive chlorine and bromine, but a much lower source of iodine. In fact, the net transfer of iodine is from the gas to the condensed phase, reﬂected by the factor of 100- to 1 000-fold enrichment of I in ﬁne-fraction marine aerosol by comparison to the I/Na ratio in seawater (Baker et al. 2000; Duce and Hoffman 1976). Irreversible accumulation of OIO and I2 O2 , formed via self-reactions and cross-reactions of IO and BrO, onto pre-existing aerosols has been assumed as the main cause of iodine enrichment of aerosol. Formation of OIO has also been implicated in new particle formation (O’Dowd et al. 2002). Self-reaction of OIO to form low-volatility iodine oxides such as I2 O4 (or [IO]+ [IO3 ]− ) has been proposed to lead to stable chain-like structures from further collisions with OIO and particle nucleation (Hoffmann et al. 2001). O’Dowd et al. (2002) included this mechanism in a simulation of the production of new particles from condensation of low-volatility iodine species on thermodynamically stable sulphate clusters. They concluded that at high iodine concentrations such particles may overcome the coagulation loss barrier, resulting in a higher probability of the new particles surviving and increasing the number and lifetime of condensation sites for production of CCN. There are still many uncertainties with the rates and mechanisms involved in iodine new particle formation. One issue is the photolysis lifetime for OIO. A high photochemical stability of OIO as found by (Cox et al. 1999) would reduce the IO concentration and subsequent ODP of iodine, with the effect being dependent upon the rates of OIO and I2 O2 uptake to aerosol. However, work by Ashworth et al. (2002) suggests a short photolysis lifetime for OIO, increasing the ODP but decreasing the potential for aerosol formation. Figure 5.6 shows a simpliﬁed schematic of iodine chemistry in the marine boundary layer. For a full review of iodine chemistry, see Carpenter (2003) and Saiz-Lopez and Plane (2004). To date (2007), no direct evidence of stratospheric ozone depletion by iodine compounds has been presented. Bosch et al. (2003) report upper limits of lower stratospheric IO (inferred lowest values: 0.10 ppt, 0.07 ppt, and 0.06 ppt at 20, 15, and 12.5 km, respectively) and OIO (inferred lowest values: 0.10 ppt, 0.06 ppt, and 0.04 ppt at 20, 15, and 12.5 km, respectively) inferred from balloon-borne solar occultation UV-visible spectroscopy. Solomon et al. (1994) speculate that iodine chemistry in combination with trends in anthropogenic chlorine and bromine may also be a factor in determining the widespread current depletion of lower stratospheric ozone by coupling the chemistry of iodine to that of bromine and chlorine: IO + ClO → I + Cl + O2 IO + BrO → I + Br + O2 The reaction of the free halogen atoms with ozone, could then lead to ozone depletion.

5.4.5

Loss processes for Perﬂuoro compounds

These compounds are extremely stable, and accordingly have extremely long atmospheric lifetimes throughout the troposphere and stratosphere (estimated to be 50 000 years for

202

Volatile Organic Compounds in the Atmosphere

CF4 ). The major sources of ﬂuorine to the stratosphere are the CFCs, which release ﬂuorine in the form of COFCl and COF2 . These carbonyl compounds are photolysed to form free ﬂuorine, which in turn is converted to the extremely stable reservoir species HF. The reported loss processes include reaction with H atoms, destruction by Lyman-α radiation at 121.6 nm, uptake by oceans and soils, processing through combustion systems and reaction with electrons and ions in the mesosphere and thermosphere.

5.4.6

Is the Montreal Protocol working? Some speciﬁc conclusions

Global emissions of ozone-depleting gases can be estimated from global measurements using inverse modelling methods. Some speciﬁc conclusions are as follows: 1.

2.

International compliance with the Montreal Protocol so far is resulting in CFC and chlorocarbon mole fractions comparable to target levels; so, the Protocol is working. Mole fractions of total chlorine contained in long-lived halocarbons (CFC-12, CFC-11, CH3 CCl3 , CCl4 , HCFC-22, CFC-113, CH3 Cl, CH2 Cl2 , CHCl3 , C2 Cl4 ) in the lower troposphere reached maximum values of about 3.6 ppb in 1993 and are beginning to slowly decrease in the global lower atmosphere. This is illustrated in Figure 5.7.

4.0

Hydrochlorofluorocarbons (HCFC-22, HCFC-141b, HCFC-142b)

3.5 Chlorinated solvents (CH3CCl3, CCl4)

Mole fraction (ppb)

3.0 2.5 2.0

Chloroflurocarbons (CFC-11, CFC-12, CFC-113, CFC-114, CFC-115)

1.5 1.0 0.5

Chloromethanes (CH3Cl, CHCl3, CH2Cl2)

0.0 1980

1985

1990 Year

1995

2000

Figure 5.7 Tropospheric chlorine loading contributions from 1978 to the end of 2001 based on the ﬂask and in situ surface measurements. The values shown are averages for the lower troposphere of the four semi-hemispheres. Results from a best-ﬁt 12-box model have been used to ﬁll in the missing data in each of the four semi-hemispheres (e.g. for the HCFCs in 0–30◦ N and 0–30◦ S).

Halogenated Volatile Organic Compounds

203

3.

The CFCs have atmospheric lifetimes consistent with destruction in the stratosphere being their principal removal mechanism. 4. Multi-annual variations in CFC and chlorocarbon emissions deduced from a global network of measurements are consistent approximately with variations estimated independently from industrial production and sales data where available; however, CFC-12 and CFC-113 show the greatest discrepancies. 5. The mole fractions of the HCFCs and HFCs, which are replacing the regulated halocarbons, are rising very rapidly in the atmosphere, but, with the exception of the much-longer-manufactured HCFC-22, they are not yet at levels sufﬁcient to contribute signiﬁcantly to atmospheric chlorine loading. These replacement species could in the future provide independent estimates of the global weighted-average OH concentration provided their industrial emissions are accurately documented. 6. In the future, analysis of pollution events measured using high-frequency, in situ measurements of CFCs and their replacements may enable emission estimates at the regional level, which, together with industrial end-use data, will be of sufﬁcient accuracy to be capable of identifying regional non-compliance with the Montreal Protocol.

5.4.7

Total chlorine and bromine in the atmosphere

The levels of total organic chorine (CCly ) and bromine (CBry ) from long-lived source gases in the troposphere are in decline. In mid-2000, CCly was ∼3.5 ppb or about 5% lower than the peak observed in 1992–1994 (see Figure 5.7). Also, by mid-2004, CBry was 0.6–0.9 ppt below the peak, with the mean decline from 1998 to 2004 being between −0.1 and −0.15 ppt Br/year. These decreases result mainly from the declines observed for CH3 CCl3 and CH3 Br, in addition to contributions from other halocarbons regulated under the Montreal Protocol and amendments. The decline observed for bromine from regulated gases is particularly signiﬁcant relative to changes in chlorine, considering that the efﬁciency for Br to deplete stratospheric ozone is 45–60 times greater than Cl. The time series of stratospheric chlorine and bromine are different from the tropospheric times series in many ways; they lag tropospheric time series due to the time taken for the halogen species to be transported to the stratosphere; in addition mixing leads to a distribution of transit times at any point in the stratosphere; ﬁnally photodissociation of organic chlorine and bromine to inorganic further complicates the issue. In an attempt to simplify the effect of chlorine and bromine species with respect to ozone depletion, two indices have been established. Effective equivalent chlorine (EECl), where the amount of chlorine and bromine measured in the lower atmosphere, is used to represent concentrations of ozone-depleting substances. It is a convenient parameter for measuring with a single number the overall potential human effect on stratospheric ozone. EECl is derived by considering the changing concentrations of halocarbon gases, measured at grounds based stations, which can affect the stratospheric ozone concentration. An index is then developed based on the ability of those gases to catalyse the destruction of ozone relative to the ability of chlorine to do so. The units of EECl are parts per trillion (ppt) by volume. EECl in tropospheric air declined from 1995 through 2000 at a mean rate of about 1.2%/year, or 24 ppt EECl/year. As of mid-2000, EECl was about 5% below the peak that was observed in 1992–1994. Whereas EECl reﬂects the time evolution of equivalent halogen only

204

Volatile Organic Compounds in the Atmosphere

in the troposphere, Effective equivalent stratospheric chlorine (EESC) provides a simple index that relates the time evolution of long-lived surface abundances of ODSs with the ozone destruction ability of stratospheric halogens that come from long-lived source gases. EESC is derived simply by adding a 3-year lag to EESC and has generally been used to relate prediction of ODS substances with future ozone depletion. To retain the simplicity of the EESC index, several assumptions must be made. These assumptions are discussed in detail in (WMO 2003).

5.5

Emission inventories

Many countries prepare national emissions inventories (NEI) with input from numerous state and local environmental agencies. These data are used for air dispersion modelling, regional strategy development, regulation setting; air toxics risk assessment, and tracking trends in emissions over time. In the United Kingdom, national atmospheric emissions inventory (NAEI) compiles estimates of emissions to the atmosphere from United Kingdom sources such as cars, trucks, power stations and industrial plants. These emissions are estimated in order to help ﬁnd ways of reducing the impact of human activities on the environment and our health. The emissions have been inventoried mainly by bottom-up approaches, which involve adding up emissions from various industrial sources, such as estimates of global halocarbons compiled by (AFEAS 2004) into production, sales into each end-use and the time schedule for atmospheric emission from each end-use (McCulloch et al. 1999a, 1999b, 2001, 2003). Uncertainties arise if the global production ﬁgures do not cover all manufacturing countries and if there are variations in the application of the end use categories. For some halocarbons, there are no global emission inventory estimates. An essential part of validating emission inventories is the comparison between emissions inventories derived using conventional approaches (bottom-up) and those from atmospheric measurements and models (top-down), methods that should give two completely independent results.

5.5.1

Bottom-up approach

The work to provide the conventionally derived emissions follows an approach that is sequential and is reﬁned by iteration. As a ﬁrst step, emissions functions are devised. The aim of this is to gain the most accurate transfer functions between production, consumption, material in use in equipment (the ‘banks’) and emissions, consistent with the databases and views on current and historic social and economic practices. These data are used, together with databases on production and consumption maintained by industry, regulatory bodies and publications to derive global and regional emissions of the target compounds. The emission functions are examined and revised regularly. The gridded databases, which are founded on estimates of emissions down to the country level and population distribution within each country, are brought up to date using the best estimates of current national emissions and population distributions. In order to reduce the uncertainty in the estimates, particularly in their distributions, emissions functions are subjected to a further iteration. Predictions of atmospheric concentrations arising from the calculated emissions are compared with atmospheric measurements.

Halogenated Volatile Organic Compounds

5.5.1.1

205

‘Enabling studies’ on emission functions and databases of activity (as production, sales or consumption)

The emission functions for halocarbon compounds such as: CFCs 11, 12 and 113; HCFCs 22, 141b and 142b, HFC-134a, CH3 CCl3 , C2 HCl3 , C2 Cl4 and CH2 Cl2 have been continually reappraised in the light of the ﬁt between calculated atmospheric concentrations and observations (McCulloch et al. 2003). Changes have been made; for example, in the case of the HCFCs, there is evidence of signiﬁcant changes in emission functions in recent years, as the using industries have managed better control of emissions. These compounds have complex emission functions that can be described in terms of an initial release followed by a somewhat smaller annual loss while the compound is being used in equipment and on its eventual disposal. Major changes to the equipment manufacturing operations have resulted in signiﬁcantly lower initial losses that are reﬂected in the most recent emission functions. Solvent emissions are relatively prompt. However, there is evidence that the reduction in emissions of CH3 CCl3 consequent on the Montreal Protocol has not occurred as fast as would be expected from such an emission function. Small current global emissions (approximately 4 Gg/year) would be consistent with a low percentage (about 5%) of CH3 CCl3 sales being to small- and medium-sized enterprises (SMEs) that held stocks far longer than the larger users (Prinn et al. 2005).

5.5.1.2

Calculations of regional emissions and further subdivision into geographically distributed (‘gridded’) emissions

In order to assist with back trajectory dispersion modelling, emissions can be distributed geographically. For example, for most of Europe emissions can be mapped onto a 1◦ latitude by 1◦ longitude grid using population density. Further subdivision into emissions from each EU member state can be accomplished partly with the data that these countries submitted to the UNFCCC and partly using national econometrics to piece together a homogeneous data set that is consistent with the total European data. The materials are emitted from use in society; in refrigeration, air-conditioning and plastic foam products (where the activities are related to human habitation and employment, particularly in SMEs), so justifying the choice of population as the general distributor. Furthermore, most countries are either sufﬁciently homogeneous or far enough away from the measuring sites, for the small anomalies inevitably introduced by using a simple single distribution function to be insubstantial. The gridded data are also reported in the repositories as maps of emissions. For example, the emissions of CFC-12 during the year 2000 are shown for Europe in Figure 5.8.

5.5.1.3

Reﬁning of these emissions data by comparison with other databases and atmospheric measurements

Comparisons of top-down to bottom-up emission estimates are numerous in the literature (O’Doherty et al. 2004; Palmer et al. 2003; Reimann et al. 2004). Some speciﬁc examples are: The Numerical Atmospheric Dispersion Modelling Environment (NAME) dispersion model, driven by 3-D synoptic meteorology from the Uniﬁed Model, has been used to determine the fraction of air arriving at Mace Head, Ireland, from different European regions over a 6-year period. These data, along with observations of pollutants at Mace Head and a best-ﬁt algorithm, have been used to derive emission estimates over Western Europe. The algorithm starts from randomly generated emission maps and iterates towards the best

206

Volatile Organic Compounds in the Atmosphere

CFC -12 Year 2000

<10−3%

<10−2% <3×10−2% <6×10−2% <10−1% <3×10−1%

<1%

Figure 5.8 Gridded distribution of CFC-12 emissions for Europe in year 2000. Percentages refer to the global total emission (134.1 Gg). Reproduced with permission from A. McCulloch, Final Report of the EU FP5 project “System for Observation of Halogenated Greenhouse Gases in Europe (SOGE)”. Project EVK2-2000-00674. This image appears in full colour in the plate section that follows page 268 as Plate 4.

solution. Using an idealised case study, it has been shown to be effective at distinguishing between distinct source regions. The technique has been applied to two ozone-depleting gases, CFC-11 and CFC-12, and two greenhouse gases, methane and nitrous oxide. The emissions derived compare favourably with existing inventories. The technique is able to provide information regarding the emission distribution across Europe and to estimate area and country contributions; information that for some species is not readily available by other means. It is a different methodology to those currently used and so is a useful tool in verifying existing inventories (Manning et al. 2003). Yokouchi et al. (2005) used aircraft monitoring over Sagami Bay, Japan, to estimate the emission ratios of halogenated VOCs from Japan. The enhanced concentrations in the boundary layer of air masses having travelled over the Japanese mainland were used for the calculation under the assumption that the air masses over Sagami Bay represented average emission ratios for anthropogenic halocarbons on a countrywide basis. Given their emission ratios, a single compound with a credible emission rate can yield the emission estimates for all the other compounds. When they employed an inventory-based emission estimate of HCFC-22 from the Pollutant Release and Transfer Register (PRTR) system 2002 of Japan (9.1 Gg/year) as the reference, the estimated emission rates of HCFC-141b, HCFC-142b, CFC-12, chloroform, and trichloroethylene for 2002 were consistent with their PRTR values within 10%. Emissions of CCl4 and CH3 CCl3 were much higher than their PRTR values, suggesting that their sources are not adequately accounted for in the current inventories. In a different type of study, Barnes et al. 2003 used correlations in pollution enhancements downwind of the New York City, Washington, DC, corridor to

Halogenated Volatile Organic Compounds

207

assess independent estimates of emissions. Atmospheric mixing ratios of carbon monoxide (CO) and C2 Cl4 were measured above the canopy at Harvard forest. Mixing ratios of gases regulated by the Montreal Protocol (CFC-11, CFC-12, CFC-113, CH3 CCl3 , and halon-1211) were referenced to CO and C2 Cl4 to determine their urban/industrial source strengths and to test existing estimates of U.S. emissions. In a recent paper by Culbertson et al. (2004), a onebox model was used to infer emissions for several CF3 -containing compounds, including CFC-115, Halon 1301, and the HFCs 23, 143a and 134a, using ﬂask samples from Oregon, Alaska and Antarctica. The emissions calculated in their study agreed generally well with emissions from other studies. They also provided ﬁrst estimates of emissions of some rarer gases, such as CFC-13 and C3 F8 . Figure 5.9 illustrates inferred emissions for several species since 1990. In general, the results presented in Figures 5.1 and 5.9 conﬁrm a clear downward trend for most compounds regulated by the Montreal Protocol. Inferred emissions for the HCFCs have been rising strongly since 1990, and those of HCFC-141b and HCFC-142b have levelled off from 2000 onwards. However, emissions of the HFCs are growing in most cases, most noticeably for HFC-134a, HFC-125 and HFC-152a.

5.6

Sampling techniques

Direct analysis of air is possible for halogenated VOCs that are present in the atmosphere at high concentration, certain CFC compounds and highly chlorinated halocarbons, such as methylchloroform and carbon tetrachloride. However, the need to increase the concentration of air samples prior to analysis (pre-concentration) exists for a majority of halogenated VOCs due to their very low concentrations in the troposphere ppb (parts per billion) to ppt. Pre-concentration allows halogenated VOCs to be measured in the normal detection range of common analytical techniques by trapping the compounds of interest preferentially to other atmospheric constituents. It is carried out either at the sampling location or after transport of the sample matrix to the laboratory, followed by gas chromatography (GC) or gas chromatography mass spectrometry (GC–MS) analysis. A great deal of research has been carried out on developing appropriate sampling and pre-concentration techniques.

5.6.1

Solid sorbent trapping

One of the most widely used methods to monitor halogenated VOCs in ambient air is the use of trapping on solid adsorbents. The method can be carried out at both ambient and reduced temperatures. Sorbent sampling at reduced temperature is different from cryogenic techniques (discussed later) that involve only solid glass beads. The adsorbed VOC can then be desorbed thermally, or by using solvent extraction and analysed by techniques such as GC and mass spectrometry (MS). The beneﬁts of this method of pre-concentration are the good capabilities for enrichment – giving high sensitivity, as well as being cost-effective and easy to use. In order for an adsorbent to be ideal for the pre-concentration of VOCs, it must have four main properties: 1. Inﬁnite breakthrough volume (BTV) 2. Complete desorption of the target compounds at moderate temperatures 3. No generation of artefacts 4. No retention of water vapour

208

Volatile Organic Compounds in the Atmosphere

300

CFC-11

250

CFC-12

CFC-13

200 150 100 50

300 200 100

CCI4

12

60 40

100

5

Halon-1211

10

Emission (kt/year)

Emission (kt/year)

Emission (kt/year)

80

150

0 1990 1992 1994 1996 1998 2000 2002

120 100

200

50

0 1990 1992 1994 1996 1998 2000 2002

0 1990 1992 1994 1996 1998 2000 2002 140

Emission (kt/year)

250

Emission (kt/year)

Emission (kt/year)

400

8 6 4

Halon-1301

4 3 2

20

2

1

0 1990 1992 1994 1996 1998 2000 2002

0 1990 1992 1994 1996 1998 2000 2002

0 1990 1992 1994 1996 1998 2000 2002

300

80

HCFC-22

40

HCFC-141b

HCFC-142b

200 150 100

60

Emission (kt/year)

Emission (kt/year)

Emission (kt/year)

250

40

20

30

20

10

50

15

HFC-23 Emission (kt/year)

Emission (kt/year)

15

0 1990 1992 1994 1996 1998 2000 2002

10

5

0 1990 1992 1994 1996 1998 2000 2002 120

HFC-125 Emission (kt/year)

0 1990 1992 1994 1996 1998 2000 2002

10

5

HFC-134a

100 80 60 40 20

0 1990 1992 1994 1996 1998 2000 2002 30

0 1990 1992 1994 1996 1998 2000 2002 7

HFC-152a Emission (kt/year)

Emission (kt/year)

SF5 NDAA in situ GC CATS NDAA in situ GC RITS NDAA flasks AGAGE 12 Box AGAGE UCI flasks UEA flasks AFEAS EDGAR

6

25 20 15 10

5 4 3 2

5

1

0 1990 1992 1994 1996 1998 2000 2002

0 1990 1992 1994 1996 1998 2000 2002

Year

0 1990 1992 1994 1996 1998 2000 2002

Year

Figure 5.9 Global annual emissions (in kt/year) inferred from the mole fractions in Figure 5.1. Emissions are estimated using a 1-box model. In addition, the AGAGE 12 box model has been used to infer emissions from their network (Prinn et al. 2000; updates provided by D. Cunnold), a time-dependent scaling for each component taking into account the vertical distribution in the troposphere and the stratosphere has been adopted in all the estimates. These scaling factors are taken from the AGAGE 12-box model. Emissions of HFC-23 are based on a 2-D model (Oram et al. 1998, with updates). The inferred emissions are compared to emissions from AFEAS and EDGAR. Reproduced with permission from IPCC/TEAP, IPCC/TEAP Special report on safeguarding the ozone layer and the global climate system: Issues related to hydroﬂuorocarbons and perﬂuorocarbons. Prepared by Working Groups I and III of the Intergovernmental Panel on Climate Change, and the Technical and Economic Assessment Panel. Cambridge University Press, UK and New York, 2005.

Halogenated Volatile Organic Compounds

209

No single adsorbent material satisﬁes all these properties, and hence the best-suited adsorbent is chosen for the compounds of interest. The issue of water retention on solid adsorbents is cause for concern because it affects the retention capacity of the adsorbent material for other compounds. There are several methods that can be used to reduce water interference. One possibility is to install a tube with hygroscopic salts [K2 CO3 , MgCO3 , Mg(ClO4 )2 ] in front of the sampling tube to dry the air sample without loss of the compounds of interest. Another method is to use a Naﬁon dryer in front of the sampling tube. The Naﬁon dryer uses a permeation distillation technique through a selectively permeable membrane.

5.6.2

Cryogenic sampling

This method of sampling is carried out at extremely low temperatures of lower than −150◦ C and, hence, incurs even greater problems with water interference than solid sorbent trapping. However, cryogenic sampling prevents the generation of artefacts caused by the adsorbent material in sorbent sampling. The cryogenic sampling technique has speciﬁc uses for the measurement of very low boiling point halocarbons, such as CF4 .

5.6.3

Canister sampling

This method is not a pre-concentration technique like the aforementioned methods. The canister method collects air samples with evacuated aluminium or stainless steel canisters. Normally the internal surfaces of the canisters have been treated in some way (electropolished, lined with fused silica) to reduce surface activity. After sampling into the canister, sorbent or cryogenic pre-concentration is then carried out. The advantage of this type of sampling is that it allows a single sample to be analysed multiple times. The disadvantages are that some compounds are not stable in the canisters, and sample loss or growth can occur over time, it is also an expensive procedure, which can require complicated cleaning of equipment.

5.6.4

Solid-phase microextraction

Solid-phase microextraction (SPME) combines sampling and pre-concentration into one step. In SPME, the analyte is adsorbed onto a ﬁne ﬁbre covered with sorbent material that projects from the needle of a syringe. Analysis is then carried out by inserting the syringe into the gas chromatograph inlet septum and injecting the ﬁbre into a heated zone. In this region, the analyte is thermally desorbed into the carrier gas ﬂow to the GC. The advantages of SPME are that it is a small, lightweight device that gives good linear response with concentration and has good sensitivity. However, the disadvantages are that quantitative information can only be derived by constructing calibration curves for each compound of interest, storage stability is poor and the collected samples are not time-integrated.

5.6.5

Passive diffusive sampling

This method is the least widely used sampling technique with only a limited number of studies having used this method. Sampling allows the monitoring of atmospheric concentrations

210

Volatile Organic Compounds in the Atmosphere

averaged over longer periods. Air concentrations are calculated based on an average uptake rate coefﬁcient, which can give rise to bias in the results.

5.7

Measurement techniques

Measurement of the halocarbons in the atmosphere has historically been based around GC techniques utilising the ECD. More recently, GC–MS has found wider use. In this section, these are the methods of measurement that will be described. There is also a wealth of more elaborate types of instrument such as the differential optical absorption spectrometer (DOAS) or Fourier transform infrared spectrometer (FTIR) and satelliteborne instrumentation; this type of instrumentation is beyond the scope of this chapter and is covered well in other publications (Brasseur et al. 1999)

5.7.1

Gas chromatography

Analysis of atmospheric constituents is conducted using a wide variety of analytical techniques, the most common of which is the use of GC. GC, in combination with a detector, has been found to be appropriate for separating and quantifying halogenated VOCs for the low atmospheric concentration levels. The principle behind chromatography is ‘the separation of components of a mixture by a series of equilibrium operations that result in separation due to partitioning between two different phases’. One of the phases is stationary (liquid or solid) with a large surface area (the column), the other is a moving phase (gas) in contact with the column. In order for a sample to be suitable for analysis by GC, it must consist of stable components that interact with the column material. The analyte components travel through the column at different rates depending on their retention times on the column. Separation of components occurs with continued eluant ﬂowing through the column, until separate bands emerge from the column, depending on their surface interactions. Helium is the most common carrier gas used in the analysis of halogenated VOCs. It is common to purify the carrier gas (N2 , H2 , He, Ar/CH4 ) prior to introduction into the GC system. Puriﬁcation is achieved by passing the carrier gas through a combination of puriﬁer traps or by cryogenic trapping. Compounds are separated due to differences in their ability to partition, or transfer between the stationary and mobile phase. This partition can be based on polarity, boiling point and occasionally chirality of the compounds. The analysis is performed with the column contained in an oven so that the temperature can be controlled. Varying the temperature of the oven allows the best possible separation of compounds to be achieved for a wide range of volatilities. Chromatographic separations can be evaluated by the shape of the peaks. Peak shapes depend on the isotherms that describe the relationship between concentration of the solute in the stationary phase and concentration of the solute in the carrier gas. There are a variety of possible detectors for GC; ﬂame ionisation detector (FID), mass spectrometer (MS), ECD and photoionisation detector (PID) being some of the main ones. A GC detector is ‘a device that senses the presence of a compound different from the carrier gas and converts this into an electronic signal’. Selection of the type of detector depends on

Halogenated Volatile Organic Compounds

211

the target compound, concentration range and whether qualitative or quantitative analysis is preferred. The usual experimental set-up for analysis of halogenated VOCs involves the use of ECD or MS detection.

5.7.2

Electron capture detector

The ECD is a highly sensitive detector capable of detecting picogram amounts of halogencontaining compounds. The high selectivity of this detector can be a great advantage in the detection of halogenated VOCs. Compared with the ﬂame ionisation detector (FID), it has much more limited linear response range, generally less than three orders of magnitude. The response can also vary signiﬁcantly with temperature, pressure and ﬂow rate. The detector uses a sealed radioactive source, Ni63 , sealed in a stainless steel cylinder and, thus, requires that certain radiological safety requirements be met. A ﬂux of beta particles generated by the Ni63 collides with the carrier gas molecules, causing them to ionise by ejecting thermal (i.e. low-energy electrons). The thermal electrons migrate to an anode (positive electrode), which generates a current signal. When a sample contains compounds that capture and so remove thermal electrons, some of the electrons are prevented from reaching the anode and consequently, results in a reduction in the baseline current. The detector electronics that maintain a constant current (of about 1 nanoampere) through the electron cloud are forced to pulse at a faster rate to compensate for the decreased number of thermal electrons. This change in pulse frequency provides the signal response for the electroncapturing compounds. Organic molecules that contain electronegative functional groups, such as halogens, phosphorous, and nitro groups, are extremely efﬁcient electron capturers.

5.7.3

Gas chromatography mass spectrometry

GC–MS is an instrumental technique, comprising a GC coupled to a mass spectrometer (MS), by which complex mixtures of chemicals may be separated, identiﬁed and quantiﬁed. This makes it ideal for the analysis of the hundreds of relatively low molecular weight compounds found in environmental materials. In order for a compound to be analysed by GC–MS, it must be sufﬁciently volatile and thermally stable. The sample solution is injected into the GC inlet where it is vapourised and swept onto a chromatographic column by the carrier gas (usually helium). The sample ﬂows through the column and the compounds comprising the mixture of interest are separated by virtue of their relative interaction with the coating of the column (stationary phase) and the carrier gas (mobile phase). The latter part of the column passes through a heated transfer line and ends at the entrance to the ion where compounds eluting from the column are converted to ions. Two potential methods exist for ion production. The most frequently used method is electron ionisation (EI), and the occasionally used alternative is chemical ionisation (CI). For EI, a beam of electrons ionise the sample molecules resulting in the loss of one electron. A molecule with one electron missing is called the molecular ion and is represented by M+ (a radical cation). When the resulting peak from this ion is seen in a mass spectrum, it gives the molecular weight of the compound. Due to the large amount of energy imparted to the molecular ion, it usually fragments, producing further smaller ions with characteristic relative abundances that provide a ‘ﬁngerprint’ for that molecular structure. This information

212

Volatile Organic Compounds in the Atmosphere

Column abundance (×1015 molecule/cm 2)

8 Pressure normalised monthly means June to November monthly means Polynomial fit to filled datapoints NPLS fit (20%)

7

CFC-12 6

5

2 HCFC-22 1

1986.0

1988.0

1990.0

1992.0

1994.0

1996.0

1998.0

2000.0

2002.0

2004.0

Calendar year

Figure 5.10 CFC-12 and HCFC-22 monthly mean time series at Jungfraujoch measured by the University of Liège, normalised to 654 hPa. Both polynomial and non-parametric least square (NPLS) ﬁts to the time bases are shown, respectively as a continuous and a dashed line. (Reproduced with permission from A. Mahieu, Final Report of the EU FP5 project “System for Observation of Halogenated Greenhouse Gases in Europe (SOGE)”. Project EVK2-2000-00674.)

may then be used to identify compounds of interest and help elucidate the structure of unknown components of mixtures. CI begins with the ionisation of methane (or another suitable gas), creating a radical which in turn will ionise the sample molecule to produce [M + H]+ molecular ions. CI is a less energetic way of ionising a molecule hence less fragmentation occurs with CI than with EI; hence, CI yields less information about the detailed structure of the molecule, but does yield the molecular ion; sometimes the molecular ion cannot be detected using EI, hence the two methods complement one another. Once ionised a small positive charge is used to repel the ions out of the ionisation chamber. The next component is a mass analyser (ﬁlter), which separates the positively charged ions according to various mass related properties depending upon the analyser used. The most common is a quadrupole mass analyser. The quadrupole consists of four parallel metal rods. A radio frequency voltage is applied across one pair of rods, and a direct current voltage across the other. Ions travel down the quadrupole in between the rods. Only ions of a certain mass to charge ratio (m/z) will reach the detector for a given ratio of voltages: other ions have unstable oscillations and will collide with the rods. This allows selection of a particular ion/group of ions, or scanning by varying the voltages. From this information, it is possible to determine with a high level of certainty the chemical composition of the original sample.

Halogenated Volatile Organic Compounds

213

220 CHCIF2

200

Volume mixing ratio (ppt)

180 160 140 120 100 80 60

VMR from total columns normalised to 654 hpa (monthly means) Linear regression to monthly mean VMRs ALEGAGE/AGAGE GLOBAL HCFC-22 (from Prinn et al. JGR, 2000) ALEGAGE/AGAGE scaled by 14% AGAGE (WMO 2002, GLOBAL)

40 20

1986.0

1988.0

1990.0

1992.0

1994.0 1996.0 Year

1998.0

2000.0

2002.0

2004.0

Figure 5.11 HCFC-22 monthly mean vmrs derived from FTIR total column measurements performed at the Jungfraujoch (46.5◦ N) are reproduced by empty circles, with error bars characterising the standard deviations around the means. The global evolution of HCFC-22 as derived from AGAGE measurements is reproduced by black crosses for comparison. This data set perfectly matches the linear regression to the Jungfraujoch vmrs after scaling by a factor of 1.14 (see grey and black traces). (Reproduced with permission from A. Mahieu, Final Report of the EU FP5 project “System for Observation of Halogenated Greenhouse Gases in Europe (SOGE)”. Project EVK2-2000-00674.)

Besides the more commonly used quadrupole mass analyser several other types of analysers are currently in use, including time-of-ﬂight, quadrupole ion trap, Fourier transform ion cyclotron resonance mass analysers, magnetic-sector, isotope ratio (IR-MS), and proton transfer (PTR-MS) types. Mass spectrometers have most commonly been used in the laboratory for analysing sample that have been collected at various locations and returned to the laboratory for analysis. However, in recent years, the robustness and low cost of benchtop MS systems has resulted in the use of mass spectrometers becoming more common on a wide variety of platforms, including remote ﬁeld stations, ships and aircraft. Much of the data from AGAGE and NOAA-GMD (National Oceanographic and atmospheric Administration/Global Monitoring Division) detailed in Section 5.4 of this chapter was acquired using in situ GC–MS.

5.7.4

Fourier transform infrared spectrometer

FTIR spectrometers are routinely operated at several sites worldwide at Ny-Ålesund/ Spitsbergen (79◦ N) and at the Jungfraujoch (46.5◦ N). Recorded spectra allow the retrieval

214

Volatile Organic Compounds in the Atmosphere

of total columns for several trace gases, for example, HCFC-22 and CFC-12, SF6 , HF. Available total column time series such as those displayed in Figure 5.10 have been consistently converted to ground-level concentrations. Validations of the resulting data sets have been performed by comparing with in situ datasets collected by the AGAGE, SOGE and NOAA/GMD networks. An illustration is given in Figure 5.11, where HCFC-22 surface data converted from Jungfraujoch total column measurements have been compared with the AGAGE global HCFC-22 time series. When accounting for a systematic bias of 14% (which is commensurate with uncertainty affecting the spectroscopic parameters used in the analysis and also partly results from comparison between global and northern mid-latitude concentrations), one obtains an excellent agreement between both datasets showing identical evolution of the HCFC-22 burden.

References Abrahamsson, K., Ekdahl, A., Collén, J. and Pedersén, M. (1995) Formation and distribution of halogenated volatile organic compounds in seawater. In: Naturally-produced organo-halogens. Edited by A. Grimvall and E.W.B. de Leer, Kluwer. AFEAS (2004) Production, sales and atmospheric releases of ﬂuorocarbons through 2001. http;// www.afeas.org Andreae, M.O. and Merlet, P. (2001) Emission of trace gases and aerosols from biomass burning. Global Biogeochemical Cycles, 15 (4): 955–66. Ashworth, S.H., Allan, B.J. and Plane, J.M. (2002) High resolution spectroscopy of the OIO radical: Implications for the ozone-depleting potential of iodine. Geophysical Research Letters, 29 (10): 1456. Ayers, G.P., Gillett, R.W., Cainey, J.M., et al. (1999) Chloride and bromide loss from sea-salt particles in southern ocean air. Journal of Atmospheric Chemistry, 33 (3): 299–319. Baker, A.R., Turner, S.M., Broadgate, W.J., et al. (2000) Distribution and sea-air ﬂuxes of biogenic trace gases in the eastern Atlantic Ocean. Global Biogeochemical Cycles, 14 (3): 871–86. Baker, J.M., Reeves, C.E., Nightingale, P.D., Penkett, S.A., Gibb, S.W. and Hatton, A.D. (1999) Biological production of methyl bromide in the coastal waters of the North Sea and open ocean of the northeast Atlantic. Marine Chemistry, 64 (4): 267–85. Barnes, D.H., Wofsy, S.C., Fehlau, B.P., et al. (2003) Urban/industrial pollution for the New York City – Washington, DC, corridor, 1996–1998: 1. Providing independent veriﬁcation of CO and PCE emissions inventories. Journal of Geophysical Research-Atmospheres, 108 (D6): JD001116. Barrie, L.A., Bottenheim, J.W., Schnell, R.C., Crutzen, P.J., Rasmussen, R.A. (1988) Ozone destruction and photochemical-reactions at polar sunrise in the lower Arctic atmosphere. Nature, 334 (6178): 138–41. Bell, N., Hsu, L., Jacob, D.J., et al. (2002) Methyl iodide: Atmospheric budget and use as a tracer of marine convection in global models. Journal of Geophysical Research-Atmospheres, 107 (D17): JD001151. Blake, N.J., Blake, D.R., Sive, B.C., et al. (1996) Biomass burning emissions and vertical distribution of atmospheric methyl halides and other reduced carbon gases in the South Atlantic region. Journal of Geophysical Research-Atmospheres, 101 (D19): 24151–64. Bloss, W.J., Lee, J.D., Johnson, G.P., et al. (2005) Impact of halogen monoxide chemistry upon boundary layer OH and HO2 concentrations at a coastal site. Geophysical Research Letters, 32 (6): GL022084. Bobrowski, N., Honninger, G., Galle, B. and Platt, U. (2003) Detection of bromine monoxide in a volcanic plume. Nature, 423 (6937): 273–6.

Halogenated Volatile Organic Compounds

215

Bosch, H., Camy-Peyret, C., Chipperﬁeld, M.P., et al. (2003) Upper limits of stratospheric IO and OIO inferred from center-to-limb-darkening-corrected balloon-borne solar occultation visible spectra: Implications for total gaseous iodine and stratospheric ozone. Journal of Geophysical Research-Atmospheres, 108 (D15): JD003078. Brasseur, G.P., Orlando, J.L. and Geofrrey, S. (1999) Atmospheric chemistry and global change. Edited by Guy P. Brasseur, John, J. Orlando, and Geoffrey, S. Tyndall, Oxford University Press (1999). Butler, J.H., Battle, M., Bender, M.L., et al. (1999) A record of atmospheric halocarbons during the twentieth century from polar ﬁrn air. Nature, 399 (6738): 749–55. Carpenter, L.J. (2003) Iodine in the marine boundary layer. Chemical Reviews, 103 (12): 4953–62. Carpenter, L.J., Penkett, W.T.S., Liss, S.A., et al. (1999) Short-lived alkyl iodides and bromides at Mace Head, Ireland: Links to biogenic sources and halogen oxide production. Journal of Geophysical Research-Atmosphere, 104: 1679–89. Carpenter, L.J., Liss, P.S. and Penkatt, S.A. (2003) Marine organohalogens in the atmosphere over the Atlantic and Southern Oceans. Journal of Geophysical Research-Atmospheres, 108 (D9): JD002769. Cicerone, R.J., Stolarsk, R.S. and Walters, S. (1974) Stratospheric ozone destruction by man-made chloroﬂuoromethanes. Science, 185 (4157): 1165–7. Class, T. and Ballschmiter, K. (1987) Chemistry of organic traces in Air. 9. Evidence of natural marine sources for chloroform in regions of high primary production. Fresenius Zeitschrift Fur Analytische Chemie, 327 (1): 40–1. Cotter, E.S.N., Booth, N.J., Canosa-Mas, C.E. and Wayne, R.P. (2001) Release of iodine in the atmospheric oxidation of alkyl iodides and the fates of iodinated alkoxy radicals. Atmospheric Environment, 35 (12): 2169–78. Cox, M.L., Sturrock, G.A., Sturrock, P.J., et al. (2003) Regional sources of methyl chloride, chloroform and dichloromethane identiﬁed from AGAGE observations at Cape Grim, Tasmania, 1998–2000. Journal of Atmospheric Chemistry, 45 (1): 79–99. Cox, R.A., Bloss, W.J., Jones, R.L. and Rowley, D.M. (1999) OIO and the atmospheric cycle of iodine. Geophysical Research Letters, 26 (13): 1857–60. Culbertson, J.A., Prins, J.M. and Grimsrud, E.P. (2000) Improvements in the detection and analysis of CF3 -containing compounds in the background atmosphere by gas chromatograpphy-highresolution mass spectrometry. Journal of Chromatography, A903(1-2) : 261–65. Culbertson, J.A., Prins, J.M., Grimsrud, E.P. Rasmussen, R.A., Khalil, M.A.K. and Shearer, M.J. (2004) Observed trends for CF3 -containing compounds in background air at Cape Meares, Oregon, Point Barrow, Alaska, and Palmer Station, Antarctica. Chemosphere, 55 (8), 1109–19. Dimmer, C.H., McCulloch, A., Simmonds, P.G., Nickless, G., Bassford, M.R. and Smythe-Wright, D. (2001) Tropospheric concentrations of the chlorinated solvents, tetrachloroethene and trichloroethene, measured in the remote northern hemisphere. Atmospheric Environment, 35 (7): 1171–82. Duce, R.A. and Hoffman, E.J. (1976) Chemical fractionation at air-sea interface. Annual Review of Earth And Planetary Sciences, 4: 187–228. Dvortsov, V.L., Geller, M.A., Solomon, S., Schaufﬂer, S.M., Atlas, E.L. and Blake, D.R. (1999) Rethinking reactive halogen budgets in the midlatitude lower stratosphere. Geophysical Research Letters, 26 (12): 1699–1702. Elkins, J.W. (1997) Climate Monitoring and Diagnostics Laboratory Summary Report No. 24, 1996– 1997. David, J. Hofman, Editor, Boulder, Colarada, Dec. 1998. Engel, A., Strunk, M., Müller, M., et al. (2002) Temporal development of total chlorine in the highlatitude stratosphere based on reference distributions of mean age derived from CO2 and SF6 . Journal of Geophysical Research-Atmospheres, 107 (D12): JD000584. Farman, J.C. and Gardiner, B.G. (1987) Ozone depletion over Antarctica. Nature, 329 (6140): 574. Farman, J.C., Gardiner, B.G. and Shanklin, J.D. (1985) Large losses of total ozone in Antarctica reveal seasonal Clox/Nox interaction. Nature, 315 (6016): 207–10.

216

Volatile Organic Compounds in the Atmosphere

Garcia, J.P., Beynemasclet, S., Mouvier, G., et al. (1992) Emissions of volatile organic-compounds by coal-ﬁred power-stations. Atmospheric Environment Part A – General Topics, 26 (9): 1589–97. von Glasow, R. and Crutzen, P.J. (2003) Tropospheric halogen chemistry. Treatise on Geochemistry, 4: 21–64. Graedel, T.E. and Crutzen, P.J. (1993) Atmospheric change; An Earth system perspective, Graedel, T.E. and Crutzen, P.J., W.H. Freeman and Co. Ltd. Graedel, T.E. and Keene, W.C. (1996) The budget and cycle of Earth’s natural chlorine. Pure and Applied Chemistry, 68 (9): 1689–97. Gschwend, P.M., Macfarlane, J.K. and Newman, K.A. (1985) Volatile halogenated organic-compounds released to seawater from temperate marine macroalgae. Science, 227 (4690): 1033–5. Harnisch, J. and Eisenhauer, A. (1998) Natural CF4 and SF6 on Earth. Geophysical Research Letters, 25 (13): 2401–4. Harnisch, J., Frische, M., Borchers, R., Eisenhauer, A. and Jordon, A. (2000) Natural ﬂuorinated organics in ﬂuorite and rocks. Geophysical Research Letters, 27 (13): 1883–6. Haselmann, K.F., Ketola, R.A., Laturnus, F., Lauritsen, F.R. and Gron, C. (2000) Occurrence and formation of chloroform at Danish forest sites. Atmospheric Environment, 34 (2): 187–93. Hoekstra, E.J., De Leer, E.W.B. and Brinkman, U.A.T. (1998) Natural formation of chloroform and brominated trihalomethanes in soil. Environmental Science and Technology, 32 (23): 3724–9. Hoffmann, T., O’Dowd, C.D. and Seinfeld, J.H. (2001) Iodine oxide homogeneous nucleation: An explanation for coastal new particle production. Geophysical Research Letters, 28 (10): 1949–52. IPCC (2001) Climate Change 2001: The Scientiﬁc Basis: Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change, edited by Houghton, T.T., Ding, Y. Griggs, D.J., Noguer, M., van der Linden, P.J., Dai, X., Maskell, K. and Johnson, CA. Cambridge University Press, Cambridge, UK, 881. Keene, W.C., Khalil, M.A.K., Erickson III, D.J., et al. (1999) Composite global emissions of reactive chlorine from anthropogenic and natural sources: Reactive chlorine emissions inventory. Journal of Geophysical Research-Atmospheres, 104 (D7): 8429–40. Keppler, F., Eiden, R., Niedan, V., Pracht, J. and Scholer, H.F. (2000) Halocarbons produced by natural oxidation processes during degradation of organic matter. Nature, 403 (6767): 298–301. Keppler, F., Eiden, R., Niedan, V. Pracht, J., and Scholer, H.F. (2001) Halocarbons produced by natural oxidation processes during degradation of organic matter (Vol. 403, pp. 298, 2000). Nature, 409 (6818): 382. Keppler, F., Harper, D.B., Rockmann, T., Moore, R.M. and Hamilton, J.T.G. (2005) New insight into atmospheric chloromethane budget gained using stable carbon isotope ratios. Atmospheric Chemistry and Physics, 5: 3899–919. Khalil, M.A.K., Rasmussen, R.A. and Hoyt, S.D. (1983) Atmospheric Chloroform (Chcl3) – Ocean air exchange and global mass balance. Tellus Series B-Chemical and Physical Meteorology, 35 (4): 266–74. Khalil, M.A.K., Moore, R.M., Harper, D.B., et al. (1999) Natural emissions of chlorine-containing gases: Reactive chlorine emissions inventory. Journal of Geophysical Research-Atmospheres, 104 (D7): 8333–46. King, D.B., Butler, J.H., Montzka, S.A., Yvon-Lewis, S.A. and Elkins, J.W. (2000) Implications of methyl bromide supersaturations in the temperate North Atlantic Ocean. Journal of Geophysical Research-Atmospheres, 105 (D15): 19763–9. Kleopfer, R.D., Easley, D.M., Haas, B.B., Deihl, T.G., Jackson, D.E. and Wurrey, C.J. (1985) Anaerobic degradation of trichloroethylene in soil. Environmental Science and Technology, 19 (3): 277–80. Ko, M.K.W. and Poulet, G. (2003) Very short-lived halogens and sulphur substances. Chapter 2, In: Scientiﬁc Assessment of Ozone Depletion; 2002, Global Ozone Research and Monitoring ProjectReport No. 47, Geneva, 2003.

Halogenated Volatile Organic Compounds

217

Koppmann, R., Johnen, F.J., Hass, B.B., Plassdulmer, C. and Rudolph, J. (1993) Distribution of methylchloride, dichloromethane, trichloroethene and tetrachloroethene over the North and South-Atlantic. Journal Of Geophysical Research-Atmospheres, 98 (D11): 20517–26. Kurylo, M.J. and Rodriguez, J.M. (1999) Short-Lived Ozone-Related Compounds. WMO Scientiﬁc Assessment of Ozone Depletion: 1998, World Meteorological Organization Global Ozone Research and Monitoring Project-Report No. 44, Geneva. Lee, J.M., Sturges, W.T., Penkett, S.A., et al. (1995) Observed stratospheric proﬁles and stratospheric lifetimes of Hcfc-141b and Hcfc-142b. Geophysical Research Letters, 22 (11): 1369–72. Lobert, J.M., Keene, W.C., Logan, J.A. and Yerich, R. (1999) Global chlorine emissions from biomass burning: Reactive chlorine emissions inventory. Journal Of Geophysical Research-Atmospheres, 104 (D7): 8373–89. Lovelock, J.E. (1962) Free electrons with thermal energy – Their generation for chemical use. Nature, 195 (4840): 488–493. Lovelock, J.E. (1963) Electron absorption detectors and technique for use in quantitative and qualitative analysis by gas chromatography. Analytical Chemistry, 35 (4): 474-5. Lovelock, J.E. (1974) Atmospheric halocarbons and stratospheric ozone. Nature, 252 (5481): 292–4. Lovelock, J.E. and Maggs, R.J. (1973) Halogenated hydrocarbons in and over the Atlantic. Nature, 241: 194–6. Manley, S.L. and delaCuesta, J.L. (1997) Methyl iodide production from marine phytoplankton cultures. Limnology And Oceanography, 42 (1): 142–7. Manning, A.J., Ryall, D.B., Derwent, R.G., Simmonds, P.G. and O’Doherty, S. (2003) Estimating European emissions of ozone-depleting and greenhouse gases using observations and a modeling back-attribution technique. Journal of Geophysical Research-Atmospheres, 108 (D14): JD002512. Mano, S. and Andreae, M.O. (1994) Emission of methyl-bromide from biomass burning. Science, 263 (5151): 1255–7. McCulloch, A. (1999) CFC and halon replacements in the environment. Journal of Fluorine Chemistry, 100 (1–2): 163–73. McCulloch, A. (2003) Chloroform in the environment: Occurrence, sources, sinks and effects. Chemosphere, 50 (10): 1291–308. McCulloch, A. and Midgley, P.M. (1996) The production and global distribution of emissions of trichloroethene, tetrachloroethene and dichloromethane over the period 1988–1992. Atmospheric Environment, 30 (4): 601–8. McCulloch, A., Aucott, M.L. and Benkovitz, C.M. (1999a) Global emissions of hydrogen chloride and chloromethane from coal combustion, incineration and industrial activities: Reactive chlorine emissions inventory. Journal of Geophysical Research-Atmospheres, 104 (D7): 8391–403. McCulloch, A., Aucott, M.L. and Graedel, T.E. (1999b) Industrial emissions of trichloroethene, tetrachloroethene, and dichloromethane: Reactive chlorine emissions inventory. Journal of Geophysical Research-Atmospheres, 104 (D7): 8417–27. McCulloch, A., Ashford, P. and Midgley, P.M. (2001) Historic emissions of ﬂuorotrichloromethane (CFC-11) based on a market survey. Atmospheric Environment, 35 (26): 4387–97. McCulloch, A., Midgley, P.M. and Ashford, P. (2003) Releases of refrigerant gases (CFC-12, HCFC-22 and HFC-134a) to the atmosphere. Atmospheric Environment, 37 (7): 889–902. McFiggans, G., Plane, J.M.C., Allan, B.J., Carpenter, L.J., Coe, H. and O’Dowd, C. (2000) A modeling study of iodine chemistry in the marine boundary layer. Journal of Geophysical Research-Atmospheres, 105 (D11): 14371–85. Midgley, P. (1997) New directions: HCFCs and HFCs: Halocarbon replacements for CFCs. Atmospheric Environment, 31 (7): 1095–6. Midgley, P.M. and McCulloch, A. (1995) The production and global distribution of emissions to the atmosphere of 1,1,1-trichloroethane (methyl chloroform). Atmospheric Environment, 29 (14): 1601–8.

218

Volatile Organic Compounds in the Atmosphere

Molina, M.J. and Rowland, F.S. (1974a) Stratospheric sink for chloroﬂuoromethanes – Chlorine atom catalyzed destruction of ozone. Bulletin Of The American Meteorological Society, 55 (5): 810–812. Molina, M.J. and Rowland, F.S. (1974b) Stratospheric sink for chloroﬂuoromethanes – Chlorine atomic-catalysed destruction of ozone. Nature, 249 (5460): 810–2. Montzka, S.A. and Fraser, P.J. (2003) Controlled substances and other source gases. WMO Scientiﬁc Assessment of Ozone Depletion: 2002, Global Ozone Research and Monitoring Project-Report No. 47, Geneva, 2003. Montzka, S.A., Butler, J.H., Hall, B.D., Mondeel, D.J. and Elkins, J.W. (2003) A decline in tropospheric organic bromine. Geophysical Research Letters, 30 (15): GL017745. Moore, R.M. and Zaﬁriou, O.C. (1994) Photochemical production of methyl-iodide in seawater. Journal of Geophysical Research-Atmospheres, 99 (D8): 16415–20. Moore, R.M. and Groszko, W. (1999) Methyl iodide distribution in the ocean and ﬂuxes to the atmosphere. Journal of Geophysical Research-Oceans, 104 (C5): 11163–71. Moore, R.M., Gut, A. and Andreae, M.O. (2005) A pilot study of methyl chloride emissions from tropical woodrot fungi. Chemosphere, 58 (2): 221–5. Mossinger, J.C., Shallcross, D.E. and Cox, R.A. (1998) UV-VIS absorption cross-sections and atmospheric lifetimes of CH2Br2, CH2I2 and CH2BrI. Journal of The Chemical Society-Faraday Transactions, 94 (10): 1391–6. O’Doherty, S., Simmonds, P.G. and Cunnold, D.M. (2001) In situ chloroform measurements at Advanced Global Atmospheric Gases Experiment atmospheric research stations from 1994 to 1998. Journal of Geophysical Research-Atmospheres, 106 (D17): 20429–44. O’Doherty, S., Cunnold, D.M. and Manning, A. (2004) Rapid growth of hydroﬂuorocarbon 134a and hydrochloroﬂuorocarbons 141b, 142b, and 22 from Advanced Global Atmospheric Gases Experiment (AGAGE) observations at Cape Grim, Tasmania, and Mace Head, Ireland. Journal of Geophysical Research-Atmospheres, 109 (D6): JD004277. Oram, D.E., Sturges, W.T., Penkett, S.A., McCulloch, A. and Fraser, P.J. (1998) Growth of ﬂuoroform (CHF3 , HFC-23) in the background atmosphere. Geophysical Research Letters, 21 (1): 35–38. O’Dowd, C.D., Jimenez, J.L. and Makela, J.M. (2002) Marine aerosol formation from biogenic iodine emissions. Nature, 417 (6889): 632–6. Olaguer, E.P. (2002) The distribution of the chlorinated solvents dichloromethane, perchloroethylene, and trichloroethylene in the global atmosphere. Environmental Science And Pollution Research, 9 (3): 175–82. Palmer, P.I., Jacob, D.J. and Mickley, L.J. (2003) Eastern Asian emissions of anthropogenic halocarbons deduced from aircraft concentration data. Journal of Geophysical Research-Atmospheres 108 (D24): JD003591. Patra, P.K., Lal, S. and Sheel, V. (2000) Chlorine partitioning in the stratosphere based on in situ measurements. Tellus Series B-Chemical And Physical Meteorology, 52 (3): 934–46. Platt, U. and Honninger, G. (2003) The role of halogen species in the troposphere. Chemosphere, 52 (2): 325–38. Prather, M. and Ehhalt, D. (2001) Climate change 2001: the scientiﬁc basis: contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change. In: IPCC (Intergovernmental Panel on Climate Change), Edited by J.T. Houghton, Y. Ding, D.J. Griggs, M. Noguer, P.J. van der Linden, X. Dai, K. Maskell, and C.A. Johnson. Cambridge University Press. Prinn, R.G., Huang, J. and Weiss, R.F. (2005) Evidence for variability of atmospheric hydroxyl radicals over the past quarter century. Geophysical Research Letters, 32 (7): L07809. Prinn, R.G., Weiss, R.F. and Fraser, P.J. (2000) A history of chemically and radiatively important gases in air deduced from ALE/GAGE/AGAGE. Journal of Geophysical Research, 105 (D14): 17751–92. Rattigan, O.V., Shallcross, D.E., and Cox, R.A. (1997) UV absorption cross-sections and atmospheric photolysis rates of CF3I, CH3I, C2H5I and CH2ICl. Journal of The Chemical Society-Faraday Transactions, 93 (16): 2839–46.

Halogenated Volatile Organic Compounds

219

Redeker, K.R., Wang, N.Y., Low, J.C., McMillan, A., Tyler, S.C. and Cicerone, R.J. (2000) Emissions of methyl halides and methane from rice paddies. Science, 290 (5493): 966–9. Reimann, S., Schaub, D. and Stemmler, K. (2004) Halogenated greenhouse gases at the Swiss High Alpine Site of Jungfraujoch (3580 m asl): Continuous measurements and their use for regional European source allocation. Journal of Geophysical Research-Atmospheres, 109 (D5): JD003923. Reimann, S., Manning, A.J. and Simmonds, P.G. (2005) Low European methyl chloroform emissions inferred from long-term atmospheric measurements. Nature, 433 (7025): 506–8. Rhew, R.C., Miller, B.R. and Bill, M. (2002). Environmental and biological controls on methyl halide emissions from southern California coastal salt marshes. Biogeochemistry, 60 (2): 141–61. Richter, U. and Wallace, D.W.R. (2004) Production of methyl iodide in the tropical Atlantic Ocean. Geophysical Research Letters, 31 (23): GL020779. Roehl, C.M., Burkholder, J.B., Moortgat, G.K., Ravishankara, A.R. and Crutzen, P.J. (1997) Temperature dependence of UV absorption cross sections and atmospheric implications of several alkyl iodides. Journal of Geophysical Research-Atmospheres, 102 (11D): 12819–29. Saiz-Lopez, A. and Plane, J.M.C. (2004) Novel iodine chemistry in the marine boundary layer. Geophysical Research Letters, 31 (4): GL019215. Sander, R., Keene, W.C. and Pszenny, A.A.P. (2003) Inorganic bromine in the marine boundary layer: A critical review. Atmospheric Chemistry And Physics, 3: 1301–36. Schaufﬂer, S.M., Heidt, L.E. and Pollock, W.H. (1993) Measurements of halogenated organiccompounds near the tropical tropopause. Geophysical Research Letters, 20 (22): 2567–70. Schaufﬂer, S.M., Atlas, E.L. and Donnelly, S.G. (2003) Chlorine budget and partitioning during the Stratospheric Aerosol and Gas Experiment (SAGE) III Ozone Loss and Validation Experiment (SOLVE). Journal of Geophysical Research-Atmospheres, 108 (D5): JD002040. Simmonds, P.G., Fenimore, D.C., Pettitt, B.C., Lovelock, J.E. and Zlatkis, A. (1967) Design of a nickel-63 electron absorption detector and analytical signiﬁcance of high temperature operation. Analytical Chemistry, 39 (12): 1428–31. Simmonds, P.G., Derwent, R.G. and Manning, A.J. (2004) AGAGE observations of methyl bromide and methyl chloride at Mace Head, Ireland, and Cape Grim, Tasmania, 1998–2001. Journal of Atmospheric Chemistry, 47 (3): 243–69. Singh, H.B., Salas, L.J. and Stiles, R.E. (1983) Selected man-made halogenated chemicals in the air and oceanic environment. Journal of Geophysical Research-Oceans And Atmospheres, 88 (NC6): 3675–83. Singh, O.N. and Fabian, P. (1999) Reactive bromine compounds. The Handbook of Environmental Chemistry – Reactive Halogen Compounds in the Atmosphere, 4: 1–43. Solomon, S., Garcia, R.R. and Ravishankara, A.R. (1994) On the role of iodine in ozone depletion. Journal of Geophysical Research-Atmospheres, 99 (D10): 20491–9. Stolarsk, R.S. and Cicerone, R.J. (1974) Stratospheric chlorine – Possible sink for ozone. Canadian Journal of Chemistry-Revue Canadienne De Chimie, 52 (8): 1610–15. Sturges, W.T., Cota, G.F. and Buckley, P.T. (1992). Bromoform emission from Arctic ice algae. Nature, 358 (6388): 660–2. Sturges, W.T., Oram, D.E., Carpenter, L.J., Penkett, S.A. and Engel, A. (2000) Bromoform as a source of stratospheric bromine. Geophysical Research Letters, 27 (14): 2081–4. Sturges, W.T., McIntyre, H.P. Penkett, S.A. (2001) Methyl bromide, other brominated methanes, and methyl iodide in polar ﬁrn air. Journal of Geophysical Research-Atmospheres, 106 (D2): 1595–606. Sturrock, G.A., Etheridge, D.M., Trudinger, C.M., Fraser, P.J. and Smith, A.M. (2002) Atmospheric histories of halocarbons from analysis of Antarctic ﬁrn air: Major Montreal Protocol species. Journal of Geophysical Research-Atmospheres, 107 (D24): JD002548. Stutz, J., Hebestreit, K., Alicke, B. and Platt, U. (1999) Chemistry of halogen oxides in the troposphere: Comparison of model calculations with recent ﬁeld data. Journal of Atmospheric Chemistry, 34 (1): 65–85.

220

Volatile Organic Compounds in the Atmosphere

Thomas, V.M., Bedford, J.A. and Cicerone, R.J. (1997) Bromine emissions from leaded gasoline. Geophysical Research Letters, 24 (11): 1371–4. Tokarczyk, R., Saltzman, E.S., Moore, R.M. and Yron-Lewis, S.A. (2003) Biological degradation of methyl chloride in coastal seawater. Global Biogeochemical Cycles, 17 (2): GB001949. Trudinger, C.M., Etheridge, D.M., Sturrock, G.A., Fraser, P.J., Krummel, P.B. and McCulloch, A. (2004) Atmospheric histories of halocarbons from analysis of Antarctic ﬁrn air: Methyl bromide, methyl chloride, chloroform, and dichloromethane. Journal of Geophysical Research-Atmospheres, 109 (D22): JD004932. United Nations Environment Programme (UNEP) (1998) Report of the Technology and Economic Assessment Panel (TEAP), 286pp., Nairobi, Kenya. Vogt, R., Sander, R., Von Glasow, R. and Crutzen, P.J. (1999) Iodine chemistry and its role in halogen activation and ozone loss in the marine boundary layer: A model study. Journal of Atmospheric Chemistry, 32 (3): 375–95. Wamsley, P.R., Elkins, J.W. and Fahey, D.W. (1998) Distribution of halon-1211 in the upper troposphere and lower stratosphere and the 1994 total bromine budget. Journal of Geophysical Research-Atmospheres, 103 (D1): 1513–26. Watling, R. and Harper, D.B. (1998) Chloromethane production by wood-rotting fungi and an estimate of the global ﬂux to the atmosphere. Mycological Research, 102: 769–87. Wayne, R.P. (2006) Chemistry of Atmospheres. 3rd Edition, Oxford University Press. Weissﬂog, L., Lange, C.A. and Pfenningsdorff, A. (2005) Sediments of salt lakes as a new source of volatile highly chlorinated C1/C2 hydrocarbons. Geophysical Research Letters, 32 (1): GL020807. Weston, R.E. (1996) Possible greenhouse effects of tetraﬂuoromethane and carbon dioxide emitted from aluminum production. Atmospheric Environment, 30 (16): 2901–10. WMO (1998) Scientiﬁc Assessment of Ozone Depletion: 1998, World Meteorological Organization Global Ozone Research and Monitoring Project-Report No. 44, Geneva, 1999. WMO (2003) Scientiﬁc Assessment of Ozone Depletion: 2002, Global Ozone Research and Monitoring Project-Report No. 47, Geneva, 2003. Wong, G.T.F. (1991) The marine geochemistry of iodine. Reviews In Aquatic Sciences, 4 (1): 45–73. Yokouchi, Y., Barrie, L.A., Toom, D. and Akimoto, H. (1996) The seasonal variation of selected natural and anthropogenic halocarbons in the Arctic troposphere. Atmospheric Environment, 30 (10–11): 1723–7. Yokouchi, Y., Noijiri, Y. and Barrie, L.A. (2000) A strong source of methyl chloride to the atmosphere from tropical coastal land. Nature, 403 (6767): 295–8. Yokouchi, Y., Ikeda, M., Inuzuka, Y. and Yukawa, T. (2002) Strong emission of methyl chloride from tropical plants. Nature, 416 (6877): 163–5. Yokouchi, Y., Inagaki, T., Yazawa, K., Tamaru, T., Enomoto, T. and Izumi, K. (2005) Estimates of ratios of anthropogenic halocarbon emissions from Japan based on aircraft monitoring over Sagami Bay, Japan. Journal of Geophysical Research-Atmospheres, 110 (D6): JD005320. Zander, R., Solomon, S. and Mahiev, E. (1996) Increase of stratospheric carbon tetraﬂuoride (CF4) based on ATMOS observations from space. Geophysical Research Letters, 23 (17): 2353–6.

Volatile Organic Compounds in the Atmosphere Edited by Ralf Koppmann Copyright © 2007 by Blackwell Publishing Ltd

Chapter 6

PAN and Related Compounds James M. Roberts

Peroxycarboxylic nitric anhydrides (PANs) are among the most unusual and important organic compounds in the atmosphere. PANs are formed in the atmosphere through the same complex chemistry, involving volatile organic compounds (VOCs) and oxides of nitrogen (NOx ), that forms ozone (O3 ) (Figure 6.1). As such, they are excellent indicators of the extent of VOC-NOx photochemistry, and the relative abundance of the different PAN compounds contains information on the mixture of VOCs that was involved in the O3 formation process. The PANs can have health and phytotoxic effects at high concentrations. In spite of their unusual and, at ﬁrst appearances, highly polar structure, R-C(O)OONO2 , the simple alkyl PAN compounds, are quite stable in the mid to upper troposphere. In fact, PAN is the most abundant odd nitrogen (NOy ) species in this environment. As a consequence, PAN compounds can be transported over quite long distances, slowly releasing NOx in the form of NO2 . This NOx source is thought to have a substantial impact on O3 in the remote troposphere. Tropospheric O3 is of importance as a greenhouse gas, and as a key ingredient in the oxidising capacity of the troposphere.

Fragmentations

Dicarbonyls, difunctional radicals

R' + CO2 NO2 NO2 R'C(O)OONO2 R'C(O)OO −NO2 OH, O2 NO

RH

OH RO2 O2

NO

O2 RO

NO2

OH R'(C O)R" or R'CH O hν

HO2

O2

R'C O

NO

HOCH2-C(OO)HR' NO2

O3 OZONIDE

Direct emissions

Figure 6.1 Gas-phase VOC-NOx photochemistry leading to PANs.

CH2 CHR'

222

Volatile Organic Compounds in the Atmosphere

The simplest PAN compound found in the atmosphere, peroxyacetic nitric anhydride, CH3 C(O)OONO2 , is commonly known by its misnomer, peroxyacetyl nitrate. It was discovered in the late 1950s by Stephens et al. (1956) during infrared spectroscopic studies of the Los Angeles smog. Stephens (1987) has recounted the early days of that research. The origin of PAN discovery is partly responsible for its name; the early IR spectroscopic analyses clearly showed the key features of the molecule, the carbonyl group and the nitrate group. The spectrum was similar, but not identical to acetyl nitrate (CH3 C(O)ONO2 ), and subsequent work revealed a peroxy linkage. Unfortunately, putting the two groups − and nitrate, − −ONO2 , leads to the wrong structure, together, peroxyacetyl, CH3 C(O)OO− CH3 C(O)OOONO2 . Domalski (1971) pointed out this nomenclature problem, and proposed that a variant of the group nomenclature be used, in which the organic moiety is speciﬁed as an acyl group {e.g. acetyl, CH3 C(O)} and the nitrogen moiety as a ‘peroxynitrate’. This nomenclature has the advantage of clarity, as the correct structure is arrived at unambiguously, but has achieved only partial acceptance, chieﬂy among European groups. Martinez (1980) asserted correctly that these compounds should be named as mixed anhydrides of peroxycarboxylic and nitric acids, and suggested that the ofﬁcial IUPAC name for PAN should be ethaneperoxoic nitric anhydride (EPNA). This name, while strictly correct, has never achieved wider usage. Roberts (1990) suggested an alternative nomenclature, for example, peroxyacetic nitric anhydride, which yields the correct structure and preserves the traditional acronym, PAN. In addition, this nomenclature is readily adapted to the larger PAN compounds such as peroxypropionic nitric anhydride, PPN, and peroxyisobutyric anhydride, PiBN. Exceptions to this acronym system have been made for peroxymethacrylic nitric anhydride, MPAN, and peroxyacrylic nitric anhydride, APAN. The anhydride nomenclature of Roberts (1990) will be used throughout this chapter. This chapter will attempt to summarise the current state of knowledge about PAN compounds in the atmosphere (Table 6.1). The ﬁrst section will discuss the chemistry of these species as it pertains to their synthesis and characterisation, and as it pertains to their behaviour in the atmosphere. The second section will discuss techniques for the measurement of PANs in the atmosphere. The third section will summarise what is currently known about the atmospheric abundance of these species. The fourth section will address the interpretation and numerical modelling that has been used to understand the role of PANs in the atmosphere. This chapter will rely heavily on previous review articles on PAN chemistry (Altshuller 1993; Gaffney et al. 1989; Kleindienst 1994; Roberts 1990; Singh 1987; Stephens 1969) hence will focus on material that has been published since those articles. Along the way, some attention will be given to areas where uncertainties still exist.

6.1 6.1.1

The chemistry of PANs Synthetic preparation

The preparation of PANs is accomplished either through wet-chemistry or photochemistry. Since the photochemical methods involve the same chemistry that operates in the atmosphere, those will be discussed in the Atmospheric Formation section and the wet-chemical methods discussed in this section. The proper class name for PANs, peroxycarboxylic nitric anhydrides is the key to their wet-chemical synthesis, the reaction of a peroxycarboxylic

PAN and Related Compounds

223

acid with a nitrating agent. These methods date back to work described by Stephens (1969), which showed that PAN could be produced by the reaction of peroxyacetic acid (PAA) with nitric acid in mixed acid solution and by reaction of PAA with silver nitrate. Louw et al. (1975) demonstrated that PANs could be synthesised by reaction of the corresponding peroxycarboxylic acid with N2 O5 in CH2 Cl2 solution. Several groups have reported the synthesis of PAN from PAA and mixed HNO3 /H2 SO4 in the presence of an alkane solvent (Gaffney et al. 1984; Nielsen et al. 1982) as a means to produce a high-purity solution. The key aspects of these syntheses are that temperatures are kept low, 0◦ C or below, and the H2 O2 or nitrating agents are adding slowly to minimise the effects of endothermic reactions. As a practical matter, the transportation, storage and disposal of commercial PAA solutions are expensive and involved, because they are toxic and mutagenic, and also because PAA decomposes upon storage. PAA can be made, when needed, from ingredients that are easier to transport and store by the methods outlined below for other peracids. The wet chemical synthesis of other PAN compounds can also be accomplished with the peroxyacid/nitration method, provided the peroxyacid can be synthesised. This is usually done by the reaction of the corresponding carboxylic anhydride with hydrogen peroxide (H2 O2 ). Such preparations are accomplished routinely for other simple linear alkyl groups, such as propionic and n-butyric, and PANs as large as n-hexanoic have been produced using the peroxyacid/nitration method. Branched-chain and unsaturated alkyl groups can also be synthesised by this method, but care must be taken to minimise decomposition and polymerisation reactions. For example, APAN and MPAN, unsaturated PANs, can be synthesised by the peroxyacid/nitration method, contrary to the suggestion of Roberts (1990), if care is taken in the addition of hydrogen peroxide to the acid anhydride and in the addition of the nitration solution (Bertman and Roberts 1991; Roberts et al. 2001b; Tanimoto and Akimoto 2001). Williams et al. (2000) found that PiBN could be synthesised from the acid anhydride, H2 O2 and strong acid nitration but there was also signiﬁcant decomposition to form 2-propyl nitrate. The aromatic PAN, peroxybenzoic nitric anhydride PBzN, has been made by reaction of peroxybenzoic acid with NaNO3 /H2 SO4 solution (Kravetz et al. 1980), and by the reaction of peroxybenzoic acid with nitronium tetraﬂuoroborate (NO2 BF4 ) (Kenley and Hendry 1982), the peroxybenzoic acid having ﬁrst been made from benzoic anhydride and H2 O2 . The complicating aspect of this synthesis is that the reactants and products are not liquid at 0◦ C. Purity of PANs synthesised by wet chemistry can range from very high for simple linear alkyl PANs to fairly low for unsaturated PANs such as APAN and MPAN. The purity of simple alkyl PANs, chieﬂy PAN and PPN, has been found to be quite high (>97% as nitrogen), provided the starting materials are pure (NO2 impurities in HNO3 can be an issue), and care is taken to separate the acid and lipid layer prior to extraction with ice water. The synthetic products of APAN, MPAN, and PiBN are known to require puriﬁcation if they are to be used for calibration or other experimental purposes. Nouaime et al. (1998) describe a silica gel column puriﬁcation method for MPAN that produces MPAN of sufﬁcient purity that direct gas-phase calibration methods can be used. Joos et al. (1986) describe a GC pre-separation method for the puriﬁcation of PAN that has been adopted and modiﬁed for use with MPAN and PiBN (Flocke et al. 2005; Williams et al. 2000). The advantage of this method is that the gas stream containing the puriﬁed MPAN is introduced to a calibration instrument and the ﬁeld instrument at the same time eliminating any uncertainties about decomposition during storage.

135.08 189.05 225.03 147.09 147.09

CH3 CH2 C(O)OONO2 CF3 CH2 C(O)OONO2 CF3 CF2 C(O)OONO2 CH2 C(CH3 )C(O)OONO2 CH3 CHCHC(O)OONO2

C3 H5 C(O)OONO2

CH3 CH(CH3 )C(O)OONO2 CH3 CH2 CH2 C(O)OONO2 (CH3 )2 CCHC(O)OONO2 C4 H7 C(O)OONO2

Peroxypropionic Peroxytriﬂuoropropionic Peroxypentaﬂuoropropionic Peroxymethacrylic Peroxycrotonic

Peroxycyclopropane Carboxylic

Peroxyisobutyric Peroxy-n-butyric Peroxy-3-methyl-butenoic Peroxycyclobutane Carboxylic

149.10 149.10 161.11 161.11

147.09

155.49 199.83 175.02 191.47 207.93 224.38 132.03 137.05 137.05 133.06

ClCH2 C(O)OONO2 BrCH2 C(O)OONO2 CF3 C(O)OONO2 CF2 ClC(O)OONO2 CFCl2 C(O)OONO2 CCl3 C(O)OONO2 NCC(O)OONO2 HOCH2 C(O)OONO2 CH3 OC(O)OONO2 CH2 CHC(O)OONO2

Peroxychloroacetic Peroxybromoacetic Peroxytriﬂuoroacetic Peroxydiﬂuorochloroacetic Peroxyﬂuorodichloroacetic Peroxytrichloroacetic Peroxycyanoacetic Peroxyglycolic Methoxy-peroxyformic Peroxyacrylic

125.01 141.47 121.05

Molecular weight

FC(O)OONO2 ClC(O)OONO2 CH3 C(O)OONO2

Formula

Peroxyﬂuoroformic Peroxychloroformic Peroxyacetic

Name

Table 6.1 Summary of known PANs

Photochemical/IR Photochemical/IR Peracid/nitration Photochemical IR, GC, MS Photochemical/IR Photochemical/IR Photochemical/IR Photochemical/IR Photochemical/IR Photochemical/IR Photochemical/IR Photochemical/IR Photochemical/IR Photochemical/ Thermal decomp Peracid/nitration/IR Photochemical/IR Photochemical/IR Photochemical/IR Photochemical/ Thermal decomp Photochemical/ Thermal decomp PTRMS Peracid/nitration/GC/ECD Peracid/nitration/IR Photochemical/IR Photochemical/ Thermal decomp PTRMS

Synthesis/evidence

Roumelis and Glavas (1992) Stephens (1964) Tuazon et al. (2005) D’Anna, et al. (2005)

D’Anna, et al. (2005)

Stephens (1964) Hurley et al. (2005) Sulbaek-Andersen et al. (2003) Tuazon and Atkinson (1990) Grosjean et al. (1994b)

Chen et al. (1996) Chen et al. (1996) Wallington et al. (1994) Zabel et al. (1994) Zabel et al. (1994) Zabel et al. (1994) Tyndall et al. (2001) Niki et al. (1987) Kirchner et al. (1997) Grosjean et al. (1994c)

Edney et al. (1979) Spence et al. (1978) Stephens (1969)

Reference

177.11 179.13 177.16 183.12 189.17 217.56 228.12 197.14 197.14 197.14 197.14 199.12 211.17 211.17 225.20 233.18 246.24 219.24

C5 H9 C(O)OONO2

HC(O)(CH2 )3 C(O)OONO2 (CH3 )3 COC(O)OONO2 CH3 (CH2 )4 C(O)OONO2 C6 H5 C(O)OONO2

C6 H11 C(O)OONO2 ClC6 H4 C(O)OONO2 O2 NC6 H4 C(O)OONO2 C6 H4 (CH3 )C(O)OONO2 C6 H4 (CH3 )C(O)OONO2 C6 H4 (CH3 )C(O)OONO2 C6 H5 CH2 C(O)OONO2 C6 H5 OC(O)OONO2 C6 H3 (CH3 )2 C(O)OONO2 C6 H3 (CH3 )2 C(O)OONO2 C6 H2 (CH3 )3 C(O)OONO2 C10 H7 C(O)OONO2 CH3 C(O)C7 H13 C(O)OONO2 CH3 (CH2 )7 C(O)OONO2

CH3 S(O)2 OONO2

Peroxycyclopentane Carboxylic

5-oxo-Peroxy-n-pentanoic t -Butoxyperoxyformic Peroxy-n-hexanoic Peroxybenzoic

Peroxycyclohexylcarboxylic m-Cl-Peroxybenzoic p-NO2 -Peroxybenzoic o-Methylperoxybenzoic m-Methylperoxybenzoic p-Methylperoxybenzoic β-Phenylperoxyacetic Phenoxyperoxyformic 2,4-Dimethylperoxybenzoic 2,5-Dimethylperoxybenzoic 2,4,6-Trimethylperoxybenz. 2-Peroxynapthoic α-Pinonaldehyde-PAN Peroxynonanoic

Peroxymethylsulfonic

157.10

175.14

163.13

CH3 (CH2 )3 C(O)OONO2

Peroxy-n-pentanoic

163.13

CH3 CH(CH3 )CH2 C(O)OONO2

Peroxy-i-pentanoic

Peracid/nitration/GC/ECD Thermal decomp Alkaline Hydrolysis Peracid/nitration/GC/ECD Thermal decomp Alkaline Hydrolysis Photochemical/ Thermal decomp PTRMS Photochemical/IR Photochemical/IR Peracid/nitration Photochemical/IR Elemental Analysis Photochemical/IR Peracid/N2 O5 Peracid/nitration Photochemical/IR Photochemical/IR Photochemical/IR Photochemical/IR Photochemical/IR Photochemical/IR Photochemical/IR Photochemical/IR Peracid/nitration Phototchemistry/IR Photochemistry Basic hydrolysis/IC Photochemistry/IR Barnes et al. (1987)

Glasson and Heuss (1977) van Swieten et al. (1978) Louw et al. (1975) Glasson and Heuss (1977) Glasson and Heuss (1977) Glasson and Heuss (1977) Glasson and Heuss (1977) Kirchner et al. (1999) Glasson and Heuss (1977) Glasson and Heuss (1977) Glasson and Heuss (1977) Louw et al. (1975) Noziere and Barnes (1998) Bowman et al. (2003)

Rogers and Rhead (1987) Kirchner et al. (1997) van Swieten et al. (1978) Heuss and Glasson (1968)

D’Anna, et al. (2005)

Grosjean et al. (1994a)

Grosjean et al. (1994a)

226

Volatile Organic Compounds in the Atmosphere

Table 6.2 Physical properties of PAN compounds Melting point, ◦ C

Compound Peroxyﬂuoroformic Peroxyacetic Peroxytriﬂuoroacetic Peroxybenzoic

6.1.2

−105 −48.5 −92

Boiling point, ◦ C(pvap , torr)

Reference

32 (760) 104 (760) 42.5 (760) 37 (0.2)

Schefﬂer et al. (1997) Kacmarek et al. (1977) Kopitzky et al. (1997) Heuss and Glasson (1968)

Physical properties

The simple alkyl PANs are relatively volatile, but are unstable in their pure form (Stephens 1969) and so have rarely been isolated in their condensed form. The chief reason for the instability of PANs is that they thermally decompose at signiﬁcant rates at moderate temperatures (see below), the products then promote explosive conditions. Consequently, there have been only a few studies of the physical properties; melting point, boiling point, of PAN-type compounds. Table 6.2 lists the compounds and properties that have been determined to date. The structures of PAN and some PAN analogues have been modelled using molecular orbital calculations (Francisco and Li 2004; Hermann et al. 2001). The structures arrived at by the two studies are very similar. The plane deﬁned by the carbonyl carbon and the oxygens bonded to it is staggered with respect to the three constituents on the other carbon. The atoms comprising the peroxynitrates group are nearly co-planar, and that plane lies almost perpendicular to the carbonyl plane.

6.1.3 6.1.3.1

Spectroscopic data Infrared

The infrared spectroscopy of PAN and its homologues has a long and storied history, being as it was, the means by which PAN was ﬁrst identiﬁed (Stephens et al. 1956). The IR spectrum contains a wealth of information, with prominent features arising from the car−NO2 symmetric (1 287–1 307 cm−1 ) and asymmetric bonyl stretch (1 799–1 910 cm−1 ), − −1 −O bond stretch (1 163 cm−1 ), and the O− −N stretch (1 696–1 760 cm ) stretches, the C− −1 (781–799 cm ). These features have been used as the sole means of identifying a number of the compounds listed in Table 6.1. Careful work has permitted IR absorption cross-sections to be used quantitatively as a means to measure PAN in the atmosphere directly or to calibrate PAN samples for other analysis methods. Allen et al. (2005) have recently re-measured the mid-infrared absorption cross-sections of PAN and compared them with previous measurements, Table 6.3. There is reasonable agreement between the studies, and the most recent one (Allen et al. 2005) indicated that acetone impurities may explain some of the discrepancies. Absorption cross-sections or integrated band strengths have been measured for other PAN-type compounds. Stephens (1964) also reported infrared absorptivities for PPN and

239 ± 4 241 ± 3

247 ± 6

Intg2

14.5 ± 0.7 15.7 ± 0.3 14.6 ± 0.5

14.3 15.8

Abs

1163

322 ± 7 356 ± 4

477 ± 9

Intg 11.2 13.6

Abs

Intg

270 ± 2 281 ± 3

405 ± 20

1302

11.3 ± 0.6 11.9 ± 0.2 11.4 ± 0.4

1 Absorptivity (×10−1 ), ppm−1 m−1 , base 10 units, 1 atm. 2 Integrated band intensity, atm−1 cm−2 .

11.5 ± 0.6 12.2 ± 0.2 11.4 ± 0.4

10.1 13.4

Abs1

794

23.6 32.6

Abs

Intg

563 ± 10 537 ± 5

808 ± 34

1741

31.0 ± 1.6 31.4 ± 0.8 30.2 ± 1.5

Band centre position, cm−1

Table 6.3 Infrared absorption constants for PAN

10.2 ± 0.5 10.9 ± 0.2 9.5 ± 0.6

10 12.4

Abs

1842

262 ± 3 260 ± 3

322 ± 9

Intg Stephens (1964) Bruckmann and Willner (1983) Gaffney et al. (1984) Niki et al. (1985) Tsalkani and Toupance (1989) Allen et al. (2005)

Reference

228

Volatile Organic Compounds in the Atmosphere

PnBN. Rogers and Rhead (1987) reported absorption coefﬁcients for peroxyglutaric nitric anhydride. Sulbaek Anderson et al. (2003) reported an integrated band strength for the NO2 asymmetric stretch in C2 F5 (O)OONO2 that was very close to that observed in PAN. −H absorption bands of PAN have been Recent measurements of the near-IR overtone C− reported in the regions 5 600–6 200 cm−1 (2ν) and 8 300–8 900 cm−1 (3ν) (Nizkorodov et al. 2005). The absorption cross-sections were 9.6% and 0.82% of the fundamental absorption, respectively. These measurements were used, along with quantum yield estimates to conclude that PAN does not undergo signiﬁcant photolysis from this process in the atmosphere.

6.1.3.2

UV-visible absorption

The absorption of PAN compounds in the UV-visible region has been used for quantiﬁcation of pure PAN samples (Talukdar et al. 1995) and provides a slow but important loss process in the upper atmosphere. The absorption spectrum of the simple PAN-type compounds thus far reported are all quite similar; a broad exponential decay between 200 and 340 nm, with a slight inﬂection at about 260 nm (Figure 6.2). This structure has been assigned to three different broadband electronic transitions in the molecular orbital analysis presented by Francisco and Li (2004). The two independent measurements of the PAN spectrum shown in Figure 6.2 are in good agreement, and the differences among the three compounds are not large, but will result in signiﬁcant differences in atmospheric lifetimes in the upper troposphere (see below). The UV absorption cross-sections of PAN (Talukdar et al. 1995) and PPN (Harwood et al. 2003) have been observed to decrease with temperature, roughly a factor of two in the actinic region, over the range 253–296◦ K.

Cross section (cm2 / molecule (×1020))

100

10

1

0.1 PAN Talukdar et al. (1995) PAN Libuda and Zabel (1995) PPN Harwood et al. (2003) FPAN Libuda and Zabel (1995)

0.01

200

220

240

260 280 300 Wavelength (nm)

320

Figure 6.2 UV-visible absorption cross-section of PAN, PPN and FPAN.

340

360

PAN and Related Compounds

229

Table 6.4 Quantum yields of NO3 from PAN and PPN photolysis Compound PAN PAN PAN PAN PPN PPN

Wavelength, nm

Yield

Reference

248 248 289 308 248 308

0.3 ± 0.1 0.19 ± 0.04 0.31 ± 0.08 0.41 ± 0.10 0.22 ± 0.04 0.39 ± 0.04

Mazely et al. (1997) Harwood et al. (2003) Flowers et al. (2005) Harwood et al. (2003) Harwood et al. (2003) Harwood et al. (2003)

The mechanism of photodissociation of PAN and PPN has been the subject of recent investigation. There are two possible channels at wavelengths >290 nm: CH3 C(O)OONO2 + hν → CH3 C(O)OO + NO2

(1)

CH3 C(O)OONO2 + hν → CH3 C(O)O + NO3

(2)

The relative importance of the two channels can have a subtle effect on the lifetime of PANs in the upper troposphere since Reaction 1 can be followed by the reverse thermal reaction to reform PAN, while Reaction 2 is followed by the immediate decomposition reaction: CH3 C(O)O + M → CH3 + CO2 + M

(3)

and irreversible loss of PAN. There have been three studies of the quantum yield of Reaction 2 for PAN and one for PPN, and the results are given in Table 6.4. The yield of Reaction 2 increases with increasing wavelength up to 0.4 in the actinic region, and the yields for PAN and PPN are the same within experimental uncertainties. PAN lifetimes were estimated to decrease 5–10% at midday in the upper troposphere as a result of this process (Harwood et al. 2003).

6.2

Atmospheric formation

The formation of PAN compounds in the atmosphere occurs through a series of reactions initiated by the production of an acyl radical, RC(O). Acyl radicals are formed primarily by reactions of aldehydes with OH radicals, but can also be produced by the photolysis of ketones or dicarbonyls (see Figure 6.1). Simple alkyl acyl radicals can decompose or react with O2 to give peroxyacyl radicals: RC(O) + M → R + C(O) + M RC(O) + O2 + M → RC(O)OO + M

(4) (5a)

with the reaction with O2 being favoured overwhelmingly in the atmosphere (Tomas et al. 2000 and references therein). There are several exceptions to this scheme. Formyl radical, HC(O), reacts with O2 to give CO and HO2 , thus accounting for the absence of peroxyformic nitric anhydride among the known PAN compounds. The carbonyl acyl radicals HC(O)C(O), CH3 C(O)C(O), CHCl2 C(O) and CCl3 C(O) are known to decompose either

230

Volatile Organic Compounds in the Atmosphere

partially or completely, rather than form peroxyacyl radicals (Mereau et al. 2001; Orlando and Tyndall 2001; Tyndall et al. 1995). Reaction 5 is known to have a low-pressure channel that involves the formation of OH radical (Tyndall et al. 1995, 1997), the direct production of which has been observed for acetyl and propionyl radicals (Baeza Romero et al. 2005; Blitz et al. 2002): RC(O) + O2 → OH + Products

(5b)

The importance of (5b) is limited to low pressures (P < 60 torr); however, lower temperatures favour Reaction 5a. Thus, although the reaction has not been studied for larger R groups, it is doubtful that the side Reaction 5b has signiﬁcance in the atmosphere. Peroxyacyl radicals are reactive with a number of species, including NO, NO2 , HO2 and other peroxy radicals (Lightfoot et al. 1992; Tyndall et al. 2001a). The reaction of RC(O)OO radicals with NO2 makes PAN compounds: RC(O)OO + NO2 + M → RC(O)OONO2 + M

(6)

The reactions of RC(O)OO with the other species lead to irreversible loss, for example, the reaction with NO RC(O)OO + NO → RC(O)O + NO2

(7)

Produces acyloxy radicals, which rapidly decompose RC(O)O + M → R + CO2 + M

(8)

The variety of possible sources of acyl radicals accounts for the different PAN compounds that are formed from the complex VOC chemistry that occurs in the troposphere. The reaction of aldehydes with OH RC(O)H + OH → RC(O) + H2 O

(9)

is one of the most common source of acyl radicals and, therefore, PANs. These reactions lead to a correspondence between the stable precursor and the stable PAN product. In contrast, there are several acyl radical sources such as the decomposition of CH3 C(O)C(O), and photolysis of methyl glyoxal and acetone, in which the product alkyl group is different from the precursor. Photolysis of α-dicarbonyls, RC(O)C(O)R is the atmospheric pathway to acyl radicals and, therefore, PANs. These compounds have reasonably strong absorptions in the region around 410 nm, with quantum efﬁciencies of 10–15% (Plum et al. 1983). Stephens (1969) describes biacetyl photolysis as a gas-phase synthetic route to making PAN. The photolysis of acetone, CH3 C(O)CH3 at wavelengths in the near UV results in the production of acetyl radical: CH3 C(O)CH3 + hν → CH3 C(O) + CH3

(10)

This can be an important source of PA radicals in the remote troposphere since acetone is among one of the longer-lived oxygenated VOCs. Acetone photolysis has become a common method of online production of PAN for calibration purposes (Meyrahn et al. 1987).

PAN and Related Compounds

231

16 14

PAN-OH reaction PAN photolysis

Altitude (km)

12

PPN photolysis

10 8 PAN thermal decomposition MPAN-OH reaction

6 4 2 0 10−9

PAN deposition 10−8

10−7 10−6 Loss rate (s−1)

10−5

10−4

Figure 6.3 Loss rates for PAN and selected PAN analogues.

6.2.1

Atmospheric losses

The loss of PANs from the atmosphere occurs through a combination of thermal decomposition, photolysis, deposition to vegetation and natural surfaces and reaction with atmospheric radicals, OH and NO3 . The relative importance of each process depends on the region of the atmosphere and the structure of the particular PAN compound being considered (Figure 6.3). Thermal decomposition is a major loss process for PANs in the lower atmosphere. Kirchner et al. (1999) have reviewed experimental measurements with the goal of achieving a systematic understanding of peroxynitrate thermal decomposition. They used the 13 C −NO2 bond strength, reasoning NMR shift of the adjacent carbon as a surrogate for the OO− that they both depend on the electron density surrounding that carbon atom. This correlation was most successful for simple hydrocarbon and oxygenated hydrocarbon side chains and least successful for halogenated formyl and acetyl peroxynitrates. The mechanism of thermal decomposition of peroxycarboxylic nitric anhydrides has been the subject of a number of experimental and theoretical studies, primarily on PAN (Miller et al. 1999, and references therein) (Table 6.5). The evidence suggests that CH3 C(O)OONO2 + M → CH3 C(O)OO + NO2 + M

(6)

is the sole pathway for this reaction. A molecular orbital study has shown that the similar bond scission channel CH3 C(O)OONO2 + M → CH3 C(O)O + NO3 + M

(11)

is nearly iso-energetic, and concluded that there must be some other barrier to dissociation from that channel. To date, there are not detailed studies of the dissociation pathways of other PAN compounds.

232

Volatile Organic Compounds in the Atmosphere

Table 6.5 Arrhenius parameters for the thermal decomposition of PANs, RC(O)OONO2 R Group

log A, 1016 , s−1

Ea , Kcal/mole

k298 K,1 atm

Reference

16.8 18.6 ± 1.6 16.4 15.7 16.7 16.4 16.3 16.8c 16.8c 16.8c 16.8e 20.6 ± 0.8 16.9 16.2 ± 0.7 23.2 ± 1.8

28.1 27.7 ± 2.3 27.5 ± 0.8 25.6 ± 1.1 26.7 ± 1.1 27.0 ± 0.5 27.7 ± 1.0 28.4 ± 1.2d 28.3 ± 1.2d 28.2 ± 1.2d 28.4 ± 1.2d 32.8 ± 1.1 27.7 ± 0.5 26.8 ± 1.0 36.6 ± 2.4

18.2 ± 5.2 18.5 ± 3.4 12.1 ± 3.1 19.4 ± 6.4 15.3 16.9 16.5 9.25 ± 0.33

29.8 ± 7.0 30.1 ± 4.6 21.6 ± 4.1 31.4 ± 8.5 26.7 27.8 26.9 ± 1.8 17.2 ± 0.5

1.43 × 10−4 1.9 × 10−2 1.7 × 10−4 8.4 × 10−4 1.1 × 10−3 4.1 × 10−4 7.5 × 10−5 8.0 × 10−5 1.0 × 10−4 1.2 × 10−4 0.8 × 10−4 3.5 × 10−4 3.5 × 10−4 3.5 × 10−4 2.3 × 10−4 1.73 × 10−4 a 2.2 × 10−4 2.7 × 10−4 1.8 × 10−4 2.4 × 10−4 5.2 × 10−5 3.1 × 10−4 5.3 × 10−4 4.3 × 10−4

Wallington et al. (1995) Spence et al. (1978) Kirchner et al. (1999) Kirchner et al. (1997) Kirchner et al. (1997) Bridier et al. (1991) Wallington et al. (1994) Zabel et al. (1994) Zabel et al. (1994) Zabel et al. (1994) Zabel et al. (1994) Grosjean et al. (1994c) Kirchner et al. (1999) Roberts and Bertman (1992) Grosjean et al. (1994b) Grosjean et al. (1994b) Grosjean et al. (1994b) Grosjean et al. (1994b) Grosjean et al. (1994a) Grosjean et al. (1994a) van Swieten et al. (1978)b Kirchner et al. (1999) Kirchner et al. (1999) Noziere and Barnes (1998)

FC(O) ClC(O) CH3 OC(O) t -C4 H9 OC(O) CH3 C(O) CF3 C(O) CF2 ClC(O) CFCl2 C(O) Cl3 C(O) CH2 CHC(O) C2 H5 C(O) CH2 C(CH3 )C(O) CH3 CHCHC(O) CH3 CH(CH3 )C(O) CH3 CH2 CH2 C(O) n-C4 H9 C(O) i-C4 H9 C(O) n-C5 H11 C(O) C6 H5 C(O) C6 H5 OC(O) α-P-PAN a Measured at 293.2 K. b In cyclohexane solvent. c Assumed value.

d Calculated from the kinetic data and assumed A factor. e Measured at 800 mbar.

It should be noted that the rate given by Reaction 6 is an upper limit to the actual loss rate in the atmosphere. This is because PAN can be reformed by Reaction 6 in the presence of NO2 . Net loss of PAN only occurs when the peroxyacyl radical is irreversibly lost through reaction or deposition. This loss usually occurs through Reaction 7. In areas where there is sufﬁcient NOx (>100 pptv), a simple correction factor, involving [NO], [NO2 ] and the known rate constants, can be calculated: {1 − (k6 [NO2 ]/(k7 [NO] + k6 [NO2 ]))}

(12)

The effect of this correct factor is to extend the lifetime of PAN, especially in the polluted boundary layer, For example, an [NO2 ]/[NO] ratio of 5, results in a net loss rate of PAN that is only 30% of that predicted by Reaction 6. An interesting possibility was raised by Stephens (1969), who proposed a cyclic decomposition pathway in part to explain the consistent observations of methyl nitrate as a major

PAN and Related Compounds

233

impurity in PAN samples. Early work purported to measure the rate of this process to be several hundred times slower than Reaction 6 (Senum et al. 1986); however, several careful systematic studies of PAN thermal decomposition as a function of the amount O2 , NO2 and NO present have shown this cyclic channel to be at least several thousand times slower than Reaction 6, if it occurs at all (Orlando et al. 1992; Roumelis and Glavas 1992a). The key to explaining the appearance of methyl nitrate in PAN experiments is in understanding that Reaction 8 for PAN CH3 C(O)O + M → CH3 + CO2 + M

(3)

results in the production of methyl radical CH3 , the subsequent reaction of which will form methyl nitrate CH3 + O2 + M → CH3 O2 + M

(12)

CH3 O2 + NO → CH3 O + NO2

(13)

CH3 O + NO2 → CH3 ONO2

(14)

Indeed, one would predict that the Cn–1 alkyl nitrate would be formed to some extent in any system where the Cn PAN compound is present. Only one study of thermal decomposition has been reported for a PAN compound in solution, peroxy-n-hexanoic nitric anhydride (van Swieten et al. 1978). The rate was slightly slower than for the shorter-chain linear alkyl PANs in the gas phase, implying that the solution-phase decomposition of PANs may take place via the same mechanism. This could have some bearing on the solution-phase chemistry of PAN, discussed below.

6.2.2

Photolysis

Photolysis rate of PANs in the atmosphere can be calculated from the UV absorption spectra and the known overall quantum efﬁciency of 1.0. Figure 6.3 shows that results of such a calculation presented by Talukdar et al. (1995) for conditions characteristic of mid-summer 30◦ N latitude. Note that the increased solar ﬂux with altitude is offset by the temperature dependence of the UV absorption cross-section, resulting in a loss rate that is virtually independent of altitude. The loss of PPN due to photolysis is also shown in Figure 6.3, as estimated from the temperature-dependent absorption data of Harwood et al. (2003) and the calculation for PAN presented by Talukdar et al. (1995).

6.2.3

Reactions with radical species

The reactions of PAN, and several PAN analogues, with trace radical species, OH, NO3 and Cl have been examined as possible atmospheric loss processes. Understandably, PAN has received the most attention, since it is usually the most abundant in the atmosphere. However, there have also been studies of PPN and MPAN reactions. The results of studies reported to date are listed in Table 6.6. The reaction of PAN with OH is at least an order of magnitude too low to be competitive with the other major loss process (Figure 6.3). The reaction of OH with PPN has not been reported but is certainly not fast enough to compete with thermal decomposition in the lower troposphere and not likely to be as fast

234

Volatile Organic Compounds in the Atmosphere

Table 6.6 Rate constants of the reactions of OH, Cl, and NO3 with PAN, PPN and MPANa Compound PAN PPN MPAN

kOH <3 × 10−14

kCl b

3.2 ± 0.8 × 10−11

kNO3 c

<7 × 10−15 c 1.14 + / − 0.12 × 10−12 d

1.45 ± 0.45 × 10−16

e

a Units of cm3 /molecule/s. b Talukdar et al. (1995).

c Wallington et al. (1994). d Orlando et al. (2002). e Canosa-Mas et al. (1999).

as photolysis at higher altitudes. Reaction of OH with MPAN is fast enough to be a major loss process in the lower troposphere. Figure 6.3 shows the loss rate estimate for MPAN under the conditions given by Talukdar, which were a diurnally averaged tropospheric OH of at most 1.2×106 molec/cm3 . Peak OH can be over an order of magnitude larger than that in the polluted boundary layer, so this process can rival thermal decomposition. Reactions of Cl atoms with PAN and PPN have been found to be quite slow relative to other Cl atom reactions, and should be unimportant in the troposphere where Cl atom concentrations are probably not larger than a few 104 molec/cm3 (Wallingon et al. 1990). The Cl atom reaction with MPAN is undoubtedly very fast, perhaps approaching the gas kinetic limit; however the abundance of Cl atom relative to OH will make this reaction unimportant as a loss for MPAN. Studies of the reactions of NO3 with PANs have been limited to the unsaturated MPAN, since the simple alkyl PANs will not react with NO3 at signiﬁcant rates. The rate constant for NO3 + MPAN, while not large relative to other NO3 rate constants, could be signiﬁcant under some circumstances. This is because NO3 is a night-time oxidant that is present only when NO2 is abundant but NO is absent, the same conditions that preserve PANs by limiting their effective thermal decomposition rates as noted above. Moreover, substantial NO3 mixing ratios (up to 100 pptv) can be present in the night-time polluted boundary layer (Brown et al. 2003).

6.2.4

Wet and dry deposition

The deposition of PAN and PAN homologues through wet and dry processes has been explored both by laboratory studies of fundamental processes and by ﬁeld measurements of apparent deposition rates. Solubility of PANs in water is the main property that governs the wet deposition of these species. The Henry’s Law equilibrium solubilities of PAN and some PAN homologues have been measured for water, aqueous solutions of some acids and salts and n-octanol. The results of those studies are listed in Table 6.7. Studies in water and aqueous solutions of acid have shown the Henry’s law coefﬁcients (H) of PANs to be low, 3–10 M/atm at temperatures characteristic of the lower troposphere. Also, the highercarbon-number homologues show a decrease in H with an increase in chain length. The low-temperature studies of PAN solubility in high-concentration H2 SO4 Zhang and Leu

PAN and Related Compounds

235

Table 6.7 Henry’s law solubility constants for PAN and PAN homologues Compound PAN

PPN MPAN PiBN PnBN

Temp, ◦ K

Solvent

H, M/atm−1

Reference

283.4 285 283 293 279 293 293 243 208 222 273 279 285 291 298 293.2 293.2 293.2 293.2

Water Water Water Water Water Water 5 M H2 SO4 5 M H2 SO4 72 wt% H2 SO4 72 wt% H2 SO4 n-Octanol n-Octanol n-Octanol n-Octanol n-Octanol Water Water Water Water

3.6 ± 0.2 8.3 ± 0.5 5±1 2.9 ± 0.06 10.3 ± 0.4 3.8 ± 0.2 2.8 ± 0.3 73 ± 4 5.6 × 104 3.1 × 103 110 ± 7.5 84 ± 7.1 61 ± 3.1 42 ± 2.1 34 ± 1.7 2.9 ± 0.06 1.7 ± 0.16 1.0 ± 0.03 2.3 ± 0.06

Lee (1984) Lee (1984) Holdren et al. (1984) Kames and Schurath (1995) Frenzel et al. (2000) Frenzel et al. (2000) Frenzel et al. (2000) Frenzel et al. (2000) Zhang and Leu (1997) Zhang and Leu (1997) Roberts (2005) Roberts (2005) Roberts (2005) Roberts (2005) Roberts (2005) Kames and Schurath (1995) Kames and Schurath (1995) Kames and Schurath (1995) Kames and Schurath (1995)

(1997) have shown much higher H’s. One study of PAN solubility in a model organic solvent, n-octanol, observed H coefﬁcients up to 110 M/atm at 273◦ K. In addition to the Henry’s law solubility, many of the studies also measured the ﬁrstorder loss rate of PAN in solution, another important parameter in estimating possible ﬂux of PAN into surface water or aerosol solutions. The measurements, listed in Table 6.8, show a range of values, from quite low for n-octanol to higher values for ﬁltered seawater. Roberts (2005) has discussed recently several features of the ﬁrst-order loss rates of PAN in water and n-octanol (Table 6.9). The aqueous solution rate constants are all higher than the gas-phase thermal decomposition rate, but trend towards that rate at the higher temperatures (see Roberts 2005 for a summary). The gas-phase thermal decomposition rate is used as a benchmark because the mechanism is a simple bond homolysis, so the rate should be similar in solution (indeed the rate determination for peroxy-n-hexanoic nitric anhydride in cyclohexane is similar to the gas-phase rates of other PANs). Also, since the reaction of PA radical with water is thought to be fast (Villalta et al. 1996), the thermal decomposition should act as a lower limit to the ﬁrst-order loss in solution. Several mechanisms may contribute to the hydrolysis of PAN, a direct reaction has been proposed (Kames and Schurath 1995), and there is evidence of a base-catalysed reaction (Lee 1984). In contrast, the ﬁrst-order loss in n-octanol is slower than the thermal decomposition rate, implying that the reactions of PAN and PA radical with n-octanol are slow (Roberts 2005). There have been fewer studies of the uptake of PAN on ice. The only reported laboratory study measured the adsorption of PAN on ice at temperatures between 218 and 250◦ K and derived an enthalpy of adsorption (Bartels-Rausch et al. 2002). Equilibrium estimates made

236

Volatile Organic Compounds in the Atmosphere

Table 6.8 Hydrolysis or ﬁrst-order loss rates of PAN in solution Temp. ◦ K 278 279 279 279 279 288 293 293 293.2 293.2 293.2 295 295 298 291 298

Solvent

Loss rate, s−1

Water Water Phosphate buffer,a pH 7.5 0.01 M NaNO2 + Phosphate buffer 5 M H2 SO4 Water Water 5 M H2 SO4 Water Synthetic seawater Filtered North seawater Water, pH 4.2–5.6 Water, pH 10 Water n-Octanol n-Octanol

1.2 ± 0.05 × 10−4 1.2 ± 0.1 × 10−4 3.1 ± 0.8 × 10−4 4.2 ± 0.5 × 10−4

Holdren et al. (1984) Frenzel et al. (2000) Frenzel et al. (2000) Frenzel et al. (2000)

0.6 ± 0.3 × 10−4 2.2 ± 0.05 × 10−4 3.2 ± 0.2 × 10−4 3.2 ± 0.7 × 10−4 3.4 ± 0.14 × 10−4 3.7 ± 0.4 × 10−4 44 ± 7 × 10−4 4.0 ± 1.2 × 10−4 1.6 × 10−2 6.8 ± 1.7 × 10−4 0.32 ± 0.25 × 10−5 5.2 ± 1.9 × 10−5

Frenzel et al. (2000) Holdren et al. (1984) Frenzel et al. (2000) Frenzel et al. (2000) Kames and Schurath (1995) Kames and Schurath (1995) Kames and Schurath (1995) Lee (1984) Lee (1984) Holdren et al. (1984) Roberts (2005) Roberts (2005)

Reference

a 0.033 M Na HPO /KH PO . 2 4 2 4

Table 6.9 Hydrolysis or ﬁrst-order loss rates of PAN homologues in water Compound PAN PPN MPAN PiBN PnBN

Temperature

Loss rate, s−1

Reference

293.2 293.2 293.2 293.2 293.2

3.1 ± 0.22 × 10−4 3.9 ± 0.2 × 10−4 5.6 ± 1.2 × 10−4 7.4 ± 0.5 × 10−4 2.4 ± 0.4 × 10−4

Kames and Schurath (1995) Kames and Schurath (1995) Kames and Schurath (1995) Kames and Schurath (1995) Kames and Schurath (1995)

from those measurements predict a relatively minor uptake of PAN on snow (7–8%) and even smaller uptake on ice particles (clouds) in the upper atmosphere (0.01–0.03%). The deposition of PAN on natural surfaces has been examined in a number of controlled experiments with natural materials and in several ﬁeld situations. Conditions have ranged from summertime vegetated ecosystems to wintertime arctic environments. There have been several controlled studies of PAN deposition to natural surfaces such as grass and soil, and fresh water (Dollard et al. 1990; Garland and Penkett 1976) that have found deposition velocities of 0.25 cm/s or less. Three studies of PAN deposition have been conducted at night, one in a corn ﬁeld found deposition velocities of 0.54 ± 0.94 cm/s (Schrimpf et al. 1996), another in natural forest setting found a deposition velocity of at least 0.5 cm/s (Shepson et al. 1992), and a study conducted at a coastal grassy ﬁeld site found night-time

PAN and Related Compounds

237

decay of PAN and O3 consistent with deposition (McFadyen and Cape 1999). Studies with vegetation (Sparks et al. 2003 and references therein) have shown that PAN is taken up through stomates and, therefore, big differences would be expected between day and night. One recent PAN deposition study was conducted in the daytime over a grassland surface and found an average deposition velocity of 0.13 ± 0.13 cm/s (Doskey et al. 2004). An eddy covariance study over a Loblolly Pine forest, using a CIMS technique (see below), found signiﬁcant uptake of PAN, PPN and MPAN by both stomatal and non-stomatal processes, which accounted for roughly 20% of the co-measured NOy ﬂux (Turnipseed et al. 2006). Farmer et al. (2006) found upward ﬂuxes of ‘total peroxynitrates’ (N2 O5 , ROONO2 and RC(O)OONO2 compounds), over Blodgett Forest in California during the daytime in summer, and attributed this to rapid production within the coniferous forest canopy, a result that has not been reported in any other study. There is a lack of studies of PAN deposition during the daytime over the variety of vegetation, for example, deciduous and coniferous forests, which are needed to fully understand this sink. The range of deposition velocities observed so far translate to time constants of 5–10 h for the loss of PAN due to dry deposition. The ﬂux of PAN to (and from) fog, snow and ice has been the subject of a number of ﬁeld studies. The low solubility of PAN in water implies that wet deposition should be a negligibly low process. Indeed, numerous sets of aircraft-based measurements have found no evidence of cloud water uptake of PAN (see below). An interesting secondary effect has been observed in the presence of fog (Roberts et al. 1996) wherein the loss of PA radical to the fog droplets (Villalta et al. 1996) is faster than the reformation of PAN via PA + NO2 (Reaction 6). The reasons this effect is observed in fog and not clouds are that temperatures are higher at the surface, contact times are longer for air parcels in fog banks and droplet surface areas tend to be higher for fog relative to clouds. The loss of PAN to snow and ice has been examined in several studies (Ford et al. 2002) conducted in Summit, Greenland, near Houghton, Michigan, and Alert, Nunavut. A controlled experiment of PAN uptake and release by new fallen snow showed that PAN was only weakly absorbed by the snow surface, consistent with a low Henry’s Law coefﬁcient and slow or non-existent snow-phase loss processes. Measurements in ambient air and snow pack showed that PAN in the snowpack interstitial air was sometimes two to three times higher than the ambient air above, implying a ﬂux out of the snowpack. In addition, measurement of PAN precursors acetaldehyde and NOx implied that the snowpack could be a source of PAN, although attempts to observe this production were inconclusive. A modelling study of the photochemistry in these arctic environments concluded that; while solubility considerations and direct measurements indicate that PAN uptake on snow is a minor effect, the production rates of PAN implied by measured acetaldehyde and NOx are inconsistent with measured PAN unless an additional large sink for PAN is operative (Dassau et al. 2004).

6.3

Measurement and calibration techniques

The measurement of PAN and PAN analogues has changed a great deal over the 50 or so years that PAN has been known. The ﬁrst detection of PAN was accomplished by IR spectroscopy, a method that while very speciﬁc, lacks the detection limit and high spatial resolution desired in a modern measurement. The advent of gas chromatographic techniques

238

Volatile Organic Compounds in the Atmosphere

provided more portability and permitted point measurements to be made, but increased the requirements for calibration. Early chromatographic techniques utilised packed columns that had issues with chemical activity towards PANs decomposition, and also had limited duty cycles due to water retention and its corresponding response on the detector. Fusedsilica capillary columns solved some of these problems because they are chemically inert and the liquid phases used with these columns have a much lower retention of water, permitting rapid sampling. The inherent sensitivity of the most popular detector, the electron capture detector (ECD), has simpliﬁed sample injection although cryogenic pre-concentration has been used to obtain lower detection limits where deemed necessary. Faster ﬂow rates and, in some cases, cryogenic sample collection have pushed the capillary GC methods to the point where PAN can be measured in less than 30 sec. Chemical ionisation mass spectrometric (CIMS) methods have recently been developed that can provide a 1 s measurement of PAN and PAN homologues, with excellent sensitivity. The measurement of PAN by IR spectroscopy has progressed, from the early days of its ﬁrst detection, through the application of Fourier transform IR spectroscopy to quantifying PAN in urban areas. Measurements require a reasonably long path, hundreds of meters up to a km or so, and have been made either with a light source and spectrometer set-up across an urban area, or with an elaborate folded path cell (Tuazon et al. 1978). The chief limitations of this method are insufﬁcient sensitivity, lack of portability, and questionable speciﬁcity for the different PAN compounds. Gas chromatographic techniques have proven to be the most practical methods for the measurement of PAN and the most common PAN analogues. The ﬁrst application of this method was reported by Darley et al. (1963) and involved the use of packed columns with polar Carbowax phases and electron capture detectors (ECDs). This method was used virtually unchanged for 25 years with the exception of a few minor improvements, most notably, more sensitive and linear ECDs, less active solid supports and system materials, and cryogenic pre-concentration to provide lower detection limits. The polar nature of the Carbowax liquid phase resulted in strong retention of water, which often limited the sample turn-around to some 15–20 min. This limitation was overcome by fairly elaborate columns switching and back-ﬂushing arrangements (Muller and Rudolph 1989). Even with these improvements it was clear that packed-column GC had a number of short-comings. High gas ﬂow rates and the broad peak proﬁles attendant to packed columns limited the sensitivity of the system since the ECD is a concentration-sensitive detector. It was also clear that some decomposition on the column was occurring in the typical packed-column systems used (Roumelis and Glavas 1989). Capillary gas chromatographic columns for PAN analyses were ﬁrst employed in the late 1980s (Roumelis and Glavas 1989; Roberts et al. 1989), greatly improving the GC/ECD analysis of PANs. The lower ﬂow rates and narrower peak proﬁles from wide-bore capillary GC separation, in combination with low temperature ECDs, have resulted in PANs analyses with detection limits of a few pptv in directly injected air samples of a few cm3 . −C4 Sample turn-around times of a few minutes for the analysis of the most common C2 − PANs have been achieved (Flocke et al. 2005). An example of a capillary GC/ECD analysis is shown in Figure 6.4, for a sample taken off the coast of New England during the summer of 2002 (see the project description in de Gouw et al. 2003). This chromatogram was obtained with a 4.5-m-long, 0.53-mm ID capillary column with a 1.0 μm coating of RTX-200 (triﬂuoropropyl silicone), with He carrier gas (15 amb cm3 /min) onto which a

2.0

MPAN rt = 3.52, 0.37 ppbv

PiBN, rt = 3.10, 0.030 ppbv

100

rt = 2.21, 0.20 ppbv

APAN, rt = 1.75 0.010 ppbv

150

PPN

200 × 103

ECD response (μV)

239

PAN, rt = 1.13, 2.28 ppbv

PAN and Related Compounds

2.5

3.0

3.5

4.0

50

0 0

1

2 3 Retention time (min)

4

5

Figure 6.4 Capillary GC/ECD chromatogram of an air sample taken off the coast of New England during the summer of 2002.

1.5 amb cm3 sample was injected. In this case, the air mass had signiﬁcant biogenic HC impact, chieﬂy from isoprene, mixed with anthropogenic pollution. As a consequence, ﬁve PAN compounds were observed, at the retention times and mixing ratios noted. Several recent efforts have focused on modifying the capillary GC separation method to provide a faster measurement for use on aircraft platforms. Marley et al. (2004) have reported the use of wide-bore capillary GC column for the rapid separation of NO2 from PAN and detection by Luminol chemiluminescence. In this case, a 30 ft length of 0.53 mm ID. Capillary coated with 3 μm of methyl silicone was used with ﬂow rates of 40–60 cm3 /min. PAN had a retention time of 24 s. A similar arrangement was reported for the analysis of up to ‘peroxybutryl nitrate’, which from the context can be assumed to be PnBN (Gaffney et al. 1998). In this case, a high ﬂow rate, 155 cm3 /min of He was used, and a sample turn-around time of 1 min was achieved. The fast separation of PANs with a GC/ECD system was demonstrated by Schneiders, Flocke and Roberts (unpublished results 2001). This was accomplished by means of cryogenic sampling with narrow-bore capillary column separation (Figure 6.5). The cryogenic sampling loop was required to obtain a narrow sample band at the head of the column. The separation column was a 0.25 mm ID column. 3.5 m long coated with 0.25 μm RTX-200 at a ﬂow rate of 15 SCCM He. The sample shown in Figure 6.5 is a prepared calibration mixture of the three compounds, PAN, PPN, MPAN, and demonstrates the sample turn-around times of 45 s are possible with this type of system.

240

Volatile Organic Compounds in the Atmosphere

500 × 103

PAN

ECD signal (μV)

400

300

200

PPN MPAN

100 0

10

20 30 40 Retention time (s)

50

60

Figure 6.5 Fast GC chromatogram.

The methods of detection of PANs in GC systems have also undergone considerable development over the last 4 decades. While ECD remains among the most sensitive techniques, several other detectors or variants have also been employed. Negative ion chemical ionisation (NICI) mass spectrometric detection has been used by Tanimoto et al. (1999) to detect PANs with sensitivities similar to those obtained with the best capillary GC/ECD systems. NICI-MS has the additional selectivity to distinguish between PANs and alkyl nitrates. A non-radioactive pulsed discharge ECD has been used by Zedda et al. (1998), along with capillary GC separation, to measure PAN and PPN. The sensitivity of the system was about a factor of 20 lower than state-of-the art GC/ECD systems, necessitating cryogenic enrichment. The primary advantage of this detector is that it is non-radioactive and therefore easier to possess and transport for ﬁeld use. The Luminol-chemiluminescence detector has also been adapted for GC use, starting with the work of Burkhardt et al. (1988). This detector relies on the reaction of oxidants, O3 and NO2 , with a solution of Luminol in front of a photomultiplier system, the details of which are not well understood for NO2 although it had been known that PAN also causes a response in this system (Wendell et al. 1983). The Luminol detection concept was taken a step further by thermal decomposition with chemical ampliﬁcation (Blanchard et al. 1993) in the presence of NO. The aforementioned work Marley et al. (2004) has brought this detection scheme up to modern standards through the use of capillary columns. GC-Luminol systems have detection limits similar to GC/ECD systems, but can suffer from nonlinearity problems analogous to the detection of NO2 with these instruments. There are several continuous techniques for the measurement of total PANs that rely on the selective thermal conversion of PANs to NO2 and sensitive and selective detection of the NO2 . The above-mentioned Luminol detector has been used in the continuous mode

PAN and Related Compounds

241

with a thermal converter, held at 110◦ C, switched in and out of the ﬂow system (Nikitas et al. 1997). It was acknowledged that the system was sensitive to NO and somewhat to CO, because of the chemical ampliﬁcation effect noted above, and ambient measurements were used to correct for those. Not mentioned was that the system will also be sensitive to N2 O5 , which can be substantial in NOx -impacted air masses at night (Brown et al. 2003), and may have a stoichiometry greater than 1. Thus, it is doubtful that this method has the selectivity necessary for broad application to ambient measurements. Another continuous PANs detection scheme has been devised based on thermal decomposition and NO2 detection by laser-induced ﬂuorescence (LIF) (Day et al. 2002). The detection of NO2 by this method has been shown to be selective and reasonably sensitive (Thornton et al. 2003). PANs were thermally decomposed selectively in fused silica tubes held at 200◦ C, and their concentration determined by subtracting the NO2 concentration measured with a room-temperature inlet. This method has been shown to be reasonably sensitive (as low as 30 pptv detection limit, 60 s averaging), with a detection limit and precision that depends on the absolute NO2 signal, since it is a difference method. The technique will also respond to N2 O5 and will, therefore, be prone to interference at night. The measurement of PANs by proton transfer reaction mass spectrometry (PTRMS) has been reported (Hansel and Wisthaler 2000). PTRMS is a general technique applicable to any compound that has a higher proton-afﬁnity than H2 O, which includes most VOCs. In the case of PAN, the ion chemistry is slightly more complicated in that protonated PAN reacts with H2 O to produce protonated peroxyacetic acid and nitric acid. This brings up the immediate observation that peroxyacetic acid will be an interferant in this measurement. In addition, acetone proton-hydrate {CH3 C(O)CH•3 H3 O+ } is also known to interfere at that mass. Hansel and Wisthaler (2000) did not report comparisons of their method with another one because they did not have any zeroing procedure to account for the above effects. Data from a later study (Hansel et al. 2002) showed signiﬁcant interference from peroxyacetic acid, and an average of 20% difference between the PTRMS and GC/ECD methods, with PTRMS values as low as −0.8 ppbv. A later set of PTRMS measurements on board a ship (de Gouw et al. 2003) showed good qualitative agreement with a GC/ECD measurement (r 2 = 0.92), although there was not an independent calibration or background measurement. Provided that the background and interference issues could be dealt with, this method could have some application to the rapid measurement of PANs at mid to high concentrations, since the sensitivity of the technique is somewhat limited (detection limit 70 pptv for 15 s averaging time). A recently developed CIMS method involves the thermal decomposition of PANs to form the corresponding peroxyacyl radical and reaction of the radical with iodide ion, I− , to form the stable carboxylate ion which is detected by mass spectrometry (Slusher et al. 2004). This technique has shown excellent selectivity and sensitivity (sub-ppt detection limits for 10 s averaging times (Swanson et al. unpublished results). The selectivity of I− reactions results in a very low background for the major PAN species, and the lack of need for any other separation presents the possibility of measuring PAN species that would not make it through a GC column. The technique does have interferences from high (10 s of ppbv) NO concentrations, because the peroxy radical is reactive with NO, however, constant standard addition of labelled PAN has essentially eliminated this interference in ambient measurements (Swanson et al. 2004). This technique has shown tremendous promise in the rapid and sensitive measurement of PANs.

242

Volatile Organic Compounds in the Atmosphere

One of the most challenging aspects of PAN measurements is producing a stable calibrated source of PANs having high, known purity. Early work on PAN sources involved photochemical production, usually with ethyl nitrite, condensation and puriﬁcation by fractional cryogenic distillation or column chromatography (Stephens 1969). Quantiﬁcation was done by IR spectroscopy. More modern strategies have focused on low-temperature liquid-phase syntheses in non-polar solvents (Nielsen et al. 1982; Gaffney et al. 1984), which have been found to yield reasonably high (>97%) purity. Furthermore, the use of highmolecular-weight solvents has allowed the preparation of gas streams containing mostly the higher-volatility PANs. Gas streams of PAN samples have been prepared by injection of a small sample of the solution into gas sampling bags or by use of a diffusion cell similar to that described by Williams et al. (2000). One problem that has been noted with the traditional tridecane/PAN mixtures in diffusion cells is that the solution is a liquid at 0◦ C and can produce unreliable emission rates when agitated, an effect that was apparently responsible for the poor results of the CITE II PAN intercomparisons (Gregory et al. 1990). This problem has been eliminated through the use of pentadecane (n-C15 H32 ) in combination with tridecane/PAN mixtures to produce a mixture that is solid at 0◦ C (Williams et al. 2000). A commercial PAN calibration source (Metcon GmbH) is available, which is based on the photolysis of acetone at 285 nm to produce an abundance of PA and other radicals, in the presence of a calibrated amount of an NO standard in a ﬂowing stream (Meyrahn et al. 1987; Warneck and Zerbach 1992). This photochemical scheme has the advantage of using a simple, easy to store, organic compound that produces a fairly clean source of PA radicals and C1 radicals (CH3 and subsequent products). This insures that the added NO will be converted to NO2 and PAN with high yield. Measurements on the commercial source have shown NO to PAN conversion efﬁciencies of 92 ± 3% (Pätz et al. 2002; Volz-Thomas et al. 2002). Modiﬁcations have been made to this commercial source to decrease the amount of Acetone added to the gas stream because the associated impurities in the acetone were found to interfere with the GC/ECD analysis using triﬂuoropropyl-silicone-coated capillary columns (Flocke et al. 2005). A photosource for PPN based on propanal photolysis has been reported by Volz-Thomas et al. (2002), and it is claimed to have essentially a 100% efﬁciency for the conversion of NO to PPN (R. Schmitt, personal communication, 2005); however, it is unclear what the chemical mechanism behind the source is since the yield for propanal photolysis to PP radical is quite small. There have been a number of calibration methods applied to PAN standards over the − years including IR spectroscopy, alkaline hydrolysis with NO− 2 /NO3 analysis, and total odd-nitrogen (NOy ) analysis. IR spectroscopy is not very sensitive, and so must be applied before dilution. This introduces an extra step, and associated possibilities for added error. Alkaline hydrolysis has been found to provide a good calibration, given that the modern PAN syntheses are quite pure (in terms of nitrogen) and the alkaline hydrolysis chemistry of PAN is well established. It has the disadvantages of requiring a fairly long sample collection step, and yields an integrated value for the source output. Calibration by NOy analysis −O3 chemiluminescence (Fahey is accomplished by catalytic conversion to NO and NO− et al. 1985). This is a sensitive and relatively selective method, which yields an online measurement of the NOy coming out of a PAN source. The method also responds to NO2 , HNO3 and RONO2 , so the sample purity must be established by other methods, or by pre-separation. Pre-separation of standard streams has been reported for PAN (Joos et al. 1986), and for some of the difﬁcult to synthesise PAN analogues, such as PiBN and MPAN

PAN and Related Compounds

243

(Williams et al. 2000), and involves separation of the standard stream on a packed or capillary GC column and detection by NOy in parallel with the analytical system being calibrated. The pre-separation GC conditions are usually set up to produce a broad peak so that the injection of the analytical GC can be timed to coincide with the plateau. The concentration of the analyte is then determined from the simultaneous NOy measurement and the relevant system ﬂow rates.

6.4

Atmospheric measurements

Ozone PAN PPN PiBN MPAN

TexAQS 2000 La Porte

Ozone (ppbv)

150

7 000

1 400

6 000

1 200

5 000

1 000

4 000 100

(PAN), pptv

200

800

3 000

600

2 000

400

1 000

200

0

0

(PPN), (PiBN), and (MPAN), pptv

Measurements of PAN date back to the late 1950s in Los Angeles (Stephens 1969). It is astounding to realise that measurable levels were observed with a method that had a detection limit of 50 ppbv. The highest levels of PANs are still observed in urban areas, where the abundance of VOC and NO sources combine to produce high photochemical O3 and PANs. The trends of PAN and O3 parallel each other in urban areas, with higher concentrations in the summertime photochemical pollution season (Roberts 1990). This trend also holds for individual days, as shown in Figure 6.6, which is a plot from the 2000 Texas Air Quality Study (TexAQS) (Roberts et al. 2001b). The data were taken at the La Porte, Texas supersite during the most intense pollution observed. The correlation between O3 and the PANs is quite tight for the ﬁrst part of the pollution episode, that is, up until 1 500 CST. However,

50

0 06:00 8/30/00

12:00

18:00

00:00 8/31/00

Time (CST)

Figure 6.6 The trend of O3 and PANs observed on 30 August 2000, during TexAQS 2000, at La Porte, TX.

244

Volatile Organic Compounds in the Atmosphere

the O3 levels continued to increase as the afternoon progressed and the PANs levels dropped slightly. This is because O3 has a much longer lifetime than PANs in the warm summertime surface layer. Thus, PANs tend to follow the photochemical production rates of radicals while O3 is more of an integral measure of photochemical production. These features are common in urban surface measurements (Glavas and Moschonas 2001; Rappengluck et al. 2000; Rubio et al. 2004). Aircraft measurements of urban plumes have observed layers of high PAN and O3 consistent with relatively rapid production within the plume (Corsmeier et al. 2002). The budget of NOy for urban areas has been examined in a few studies. Of interest here is the fraction of NOy that is represented by PAN. Measurements in Clarement, CA, in 1980 resulted in a ratio of PAN/NOy that averaged 12% (Grosjean 1983). More recent measurements of PAN and NOy in Nashville, TN, showed a diurnal cycle in PAN/NOy that varied from an average of about 3% at night to 15% at midday, with an overall average of 10% (Roberts et al. 2002). The relative abundances of the PANs species have interesting implications for the VOC portion of the VOC-NOx photochemistry (Williams et al. 1997). Urban VOC sources are typically dominated by anthropogenic sources (AHC), which are sources of PAN, but virtually the only source of the higher alkyl PANs, PPN, and PiBN. Conversely, there are urban areas such as Atlanta, GA, Nashville, TN, Boston, MA, that also have signiﬁcant biogenic hydrocarbon (BHC) sources, chieﬂy isoprene, which is a source of PAN and MPAN (Roberts et al. 1998, 2002; Warneke 2004; Williams et al. 1993). The ratio of PPN to PAN in urban areas is a good indicator of the VOCs that contributed to the VOC-NOx photochemistry, with ratios of 15–20% characteristic of AHC-dominated photochemistry (Grosjean et al. 1996; Roberts et al. 1998, 2002). Contributions of BHC chemistry tend to form PAN and MPAN and lower the PPN/PAN ratio (Ridley 1991). MPAN/PAN ratios as high as 25% have been observed in an urban area signiﬁcantly impacted by isoprene sources (Roberts et al. 2002). Much less is known about other PAN species in urban areas because of the much lower concentrations typical of those compounds and the limits of analytical methods that have been employed. PiBN and PnBN, for example, have been observed in a few studies in urban area (Gaffney et al. 1999; Roberts et al. 2002, 2003), and generally have concentration of 3% or less of PAN. These species are formed from longer-chain and 2-methyl alkanes, and there is some evidence from the TexAQS study that anomalously high sources of precursors, such as isobutane, could be related to unusually high PiBN/PAN ratios observed at times in that study. APAN was also observed in TexAQS at unusually high concentrations due to the local petrochemical sources of 1,3-butadiene and acrolein (Roberts et al. 2001b). A number of studies have focused on the atmospheric chemistry occurring in wider regions impacted by urban and continental sources of VOCs and NOx . The mixing ratios of PAN that have been observed have ranged from below 5 pptv to upwards of 5 ppbv. It is clear that PANs are a signiﬁcant fraction of NOy present in such regionally impacted air masses. A summary of NOy and speciated NOy measurements at North American sites has shown the fraction PAN/NOy to have a median of 15–25% (Parrish et al. 1993; Thornberry et al. 2001). Another study in the Vancouver and the lower Fraser Valley, B.C. showed average PAN/NOy values <5% (Hayden et al. 2004). A set of PAN and NOy measurements at Schauinsland, Germany, has shown that PAN/NOy can be as high as 50% in air masses transported up to the site from Freiberg (Pätz et al. 2000). Several competing factors are at play in determining the PAN/NOy ratio at the surface. Production rates are limited

PAN and Related Compounds

245

primarily by the abundance and reactivity of VOCs, since NOx is usually in fairly high abundance (>1 ppbv) in regionally impacted air masses. Loss rates are generally controlled by temperature in the daytime and by deposition at night. The dynamics of the planetary boundary layer affect both production and loss in that the summertime, daytime PBL tends to be mixed on about a 0.5 h timescale, effectively integrating production and loss over the 1–1.5 km depths characteristic of the PBL. Conversely, stable layers close to the ground during night-time limit mixing and can result in PANs concentrations of essentially zero due to deposition or thermal decomposition in combination with soil NO emissions. The relative abundances of PAN compounds at regionally representative sites have been determined mostly for PAN and PPN, and more recently for MPAN, APAN and some of the butyl PANs. The relative abundances of the PANs can yield information in regards to the relative importance of AHC and BHC in the production of ozone. The relative concentrations of PPN and PAN have been observed in numerous studies of regional pollution (Kourtidis et al. 1993; Nouaime et al. 1998; Pippin et al. 2001; Ridley et al. 1990; Roberts et al. 1998; Schmidt et al. 1998; Singh and Salas 1989; Tanimoto 2000; Williams and Grosjean 1991; Williams et al. 1997). The systematics of these observed PPN/PAN ratios reﬂect the fact that AHC sources produce ratios in the range of 12–20%, while BHC chemistry, chieﬂy isoprene, produces PAN but not PPN, thus lowering observed PPN/PAN in those air masses. The different loss processes affect PPN/PAN in opposite ways; thermal decomposition is about 15% faster for PAN and so will increase PPN/PAN, while photolysis of PPN (important for very cold environments or high altitude) is about twice as fast as PAN, so will decrease PPN/PAN. The relative abundance of MPAN and PAN is affected by the presence of active isoprene chemistry and the fact that MPAN has an additional loss due to reaction with OH radical (Orlando et al. 2002). Consequently, the ratio of MPAN/PAN ranges from a high of about 25% down. The distance scale of impact at regional sites is quite broad. As mentioned before, local sources of isoprene have been observed to contribute to local PAN (and MPAN) formation, particularly in the Eastern and South-eastern North America (Nouaime et al. 1998; Roberts et al. 1998; Williams et al. 1997). Conversely, there have been observations of fairly longrange impact in PAN measurements made at regionally impacted sites (Nielsen et al. 1981; Rappenglück et al. 2004). In these cases, transport in relatively cool layers above the surface layer was suspected. Parrish et al. (2004b) has examined cases where the transport of air out of the PBL to the free troposphere was observed using CO as a tracer and found that the percentage of NOy represented by PAN ranged from 25 to almost 60%. There are only a few sets of PAN measurements at regionally representative sites that extend for periods long enough to detect annual growth trends. The data set from Delft in the Netherlands has been discussed previously (R. Guicherit, as given in Roberts 1990) and shows an increase in PAN of approximately 10% per year in air masses coming from the North Sea, between 1973 and 1985. Another data set from Kemjimkujik National Park in Nova Scotia has shown an increase in PAN of about a factor of 1.9 between 1984 and 1989, corresponding to an annual trend of approximately 14% (Bottenheim et al. 1994). PAN was measured for 5 years on Tenerife in the Atlantic, however, reports on this data set have not mentioned any annual trend (Schmitt and Volz-Thomas 1997). These are intriguing observations in that they imply large-scale changes in the odd-nitrogen chemistry of the Northern Hemisphere. There are other glimpses of such possible changes. For example, Parrish et al. (2004a) note possible increases in PAN in air masses off the west coast of

246

Volatile Organic Compounds in the Atmosphere

California over the period 1985 to 2002. This is a possible signature of the increase in odd-nitrogen in the Northern Paciﬁc due to increases in Asian emissions. Measurements of PANs in remote environments have include both ground and aircraft campaigns. The geographic extent of ground measurements has ranged from the north of the Arctic Circle (Spitsbergen), to the South Pole (Roberts et al. 2007). The general trends in remote PAN concentrations have been summarised by Roberts 1990, and the more recent observations give a more complete picture. Remote Northern Hemisphere measurements show a distinct springtime maximum similar to that ﬁrst reported by Penkett and Brice (1986). This feature is observed in the 5-year data set obtained at Kemjimkujik in Nova Scotia, 44◦ 26’◦ N (Bottenheim et al. 1994) in a nearly 3-year data set at Ny-Alesund, Svalbard, 78◦ 54’N in the European Arctic (Beine and Krognes 2000), and in a 5-year data set from the Izana site on Tenerife (Schmitt and Volz-Thomas 1997). The explanation for this seasonal dependence is that precursors build up somewhat in the Northern Hemisphere winter and are available for PAN production as the springtime rise in photochemical activity happens, but at the same time, temperatures are relatively still low, so PAN loss rates are lower than summertime. This springtime PAN maximum is coincident with a less pronounced ozone maximum, although the causes of that are less clear since springtime tends to be the season of highest Stratosphere–Troposphere exchange. An interesting test of the increased precursor factor would be a long-term set of PAN measurements in the remote Southern Hemisphere; however, none have been reported to date. A survey of PAN measurements at remote surface sites shows a general decreasing trend with latitude. The extremes of this trend can be seen in measurements made in the Arctic and Antarctic during their respective summers, Table 6.10. The Northern Polar Regions appear to be about a factor of 4–5 higher in PAN than the Southern polar region. There Table 6.10 Measurements of PAN in the Arctic and Antarctic during the summer Site

Time period

Ny-Alesund, 78.54◦ N

Summer 94–97

Khatanga, Siberia, 73.8◦ N Summit, Greenland, 72.33◦ N Summit, Greenland, 72.33◦ N Neumayer, Antarctic, 70.39◦ S South Pole, Antarctic, 90◦ S

28 June 1994

Minimuma Maximum

27.5

1 610

Mean Median (std dev.) 92.2 (37.6)

89.4

60

Reference

Beine and Krognes (2000) Jaeschke et al. (1999)

29 June–7 July 1998

∼20

150

52 (19)

Ford et al. (2002)

3–17 July 1999

∼20

150

74 (26)

Ford et al. (2002)

1–28 February 1999

<5

47.9

13.1 (7.3)

11.5

Jacobi et al. (2000)

25 November–26 December 2003

<1

27.3

15.5 (4.3)

15.8

Roberts et al. (2007)

a Mixing ratios in pptv.

PAN and Related Compounds

247

are several other shipboard data sets involving North–South transects of the Atlantic Ocean that can be added to the data in Table 6.10 to ﬁll out the picture. A summertime study in the North Atlantic showed mixing ratios between <1 pptv and 40 pptv with a general N–S gradient between 60◦ and 7◦ N (Gallagher et al. 1990). A study in the Atlantic in September and October, 1989, showed a gradient between 52◦ N and 31◦ S with values in the range of 100–200 pptv at 42–45◦ N down to below detection limit 0.4 pptv between 30◦ N and 30◦ S and a few values between 10 and 100 pptv close to Brazil at 31◦ S (Muller and Rudolph 1992). Transects of the Atlantic Ocean from 25 May to 21 June 1998 (South Africa to Bremerhaven), and 1 March to 16 March 1999 (Nuemayer to South Africa), provide a time series of PAN measurements in the Southern Ocean as well as the mid-latitudes (Jacobi and Schrems 1999; Jacobi et al. 1999). The leg between Antarctica and South Africa showed a mean PAN of 18 pptv below 55◦ S and 62 pptv between 50◦ and 35◦ S, while the cruise between South Africa and Germany showed a gradient with ranges from below detection limit (5 pptv) to about 19 pptv in the south, and tropical, Atlantic to hundreds of pptv in the North Atlantic. There are not comparable transects of the Paciﬁc Ocean. The only one that has been reported, went from Seattle, WA (48◦ N), down to Punta Arenas, Chile (48◦ S, from 11 November to 28 December 1981 (Singh et al. 1986). PAN values ranged from averages of 40–80 pptv at 25–48◦ N to 2–10 pptv from 10◦ N to 48◦ S. Taken as a whole, the data in Table 6.10 and the ship transects give a consistent picture with the highest PAN values in the northern mid to high latitudes, low PAN values in the tropics, due to short thermal lifetimes, and slightly higher PAN values at the southern high latitudes. Aircraft measurements, most of them conducted in the past 15 years, have provided a more complete picture of the global transport of PAN and the importance of PAN as an NOy species. The production of PAN is driven by VOC-NOx photochemistry, hence, is heavily weighted towards the polluted continental boundary layer, although there will be contribution from remote air masses in which NOx is sufﬁciently high (20–50 pptv). Understanding the transport of PAN, therefore, is intimately related to understanding the vertical transport of the lower atmosphere. The export of NOy to the free troposphere has been estimated to be 10–25% of NOx emitted at the surface (Parrish et al. 2004b). The mechanisms for such transport include turbulent diffusion at the top of the CBL, a weak but fairly pervasive process, and frontal passage characterised by ‘warm conveyor belt’ (WCB) transport (Cooper et al. 2001), which is a much more efﬁcient, albeit stochastic, process. There are several examples of this WCB process resulting in efﬁcient vertical and subsequent long-distance transport of PAN. One excellent example was observed during the ‘textbook’ WCB case of 11 September 1997 described by Cooper et al. (2001). PAN measurements made during that same ﬂight (see Williams et al. 2000 for a summary of the 1997 data), illustrated in Figure 6.7 along with the ﬂight path, showed values up to nearly 600 pptv in air masses that had been lifted by the WCB and were being transported easterly out over the North Atlantic. These PAN values were 50% or more of the NOy observed in these air parcels. Another example of the WCB transport mechanism was observed during a study of the springtime transport of Asian emissions conducted off the west coast of North America in 2002. Cooper et al. (2004) discuss several cases of this WCB process on the east coast of Asia that resulted in the long-range transport of well-deﬁned plumes of photochemical products. Two contrasting cases of this transport were examined, a case where the transport took place at high altitude, in which PAN remained high and was the most abundant NOy species, and a case in which the impacted air mass subsided to altitudes low enough that

248

Volatile Organic Compounds in the Atmosphere

Latitude (°N)

55

50

45

−75

−70 −65 −60 Longitude (°W)

−55

0

100 200 300 400 500 600 (PAN), pptv

0

1

2 3 4 5 6 7 Pressure altitude (km)

8

Figure 6.7 Measurements of PAN made on the 11 September 1997 ﬂight during NARE 97. PAN values are coded by colour, and altitude is coded by marker size as indicated by the scale. This image appears in full colour in the plate section which follows page 268 as Plate 5.

PAN could thermally decompose with resulting formation of HNO3 and O3 (Nowak et al. 2004; Roberts et al. 2004). The implications of O3 production in these cases are discussed by Hudman et al. (2004). Similar points about PAN decomposition and the production of ozone in the North Paciﬁc springtime have been made by Kotchenruther et al. (2001). The modelling study of Schultz et al. (1998) implied that the transport of North American emissions to Tenerife was accompanied by signiﬁcant decomposition of PAN, in descending air, to form NOx , an effect that served to keep the photochemical production and loss of O3 roughly in balance. Another pertinent transport process involved convection driven by intense biomass burning. Correlations presented by Lavoué et al. (2000) show that the height at which ﬁre emissions are injected in the troposphere is proportional to ﬁre intensity and that very intense crown ﬁres can inject emissions up to the tropopause. These emissions of course include particles, NOx and VOCs, and oxygenated VOCs (OVOC) in particular, providing the ingredients for the production of PAN and PAN homologues. Elevated PAN has been observed in the free troposphere in air masses impacted by African savannah ﬁres (Singh et al. 1996b). Some of the plumes observed during ITCT 2K2 were impacted by biomass burning as indicated by elevated acetonitrile (de Gouw et al. 2004), and had PAN levels over 600 pptv almost 10 days after initial emissions (Roberts et al. 2004). Several general aspects of aircraft PAN measurements can be gleaned from the various data sets in the literature. The North–South trends observed in the surface data seem to be

PAN and Related Compounds

249

reﬂected in the aircraft data (Rudolph et al. 1987; Talbot et al. 1996; Singh et al. 1996a, 1998; Russo et al. 2003), with the highest levels observed it the NH mid to high latitudes, and the lowest levels over the tropics. The altitude dependence of PAN concentrations generally follows the dependence of lifetime on altitude, the highest concentrations are observed between 5 and 10 km. PAN is generally low in the uppermost troposphere because there are only weak photochemical sources in this environment. In addition, there appears to be fairly low mixing ratios of PAN in the stratosphere 30–90 pptv (Singh et al. 2000), so that Strat-Trop exchange would result in somewhat reduced PAN in the upper troposphere. PAN is generally the most abundant NOy species in the mid to upper troposphere, with some dependence on latitude and air mass origin. Studies have been conducted in environments ranging from the tropical South Atlantic to the high-latitude Northern Hemisphere (Roberts et al. 2004; Singh et al. 1994, 1998; Talbot et al. 1996, 1999, 2003). The general features observed in PAN mixing ratios are mirrored in the PAN/NOy data; boundary layer PAN/NOy tends to range from very low <5% in the warm MBL to up to 20% in air masses impacted by continental emissions, above the boundary layer PAN/NOy can vary from fairly low (20%) to fairly high (65%) fractions of NOy , and the highest altitudes (>7 km) have PAN/NOy ratios that are general above 50%, and sometimes close to 100% in air masses that have experienced convective transport associated with precipitation. The breadth of these studies make it clear that PAN is a very important vector for the global transport of NOy and understanding its transport and losses, especially its conversion back to NOx is a vital aspect of understanding global oxidising capacity of the atmosphere.

6.5

Modelling and interpretation of ambient measurements

A quantitative assessment of the importance of PANs in the atmosphere requires numerical modelling. Such models can be as simple as attempts to correlate two or more quantities, or as elaborate as the incorporation of real-time chemistry into global circulation models (GCMs). Several examples of simple models as applied to PANs and PANs precursors will be given in this section. The chemical mechanisms used to model PANs in urban and regional settings will be discussed. Several studies about the impact of PAN on the global atmosphere will also be summarised. A number of issues of importance in the chemistry of PANs in the troposphere will be discussed, including using PANs to discern the importance of isoprene as a source of PAN and O3 , PAN production by lightning and biomass burning NOx , and PAN as a source of NOx to the background troposphere.

6.5.1

Simple models

Several simple models have been used to examine the interrelationships of PAN, PPN and MPAN and their precursors to see if ambient measurements can yield quantitative information on the participation of biogenic hydrocarbons (BHC) in O3 production. A linear-combination model (Williams et al. 1997) was used to describe measured (PAN) as the sum of contributions from BHC and AHC chemistries, as represented by MPAN and PPN, respectively. [PAN] = a[MPAN] + b[PPN]

(15)

250

Volatile Organic Compounds in the Atmosphere

These two indicator species were chosen because MPAN is known to be made exclusively from methacrolein and therefore isoprene photooxidation, and PPN is known to have almost exclusively anthropogenic HC sources, while PAN is produced by both chemistries. The model was conﬁned to only instances when production was thought to dominate, but at the time, loss processes for all three PAN compounds were thought to be dominated by thermal decomposition, which was essentially the same rate for all three compounds. The coefﬁcients in Equation (15) were determined from the ambient data by least-squares ﬁts, and a check on how effective the model was at describing the atmosphere was obtained by the correlation of measured PAN with that estimated from measured MPAN and PPN and the coefﬁcients. Those correlations generally had R 2 of 0.90 or more (Roberts et al. 1998, 2002). The linear combination, model could be applied to O3 production because of the correlations of O3 and PAN that were observed in the ambient data sets, the amount of O3 attributed from each chemical regime was estimated from the ambient MPAN and PPN comeasured with O3 , and ﬁt coefﬁcients, after subtracting the O3 background (35–40 ppbv) apparent in the O3 -PAN correlation. Results of such analyses for the greater Nashville region showed that, on average, AHC chemistry was responsible for 80% of the O3 formed above background, and BHC chemistry was responsible for the other 20% (Roberts et al. 2002). However, there were instances when BHC chemistry contributed up to 75% of the O3 observed in metropolitan Nashville, which were related to a power plant plume that had passed over a high isoprene source region west of the city. This represented direct evidence of the importance of BHC in O3 formation and underscores the need to consider that reactive VOC source in formulation of O3 control strategies. A re-measurement of the MPAN + OH rate constant (Orlando et al. 2002) has reduced the quantitative utility of this analysis. Aldehydes and their associated PAN compounds are related to each other through the chemistry outlined in Figure 6.1. There have been several attempts to quantify this relationship through the use of simple kinetic analysis and observable quantities. Sillman et al. (1990) used a simple set of equations to describe PAN production from acetaldehyde in the presence of NOx : CH3 C(O)H + OH + O2 → CH3 C(O)OO + H2 O

k1

CH3 C(O)OO + NO2 → CH3 C(O)OONO2

k2

CH3 C(O)OONO2 + M → CH3 C(O)OO + NO2 + M k3 CH3 C(O)OO + NO → Loss

k4

and made a steady state assumption for PAN to arrive at an expression for PAN (Sillman et al. 1990; Equation 6.9), which is rearranged here to provide an expression for the ratio [PAN]/[Acetaldehyde]: [PAN]/[Acetaldehyde] = k1 k2 /k3 k4 × [OH]{[NO2 ]/[NO]}

(16)

The appropriateness of a steady-state assumption for PAN is questionable and will certainly depend on temperature. Only slightly lower temperatures will result in the lifetime of PAN being longer than that of acetaldehyde, clearly a situation in which [PAN] will not reach steady state. These kinds of considerations, albeit with respect to a different chemical system, led to the development of the ‘Sequential Reaction Model’ (Bertman et al. 1995). Roberts et al. (2001a) applied this model to the reaction of aldehydes to form PANs

PAN and Related Compounds

251

10

8 hrs

8 7 6 5 4

4 hrs

3 2

(PPN)/(Propanal)

2 hrs 1 8 7 6 5 4

1 hr

3

0.5 hrs

2

0.1

0.25 hrs

8 7 6 5 4 3 2 3

4

5

6 7 8 9

0.1

0

25

2

3

4

5

6 7 8 9

(PAN)/(Ethanal)

2

3

1

50 75 100 (Ozone), ppbv

125

Figure 6.8 The ratio [PPN]/[propanal] vs [PAN]/[ethanal]. The individual points are from the NEAQS 2002 study for samples taken near the coast of New England, colour coded by the co-measured O3 mixing ratios. Also shown are the data from the Nashville 1999 Study (squares), and TexAQS 2000 Study (circles) plotted in 8 bins, with error bars denoting the extent (maximum and minimum) of each bin on the horizontal scale, and the standard deviation of each bin on the vertical scale. The solid diamond is the point derived from the steady state calculation of Sillman et al. (1990). This image appears in full colour in the plate section which follows page 268 as Plate 6.

and the subsequent loss of PANs due to thermal decomposition. In this case, the timedependent equations were solved analytically after making the simplifying assumption of no dilution (a reasonable approach if considering a ratio). The following expression was arrived at: [PAN]/[Aldehyde] = βkA /(kB − kA ){1 − exp(kA − kB )t }

(17)

where β = 1/(1 + k4 [NO]/k2 [NO2 ]), kA = k1 [OH], and kB = k3 (1 − β). Equation (17) also assumes that the ratio is zero at t = 0. Ratios derived from ambient data can then be plotted against some objective measure of time if available, or more easily, one ratio can be plotted against another, as in Figure 6.8. Data from three different ﬁeld projects are shown in Figure 6.8, along with a calculated line derived from measured rate constants and branching ratios, and the average conditions; T = 303◦ K, [OH] = 6 × 106 , and NO/NO2 = 0.1, appropriate to the NEAQS 2002 study. These average conditions were

252

Volatile Organic Compounds in the Atmosphere

used to establish the times shown on the plot; the position of the line is relatively insensitive to the conditions used as explained by Roberts et al. (2001a). Also plotted are the ratios obtained from Equation 16 for the same rate constants and conditions used to calculate the kinetic line. It is clear that the single point thus obtained does not yield much information relative to the range of observed ratios. The data from TexAQS 2000 lie on the model line, implying that the chemistry in that environment is reasonably described by the simple model. The data from the Nashville 99 study and NEAQS 2002 study lie off the line systematically. The direction of this deviation is consistent with the presence of sources of PAN in addition to the reaction of acetaldehyde with OH. This has been attributed to isoprene photochemistry, since that mechanism is known to form PAN through the photolysis of methylglyoxal, a pathway that does not produce acetaldehyde. The observations in Figure 6.8 are therefore consistent with what is known about the importance of isoprene in NOx VOC photochemistry in the Nashville and NEAQS studies and the lack of isoprene impact during the TexAQS 2000 campaign (Roberts et al. 2006b).

6.5.2

Photochemical models

The heart of a good box or regional model is its chemical mechanism. There is always a trade off between the completeness of a chemical mechanism and the computational resources required to run it. There are only a few mechanisms used for routine regional air quality modelling in the United States: Carbon Bond-4 (CB-4) (Byun and Ching 1999), Regional Acid Deposition Model-2 (RADM-2) (Byun and Ching 1999) and its update, the Regional Atmospheric Chemistry Mechanism (RACM) (Stockwell et al. 1997) and the Statewide Air Pollution Research Center (SAPRC-99) model (Carter 2000). These models vary in the way that they treat aldehyde and PANs chemistry. CB-4 and RADM2 lump aldehydes and PANs together so that PAN is not differentiated from higher-carbon-number PANs, with the addition that RADM2 separates an unsaturated carbonyl PAN that is formed from aromatic HC chemistry via an unsaturated dicarbonyl. The RACM model lumps the simple alkyl aldehydes and PAN together, and combines the unsaturated aldehydes from isoprene and dienes into one species and then combines the PAN formed from that with the PAN that would be formed from unsaturated dicarbonyls. The SAPRC99 mechanism has an explicit treatment of PAN and the acetaldehyde sources in the model are dealt with by estimating the amount that will be formed from lumped higher HC species, PPN and higher PANs are lumped together, and isoprene and MPAN are treated explicitly. The compromises inherent in most of these mechanisms translates into an overestimations of modelled PAN compared to ambient measurements of about a factor of 2 in regional models run for eastern North America (S. McKeen, personal communication). There have been a number of studies in which the VOC-NOx and PANs chemistry has been treated more explicitly. These so-called Master Mechanisms provide a means to explore which compounds are most important in ozone, and secondary organic aerosol production, and can give a much better description of PAN chemistry. The NCAR Master Mechanism effort has recently been described by Aumont et al. (2005) and the University of Leeds Master Chemical Mechanism (MCM) has been described by Saunders et al. (2003) and Jenkin et al. (2003). These mechanisms, while not currently compatible with full 3-D chemical transport modelling efforts, can be used to examine interesting features of PANs chemistry.

PAN and Related Compounds

253

For example, the MCM was used to calculate Photochemical Ozone Creation Potentials (POCP) and Photochemical PAN Creation Potentials (PPCP) for individual hydrocarbons (Derwent et al. 1998). PPCPs were calculated from the change in PAN due to an incremental addition of a given organic compound to a standard modelling scenario divided by that for the same mass of propylene, multiplied by 100. Most organic compounds have PPCPs below 100, which is understandable since propylene reacts fairly rapidly with OH and the ensuing photochemistry produces PAN readily. The highest PPCP were observed for 2-butenes and pentenes, and trimethyl benzenes, and were in the range (113–138). The lowest PPCPs were observed for benzaldehyde (−4) and small alkanes (0.9–17). There was some correlation between POCPs and PPCPs for some alkenes and alkyl aromatics, with some notable exceptions such as ethylene, which is used as the basis for POCP, but has no direct mechanistic path to PAN. Isoprene is an interesting case since it is the single most abundant HC emitted to the atmosphere and has a POCP of 109 and a PPCP of 77.4. The mass-weighted average of anthropogenic hydrocarbons found in tunnel studies in the United States (Harley et al. 2001; Kean et al. 2001) has a weighted PPCP of approximately 50. This has interesting implications for a region such as Eastern United States, where summertime isoprene emissions can be as large or larger than the anthropogenic HCs (Geron et al. 1994). In such an instance, the production of PAN from isoprene can be several times higher than that from AHC, which is consistent with the deviations of the PAN/ethanal ratios from the expected model line noted previously for the Nashville 99 and NEAQS2002 studies. The GEOS-CHEM model was used by Horowitz et al. (1998) to model PAN production in summertime North America. Two scenarios, AHC only and isoprene only, were examined in the study and it was found that the isoprene only scenario produced more than twice the PAN than the AHC-only scenario did. It should be noted that like O3 , there are reasons to believe that PAN production is not necessary linear with total VOC loading, making the GEOS-CHEM results perhaps a slight over estimate. Nevertheless, these results are essentially consistent with the sequential reaction treatment of eastern US data and the PPCP analysis. Isoprene-NOx photochemistry is a substantial source of PAN over the summertime continents increasing PAN by factors of 2 to 3 over what AHC chemistry produces.

6.5.3

Global PAN

The effect of PAN on the global atmosphere has been a subject of emerging interest. Much of this interest is focused on the role of PAN as a means of transporting NOx throughout the troposphere. The questions that arise in this regard are; what is the signiﬁcance of transport out of the polluted continental boundary layer as a source of PAN to the global atmosphere and how does biomass burning affect this? How important is lightning in the production of PAN? What are the roles of isoprene and oxygenated VOCs (OVOC) in the global production of PAN? A number of studies have dealt with these issues and are summarised below. Early attempts to model PAN in the global atmosphere were reported by Moxim et al. (1996), who added PAN to their global chemical transport model. The organic side of the PAN formation in that study was driven entirely by ethane and propane, scaled to

254

Volatile Organic Compounds in the Atmosphere

reproduce the PAN levels observed at rural sites in North America. As such, this chemistry will not accurately simulate the reactivity of the actual atmosphere. Nevertheless, there were features in the global PAN ﬁelds that qualitatively reproduced observations. Late-winter, early-spring PAN maxima were apparent in the zonal averages. In general, the model was higher than observations in the remote troposphere with the exception of studies in which there were acknowledged impacts from biomass burning. The model did not include the up-to-date, temperature-dependent photolysis cross-sections of Talukdar et al. (1995), but did include thermal decomposition rates calculated from temperature ﬁelds. As a consequence, Moxim et al. (1996) predicted signiﬁcant transport of PAN to the remote troposphere, especially the Northern Hemisphere Marine Boundary Layer, with a corresponding regeneration of NOx . This effect resulted in as much as 75 pptv NOx in some instances, more than sufﬁcient to shift that region from an O3 -loss to an O3 -producing regime. A study by Thakur et al. (1999) compared global PAN measurements with the results of two different global models; Harvard/Giss (Wang et al. 1998), and IMAGES (Muller and Brasseur 1995). Both models utilised an abbreviated chemistry scheme that lumped larger hydrocarbons and some PANs, but had isoprene simulated as a separate species. In general, PAN was over-predicted, especially in the upper troposphere. Similar results were observed in the global modelling results from the MATCH v3 model presented by Von Kuhlmann et al. (2003) wherein the model over-predicted PAN in the remote troposphere, and more often at higher altitudes. The MATCH results for NH surface sites also showed an overprediction; however, the model did simulate the springtime maximum that is observed in the background troposphere. Jaffe et al. (1997) described the results of a global chemical transport model that included isoprene chemistry, but limited higher HC chemistry, and apparently no PAN photolysis in the upper atmosphere. They found that the model agreed within a factor of 2 with surface PAN observations, reasonably described the springtime maximum, but slightly over-predicted PAN/NOy at remote sites. Sudo et al. (2002) presented a global chemical model that contained isoprene and lumped HC chemistry for HC > C4 , and found that the model generally compared well with observations, with a tendency to over-predict PAN in the mid to upper troposphere. The effects of lightning-induced NOx , and isoprene as a reactive carbon source, have been examined in several global chemical model studies. Lightning has a profound effect on modelled PAN in the mid-latitude mid-troposphere; Tie et al. (2001) estimated that a total lightning source of 7 TgN/year resulted in a doubling of PAN compared to the case with no lightning, and Labrador et al. (2005) found similar results for their case in which 5 TgN/year is injected into the upper layers of cloud anvils. These enhancements amount to mixing ratios of 100 pptv (PAN) or more in the mid-troposphere, and the work of Tie et al. (2001) predicts that recycling of this PAN to NOx over the remote MBL can result in a NOx enhancement as high as 10 pptv. The effect of isoprene on global PAN is hinted at by the results of Horowitz et al. (1998) discussed above. The sheer magnitude of the global isoprene ﬂux, 500 TgC/year (Guenther et al. 1995), combined with its reactivity makes this a potentially important global source. The other interesting aspect of the isoprene source is that the net formation of PAN also depends very much on the availability of NOx . One can imagine large geographic areas, such as the tropics, where there is a large ﬂux of isoprene but insufﬁcient NOx to produce much PAN. The modelling study of von Kuhlmann et al. (2004) examined the sensitivity of PAN to isoprene chemistry on a global scale. This study did

PAN and Related Compounds

255

not include AHCs, so Northern Hemisphere continental results are not really meaningful. However, several results from the tropics were interesting; the differences in modelled PAN among the different chemical mechanisms used were similar to the differences obtained using different isoprene emission scenarios, isoprene chemistry appears to be responsible for roughly 50% of the PAN observed in the tropics, and the lightning NOx source is a very important contributor to tropical PAN. The effects of biomass burning (BB) on PAN has been modelled using modiﬁed plume models. Chatﬁeld et al. (1996) examined the contribution of BB to photochemical plumes observed off the coast of Africa during TRACE A, and found that substantial PAN (>400 pptv) can be formed and ends up being lofted to as high as 7 km. This agreed with some observations downwind of the BB areas, but under-predicted PAN in some of the more intense plumes observed closer to the actual burning. Trentmann et al. (2003) used a plume model to simulate a BB plume. Their plume model produced as much as 5.5 ppbv a relatively short distance (6–8 km) downwind of the ﬁre. The PAN formed was approximately 5% of the NOy at this point in the plume. Several aspects of BB plumes make it difﬁcult to model the effects of BB on the global atmosphere; ﬁres are usually sub-gridscale in size, so their chemistry is probably not well represented by models that spread those emission over a larger grid box, intense ﬁres generate their own convective cells that can be quite intense in nature, lofting the emissions up to 10 km or more in some cases (Lavoué et al. 2000), and the nature and intensity of NOx and VOC emissions are location- and time-dependent. Nevertheless, the magnitude of this nitrogen source, 5.6 TgN/year (Lobert et al. 1990), and its association with tropical and otherwise low-NOx areas, makes PAN production from BB an important process to understand. There is one last subject in the area of photochemical modelling of PAN that bears some examination. Several studies have found inconsistencies between the measured levels of OVOC, NOx and PAN and known chemistry. Speciﬁcally, measurements of the three species; PAN, Acetaldehyde, NOx are mutually incompatible given known chemistry; either PAN is too low or acetaldehyde is too high. This problem appears in measurements over the Paciﬁc Ocean (Flocke et al. 2002; Singh et al. 2004; Staudt et al. 2003) and in the Arctic, at or just below the snow surface (Dassau et al. 2004). The Paciﬁc measurements have some additional constraints in that PAN and NOx are part of the measured total NOy , the balance of which appears reasonable. In addition, the production of PAN from acetaldehyde would deplete NOx to levels signiﬁcantly below what is observed. A modelling study of the Arctic surface measurements hypothesises that there is an unknown loss process for PAN, but it is difﬁcult to imagine what that might be. More likely is that acetaldehyde measurements are too high, perhaps because of interferences introduced in the sampling process (Northway et al. 2004).

6.6

Conclusions

Peroxyacetic nitric anhydride is one of the most important species in the lower atmosphere. It sits at the junction of the VOC-NOx photochemistry, the very chemistry that is responsible for much of the human impacts on the atmosphere, including ozone and aerosol formation. There are a number of PAN homologues such as PPN and MPAN, which can yield valuable information on the details of the VOC-NOx photochemistry that has taken place in a given

256

Volatile Organic Compounds in the Atmosphere

environment. PAN is often the most abundant odd nitrogen compound, especially in the mid to upper troposphere, because of its thermal stability and slow photolysis in these environs. As a result, the transport of PAN and decomposition to NOx is a signiﬁcant source of NOx to the remote atmosphere. A number of unanswered questions remain about PAN and its homologues. The reaction of acyl radicals with O2 apparently has another channel that does not lead to RC(O)OO formation, and it is not clear how important that may be for larger (>C3 ) PANs. It is clear from recent measurements that there are probably differences among PANs in thermal decomposition rates, some as large as 25%. More careful studies, perhaps involving relative rate measurements, might improve the precision of those measurements. Techniques for the measurement of PANs have continued to evolve, with emphasis on faster methods such as capillary and fast capillary GC methods, and chemical ionisation mass spectrometric methods. Calibration methods are becoming more sophisticated, with the advent of efﬁcient photochemical schemes that permit the source output to be calibrated in terms of a stable, well-known quantity, a nitric oxide standard. Our knowledge of PAN and PAN homologue mixing ratios in the atmosphere continues to be heavily weighted towards the Northern Hemisphere, continental sites during the summer. A survey of global measurements of PANs indicates some interesting features, a general gradient from high values in the mid to high latitudes to lower values in the tropics, especially in the Northern Hemisphere. Springtime maxima are observed in numerous measurements at rural and remote sites. Much less is known about PAN in the tropics and the Southern Hemisphere. The tropics and the remote MBL are two areas in which the source of NOx represented by PAN transport and decomposition will have the largest effect on the balance of O3 production or destruction. Recent measurements have ﬂeshed-out our understanding of the uptake of PAN by vegetation and have re-enforced the need for more studies in more representative ecosystems during the daytime growing season. There have been great improvements in the modelling and interpretation of ambient measurements of PANs. Simple models involving PAN, PPN, MPAN and aldehydes have provided observational evidence for the importance of isoprene in the formation of regional ozone over North America. Three-dimensional photochemical models used for air quality prediction and assessment are improving in their description of PAN, but still have problems in comparison to the actual ambient measurements, and the mechanisms for higher PANs are generally not adequate. Global models have also shown great improvements, with more realistic chemical mechanisms coupled with descriptions of global transport. These models can simulate several aspects of PAN observations, such as the springtime maximum, higher PAN and PAN/NOy in the mid to upper troposphere, and transport to the remote troposphere. There is evidence of the importance of isoprene, and lightning and biomass burning and NOx , in the production of PAN, although the magnitudes of these processes need to be better deﬁned. Another remaining puzzle is that models of known chemistry cannot reconcile acetaldehyde, NOx and PAN measurements in the mid and upper troposphere and over the Greenland ice sheet.

Acknowledgements I thank Andreas Schneiders and Frank Flocke for the use of unpublished results.

PAN and Related Compounds

257

References Allen, G., Remedios, J.J., Newnham, D.A., Smith, K.M. and Monks, P.S. (2005) Improved midinfrared cross-sections for peroxyacetyl nitrate (PAN) vapour. Atmospheric Chemistry and Physics, 5: 47–56. Altshuller, A.P. (1993) PANs in the atmosphere. Journal of Air and Waste Management Association, 43: 1221–30. Aumont, B., Szopa, S. and Madronich, S. (2005) Modelling the evolution of organic carbon during its gas-phase tropospheric oxidation: Development of an explicit model based on a self generating approach. Atmospheric Chemistry and Physics Discussion, 5: 703–54. Baeza Romero, M.T., Blitz, M.A., Heard, D.E., Pilling, M.J., Price, B. and Seakins, P.W. (2005) OH formation from the C2 H5 CO + O2 reaction: An experimental marker for the propionyl radical. Chemical Physics Letters, 408: 232–6. Barnes, I., Bastian, V., Becker, K.H. and Niki, H. (1987) FTIR spectroscopic studies of the CH3 S + NO2 reaction under atmospheric conditions. Chemical Physics Letters, 140: 451–7. Bartels-Rausch, T., Eichler, B., Zimmerman, P., Gaggeler, H.W. and Ammann, M. (2002) The adsorption enthalpy of nitrogen oxides on crystalline ice. Atmospheric Chemistry and Physics, 2: 235–47. Beine, H.J. and Krognes, T. (2000) The seasonal cycle of peroxyacetyl nitrate (PAN) in the European Arctic. Atmospheric Environment, 34: 933–40. Bertman, S.B. and Roberts, J.M. (1991) A PAN analog from isoprene photooxidation. Geophysical Research Letters, 18: 1461–4. Bertman, S.B., Roberts, J.M., Parrish, D.D., et al. (1995) Alkyl nitrates as a measure of air mass evolution. Journal of Geophysical Research, 100: 22805–14. Blanchard, P., Shepson, P.B., Schiff, H.I. and Drummond, J.W. (1993) Development of a gas chromatograph for trace level measurement of peroxyacetyl nitrate using chemical ampliﬁcation. Analytical Chemistry, 65: 2472–7. Blitz, M.A., Heard, D.E. and Pilling, M.J. (2002) OH formation from CH3 CO + O2 : A convenient experimental marker for the acetyl radical. Chemical Physics Letter, 365: 374–9. Bottenheim, J.W., Sirois, A., Brice, K.A. and Gallant, A.J. (1994) Five years of continuous observations of PAN and ozone at a rural location in Eastern Canada. Journal of Geophysical Research, 99: 5333–52. Bowman, J.H., Barket Jr., D.J. and Shepson, P.B. (2003) Atmospheric chemistry of nonanal. Environmental Science and Technology, 37: 2215–25. Bridier, I., Caralp, F., Loirat, H., et al. (1991) Kinetic and theoretical studies of the reactions CH3 C(O)O2 + NO2 + M = CH3 C(O)O2 NO2 + M between 248 and 393 K and between 30 and 760 torr. Journal of Physical Chemistry, 95: 3594–600. Brown, S.S., Stark, H., Ryerson, T.B., et al. (2003) Nitrogen oxides in the nocturnal boundary layer: Simultaneous in situ measurements of NO3 , N2 O5 , NO2 , NO, and O3 . Journal of Geophysical Research, 108 (D9): 4299. Bruckmann, P.W. and Willner, H. (1983) Infrared spectroscopic studies of peroxyacetyl nitrate (PAN) and its decomposition products. Environmental Science and Techonology, 17: 352–57. Burkhardt, M.R., Maniga, N.I., Stedman, D.H. and Paur, R.J. (1988) Gas chromatographic method for measuring nitrogen dioxide and peroxyacetyl nitrate in air without compressed gas cylinders. Analytical Chemistry, 60: 816–19. Byun, D.W. and Ching, J.K.S. (1999) Science Algorithms of the EPA Models-3 Community Multiscale Air Quality (CMAQ) Modeling System, Environmental Protection Agency Report, EPA/600/R-99/030. Canosa-Mas, C.E., King, M.D., Shallcross, D.E. and Wayne, R.P. (1999) Kinetic investigation of the reaction between the NO3 radical and peroxymethacrylic nitric anhydride (MPAN). Physical Chemistry Chemical Physics, 1: 2411–14.

258

Volatile Organic Compounds in the Atmosphere

Carter, W.P.L. (2000) Documentation of the SAPRC-99 Chemical Mechanism for VOC Reactivity Assessment, ﬁnal report to the California Air Resources Board. Chatﬁeld, R.B., Vastano, J.A., Singh, H.B. and Sachse, G. (1996) A general model of how ﬁre emissions and chemistry produce African/oceanic plumes (O3 , CO, PAN, Smoke) in TRACE A. Journal of Geophysical Research, 101: 24279–306. Cooper, O.R., Moody, J.L., Parrish, D.D., et al. (2001) Trace gas signatures of the airstreams within North Atlantic cyclones: Case studies from the North Atlantic Regional Experiment (NARE’97) Aircraft Intensive. Journal of Geophysical Research, 106: 5437–56. Cooper, O.R., Forster, C., Parrish, D., et al. (2004) A case study of trans-Paciﬁc warm conveyor belt transport: The inﬂuence of merging airstreams on trace gas import to North America. Journal of Geophysical Research, 109: D23S08. Corsmeier, U., Kalthoff, N., Vogel, B., et al. (2002) Ozone and PAN formation inside and outside of the Berlin plume – Process analysis and numerical process simulation. Journal of Atmospheric Chemistry, 42: 289–321. D’Anna, B., Wisthaler, A., Andreasen, Ø., et al. (2005) Atmospheric chemistry of C3 –C6 cycloalkanecarbaldehydes. Journal of Physical Chemistry, 109: 5104–18. Darley, E.F., Kettner, K.A. and Stephens, E.R. (1963) Analysis of peroxyacyl nitrates by gas chromatography with electron capture detection. Analytical Chemistry, 35: 589–91. Dassau, T.M., Shepson, P.B., Bottenheim, J.W. and Ford, K.M. (2004) Peroxyacetyl nitrate photochemistry and interactions with the Arctic surface. Journal of Geophysical Research, 109: D18302. Day, D.A., Wooldridge, P.J., Dillon, M.B., Thornton, J.A. and Cohen, R.C. (2002) A thermal dissociation laser-induced ﬂuorescence instrument for in situ detection of NO2 , peroxy nitrates, alkyl nitrates, and HNO3 . Journal of Geophysical Research, 107 (D6): D000779. Derwent, R.G., Jenkin, M.E., Saunders, S.M. and Pilling, M.J. (1998) Photochemical ozone creation potentials for organic compounds in Northwest Europe calculated with a master chemical mechanism. Atmospheric Environment, 32: 2429–41. Dollard, G.J., Jones, B.M.R. and Davies, T.J. (1990) Dry Deposition of HNO3 and PAN. AERE Report R13780. Harwell Lab, Oxfordshire, UK. Domalski, E.S. (1971) Thermochemical properties of peroxyacetyl (PAN) and peroxybenzoyl nitrate (PBN). Environmental Science and Technology, 5: 443–4. Doskey, P.V., Kotamarthi, V.R., Fukui, Y., Cook, D.R., Breitbeil III, F.W. and Wesely, M.L. (2004) Airsurface exchange of peroxyacetyl nitrate at a grassland site. Journal of Geophysical Research, 109: D10310. Edney, E.O., Spence, J.W. and Hanst, P.L. (1979) Peroxy nitrate air pollutants: Synthesis and thermal stability. In: D. Grosjean (Ed.) Nitrogenous Air Pollutants. Ann Arbor Science Publishers, Ann Arbor, MI. Fahey, D.W., Eubank, C.S., Hübler, G. and Fehsenfeld, F.C. (1985) Evaluation of a catalytic reduction technique for the measurement of total reactive odd-nitrogen, NOy , in the atmosphere. Journal of Atmospheric Chemistry, 3: 435–68. Farmer, D.K., Wooldridge, P.J. and Cohen, R.C. (2006) Application of thermal dissociation-laser induced ﬂuorescence (TD-LIF) to measurement of HNO3 , alkyl nitrates, peroxynitrates, and NO2 ﬂuxes using eddy covariance. Atmospheric Chemistry and Physics Discussion, 6: 2923–64. Flocke, F.M., Weinheimer, A.J., Kondo, Y., et al. (2002) Observations of PANs on board the NASA P-3 during TRACE-P, American Geophysical Union National Meeting, December 6–10, San Francisco. Flocke, F.M., Weinheimer, A.J., Swanson, A.L., Roberts, J.M., Schmitt, R. and Shertz, S. (2005) On the measurement of PANs by gas chromatography and electron capture detection. Journal of Atmospheric Chemistry, 52: 19–43. Flowers, B.A., Angerhofer, M.E., Simpson, W.R., Nakayama, T. and Matsumi, Y. (2005) Nitrate radical quantum yield from peroxyacetyl nitrate photolysis. Journal of Physical Chemistry A, 109: 2552–8.

PAN and Related Compounds

259

Ford, K.M., Campbell, B.M., Shepson, P.B., et al. (2002) Studies of peroxyacetyl nitrate (PAN) and its interaction with the Snowpack at Summit Greenland. Journal of Geophysical Research, 107 (D10): D000547. Francisco, J.S. and Li, Y. (2004) Ab initio study of the electronic spectrum of peroxyacetyl nitrate. Journal of Chemical Physics, 121: 6298–301. Frenzel, A., Kutsuna, S., Takeuchi, K. and Ibusuki, T. (2000) Solubility and reactivity of peroxyacetyl nitrate (PAN) in dilute aqueous salt solutions and in sulphuric acid. Atmospheric Environment, 34: 3641–4. Gaffney, J.S., Fajer, R. and Senum, G.I. (1984) An improved procedure for high purity gaseous peroxyacetyl nitrate production: Use of heavy lipid solvents. Atmospheric Environment, 18: 215–18. Gaffney, J.S., Marley, N.A. and Prestbo, E.W. (1989) Peroxyacyl nitrates (PANs): Their physical and chemical properties. In: O. Hutzinger (Ed.) The Handbook of Environmental Chemistry, 4B. Springer-Verlag, Berlin, pp. 1–38. Gaffney, J.S., Bornick, R.M., Chen, Y.-H. and Marley, N.A. (1998) Capillary gas chromatographic analysis of nitrogen dioxide and PANs with luminol chemiluminescent detection. Atmospheric Environment, 32: 1445–54. Gaffney, J.S., Marley, N.A., Cunningham, M.M. and Doskey, P.V. (1999) Measurements of peroxyacyl nitrates (PANS) in Mexico City: Implications for megacity air quality impacts on regional scales. Atmospheric Environment, 33: 5003–12. Gallagher, M.S., Carsey, T.P. and Farmer, M.L. (1990) Peroxyacetyl nitrate in the North Atlantic marine boundary layer. Global Biogeochemical Cycles, 4: 297–308. Garland, J.A. and Penkett, S.A. (1976) Absorption of peroxyacetyl nitrate and ozone by natural surfaces. Atmospheric Environment, 10: 1127–31. Geron, C.A., Guenther, A. and Pierce, T. (1994) An improved model for estimating emissions of volatile organic compounds from forests in the eastern United States. Journal of Geophysical Research, 99: 12773–92. Glavas, S. and Moschonas, N. (2001) Determination of PAN, PPN, PnBN and selected pentyl nitrates in Athens, Greece. Atmospheric Environment, 35: 5467–75. Glasson, W.A. and Huess, J.M. (1977) Synthesis and evaluation of potential eye irritants. Environmental Science and Technology, 11: 395–8. de Gouw, J.A., Goldan, P.D., Warneke, C., et al. (2003) Validation of proton transfer reactionmass spectrometry (PTR-MS) measurements of gas-phase organic compounds in the atmosphere during the New England Air Quality Study (NEAQS) in 2002. Journal of Geophysical Research, 108: D003863. de Gouw, J.A., Cooper, O.R., Warneke, C., et al. (2004) Chemical composition of air masses transported from Asia to the U.S. West Coast during ITCT2K2: Fossil fuel combustion versus biomass-burning signatures. Journal of Geophysical Research, 109: D23S20. Gregory, G.L., Hoell Jr., J.M., Ridley, B.A., et al. (1990) An intercomparison of airborne PAN measurements. Journal of Geophysical Research, 95: 10077–88. Grosjean, D. (1983) Distribution of atmospheric nitrogenous pollutants at a Los Angeles area smog receptor site. Environmental Science and Technology, 17: 13–19. Grosjean, D., Grosjean, E. and Williams II, E.L. (1994a) Formation and thermal decomposition of butyl-substituted peroxyacyl nitrates: n-C4 H9 C(O)OONO2 and i-C4 H9 C(O)OONO2 . Environmental Science and Technology, 28: 1099–105. Grosjean, D., Grosjean, E. and Williams II, E.L. (1994b) Thermal decomposition of C3 -substituted peroxyacyl nitrates. Research on Chemical Intermediates, 20: 447–61. Grosjean, D., Grosjean, E. and Williams II, E.L. (1994c) Thermal decomposition of PAN, PPN, and vinyl PAN. Journal of Air and Waste Management Association, 44: 391–6. Grosjean, E., Grosjean, D., Fraser, M.P. and Cass, G.R. (1996) Air quality model evaluation data for organics. 3. Peroxyacetyl nitrate and peroxypropionyl nitrate in Los Angeles air. Environmental Science and Technology, 30: 2704–14.

260

Volatile Organic Compounds in the Atmosphere

Guenther, A., Hewitt, C.N., Erickson, D., et al. (1995) A global model of natural volatile organic compound emissions. Journal of Geophysical Research, 100: 8873–92. Hansel, A. and Wisthaler, A. (2000) A method for real-time detection of PAN, PPN, and MPAN in ambient air. Geophysical Research Letter, 27: 895–8. Hansel, A., Wisthaler, A., Flocke, F., et al. (2002) An Intercomparison of Airborne VOC and PAN Measurements. Presented at the Fall Meeting of the American Geophysical Union, 6–10 December, San Francisco, CA. Harley, R.A., McKeen, S.A., Pearson, J., Rodgers, M.O. and Lonneman, W.A. (2001) Analysis of motor vehicle emissions during the Nashville/Middle Tennessee Ozone Study. Journal of Geophysical Research, 106: 3559–67. Harwood, M.H., Roberts, J.M., Frost, G.J., Ravishankara, A.R. and Burkholder, J.B. (2003) Photochemical studies of CH3 C(O)OONO2 (PAN) and CH3 CH2 C(O)OONO2 (PPN): NO3 quantum yields. Journal of Physical Chemistry A, 107: 1148–54. Hayden, K.L., Anlauf, K.G., Li, S.-M., et al. (2004) Characterization of gaseous nitrogen oxides in the lower Fraser Valley during Paciﬁc 2001. Atmospheric Environment, 38: 5811–23. Hermann, A., Niemeyer, J., Mack, H.-G., et al. (2001) Gas-phase structures of acetyl peroxynitrate and triﬂuoroacetyl peroxynitrate. Inorganic Chemistry, 40: 1672–6. Holdren, M.W., Spicer, C.W. and Hales, J.M. (1984) Peroxyacetyl nitrate solubility and decomposition rate in acidic water. Atmospheric Environment, 18: 1171–3. Horowitz, L.W., Liang, J., Gardner, G.M. and Jacob, D.J. (1998) Export of reactive nitrogen from North America during summertime: Sensitivity to hydrocarbon chemistry. Journal of Geophysical Research, 103: 13451–76. Hudman, R.C., Jacob, D.J., Cooper, O.R., et al. (2004) Ozone production in transpaciﬁc Asian pollution plumes and implications of ozone air quality in California. Journal of Geophysical Research, 109: D23S10. Heuss, J.M. and Glasson, W.A. (1968) Hydrocarbon reactivity and eye irritation. Environmental Science and Technology, 2: 1109–16. Hurley, M.D., Misner, J.A., Ball, J.C., et al. (2005) Atmospheric chemistry of CF3CH2CH2OH: Kinetics, mechanisms, and products of Cl atom and OH radical initiated oxidation in the presence and absence of NOx . Journal of Physical Chemistry, 109: 9816–26. Jacobi, H.-W. and Schrems, O. (1999) Peroxyacetyl nitrate (PAN) distribution over the Southern Atlantic Ocean. Physical Chemistry Chemical Physics, 1: 5517–21. Jacobi, H.-W., Weller, R., Bluszcz, T. and Schrems, O. (1999) Latitudinal distribution of peroxyacetyl nitrate (PAN) over the Atlantic Ocean. Journal of Geophysical Research, 104: 26901–12. Jacobi, H.-W., Weller, R., Jones, A.E., Anderson, P.S. and Schrems, O. (2000) Peroxyacetyl nitrate (PAN) concentrations in the Antarctic troposphere measured during the photochemical experiment at Neumayer (PEAN’99). Atmospheric Environment, 34: 5235–47. Jaeschke, W., Salkowski, T., Dierssen, J.P., Trumbach, J.V., Krischke, U. and Gunter, A. (1999) Measurement of trace substances in the Arctic troposphere as potential precursors and constituents of Artic haze. Journal of Atmospheric Chemistry, 34: 291–319. Jaffe, D.A., Berntsen, T.K. and Isaksen, I.S.A. (1997) A global three-dimensional chemical transport model 2. Nitrogen oxides and nonmethane hydrocarbon results. Journal of Geophysical Research, 102: 21281–96. Jenkin, M.E., Saunders, S.M., Wagner, V. and Pilling, M.J. (2003) Protocol of the development of the master chemical mechanism, MCM v3 (Part B): Tropospheric degradation of aromatic volatile organic compounds. Atmospheric Chemistry and Physics, 3: 181–93. Joos, L.F., Landolt, W.F. and Leuenberger, H. (1986) Calibration of peroxyacetyl nitrate measurements with an NOx analyzer. Environmental Science and Technology, 20: 1269–73. Kacmarek, A.J., Solomon, I.J. and Lustig, M. (1977) Preparation and properties of peroxyacetyl nitrate. Journal of Inorganic and Nuclear Chemistry, 40: 574–6.

PAN and Related Compounds

261

Kames, J. and Schurath, U. (1995) Henry’s law and hydrolysis-rate constants for peroxyacyl nitrates (PANs) using a homogenous gas-phase source. Journal of Atmospheric Chemistry, 21: 151–64. Kean, A.J., Grosjean, E., Grosjean, D. and Harley, R.A. (2001) On-road measurement of carbonyls in California light-duty vehicle emissions. Environmental Science and Technology, 35: 4198–204. Kenley, R.A. and Hendry, D.G. (1982) Generation of peroxy radicals from peroxynitrates (ROONO2 ). Decomposition of peroxybenzoyl nitrate (PBzN). Journal of American Chemical Socety, 104: 220–4. Kirchner, F., Thuner, L.P., Barnes, I., Becker, K.H., Donner, B. and Zabel, F. (1997) Thermal lifetimes of peroxynitrates occurring in the atmospheric degradation of oxygenated fuel additives. Environmental Science and Technology, 31: 1801–4. Kirchner, F., Mayer-Figge, A., Zabel, F. and Becker, K.H. (1999) Thermal stability of peroxynitrates. International Journal of Chemical Kinetics, 31: 127–144. Kleindienst, T.E. (1994) Recent developments in the chemistry and biology of peroxyacetyl nitrate. Research on Chemical Intermediates, 20: 335–84. Kopitzky, R., Beukleke, M., Balzer, G. and Willner, H. (1997) Properties of trifuoroacetyl peroxynitrate, CF3 C(O)OONO2 . Inorganic Chemistry, 36: 1994–7. Kotchenruther, R.A., Jaffe, D.A. and Jaegle, L. (2001) Ozone photochemistry and the role of peroxyacetyl nitrate in the springtime northeastern Paciﬁc troposphere: Results from the photochemical ozone budget of the Eastern North America atmosphere (PHOBEA) campaign. Journal of Geophysical Research, 106: 28731–41. Kourtidis, K.A., Fabian, P., Zerefos, C. and Rappenglück, B. (1993) Peroxyacetyl nitrate (PAN), peroxypropionyl nitrate (PPN) and PAN/ozone ratio measurements at three sites in Germany. Tellus, 45B: 442–57. Kravetz, T.M., Martin, S.W. and Mendenhall, G.D. (1980) Synthesis of peroxyacetyl and peroxyaroyl nitrates. Complexation of peroxyacetyl nitrate with benzene. Environmental Science and Technology, 14: 1262–4. von Kuhlmann, R., Lawrence, M.G., Crutzen, P.J. and Rasch, P.J. (2003) A model of studies of tropospheric ozone and non-methane hydrocarbons: Model evaluation of ozone-related species. Journal of Geophysical Research, 108 (D23): 4729. von Kuhlmann, R., Lawrence, M.G., Poschl, U. and Crutzen, P.J. (2004) Sensitivities in global scale modeling of isoprene. Atmospheric Chemistry and Physics, 4: 1–17. Labrador, L.J., von Kuhlmann, R. and Lawrence, M.G. (2005) The effects of lightning-produced NOx and its vertical distribution on atmospheric chemistry: Sensitivity simulations with MATCHMPIC. Atmospheric Chemistry and Physics, 5: 1815–34. Lavoué, D., Liousse, C., Cachier, H., Stocks, B.J. and Goldammer, J.G. (2000) Modeling of carbonaceous particles emitted by boreal and temperate wildﬁres at northern latitudes. Journal of Geophysical Research, 105: 26871–90. Lee, Y.-N. (1984) Kinetics of some aqueous-phase reactions of peroxyacetyl nitrate. Gas-Liquid Chemistry of Natural Waters. Brookhaven National Laboratory, BNL-51757. Libuda, H.G. and Zabel, F. (1995) UV absorption cross sections of acetyl peroxynitrate and triﬂuoroacetyl peroxynitrate at 298K. Ber. Bunsenges. Phys. Chem., 99: 1205–13. Lightfoot, P.D., Cox, R.A., Crowley, J.N., et al. (1992) Organic peroxy radicals: Kinetics, spectroscopy and tropospheric chemistry. Atmospheric Environment, 26A: 1805–961. Lobert, J.M., Scharffe, D.H., Hoa, W.M. and Crutzen, P.J. (1990) Importance of biomass burning in the atmospheric budgets of nitrogen-containing gases. Nature, 346: 552–4. Louw, R., Sluis, G.J. and Vermeeren, H.P.W. (1975) Facile preparation of peroxyacyl nitrates. Journal of American Chemical Society, 97: 4396–7. Marley, N.A., Gaffney, J.S., White, R.V., et al. (2004) Fast gas chromatography with luminol chemiluminescence detection for the simultaneous determination of nitrogen dioxide and peroxyacetyl nitrate in the atmosphere. Review of Scientiﬁc Instruments, 75: 4595–605.

262

Volatile Organic Compounds in the Atmosphere

Martinez, R.I. (1980) A systematic nomenclature for the ‘Peroxyacyl Nitrates’, the functional and structural misnomers for anhydride derivatives of nitrogen oxo acids. International Journal of Chemical Kinetics, 12: 771–5. Mazely, T.L., Friedl, R.R. and Sander, S.P. (1997) Quantum yield of NO3 from peroxyacetyl nitrate photolysis. Journal of Physical Chemistry A, 101: 7090–7. McFadyen, G.G., and Cape, J.N. (1999) Physical and chemical inﬂuences on PAN concentrations at a rural site. Atmospheric Environment, 33: 2929–40. Mereau, R., Rayez, M.-T., Rayes, J.-C., Caralp, F. and Lesclaux, R. (2001) Theoretical study on the atmospheric fate of carbonyl radicals: Kinetics of decomposition reactions. Physical Chemistry Chemical Physics, 3: 4712–17. Meyrahn, H., Helas, G. and Warneck, P. (1987) Gas chromatographic determination of peroxyacetyl nitrate: Two convenient calibration techniques. Journal of Atmospheric Chemistry, 5: 405–15. Miller, C.E., Lynton, J.I., Keevil, D.M. and Francisco, J.S. (1999) Dissociation pathways of peroxyacetyl nitrate (PAN). Journal of Physical Chemistry, 103: 11451–9. Moxim, W.J., Levy, H. and Kasibhatla, P.S. (1996) Simulated global tropospheric PAN: Its transport and impact on NOx . Journal of Geophysical Research, 101: 12621–38. Muller, J.-F. and Brasseur, G. (1995) IMAGES: A three dimensional chemical transport model of the global troposphere. Journal of Geophysical Research, 100: 16445–90. Muller, K.P. and Rudolph, J. (1989) An automated technique for the measurement of peroxyacetylnitrate in ambient air at ppb and ppt levels. International Journal of Environmental Analytical Chemistry, 37: 253–62. Muller, K.P. and Rudolph, J. (1992) Measurements of peroxyacetylnitrate in the marine boundary layer over the Atlantic. Journal of Atmospheric Chemistry, 15: 361–7. Nielsen, T., Samuelsson, U., Grennfelt, P. and Thomsen, E.L. (1981) Peroxyacetyl nitrate in long-range transported air, Nature, 293: 553–5. Nielsen, T., Hansen, A.M. and Thomsen, E.L. (1982) A convenient method for the preparation of pure standards of peroxyacetyl nitrate for atmospheric analysis. Atmospheric Environment, 16: 2447–50. Niki, H., Maker, P.D., Savage, C.M. and Breitenbach, L.P. (1985) An FTIR spectroscopic study of the reactions Br + CH3 CHO → HBr + CH3 CO and CH3 C(O)OO + NO2 ↔ CH3 C(O)OONO2 (PAN). International Journal of Chemical Kinetics, 17: 525–34. Niki, H., Maker, P.D., Savage, C.M. and Hurley, M.D. (1987) Fourier transform infrared study of the kinetics and mechanisms for the Cl-atom and HO-radical-initiated oxidation of glycolaldehyde. Journal of Physical Chemistry, 91: 2174–8. Nikitas, C., Clemitshaw, K.C., Oram, D.E. and Penkett, S.A. (1997) Measurements of PAN in the polluted boundary layer and free troposphere using a luminol-NO2 detector combined with a thermal converter. Journal of Atmospheric Chemistry, 28: 339–59. Nizkorodov, S.A., Crounse, J.D., Fry, J.L., Roehl, C.M. and Wennberg, P.O. (2005) Near-IR photodissociation of peroxy acetyl nitrate. Atmospheric Chemistry and Physics, 5: 385–92. Northway, M.J., de Gouw, J.A., Roberts, J.M., et al. (2004) Evaluation of the role of heterogeneous oxidation of alkenes in the detection of atmospheric acetaldehyde. Atmospheric Environment, 38: 6017–28. Nouaime, G., Bertman, S.B., Seaver, C., et al. (1998) Sequential oxidation products from tropospheric isoprene chemistry : MACR and MPAN at a NOx -rich forest environment in the Southeastern United States. Journal of Geophysical Research, 103: 22,463–71. Nowak, J.B., Parrish, D.D., Neuman, J.A., et al. (2004) Gas phase chemical characteristics of Asian pollution plumes observed during ITCT2K2 over the Eastern North Paciﬁc Ocean. Journal of Geophysical Research, 109: D23S19. Noziere, B. and Barnes, I. (1998) Evidence for formation of a PAN analogue of pinonic structure and investigation of its thermal stability. Journal of Geophysical Research, 103: 25587–97.

PAN and Related Compounds

263

Orlando, J.J. and Tyndall, G.S. (2001) The atmospheric chemistry of the HC(O)CO radical. International Journal of Chemical Kinetics, 33: 149–56. Orlando, J.J., Tyndall, G.S. and Calvert, J.G. (1992) Thermal decomposition pathways for peroxyacetyl nitrate (PAN): Implications for atmospheric methyl nitrate levels. Atmospheric Environment, 26A: 3111–18. Orlando, J.J., Tyndall, G.S., Bertman, S.B., Chen, W. and Burkholder, J.B. (2002) Rate coefﬁcient for the reaction of OH with CH2 = C(CH3 )C(O)OONO2 (MPAN). Atmospheric Environment, 36: 1895–900. Parrish, D.D., Buhr, M.P., Trainer, M., et al. (1993) The total reactive oxidized nitrogen levels and the partitioning between the individual species at six rural sites in Eastern North America. Journal of Geophysical Research, 98: 2927–39. Parrish, D.D., Dunlea, E.J., Atlas, E.L., et al. (2004a) Changes in the photochemical environment of the temperate North Paciﬁc troposphere in response to increasing Asian emissions. Journal of. Geophysical Research, 109: D23S18. Parrish, D.D., Ryerson, T.B., Holloway, J.S., et al. (2004b) Fraction and composition of NOy transported in air masses lofted from the North American continental boundary layer. Journal of Geophysical Research, 109: D09302. Pätz, H.-W., Corsmeier, U., Glaser, K., et al. (2000) Measurements of trace gases and photolysis frequencies during SLOPE96 and a coarse estimate of the local OH concentration from HNO3 formation. Journal of Geophysical Research, 105: 1563–83. Pätz, H.-W., Lerner, A., Houben, N. and Volz-Thomas, A. (2002) Validierung eines neuen Verfahrens zur Kalibrierung von Peroxiacetylnitrat (PAN)-Analysatoren. Gefahrstoffe – Reinhaltung der Luft, 62: 215–20. Penkett, S.A. and Brice, K.A. (1986) The spring maximum in photo-oxidants in the Northern Hemisphere troposphere. Nature, 319: 655–7. Pippin, M., Bertman, S., Thornberry, T., Town, M., Carroll, M.A. and Sillman, S. (2001) Seasonal variations of PAN, PPN, and O3 at the upper Midwest PROPHET site. Journal of Geophysical Research, 106: 24451–63. Plum, C.N., Sanhueza, E., Atkinson, R., Cater, W.P.L. and Pitts Jr., J.N. (1983) OH radical rate constants and photolysis rates of α-dicarbonyls. Environmental Science and Technology, 17: 479–84. Rappenglück, B., Oyola, P., Olaeta, I., and Fabian, P. (2000) The evolution of photochemical smog in the metropolitan area of Santiago de Chile. Journal of Applied Meteorology, 39: 275–90. Rappenglück, B., Forster, C., Jakobi, G. and Pesch, M. (2004) Unusually high levels of PAN and ozone over Berlin, Germany, during the nighttime on August 7, 1998. Atmospheric Environment, 38: 6125–34. Ridley, B.A. (1991) Recent measurements of oxidized nitrogen compounds in the troposphere. Atmospheric Environment, 25A: 1905–26. Ridley, B.A., Shetter, J.D., Walega, J.G., et al. (1990) The behavior of some organic nitrates at Boulder and Niwot Ridge, Colorado. Journal of Geophysical Research, 95: 13949–61. Roberts, J.M. (1990) The atmospheric chemistry of organic nitrates. Atmospheric Environment, 24A: 243–87. Roberts, J.M. (2005) Measurement of the Henry’s law coefﬁcient and ﬁrst order loss rate of PAN in n-octanol. Geophysical Research Letters, 32: L08803. Roberts, J.M. and Bertman, S.B. (1992) The thermal decomposition of peroxyacetic nitric anhydride (PAN) and peroxymethacrylic nitric anhydride (MPAN). International Journal of Chemical Kinetics, 24: 297–307. Roberts, J.M., Fajer, R.W. and Springston, S.R. (1989) Capillary gas chromatographic separation of alkyl nitrates and peroxycarboxylic nitric anhydrides. Analytical Chemistry, 61: 771–2. Roberts, J.M., Parrish, D.D., Norton, R.B., et al. (1996) Episodic removal of NOy species from the marine boundary layer over the North Atlantic. Journal of Geophysical Research, 101: 28947–60.

264

Volatile Organic Compounds in the Atmosphere

Roberts, J.M., Williams, J., Baumann, K., et al. (1998) Measurements of PAN, PPN and MPAN made during the 1994 and 1995 Nashville intensives of the southern oxidant study: Implications for regional ozone production from biogenic hydrocarbons. Journal of Geophysical Research, 103: 22473–90. Roberts, J.M., Stroud, C.A., Jobson, B.T., et al. (2001a) Application of a sequential reaction model to PANs and aldehyde measurements in two urban areas. Geophysical Research Letters, 28: 4583–6. Roberts, J.M., Flocke, F., Weinheimer, A., et al. (2001b) Observations of APAN during TEXAQS 2000. Geophysical Research Letters, 28: 4195–8. Roberts, J.M., Flocke, F.M., Stroud, C.A., et al. (2002) Ground-based measurements of peroxycarboxylic nitric anhydrides (PANs) during the 1999 southern oxidants study Nashville intensive. Journal of Geophysical Research, 107 (D21): 4554. Roberts, J.M., Jobson, B.T., Kuster, W., et al. (2003) An examination of the chemistry of peroxycarboxylic nitric anhydrides (PANs) and related VOC chemistry during TexAQS 2000 using ground-based measurements. Journal of Geophysical Research. 108 (D16): 4495. Roberts, J.M., Flocke, F., Chen, G., et al. (2004) Measurement of peroxycarboxylic nitric anhydrides (PANs) during the ITCT2K2 aircraft intensive experiment. Journal of Geophysical Research, 109: D23S21. Roberts, J.M., Flocke, F., Warshawsky, M., et al. (2007) Measurements of PAN at the south pole during ANTCI 2003, Atmospheric Environment, submitted. Roberts, J.M. (2006) Analysis of the isoprene chemistry observed during the New England Air Quality Study (NEAQS) 2002 intensive experiment. Journal of Geophysical Research, 111: D23S12. Rogers, J.D. and Rhead, L.A. (1987) Peroxyglutaryl nitrate: Formation and infrared spectrum. Atmospheric Environment, 21: 2519–23. Roumelis, N. and Glavas, S. (1989) Decomposition of peroxyacetyl nitrate and peroxypropionyl nitrate during gas chromatographic determination with a wide-bore capillary and two packed columns. Analytical Chemistry, 61: 2731–4. Roumelis, N. and Glavas, S. (1992a) Thermal decomposition of peroxyacetyl nitrate in the presence of O2 , NO2 , and NO. Monatshefte fur Chemie, 123: 63–72. Roumelis, N. and Glavas, S. (1992b) Formation of peroxyacyl- and alkyl nitrates in irradiations of gasoline vapours under atmospheric conditions. Environmental Technology, 13: 189–94. Rubio, M.A., Oyola, P., Gramsch, E., Lissi, E., Pizarro, J. and Villena, G. (2004) Ozone and peroxyacetylnitrate in downtown Santiago, Chile. Atmospheric Environment, 38: 4931–9. Rudolph, J., Vierkorn-Rudolph, B. and Meixner, F.X. (1987) Large-scale distribution of peroxyacetylnitrate results from the STRATOZ III ﬂights. Journal of Geophysical Research, 92: 6653–61. Russo, R.S., Talbot, R.W., Dibb, J.E., et al. (2003) Chemical composition of Asian continental outﬂow over the Western Paciﬁc: Results from transport and chemical evolution over the Paciﬁc (TRACE-P). Journal of Geophysical Research, 108 (D20): 8804. Ryerson, T.B., Trainer, M., Angevine, W.M., et al. (2003) Effect of petrochemical industrial emissions of reactive alkenes and NOx on tropospheric ozone formation in Houston, TX. Journal of Geophysical Research, 108: D003070. Saunders, S.M., Jenkin, M.E., Derwent, R.G. and Pilling, M.J. (2003) Protocol for the development of the master chemical mechanism, MCM v3 (Part A): Tropospheric degradation of non-aromatic volatile organic compounds. Atmospheric Chemistry and Physics, 3: 161–80. Schefﬂer, D., Schaper, I., Willner, H., Mack, H.-G. and Oberhammer, H. (1997) Properties of ﬂuorocarbonyl peroxynitrate, Inorganic Chemistry, 36: 339–44. Schmidt, R.W.H., Slemr, F. and Schurath, U. (1998) Airborne peroxyacetyl nitrate (PAN) and peroxypropionyl nitrate (PPN) measurements during TRACT 1992 (1998). Atmospheric Environment, 32: 1203–27.

PAN and Related Compounds

265

Schmitt, R. and Volz-Thomas, A. (1997) Climatology of ozone, PAN, CO, and NMHC in the free troposphere over the Southern North Atlantic. Journal of Atmospheric Chemistry, 28: 245–62. Schrimpf, W., Lienaerts, K., Muller, K.P., et al. (1996) Dry deposition of peroxyacetyl nitrate (PAN): Determination of its deposition velocity at night from measurements of the atmospheric PAN and 222 Radon concentration gradient. Geophysical Research Letters, 23: 3599–602. Schultz, M., Schmitt, R., Thomas, K. and Volz-Thomas, A. (1998) Photochemical box modeling of long-range transport from North America to Tenerife during the North Atlantic Regional Experiment (NARE) 1993. Journal of Geophysical Research, 103 (D11): 13477–88. Senum, G.I., Fajer, R. and Gaffney, J.S. (1986) Fourier transform infrared spectroscopic study of the thermal stability of peroxyacetyl nitrate. Journal of Physical Chemistry, 90: 152–6. Shepson, P.B., Bottenheim, J.W., Hastie, D.R. and Venkatram, A. (1992) Determination of the relative ozone and PAN deposition velocities at night. Geophysical Research Letters, 19: 1121–4. Sillman, S., Logan, J.A. and Wofsy, S.C. (1990) The sensitivity of ozone to nitrogen oxides and hydrocarbons in regional ozone episodes. Journal of Geophysical Research, 95: 1837–51. Singh, H.B. (1987) Reactive nitrogen in the troposphere, chemistry and transport of NOx and PAN. Environmental Science and Technology, 21: 320–7. Singh, H.B. and Salas, L.J. (1989) Measurements of peroxyacetyl nitrate (PAN) and peroxypropionyl nitrate (PPN)) at selected urban, rural and remote sites. Atmospheric Environment, 23: 231–8. Singh, H.B., Salas, L.J. and Viezee, W. (1986) Global distribution of peroxyacetyl nitrate. Nature, 321: 588–91. Singh, H.B., Herlth, D., O’Hara, D., et al. (1994) Summertime distribution of PAN and other reactive nitrogen species in the northern high-latitude atmosphere of Eastern Canada. Journal of Geophysical Research, 99: 1821–35. Singh, H.B., Herlth, D., Kolyer, R., et al. (1996a) Reactive nitrogen and ozone over the Western Paciﬁc: Distribution, partitioning, and sources. Journal of Geophysical Research, 101: 1793–808. Singh, H.B., Herlth, D., Kolyer, R., et al. (1996b) Impact of biomass burning emissions on the composition of the South Atlantic troposphere: Reactive nitrogen and ozone. Journal of Geophysical Research, 101: 24203–19. Singh, H.B., Viezee, W., Chen, Y., et al. (1998) Latitudinal distribution of reactive nitrogen in the free troposphere over the Paciﬁc Ocean in late Winter/early Spring. Journal of Geophysical Research, 103: 28237–46. Singh, H., Chen, Y., Tabazadeh, A., et al. (2000) Distribution and fate of selected oxygenated organic species in the troposphere and lower stratosphere over the Atlantic. Journal of Geophysical Research, 105: 3795–805. Singh, H.B., Salas, L.J., Chatﬁeld, R.B., et al. (2004) Analysis of the atmospheric distribution, sources, and sinks of oxygenated volatile organic chemicals based on measurements over the Paciﬁc during TRACE-P. Journal of Geophysical Research, 109: D15S07. Slusher, D.L., Huey, L.G., Tanner, D.J., Flocke, F.M. and Roberts, J.M. (2004) A thermal dissociationchemical ionization mass spectrometry (TD-CIMS) technique for the simultaneous measurement of peroxyacyl nitrates and dinitrogen pentoxide. Journal of Geophysical Research, 109: D19315. Sparks, J.P., Roberts, J.M. and Monson, R.K. (2003) The uptake of gaseous organic nitrogen by leaves: A signiﬁcant global nitrogen transfer process. Geophysical Research Letters, 30: 2189. Spence, J.W., Edney, E.O. and Hanst, P.L. (1978) Peroxychloroformyl nitrate: Synthesis and thermal stability. Chemical Physics Letters, 56: 478–83. Staudt, A.C., Jacob, D.J., Ravetta, F., et al. (2003) Sources and chemistry of nitrogen oxides over the tropical Paciﬁc. Journal of Geophysical Research, 108 (D2): 8239. Stephens, E.R. (1964) Absorptivities for infrared determination of peroxyacyl nitrates. Analytical Chemistry, 36: 928–9. Stephens, E.R. (1969) The formation, reactions and properties of peroxyacyl nitrates (PANs) in photochemical air pollution. Advances in Environmental Science, 1: 119–46.

266

Volatile Organic Compounds in the Atmosphere

Stephens, E.R. (1987) Smog studies of the 1950’s. EOS Transformation, 68: 89–92. Stephens, E.R., Hanst, P.L., Doerr, R.C. and Scott, W.E. (1956) Reactions of nitrogen dioxide and organic compounds in air. Industrial and Engineering Chemistry, 48: 1498–504. Stockwell, W.R., Kirchner, F., Kuhn, M. and Seefeld, S. (1997) A new mechanism for atmospheric chemistry modeling. Journal of Geophysical Research, 102: 25847–79. Sudo, K., Takahashi, M. and Akimoto, H. (2002) CHASER: A global chemical model of the troposphere 2. Model results and evaluation. Journal of Geophysical Research, 107 (D21): 4586. Sulbaek Andersen, M.P., Hurley, M.D., Wallington, T.J., et al. (2003) Atmospheric chemistry of C2 F5 CHO: Reaction with Cl atoms and OH radicals, IR spectrum of C2 F5 C(O)OONO2 . Chemical Physics Letters, 379: 28–36. Swanson, A.L., Flocke, F., Roberts, J., et al. (2004) Characterization of a TD-CIMS Instrument for Aircraft Measurements of PeroxyAcetyl Nitrate (PAN) Compounds. Presented at the Fall Meeting of the American Geophysical Union, 13–17 December, San Francisco, CA. van Swieten, A.P., Louw, R. and Arnoldy, P. (1978) Chemistry of peroxyacyl nitrates I. Thermal decomposition in cyclohexane solution: Effect of added nitric oxide and of nitrogen dioxide. Reclueil des Travaux Chimiques des Pays-Bas, 97: 61–65. Talbot, R.W., Bradshaw, J.D., Sandholm, S.T., et al. (1996) Chemical characteristics of continental outﬂow over the tropical South Atlantic Ocean from Brazil and Africa. Journal of Geophysical Research, 101: 24187–202. Talbot, R.W., Dibb, J.E., Scheuer, E.M., et al. (1999) Reactive nitrogen budget during the NASA SONEX mission. Geophysical Research Letters, 26: 3057–60. Talbot, R.W., Dibb, J.E., Scheuer, E.M., et al. (2003) Reactive nitrogen in Asian Continental Outﬂow over the Western Paciﬁc: Results from the NASA transport and chemical evolution over the Paciﬁc (TRACE-P) airborne mission. Journal of Geophysical Research, 108 (D20): 8803. Talukdar, R.K., Burkholder, J.B., Schmoltner, A.-M., Roberts, J.M., Wilson, R. and Ravishankara, A.R. (1995) An investigation of the loss processes for peroxyacetyl nitrate in the atmosphere: UV photolysis and reaction with OH. Journal of Geophysical Research, 100: 14163–73. Tanimoto, H. (2000) The Seasonal Variation of Atmospheric Peroxyacetyl Nitrate (PAN) in East Asia Observed by GC/NICI-MS Technique. PhD Thesis, University of Tokyo, Tokyo, Japan. Tanimoto, H. and Akimoto, H. (2001) A new peroxycarboxylic nitric anhydride identiﬁed in the atmosphere: CH2 = CHC(O)OONO2 (APAN). Geophysical Research Letters, 28: 2831–4. Tanimoto, H., Hirokawa, J., Kajii, Y. and Akimoto, H. (1999) A new measurement technique of peroxyacetyl nitrate at parts per trillion by volume levels: Gas chromatography/negative ion chemical ionization mass spectrometry. Journal of Geophysical Research, 104: 21343–54. Thakur, A.N., Singh, H.B., Mariani, P., et al. (1999) Distribution of reactive nitrogen species in the remote free troposphere: Data and model comparisons. Atmospheric Environment, 33: 1403–22. Thornberry, T., Carroll, M.A., Keeler, G.J., et al. (2001) Observations of reactive oxidized nitrogen and speciation of NOy during the PROPHET summer 1998 intensive. Journal of Geophysical Research, 106: 24359–86. Thornton, J.A., Wooldridge, P.J., Cohen, R., et al. (2003) Comparisons of in situ and long path measurements of NO2 in urban plumes. Journal of Geophysical Research, 108 (D16): 4496. Tie, X., Zhang, R., Brasseur, G., Emmons, L. and Lei, W. (2001) Effects of lightning on reactive nitrogen and nitrogen reservoir species in the troposphere. Journal of Geophysical Research, 106: 3167–78. Tomas, A., Villenave, E. and Lesclaux, R. (2000) Kinetics of the (CH3 )2 CHCO and (CH3 )3 CCO radical decomposition: Temperature and pressure dependence. Physical Chemistry Chemical Physics, 2: 1165–74. Trentmann, J., Andreae, M.O. and Graf, H.-F. (2003) Chemical processes in a young biomass-burning plume. Journal of Geophysical Research, 108 (D22): 4705. Tsalkini, N. and Toupance, G. (1989) Infrared absorptivities and integrated band intensities for gaseous peroxyacetyl nitrate (PAN). Atmospheric Environment, 23: 1849–54.

PAN and Related Compounds

267

Tuazon, E.C. and Atkinson, R. (1990) A product study of the gas phase reaction of methacrolein with the OH radical in the presence of NOx . International Journal of Chemical Kinetics, 22: 591–602. Tuazon, E.C., Graham, R.A., Winer, A.M., Easton, R.R., Pitts Jr., J.N. and Hanst, P.L. (1978) A kilometer pathlength Fourier-transform infrared system for the study of trace pollutants in ambient and synthetic atmospheres. Atmospheric Environment, 12: 865–75. Tuazon, E.C., Aschmann, S.M., Nishino, N., Arey, J. and Atkinson, R. (2005) Kinetics and products of the OH radical-initiated reaction of 3-methyl-2-butenal. Physical Chemistry Chemical Physics, 7: 2298–304. Turnipseed, A.A., Huey, L.J., Nemitz, E., et al. (2006) Eddy covariance ﬂuxes of peroxyacetyl nitrates (PANs) and NOy to a coniferous forest. Journal of Geophysical Research, 111: D09304. Tyndall, G.S., Staffelbach, T.A., Orlando, J.J. and Calvert, J.G. (1995) Rate coefﬁcients for the reactions of OH radicals with methylglyoxal and acetaldehyde. International Journal of Chemical Kinetics, 27: 1009–20. Tyndall, G.S., Orlando, J.J., Wallington, T.J. and Hurley, M.D. (1997) Pressure dependence of the rate coefﬁcients and product yields for the reaction of CH3 CO radicals with O2 . International Journal of Chemical Kinetics, 29: 655–63. Tyndall, G.S., Cox, R.A., Granier, C., et al. (2001a) Atmospheric chemistry of small organic peroxy radicals. Journal of Geophysical Research, 106: 12157–82. Tyndall, G.S., Orlando, J.J., Wallington, T.J. and Hurley, M.D. (2001b) Products of the chlorine-atom and hydroxyl-radical-initiated oxidation of CH3 CN. Journal of Physical Chemistry A, 105: 5380–4. Villalta, P.W., Lovejoy, E.R. and Hanson, D.R. (1996) Reaction probability of peroxyacetyl radical on aqueous surfaces. Geophysical Research Letters, 23: 1765–8. Volz-Thomas, A., Xueref, I. and Schmitt, R. (2002) Automatic gas chromatograph and calibration system for ambient measurements of PAN and PPN. Environmental Science and Pollution Research, 9: 72–6. Wallington, T.J., Andino, J.M., Ball, J.C. and Japar, S.M. (1990) Fourier transform infrared studies of the reaction of Cl atoms with PAN, PPN, CH3 OOH, HCOOH, CH3 COCH3 , and CH3 COC2 H5 at 295 ± 2K. Journal of Atmospheric Chemistry, 10: 301–13. Wallington, T.J., Sehested, J. and Nielsen, O.J. (1994) Atmospheric chemistry of CF3 C(O)O2 radicals. Kinetics of their reaction with NO2 and kinetics of the thermal decomposition of the product CF3 C(O)O2 NO2 . Chemical Physics Letters, 226: 563–9. Wallington, T.J., Schneider, W.F., Mogelberg, T.E., Neilsen, O.J. and Sehested, J. (1995) Atmospheric chemistry of FCOx radicals: Kinetic and mechanistic study of the FC(O)O2 + NO2 reaction. International Journal of Chemical Kinetics, 27: 391–402. Wang, Y., Jacob, D.J. and Logan, J.A. (1998) Global simulation of tropospheric O3 -NOx -hydrocarbon chemistry, 1. Model formulation. Journal of Geophysical Research, 103: 10713–25. Warneck, P. and Zerbach, T. (1992) Synthesis of peroxyacetyl nitrate in air by acetone photolysis. Environmental Science and Technology, 26: 74–9. Warneke, C. (2004) Comparison of day and nighttime oxidation of biogenic and anthropogenic VOCs along the New England Coast in Summer during NEAQS 2002. Journal of Geophysical Research, 109: D10309. Wendell, G.J., Stedman, D.H., Cantrell, C.A. and Damrauer, L. (1983) Luminol-based nitrogen dioxide detector. Analytical Chemistry, 55: 937–40. Williams II, E.L. and Grosjean, D. (1991) Peroxypropionyl nitrate at a Southern California mountain forest site. Environmental Science and Technology, 25: 653–9. Williams, E.L., Grosjean, E. and Grosjean, D. (1993) Ambient levels of the peroxyacyl nitrates, PAN, PPN, and MPAN in Atlanta GA. Journal of Air and Waste Management Association, 43: 873–9. Williams, J., Roberts, J.M., Fehsenfeld, F.C., et al. (1997) Regional ozone from biogenic hydrocarbons deduced from airborne measurements of PAN, PPN, and MPAN. Geophysical Research Letters, 24: 1099–102.

268

Volatile Organic Compounds in the Atmosphere

Williams, J., Roberts, J.M., Bertman, S.B., et al. (2000) A method for the airborne measurement of PAN, PPN and MPAN. Journal of Geophysical Research, 105: 28943–60. Zabel, F., Kirchner, F. and Becker, K.H. (1994) Thermal decomposition of CF3 C(O)O2 NO2 , CClF2 C(O)O2 NO2 , CCl2 FC(O)O2 NO2 , and CCl3 C(O)O2 NO2 . International Journal of Chemical Kinetics, 26: 827–45. Zedda, D., Keigley, G.W., Joseph, D.W. and Spicer, C.W. (1998) Development of a new high sensitivity monitor of peroxyacetyl nitrate and results from the West-central Mediterranean region. In: C.A. Brebbia, C.F. Ratto and H. Power (Eds) Air Pollution VI. WIT Press, Southampton, UK, pp. 80–8. Zhang, R. and Leu, M.-T. (1997) Heterogeneous interaction of peroxyacetyl nitrate with liquid sulfuric acid. Journal of Geophysical Research, 102: 8837–43.

Volatile Organic Compounds in the Atmosphere Edited by Ralf Koppmann Copyright © 2007 by Blackwell Publishing Ltd

Chapter 7

Organic Nitrates Paul B. Shepson

7.1

Introduction

An unwanted side result of the Earth’s atmosphere’s natural photochemical cleansing mechanism is the production of tropospheric ozone. Tropospheric ozone concentrations in the northern hemisphere have increased over the past 100 years by approximately two to three times (Crutzen and Lelieveld 2001; Marenco et al. 1994; Volz and Kley 1988), although there is a great deal of regional and temporal variability (Fiore et al. 2003). It is expected to increase slowly in this century, in signiﬁcant part due to increasing NOx emissions in South-east Asia and the Indian subcontinent (Lelieveld and Dentener 2000). Ozone in the troposphere is produced as a result of the oxidation of NO to NO2 by organic peroxy radicals (cf. Hauglustaine et al. 1996) and by HO2 . Since NOx is required for ozone production, the fate and lifetime of NOx is important to understanding the global distribution of ozone and its precursors. Organic peroxy radicals are produced via OH-induced attack on volatile organic compounds (VOCs), which then react with NO, NO2 or other peroxy radicals, as shown in reactions 1–7. OH + VOC(+O2 ) → RO2

(1)

RO2 + NO → ROONO∗

(2)

ROONO∗ → RO• + NO2

(3a)

ROONO∗ (+M) → RONO∗2 (organic nitrate)

(3b)

RO2 + NO2 ↔ ROONO2 (peroxynitrate)

(4,−4)

RO + O2 → R CHO + HO2

(5)

R CHO + OH(+O2 ) → R C(O)OO• + H2 O

(6)

R C(O)OO• + NO2 ↔ R C(O)OONO2 (peroxyacyl nitrate)

(7,−7)

In the process, a variety of organic nitrogen compounds can be produced that are either temporary or permanent sinks for NOx . As shown in reaction sequence 1–7, three different types of organic nitrogen are produced, speciﬁcally organic nitrates of the formula RONO2 (after collisional deactivation, discussed below), peroxynitrates of the formula ROONO2 and peroxyacyl nitrates (PAN) of the formula RC(O)OONO2 (often referred to

270

Volatile Organic Compounds in the Atmosphere

as PAN compounds, see Chapter 6). These three types of organic nitrogen are really quite chemically distinct. Reaction −4 involving thermal decomposition by breaking the weak −N bond is quite fast. For CH3 OONO2 , k−4(298) = 16.8/s, while k−4(228) = 2.8 × 10−4 /s O− (at 8 km, P = 0.37 atm), so at low temperatures, these compounds can be important, while they represent an insigniﬁcant component of odd nitrogen at normal low-latitude surface temperatures. For PAN, CH3 C(O)OONO2 , this species is considerably more stable, with k−7(298) = 6.1 × 10−4 /s (τ = 28 min), while k−7(228) = 5.0 × 10−10 /s (τ = 63 years). So, PAN can be a sink or a source of NOx , depending on transport and temperature. For the organic nitrates, they are effectively inﬁnitely thermally stable (in the gas phase), and so, in the absence of further chemical reaction, production of these species is a NOx sink. Thus, due to their very different properties, it is highly inappropriate to consider them together. Here, we discuss the chemistry of the organic nitrates. While the lifetime and pressure stabilisation kinetics of the vibrationally excited peroxy nitrite (ROONO∗ ) is currently unclear, reaction 3 is believed to occur very fast (Barker et al. 2003), producing a signiﬁcantly vibrationally excited product RONO∗2 in reaction 3b, which can then be either collisionally stabilised, or dissociated: RONO∗2 + M → RONO2

(8)

RONO∗2 → RO + NO2

(9)

Ozone is produced as a result of reaction 3a, followed by the photolysis of NO2 : NO2 + hν → NO + O(3 P)

(10)

O(3 P) + O2 (+M) → O3

(11)

The alkoxy radical produced in reaction 3a normally reacts with O2 to produce a carbonyl compound and HO2 , as recently discussed by Orlando et al. (2003). HO2 reaction with NO then propagates the chain, as it regenerates the OH catalyst: RO + O2 → RC(O)R + HO2

(12)

HO2 + NO → OH + NO2

(13)

This sequence demonstrates why oxidation of VOCs can produce as much as two ozone molecules per VOC consumed, because 2NO molecules are oxidised in reactions 2 (followed by 3a) and 13. However, in this sequence, the possible production of the organic nitrate is a chain-terminating step, as it consumes both radicals and NOx . The observation that reaction 3b can occur was ﬁrst reported for laboratory experiments in the Atkinson laboratory, for alkyl peroxy radicals, by Darnall et al. (1976). They detected the organic nitrates using gas chromatography (GC). Since that time, organic nitrate production has been observed for alkenes (O’Brien et al. 1998; Shepson et al. 1985), aromatics and haloalkanes (Espada et al. 2005), ethers, glycol ethers (Espada et al. 2005) and aldehydes (Hurst et al. 2003). Alkyl nitrates were ﬁrst detected in the atmosphere by Atlas (1988). Since organic nitrates sequester NOx , measurement of their concentration and fate is important to understanding the fate and long-range transport of NOx (Buhr et al. 1990; Kastler et al. 2000; O’Brien et al. 1997; Ridley et al. 1991). Organic nitrate measurements have been used to assess air mass photochemical age (Bertman et al. 1995) and to infer the distribution of organic peroxy

Organic Nitrates

271

radicals involved in ozone production (O’Brien et al. 1995). Recently, it has been shown that organic nitrates can represent as much as ∼15% of NOy in forest-impacted environments (Day et al. 2003). It is thus clear that understanding organic nitrogen formation chemistry and fate is important. Here, we review what is currently known about the production mechanism, distribution, atmospheric fate and measurement methods for organic nitrates in the atmosphere.

7.2

Production mechanism

Prediction of the fate and distribution of atmospheric nitrogen demands an understanding of the fundamentals of the mechanisms. Here, we discuss what is known about the production mechanism in general terms, for a variety of precursor radical types.

7.2.1

Pressure and temperature dependence

Since NO is a free radical (i.e. odd-electron) species, it, like NO2 , can react with other free radicals to produce an adduct. Thus, reaction 3 above proceeds as shown in Figure 7.1. It is envisioned that the initially formed vibrationally excited peroxy nitrite must be collisionally deactivated, or the energy internally distributed so that the highly strained three-member transition state can develop, rather than proceeding as in reaction 3a above. Since there is a competition between decomposition of the peroxy nitrite via reaction 3a, and rearrangement to produce the nitrate via 3b, and since this depends on the amount and disposition of internal energy in the initially formed peroxy nitrite, one would expect a positive pressure dependence to the branching ratio k3b /k3 , and a negative temperature dependence to this ratio. Indeed, Atkinson et al. (1983, 1984) have reported the temperature and pressure dependence for a number of alkyl peroxy radicals. As an example, for 2-pentyl peroxy radicals, the room temperature branching ratio (deﬁned in practice as the fraction of RO2 + NO reactions that lead to RONO2 production) increases from about 3% at 2 × 1018 molecules/cm3 to about 13% at 2.5 × 1019 molecules/cm3 . For these radicals, the branching ratio also increases slightly with decreasing temperature. In Figure 7.2 at right, we show the measured branching ratio for 2-pentyl peroxy radicals as a function of pressure and temperature, using the data presented in Carter and Atkinson (1989). It is clear that the reaction path depicted in Figure 7.1 to produce the RONO2 species is complex and poorly understood. Barker et al. (2003) apply statistical reaction dynamics

R

O

O• + •N=O → R O

R

O

N

← O R

Figure 7.1 Simpliﬁed nitrate production mechanism.

∗

O

O N=O M↓

O O

N=O

272

Volatile Organic Compounds in the Atmosphere

0.18 0.16

284 K

0.14

300 K

k3b/k3

0.12 0.10 327 K

0.08

337 K 0.06 0.04 0.02 0.00 0.0

0.5

1.0 Total pressure

1.5 molecules/cm3

2.0

2.5

(×1019 )

Figure 7.2 P and T dependence of branching ratio for 2-pentyl peroxy radicals. Adapted from Carter and Atkinson (1989).

calculations (master equations) applied to the peroxy nitrite reactions, to simulate the measured pressure dependence of the alkyl nitrates. They ﬁnd that the pernitrite is very short-lived and reacts via a rather loose transition state. The pressure dependence of the branching ratios was suggested to result from collisional stabilisation of the vibrationally excited nitrate, which seems to be surprisingly ineffective, that is, the yields are pressure dependent even for relatively large nitrates such as the pentyl nitrates. Zhang et al. (2002), however, present quite a different interpretation of the existing pressure and temperature dependence of the yields (as measured by the Atkinson group), which involves peroxy nitrite rearrangement to form the nitrate from only the trans-conformer (but not the cis-, which they assume only decomposes via reaction 3a). They also assume that the observed pressure dependence of the yields arises from stabilisation of the peroxy nitrite, rather than of the excited nitrate. In their model, the barrier to isomerisation lies below the dissociation (to RO + NO2 ) energy. While the Zhang et al. (2002) model involves a more slowly reacting ROONO intermediate, and a tighter transition state than that assumed by Barker et al. (2003), it does a better job at ﬁtting the low- and high-pressure extrema in the data. They assume that the high pressure limiting branching ratio is deﬁned by the cis–trans branching ratio. The apparent discrepancy in treatment of this system by the two groups reﬂects the fact that neither model is well constrained by observational data, in particular at very high and low pressures, and low temperatures. Work by Ellison et al. (2004) using FONO as a more tractable model of ROONO involved the use of coupled cluster ab initio methods to investigate rearrangement of FONO to FNO2 . Their results were consistent with a model involving a rearrangement barrier (of ∼22 kcal/mol) and two interacting diabatic states that correlate the trans-FONO to the ground and excited states of NO2 . While this study helps us understand the possible nature of the transition state, it

Organic Nitrates

273

is clear that experimental data regarding the structures, spectroscopy and reactivity of the peroxy nitrite intermediates themselves are critically needed. Interestingly, Dahl et al. (2003) reported measured values for the branching ratios for alkyl nitrate formation in aqueous solution, and obtained values of 0.23 ± 0.04, 0.67 ± 0.03, 0.47 ± 0.03 and 0.24 ± 0.02 for methyl, ethyl, 2-propyl, and n-propyl peroxy radicals, respectively, that is, an order of magnitude or more larger than in the gas phase at atmospheric pressure. Since the Henry’s law constants for simple light alkyl nitrates are so small (2.0 and 1.6 M/atm for methyl and ethyl nitrates, respectively (Kames and Schurath 1992), once produced, the light alkyl nitrates would rapidly evolve to the gas phase. This may then help explain the relatively large atmospheric concentrations of light alkyl nitrates, as discussed below. If the model of Zhang et al. (2002) is correct, these very large branching ratios may reﬂect the fraction of the peroxy nitrite that is in the trans-conformation. The apparent zero-pressure intercept in Figure 7.2 is discussed by Zhang et al. (2002) as a result of a ‘double fall-off ’ in the pressure dependence, from excited nitrate stabilisation at very low pressures, and peroxy nitrite stabilisation at higher pressures.

7.2.2

Dependence of branching ratios on peroxy radical structure

The branching ratio k3b /k3 has been measured for many peroxy radicals at atmospheric pressure (discussed here as if all RONO∗2 will be stabilised rather than dissociate, as assumed by Zhang et al. (2002) for all but very low pressures), and is dependent on the ability to distribute the reaction energy of the peroxy nitrite (or excited nitrate) away from the critical dissociative vibrational mode. Thus, the branching ratio is dependent on the number of vibrational modes in, or size of, the organic peroxy radical. For simple alkyl peroxy radicals, this has been recently discussed by Arey et al. (2001), who ﬁnd that the branching ratio −CH2 − − group, from 0.039 for 2-propyl peroxy to 0.235 for 2-octyl increases by ∼30% per − peroxy radicals. For ethyl peroxy radicals, the Ranschaert et al. (2000) data leads to a branching ratio of only 0.0078 at 298 K and the experimental pressure of 100 torr. This value is consistent with that reported by Atkinson et al. (1982a) of ≤0.014. Unfortunately, the pressure dependence for this peroxy radical has not been studied. For methyl peroxy radicals, the branching ratio is very small; Barker et al. (2003) estimate a value of 0.0005, which is consistent with the estimate from Flocke et al. (1998) of 0.00015–0.0003. It has been shown that the branching ratio k3b /k3 is signiﬁcantly impacted by substituents on the R group. O’Brien et al. (1998) have shown that peroxy radicals with a β-hydroxy group have a branching ratio for organic nitrate formation that is ∼0.3–0.5 that for the unsubstituted analogues. For example, the branching ratio for 2-hexyl peroxy radicals is 0.15, while it is 0.055 for 1-hydroxy-2-hexyl peroxy radicals. As discussed in O’Brien et al. −O bond in the peroxy nitrite, (1998), this may occur as a result of the weakening of the O− through hydrogen bond formation. This effect is very important, as it is relevant to atmospheric oxidation of alkenes, which occurs through OH and O2 addition across the double bond (Atkinson 2000), and because, globally, alkenes are often the most important VOCs in the boundary layer (Cantrell et al. 1992; Guenther et al. 1995; Sumner et al. 2001). It is particularly important to determine the branching ratio for biogenic VOCs, for which there are ample amounts of contradictory information. For isoprene, a very important biogenic VOC, Tuazon and Atkinson (1990) reported an organic nitrate yield from OH

274

Volatile Organic Compounds in the Atmosphere

reaction with isoprene in the range 8–13%, using Fourier transform infrared (FTIR) determ−NO2 stretching ination, using an average value for the absorption coefﬁcient for an RO− mode. Chen et al. (1998) reported a total ‘isoprene nitrate’ yield of 4.4%, using GC with a ‘luminol nitrate detector’. Chuong et al. (2002) estimated the isoprene nitrate yield as 15%, by simulating the radical propagation from OH reaction with isoprene in a ﬂow tube study. Sprengnether et al. (2002) reported an isoprene nitrate yield of 12% in a high-pressure ﬂow tube study, using FTIR detection of the ‘prompt’ nitrates. Interestingly, however, a yield this large seems inconsistent with the results of O’Brien et al. (1998), since most of the isoprene nitrates are of the α,β-hydroxy nitrate variety (Giacopelli et al. 2005), and the branching ratios would thus be expected to be smaller than those for similar-size alkyl peroxy radicals (i.e. about 11% for pentyl peroxy radicals; Arey et al. 2001). For α-pinene, the yield was determined by Nozière et al. (1999) to be ∼18%, while Aschmann et al. (2002) reported a nitrate yield of ∼1%. It is not clear what may be the cause of these discrepancies, although it is clear that OH reaction with these oleﬁnic VOCs produces multiple isomeric nitrate products (Giacopelli et al. 2005), which are adsorptive, reactive and difﬁcult to detect. Clearly, more laboratory work is necessary on the production yields for these compounds. Other substituents also have an impact on the branching ratios. For example, Espada et al. (2005) report that the branching ratio for 1-bromo-2-propyl peroxy radicals is 0.022, while that for 2-propyl peroxy radicals is 0.039 (Arey et al. 2001). Espada et al. (2005) −O bond via an inductive effect from the Br atom. attribute this to a weakening of the O− −OR substituent can also lead to lower values for Espada et al. (2005) found that an ether − k3b /k3 , if the ether oxygen is alpha to the peroxy nitrite group. For example, they found that the branching ratio for CH3 CH2 CH(OO• )OCH3 radicals was 0.006, much smaller than the value of 0.07 for CH3 CH2 CH(OO• )CH3 radicals. This was also attributed to an inductive effect caused by the O atom. −ONO2 group, then they When these substituents are not on a carbon adjacent to the − have the effect of increasing the branching ratio. For example, for Br-CH2 CH2 CH2 OO radicals, the branching ratio is 0.029 (Espada et al. 2005), while Atkinson et al. (1982a) report a value 0.023 for CH3 CH2 CH2 OO• radicals. Espada et al. (2005) report a branching ratio for CH3 OCH2 CH(OO• )CH3 radicals of 0.11, 60% larger than that for CH3 CH2 CH(OO• )CH3 −O− − group has an effect similar to that found for increasing the radicals; in this case, the − number of methylene groups (Arey et al. 2001). A similar impact was found by Espada −OH groups. A summary of these substituent effects is presented et al. (2005) for distant − in Figure 7.3.

7.3

Measurement methods

Alkyl nitrates have been detected and quantiﬁed in the gas phase by a number of methods. Darnall et al. (1976) and the Atkinson group (Atkinson et al. 1982a) detected and identiﬁed these compounds in laboratory experiments using GC/FID and GC/MS techniques. That same group has also conducted measurements of total organic nitrates via FTIR (Tuazon and Atkinson 1990). Although the absorption coefﬁcient is well known, this method only allows determination of the sum of the organic nitrate concentrations. Most −NO2 band centred at ∼1290/cm, using of the measurement methods have relied on the O−

Organic Nitrates

0.30

275

• Arey et al. (2001) (secondary alkyl ROO ) O' Brien et al. (1998) ( – OH secondary ROO•) • Etthers, ---- OO is not

0.25

-peroxy ethers

k3b/k3 for ROO•/S

Br-alkyl peroxy radicals

0.20

0.15 • CH3CH(OO )CH2OCH3

0.10

C4H9OCH(OO•)CH3

0.05

• C4H9OCH(OO )CH2OH

•

CH2BrCH2CH2OO CH2BrCH(OO•)CH3

0.00

CH3OO•

1

2

3

4

5

6

7

8

9

Number of carbons −OO Figure 7.3 Atmospheric pressure branching ratios vs radical size. For alkyl peroxy radicals, the − group is at C-2.

integrated absorption coefﬁcients for this band of 2.8 × 10−17 cm/molecule (Tuazon and Atkinson 1990) and 2.4 × 10−17 cm/molecule (Sprengnether et al. 2002) for this mixed nitrate stretch feature. As discussed by Tuazon and Atkinson (1990), the absorption coefﬁcient seems quite independent of substituents. While this method is useful for laboratory studies, it does not have the low ppt limit of detection needed for ambient measurements. For ambient measurements, the most common method has been capillary GC with electron capture detection (Atlas 1988; Buhr et al. 1990; Giacopelli et al. 2005; Kastler et al. 2000; O’Brien et al. 1995). The electron capture cross-sections are quite large for organic nitrates, leading to excellent limits of detection. For example, Giacopelli et al. (2005) have reported limits of detection for individual isoprene nitrates of less than 1 ppt. Relative electron capture detector (ECD) sensitivities have been presented for some alkyl and hydroxy alkyl nitrates by Muthuramu et al. (1993). All of the ambient measurement methods to date have involved chromatography, because of the low concentrations and wide variety of organic nitrates that are generally detectable (Fischer et al. 2000; O’Brien et al. 1995). In most cases, the analytical column of choice has been a bonded semi-polar phase, for example, a bonded version of OV1701. In most cases, direct injection provided sufﬁcient limits of detection (cf. Bertman et al. 1995), although the original method of Atlas (1988) involved preconcentration into small charcoal tubes. Some investigators have conducted preconcentration using Tenax (Fischer et al. 2000; Grossenbacher et al. 2001) and silica gel (Schneider et al. 1998). While detection has been most frequently conducted via electron capture detection, negative ion chemical ionisation mass spectrometry has been successfully employed for the determination of organic nitrates (Calderara et al. 2004). Flocke et al. (1991) used an Au-tube-based NOy instrument

276

Volatile Organic Compounds in the Atmosphere

for detection of alkyl nitrates in the Schauinsland. Hao et al. (1994) have described the development of the postcolumn pyrolyser that quantitatively converts organic nitrates to NO2 , allowing selective detection of organic nitrates as NO2 via luminol chemiluminescence. That method was used for ambient measurements of alkyl and hydroxy nitrates (O’Brien et al. 1995, 1997) and isoprene nitrates (Grossenbacher et al. 2001, 2004). Thermal decomposition to NO2 has recently been used by Day et al. (2003) to determine the sum of all organic nitrates via a series of heated inlets, followed by laser-induced ﬂuorescence (LIF) detection of the NO2 . It is relatively straightforward to calibrate for simple alkyl nitrates and hydroxy nitrates through synthesis of authentic standards. Simple alkyl nitrates can be prepared either from the nitrate substitution of alkyl bromides (Flocke et al. 1991; Muthuramu et al. 1993), or from the nitration of alcohols (Kames et al. 1993; Roberts 1990). Muthuramu et al. (1993) have described preparation of α,β-hydroxynitrates via HNO3 ring-opening of the appropriate epoxide. Hydroxy nitrates have also been prepared using mild nitration reagents applied to diols by Treves et al. (2000).

7.4

Atmospheric measurements

To date, there have been numerous measurements of organic nitrates in the atmosphere, although the vast majority of measurements have been of the simple alkyl nitrates. As it is impossible to discuss them all here, we summarise in Table 7.1 the measurement campaigns that included organic nitrate quantitation, with information about the measurement location and nature of the species measured. As discussed by O’Brien et al. (1995), the production rate for a speciﬁc organic nitrate (i) can be expressed as in Equation 7.1, where αi is the fraction of time the appropriate RONO2 (i)Production Rate = k1 · [OH] · [VOC] · αi · (k3b /k3 )i · βi

(7.1)

precursor radical is produced from OH reaction with the precursor VOC, and βi is the fraction of the time the precursor RO2 radical reacts with NO. Thus, the rate depends in general on the product kOH · [VOC], on the size of the RO2 radical, and on the fraction of the time in which the precursor RO2 is produced. At very low [NOx ] conditions, the organic nitrate production rate can be low because of reactions such as 14 and 15 below (Grossenbacher et al. 2004). RO2 + HO2 → ROOH + O2

(14)

RO2 + RO2 → 2RO + O2

(15)

This equation helps explain some of the details of the observations described here.

7.4.1

Alkyl nitrates

Atlas (1988) reported the ﬁrst measurements of alkyl nitrates in the atmosphere, via samples obtained from a ship cruise in the North Paciﬁc Ocean region. He identiﬁed a series of

Organic Nitrates

277

−C7 alkyl nitrates, and suggested that such nitrates would be a relatively stable NOx C3 − reservoir. Interest in measurements of organic nitrates increased signiﬁcantly after reports (Fahey et al. 1986; Parrish et al. 1992) that a full accounting of the components of total reactive nitrogen, that is, NOy , was not possible, leading to the conclusion that there was a component of NOy that was ‘missing’. Buhr et al. (1990) found a ∼15% shortfall in NOy at Scotia, PA, and found that simple propyl–pentyl nitrates represented 1–2% of NOy . They found that the shortfall maximised in mid-afternoon, and concluded that the shortfall −OH substituted) organic nitrates. species behaved as expected for multifunctional (e.g. − −C8 alkyl nitrates in the Black Forest, Flocke et al. (1991) conducted measurements of C1 − and also found that these compounds represented 1–2% of NOy on average, although a fraction as high as 7% was measured. They used the regression of the sum of alkyl nitrates vs O3 to conclude that HO2 and CH3 OO are the likely dominant ozone precursors. However, this conclusion requires that all organic nitrates be identiﬁed. Atlas et al. (1992) measured propyl–pentyl alkyl nitrates at Mauna Loa, ﬁnding a total concentration ranging from ∼1 to 10 ppt, and again accounting for 1–2% of NOy , which also cannot account for the missing NOy at Mauna Loa. Bertman et al. (1995) have shown that the ratio [alkyl nitrate]/[parent alkane] can be used as a photochemical clock. Furthermore, they used deviations from the theoretical behaviour to infer that there are additional sources of the light alkyl nitrates, for example, ethyl nitrate, beyond OH + ethane. That issue has been discussed by others, for example, Flocke et al. (1998) and O’Brien et al. (1995), who have hypothesised that smaller alkyl radicals are produced from decomposition of larger alkoxy radicals (Orlando et al. 2003). However, it is also possible to produce them from decomposition of peroxyacyl radicals, for example, as shown in reactions 16 and 17 below, for the case of peroxypropionyl radical conversion to ethyl radicals. CH3 CH2 C(O)OO• + NO → CH3 CH2 C(O)O• + NO2 •

CH3 CH2 C(O)O →

CH3 CH•2

+ CO2

(16) (17)

The contribution of alkyl nitrates to NOy is much larger in polar environments. Bottenheim et al. (1993) found that alkyl nitrates represented as much as 20% of NOy in the remote Arctic troposphere. The same result was found by Muthuramu et al. (1993), that is, alkyl nitrates accounted for 7–20% of NOy during the polar sunrise period of 1992. This muchlarger contribution of simple alkyl nitrates to NOy is to a small extent related to greater production yield at lower temperatures. However, it largely results from the fact that during long-range transport of polluted air masses to Alert, polar, soluble NOy components such as HNO3 and HO2 NO2 are removed via dry and wet deposition, and NO and NO2 are slowly converted to removable products via dark reaction with O3 . Thus, in the Arctic winter, most of what remains are the water insoluble PAN and alkyl nitrates (Honrath et al. 1996). Muthuramu et al. (1993) used the alkyl nitrate decay after polar sunrise at Alert to infer their destruction via Cl atom reaction. Talbot et al. (2000) show clear evidence for a marine source of alkyl nitrates (largely methyl and ethyl nitrate), from measurements in the South Paciﬁc Ocean. This is consistent with the measurements of Dahl et al. (2003), who found that irradiated seawater samples produced ethyl-, 2-propyl-, and 1-propyl nitrates. Based on the distribution of these nitrates, Dahl et al. suggested that they were produced as a result of OH radical reaction with dissolved organic matter. These observations are quite interesting in light of the observations of

Table 7.1 Summary of ambient measurements of organic nitrates Site location

Measurement dates (month/year)

North Paciﬁc Ocean Scotia, PA (forest) Niwot Ridge, CO Schauinsland, Germany (forest) Julich, Germany Mauna Loa, Hawaii (remote/free troposphere) Equatorial Paciﬁc Ocean Alert, Nunavut (Arctic) Egbert, Ontario (rural) Arctic Ocean (aircraft) Alert, Nunavut Kinterbish, AL Hastings, ON (rural)

4–7/1986 7–8/1988 6–7/1987 5–6/1989 11/1988 to 4/1989 5–6/1988

Poker Flats, AK (Arctic) Schauinsland, Germany Chebogue Pt, NS Eastern North America downwind Hawaii (aircraft) Rural/suburban Northern California Atlantic Ocean (ship)

3–5/1993 6/1990 to 5/1991 8–9/1993

2–3/1990 3–4/1988 3–4/1990 4/1992 1–4/1992 5–6/1992 8/1992

Species measured∗

Concentration range (ppt)

Percentage of NOy

−C5 C4 − −C5 C2 − n-C3 , 2-C4 −C5 C1 − −C8 C1 − −C5 C3 −

0.8–37 2–200 0.4–12 5–180 35–530 0.9–11

— 1–1.5 0.5 1.0 — 1.4

−C5 C2 − −C7 C3 − −C6 C3 − −C6 C3 − −C7 C3 − −C5 C2 − −C6 alkyl C3 − −C4 hydroxy nitrates C2 − −C6 C2 − −C8 C1 − −C5 C1 −

2–19 92–251 29–87 36–61 34–128 10–154 12–140

— ∼20 1.8 — 6–20 0.6–4.2 0.5–3.0

Atlas et al. (1993) Bottenheim et al. (1993) Shepson et al. (1993) Leaitch et al. (1994) Muthuramu et al. (1993) Bertman et al. (1995) O’Brien et al. (1995)

11–66 30–630 4–50

— 0.5–10 —

Beine et al. (1996) Flocke et al. (1998) Roberts et al. (1998)

4–5/1992 8/1995

−C5 C1 − −C12 C3 −

4–30 3–54

<5% —

5–6/1994

−C12 C3 −

3–11

—

Reference

Atlas (1988) Buhr et al. (1990) Ridley et al. (1990) Flocke et al. (1991) Flocke et al. (1991) Atlas et al. (1992)

Ridley et al. (1997) Schneider et al. (1998) Schneider and Ballschmiter (1999)

Suburban Vancouver

7–8/1993

Ulm, Germany North Atlantic cruise

5/1998 10–11/1996

Ulm, Salt Lake City, Las Vegas

North Atlantic (NARE) Pellston, MI (PROPHET) Pellston, MI (PROPHET) Atlantic Ocean cruise Arctic (aircraft/TOPSE) Blodgett Forest, CA Houston, TX Summit, Greenland (glacier) Paciﬁc Ocean (TRACE-P) Rural Tennessee Pellston, MI (PROPHET) Tropical Paciﬁc Ocean Hong Kong suburban

5–8/1998

1997 7–8/1998 7–8/1998 9–10/1999 and 9–10/2000 2–5/2000 10/2000 to 12/2001 8–9/2000 6/1997 to 6/1998 2–4/2001 6–7/1999 7–8/2000 5–7/2004 8/2001 to 12/2002

∗ Refers to alkyl nitrates, unless otherwise speciﬁed.

−C6 alkyl nitrates C2 − and hydroxy nitrates Isoprene nitrates −C13 alkyl nitrates, C1 − −C6 dinitrates, C3 − −C6 hydroxynitrates C2 − −C13 alkyl nitrates, C6 − −C6 dinitrates, C3 − −C6 hydroxynitrates C2 − −C5 C3 − −C5 C3 − Isoprene nitrates Methyl and ethyl nitrate −C4 C1 − Total organic nitrates Total organic nitrates −C4 C1 − −C5 C1 − Isoprene nitrates Isoprene nitrates −C3 C1 − (air and ocean) −C5 C1 −

40–570

0.1–2.5

O’Brien et al. (1997)

— 1–3

— —

Werner et al. (1999) Fischer et al. (2000)

12–50

—

Kastler et al. (2000)

∼1–200 3–66 0.5–35 3–52 9–50 ∼50–700 ∼0–3 000 9–34 2–200 0–200 0–80 ∼6–78

— 0.5–3.0 0.5–4.0 — — ∼ 10–20 ∼0–11 — — ∼0–5 — —

7–229

1–6

Stroud et al. (2001) Ostling et al. (2001) Grossenbacher et al. (2001) Chuck et al. (2002) Blake et al. (2003) Day et al. (2003) Rosen et al. (2004) Swanson et al. (2003) Simpson et al. (2003) Grossenbacher et al. (2004) Giacopelli et al. (2005) Dahl et al. (2005) Simpson et al. (2006)

280

Volatile Organic Compounds in the Atmosphere

Swanson et al. (2002), who found evidence of substantial production of light alkyl nitrates in the near-surface snowpack at Summit, Greenland. The observed distribution of light alkyl nitrates was very much consistent with what Dahl et al. (2003) found, and implies that the same photochemistry is occurring in the snowpack quasi-liquid layer, again involving dissolved organic matter (cf. Dominé and Shepson 2002).

7.4.2

Multifunctional organic nitrates

O’Brien et al. (1995, 1997) ﬁrst reported ambient measurements of the hydroxy nitrate products of OH reaction with alkenes. They measured the hydroxy nitrates from ethene, propene and the butenes; however, the sum of 18 different alkyl and hydroxy nitrates still did not account for more than 3% of NOy . O’Brien et al. (1995) showed that the measurements of the organic nitrates, taken with their production yield, can be used to determine the relative importance of organic peroxy radicals in ozone production. They also used the measured VOC concentrations at their rural Ontario site to calculate that isoprene nitrates should, in the summer time continental boundary layer, be the dominant organic nitrates. Fischer et al. (2000) and Kastler et al. (2000) have reported an impressive array of alkyl and multifunctional organic nitrate concentrations for both urban and remote marine locations. Fischer et al. (2000) reported the latitudinal gradients, from 67◦ N to 50◦ S in an Atlantic Ocean cruise from the Polarstern. In general, levels of light alkyl nitrates were very low in equatorial regions, and much higher in high North and South latitudes, due to faster destruction rates, and lower production rates in low latitudes. They also ﬁnd relatively high concentrations of long-chain primary (i.e. compared to secondary nitrates) alkyl nitrates in aged marine air. Hurst et al. (2003) discuss how 1-octyl nitrate can be produced from the oxidation of nonanal, which has been detected in a wide variety of environments, and is believed to be biogenic in origin. Fischer et al. (2000) also identiﬁed 24 alkyl dinitrates, which can be produced from addition of NO3 to alkenes (Shepson et al. 1985), and from OH abstraction from alkyl nitrates. They also identiﬁed seven different hydroxy nitrates. As discussed above, for continental environments during the growing season, one of the most prevalent types of organic nitrates should be isoprene nitrates, produced, for example, as shown in Figure 7.4. Werner et al. (1999) ﬁrst identiﬁed isoprene nitrates in ambient air. Grossenbacher et al. (2001, 2004) conducted quantitative measurements of

+ NO

OH + O2 OO•

ONO2

+ OH

OH Figure 7.4 An example of the isoprene nitrate production mechanism.

OH

Organic Nitrates

281

isoprene nitrates in 1998 as part of the Program for Research on Oxidants: Photochemistry, Emissions, and Transport (PROPHET) campaign in northern Michigan, and in 1999, as part of the Southern Oxidants Study in Tennessee. They found that isoprene nitrates represented as much as ∼5% of NOy , although often considerably less than expected. The sum of the isoprene nitrates and alkyl nitrates at the PROPHET site still represented only ∼1–4% of NOy (Ostling et al. 2001; Thornberry et al. 2001). However, Giacopelli et al. (2005) discussed that this is likely, at least in part, due to the high reactivity of the isoprene nitrates, due to the remaining double bond. Indeed, they discuss that much of the organic nitrate present at the PROPHET site may be secondary nitrates produced from the OH- and O3 -induced oxidation of the primary isoprene nitrates. Day et al. (2002) developed a method for determining the total concentration of organic nitrates in air. The method involves thermal decomposition of the nitrates, followed by laser-induced ﬂuorescence detection of the NO2 produced. This has been applied to measurements in a variety of environments (Day et al. 2003; Rosen et al. 2004). Day et al. (2003) found that, at the Blodgett forest site in California, organic nitrates represented as much as 15% of NOy during summer months. This is much larger than what has been observed under similar environmental conditions, from summing the individually determined organic nitrates, and implies that much of the organic nitrates, and thus much of the organic nitrate precursors, are unidentiﬁed. That much of the reactive VOCs in ambient air are also unidentiﬁed is becoming clear. DiCarlo et al. (2004) measured the total ‘OH reactivity’ (via sampling ambient air through a ﬂow tube, and measuring the decay rate of OH radicals within that ﬂow tube), and compared it with the sum of the OH reactivities (i.e. ki,OH· [VOC]i ) from the measured VOC concentrations and found that a signiﬁcant fraction of the source of OH reactivity could not be ascribed to measured VOCs. More importantly, they found that the missing reactivity increased exponentially with ambient temperature, similar to the known exponential temperature-dependence of terpene emissions. At the Blodgett Forest site, Kurpius and Goldstein (2003) found that O3 dry deposition is dominated by gas-phase chemistry at the canopy height, and concluded that this must occur via reaction with terpenoid VOCs. More recently, Holzinger et al. (2004) observed a variety and substantial concentrations of biogenic oxidation products just above the Blodgett Forest canopy, and inferred ‘very reactive’ biogenic VOC (VRBVOC) emission rates as much as an order of magnitude greater than that for the sum of the pinenes. It is speculated that the VRBVOCs may be sesquiterpenes (C15 H24 ). It seems quite plausible that such compounds might form a signiﬁcant part of the unknown organic nitrates measured by Day et al. (2003), since their concentrations are high, and the branching ratio (k3b /k3 ) for RONO2 formation for sesquiterpenes is likely to be quite large. As an example, we show in Figure 7.5 a partial oxidation mechanism for β-caryophyllene, an example of a VRBVOC. When OH reacts with such BVOCs, it will most often (but not solely) add across a double bond and, most frequently, one that is internal. In the case shown in Figure 7.5, a tertiary nitrate is produced. Considerable efforts would need to be applied for the identiﬁcation and quantitative determination of such species. As shown here, it is likely that this oxidation leads to Secondary Organic Aerosol (SOA). Indeed, it is now known that even oxidation of the BVOC isoprene can lead to SOA production, as observed by Kroll et al. (2005). It is possible that organic nitrate formation may be involved in this SOA production. As discussed by Kroll et al. (2005), it is likely that SOA formation from isoprene results from oxidation of both double bonds. So, for

282

Volatile Organic Compounds in the Atmosphere

OO• (+ O2)

+ OH

HO

β-Caryophyllene

+ NO

NO2 O HO

ONO2 O•

.

HO Tertiary hydroxy nitrate

HO

+ OH, NO + O2 HO2

Very low V.P. nitrates, or carbonyl compounds

O O

+ OH, NO2 + O3

Very low V.P. PAN compound

Very low V.P. RCOOH’s Aerosols

Figure 7.5 An example of the mechanism for organic nitrate production from OH + β-caryophyllene.

example, this could occur as shown in reactions 18–20 below, in which a multiple functional group product of very low vapour pressure is produced, which would readily partition to the particle phase. OH + C5 H8 (+O2 ) → HOCH2 C(OO• )(CH3 )CH==CH2

(18)

HOCH2 C(OO• )(CH3 )CH==CH2 + NO → HOCH2 C(ONO2 )(CH3 )CH==CH2 (19) HOCH2 C(ONO2 )(CH3 )CH==CH2 + O3 →→ HOCH2 C(ONO2 )(CH3 )C(O)OH + HCHO

(20)

Quantitative ambient measurement of such species would be quite difﬁcult, and considerable analytical methods development would be necessary to enable such measurements.

7.5

Fate

It is known that organic nitrates can be oxidised in such a way as to release NO2 back to the atmosphere. For example, Grossenbacher et al. (2001) discussed such a mechanism for isoprene nitrates. We have found in experiments conducted in this group that Cl atom reaction with 2-propyl nitrate results in production of NO2 with 26% yield. Furthermore,

Organic Nitrates

283

kOH(× 10−12 cm3/molecule/s)

8

Alkyl nitrates (secondary for C3 and greater) n-alkanes

6

2-nitrooxy-1-propanol and 2-nitrooxy-1-butanol

4

2

0 1

2

3

4

5

6

7

8

Carbon number Figure 7.6 OH rate constants for n-alkanes and the corresponding saturated alkyl nitrates. Data from Nielsen et al., 1991; Vaghjiani and Ravishankara, 1991; Becker and Wirtz, 1989; Atkinson et al., 1982b; Donahue et al., 1998; Treves and Rudich, 2003.

it is believed that uptake of some types of organic nitrogen by vegetation can occur through the stomata, where that nitrogen may be utilised, thus impacting carbon sequestration in forests (Sparks et al. 2003). Thus, knowledge of the atmospheric fate of organic nitrates is important. Organic nitrates can be removed from the atmosphere via photolysis, OH reaction, dry and wet deposition, and, in the case of oleﬁnic nitrates such as those shown in Figures 7.4 and 7.5, reaction with O3 and NO3 . Thus, the time rate of change of [RONO2 ] (in terms of chemical tendency, that is, ignoring mixing and transport) can be expressed as in Equation 7.2, where: d[RONO2 ]/dt = Production Rate kox · [ox]i + JRONO2 + kDD + kWD · [RONO2 ] −

(7.2)

i

Production Rate is from Equation 7.1; ‘ox’ is any of OH, O3 or NO3 ; J is the photodissociation rate coefﬁcient; and kDD and kWD are ﬁrst-order loss frequencies for dry and wet deposition, respectively. For simple alkyl nitrates, photolysis and OH reaction will be the most important loss processes. OH reaction proceeds both by addition to the nitrooxy group, and by hydrogen atom abstraction (Atkinson et al. 1982b; Nielsen et al. 1991). However, the nitrooxy group has a deactivating effect with respect to H-atom abstraction (Neeb 2000). This is shown clearly in Figure 7.6, in which OH rate constants are plotted −ONO2 group for n-alkanes, and the corresponding simple n-alkyl nitrates (where the − is at C-2). Thus, the OH rate constant is 50 times faster for methyl nitrate compared to

284

Volatile Organic Compounds in the Atmosphere

First-order loss frequency (× 10 −6)/s

3.5 kOH[OH]/s

3.0

J, alkyl nitrate/s J, hydroxy nitrate/s

2.5 2.0 1.5 1.0 0.5 0.0 1

2

3

4 Carbon number

5

6

7

Figure 7.7 Comparison of photolysis frequencies (40◦ N, 1 July) and OH reaction rate for alkyl nitrates. Photolysis frequencies from Nielsen et al., 1991; Roberts and Fajer, 1989; Atkinson et al., 1982a; Becker and Wirtz, 1989.

methane, but when abstraction becomes more important than addition at the C3 ’s, the OH rate constants for the simple alkanes are faster. For the hydroxy nitrates produced from OH addition to alkenes (or from alkoxy radical rearrangement), the OH rate constants are considerably larger, as discussed by Treves and Rudich (2003), and shown in Figure 7.6, for the hydroxy nitrates produced from OH addition to the terminal carbon in propene and 1-butene. If we assume the global average [OH] of 1.0 × 106 (Prinn et al. 1995), we can compare the ﬁrst-order loss rates by OH reaction and photolysis for the simple alkyl nitrates, as a function of size. This is shown in Figure 7.7, in which we present the photodissociation rate coefﬁcient as calculated for 40◦ N and 1 July. These data are for linear alkyl nitrates, −ONO2 group at C-2. Up through C4 nitrates, photolysis dominates, with the with the − combined lifetime ranging from 13 (methyl nitrate) to 7 (2-butyl nitrate) days. For larger alkyl nitrates, reaction with OH will dominate. As discussed above, multifunctional nitrates can be produced from NO3 reaction with alkenes (producing dinitrates and carbonyl nitrates), from OH reaction with alkenes, and from δ-H atom internal abstraction by alkoxy radicals (Orlando et al. 2003; Shepson et al. 1985; Treves and Rudich 2003). For the carbonyl nitrates and dinitrates, the photodissociation rate coefﬁcients are generally larger than for the simple alkyl nitrates. Barnes et al. (1993) measured the ultraviolet (UV) spectra for dinitrates and keto-nitrates, and found that the photolysis frequencies are approximately two and ﬁve times faster, respectively, than for the simple alkyl nitrates. This is consistent with the measurements of Roberts and Fajer (1989), who calculated a J -value for nitrooxyacetone for 40◦ N, 1 July, of 1.4 × 10−5 /s,

Organic Nitrates

285

that is, over ten times greater than that for 2-propyl nitrate. Thus, for these compounds, photolysis is likely to be the dominant atmospheric removal mechanism, with atmospheric lifetimes of roughly 6 and 1–2 days, respectively, for dinitrates and ketonitrates. There is relatively little information about the photodissociation rates for hydroxy nitrates. From the data in Roberts and Fajer (1989), the photolysis frequency for nitrooxyethanol is about three times smaller than that for simple alkyl nitrates; however, much more data is needed for these important organic nitrates. For hydroxy nitrates, wet and dry deposition is likely to be very important. This has been discussed by Shepson et al. (1996) for α,β-hydroxy nitrates (i.e. those produced from OH addition to alkenes), and by Treves et al. (2000) for δ-hydroxy nitrates that are produced from alkoxy radical isomerisation. As discussed by these groups, the Henry’s law coefﬁcients for the hydroxy nitrates vary from roughly 10 000 to 40 000 M/atm. These values are much larger than those measured for the simple alkyl nitrates (Kames and Schurath 1992), for which rainout will be an unimportant removal process. The Henry’s law constants for the hydroxynitrates are highly temperature dependent, increasing by a factor ∼1.7 as T decreases from 298 to 293 K (Shepson et al. 1996). Shepson et al. (1996) and Treves et al. (2000) used the method of Brimblecombe and Dawson (1984) to calculate that the wet deposition lifetimes for these species are on the order of 3–7 days. For these compounds, photolysis is relatively unimportant, and their removal is dominated by OH reaction and wet deposition, depending, of course, on atmospheric conditions, with lifetimes ranging from 1 to 3 days. Dry deposition is not well studied, but expected to be a minor (but not unimportant) loss process, compared to wet deposition and OH radical reaction. For the carbonyl nitrates and dinitrates, it is expected that photolysis will be the dominant removal mechanism. For biogenic nitrates with a remaining double bond, reaction with OH and O3 will be the dominant removal mechanism. The relative importance of each depends on whether the double bond is internal or external, as the internal oleﬁns react signiﬁcantly faster with O3 (Aschmann and Atkinson 1994; Grosjean and Grosjean 1996). NO3 reaction is likely to be unimportant for biogenically derived oleﬁnic organic nitrates, because this reaction will not compete with OH and O3 reaction during the day, and, because their lifetimes are very short (Giacopelli et al. 2005), their concentrations will be very low at night when NO3 is relatively high in concentration. For these compounds, it is likely that dry deposition is much more important, as they tend to be conﬁned to the boundary layer near the source of their short-lived precursors; the ﬁrst-order dry deposition loss frequencies may be on the order of 6 × 10−6 /s from the boundary layer. For these compounds, it is important to know their oxidation products and mechanisms, as they are rapidly converted to secondary product nitrates. Thus, in forest environments, much of the organic nitrate may be highly oxidised (Giacopelli et al. 2005).

7.6

Conclusions

It now appears that it is indeed the case that organic nitrates represent a signiﬁcant fraction of atmospheric NOy , depending on how processed the air mass is, and proximity to reactive VOC sources and NOx . It is likely that the importance of the organic nitrates as NOx reservoirs has previously been underestimated, because of the early focus on alkyl nitrates,

286

Volatile Organic Compounds in the Atmosphere

and because of the difﬁculty of quantitatively measuring the enormous suite of organic nitrates that are often present in the atmosphere. Progress in this regard is afforded in the method of Day et al. (2002, 2003); however, identifying the components of their signal represents a signiﬁcant challenge. That challenge may be met through the development of analytical methods that are non-chromatographic, such as modiﬁcations to the proton transfer reaction mass spectrometry (PTRMS) method (Lindinger et al. 2005), that enable multidimensional analysis. However, given the enormous variety of structural types of organic nitrates, it would seem hopeless, with existing technology, to be able to quantify the majority of atmospheric organic nitrates of importance on an individual nitrate basis. Thus, it is very important to emphasise the need for a fundamental understanding of the reaction mechanism, so that a predictive capability with respect to the organic nitrate yields can be achieved. It is thus essential that kinetic data as a function of pressure and temperature for a variety of peroxy radicals be obtained, and that we develop the capability to detect the peroxy nitrite intermediate.

References Arey, J., Aschmann, S.M., Kwok, E.S.C. and Atkinson, R. (2001) Alkyl nitrate, hydroxyalkyl nitrate, −C8 n-alkanes. Journal and hydroxycarbonyl formation from the NOx -air photooxidations of C5 − of Physical Chemistry A, 105: 1020–7. Aschmann, S.M., Atkinson, R. and Arey, J. (2002) Products of reaction of OH radicals with alpha-pinene. Journal of Geophysical Research, 107: (D14). Atkinson, R. (2000) Atmospheric chemistry of VOCs and NOx . Atmospheric Environment, 4: 2063–101. Atkinson, R., Aschmann, S.M., Carter, W.P.L., Winder, A.M. and Pitts J.N. Jr. (1982a) Alkyl nitrate −C8 n-alkanes. Journal of Physical Chemistry, formation form the NOx -air photooxidations of C2 − 86: 4563–9. Atkinson, R., Aschmann, S.M., Carter, W.P.L. and Winer, A.M. (1982b) Kinetics of the gas-phase reactions of OH radicals with alkyl nitrates at 299 ± 2 K. International Journal of Chemical Kinetics, 14: 919–26. Atkinson, R., Carter, W.P.L. and Winer, A.M. (1983) Effects of temperature and pressure on alkyl nitrate yields in the NOx photooxidations of normal-pentane and normal-heptane. Journal of Physical Chemistry, 87: 2012–18. Atkinson, R., Aschmann, S.M., Carter, W.P.L., Winder, A.M. and Pitts J.N. Jr. (1984) Formation of alkyl nitrates from the reaction of branched and cyclic alkyl peroxy-radicals with NO. International Journal of Chemical Kinetics, 16: 1085. Atkinson, R., Aschmann, S.M. and Winer, A.M. (1987) Alkyl nitrate formation from the reaction of a series of branched RO2 radicals with NO as a function of temperature and pressure. Journal of Atmospheric Chemistry, 5: 91. Atkinson, R., Tuazon, E.C. and Aschmann, S.M. (1995) Products of the gas-phase reactions of O-3 with alkenes. Environmental Science and Technology, 29: 1860–66. Atlas, E. (1988) Evidence for ≥C-3 alkyl nitrates in rural and remote atmospheres. Nature, 331: 426–8. Atlas, E., Schaufﬂer, S.M., Merrill, J.T., et al. (1992) Alkyl nitrate and selected halocarbon measurements at mauna loa observatory, Hawaii. Journal of Geophysical Research, 97: 16, 10331–48. Atlas, E., Pollock, W., Greenberg, J., et al. (1993) Alkyl nitrates, nonmethane hydrocarbons, and halocarbon gases over the equatorial Paciﬁc during SAGA-3. Journal of Geophysical Research, 98: 16, 933–47.

Organic Nitrates

287

Barker, J.R., Lohr, L.L., Shroll, R. and Reading, S. (2003) Modeling the organic nitrate yields in the reaction of alkyl peroxy radicals with nitric oxide. 2. Reaction simulations. Journal of Physical Chemistry A, 107: 7434–44. Barnes, I., Becker, K.H. and Zhu, T. (1993) Near UV absorption-spectra and photolysis products of difunctional organic nitrates – possible importance as NOx reservoirs. Journal of Atmospheric Chemistry, 17: 353–73. Becker, K.H. and Wirtz, K. (1989) Gas-phase reactions of alkyl nitrates with hydroxyl radicals under tropospheric conditions in comparison with photolysis. Journal of Atmospheric Chemistry, 9: 419–33. Beine, H.J., Jaffe, D.A., Blake, D.R., et al. (1996) Measurements of PAN, alkyl nitrates, ozone, and hydrocarbons during spring in interior Alaska. Journal of Geophysical Research, 101: 12613–19. Bertman, S.B., Roberts, J.M., Parrish, D.D., et al. (1995) Evolution of alkyl nitrates with air-mass age. Journal of Geophysical Research, 100: 22805–13. Blake, N.J., Blake, D.R., Sive, B.C., et al. (2003) The seasonal evolution of NMHCs and light alkyl nitrates at middle to high northern latitudes during TOPSE. Journal of Geophysical Research, 108. Bottenheim, J.W., Barrie, L.A. and Atlas, E. (1993) The partitioning of nitrogen-oxides in the lower arctic troposphere during Spring 1988. Journal of Atmospheric Chemistry, 17: 15–27. Brimblecombe, P. and Dawson, G.A. (1984) Wet removal of highly soluble gases. Journal of Atmospheric Chemistry, 2: 95–107. Buhr, M.P., Parrish, D.D., Norton, R.B., Fehsenfeld, F.C., Sievers, R.E. and Roberts, J.M. (1990) Contribution of organic nitrates to the total reactive nitrogen budget at a rural Eastern United States site. Journal of Geophysical Research, 95: 9809–16. Calderara, S., Gardebas, D., Martinez, F. and Khong, S.P. (2004) Organic explosives analysis using on column-ion trap EI/NICI GC-MS with an external source. Journal Forensic Science, 49: 1005–8. Cantrell, C.A., Lind, J.A., Shetter, R.E., et al. (1992) Peroxy-radicals in the rose experiment – Measurement and theory, Journal of Geophysical Research, 97: 20671–86. Carter, W.P.L. and Atkinson, R. (1989) Alkyl nitrate formation from the atmospheric photooxidation of alkanes: A revised estimation method. Journal of Atmospheric Chemistry, 8: 165–73. Chen, X., Hulbert, D. and Shepson, P.B. (1998) Measurement of the organic nitrate yield from OH reaction with isoprene. Journal of Geophysical Research, 103: 25563–8. Chuck, A.L., Turner, S.M. and Liss, P.S. (2002) Direct evidence for a marine source of C-1 and C-2 alkyl nitrates. Science, 297: 1151–4. Chuong, B. and Stevens, P.S. (2002) Measurements of the kinetics of the OH-initiated oxidation of isoprene. Journal of Geophysical Research, 107: Art.No. 4162. Crutzen, P.J. and Lelieveld, J. (2001) Human impacts on atmospheric chemistry. Annual Review of Earth and Planetary Sciences, 29: 17–45. Dahl, E.E., Saltzman, E.S. and deBruyn, W.J. (2003) The aqueous phase yield of alkyl nitrates from ROO + NO: Implications for photochemical production in seawater. Geophysical Research Letters, 30. Dahl, E.E., Yvon-Lewis, S.A. and Saltzman, E.S. (2005) Saturation anomalies of alkyl nitrates in the tropical Paciﬁc Ocean. Geophysical Research Letters, 32. Darnall, K.R., Carter, W.P.L., Winer, A.M., Lloyd, A.C. and Pitts J.N. Jr. (1976) Importance of −C6 alkane photooxidations under simulated RO2 + NO in alkyl nitrate formation from C4− atmospheric conditions. Journal of Physical Chemistry, 80: 1948–50. Day, D.A., Wooldridge, P.J., Dillon, M.B., et al. (2002) A thermal dissociation laser-induced ﬂuorescence instrument for in situ detection of NO2 , peroxy nitrates, alkyl nitrates, and HNO3 . Journal of Geophysical Research, 107. Day, D.A., Dillon, M.B., Wooldridge, P.J., et al. (2003) On alkyl nitrates, O−3 , and the ‘missing NOy ’, Journal of Geophysical Research, 108.

288

Volatile Organic Compounds in the Atmosphere

DiCarlo, P., Brune, W.H., Martinez, M., et al. (2004) Missing OH reactivity in a forest: Evidence for unknown reactive biogenic VOCs. Science, 304: 722–5. Dominé, F. and Shepson, P.B. (2002) Air-snow interactions and their impact on atmospheric chemistry. Science, 297: 1506–10. Donahue, N.M., Anderson, J.G. and Demerjian, K.L. (1998) New rate constants for ten OH alkane reactions from 300 to 400 K: An assessment of accuracy. Journal of Physical Chemistry A, 102: 3121–6. Ellison, G.B., McCoy, A.B., Stanton, J.F. and Szalay, P.G. (2004) Unimolecular rearrangement of transFONO to FNO2 . A possible model system for atmospheric nitrate formation. Journal of Physical Chemistry A, 108: 7639–42. Espada, C. and Shepson, P.B. (2005) The production of organic nitrates from atmospheric oxidation of ethers and glycol ethers. International Journal of Chemical Kinetics, 37: 686–99. Espada, C., Grossenbacher, J., Ford, K., Couch, T. and Shepson, P.B. (2005) The production of organic nitrates from various anthropogenic volatile organic compounds. International Journal of Chemical Kinetics, 37: 675–85. Fahey, D.W., Hubler, G. and Parrish, D.D. (1986) Reactive nitrogen species in the troposphere: Measurements of NO, NO2 , HNO3 , particulate nitrate, peroxyacetyl nitrate (PAN), O3 , and total reactive odd nitrogen at Niwot Ridge, Colorado. Journal of Geophysical Research, 91: 9781–93. Fiore, A., Jacob, D.J., Liu, H., Yantosca, R.M., Fairlie, T.D. and Li, Q. (2003) Variability in surface ozone background over the United States: Implications for air quality policy. Journal of Geophysical Research, 108. Fischer, R.G., Kastler, J. and Ballschmiter, K. (2000) Levels and pattern of alkyl nitrates, multifunctional alkyl nitrates, and halocarbons in the air over the Atlantic Ocean. Journal of Geophysical Research, 105: 14473–94. Flocke, F., Volz-Thomas, A. and Kley, D. (1991) Measurements of alkyl nitrates in rural and polluted air masses. Atmospheric Environment, 25: 1951–60. Flocke, F., Atlas, E., Madronich, S., et al. (1998) Observations of methyl nitrate in the lower stratosphere during STRAT: Implications for its gas phase production mechanisms. Geophysical Research Letters, 25: 1891–4. Giacopelli, P., Ford, K. and Shepson, P.B. (2005) A comparison of measured and simulated isoprene nitrates above a forest canopy. Journal of Geophysical Research, 110: D01304. Grosjean, E. and Grosjean, D. (1996) Rate constant for the gas-phase reaction of ozone 1,1disubstituted alkenes. International Journal of Chemical Kinetics, 28: 911–18. Grossenbacher, J.W., Shepson, P.B., Thornberry, T., et al. (2001) Measurements of isoprene nitrates above a forest canopy. Journal of Geophysical Research, 106: 24429–38. Grossenbacher, J.W., Barket, D.J., Shepson, P.B., Carroll, M.A., Olszyna, K.G. and Apel, E. (2004) A comparison of isoprene nitrate concentrations at two forest-impacted sites. Journal of Geophysical Research, 109: D11311. Guenther, A., Hewitt, C.N., Erickson, D., et al. (1995) A global-model of natural volatile organiccompound emissions. Journal of Geophysical Research, 100: 8873–92. Hauglustaine, D.A., Madronich, S., Ridley, B.A., et al. (1996) Observed and model-calculated photostationary state at Mauna Loa observatory during MLOPEX 2. Journal of Geophysical Research, 101: 14681–96. Holzinger, R., Lee, A., PawU, K.T. and Golstein, A.H. (2004) Observations of oxidation products above a forest imply biogenic emissions of very reactive compounds. Atmospheric Chemistry and Physics Discussion, 4: 5345–65. Honrath, R.E., Hamlin, A.J. and Merrill, J.T. (1996) Transport of ozone precursors from the Arctic troposphere to the North Atlantic region. Journal of Geophysical Research, 101: 29335–51. Hurst, J.M., Barket, D.J. and Shepson, P.B. (2003) The atmospheric chemistry of nonanal. Environmental Science and Technology, 37: 2218–25.

Organic Nitrates

289

Kames, J. and Schurath, U. (1992) Alkyl nitrates and bifunctional nitrates of atmospheric interest – Henry law constants and their temperature dependencies. Journal of Atmospheric Chemistry, 15: 79–95. Kames, J., Schurath, U., Flocke, F., et al. (1993) Preparation of organic nitrates from alcohols and N2 O5 for species identiﬁcation in atmospheric samples. Journal of Atmospheric Chemistry, 16: 349–59. Kastler, J., Jarman, W. and Ballschmiter, K. (2000) Multifunctional organic nitrates as constituents in European and US urban photo-smog. Fresenius’ Journal of Analytical Chemistry, 368: 244–9. Kroll, J.H., Ng, N.L., Murphy, S.M., Flagan, R.C. and Seinfeld, J.H. (2005) Secondary organic aerosol formation from isoprene photooxidation. Geophysical Research Letters, 32: L18808. Kurpius, M.R. and Goldstein, A.H. (2003) Gas-phase chemistry dominates O3 loss to a forest, implying a source of aerosols and hydroxyl radicals to the atmosphere. Geophysical Research Letters, 30. Leaitch, W.R., Barrie, L.A., Bottenheim, J.W. et al. (1994) Airborne observations related to ozone depletion at polar sunrise. Journal of Geophysical Research, 99: 499–525. Lelieveld, J. and Dentener, F.J. (2000) What controls tropospheric ozone? Journal of Geophysical Research, 105: 3531–5. Lindinger, C., Pollien, P., Ali, S., Yeretzian, C., Blank, I. and Mark, T. (2005) Unambiguous identiﬁcation of volatile organic compounds by proton-transfer reaction mass spectrometry coupled with GC/MS. Analytical Chemistry, 77: 4117–24. Marenco, A., Gouget, H., Nedelec, P., Pages, J.P. and Karcher, F. (1994) Evidence of a long-term increase in tropospheric ozone from Pic du Midi data series – Consequences – Positive radiative forcing. Journal of Geophysical Research, 99: 16617–32. Muthuramu, K., Shepson, P.B. and O’Brien, J.M. (1993) Preparation, analysis, and atmospheric production of multifunctional organic nitrates. Environmental Science and Technology, 27: 1117–24. Neeb, P. (2000) Structure-reactivity based estimation of the rate constants for hydroxyl radical reactions with hydrocarbons. Journal of Atmospheric Chemistry, 35: 295–315. Nielsen, O.J., Sidebottom, H.W. and Donlon, M. (1991) An absolute and relative rate study of the gas phase reaction of OH radicals and Cl atoms with n-alkyl nitrates. Chemical Physics Letter, 178: 163–70. Nozière, B., Barnes, I. and Becker, K.H. (1999) Product study and mechanisms of the reactions of α-pinene and of pinonaldehyde with OH radicals. Journal of Geophysical Research, 104: 23645–56. O’Brien, J.M., Shepson, P.B., Muthuramu, K., et al. (1995) Measurements of alkyl and multifunctional organic nitratesat a rural site in Ontario. Journal of Geophysical Research, 100: 22795–804. O’Brien, J.M., Shepson, P.B., Biesenthal, T., et al. (1997) Production and distribution of organic nitrates, and their relationship to carbonyl compounds in an urban environment. Atmospheric Environment, 31: 2059–69. O’Brien, J.M., Czuba, E., Hastie, D.R., Francisco, J.S. and Shepson, P.B. (1998) Determination of the −C6 alkenes with OH in the presence of NO. Journal hydroxy nitrate yields from the reaction of C2 − of Physical Chemistry, 102: 8903–8. Orlando, J.J., Tyndall, G.S. and Wallington, T.J. (2003) The atmospheric chemistry of alkoxy radicals. Chemical Review, 103: 4657–89. Ostling, K., Kelly, B., Bird, S., et al. (2001) Fast-turnaround alkyl nitrate measurements during the PROPHET 1998 summer intensive, Journal of Geophysical Research, 106: 24439–49. Prinn, R.G., Weiss, R.F., Miller, B.R., et al. (1995) Atmospheric trends and lifetime of CH3 CCL3 and global OH concentrations. Science, 269: 187–92. Ranschaert, D.L., Schneider, N.J. and Elrod, M.J. (2000) Kinetics of the C2 H5 O2 + NOx reactions: Temperature dependence of the overall rate constant and the C2 H5 ONO2 branching channel of C2 H5 O2 + NO. Journal of Physical Chemistry, 104: 5758–65.

290

Volatile Organic Compounds in the Atmosphere

Ridley, B.A. (1991) Recent measurements of oxidized nitrogen-compounds in the troposphere. Atmospheric Environment, 25: 1905–26. Ridley, B.A., Shetter, J.D., Walega, J.G., et al. (1990) The behavior of some organic nitrates at boulder and Niwot Ridge, Colorado. Journal of Geophysical Research, 95: 13949–61. Ridley, B.A., Atlas, E.L., Walega, J.G., et al. (1997) Aircraft measurements made during the spring maximum of ozone over Hawaii: peroxides, CO, O−3 , NOy , condensation nuclei, selected hydrocarbons, halocarbons, and alkyl nitrates between 0.5 and 9 km altitude, Journal of Geophysical Research, 102: 18935–61. Roberts, J.M. (1990) The atmospheric chemistry of organic nitrates. Atmospheric Environment, 24: 243–87. Roberts, J.M. and Fajer, R.W. (1989) UV absorption cross-sections of organic nitrates of potential atmospheric importance and estimation of atmospheric lifetimes. Environmental Science and Technology, 24: 945–51. Roberts, J.M., Bertman, S.B., Parrish, D.D., et al. (1998) Measurement of alkyl nitrates at Chebogue point, Nova Scotia during the 1993 North Atlantic Regional Experiment (NARE) intensive. Journal of Geophysical Research, 103: 13569–80. Rosen, R.S., Wood, E.C., Wooldridge, P.J., et al. (2004) Observations of total alkyl nitrates during Texas Air Quality Study 2000: Implications for O−3 and alkyl nitrate photochemistry. Journal of Geophysical Research, 109: D07303. −C14 alkyl nitrates in remote south Atlantic air. Schneider, M. and Ballschmiter, K. (1999) C3 − Chemosphere, 38: 233–44. Schneider, M., Luxenhofer, O., Deissler, A. and Ballschmiter, K. (1998) C-1-C-15 alkyl nitrates, benzyl nitrate, and bifunctional nitrates: Measurements in California and South Atlantic air and global comparison using C2 Cl4 and CHBr3 as marker molecules. Environmental Science and Technology, 32: 3055–62. Shepson, P.B., Anlauf, K.G., Bottenheim, J.W. (1993) Alkyl nitrates and their contribution to reactive nitrogen at a rural site in Ontario, Atmospheric Environment, 27A: 749–57. Shepson, P.B., Edney, E.O., Kleindienst, T.E., Pittman, J.H., Namie, G.R. and Cupitt, L.J. (1985) The production of organic nitrates from OH and NO3 reaction with propylene. Environmental Science and Technology, 19: 849. Shepson, P.B., Mackay, E. and Muthuramu, K. (1996) Henry’s law constants and removal processes for several atmospheric β-hydroxy alkyl nitrates. Environmental Science and Technology, 30: 3618–23. Simpson, I.J., Blake, N.J., Blake, D.R., et al. (2003) Photochemical production and evolution of selected −C5 alkyl nitrates in tropospheric air inﬂuenced by Asian outﬂow. Journal of Geophysical C2 − Research, 108. −C5 alkyl Simpson, I.J., Wang, T., Guo, H., et al. (2006) Long term atmospheric measurements of C1 − nitrates in the Pearl River delta region of southeast China. Atmospheric Environment, 40: 1619–32. Sparks, J.P., Roberts, J.M. and Monson, R.K. (2003) The uptake of gaseous organic nitrogen by leaves: A signiﬁcant global nitrogen transfer process. Geophysical Research Letters, 30 (23). Sprengnether, M., Demerjian, K.L., Donahue, N.M. and Anderson, J.G. (2002) Product analysis of the OH oxidation of isoprene and 1,3-butadiene in the presence of NO. Journal of Geophysical Research, 107: (D15), 4268. Stroud, C.A., Roberts, J.M., Williams, J., et al. (2001) Alkyl nitrate measurements during STERAO 1996 and NARE 1997: Intercomparison and survey of results. Journal of Geophysical Research, 106: 23043–53. Sumner, A.L., Couch, T.L., Shepson, P.B., et al. (2001) A study of the formaldehyde budget above a forest canopy. Journal of Geophysical Research, 106: 24387–405. Swanson, A.L., Blake, N.J., Dibb, J.E., et al. (2002) Photochemically induced production of CH3 Br, CH3 I, C2 H5 I, ethene, and propene within surface snow at Summit, Greenland. Atmospheric Environment, 36: 2671–82.

Organic Nitrates

291

Swanson, A.L., Blake, N.J., Atlas, E., Flocke, F., Blake, D.R. and Rowland, F.S. (2003) Seasonal variations −C4 nonmethane hydrocarbons and C1 − −C4 alkyl nitrates at the Summit research station in of C2 − Greenland. Journal of Geophysical Research, 108 (D2). Talbot, R.W., Dibb, J.E., Scheuer, E.M., et al. (2000) Tropospheric reactive odd nitrogen over the South Paciﬁc in austral springtime. Journal of Geophysical Research, 105: 6681. Thornberry, T., Carroll, M.A., Keeler, G.J., et al. (2001) Observations of reactive oxidized nitrogen and speciation of NOy during the PROPHET summer 1998 intensive. Journal of Geophysical Research, 106: 24359–68. Treves, K., Shragina, L. and Rudich, Y. (2000) Henry’s law constants of some beta-, gamma-, and delta-hydroxy alkyl nitrates of atmospheric interest. Environmental Science and Technology, 34: 1197–203. Treves, K. and Rudich, Y. (2003) The atmospheric fate of C-3–C-6 hydroxyalkyl nitrates. Journal of Physical Chemistry, 107: 7809–17. Tuazon, E.C. and Atkinson, R. (1990) A product study of the gas-phase reaction of isoprene with the OH radical in the presence of NOx . International Journal of Chemical Kinetics, 22: 1221–36. Vaghjiani, G.L. and Ravishankara, A.R. (1991) New measurement of the rate coefﬁcient for the reaction of OH with methane, Nature, 350: 406–9. Volz, A. and Kley, D. (1988) Evaluation of the Montsouris series of ozone measurements made in the 19th-century. Nature, 332: 240–2. Werner, G., Kastler, J., Looser, R. and Ballschmiter, K. (1999) Organic nitrates of isoprene as atmospheric trace compounds. Angewandte Chemie, 38: 1634–7. Zhang, D., Zhang, R.Y., Park, J., et al. (2002) Hydroxy peroxy nitrites and nitrates from OH initiated reactions of isoprene. Journal of the American Chemical Society, 124: 9600–5.

Volatile Organic Compounds in the Atmosphere Edited by Ralf Koppmann Copyright © 2007 by Blackwell Publishing Ltd

Chapter 8

High-Molecular-Weight Carbonyls and Carboxylic Acids Paolo Ciccioli and Michela Mannozzi

8.1

Introduction

Fifteen years ago, lists of volatile organic compounds (VOCs) requiring intensive monitoring in space and time were prepared in the United States (Purdue et al. 1992) and Europe (Kotzias and Hjorth 1991). Basically, target species were non-polar compounds characterised by high ozone-creating potentials (Derwent and Jenkin 1991) and toxic properties (Speijers 1993). The number of compounds to be monitored was deﬁned on the basis of the best available techniques for their collection and analysis (Purdue et al. 1992). It was concluded that 45 VOCs with two to nine carbon atoms in the molecule could have been quantiﬁed with high accuracy and in a short time (Purdue et al. 1992). They were mainly alkanes, alkenes, dienes and substituted arenes (Derwent and Jenkin 1991; Speijers 1993). The only polar compounds included in the list were formaldehyde, acetaldehyde and acetone (Purdue et al. 1992). They were selected because of their widespread emission and strong photochemical production (Finalyson-Pitts and Pitts 2000; UK PORG 1994). Monitoring of polar VOCs with a number of carbon atoms larger than 3, although desired (Kotzias and Hjorth 1991), was not considered of high priority. Many of these compounds were, indeed, believed to be carbonyls of photochemical origin characterised by rather short atmospheric lifetimes. They should not have substantially contributed the −C11 fraction). semi-volatile fraction of VOCs ranging between 4 and 11 carbon atoms (C4 − Their impact on ozone production could have been assessed by mathematical models describing the emission, transport and chemical reactions of parent VOCs (Derwent et al. 1996). When high-resolution gas chromatographic–mass spectrometric (GC–MS) techniques −C14 fraction (Ciccioli were allowed to identify and quantify all VOCs present in the C4 − et al. 1992; Helmig and Greenberg 1994), it was found that polar compounds in general, but particularly carbonyls with 4–11 carbon atoms in the molecule, were, by far, the dominant components present in many tropospheric sites (Ciccioli et al. 1993a, 1993b, 1994, 1996; Greenberg and Zimmerman 1984; Helmig et al. 1996; Yokouchi et al. 1990). Although their contribution was higher in rural, forest and remote areas (Bowman et al. 2003; Ciccioli et al. 1993b, 1994, 1996; Greenberg and Zimmerman 1984; Helmig et al. 1996; McClenny et al. 1998; Yokouchi et al. 1990), levels measured in urban and suburban air sheds were

High-Molecular-Weight Carbonyls and Carboxylic Acids

293

not negligible with respect to non-polar VOCs (Ciccioli et al. 1993a; Grosjean et al. 1996, 2002). For more than a decade, the ubiquitous occurrence of these polar compounds and their high tropospheric levels were a matter of controversy. Some authors believed that they mainly originated from sampling artefacts (Helmig et al. 1996; McClenny et al. 1998); others believed that unidentiﬁed sources existed for them (Bowman et al. 2003; Ciccioli et al. 1993a; Yokouchi et al. 1990). Scientists supporting the second hypothesis suggested that emission of these polar VOCs was widespread and high enough to largely compensate for their removal from the atmosphere (Bowman et al. 2003; Ciccioli et al. 1993a; Grosjean et al. 1996). Because of this, they could have substantially contributed to the build-up of ozone and photochemical oxidants in the troposphere (Ciccioli et al. 1993a; Grosjean et al. 1996; Yokouchi et al. 1990). Some indications were given on the possible sources emitting high-molecular-weight carbonyls (Bowman et al. 2003; Ciccioli et al. 1993a, 1993b; Grosjean et al. 1996; Yokouchi et al. 1990), but very few studies were available to support their existence. Although there was evidence of sampling artefacts (McClenny et al. 2001; Pires and −C11 carbonyls Carvalho 1998; Roberts et al. 1984), the question of the presence of C4 − in air was not solved. The production of these compounds in the adsorption media was, in fact, highly dependent on the sorbent material used and the ozone content in air (Ciccioli et al. 2002). Recent data obtained using artefact-free techniques, including direct analysis by proton-transfer reaction mass spectrometry (PTR-MS) (Karl et al. 2001), have shown that the atmospheric levels of high-molecular-weight carbonyls do not differ so much from those recorded in some of the early studies. If artefacts occurred in the past, their extent was not always high enough to question the ubiquitous occurrence of these polar compounds in air and the existence of strong and widespread tropospheric sources. Many of these sources have now been identiﬁed and emission rates accurately determined. The available data indicate that emission of high-molecular-weight polar VOCs can justify the observed levels. Today, enough information exists to dedicate a speciﬁc chapter of this book to the sources, atmospheric levels and reactivity of the most important polar VOCs belonging to the −C11 fraction. The focus here will be on carbonyl compounds and monocarboxylic acids. C4 − Methods for their determination in air and emission sources will also be discussed. Information provided in this chapter will be complementary to that on very volatile carbonyls and acidic compounds, whose sources, atmospheric levels and reactivity will be discussed in other parts of this book. In this chapter, the discussion will be limited to compounds with 11 carbon atoms because polar VOCs with higher molecular weight, although present in some emission sources, will never reach detectable concentrations in the troposphere.

8.2 8.2.1

Sources Man-made sources

−C11 carbonyls and Among the numerous man-made activities acting as a source for C4 − monocarboxylic acids, the most intense and widespread are: biomass burning, biofuel and charcoal combustion, vehicular exhaust emission and food cooking. A speciﬁc section will

294

Volatile Organic Compounds in the Atmosphere

be dedicated to each one of them, whereas a separate section will be devoted to man-made sources whose emission rates are small or not well deﬁned yet.

8.2.1.1

Biomass burning and biofuel and charcoal combustion

High-molecular-weight carbonyls and carboxylic acids detected in the gas-phase emission of biomass burning and biofuel and charcoal combustion are listed in Tables 8.1 and 8.2. Their emission rates are expressed in mg per kg of dry matter burned (mg/kg). Values were obtained by combining the contributions from the ﬂaming and smouldering phases of combustion (Andreae and Merlet 2001; Ciccioli et al. 2001; Schauer et al. 2001). While Table 8.1 summarises the results of laboratory studies performed in burning facilities, Table 8.2 reports composite data obtained by combining ﬁeld and laboratory observations from different research teams. These latter data are more representative of the emission occurring at ecosystem level, where various materials containing different amounts of water are combusted. A comparison between the two tables indicates that the great detail of laboratory studies, very useful for identiﬁcation and quantiﬁcation purposes, is partly lost in ﬁeld investigations. This can be explained by considering that the dynamic of ﬁres in forest or agricultural areas is not the same as that generated in a burning facility. Also the water content and composition of fuels are different, as well as the sampling protocol used for VOC collection. While in laboratory studies VOCs are sampled in a chimney under controlled ﬂow and temperature conditions, ﬁeld sampling is carried out in the open atmosphere and at variable distances from the ﬁre. The sampling distance depends upon the temperature, direction and combustion phase of the ﬁre. Because of this, transport, chemical reactions and physical transformations can take place before sampling occurs. Components can be selectively removed from the atmosphere by oxidation, adsorption and partition processes. The situation is complicated by the fact that, in a real ﬁre, compounds emitted in the ﬂaming phase can mix and interact with those released during the smouldering and ignition phases (Andreae and Merlet 2001). These effects might explain why many compounds detected in laboratory conditions are not seen in the ﬁeld. In spite of this, ﬁeld data are crucial to estimate the emission of polar VOCs from these sources, since they better reﬂect the dynamics of ﬁres and account for the complexity of burned material present in a given ecosystem. By combining the data from the two tables, we can say that biomass burning and biofuel and charcoal combustion are all strong sources of high-molecular-weight carbonyls in air. In addition to n-alkenals, n-alkanals and saturated and unsaturated linear ketones, they also emit metacrolein (MAC), methylvinylketone (MVK), methylethylketone (MEK), furaldehydes, benzaldehyde and 6-methyl-5-heptene-2-one (6-MHO). Particularly high are the emission rates of MEK, furaldehydes, benzaldehyde and 2-3-butanedione. These latter species reach emission values comparable to those of many pyrogenic alkanes, alkenes and arenes present in forest ﬁres and biofuel combustion (Andreae and Merlet 2001). The emission of organic acids was detected only by one research group in two biomassburning experiments performed in a combustion facility (Ciccioli et al. 2001). Their presence in the emission can depend on several factors, such as the type of biomass burned, the combustion conditions and the sampling and analytical procedures adopted. Whatever

Carboxylic acids Propanoic acid 2-Propenoic acid Butanoic acids Pentanoic acids Hexanoic acid Heptanoic acid Octanoic acid Nonanoic acid Decanoic acid

Ketones Methylethylketone (2-butanone, MEK) Methylvinylketone (3-buten-2-one, MVK) 2,3 Butanedione Pentenones Pentanones Hexanones Heptanones 6-Methyl-5-heptene-2-one (6-MHO) Octanones

Aldehydes 2-Methyl-2-propenal (methacrolein, MAC) 2-Butenal (Crotonaldehyde) Butanal + isobutanal 2-Oxo-butanal Pentanals Hexanals Heptanals Octanals Furaldehydes Benzaldehyde m –o Tolualdehyde 2,5 Dimethyl benzaldehyde

Type of burned material References

115

20

20

215

21

16

49 12 12 77

198 31 337 28 189 626

177 62 194 88 90 77

276 96 241 32 418 419

44

Eucalyptus Schauer et al. (2001)

9

American oak Schauer et al. (2001)

23

American pine Andreae and Merlet (2001)

2

3 3 4 6 206 69

6

90 39 14 5 4

228 94 68 52 26 14 24 39 30

147 23 15 15 22 130 32

891

74

71

1

207

85

108 48

22 7

9

Wheat straw Ciccioli et al. (2001)

5 11

12

Mediterranean pine Ciccioli et al. (2001)

16

4

40

680

30–60

220 2–3

4–9 3

40–50

Biofuel burning Andreae and Merlet (2001)

−C11 carbonyl compounds and carboxylic acids from biomass burning and biofuel combustion Table 8.1 Gas-phase emission rates (mg/kg of dry mass) of C4 −

296

Volatile Organic Compounds in the Atmosphere

−C11 carbonyl compounds from Table 8.2 Gas-phase emission rates (mg/kg of dry mass) of C4 − biomass burning and charcoal combustion Type of burned material Aldehydes Butanal + isobutanal Hexanals Heptanals Furaldehydes Benzaldehyde Ketones Methylethylketone (2-butanone, MEK) 2,3 Butanedione Pentanones Heptanones Octanones

Savannah and grassland

Tropical forests

Extratropical forests

Agricultural residues

Charcoal burning

53 2–24 3 230 29

71 31 3 370 27

210 20 4 290–630 36

21 12 1 370 9

200 40 8 720 70

260

430

17–74

440

830

1–2 6 15

92 28 2 19

350–1 500 90

900 7 2 20

1 800 90 10 40

Data from Andreae and Merlet (2001).

the reason is, data suggest that emission of high-molecular-weight carboxylic acids also takes place to some extent. The importance of biomass burning in determining the levels of high-molecular-weight carbonyls and acids in the troposphere arises from to the possibility that emitted products have to travel thousands of kilometres before deposition occurs (Ramanathan et al. 2001). Long-range transport takes place when huge amounts of gases, vapours and particles are released into the atmosphere. Since their injection into a cloud decreases the size of water droplets and increases the temperature of the cloud, turbulent motions are generated that are able to bring pollutants at high altitudes (5–7 km from the Earth’s surface) (Ramanathan et al. 2001). Compounds will remain and react inside the cloud until a critical diameter of water droplets is reached (Ramanathan et al. 2001). Depending on the meteorological conditions, several days are needed before products are removed by wet deposition. Generation of polluted clouds explains why higher-molecular-weight polar compounds can be detected in high-elevation sites (Ciccioli et al. 1993b; Karl et al. 2001).

8.2.1.2

Vehicular exhaust emission

Vehicular exhaust is another important source of high-molecular-weight carbonyls and carboxylic acids (Kawamura et al. 2000). Very often, it is the main source present in urban and suburban areas. Table 8.3 reports the gas-phase emission rates measured in the tail pipe of two gasoline-fuelled cars and a medium-duty diesel truck. Both vehicles were run in a test facility simulating the engine regime followed in a typical urban cycle (Schauer et al. 1999a, 2002a). In the same table, the emission factors of light- and heavy-duty vehicles determined in a tunnel experiment are also displayed (Grosjean et al. 2001; Kean et al. 2001). All emission data are given in mg of fuel consumed per kilometre run by the car (mg/km).

41.08 19.00 7.30 2.60 2.30 2.20 2.50 1.70 159.00 145.00 45.00 145.00

0.49

0.3 0.12 0.19 0.03 0.05 0.04

(0.12) (0.05) (0.05) (0.06) 0.01 0.03 0.12

0.47

0.08 0.08

32.00

114.00 31.00

1.76 0.06

1.27 1.34 1.34 0.24

23.00

Gasoline-powered without catalyst Tailpipe Schauer et al. (2002a)

0.30

Gasoline-powered with catalyst Tailpipe Schauer et al. (2002a)

Values in brackets are estimated by extrapolating data of Kawamura et al. (2000). Abbreviations: LDV = Light duty vehicles, HDV = Heavy duty vehicles. d.n.q. = detected but not quantiﬁed.

Ketones Methylethylketone (2-butanone, MEK) 2-Pentanone 4-Methyl-2-pentanone Carboxylic Acids Butanoic acids Pentanoic acid Hexanoic acid Heptanoic acid Octanoic acid Nonanoic acid Benzoic acid Methylbenzoic acids

Aldehydes 2-Methyl-2-propenal (methacrolein, MAC) 2-Butenal (crotonaldehyde) Butanals 2-Oxo-butanal Pentanals Hexenals Hexanals Heptanals Octanals Nonanal Decanal Undecanal Furaldehyde Benzaldehyde m + p -Tolualdehyde o -Tolualdehyde 2,5 Dimethylbenzaldehyde 2,4 Dimethylbenzaldehyde 2,4,6-Trimethylbenzaldehyde

Type of determination References

Type of vehicle

0.13 0.24 1.26 0.77

7.50

4.10

3.80

2.20 3.20 3.10 4.40 2.80 2.60

13.40 1.30

4.00

Medium-duty diesel Tailpipe Schauer et al. (1999a)

−C11 carbonyls and carboxylic acids from vehicular emission Table 8.3 Gas-phase emission rates (mg/km) of C4 −

0.75 0.97

0.04 0.01

0.86 1.57 31.60 0.81 0.23 0.23

1.07 0.22 0.08 1.05 0.3 0.68 0.03 d.n.q. d.n.q. d.n.q. d.n.q.

0.70

Tunnel Grosjean et al. (2001)

HDV

0.12

0.09 0.09 0.02 0.05 0.01 0.04

0.15 0.06 0.01 0.13 0.04 0.04 0.18 d.n.q. d.n.q. d.n.q. d.n.q.

0.10

Tunnel Grosjean et al. (2001)

LDV

298

Volatile Organic Compounds in the Atmosphere

Also in this case, emission rates determined in laboratory test facilities differ from those measured on the road. Again, strong discrepancies are observed for the heavier components. This mainly happens because the engine regimes of urban cycle simulations seldom match those of vehicles passing through a tunnel. If the average speed in the tunnel is higher than that simulated by the urban cycle, the internal combustion of vehicles passing through it is more efﬁcient, and reduced emission of particles and heavier components takes place (Di Lorenzo et al. 1991). In addition, accumulation of water vapour and soot particles inside the tunnel can also promote the selective removal of heavier components from the air by partition, condensation and adsorption processes. The removal rates will largely depend on the internal volume of the tunnel and its ventilation rate. Which one of the two methods is more representative of the real emission is still an open question. While tunnel studies provide values that better reﬂect the composition of the vehicular ﬂeet running on a given area, urban cycle simulations supply data that more accurately approach those of vehicles experiencing real trafﬁc conditions. A comparison between the relative emission strength of biomass burning and internal combustion indicates that the former source exhibits emission rates that are two to three times higher than the latter. Such a comparison was obtained by converting the data displayed in Table 8.3 into mg/L of fuel. This was only possible for the tunnel experiment, for which conversion factors of 15.4 and 3.3 were provided for light- and heavy-duty vehicles, respectively (Kean et al. 2001). By looking at the data shown in the table, it appears that the most abundant highmolecular-weight carbonyls released by vehicular exhaust are aromatic aldehydes, alkanals and some ketones with four and ﬁve carbon atoms. MAC and MEK are also present in substantial amounts, whereas MVK and furaldehydes are lacking in this source. As far as monocarboxylic acids are concerned, some of the data were obtained from the cited literature; others were estimated from the work of Kawamura et al. (2000). Calculations were made under the assumption that the ratio between carbonyl and carboxylic acids that was measured in the tail pipe of different cars was the same as that occurring in an urban cycle experiment. These data are reported in brackets in the table. Particularly interesting among the detected compounds is the presence of benzoic and methylbenzoic acids, which are not released from biomass burning. These data highlight the importance of abatement devices and of the ﬂeet composition in determining the levels of carbonyls and organic acids in the atmosphere. Higher pollution levels will be reached in countries where the ﬂeet is old and transport based on trucks equipped with diesel engines.

8.2.1.3

Food cooking

Food cooking is deﬁnitely another important and widespread tropospheric source of −C11 carbonyls and carboxylic acids in air. In one of the early papers (Ciccioli et al. C4 − 1993a) food cooking was reported to be the main source of semi-volatile carbonyls in urban and suburban areas. Gas-phase emissions of carbonyls and acids from meat charbroiling and vegetable cooking are shown in Table 8.4 (Schauer et al. 1999b, 2002b). Data are given in mg of VOC per kg of food cooked (mg/kg). As can be seen, the pattern of these sources is deﬁnitely dominated by linear carbonyls and carboxylic acids. Important is the presence of substantial amounts of unsaturated compounds (mainly n-alkenals) because, as we shall

High-Molecular-Weight Carbonyls and Carboxylic Acids

299

−C11 carbonyls and carboxylic acids measured in food cooking Table 8.4 Gas-phase emissions rates of C4 − operations. Data are expressed in mg/kg of food cooked Type of food Type of cooking Type of seed oil used References Aldehydes 2-Methyl-2-propenal (methacrolein, MAC) 2-Butenal (Crotonaldehyde) Butanal Hexanals Heptanals Octanals Nonanal Decenals Decanal Undecenals Undecanal Furaldehyde Ketones 2-Nonanone 2-Decanone 2-Undecanone Carboxylic acids Heptanoic acid Octanoic acid Nonanoic acid

Meat Charbroiling — Schauer et al. (1999b)

Vegetables Stir-frying Soybean Schauer et al. (2002b)

Vegetables Stir-frying Canola Schauer et al. (2002b)

Potatoes Deep-frying Hydrogenated oil Schauer et al. (2002b)

52.00

5.50

1.10

0.80

495.00

29.10

24.10

5.20

373.00 203.00 125.00 146.00 141.00 98.20 70.00 88.60 35.00 81.50

19.70 4.10 4.30 7.90 12.40 16.10 5.20 18.40 3.00

17.40 6.40 8.00 9.70 14.80 26.40 1.09 29.40 0.20

4.50 6.70 5.20 5.70 13.50 2.63 2.90 2.88 1.20

77.30 65.50

3.30 2.60 2.31

3.13

0.08 0.59 0.14

26.00 29.40 42.00

5.94 5.93 11.89

0.53 4.33 12.2

0.37 0.64 3.27

see later, they are highly reactive compounds. It is worth noting that emission of carbonyls is not restricted to meat cooking but also occurs when various kinds of ﬁsh are grilled or fried (Yasuhara and Shibamoto 1995). Emission from cooking is so important in determining the levels of carbonyls in air, that Feng and co-workers (Feng et al. 2005) identiﬁed emission from restaurants as one of the major sources of these compounds in the urban area of Guangzhou (Canton), China. In considering the impact of food cooking on the atmosphere, we should keep in mind that this type of emission is often associated with that of biofuel and charcoal combustion. Emission rates much higher than those reported in the table are likely to occur in poor countries where combustion is not performed under controlled conditions.

8.2.1.4

Other sources

−C11 carbonyl compounds have been reported from indoor Substantial emissions of C4 − sources. In addition to food cooking, carbonyl compounds are generated from cigarette

300

Volatile Organic Compounds in the Atmosphere

smoke (Shaughnessy et al. 2001), and evaporation of household products (Knöppel and Schauenburg 1989), plastic packages (Ezquerro et al. 2002) and ﬂooring materials (such as linoleum) (Jensen et al. 1995). Jurvelin et al. (2003) found that in Finland indoor emission is the main source of benzaldehyde, pentanal, hexanal, octanal and nonanal in air. With the exception of nonanal, strong emission of high-molecular-weight carbonyls from indoor sources was also observed in a study conducted in China (Feng et al. 2004). MEK, cyclohexanone and benzaldehyde are also emitted from chemical plants using solvents for the production of polymeric materials, dyes, varnishes and perfumes. Because of the widespread use of carbonyl compounds in plastic and household products, carbonyls are found at trace levels in the gas produced by controlled (Parker et al. 2002) and uncontrolled landﬁlls (Yassaa et al. 2001). Finally, 6-MHO, geranyl acetone (GA) and some n-alkanals can be released by food, beverages and cigarettes because these substances are in the list of permitted additives of the Food and Drug Administration (FDA) of the United States. At ppmv levels, they give the taste of fruits to beverages and candies or a pleasant smell to tobacco smoke.

8.2.2 8.2.2.1

Biogenic sources Terrestrial vegetation

Since the emission of polar VOCs from terrestrial plants has been summarised in the excellent reviews made by Puxbaum et al. (1997) and, more recently by Kesselmeier and Staudt (1999), we will concentrate here on those aspects not covered in these reviews. Although emissions of higher-molecular-weight carbonyl compounds were reported since 1992 (Isidorov 1992; Winer et al. 1992), no individual emission rates were given in these early studies. Only in 1995 (König et al. 1995), the ﬁrst report appeared in which speciated emission of higher-molecular-weight carbonyls from natural and agricultural species was given. Among the eight vegetation species investigated (grape, rape, rye, wheat, beech, hornbeam, birch and an European oak), six aldehydes (hexanal, 2-E-hexenal, heptanal, octanal and nonanal) and four ketones (butanone, 2-pentanone, 6-MHO and menthone) were identiﬁed and quantiﬁed. Butanone was found in grape and birch, 2-pentanone in wheat and 6-MHO and menthone in birch. As far as n-alkanals are concerned, hexanal was found in all plants but the other compounds were detected only in the emission of Quercus petraea, a quite common oak species in Europe. In all cases the emission rates of individual carbonyls ranged between 0.002 and 0.04 μg/g DW (dry weight)h, a value extremely small if compared to that of isoprenoids (mainly isoprene and monoterpenes) released from emitting plants (0.1–20 μg/g DW h). Grassland emission was also analysed in the same study, and three ketones (butanone, 2-pentanone and camphor) and three aldehydes (butanal, hexanal and 2-E-hexenal) were detected. Also in this case emission rates were very small and ranged between 0.001 and 0.01 μg/g DW h. High-molecular-weight carbonyls were identiﬁed by Kirstine et al. (1998) in grass and clover emissions. Among them, ﬁve aldehydes (MAC, 3-methylbutanal, pentanal, nonanal and benzaldehyde) and two ketones (2,3-butanedione and 4-methyl2-pentanone) were detected. Emission rates from grass did not differ too much from those

High-Molecular-Weight Carbonyls and Carboxylic Acids

301

recorded in the European study. Clover was found to produce lower emission than grass. When compared to very volatile compounds, the emission rates of high-molecular-weight carbonyls from clover and grass were from one to two orders of magnitude lower than those of acetaldehyde or acetone. Low emission rates (0.01 and 0.02 μg/g DW h) of higher-molecular-weight n-alkanals and 6-MHO obtained from enclosure experiments were reported by Possanzini et al. (2000) from Valencia navel late, a citrus plant species originated from Spain. They found that the emission of these carbonyl compounds was depending upon the content of ozone to which the plant was exposed. The emission of heptanal, octanal and 6-MHO increased when the content of ozone in the branch enclosure changed from about 10 to 60 ppbv. It was suggested that ozone could have stimulated the emission of some alkanals, in addition to release 6-MHO by ozonolysis of leaf lipids (Fruekilde et al. 1997). Emission of high-molecularweight alkanals by ozonolysis was also hypothesised by Bowman et al. (2003) based on the observations made at Dikson, a rural site in the United States. Compelling evidence of the ozone effect has been recently provided by the studies carried out by Wildt et al. (2003) in laboratory studies under controlled conditions. By investigating various plant species, it was found that ozone stimulated the emission of n-alkanals from hexanal to nonanal, and that the emission was light and temperature dependent. The proﬁles obtained in these experiments closely approached those recorded in a forest area of Central Italy, where high ozone levels were measured (Ciccioli et al. 1993a). However, even in the presence of ozone, emission rates of these alkanals remained rather small and never exceeded 0.04 μg/g DW h. The light dependency of the emissions suggested that these compounds were formed inside the leaf. Higher emission rates of nonanal and decanal (0.1–20 μg/g DW h) from several Mediterranean plant species were only reported in one ﬁeld study (Owen et al. 1997). At least for two of the plant species investigated (Quercus ilex and Pinus Pinea); the results did not agree with other ﬁeld and laboratory investigations, where very low emissions of semi-volatile carbonyls were detected (Ciccioli et al. 1997; Kesselmeier et al. 1997). −C11 alkanals Based on the available information, it can be concluded that emission of C4 − and some linear and branched ketones from terrestrial plants is widespread, but small and ozone dependent. So far, no pathways supporting the de novo synthesis of these components have been identiﬁed (see also Chapter 4). Biochemical pathways leading to the formation of carbonyl compounds having a terpenoid structure (such as menthone, carvone and camphor) are, instead, fairly well known. Production of these monoterpenes is strictly related to photosynthesis and mainly occurs in the leaf chloroplasts (Ciccioli et al. 1997; Fall 1999; Kesselmeier et al. 1997; Loreto et al. 1996). As other monoterpenes, these compounds are formed from geranylpyrophosphate (GPP), mainly through the so-called DOX pathway (Lichtenthaler 1999). Since speciﬁc enzymes are required for the biosynthesis of different monoterpenes, not all plants are able to produce and emit these compounds. Emission is light and temperature dependent if monoterpenes are not stored in specialised organs (such as glands or resin ducts) (Ciccioli et al. 1997; Loreto et al. 1996). When products are stored in internal or external organs isolated from the stomatal chamber, their emission is only driven by the temperature (Fall 1999). In terrestrial plants, the emission of carbonyls having a terpenoid structure is usually one to two orders of magnitude smaller than that of the most commonly emitted monoterpenes, such as α-pinene, β-pinene, limonene and sabinene. Only in two plant species

302

Volatile Organic Compounds in the Atmosphere

growing in southern California, camphor was found to be the most abundant emitted compound (Arey et al. 1995). Emission rates of camphor from Black Sage (Salvia mellifera) and California Sagebush (Artemisia californica) were so high (>10 μg/g DW h) that this compound was detected at levels of about 0.01 μg/m in suburban and high-elevation sites near Los Angeles (Reissell and Arey 2001). Also known is the mechanism responsible for the huge emissions (>100–200 μg/g DW h) of 2-E-hexenal from injured plants. This carbonyl compound is always released together with comparable amounts of 2-E-hexenol and 2-E-hexenyl acetate (Fall 1999). It occurs in any plant, regardless of its capability to produce and emit isoprenoids. When wounding takes place, a pathway stimulating the production of hormones is activated (Fall 1999) (see also Chapter 4). The biosynthesis of such hormones, whose main function is to repair damages produced in the plant tissue, leads to the emission of the three polar VOCs from the wound. Their release is responsible for the intense smell generated during grass cutting. In vegetation species equipped with storage compartments where monoterpenes are stored, emission of wounding products is concurrent with that of monoterpenes. It is an important source in case plants are subject to herbivore attack. In some cases, plants use this emission as a defence system (Kessler and Baldwin 2001). Chemicals released from the wound are used as signalling compounds to indicate the presence of the herbivore to its natural enemies. In this way, the herbivore can be killed and the plant saved from further damages. It has recently been shown that pathogen attack can also induce the emission of n-alkanals from hexanal to nonanal from plants (Wildt et al. 2003). However, it is not clear if this emission is produced by vegetation or derived from the enzymatic transformation of plant products assimilated by the insect.

8.2.2.2

Insects and animals

Many insects use high-molecular-weight VOCs for communication purposes (Tillman et al. 1999). Signalling compounds (also called semiochemicals) are released to stimulate reproduction, defence and aggregation. A detailed database of semiochemicals, mainly pheromones, has been recently prepared by El-Sayed (2005). From this list, it appears that 2-methyl-butanal, 2-methyl pentanal, 6-MHO (also called sulcatone by the entomologists), GA and other higher alkanals and ketones are quite common pheromones. They are released by some insects belonging to the families of the orders of heteroptera, hymenoptera and coleoptera. In particular, 6-MHO was found in the emission of wheat infested by aphids (Quiroz et al. 1997) or by wheat stem sawﬂy (Cephus cinctus) (Peck 2004). In the aphid Alloxysta vitrix, the release of 6-MHO is used to attract females and repel males, thus allowing an even distribution of females in the ﬁeld (Micha et al. 1993). 6-MHO is also emitted by the male of an engraver insect infecting living poplars (male Platypus mutatus Chapuis) (Audino Gonzalez et al. 2005). Behavioural assays show that females are more attracted than males to galleries with boring males inside, and that their antennas positively respond to 6-MHO. According to a recent study, 6-MHO is also produced and released in substantial amounts by large animals, such as cows (Birkett et al. 2004). It exploits the function to repel ﬂies capable of transmitting severe diseases to these animals.

High-Molecular-Weight Carbonyls and Carboxylic Acids

8.2.3 8.2.3.1

303

Photochemical production Gas-phase oxidation mechanisms

As already mentioned in the introduction, oxidation of polar and non-polar VOCs is a source of carbonyls in the troposphere. The most common starting process is the reaction of VOCs with OH radicals (Finalyson-Pitts and Pitts 2000). Organic radicals can be formed either by H abstraction (R• ) or OH addition (HOR• ) (Finalyson-Pitts and Pitts 2000). Addition of oxygen to the primary radical leads to organic hydroperoxy- (HORO•2 ) and peroxy-radicals (RO•2 ). They can react with NO and other peroxy radicals (HO2 , HORO2 , RO2 ). When sufﬁcient amounts of NO are present in air, the preferred process is the oxidation of NO to NO2 . It gives rise to organic oxy radicals (HORO, RO• ) as products (Finalyson-Pitts and Pitts 2000). Association with NO to produce organic nitrates can also take place, but only in NO-rich environments (Finalyson-Pitts and Pitts 2000). The formation of organic oxy radicals is the fundamental step leading to carbonyl compounds. −C bond of HORO and RO, or by reaction They can be directly generated by scission of a C− of these radicals with oxygen (Finalyson-Pitts and Pitts 2000). Various carbonyls can be formed in this step because isomerisation is possible. In all cases, formation of carbonyl compounds is concurrent with the release of HO2 radicals from the reaction mixture. Reaction with OH radicals is faster in compounds in which addition of OH radicals is dominant over H abstraction (Finalyson-Pitts and Pitts 2000). This explains why the highest reaction rates are observed in oleﬁns, particularly those having more than one double bond in the molecule (Finalyson-Pitts and Pitts 2000). In nature, most of them are isoprenoid compounds synthesised by terrestrial plants (Guenther 1999) or insects (Tillman et al. 1999). On a global scale, isoprenoids released from terrestrial plants account for about 65% of total reactive emission (Guenther 1999). Oleﬁn compounds can also react with ozone to produce carbonyls and, in some instances, carboxylic acids (Finalyson-Pitts and Pitts 2000). In this reaction, formation of a primary ozonide by addition to the double bond is followed by a decomposition process, giving rise to two carbonyl compounds and two energy rich biradicals (Criegee’s intermediates) (Finalyson-Pitts and Pitts 2000). These latter species, can lead to various products after isomerisation or loss of excited oxygen and OH radicals (Finalyson-Pitts and Pitts 2000). As we shall see later, the two carbonyl compounds formed in the ozonolysis of oleﬁns can further react with OH radicals and ozone or decompose by ultraviolet (UV) photolysis. Carbonyl compounds are also generated by reaction of VOCs with NO3 radicals (Peck 2004). Although these radicals are much more reactive than OH radicals, their production is small and occurs only at night.

8.2.3.2

Carbonyls from gas-phase oxidation reactions

Detailed investigations (Ciccioli et al. 1999) have shown that in areas experiencing high emission from vehicular sources and fossil fuel combustion, the high molecular fraction of VOC contains about 76 alkanes, 54 alkenes and 11 dienes. All of them can react with OH radicals and/or ozone to form carbonyl compounds as primary products. However, in each one of these classes the high-molecular-weight fraction accounts for a rather small part of the total mass (5–10%), and concentrations usually follow an exponential decay with the number of carbon atoms in the molecule (Andreae and Merlet 2001; Sagebiel et al. 1996).

304

Volatile Organic Compounds in the Atmosphere

Higher homologues are thus present at trace levels in emission sources and in heavily polluted air sheds. The amount of high-molecular-weight carbonyls that can be produced by oxidation of VOCs present in these sources is thus small. The example of 1-octene serves well to illustrate this point. Oxidation of this compound is known to produce hexanal by ozonolysis (Grosjean and Grosjean 1996). Although the reaction is fast, the amount of hexanal formed is limited because the precursor is present in the atmosphere at trace levels (tenths of pptv), and part of it reacts with OH radicals to give different carbonyl products. Based on these considerations, the proﬁle of high-molecular-weight carbonyls should approximately follow that of their precursors. The bulk of products will be concentrated in the C4 –C5 range. Due to the complexity of precursors and products, it is not possible to say how many of them will be linear or branched. Carbonyl compounds higher than C5 can be generated by oxidation of n-alkanes, but their production rates are small, due to the low concentrations and limited reactivity of these VOCs in air. The chemistry of arenes released most by vehicular exhaust emission, fossil fuel combustion and biomass burning (Rainer and Obermeier 1999) is also rather complex. In the atmosphere about 40 different components belonging to this class have been identiﬁed in very polluted air sheds (Ciccioli et al. 1999). The arene fraction is comprised of benzene and its upper homologues having methyl- and/or alkyl groups attached to the benzene ring. The highest molecular weight compounds of this class are substituted arenes having alkyl groups with ﬁve carbon atoms in the benzene ring (Ciccioli et al. 1999). However, the most abundant members in the emission are benzene, toluene, xylenes and ethylbenzene (Rainer and Obermeier 1999). Basically, arene oxidation occurs by reaction with OH radicals (Finalyson-Pitts and Pitts 2000). Ozonolysis is limited to substituted components having unsaturated alkyl groups in the benzene ring (such as styrene) (Finalyson-Pitts and Pitts 2000). The reactivity towards the reaction with OH radicals is known to rapidly increase with the molecular weight (Derwent and Jenkin 1991; Derwent et al. 1996). Benzene is therefore the least reactive arene in air. Although reaction constants are known (Derwent and Jenkin 1991; Derwent et al. 1996), the oxidation mechanism of arenes is not well understood yet and is still a matter of investigation (Finalyson-Pitts and Pitts 2000). Many products are formed in the OH-initiated reaction, and only some of them are carbonyls. They are generated at small yields or are so reactive that they rapidly undergo further degradation. Few of them can be detected in the atmosphere. The complexity of the arene chemistry is well illustrated by the case of toluene. Its reaction with OH radicals gives rise to a large number of different compounds (FinalysonPitts and Pitts 2000). According to recent views (Bloss et al. 2005), these products can be assigned to four main pathways. One originated from H-abstraction and three from the OH-radical addition. While the ﬁrst pathway accounts for the formation of small amounts of benzaldehyde (7%), the other three explain the formation of secondary aromatic compounds and smaller oxygenated compounds. Production of metylbenzoquinone, o-cresol and its derivatives (nitrocresol, hydroxycresols) occurs through the so-called phenolic pathway. Formation of α-dicarbonyl products and assumed co-products (α, β-unsaturatedγ -dicarbonyls and furanones) is originated from the so-called peroxide-bicyclic pathway. This latter route accounts for 65% of products formed. Basically the only high-molecularweight carbonyls that can somehow be related to arene oxidation are aromatic aldehydes and furanones.

High-Molecular-Weight Carbonyls and Carboxylic Acids

305

In addition to precursors of anthropogenic origin, high-molecular-weight carbonyls are also formed by oxidation of biogenic compounds, mainly isoprenoids. Among them, isoprene is by far the most important non-methane VOCs emitted at a global scale (Guenther 1999). By reaction with OH-radicals and ozone, isoprene is converted into MVK and MAC (Atkinson and Arey 2003). Consistent yields have been reported for both reactions. While in the OH-initiated reaction yields of about 0.32 and 0.22 were obtained for MVK and MAC, respectively, reaction yields of about 0.16 and 0.4 were reported for the ozonolysis. Since MVK and MAC are also present in biomass burning and vehicular emission, they can be considered speciﬁc tracers of isoprene oxidation only in forest and rural areas where anthropogenic pollution is small. Gas-phase oxidation of monoterpenes, which are the second most abundant biogenic components emitted at a global scale (Guenther 1999), gives rise to a large variety of highmolecular-weight carbonyls and carboxylic acids (Atkinson and Arey 2003). The type of products formed depends upon the structure of the monoterpene compound considered (Atkinson and Arey 2003). Figure 8.1 shows the formation of acetone and 4-methyl-3cycloxene-1-one from the reaction of OH radicals and ozone with terpinolene (Reissell et al. 1999). In this reaction, addition to the double bond located in the six-ring system

O3 + O O

O

∗ •

•

+ [(CH3)2COO]∗ CH3C(O)CH3 + •

O

Other products

OO•

Other products

NO •

OH + •

OH

O• OH

NO2

Decomposition •

+ CH3C(OH)CH3 O2 O CH C(O)CH + HO 3 3 2 Figure 8.1 Reaction schemes of the reaction of ozone and OH radicals with terpinolene.

306

Volatile Organic Compounds in the Atmosphere

can also take place. It produces a dicarbonyl compound (terpinolenonaldehyde) at a lower yield. Oxidation pathways analogous to terpinolene are followed by ocimene and myrcene (Reissell et al. 1999), emitted in large amounts by two plant species very common in the Mediterranean area (Pinus pinea and Quercus ilex) (Ciccioli et al. 1997; Kesselmeier et al. 1997; Loreto et al. 1996; Staudt et al. 1997). The reaction schemes of bicyclic monoterpenes, such as α-pinene and β-pinene, are deﬁnitely more complex than the one shown in Figure 8.1, because H atom abstraction and OH radical addition are both possible in the OH-initiated reaction (Aschmann et al. 1998). While the former process accounts for the formation of acetone, the second pathway (which is the preferred one) leads to pinonaldehyde (in the case of α-pinene) and nopinone (in the case β-pinene), as major products. In the presence of sufﬁcient levels of NO, formation of hydroxy and dihydroxycarbonyls and nitrated products has been reported (Aschmann et al. 1998). Pinonaldehyde and nopinone are also the main oxidation products formed in the ozonolysis reaction of these monoterpenes (Atkinson and Arey 2003). They are produced together with and a series of aldehydes and mono- and dicarboxylic acids. Figure 8.2 summarises the ﬁrst generation carbonyl compounds formed in the oxidation reactions of the most common monoterpenes emitted by terrestrial plants. Also shown is the range of yields that were measured in different smog chamber experiments (Atkinson and Arey 2003). Since in many instances primary carbonyl products have structural features that can be related to their precursors, oxidation processes in the atmosphere can be followed by looking at the product-precursor relationship. A study of this type was performed in Finland within the frame of the European Project OSOA (Boy et al. 2004).

8.2.3.3

Carboxylic acids from gas-phase oxidation reactions

As already mentioned, small amounts of carboxylic acids can be formed in the ozonolysis of some monoterpenes. Acids have been detected in O3 -initiated reaction of α-pinene, β-pinene, sabinene and 3-carene (Atkinson and Arey 2003). They are generated by stabilisation and reaction of the energy-rich Criegree intermediates. Since primary acid products contain two polar groups, they exhibit a rather low vapour pressure and will never contribute in a signiﬁcant way to the VOC fraction in air. In the OSOA study performed in Finland, more than 90% of pinonic acid was detected in the aerosol phase (Boy et al. 2004). The only high-molecular-weight carboxylic acids having sufﬁcient volatility to contribute −C11 fraction of VOCs are alkanoic acids and benzoic acid (Kean et al. 2001). In to the C4 − a study conducted in the Los Angeles area, it was found that photochemical production of n-alkanoic acids was comparable to direct emission (Kawamura et al. 2000). Although most of the photochemical production gave rise to low-molecular-weight compounds, a minor −C11 fraction. The photochemical origin of butyric portion also contributed to the C4 − to decanoic acids was assessed through a correlation analysis performed on air samples. These monocarboxylic acids were believed to originate from the ozonolysis of large oleﬁns emitted by vehicular exhaust emission and biogenic sources, through as reaction sequence similar to the one leading to the formation of pinonic or norpinonic acids from α- and β-pinene. Biogenically emitted oleﬁns were indicated as precursors of nonanoic acid. The photochemical origin of high-molecular-weight alkanoic acids from energy-rich Crieege intermediates, although reasonable, has not been conﬁrmed by laboratory experiments and

High-Molecular-Weight Carbonyls and Carboxylic Acids

307

Yield OH O CHO

-Pinene

Yield O3

OH

O

O3

0.25–0.79 0.16–0.4

0.28–0.87 0.1–0.5 -Pinene

Pinonaldehyde

Nopinone O

O CHO

3-Carene

0.17

0.46–0.5

<0.02

0.2–0.36

0.31–0.34 <0.08–0.44

Coronaldehyde

Sabinene

Sabinaketone O

O CHO

0.29

0.01–0.04

Camphene

Camphenilone

Limononaldehyde O

Limonene

0.17–0.2

O Limonaketone

-Phellandrene

0.19

Myrcene

0.29

0.01

0.7

<0.02

4-isopropyl-2-cyclohexen-1-one

0.24–0.26

O 4-Vinyl-4-pentenal

0.29

0.4

O 4-Methyl-3-cyclohexene-1-one

O CHO

0.33

0.1

Terpinolene O cis-Ocimene

4-Methyl-3, 5-hexadienal

Terpinolenonaldehyde

Figure 8.2 Main carbonyl products from the reaction of selected monoterpenes with ozone and OH radicals and range of yields measured in different smog chamber experiments. Data are from Atkinson and Arey (2003).

308

Volatile Organic Compounds in the Atmosphere

O R

O O NO2 NO2 NO

•OH

O R

H

R

O •

O2

O

O O O•

R

R• + CO2

O•

R NO2

R = alkyl group, for example, C6H13 R'H, R"CHO

HO2• RO •

R" = H, alkyl group

2

O R

+ OH

R'• • R"CO

Figure 8.3 Reaction schemes for the formation of acids and PANs from carbonyl compounds. For the sake of clarity, the pathway leading to nitrate organic products has been omitted. For this pathway, see the text.

should be taken with some caution until precursors are positively identiﬁed and reaction yields measured. As illustrated in Figure 8.3, formation of carboxylic acids can also take place by reaction of OH radicals with aldehydes (Finalyson-Pitts and Pitts 2000). H-abstraction from a saturated carbonyl produces a peroxyacyl radical able to generate carboxylic acids through different pathways. As indicated in Figure 8.3, production of a carboxylic acid is competitive with the equilibrium reaction forming peroxyacyl nitrates (PANs). It is evident from the scheme that acid production will take place when decomposition of PANs largely dominates over formation. This occurs when high temperatures and NOx -limited conditions are established in the atmosphere (Finalyson-Pitts and Pitts 2000). These conditions are usually observed in rural or forest areas during the summer season (Finalyson-Pitts and Pitts 2000). The −C6 alkanoic acids pathway shown in Figure 8.3 can explain the small production of C4 − observed in polluted air masses moving from the metropolitan area of Tokyo to the internal areas of Japan, covered mostly by forests (Satsumabayashi et al. 1995). Whichever pathway is followed, gas-phase production of high-molecular-weight organic acids is always much smaller than that of corresponding carbonyls and, although ubiquitous, gives rise only to n-alkanoic acids as volatile products. It is deﬁnitely a negligible source when compared to the anthropogenic ones.

8.2.3.4

Carbonyls form ozone-surface reactions

Although formation of alkanals from ozonolysis of lipids present at the sea surface was already hypothesised in the early 1990’s (Yokouchi et al. 1990), the ﬁrst compelling evidence

High-Molecular-Weight Carbonyls and Carboxylic Acids

309

of carbonyls production from ozone-surface reactions was provided Fruekilde et al. (1997). They showed that the ozonolysis of waxes covering the surface of plant leaves gives rise to the production of 6-MHO, GA, 4-oxopentanal (4-OPA), some alkanals and acetone. The common precursor of these products is squalene, an alkene compound having 30 carbon atoms and six double bonds. It is present in plant and skin lipids of men and animals. The matter of ozone-surface reactions has been recently reviewed by Rudich (2003). In this publication, emphasis is given to the ozonolysis of the hydrophobic organic matter accumulated at the air-water interface. Data show that carbonyls are formed together with hydrophilic compounds. Experiments carried out by Wadia et al. (2000) indicate that the ozonolysis of unsaturated phospholipids in water produces nonanal at yields slightly lower than 50%. Reaction rates are higher than those usually observed in the gas phase, although the reaction mechanism is basically the same (Wadia et al. 2000). The only difference is that Criegee’s intermediates can better transfer the excess of energy to the liquid matrix, and hydrophobic forces increase the probability of product recombination. In these reactions, high-molecular-weight carbonyls are formed only from compounds with internal double bonds. This is the reason why ozonolysis of linoleic and oleic acids at the water surface produces high-molecular-weight alkanals and alkenals (Rudich 2003). Compounds with double bonds in terminal positions can only produce formaldehyde. These laboratory experiments provide an alternative pathway to the formation of n-alkanals at the sea-surface microlayer that until recently has been exclusively attributed to UV photolysis (Zhou and Mopper 1997). At the moment, it is impossible to say which one of the two processes is the most responsible for the formation of high-molecular-weight carbonyls. It is even impossible to know if these two processes are actually part of the same mechanism. Although the bulk of hydrophilic compounds produced by these reactions will remain in the liquid phase, carbonyl products are sufﬁciently volatile to be partitioned with the atmosphere. Thus, oceans and large water reservoirs can act as sources of highermolecular-weight carbonyls in the atmosphere. In particular, marine aerosol can generate −C11 carbonyls because of the larger surface area exposed to the atmohigh amounts of C4 − sphere and solar radiation. Oxidation reactions of this type could explain the high levels of n-alkanals detected at Mauna Loa, a remote site in the Paciﬁc Ocean (Helmig et al. 1996).

8.3 8.3.1

Atmospheric levels Main carbonyls in air

Tables 8.5a–g provide a general overview of the composition and tropospheric concentrations of higher-molecular-weight carbonyls. All data are expressed in μg/m3 . When a conversion was made from ppbv, ppbvC or molecules/cm, it was assumed that sampled volumes were all normalised to standard conditions of pressure and temperature. With the exception of two Greek sites, all data were shown to a minimum value of 0.01 μg/m3 . When values lower than 0.01 μg/m3 were reported, they are indicated in the tables as <0.01 or as 0.01 μg/m3 . For the accuracy and uncertainties of determinations, the reader is referred to the original literature. It is important to stress that missing values in the tables do not necessarily mean that components were not present in air.

−C11 carbonyl compounds in urban, suburban, rural, forest and remote areas and methods used for their Table 8.5 Gas-phase concentrations (μg/m3 ) of C4 − sampling and detection (a) Site Country Type of site Sampling method Detection and separation method Time of the year References 2-Methyl-2-propenal (methacrolein, MAC) 2-Butenal (Crotonaldehyde) Butanal Pentanal 4-Oxo-pentanal (4-OPA) Hexanal Heptanal Octanal Nonanal Decanal Undecanal Benzaldehyde m–o Tolualdehyde p-Tolualdehyde Pinonaldehyde

Algiers Algeria Urban 3 and 5 a and c

Ouargla Algeria Urban 3 and 5 a and c

Los Angeles United States Urban 3 c and d

New Jersey United States Urban 3 c

Rio de Janeiro Brazil Urban 3 c and d

Porto Alegre Brazil Urban 3 c and d

Tsukuba Japan Urban 2 b

Winter Cecinato et al. (2002)

Winter Cecinato et al. (2002)

Summer Grosjean et al. (1996)

Summer Zhang et al. (1994)

Summer Grosjean et al. (2002)

Summer Grosjean et al. (1999)

Summer Yokouchi et al. (1990)

0.06 1.09 1.14

0.33 0.15

0.94 2.28 1.57

1.47 2.81

0.30 0.48 0.27

0.97 0.56 0.22 0.56 0.29

0.48 0.38 0.28 0.48 0.48

2.44

0.33

1.88 3.26 3.03 6.91 3.69 2.20 1.04 1.29

0.16 1.22 1.23 0.98 0.15 0.74 0.87 0.73 0.21 0.54 2.51 2.41

2.45

1.08

0.27 0.08 0.09 2.09 1.47 0.11 1.08 0.40 0.05

0.38 0.41 1.16 0.96 0.22

Methylethylketone (2-butanone, MEK) Methylvinylketone (3-buten-2-one, MVK) Pentanone Cyclohexanone 2-Hexanone 2-Heptanone 3-Heptanone 6-Methyl-5-hepten-2-one (6-MHO) Geranyl acetone (GA) Nopinone Camphor (b) Site Country Type of site Sampling method Detection and separation method Time of the year Reference 2-Methyl-2-propenal (methacrolein, MAC) 2-Butenal (Crotonaldehyde) Butanal Pentanal 4-Oxo-pentanal (4-OPA)

0.33

0.18

3.86

1.42

0.88

2.09

Hong Kong China Urban 3 c

Guangzhou China Urban 3 c

Rome Italy Urban 1 a

Athens Greece Urban 3 c

Darmstadt Germany Urban 5 a

Leipzig Germany Urban 3 c

Leipzig Germany Urban 3 c

Annual averages Ho et al. (2002)

Summer Feng et al. (2005)

Summer Ciccioli et al. (1993a) and∗

Summer Bakeas et al. (2003)

Summer Schlomski et al. (1997)

Winter Muller (1997)

Summer Muller (1997)

0.70 0.37

0.50 0.24

0.33

0.18 0.22

0.29 1.24 0.88

0.52 0.83 1.12

1.8 7.4

6.81 5.73

Table 8.5 (Continued) (b) Continued Hexanal Heptanal Octanal Nonanal Decanal Undecanal Benzaldehyde m–o Tolualdehyde p-Tolualdehyde Pinonaldehyde Methylethylketone (2-butanone, MEK) Methylvinylketone (3-buten-2-one, MVK) Pentanone Cyclohexanone 2-Hexanone 2-Heptanone 3-Heptanone 6-Methyl-5-hepten-2-one (6-MHO) Geranyl acetone (GA) Nopinone Camphor (c) Site Country Type of site Sampling method Detection and separation method Time of the year

0.50

0.81 0.38 0.41 1.96 0.46

1.56 2.09 2.76 2.32 0.75

3.2

0.90 0.21

1.29 1.18

0.75

1.6 3.9 2.3

0.65

5.23

1.05

4.66 9.51 9.87 0.38

0.32

0.57

0.63

0.68

0.49

Montelibretti Italy Suburban Rural 1+3 a and b Summer

0.73

1.98

2.04

Helsinki Finland Urban

Helsinki Finland Urban

Algiers Area Algeria Suburban

Algiers Area Algeria Suburban

Asuza United States Suburban

3 d

3 d

3 and 5 a and c

3 and 5 a and c

1 a

Montelibretti Italy Suburban Rural 1 a

Winter

Summer

Winter

Summer

Summer

Summer

References

Hellén et al. (2004)

Hellén et al. (2004)

2-Methyl-2-propenal (methacrolein, MAC) 2-Butenal (Crotonaldehyde) Butanal Pentanal 4-Oxo-pentanal (4-OPA) Hexanal Heptanal Octanal Nonanal Decanal Undecanal Benzaldehyde m–o Tolualdehyde p-Tolualdehyde Pinonaldehyde Methylethylketone (2-butanone, MEK) Methylvinylketone (3-buten-2-one, MVK) Pentanone Cyclohexanone 2-Hexanone 2-Heptanone 3-Heptanone 6-Methyl-5-hepten-2-one (6-MHO) Geranyl acetone (GA) Nopinone Camphor

0.02

0.02

0.07 0.07

0.06 0.05

0.03 0.02 0.05 0.12 0.13 0.01 0.07

0.06 0.05 0.30 0.20 0.30 0.02 0.03

0.05

Cecinato et al. (2002)

Cecinato et al. (2002)

Reissell and Arey (2001)

Ciccioli et al. (1993a)

1.14 0.13 0.49 0.37

0.10 0.39 0.30

1.09 0.26 0.29 0.93 0.53

0.99 0.76 0.31 0.70 0.19

0.94

1.21

0.46

0.69

0.50 0.20 1.11 0.87 1.11 2.38 3.12 1.38

2.03

0.02

0.02

0.02

0.02

Possanzini et al. (2000)

1.78

0.01

0.40 0.30 0.80 1.00 2.40 0.20

Table 8.5 (Continued) (d) Site Country Type of site Sampling method Detection and separation method Time of the year References 2-Methyl-2-propenal (methacrolein, MAC) 2-Butenal (Crotonaldehyde) Butanal Pentanal 4-Oxo-pentanal (4-OPA) Hexanal Heptanal Octanal Nonanal Decanal Undecanal Benzaldehyde m–o Tolualdehyde p-Tolualdehyde Pinonaldehyde Methylethylketone (2-butanone, MEK) Methylvinylketone (3-buten-2-one, MVK)

Montelibretti Italy Suburban Rural 2 and 5 a and c

Montelibretti Italy Suburban Rural 2 and 5 a and c

Likovrisi Greece Suburban Rural 3 c

Hendersonville United States Rural 2 c

Dickson United States Forest Rural 2 c

Summer

Winter

Summer

Summer

Summer

Cecinato et al. (2001)

Cecinato et al. (2001)

Bakeas et al. (2003)

McClenny et al. (1998)

Bowman et al. (2003)

0.43

1.00 0.45

0.17

1.6

0.70 0.71 1.46 1.05 0.93 0.30

0.14 0.06 0.19 0.66 0.79 0.29

Rondonia Brazil Forest 1 c Spring Autumn Kesselmeier et al. (2002)

Spring *

2.38

0.19 0.12

0.16

2.4

0.28 0.12 0.27 0.46

0.42 0.30 0.40 1.26

1.1

0.93

0.37 0.26

0.43 0.28

0.23 0.24 0.26 0.21 0.10

0.25 0.32 0.43 0.55 0.64

0.28

0.31

2.5 0.64 0.78

Okinawa Japan Forest Coastal 1 a

5.81

Pentanone Cyclohexanone 2-Hexanone 2-Heptanone 3-Heptanone 6-Methyl-5-hepten-2-one (6-MHO) Geranyl acetone (GA) Nopinone Camphor (e) Site Country Type of site Sampling method Detection and separation method Time of the year References 2-Methyl-2-propenal (methacrolein, MAC) 2-Butenal (Crotonaldehyde) Butanal Pentanal 4-Oxo-pentanal (4-OPA) Hexanal Heptanal Octanal Nonanal Decanal Undecanal Benzaldehyde

Uryo Japan Forest

Uryo Japan Forest 4 b

Burriana Spain Rural Coastal 1+3 a and c

Castelporziano Italy Rural Coastal 1 a

4 b Summer Matsunaga et al. (2004)

Summer Matsunaga et al. (2003, 2004)

Summer Possanzini et al. (2000)

Summer *

5.01

0.20

Melpitz Germany Rural

Melpitz Germany Rural

Melpitz Germany Rural

3 c

3 c

1 a

Winter Muller (1997)

Summer Muller (1997)

Summer OSOA-Final report and Hoffmann (1999) 0.07

0.22 0.51 0.15

0.10 0.04

0.68

0.11 0.12

0.74 0.99 2.80 2.78 3.72 1.92

0.08 0.16 0.02 0.28 0.43 0.78 1.75 2.82 1.63 0.39

0.23 0.16

0.26 0.16

0.03 0.59 0.17 0.80 0.17 0.16

0.20

0.20

Table 8.5 (Continued) (e) Continued m–o Tolualdehyde p-Tolualdehyde Pinonaldehyde Methylethylketone (2-butanone, MEK) Methylvinylketone (3-buten-2-one, MVK) Pentanone Cyclohexanone 2-Hexanone 2-Heptanone 3-Heptanone 6-Methyl-5-hepten-2-one (6-MHO) Geranyl acetone (GA) Nopinone Camphor (f) Site Country Type of site Sampling method Detection and separation method Time of the year References

0.45

0.49 0.08

1.47

0.22

0.47

0.52 0.32

Pabstthum Germany Rural 5 a

Birkenes Norway Forest 3 c

Birkenes Norway Forest 1 a

Hyytiälä Finland Forest 3 c

Hyytiälä Finland Forest 1 a

Hyytiälä Finland Forest 3 d

Alert Canada Remote 1 a

Summer Moortgat et al. (2002)

Spring + Summer CATOME-Final Report and Dye (1999)

Spring + Summer CATOME-Final Report and Dye (1999)

Summer Hellén et al. (2004)

Summer *

Spring Hellén et al. (2004)

Winter *

2-Methyl-2-propenal (methacrolein, MAC) 2-Butenal (Crotonaldehyde)

0.08

0.04

Butanal Pentanal 4-Oxo-pentanal (4-OPA)

0.25 0.27

0.06 0.15

0.27 0.22

0.16 0.14

0.06 0.05

Hexanal Heptanal Octanal Nonanal Decanal Undecanal Benzaldehyde m–o Tolualdehyde p-Tolualdehyde Pinonaldehyde Methylethylketone (2-butanone, MEK) Methylvinylketone (3-buten-2-one, MVK) Pentanone Cyclohexanone 2-Hexanone 2-Heptanone 3-Heptanone 6-Methyl-5-hepten-2-one (6-MHO) Geranyl acetone (GA) Nopinone Camphor

0.67 0.35 0.23 0.55 0.30 0.18

0.52

0.13

0.06 0.03 0.06 0.08 0.01 0.02 0.02

0.17 0.09 0.10 0.12 0.12

0.16

0.13

0.08 0.24

0.05 0.04 0.08 0.08 0.12

0.05 0.08

0.06

0.02

0.02 0.01 0.05 0.03 0.02

0.21

0.17

0.09

0.01

0.02

0.02 0.02 0.02 0.12

0.04

0.01

0.01

<0.01

(g) Site

Mauna Loa

Mace Head

Sahara desert

K2 Pyramid

K2 Pyramid

Country Type of site

United States Remote

Ireland Coastal

Niger Remote

Sampling method Detection and separation method Time of the year References

2 a

2 a

1 a

Nepal Remote High Altitude 1 a

Nepal Remote High Altitude 1 a

Annual averages Helmig et al. (1996)

Autumn Sartin et al. (2001)

Spring *

Autumn Ciccioli et al. (1993b)

Autumn Ciccioli et al. (1993b)

Sonnblick Observatory Austria High Altitude 6 e Fall-Winter Karl et al. (2001)

Table 8.5 (Continued) (g) Continued 2-Methyl-2-propenal (methacrolein, MAC) 2-Butenal (Crotonaldehyde) Butanal Pentanal 4-Oxo-pentanal (4-OPA) Hexanal Heptanal Octanal Nonanal Decanal Undecanal Benzaldehyde m–o Tolualdehyde p-Tolualdehyde Pinonaldehyde Methylethylketone (2-butanone, MEK) Methylvinylketone (3-buten-2-one, MVK) Pentanone Cyclohexanone 2-Hexanone 2-Heptanone 3-Heptanone 6-Methyl-5-hepten-2-one (6-MHO) Geranyl acetone (GA) Nopinone Camphor

0.13 0.13

0.03 0.03 >0.01 >0.01

0.10 0.24

2.3 1.99 2.96

0.08

7.36 4.22 5.13 6.64 5.29

1.84

1.65

0.12

2.07

0.17

0.65

0.20

2.66 1.90 1.70 2.36

0.08 <0.01 0.68

0.21

0.26

* Original data from the authors. Sampling method: (1) Traps coated with graphitic carbon adsorbents; (2) Traps coated with Tenax adsorbents; (3) Traps ﬁlled with C18 supports coated with 2,4-dinitrophenylhydrazine DNPH; (4) Annular denuder coated with o -benzylhydroxyl ammonium (BHA); (5) Traps ﬁlled with C18 supports coated with pentaﬂuorophenylhydrazione (PFPH) and (6) Direct inlet. Separation and detection method: (a) GC–MS; (b) GC-FID; (c) HPLC-UV; (d) HPLC–MS and (e) PTR-MS.

High-Molecular-Weight Carbonyls and Carboxylic Acids

319

Together with the aerometric concentrations, the tables report the methods that were used for the sampling, separation and detection of high-molecular-weight carbonyls. To highlight seasonal variations, special attention was paid to ﬁnd sites in which samples were collected at different times of the year. Efforts were also made to ﬁnd sites where data were collected by different groups, but at the same time of the year. Most of the values in the tables refer to daily or weekly averaged values, only in few cases annual means or single measurements are reported. Results show that n-alkanals from butanal to undecanal, when detected, were the domin−C11 fraction. Nonanal, and sometimes decanal, ant carbonyl compounds present in the C4 − were often the major compounds. Benzaldehyde was also detected at levels comparable to those of nonanal in many sites. Tolualdehydes were detected only in three urban areas and in one suburban site. At least in one of them (Athens), their occurrence might arise from sampling in a busy road, too close to trafﬁc emissions. Crotonaldehyde was mainly detected in urban areas and at concentrations lower than nonanal. Only in the rural site of Melpitz, its levels exceeded those of other carbonyl compounds. MAC and MVK were detected only in a limited number of sites. They were abundant in rural forest areas in which substantial emission of isoprene occurred (Rondonia, Hendersonville, Azusa). In European forest sites, where isoprene emission from terrestrial vegetation is lower than in the United States, and in tropical regions, concentrations of MAC and MVK, when detected, were much smaller than those of nonanal. Cyclohexanone, 2-hexanone, 2-heptanone and 3-heptanone were found only in few sites, and at least in two of them (Los Angeles and the K2-Pyramid), their presence in air was directly associated with anthropogenic sources. Pentanone was only found by one research group in Melpitz and Leipzig. Only few data are available on the levels of 6-MHO in air, and most are from the same group. When detected, this compound often exceeded the levels of nonanal. Only in Hyytiälä, Uryo and Castelporziano was it found to be less. 6-MHO was detected and quantiﬁed together with GA at Mace Head. At this site, GA was about 30% of 6-MHO. 4-OPA, one of the oxidation products of 6-MHO (Matsunaga et al. 2004), was found at one forest site of the Hokkaido Island and in Castelporziano. Only in Japan, the levels of 4-OPA were comparable or higher than those of nonanal. Pinonaldehyde, nopinone and camphor were detected only at few sites, and generally at levels equal or lower than 0.01 μg/m3 . Only in the Los Angeles area, one group reported larger concentrations of camphor. The data given in the tables conﬁrm that highmolecular-weight carbonyls, in general, and particularly n-alkanals are deﬁnitely important constituents of the atmosphere. At some sites (Kesselmeier et al. 2002; Possanzini et al. 2000), they account for 24 (Hellén et al. 2004) to 36% (Possanzini et al. 2000) of the whole mass of carbonyls in air.

8.3.1.1

Levels and trends at urban sites

The concentrations recorded in urban areas reﬂect the large differences in the source patterns and emission strengths existing between the investigated sites. While in some urban areas (such as Los Angeles, Rome, Athens, Rio de Janeiro and Porto Alegre) vehicular exhaust −C11 carbonyls emission followed by food cooking are by far the dominant sources of C4 − in air, in other sites (such as Guangzhou) food cooking, biomass and coal combustion, and

320

Volatile Organic Compounds in the Atmosphere

indoor pollution might have an impact comparable or even higher than that of vehicular exhaust emission. The population density and meteorological conditions are also important in determining the observed levels. The fact that most of the data shown in the tables were collected in the summer season, when the photochemistry plays a decisive role in the production and proliferation of carbonyls, does not help to rationalise the observations. We can only say that the high concentrations measured in Los Angeles, Rome and Athens are consistent with the high emission rates and severe smog episodes occurring in these urban areas (Ciccioli et al. 1993a, 1999; Grosjean et al. 1996). The summer levels measured in Helsinki ﬁt also well with the limited emission and photochemical production occurring in this city. The only information providing some indications on the impact of photochemistry on the levels of high-molecular-weight carbonyls comes from the data collected in Helsinki, where winter and summer values have been reported. The data show that summer levels −C11 n-alkanals were two to four times higher than those occurring in wintertime. of C6 − The increase of these compounds was mirrored by a small decay in butanal and pentanal and by the disappearance of MEK in air. No detectable seasonal variations were observed in MAC and MVK. The trend of MEK clearly reﬂects the decrease in emission from manmade sources, particularly vehicular exhaust emission and wood and biofuel combustion. The constant mixing ratios of MAC and MVK, suggests that reduced emission from manmade sources in summer is fully compensated for by photochemical oxidation of isoprene released from vegetation surrounding the city. The small decrease in butanal and pentanal indicates that, in this case, vegetation emission and photochemical production were not strong enough to compensate for the reduced supply from man-made sources. It is worth noting that, in this site, the contribution of the high molecular fraction to the total content −C11 fraction accounted for of carbonyls showed a rather small seasonality. While the C4 − about 26% of the total carbonyl content in the winter season, in the summer period it was a little bit less than 24% (Hellén et al. 2004).

8.3.1.2

Levels and trends at suburban, rural and forest sites

Data collected in suburban, rural and forest sites basically conﬁrm the early observations −C11 carbonyls in these areas. At some suburban indicating the presence of high levels of C4 − sites, such as Montelibretti, transport is the process most responsible for the levels observed. The effect is more evident in the summer period when the activation of a sea-breeze circulation pushes the precursors and products of photochemical pollution accumulated in urban area of Rome in the Tiber valley (Ciccioli et al. 1993a, 1999). During this time of the year, the arrival of the urban plume in Montelibretti is clearly seen by the sudden increase in the concentrations of ozone, PANs, nitric acid, carbonyls and ﬁne particles (Ciccioli et al. 1993a, 1999). Maximum values of all these pollutants are reached between 1 and 3 pm, depending on the intensity of the sea-breeze (Ciccioli et al. 1999). However, high concentrations of −C11 carbonyls last until the late afternoon, when the levels of ozone and photochemC4 − ical oxidants have already decreased (Ciccioli et al. 1999). It is possible that the arrival of ozone stimulates local emission of carbonyls from vegetation or that carbonyls come from marine air masses in which surface–ozone reactions occurred (Ciccioli et al. 1999). The combined effect of transport and photochemical pollution is so high that summer levels of high-molecular-weight carbonyls recorded in Montelibretti are, three to seven times higher

High-Molecular-Weight Carbonyls and Carboxylic Acids

321

than those measured in wintertime. However, this strong seasonality is not common to all suburban sites. For instance, in the suburban site of Algiers, wintertime values of carbonyls are generally higher than those measured in summer. Only heptanal follows an opposite trend. In Melpitz, no substantial differences are observed between summer and winter values. This suggests that, in these sites, no substantial seasonal changes occur in man-made emissions or, if they do, they are compensated for by photochemical production. −C11 carbonyls in rural and forest areas The main processes driving the levels of C4 − are vegetation emission, photochemical production and transport. The ﬁrst two processes are important sources of carbonyls in summer, when the highest values of solar radiation intensities and temperatures are reached. In addition to efﬁciently oxidising isoprene and monoterpenes, photochemical oxidants can also stimulate production of n-alkanals through different mechanisms. Because of this, high levels of carbonyls can be reached in rural-forest areas during summer. Daily trends recorded in Dickson (Bowman et al. 2003) and Hendersonville (McClenny et al. 1998) indicate that summer vegetation emission of n-alkanals is sufﬁciently high to largely compensate for their removal by gas-phase reactions and UV photolysis. The strong dependence of vegetation emission from solar radiation intensity, temperature and the oxidising capacity of the atmosphere are demonstrated by the latitudinal variations −C11 carbonyls occurring in rural-forest sites. Summer concentrations increase from of C4 − ecosystems located at high latitudes of the northern hemisphere (Uryo, Birkenes, Hyytiälä) to those located at low latitudes (Dickson, Hendersonville, Burriana and Castelporziano). Alkanals, MAC, MVK and in some cases 6-MHO, GA and 4-OPA together with primary −C11 products of monoterpene oxidation are the dominant components found in the C4 − fraction. MEK is present only in few of them. MAC and MVK are, by far, the most abundant high-molecular-weight carbonyls found in the tropical site of Rondonia, where isoprene emission was extremely high. It should be noted, however, that data collected at this site refer to a time of the year in which no biomass burning occurred and ozone levels were near to the background levels (20 ppbv) (Kesselmeier et al. 2002). This explains why levels of n-alkanals were lower than those recorded in temperate forests. In wintertime, when both vegetation emission and the oxidising capacity of the atmosphere are minimised, concentrations of high-molecular-weight carbonyls are determined mostly by anthropogenic sources. Carbonyls can be locally emitted or transported from urban or industrial areas. Using radiocarbon data, Larsen et al. (2001) found that, in Castelporziano, 40–70% of the winter carbonyl fraction was coming from fossil fuel combustion. It is conceivable to think that the wintertime dominance of anthropogenic carbonyls is more pronounced at higher latitudes, where temperatures are low and exposure to photosynthetically active radiation (PAR) are restricted to a few hours. At these sites, more than 95% of high-molecular-weight carbonyls present in wintertime are likely to be of anthropogenic origin. Due to the low population density, most of these carbonyls come from long-range transport. At the moment, the only homogeneous data providing some information on the seasonality of high-molecular-weight carbonyls in rural forest sites of Europe are those displayed in Table 8.6. It summarises monthly averaged values of MAC, MEK, hexanal and benzaldehyde that were measured in ﬁve environmental monitoring European program (EMEP) stations of Northern and Central Europe (Solberg 2005). Although limited to few compounds, these data provide useful indications on the seasonal behaviour occurring at different latitudes.

0.08 0.07 0.15 0.15 — 0.10 0.13 0.11 0.12 0.12 0.04 0.05

0.28 0.51 0.27 0.23 0.08 0.16 0.35 0.15 0.33 0.21 0.16 0.25

January February March April May June July August September October November December

January February March April May June July August September October November December

0.40 0.71 0.55 0.58 0.39 0.42 0.84 0.94 0.70 0.52 0.56 0.54

0.07 0.13 0.06 0.09 0.13 0.18 0.17 0.27 0.16 0.10 0.08 0.09

Košetice (Czech Republic)

Donon (France) 0.05 — 0.11 0.22 0.16 0.15 0.12 0.21 0.26 0.04 0.02 0.02

0.16 — 0.37 0.85 0.18 0.54 0.35 0.51 0.45 0.39 0.29 0.31

La Tardiere (France) Hexanal 0.07 0.03 0.05 0.09 0.06 0.07 0.07 0.15 0.07 0.03 0.02 0.02 MEK 2.24 0.41 0.62 0.73 0.42 0.51 0.98 0.89 0.83 0.65 0.40 0.89

Data are all expressed in (μg/m3 ). They were taken from Solberg (2005).

Utö (Finland)

Station

2.21 0.12 — 0.46 0.51 0.41 1.21 0.53 0.32 0.33 0.28 0.43

0.22 0.57 0.12 0.25 0.22 0.07 0.17 0.27 0.20

0.10 0.06

Peyrusse Vieille (France)

— — — — 0.05 — — — — — — —

— 0.07 0.05 — — — 0.09 — — — 0.04 —

Utö (Finland)

0.01 — — — — 0.04 — 0.07 0.03 — — —

0.04 0.08 0.05 — 0.05 0.02 — — 0.03 — 0.05 0.05

Košetice (Czech Republic)

MAC 0.02 0.02 0.01 0.01 0.02 0.26 0.47 0.78 0.11 0.05 0.02 0.02

Benzaldehyde 0.09 0.10 0.09 0.06 0.07 0.13 0.06 0.11 0.07 0.06 0.09 0.05

La Tardiere (France)

Table 8.6 Seasonal trends of MEK, MAC, hexanal and benzaldehyde recorded in some EMEP stations of Central Europe in the year 2003

0.02 — 0.05 0.09 0.18 0.97 0.90 0.89 0.26 0.04 0.03 0.01

0.03 — 0.06 0.06 0.04 0.06 0.05 0.06 0.21 0.04 0.02 0.03

Donon (France)

0.01 0.01 — 0.02 0.09 0.79 0.50 1.45 0.33 0.06 0.04 0.02

0.03 0.02 — 0.06 0.04 0.05 0.05 0.08 0.03 0.03 0.03 0.02

Peyrusse Vieille (France)

High-Molecular-Weight Carbonyls and Carboxylic Acids

323

The highest increase in the summer concentrations of MAC and, to a lesser extent, hexanal and benzaldehyde, is observed in the southern continental sites of Europe, where the emission and photochemical production is stronger. At the Finnish station, located in a small island of the Baltic Sea, the seasonality of hexanal is small and concentrations of benzaldehyde and MAC are often below the detection limits. The fact that at all sites MEK follow a different trend than the other carbonyls is consistent with its anthropogenic origin.

8.3.1.3

Levels and trends at remote sites

Data collected in the Sahara desert of Niger indicate that only few carbonyls are present at remote sites in which biogenic and anthropogenic emission is negligible, transport limited and UV photolysis highly efﬁcient. In such desert areas, molecular weight carbonyls higher than heptanal are deﬁnitely below the detection limits and the others are present at tiny levels. Also low are the concentrations found in wintertime in Alert (Canada), but compounds up to nonanal can still be detected at this remote site. They are on average two to three times smaller than those measured in Hyytiälä during the spring season. Since no sources exist for these compounds in wintertime and no photochemistry takes place, their presence in the atmosphere can only be attributed to long-range transport. Long-range transport is also responsible for the substantial levels of carbonyls detected in high elevation sites (>3 000 m asl). Data collected at the K2 Pyramid in Nepal give an idea of the dramatic impact that polluted plumes (or clouds) have in affecting the levels of high-molecular-weight carbonyls. Aircraft measurements have shown (Lelieveld et al. 2001) that such plumes can easily be formed over the Indian peninsula during the monsoon season. They can move toward the Himalaya when prevalent winds blow from south-east. Similar effects have been observed in the Alps when stagnant conditions promote the formation of polluted plumes over large portions of continental Europe (Karl et al. 2001). High-molecular-weight carbonyls accumulated in these plumes can be transported to high altitudes when strong slope winds are activated. This happens when the inversion layer, below which pollutants are accumulated, becomes higher than the top of the mountains (Karl et al. 2001). During one of these episodes, substantial amounts of alkanals were found in the Sonnblick observatory, Austria. The ascent of high molecular carbonyls to high elevation sites of Europe has been recently demonstrated by the presence of about 0.23 μg/m of heptanal in the air masses moving over mount Schmücke, Germany (Muller et al. 2005).

8.3.2

Main carboxylic acids in air and their atmospheric levels

The database of higher-molecular-weight alkanoic acids is extremely limited, and the few existing measurements deal with urban areas. Table 8.7 shows typical concentrations of alkanoic acids measured in southern California and central Germany. In general, levels are one order of magnitude smaller than corresponding alkanals. These ratios reﬂect those found in some emission sources. If photochemical production occurs, it appears to be proportional to that leading to high-molecular-weight alkanals. From the data of Satsumabayashi et al. (1995), it seems that such proportionality also holds in aged air masses where NOx -limited conditions are established. However, the detail of data is not

324

Volatile Organic Compounds in the Atmosphere

−C10 alkanoic acids Table 8.7 Gas-phase concentrations (μg/m) of C4 − detected in urban and suburban sites Site

References i-C4 C4 C5 C6 C7 C8 C9 C10 Benzoic acid

Downtown Los Angeles (United States) Kawamura et al. (2000)

West Los Angeles (United States)

0.213 0.109 0.070 0.147 0.068 0.091 0.048 0.028 0.051

0.027 0.071 0.034 0.046 0.022 0.048 0.018 0.007 0.024

Kawamura et al. (2000)

Downtown, Darmstadt (Germany) Schlomski et al. (1997) 0.121 0.429 0.079 0.077 0.075 0.634

sufﬁcient to draw some deﬁnite conclusions. Higher alkanoic acid/carbonyl ratios can only be observed in sites impacted by aged plumes or clouds in which pollution is generated by biomass burning and food cooking. This is the case of the K2 Pyramid where alkanoic acids were present at levels approaching those of alkanals.

8.4 8.4.1

Reactivity and impact on the atmosphere Reactions and products

The impact of high-molecular-weight carbonyls and alkanoic acids on the quality of the atmosphere is measured by their capability to produce ozone, photochemical oxidants and particles. As far as gas-phase reactions are concerned, carbonyls and acids react with OH radicals following the same basic rules described in Section 2.3.1. Compounds with double bonds (such as 6-MHO, MAC and MVK) have also the possibility to react with ozone. Oxidation of carbonyls is particularly efﬁcient in the ozone formation as it gives rise to products able to efﬁciently convert NO into NO2 . PANs, alkyl nitrates and acids are formed under certain oxidising conditions. An overview of the reactivity of carbonyls belonging to −C11 fraction is presented in Table 8.8, where the lifetimes for the reaction with OH the C4 − radicals and ozone are reported. For the sake of comparison, lifetimes of some common VOCs are also displayed in the same table. In addition, to react with ozone and radical species, carbonyl compounds can also absorb sufﬁcient energy in the near UV to photodissociate. Various reactions are possible (Finalyson-Pitts and Pitts 2000). The most important ones are the so-called Norrish type I and II photo dissociation of aldehydes (Finalyson-Pitts and Pitts 2000; Tadic et al. 2002). In the Norrish type I photodissociation scheme, two radicals are formed. One is the formyl radical (HCO• ); the other is an alkyl radical (R• ). Both of them can produce peroxy

Table 8.8 Lifetimes of higher-molecular-weight carbonyls and some common VOCs for the reaction of OH radicals and ozone. For some oxidation products of monoterpenes are also indicated in brackets the precursors τOH

Carbonyl compound

[OH] = 2 × 106 molecule/cm Time Unit

High-molecular-weight carbonyl compounds Butanal 5.7 MEK 4.8 MAC 4.8 MVK 7 Pentanal 5.3 Hexanal 5.1 6-MHO 55 Heptanal 5 Octanal 4.8 Nonanal 4.8 Camphor 2.5 Pinonaldehyde (α -Pinene) 3.3 10 Nopinone (α - and β -Pinene) Limona aldehyde (Limonene) Limona ketone (Limonene) Caronaldehyde (-3-Carene) Sabinaketone (Sabinene) Camphenilone (Camphene) 4-Methyl-3-cyclohexen-1-one (terpinolene) (3Z)-4-Methylhexa-3,5-dienal ((3Z)-4-ocimene) (3E)-4-Methylhexa-3,5-dienal ((3E)-4-ocimene) 4-Vinyl-4-pentenal (Myrcene) Benzaldehyde Other VOCs Isoprene Butane Nonane α -Pinene Benzene

References

3

References

h h h day day h

Atkinson et al. (2005) Atkinson et al. (2005) Atkinson et al. (2005) Atkinson et al. (2005) D’Anna et al. (2001) D’Anna et al. (2001) Smith et al. (1996) Plagens (2001) Plagens (2001) Plagens (2001) Reissell et al. (2001) Alvarado et al. (1998) Calogirou et al. (1999) and Atkinson and Aschmann (1993) Calogirou et al. (1999) Atkinson and Aschmann (1993) Alvarado et al. (1998) Alvarado et al. (1998) Atkinson and Aschmann (1993) Baker et al. (2004)

52

min

Baker et al. (2004)

9.6

h

Baker et al. (2004)

33

min

Baker et al. (2004)

7.4

h

Baker et al. (2004)

54 19.6

min h

Baker et al. (2004) Hellén et al. (2004)

1.1

day

Baker et al. (2004)

1.4 2.5 13.9 2.6 4.8

h day h h day

Atkinson et al. (2005) Finalyson-Pitts and Pitts (2000) Finalyson-Pitts and Pitts (2000) Atkinson et al. (2005) Finalyson-Pitts and Pitts (2000)

1.3

day

Atkinson et al. (2005)

4.4

h

Atkinson et al. (2005)

1.3 1.1 2.9 2.3 2.3 1.3

h day h h h h min h h h day h h

τO

[O3 ] = 7 × 1011 molecule/cm Time Unit

Atkinson et al. (2005) 14 3.1

day day

Atkinson et al. (2005) Atkinson et al. (2005)

h

Smith et al. (1996)

>0.65 >2.3 >8

year year year

2 2.6 >2.3 >0.9 >0.9 5.7

day h year year year h

Reissell et al. (2001) Alvarado et al. (1998) Calogirou et al. (1999) and Atkinson and Aschmann (1993) Calogirou et al. (1999) Atkinson and Aschmann (1993) Alvarado et al. (1998) Alvarado et al. (1998) Atkinson and Aschmann (1993) Baker et al. (2004)

1

326

Volatile Organic Compounds in the Atmosphere

Table 8.9 Lifetimes due to the photolysis of higher-molecular-weight carbonyls estimated in Hyytiälä (Finland) and Valencia (Spain) and their comparison with those of the reaction OH radicals Lifetime Site

τphot Hyytiälä (Hellén et al. (2004))

τphot Valencia (Moortgat (2001))

Season Solar radiation information Carbonyl compound

Spring 61.30◦ N, Zenith angle 60◦ Time

Unit

Summer 40◦ N, Zenith angle 16.9◦ Time

Formaldehyde Acetaldehyde MAC Propanal Butanal 2-Methylbutanal Pentanal Hexanal Nonanal Pinonaldehyde MEK MVK Nopinone Limonaketone

8 8.1 12 2.4 1.6

h day day day day

1.5 1.3 1.6

day day day

2.9–29

day

>8 >8

day day

τOH Assuming [OH] = 2 × 106 molecule/cm

Unit

Time

Unit

5 4 >6 1.2 1.2 7.3 17 17 1 1

h day day day day h h h day day

>6 >4 >4

day day day

16 9.2 4 7 5.7 6 5.3 5.1 4.8 3.3 4.8 7 10 1.1

h h h h h h h h h h day h h h

radicals (HO2 and RCO2 ) by reaction with O2 , able to convert NO to NO2 . As mentioned earlier, alkylperoxy radicals can further react (or decompose), giving rise to HO2 radicals and smaller molecular weight carbonyls, as products. In the Norrish II type of reaction two stable products usually an oleﬁn and an alkanal are formed (Finalyson-Pitts and Pitts −C7 n-alkanals, ethenol and an oleﬁn were the com2000). However, in the photolysis of C5 − pounds formed (Tadic et al. 2002), both of which can be quite efﬁcient in the production of ozone. The effective quantum yields of high-molecular-weight carbonyls are much smaller than −C9 those of formaldehyde, acetaldehyde and acetone (Moortgat 2001). In the case of C4 − n-alkanals, values are in the range of 0.25–0.3. For MAC and MVK, they are even smaller (0.004). Only in the case of α-branched alkanals values ranging between 0.2 and 0.7 were reported (Moortgat 2001). These data suggest that photodissociation of high-molecularweight carbonyls is not as efﬁcient as the reaction with OH radicals in the ozone production, even at noontime of summer days and at low latitudes. This conclusion is summarised in Table 8.9, in which the lifetimes due to photolysis calculated at two sites are compared with those of the reaction with OH radicals. Results indicate that photodissociation of highmolecular-weight carbonyls can take some part in the ozone production only for a limited number of species and under high solar radiation intensities.

High-Molecular-Weight Carbonyls and Carboxylic Acids

8.4.2

327

Ozone and photochemical oxidant production

A reasonable assessment of the contribution of higher-molecular-weight carbonyl to ozone and total alkyl nitrates (TAN) has been made in two recent publications (Cleary et al. 2005; Hellén et al. 2004). In both cases a relevant contribution was found. By scaling the OH-reactivity of different VOCs with respect to formaldehyde, it was estimated that about 52% of the OH radicals present in Hyytiälä were removed by carbonyl compounds (Hellén et al. 2004). About half of this portion was assigned to high-molecularweight carbonyls, although their mass was less than 24% of the total carbonyl fraction. −C11 carbonyls accounted for about 26% of the ozone proCalculations showed that C4 − duced by all VOCs. The compound contributing most to ozone formation was 6-MHO, present at trace levels in the VOC mixture (<1% in mass). A substantial contribution to the production of ozone and TAN was also found in a study conducted in the urban plume of Sacramento, in California (Cleary et al. 2005). TAN were investigated because, in polluted plumes of the United States, they account for the largest portion of NOy (Day et al. 2003). In the Sacramento study, production rates of both ozone and TAN were estimated for about 56 different components of biogenic, anthropogenic and photochemical origin. The mixture was comprised of VOCs with a number of carbon atoms ranging from 1 to 9. It included alkanes, alkenes, alkynes, aromatics, isoprene and its oxidation products, monoterpenes, n-aldehydes and other oxygenated VOCs. Also in this case, reaction with OH radicals was considered the process contributing most to ozone and TAN formation. The contribution of photolysis was calculated only for formaldehyde and acetone. The results obtained in this study are summarised in Figure 8.4. They show that high-molecular-weight carbonyls from C4 to C9 accounted for about 21% of the total ozone production and for about 29% of the total TAN production, although they were present in the VOC mixture at quite low levels (6%). In this study, the higher-molecular-weight carbonyls with the highest contribution (>30%) to ozone and TAN formation were MAC followed by nonanal and MVK. The contribution of higher-molecular-weight carbonyls to PANs has been accurately measured only for products originated from isobutanal (PiBN, peroxyisobutyric nitric anhydride), and MAC (MPAN, peroxymethacrylic nitric anhydride) (Roberts et al. 1998). It was found to be less than 8% (on a molar basis) of the total fraction in polluted plumes in Texas and Tennessee. It is also possible that other compounds, particularly nonanal and decanal, contribute to the total burden of PANs (Bowman et al. 2003), but no data exist on the atmospheric levels of these peroxyacylnitrates. It should be noted that PANs formed from high-molecular-weight compounds, in which addition of OH radicals prevails over H-abstraction, usually contain one or more hydroxyl groups in the molecule, and they will be easily dissolved in the water ﬁlm covering ﬁne aerosols. The same considerations hold for nitrates originated from ﬁrst generation products of unsaturated, cyclic and aromatic carbonyls.

8.4.3

Secondary organic aerosol formation

It is now recognised that about 78% of the mass production of secondary organic aerosols (SOA) in the troposphere comes from monoterpene oxidation products, and that SOA is

328

Volatile Organic Compounds in the Atmosphere

100

Percent

80 HMW Carbonyls

60

LMW Carbonyls Other VOC

40

20

0 Composition

O3 production

AN production

Figure 8.4 Contribution of lower- (LMW) and higher-molecular-weight carbonyls (HMW) to the VOC composition and to the ozone and alkyl nitrate (AN) production in the urban plume of Sacramento, California. Data are from Cleary et al. (2005).

an important fraction of all particulate matter (Chung and Seinfeld 2002). However, the contribution of higher-molecular-weight carbonyls to SOA is still controversial. According to Bonn et al. (2004), biogenic hydroperoxides are the compounds with the highest contribution (63%) to both SOA formation and SOA mass production. PANs and nitrates contribute for about 11–12%. Carbonyls formed from monoterpene oxidation should not contribute at all to SOA formation and SOA mass production because their vapour pressure is too high and water solubility too low to efﬁciently transfer them into the water ﬁlm covering the aerosol phase. Based on these considerations, these carbonyls should be almost exclusively found in the gas-phase. High concentrations should be, thus, measured in forest areas, where the emission of parent monoterpenes and the oxidation capacity of the atmosphere are high. These conclusions are in contrast to the existing observations indicating low levels of these compounds in forest areas. Even the most volatile product (nopinone) is detected at a few sites and always at small concentrations. For this reason, some authors (Jang et al. 2002; Tolocka et al. 2004) believe that carbonyls formed from monoterpene oxidation are heavily involved in both SOA formation and SOA mass production. These conclusions are supported by laboratory experiments indicating strong oxidation of carbonyls in the water ﬁlm covering the aerosol surface (Jang et al. 2002; Tolocka et al. 2004). Efﬁcient transfer in the liquid phase occurs because carbonyls are converted into highly watersoluble oligomers (Jang et al. 2002). Oxidation preferentially occurs in acid solutions and is catalysed by some metal ions (like iron) very abundant in atmospheric particles (Jang et al. 2002). Precursors of water soluble oligomers are hydrophilic compounds formed by the keto–enolic equilibrium of carbonyls in acid solutions (Jang et al. 2002). Substantial formation of numerous oligomeric compounds has been reported by Tolocka et al. (2004) in the ozonolysis reaction of α-pinene carried out in the presence of seed particles. To identify such large and highly polar molecules, electron-spray Fourier-transform ion-cyclotronmass-spectrometry (ESI-FTCIP) was used. Oligomeric compounds attributed to isoprene

High-Molecular-Weight Carbonyls and Carboxylic Acids

329

oxidation have recently been found in the water extracts of particles collected in the Amazon region (Claeys et al. 2004). While these oxidation mechanisms explain well how carbonyls contribute to SOA mass production, it is not clear how they can promote SOA formation. It is not yet proven that changes in the hygroscopic properties of SOA caused by the accumulation of oligomers in the water ﬁlm make the aerosols to act as cloud condensation nuclei. Probably, the oxidation mechanism of carbonyls is more complex than what is presently believed and involves the formation of additional compounds affecting the hygroscopic features of SOA. This conclusion is supported by recent experiments indicating the active involvement of OH radicals in SOA formation. Lower yields of SOA were observed in the ozonolysis of monoterpenes when OH radicals were removed from the reaction chamber (Iinuma et al. 2005). How these radicals affect the oxidation processes of carbonyls in the liquid phase and the hygroscopic feature of SOA is still a matter of investigation. High-molecular-weight alkanoic acids have been found in particles emitted from many anthropogenic sources, especially food cooking (Schauer et al. 1999b, 2002b) and biomass burning (Schauer et al. 2001). Low contents were detected, instead, in the extracts of particulate matter released from vehicular emission (Kawamura et al. 2000; Schauer et al. 1999a, 2002a). No speciﬁc studies have been made to quantify the contribution of alkanoic acids to SOA. They should contribute to SOA mass production because they are much more soluble than carbonyl compounds in water solutions. However, their contribution should be much smaller than that of acidic compounds containing more than one polar group in the molecule (such as dicarboxylic acids or the ones containing a carbonyl and a carboxylic group in the molecule), as they have much lower vapour pressure and lower water solubility. They are ubiquitous constituents of atmospheric particulate matter but they are present at very low concentrations.

8.5 8.5.1

Sampling and analysis Carbonyl compounds

Basically, only limited advances in the collection and analysis of higher-molecular-weight carbonyl compounds have occurred in the past 10 years. The most widely used methods remain those based on the preconcentration of carbonyls on traps ﬁlled with solid sorbents, followed by their analysis by capillary GC or HPLC. Liquid extraction or thermal desorption can be used to recover VOCs from the adsorption media. Thermal desorption is usually performed on uncoated cartridges ﬁlled with graphitic carbon adsorbents (Ciccioli et al. 1992, 1993a, 1993b, 1994, 2002) and porous polymers of the Tenax family (Bowman et al. 2003; Ciccioli et al. 2002; Helmig et al. 1996; McClenny et al. 1998). It is obtained by increasing the temperature of the trap up to 200–250◦ C. In this step, the direction of the ﬂow rate of the carrier gas is opposite to that used during sampling. Desorbed compounds are cryofocussed on a capillary tube kept at very low temperature (typically −150◦ C), before they are injected into the analytical column. This step is necessary to avoid band broadening of the sample, which drastically reduces the efﬁciency and selectivity of capillary columns.

330

Volatile Organic Compounds in the Atmosphere

The use of ozone scrubbers is particularly important with Tenax adsorbents because ozonolysis of the sorbent matrix produces large amounts of benzaldehyde, phenol and acetophenone (Ciccioli et al. 2002; Helmig 1997 and reference therein). Formation of n-alkanals by ozonolysis of some Tenax adsorbents has also been reported (Roberts et al. 1984). As far as carbon materials are concerned, the removal of ozone is mandatory only with certain types of traps, because partial decomposition of n-alkanals has been observed (McClenny et al. 2001). Traps exhibiting these features are those in which a graphitic carbon adsorbent (Carbopack B) is placed in series with a carbon molecular sieve (usually Carbosieve S III). Although recommended by the U.S. EPA for certain applications, this combination was already found not suitable for the collection of polar VOCs in air, especially in humid environments (Ciccioli et al. 1992, 2002). Better results can be obtained with traps ﬁlled with two or three graphitic carbon adsorbents with surface areas ranging from 15 to 250 m2 /g (Ciccioli et al. 2002). VOCs retained on this type of traps show a better resistance to ozone and atmospheric contaminants present in atmospheric water. The quality of the sample can be preserved at levels higher than 90 ppbv (Ciccioli et al. 2002). Above these values, the use of scrubbers is recommended, especially in very humid environments. Since ozone scrubbers can affect the quality of the sample, special care should be paid in their selection (Helmig 1997). At the moment, the simplest ozone scrubber that better preserves the quality of the sample is the one obtained by impregnating a glass-coated ﬁlter with a dilute solution of sodium thiosulfate (Helmig 1997). Although less practical, the addition of NO to the air stream entering the trap is also efﬁcient, and preserves well the quality of the sample (Helmig 1997). With adsorption traps ﬁlled with porous polymers and graphitic carbon materials, volumes ranging from 2 to 5 L are sufﬁcient for the analysis of high-molecular-weight compounds because the whole sample can be injected into the capillary column. Positive identiﬁcation of eluted compounds is achieved by knowing the retention times of VOCs on the capillary column used and by selecting mass spectrometry (MS) as detection system (Ciccioli et al. 2002). The mass spectrometer can be run in the scan mode or in selected ion detection. Relative retention indices and selective ions for the identiﬁcation and quantiﬁcation of higher-molecular-weight carbonyls and acids by GC–MS have been recently published by Ciccioli et al. (2002). The database contains information for the identiﬁcation and quantiﬁcation of more than 600 VOCs analysed on a DB-1 capillary column. Elution of all VOCs from C4 to C14 is achieved by increasing the column temperature from 30◦ C to 250◦ C. Providing that the same column is used, the database can help in the identiﬁcation of VOCs when other GC detectors such as a ﬂame ionisation detector (FID) are used. Methods based on 2,4-dinitrophenylhydrazine (DNPH)-coated cartridges are also widely used for the selective collection of high-molecular-weight carbonyls (Cecinato et al. 2002; Bakeas et al. 2003; Grosjean et al. 1996, 1999, 2002; Hellén et al. 2004; Ho et al. 2002; Muller 1997; Zhang et al. 1994). With respect to those using uncoated material, they offer the advantage that all carbonyl compounds, including the very volatile ones, can be detected at once. This method does not provide, however, any information on the other VOCs present in the sample, because selective retention is achieved by converting carbonyls into hydrazone derivatives. With these cartridges, the use of an ozone scrubber is mandatory −C11 alkanals, occur to a large because sampling artefacts, simulating the presence of C4 − extent even at low ozone levels (Pires and Carvalho 1998). The most widely used ozone scrubbers are those made by glass or copper tubes internally coated with potassium iodide

High-Molecular-Weight Carbonyls and Carboxylic Acids

331

(Pires and Carvalho 1998). In some instances, annular denuders coated with the same material also have been successfully used for ozone removal (Cecinato et al. 2002). With conventional DNPH-coated cartridges, sample recovery is accomplished by liquid extraction, using acetonitrile or methanol as eluants. Since only small aliquots of the liquid extract can be injected in the HPLC column, high sampling volumes are needed to meet the sensitivity requirements of the detection system. Depending on the concentrations of carbonyls in air, sampled volumes range from 0.15 to 1.5 m3 (Cecinato et al. 2002; Bakeas et al. 2003; Grosjean et al. 1996, 1999, 2002; Hellén et al. 2004; Ho et al. 2002; Muller 1997; Zhang et al. 1994). High sampling ﬂow rates must be used when diurnal cycles of carbonyls need to be followed. Gradient elution is necessary to quantify the high molecular fraction of carbonyls by HPLC. Since the best selectivity is obtained on columns working in reversed-phase liquid chromatography, the analysis starts with organic eluants (usually acetonitrile or methanol) containing high percent of water (usually 64% v/v). The water content is then gradually decreased until 100% of the organic solvent is passed through the column (Grosjean et al. 1996, 2002; Pires and Carvalho 1998). For a complete elution of carbonyls containing 11 carbon atoms in the molecule, the ﬂow of the pure organic solvent must be maintained for some time. With conventional HPLC columns, with an internal diameter of 4.6 mm and a length of 250 mm, ﬂow rates of the eluant typically range between 1 and 1.5 ml/min. UV-visible absorption is the cheaper and commonly used method for the detection of hydrazones. It can be performed at various wavelengths (360, 385 and 430 nm), but the most intense absorption band occurs at around 360 nm. Pure hydrazone standards are needed for identiﬁcation and quantiﬁcation purposes. After the pioneering work made by Grosjean and co-workers (Grosjean et al. 1996, 1999), MS has become more and more used for the positive identiﬁcation of DNPH-derivatives eluted from HPLC columns. Selective identiﬁcation can be achieved by atmospheric pressure mass spectrometry working in negative ion recording (Grosjean et al. 1999; Kolloker et al. 1998). Particularly useful for identiﬁcation purposes is the use of MS–MS techniques, allowing the selective fractionation of parent and secondary ions into fragments providing structural information on the eluted compound (Kolloker et al. 1998). With the advent of instruments equipped with ion-trap sources connected to quadrupole analysers, this method has become affordable to many research laboratories. An instrument of this type has been used for the analysis of samples collected in Helsinki and Hyytiälä (Hellén et al. 2004). Very recently, attempts have been made to develop DNPH-coated cartridges suitable for thermal desorption. This task has been successfully accomplished by depositing small amounts of DNPH on Tenax GC (Ho and Yu 2004). The analysis of hydrazones is carried by GC–MS. Very good proﬁles of carbonyls from acetaldehyde to hexanal have been obtained with this method (Ho and Yu 2004). However, the electron beam of the MS must be shut off during the elution of unreacted DNPH. Since the reagent generates a large peak overlapping with formaldehyde, this compound cannot be quantiﬁed by GC–MS. To avoid overloading of the capillary column and the appearance of DNPH impurities in the chromatogram, the reagent must be extremely pure and the amount deposited on the Tenax surface accurately calculated (Ho and Yu 2004). Another derivatising agent allowing the selective collection of carbonyl compounds in air is perﬂuorophenylhydrazine (PFPH) (Cecinato et al. 2001, 2002; Schlomski et al. 1997).

332

Volatile Organic Compounds in the Atmosphere

It has the same efﬁciency as DNPH but hydrazones formed have a much higher volatility and a better thermal stability. This makes their analysis by GC–MS easier. Sample volumes needed for the analysis are smaller than those required with conventional DNPH-coated cartridges (Cecinato et al. 2001). The only limitation associated with the use of PFPH is that reaction with carbonyls often gives rise to two different hydrazones (the E and Z isomers) (Cecinato et al. 2001). Only compounds that are symmetrical with respect to the functional group (such as formaldehyde and acetone) form single products. Studies performed on the ozone interferences (Cecinato et al. 2001) have shown that an ozone scrubber is needed to prevent artefacts formation (Cecinato et al. 2001, 2002). Only one attempt has been made to combine conventional DNPH-coated cartridges with uncoated adsorption traps ﬁlled with graphitic carbon materials (Possanzini et al. 2000). The system performed quite well in the analysis of carbonyls and high-molecularweight VOCs, but the identiﬁcation and quantiﬁcation of the whole sample required the use of different analytical systems. Moreover, front traps, made by tubes ﬁlled with carbon materials, needed to be changed when 10 L of air were passed through the sampling system. Denuders coated with o-benzylhydroxyl ammonium chloride have been successfully used for the gas-phase collection of 4-OPA, nonanal, decanal, pinonic acid and few other volatile carbonyls (Matsunaga et al. 2003, 2004a, 2004b; OSOA-Final report and Hoffmann 1999). The benzylhydroxyl oximes formed by reaction with carbonyl compounds were extracted with ethyl acetate, and the solution analysed by GC-FID. To prevent degradation of oximes by ozone, a ﬂow of NO was added to the air stream entering the denuder system. So far direct analysis of higher-molecular-weight carbonyls has been performed only with PTR-MS (Karl et al. 2001). Basically, the instrument is a mass spectrometer working with chemical ionisation. High sensitivity is achieved by using a ﬂow tube to obtain very high ionisation efﬁciencies of the molecules to be identiﬁed. H3 O+ ions generated in the source are mixed with the air sample inside the ﬂow tube. Efﬁcient oxidation of carbonyl compounds with more than one carbon atom takes place through an acid-base reaction, in which a proton is transferred to the neutral molecule. From this reaction, a parent ion with a mass of (M + 1) is generated. Since the energy transferred by the proton to the neutral molecule is on the order of a few eV, fragmentation of the parent ions is very small or lacking. The main limit of this instrument is the impossibility to distinguish isobaric compounds, as they generate parent ions of the same mass/charge ratio. Therefore, carbonyls such as MAC and MVK cannot be individually quantiﬁed. The same holds true for hexanals and hexanols. So far, this instrument has been successfully applied to the direct determination of alkanals only in air sheds in which no interfering compounds were present (Karl et al. 2001). Even with these limitations, monitoring with PTR-MS is extremely advisable because the measurements are instantaneous and do not require any ozone scrubber. The ideal situation is to use PTR-MS in parallel with conventional methods to check the levels of carbonyls and the possible occurrence of sampling artefacts.

8.5.2

Carboxylic acids

For high-molecular-weight alkanoic acids, the simplest and reliable sampling method remains the one based on ﬁlters coated with alkaline agents (usually KOH), placed after the ﬁlter used for particle collection (Satsumabayashi et al. 1995). Acids can be analysed by

High-Molecular-Weight Carbonyls and Carboxylic Acids

333

GC-FID or GC–MS, after derivatisation. Denuder sampling has also been used for the collection of these compounds, but only in emission sources (Schauer et al. 1999a, 2001, 2002a, 2002b). Although PTR-MS has a great potential in the analysis of acids, no applications have been reported so far.

8.5.3

Calibration methods

Calibration of the instrumentation is a critical step in the analysis of high-molecular-weight carbonyls and carboxylic acids as it affects both the accuracy and precision of determinations. It has been shown that good gaseous standards of alkanals up to nonanal can be obtained (Ciccioli et al. 2002) by using the system proposed by Gautrois and Koppmann (1999). A constant source of VOCs is obtained by diffusion of vapours through a capillary tube in equilibrium with the liquid. Vapours emitted are mixed with the ﬂow of an inert gas (usually helium or nitrogen) to generate gas mixtures at known concentrations. A reliable source is obtained by keeping the vials at constant temperature and pressure. A schematic diagram of the apparatus for the generation of standard mixtures of VOCs, including the systems needed for the addition of ozone and water vapours, can be found in Ciccioli et al. (2002), Gautrois et al. (1999) and Larsen et al. (1997). It is worth noting that this calibration method generates primary standards because the mass of VOCs released by capillary diffusion can be accurately determined by measuring the weight losses of liquid from the vials. With preconcentration devices, adsorption of few μL of liquid standard solutions, containing the compounds of interest at ppmv levels, is an alternative calibration method. The solvent, usually a chlorinated solvent, can be removed by passing adequate volumes of clean air (or helium) through the trap. Very good consistency has been reported between this method and the one based on diffusion tubes (Ciccioli et al. 2002). The use of liquid solutions is particularly suitable for carboxylic acids (Ciccioli et al. 2002). Gas cylinders ﬁlled with standard gaseous mixtures are not reliable for the calibration of high-molecular-weight carbonyls and carboxylic acids, because selective adsorption and condensation can take place in the container walls. Concentrations can drastically change as a function of the time and ambient temperature.

8.6

Conclusions

After 15 years of data collection and analysis, the hypothesis of the ubiquitous occurrence of high-molecular-weight carbonyls in the troposphere is now a consolidated reality. It is supported by reliable data on the emission and a better knowledge of photochemical processes leading to their formation, either in the gas phase or in condensed phases. These compounds deﬁnitely contribute to the production of ozone, PANs, TAN and, very likely, to the mass of SOA. Although their sampling is not performed on a routine basis, more and more laboratories now start to recognise their importance, and actions are taken to monitor them. However, research is lacking on new methods for their collection and detection in air. Data on alkanoic acids are still extremely limited. This is only partly justiﬁed by their low atmospheric reactivity and aerometric levels. Also, in this case, almost no research has been performed in the development of new methods for their collection and analysis.

334

Volatile Organic Compounds in the Atmosphere

References Alvarado, A., Arey, J. and Atkinson, R. (1998) Kinetics of the gas-phase reactions of OH and NO3 radicals and O3 with the monoterpene reaction products pinonaldehyde, caronaldehyde, and sabinaketone. Journal of Atmospheric Chemistry, 31: 281–97. Andreae, M.O. and Merlet, P. (2001) Emission of trace gases and aerosols from biomass burning. Global Biogeochemical Cycles, 15: 955–66. Arey, J., Crowley, D.E., Crowley, M., Resketo, M. and Lester, J. (1995) Hydrocarbons emissions from natural vegetation in California’s South coast air basin. Atmospheric Environment, 29: 2977–88. Aschmann, S.M., Reissell, A., Atkinson, R. and Arey, J. (1998) Products of the gas phase reactions of the OH radical with α- and β-pinene in the presence of NO. Journal of Geophysical Research, 103: 25553–61. Atkinson, R. and Arey, J. (2003) Gas-phase tropospheric chemistry of biogenic volatile organic compounds: A review. Atmospheric Environment, 37: S197–219. Atkinson, R., Baulch, D.L., Cox, R.A., et al. (2005) Evaluated kinetic and photochemical data for atmospheric chemistry: Volume II – reactions of organic species. Atmospheric Chemistry and Physics Discussion, 5: 6295–7168. Atkinson, R. and Aschmann, S.M. (1993) Atmospheric chemistry of the monoterpene reaction products nopinone, camphenilone, and 4-acetyl-1-methylcyclohexene. Journal of Atmospheric Chemistry, 16: 337–48. Audino Gonzalez, P., Villaverde, R., Alfaro, R. and Zerba, E. (2005) Identiﬁcation of volatile emissions from Platypus mutatus (sulcatus) (Coleoptera: Platypodidae) and their behavioral activity. Journal of Economic Entomology, 98: 1506–9. Bakeas, E.B., Argyris, I.D. and Siskos, A.P. (2003) Carbonyl compounds in the urban environment of Athens, Greece. Chemosphere, 52: 805–13. Baker, J., Arey, J. and Atkinson, R. (2004) Kinetics of the gas-phase reactions of OH radicals, NO3 radicals and O3 with three C7 -carbonyls formed from the atmospheric reactions of myrcene, ocimene and terpinolene. Journal of Atmospheric Chemistry, 48: 241–60. Birkett, M.A., Agelopoulos, N., Jensen, K.M., et al. (2004) The role of volatile semiochemicals in mediating host location and selection by nuisance and disease-transmitting cattle ﬂies. Medical and Veterinary Entomology, 18: 313–22. Bloss, C., Wagner, V., Jenkin, M.E., et al. (2005) Development of a detailed chemical mechanism (MCMv3.1) for the atmospheric oxidation of aromatic hydrocarbons. Atmospheric Chemistry and Physics, 5: 641–64. Bonn, B., von Kuhlmann, R. and Lawrence, M.G. (2004) High contribution of biogenic hydroperoxides to secondary organic aerosol formation. Geophysical Research Letters, 31: L10108. Bowman, J.H., Barket, D.J. Jr, and Shepson, P.B. (2003) Atmospheric chemistry of nonanal. Environmental Science and Technology, 37: 2218–25. Boy, M., Petäjä, T., Dal Maso, M., et al. (2004) Overview of the ﬁeld measurement campaign in Hyytiälä, August 2001 in the framework of the EU project OSOA. Atmospheric Chemistry and Physics, 4: 657–78. Calogirou, A., Jensen, N.R., Nielsen, J.C., Kotzias, D. and Hjorth, J. (1999) Gas-phase reactions of nopinone, 3-Isopropenyl-6-oxo-heptanal, and 5-Methyl-5-vinyltetrahydrofuran-2-ol with OH, NO3 , and Ozone. Environmental Science and Technology, 33: 453–60. CATOME-Final Report and Dye, C.M. (ed.) (1999) CEC Contract No. ENV4-CT-97-0416-CATOME, NILU, Lillestrom, Norway. Cecinato, A., Di Palo, V., Mabilia, R. and Possanzini, M. (2001) Pentaﬂuorophenylhydrazine as a coating reagent for the HRGC-MS determination of semi-volatile carbonyl compounds in air. Chromatographia, 54: 263–7.

High-Molecular-Weight Carbonyls and Carboxylic Acids

335

Cecinato, A., Yassaa, N., Di Palo, V. and Possanzini, M. (2002) Observation of volatile and semivolatile carbonyls in an Algerian urban environment using dinitrophenylhydrazine/silica-HPLC and pentaﬂuorophenylhydrazione/silica-GC-MS. Journal of Environmental Monitoring, 4: 1–7. Chung, S.H. and Seinfeld, J.H. (2002) Global distribution and climate forcing of carbonaceous aerosols. Journal of Geophysical Research, 107: 4407. Ciccioli, P., Brancaleoni, E., Cecinato, A., Sparapani, R. and Frattoni, M. (1993b) Identiﬁcation and determination of anthropogenic volatile organic compounds in forest areas of northern and southern Europe and a remote site of the Himalaya region by high resolution gas chromatography-mass spectrometry. Journal of Chromatography A, 643: 55–69. Ciccioli, P., Brancaleoni, E., Frattoni, M., Cecinato, A. and Brachetti, A. (1993a) Ubiquitous occurrence of semi-volatile carbonyl compounds in tropospheric samples and their possible source. Atmospheric Environment, 27A: 1891–901. Ciccioli, P., Cecinato, A., Brancaleoni, E., Brachetti, A., Frattoni, M. and Sparapani, R. (1994) Composition and distribution of polar and non-polar VOCs in urban, rural, forest and remote areas. In: A. Angeletti and G. Restelli (Eds) Proceedings of the sixth European Symposium, EUR Report15609 EN, European Communities, Brussels, pp. 549–67. Ciccioli, P., Brancaleoni, E. and Frattoni, M. (1999) Reactive hydrocarbons in the atmosphere at urban and regional scales. In: C.N. Hewitt (Ed.) Reactive Hydrocarbons in the Atmosphere. Academic Press, San Diego, CA, pp. 159–207. Ciccioli, P., Brancaleoni, E., Frattoni, M., Cecinato, A. and Pinciarelli, L. (2001) Determination of volatile organic compounds (VOC) emitted from biomass burning of Mediterranean vegetation species. Analytical Letters, 34: 939–55. Ciccioli, P., Brancaleoni, E., Frattoni, M. and Maris, C. (2002) Sampling of atmospheric volatile organic compounds (VOCs) with sorbent tubes and their analysis by GC-MS. In: F.R. Burden, I. McKelvie, U. Forstner and A. Guenther (Eds) Environmental Monitoring Handbook. McGraw-Hill, New York, pp. 21.1–85. Ciccioli, P., Cecinato, A., Brancaleoni, E. and Frattoni, M. (1996) Occurrence of oxygenated volatile organic compounds (VOC) in Antartica. International Journal of Analytical Chemistry, 62: 245–53. Ciccioli, P., Cecinato, A., Brancaleoni, E. and Frattoni, M. (1992) Use of adsorption traps combined with high-resolution gas chromatography-mass spectrometry for the analysis of polar and non−C14 volatile organic compounds involved in photochemical smog formation. Journal of polar C4 − High Resolution Chromatography and Chromatography Communications, 15: 75–84. Ciccioli, P., Fabozzi, C., Brancaleoni, E., et al. (1997) Use of the isoprene algorithm for predicting the monoterpene emission from the Mediterranean holm oak Quercus ilex L.: Performances and limits of this approach. Journal of Geophysical Research, 102: 23319–28. Claeys, M., Graham, B., Vas, G., et al. (2004) Formation of secondary organic aerosols through photooxidation of isoprene. Science, 303: 1173–6. Cleary, P.A., Murphy, J.G., Wooldridge, P.J., et al. (2005) Observations of total alkyl nitrates within the Sacramento urban plume. Atmospheric Chemistry and Physics Discussion, 5: 4801–43. D’Anna, B., Andresen, O., Gefen, Z. and Nielsen, C.L. (2001) Kinetic study of OH and NO3 radical reactions with 14 aliphatic aldehydes. Physical Chemistry Chemical Physics, 3: 3057–63. Day, D.A., Dillon, M.B., Wooldridge, P.J., et al. (2003) On alkyl nitrates, O3, and the ‘missing NOy ’. Journal of Geophysical Research, 108 D16, 4501. Derwent, R.G. and Jenkin, M.E. (1991) Hydrocarbons and the long range transport of ozone and PAN across Europe. Atmospheric Environment, 25A: 1661–78. Derwent, R.G., Jenkin, M.E. and Saunders, S.M. (1996) Photochemical ozone creation potentials for a large number of reactive hydrocarbons under European conditions. Atmospheric Environment, 25A: 181–99.

336

Volatile Organic Compounds in the Atmosphere

Di Lorenzo, A., Polletta, A., Ciccioli, P., Brancaleoni, E. and Cucinato, A. (1991) Emission of PAH from a light-duty diesel car as a function of the engine operating conditions: A possible approach for investigating the parameters affecting formation of toxic components in soot. In: M. Cooke, K. Loening and J. Merritt (Eds) Polynuclear Aromatic Hydrocarbons: Measurements, Means and Metabolism. Battelle Press, Columbus, OH, pp. 239–58. El-Sayed, A.M. (2005) The pherobase: Database of insect pheromones and semichemicals. http:/www. Pherobase/compound/15 November 2005. Ezquerro, O., Pons, B. and Tena, M.T. (2002) Development of a headspace solid-phase microextraction–gas chromatography–mass spectrometry method for the identiﬁcation of odour-causing volatile compounds in packaging materials. Journal of Chromatography A, 963: 381–92. Fall, R. (1999) Biogenic emissions of volatile organic compounds from higher plants. In: C.N. Hewitt (Ed.) Reactive Hydrocarbons in the Atmosphere. Academic Press, San Diego, CA, pp. 41–96. Feng, Y., Wen, S., Chen, Y., et al. (2005) Ambient levels of carbonyl compounds and their sources in Guangzhou, China. Atmospheric Environment, 39: 1789–800. Feng, Y., Wen, S., Wang, X., et al. (2004) Indoor and outdoor carbonyl compounds in the hotel ballrooms in Guangzhou, China. Atmospheric Environment, 38: 103–12. Finalyson-Pitts, J.B. and Pitts, J.N. (2000) Chemistry of the Upper and Lower Atmosphere-Theory, Experiments and Applications. Academic Press, San Diego, CA. Fruekilde, P., Hjorth, J., Jensen, N.R., Kotzias, D. and Larsen, B. (1997) Ozonolysis at vegetation surface: A source of acetone, 4-oxopentanal, 6-methyl-5-heptene-2-one and geranyl acetone in the troposphere. Atmospheric Environment, 32: 1893–902. Gautrois, M. and Koppmann, R. (1999) Diffusion technique for the production of gas standards for atmospheric measurements. Journal of Chromatography A, 848: 239–49. Greenberg, J. and Zimmerman, P.R. (1984) Non methane hydrocarbons in remote, tropical, continental, and marine atmosphere. Journal of Geophysical Research, 89: 4767–78. Grosjean, D., Grosjean, E. and Gentler, A.W. (2001) On road emissions carbonyls from light-duty and heavy-duty vehicles. Environmental Science and Technology, 35: 45–53. Grosjean, D., Grosjean, E. and Moreira, L.F.R. (2002) Speciated ambient carbonyls in Rio de Janeiro, Brazil. Environmental Science and Technology, 36: 1389–95. Grosjean, E. and Grosjean, D. (1996) Carbonyl products from the gas-phase reaction of ozone with 1-alkenes. Atmospheric Environment, 24: 4107–13. Grosjean, E., Green, P.G. and Grosjean, D. (1999) Liquid Chromatography analysis of carbonyl (2,4-Dinitrophenyl)hydrazones with detection by diode array ultraviolet spectroscopy and by atmospheric pressure negative chemical ionization mass spectrometry. Analytical Chemistry, 70: 1851–61. Grosjean, E., Grosjean, D., Fraser, M.P. and Cass, G.R. (1996) Air quality model evaluation data for organics. 2. C1 -C14 carbonyls in Los Angeles air. Environmental Science and Technology, 36: 1389–95. Guenther, A. (1999) Modeling biogenic volatile organic compound emissions to the atmosphere. In: C.N. Hewitt (Ed.) Reactive Hydrocarbons in the Atmosphere. Academic Press, San Diego, CA, pp. 97–118. Hellén, H., Hakola, H., Reissell, A. and Ruuskanen, T.M. (2004) Carbonyl compounds in boreal coniferous forest air in Hyytiala, Southern Finland. Atmospheric Chemistry and Physics, 4: 1771–80. Helmig, D. and Greenberg, J.P. (1994) Automated in situ gas chromatographic-mass spectrometric analysis of ppt level volatile organic trace gases using multistage solid adsorbing trapping. Journal of Chromatography A, 677: 123–32. Helmig, D. (1997) Ozone removal techniques in the sampling of atmospheric volatile organic trace gases. Atmospheric Environment, 31: 3635–51.

High-Molecular-Weight Carbonyls and Carboxylic Acids

337

Helmig, D., Pollock, W., Greemberg, J. and Zimmerman, P. (1996) Gas chromatography mass spectrometry analysis of volatile organic trace gases at Mauna Loa Obseervatory, Hawaii. Journal of Geophysical Research, 101: 14697–710. Ho, K.F., Lee, S.C., Louie, P.K.K. and Zou, S.C. (2002) Seasonal variation of carbonyl compound concentrations in urban area of Hong Kong. Atmospheric Environment, 36: 1259–65. Ho, S.S.A. and Yu, J.Z. (2004) Determination of airborne carbonyls: Comparison of a thermal desorption/GC method with the standard DNPH/HPLC method. Environmental Science and Technology, 38: 862–70. Iinuma, Y., Boge, O., Miao, Y., Siero, B., Gnauk, T. and Herrmann, H. (2005) Laboratory studies on secondary organic aerosol formation from terpenes. Faraday Discussions, 130: 279–94. Isidorov, V.A. (1992) Non-methane hydrocarbons in the atmosphere of boreal forests: Composition, emission rates, estimation of regional emission and photocatalytic transformation. Ecology Bulletins, 42: 71–6. Lelieveld, J., Crutzen, P.J., Ramanathan, V., et al. (2001) The Indian Ocean experiment: Widespread air pollution from South and Southeast Asia. Science, 291: 1031–6. Jang, M., Czoschke, N.M., Lee, S. and Kamens, R.M. (2002) Heterogeneous atmospheric aerosol production by acid-catalyzed particle-phase reactions. Science, 298: 814–17. Jensen, B., Wolkoff, P. and Willkins, C.K. (1995) Characterization of linoleum. Part 2: Preliminary odor evaluation. Indoor Air, 5: 44–9. Jurvelin, J.A., Edwards, R.D., Vartainen, M., Pasanen, P. and Jantunen, M. (2003) Residential indoor, outdoor, and workplace concentrations of carbonyl compounds: Relationships with personal exposure concentrations and correlation with sources. Journal of the Air and Waste Management Association, 53: 560–73. Karl, T., Crutzen, P.J., Mandl, M., et al. (2001) Variability-lifetime relationship of VOCs observed at the Sonnblick Observatory 1999 Estimation of HO-densities. Atmospheric Environment, 35: 5287–300. −C10 monoKawamura, K., Steinberg, S. and Kaplan, I.R. (2000) Homologous series of C1 − −C6 carbonyls in Los Angeles air and motor vehicle exhaust. Atmospheric carboxylic acids and C1 − Environment, 34: 4175–91. Kean, A.J., Grosjean, E., Grosjean, D. and Harley, A.R. (2001) On road emissions carbonyls from light-duty and heavy-duty vehicles. Environmental Science and Technology, 35: 45–53. Kesselmeier, J., Bode, K., Hofmann, U., et al. (1997) Emission of short chained organic acids, aldehydes and monoterpenes from Quercus ilex L. and Pinus Pinea L. in relation to physiological activity, carbon budget and emission algorithms. Atmospheric Environment, 31 (SI): 119–34. Kesselmeier, J., Kuhn, U., Rottembeger, S., et al. (2002) Concentrations and species composition of atmospheric volatile organic compounds (VOCs) as observed during the wet and dry season in Rondonia (Amazonia). Journal of Geophysical Research, 107: D20, 8053. Kesselmeier, J. and Staudt, M. (1999) Biogenic Volatile Organic Compounds (VOC): An overview on emission, physiology and ecology. Journal of Atmospheric Chemistry, 33: 23–88. Kessler, A. and Baldwin, I.T. (2001) Defensive function of herbivore-induced plant volatile emissions in nature. Science, 291: 2141–3. Kirstine, W., Galbally, I., Ye, Y. and Hooper, M. (1998) Emissions of volatile organic compounds (primarily oxygenated species) from pasture. Journal of Geophysical Research, 103: 10605–19. Knöppel, H. and Schauenburg, H. (1989) Screening of household products for the emission of volatile organic compounds. Environment International, 15: 413–18. Kolloker, S., Oehme, M. and Dye, C. (1998) Structure elucidation of 2-dinitrophenylhydrazone derivatives of carbonyl compounds in ambient air by HPLC-MS and multiple MS-MS using atmospheric chemical ionization in negative ion mode. Analytical Chemistry, 70: 1979–85. König, G., Brunda, M., Puxbaum, H., Hewitt, C.N., Duckham, S.C. and Rudolph, J. (1995) Relative contribution of oxygenated hydrocarbons to the total biogenic VOC emissions of

338

Volatile Organic Compounds in the Atmosphere

selected mid-European agricultural and natural plant species. Atmospheric Environment, 29: 861–74. Kotzias, D. and Hjorth, J. (1991) Sampling and analysis of selected volatile organic compounds (VOC) relevant to the formation of photochemical oxidants. Pollution Atmospherique, 131, numéro special: 209–16. Larsen, B., Bomboi-Mingarro, T., Brancaleoni, E., et al. (1997) Sampling and analysis of terpenes in air. An interlaboratory comparison. Atmospheric Environment, 31 (SI): 35–49. Larsen, B.R., Tudos, A., Slanina, J., Van der Borg, K. and Kotzias, D. (2001) Quantiﬁcation of airborne fossiland biomass carbonylic carbon by combined radiocarbon and liquid chromatography mass spectrometry. Atmospheric Environment, 35: 5695–707. Lichtenthaler, H.K. (1999) The 1-deoxy-D-xylulose-5-phosphate pathway of isoprenoid biosynthesis in plants. Annual Review of Plant Physiology and Plant Molecular Biology, 50: 47–65. Loreto, F., Ciccioli, P., Brancaleoni, E., Cecinato, A., Frattoni, M. and Sharkey, T. (1996) Different sources of reduced carbon contribute to form three classes of terpenoid emitted by Quercus ilex L. leaves. Proceedings of the National Academy of Science USA, 93: 9966–9. Matsunaga, S., Mochida, M. and Kawamura, K. (2003) Growth of organic aerosols by biogenic semivolatile carbonyls in the forestal atmosphere. Atmospheric Environment, 37: 2045–50. Matsunaga, S., Mochida, M. and Kawamura, K. (2004) High Abundance of gaseous and particulate 4-oxopentanal in the forestal atmosphere. Chemosphere, 55: 1143–7. Matsunaga, S., Mochida, M. and Kawamura, K. (2004) Variation in the atmospheric concentration of biogenic carbonyl compounds and their removal processes in the Northern forest at Moshiri, Hokkaido Island Japan. Journal of Geophysical Research, 109: D04302. McClenny, W.A., Colon, M. and Oliver, K.D. (2001) Ozone reaction with n-aldehydes (n = 4–10), benzaldehyde, ethanol, isopropanol, and n-propanol adsorbed on a dual-bed graphitized carbon– carbon molecular sieve adsorbent cartridge. Journal of Chromatography A, 929: 89–100. McClenny, W.A., Daughtrey, E.H. Jr, Adams, J.R., Oliver, K.D. and Kronmiller, K.G. (1998) Volatile organic compound concentration patterns at the New Hendersonville monitoring site in the 1995 Southern Oxidants Study in the Nashville, Tennessee, area. Journal of Geophysical Research, 103: 22509. Micha, S.G., Stammel, J. and Holler, C. (1993) 6-Methyl-5-heptene-2-one, a putative sex and spacing pheromone of the aphid hyperparasitoid, Alloxysta victrix (Hymenoptera: Alloxystidae). European Journal of Entomology, 90: 439–42. Moortgat, G.K. (2001) Important photochemical processes in the atmosphere. Pure and Applied Chemistry, 73: 487–90. Moortgat, G.K., Grossmann, D., Boddemberg, A., et al. (2002) Hydrogen peroxide, organic peroxides and higher carbonyl compounds determined during the BERLIOZ campaign. Journal of Atmospheric Chemistry, 42: 443–63. Muller, K. (1997) Determination of aldehydes and ketones in the atmosphere: A comparative long time study at an urban and a rural site in Eastern Germany. Chemosphere, 35: 2093–106. Muller, K., van Pinxteren, D., Plewka, A., et al. (2005) Aerosol characterisation at the FEBUKO upwind station Goldlauter (II): Detailed organic chemical characterization. Atmospheric Environment, 39: 4219–31. OSOA-Final report and Hoffmann, T. (Ed.) (1999) Origin and Formation of Secondary Organic Aerosol. CEC Contract No. EVK2-CT-1999-00016, ISAS, Dortmund, Germany. Owen, S., Boissard, C., Street, R.A., Duckham, C., Csiky, O. and Hewitt, C.N. (1997) Screening of 18 Mediterranean plant species for volatile organic compound emissions. Atmospheric Environment, 31 (SI): 101–18. Parker, T., Dottridge, J. and Kelly, S. (2002) Investigation of the Composition and Emissions of Trace Components in Landﬁll Gas. R&D Technical Report P1-438/TR Environment Agency, Bristol, UK.

High-Molecular-Weight Carbonyls and Carboxylic Acids

339

Peck, G.E. (2004) Potential Semiochemicals of Wheat (Triticum aestivum L.) Induced by Oviposition and Feeding of the Wheat Stem Sawﬂy Cephus Cinctus Norton (Hymenoptera: Cephidae). Master of Science Thesis, Montana State University, Bozeman, MO. Pires, M. and Carvalho, L.R.F. (1998) An artifact in air carbonyls sampling using C18 DNPH-coated cartridge. Analytica Chimica Acta, 367: 223–31. Plagens, H. (2001) Untersuschungen zum atmosparenchemischen Abbau langkettiger Aldehyde, Doctorat Thesis in Chemistry, Bergischen Universitat Gesamthochschule Wuppertal, Wuppertal, available on-line at: http://elpub.bib.uniwuppertal.de/edocs/dokumente/fb09/diss2001/ plagens;internal&action = buildframes.action. Possanzini, M., Di Palo, V., Brancaleoni, E., Frattoni, M. and Ciccioli, P. (2000) A train of DNPH-coated cartridges for the determination of carbonyls from C1 to C12 in air and emission samples. Atmospheric Environment, 34: 5311–18. Purdue, L.J., Reagan, J.A., Lonnemmand, W.A., et al. (1992) Atlanta Ozone Precursor Study. Report. EPA/ 600-R-92/157, U.S. Environmental Protection Agency, Research Triangle Park, NC. Puxbaum, H. (1997) Biogenic emissions of alcohols, ester, ether and higher aldehydes. In: G. Helas, S. Slanina and R. Steinbrecher (Eds) Biogenic Volatile Organic Compounds in the Atmosphere– Summary of Present Knowledge. SPB Academic Publishers, Amsterdam, The Netherlands, pp. 79–99. Quiroz, A., Wadhams, L.J., Niemeyer, H.M., Pettersson, J. and Pickett, J.A. (1997) Semiochemical mediating spacing behavior of bird cherry-oat aphid, Rhopalosiphum padi, feeding on cereals. Journal of Chemical Ecology, 23: 2599–607. Rainer, F. and Obermeier, A. (1999) Anthropogenic emissions of volatile organic compounds. In: C.N. Hewitt (Ed.) Reactive Hydrocarbons in the Atmosphere. Academic Press, San Diego, CA, pp. 1–39. Ramanathan, V., Crutzen, P.J., Kiehl, J.T. and Rosenfeld, D. (2001) Aerosols, climate, and the hydrological cycle. Science, 294: 2119–24. Reissell, A. and Arey, J. (2001) Biogenic volatile organic compounds at Azusa and elevated sites during the 1997 Southern California ozone study. Journal of Geophysical Research, 106: 1607–21. Reissell, A., Herry, C., Aschmann, S.M., Atkinson, R. and Arey, J. (1999) Formation of acetone from the OH-radical- and O3 -initited reactions from a series of monoterpenes. Journal of Geophysical Research, 104: 13869–79. Reissell, A., Arey, J. and Atkinson, R. (2001) Atmospheric chemistry of camphor. International Journal of Chemical Kinetics, 33: 56–63. Roberts, J.M., Fehsenfeld, F.C., Albritton, D.L. and Sievers, R.E. (1984) Sampling and analysis of monoterpene hydrocarbons in the atmosphere with Tenax gas chromatographic porous polymer. In: L.H. Keith (Ed.) Identiﬁcation and Analysis of Organic Pollutants in Air. Butterworth: Wooburn, pp. 371–87. Roberts, J.M., Williams, J., Baumann, K., et al. (1998) Measurements of PAN, PPN and MPAN made during the 1994 and 1995 Nashville intensives of the southern oxidant study: Implications for regional ozone production from biogenic hydrocarbons. Journal of Geophysical Research, 103: 22473–90. Rudich, Y. (2003) Laboratory perspectives on the chemical transformations of organic matter in atmospheric particles. Chemical Reviews, 103: 5097–124. Sagebiel, J.C., Zielinska, B., Pierson, W.R. and Gertler, A.W. (1996) Real world emissions and calculated reactivities of organic species from motor vehicles. Atmospheric Environment, 12: 2287–844. Sartin, J.H., Halsall, C.J., Davison, B., Owen, S. and Hewitt, N.C. (2001) Determination of biogenic −C16 ) in the coastal atmosphere of Mace Head, Ireland. Analytica volatile organic compounds (C8 − Chimica Acta, 428: 61–72.

340

Volatile Organic Compounds in the Atmosphere

Satsumabayashi, H., Kurita, H., Chang, Y.S., Carnichael, M.R. and Ueda, H. (1995) Photochemical formation of lower aldehydes and lower fatty acids under long-range transport in central Japan. Atmospheric Environment, 29: 255–66. Schauer, J.J., Kleeman, M.J., Cass, G.R. and Simoneit, B.R.T. (2002a) Measurement of emissions −C32 organic compounds from gasoline-powered vehicles. from air pollution sources. 5. C1 − Environmental Science and Technology, 36: 1169–80. Schauer, J.J., Kleeman, M.J., Cass, G.R. and Simoneit, B.R.T. (2002b) Measurement of emissions from −C27 organic compounds from cooking with seed oils. Environmental air pollution sources. 4. C1 − Science and Technology, 36: 567–75. Schauer, J.J., Kleeman, M.J., Cass, G.R. and Simoneit, B.R.T. (2001) Measurement of emissions −C29 Organic compounds from ﬁreplace combustion wood. from air pollution sources. 3. C1 − Environmental Science and Technology, 35: 1716–28. Schauer, J.J., Kleeman, M.J., Cass, G. and Simoneit, B.T. (1999a) Measurement of emissions from air pollution sources. 2. C1 through C30 organic compounds from medium duty diesel trucks. Environmental Science and Technology, 33: 1578–87. Schauer, J.J., Kleeman, M.J., Cass, G. and Simoneit, B.T. (1999b) Measurement of emissions from air pollution sources. 1. C1 through C29 organic compounds from meat charbroiling. Environmental Science and Technology, 33: 1566–77. Schlomski, S., Kibler, M., Ebert, P., et al. (1997) Multiphase chemistry of organic acids and carbonyl compounds. In: B. Larsen, B. Versino and G. Angeletti (Eds) Proceedings of the seventh European Symposium Physico-Chemical Behaviour of Atmospheric Pollutants, EUR 17428 EN, European Communities, Brussels, pp. 557–61. Shaughnessy, R.J., McDaniels, T.J. and Weschler, C.J. (2001) Indoor chemistry: Ozone and volatile organic compounds found in tobacco smoke. Environmental Science and Technology, 35: 2758–64. Smith, A.M., Rigler, E., Kwok, E.S.C. and Atkinson, R. (1996) Kinetics and products of the gas-phase reactions of 6-methyl-5-hepten-2-one and trans-cinnamaldehyde with OH and NO3 radicals and O3 at 296◦ K. Environmental Science and Technology, 30: 1781–5. Solberg, S. (2005) VOC measurements 2003. EMEP/CCC-Report 10/2005, NILU, Kjeller, Norway. Speijers, G.J.A. (1993) VOCs and the environment and public health-health effects. In: Th.H.J. Bloemen and J. Burn (Eds) VOCs in the Environment. Chemistry and Analysis. Blackie Academical & Professional: London, pp. 24–91. Staudt, M., Bertin, N., Hansen, U., et al. (1997) Seasonal and diurnal patterns of monoterpene emissions from Pinus pinea (L.) under ﬁeld conditions. Atmospheric Environment, 31 (SI): 145–6. Tadic, M.J., Juranic, O.I. and Moortgat, G.K. (2002) Photooxidation of n-heptanal in air: Norrish type I and II processes and quantum yield total pressure dependency. Journal of Chemical Society Perkin Transaction, 2: 135–40. Tillman, A.J., Seybold, S.J., Jurenka, R.A. and Blomquist, G.J. (1999) Insect pheromones – an overview of biosynthesis and endocrine regulation. Insect Biochemistry and Molecular Biology, 29: 481–514. Tolocka, M.P., Myoseon, J., Ginter, J.M., Cox, J.F., Kamens, R.M. and Johnston, M.V. (2004) Formation of oligomers in secondary organic aerosols. Environmental Science and Technology, 38: 1428–34. UK PORG (1994) Ozone in the United Kingdom 1993. Third Report of the United Kingdom Photochemical Oxidants Review Group. Air Quality Division, Department of the Environment, London. Wadia, Y., Tobias, D.J., Stafford, R. and Finlayson-Pitts, B.J. (2000) Real-time monitoring of the kinetics and gas-phase products of the reaction of ozone with 1-oleoyl-2-palmitoyl-sn-glycero-3phosphocholine at the air/water interface. Langmuir, 16: 9321–30. Wildt, J., Kobel, K., Schuh-Thomas, G. and Heiden, C. (2003) Emissions of oxygenated volatile organic compounds from plants Part II: Emissions of saturated aldehydes. Journal of Atmospheric Chemistry, 45: 173–96.

High-Molecular-Weight Carbonyls and Carboxylic Acids

341

Winer, A., Arey, J., Atkinson, R., et al. (1992) Emission rates of of organics from from vegetation in California’s central valley. Atmospheric Environment, 26A: 2647–59. Yassaa, N., Meklati, B.Y., Brancaleoni, E., Frattoni, M. and Ciccioli, P. (2001) Polar and non-polar volatile organic compounds (VOCs) in urban Algiers and saharian sites of Algeria. Atmospheric Environment, 35: 787–801. Yasuhara, A. and Shibamoto, T. (1995) Quantitative analysis of volatile aldehydes formed from various kinds of ﬁsh ﬂesh during heat treatment. Journal of Agricultural and Food Chemistry, 43: 94–7. Yokouchi, Y., Mukai, H., Makajima, K. and Ambe, Y. (1990) Semi-volatile aldehydes as predominant organic gases in remote areas. Atmospheric Environment, 24: 439–42. Zhang, J., He, Q. and Lioy, P.J. (1994) Characteristics of aldehydes: Concentrations, sources and exposure for indoor and outdoor residential microenvironments. Environmental Science and Technology, 28: 146–52. Zhou, X. and Mopper, K. (1997) Photochemical production of low-molecular-weight carbonyl compounds in seawater and surface microlayer and their air-sea exchange. Marine Chemistry, 56: 201–13.

Volatile Organic Compounds in the Atmosphere Edited by Ralf Koppmann Copyright © 2007 by Blackwell Publishing Ltd

Chapter 9

Organic Aerosols Thorsten Hoffmann and Jörg Warnke

9.1

Introduction

Atmospheric aerosols interact both directly and indirectly with the Earth’s radiation budget and climate. As a direct effect, the aerosols scatter or absorb sunlight. As an indirect effect, aerosols in the lower atmosphere can modify number and size of cloud droplets, changing how the clouds reﬂect and absorb sunlight, thereby affecting the Earth’s radiation budget. Aerosols also can act as sites for chemical reactions to take place (heterogeneous chemistry). Hence, they play an important role in global climate and atmospheric chemistry. Furthermore, atmospheric aerosols affect our environment at the local and regional levels. Aerosols are now becoming recognised as a signiﬁcant health problem, especially in regard to respiratory diseases. The formation of organic aerosols from the oxidation of hydrocarbons is only one but important pathway that determines the overall composition of atmospheric aerosols. In general, the volatile aerosol precursors are ﬁrst degraded in the gas phase by bimolecular reactions with radicals or ozone or by photolysis, followed by the formation of products with a lower volatility. These products are higher functionalised compounds with hydroxyl, carbonyl, carboxyl groups or groups containing heteroatoms, which will either condense on existing particles or even form new aerosol particles. To distinguish this fraction of tropospheric aerosols from the direct input of particulate organics into the atmosphere, it is speciﬁed as secondary organic aerosol (SOA). Since primary and SOA particles are closely linked in several aspects about their potential role and signiﬁcance in atmospheric chemistry as well as their analytical treatment, we consider both primary and secondary contributions in this chapter. However, emphasis lies on the aerosol components from hydrocarbon oxidation.

9.1.1

Historical

The optical effects of atmospheric aerosol particles formed from the oxidation of volatile organic compounds (VOCs) were recognised quite early by humans. Sha-co-na-qe ‘Place of Blue Smoke’ was the name given for the Great Smoky Mountains by the Cherokee Indians, of course, without knowing the cause of the haziness often observed in the summertime above the forested Appalachian Highlands. This phenomenon of forested regions – that they are frequently enveloped in a blue haze or smoke – is created by Rayleigh scattering of submicrometre particles. Numerous other forested mountain sites are named after this

Organic Aerosols

343

phenomenon: ‘Blue Mountains’, ‘Blue Ridges’ or ‘Smoky Mountains’ can be found, for example, in India close to the Burman border, in Jamaica, at the east coast of Australia and even several times in North America (Montana, Oregon, Idaho, Maine, Pennsylvania, Tennessee). The ﬁrst connection between VOCs and the formation of atmospheric particles was probably made by Arie Haagen-Smit at Caltech in 1952 in a strongly anthropogenically inﬂuenced environment. Studying various aspects of the Los Angeles smog formation, he not only explained ozone and peroxide formation by the photochemistry of the released hydrocarbons and nitrogen oxides, but also related the decrease in visibility during smog episodes to the condensation of aldehydes and acids formed in the oxidation of the hydrocarbons. In 1960, F.W. Went, director of the Missouri Botanical Garden and former colleague of Haagen-Smit at Caltech, published an extensive article in Nature titled ‘Blue hazes in the atmosphere’. Based on his observations while staying in countryside and on his everyday experiences, as well as his knowledge about secondary plant products, he ﬁnally also connected the occurrence of the natural phenomena with the volatilisation and gas-phase oxidation of terpenes from terrestrial vegetation. However, during these ﬁrst years of atmospheric chemistry and the following decades, the main interest in the VOC chemistry was focused on gas-phase photochemistry. At about 1990, increasing interest arose to understand the aerosol formation behaviour of hydrocarbons, driven by the awareness of the role of natural and anthropogenic aerosols in the radiative properties of the atmosphere and the Earth’s climate. Moreover, the ozone hole research in the Antarctic clearly showed that heterogeneous reactions on surfaces of air suspended matter can inﬂuence gas-phase processes. Another major driver to investigate the origin and formation of aerosol particles in the last decades has been their effect on human health. It has been shown that cardio-pulmonary diseases and mortality are related to the presence of ﬁne particulate matter (Dockery et al. 1993; Laden et al. 2000; Mar et al. 2000; Tsai et al. 2000). As a consequence of the increasing scientiﬁc interest to understand aerosol-forming atmospheric processes, new instrumental techniques for particle analysis have been developed at a rapid rate in order to produce methods with lower detection limits, shorter temporal resolution and increased selectivity. Especially mass spectrometric online techniques have been developed during the last years, as reviewed by Sipin et al. (2003).

9.1.2

Sources and sinks of atmospheric particles

Particles in the atmosphere are divided into primary and secondary particles according to their formation processes. Primary particles are released directly into the atmosphere, whereas secondary particles are produced within the atmosphere as a consequence of the conversion of volatile precursors into low- or non-volatile substances that form new particles. Formation processes of primary particles are basically mechanical production (abrasion, suspension and sea spray) and production during combustion processes (condensation of hot vapours or formation inside ﬂames, such as soot particles) (Seinfeld and Pandis 1998). In general, mechanical processes create coarse particles whereas combustion processes create ﬁne particles. Particles are called “ﬁne particle” if their diameter is below 1 or 2.5 μm, respectively, depending on the

344

Volatile Organic Compounds in the Atmosphere

Coarse particles

Fine particles

Primary (mechanical production)

Primay (combustion) Secondary (nucleation and condensational growth) Nucleation (Molecules)

0.1 nm

Cluster

1 nm

Condensational growth Nucleation Aitken Accumulation mode mode mode 10 nm

100 nm

1 μm

Coarse mode 10 μm 100 μm Particle diameter

Figure 9.1 Size modes of atmospheric particles and important sources.

deﬁnition. In atmospheric measurements the considered size class of atmospheric aerosols is often described by the maximum diameter of sampled particles. A size class is then characterised by e.g. PM10 which means the upper diameter of sampled “particulate matter” is 10 μm. This term will be used frequently in this chapter (refer also to Section 9.3.1). Secondary atmospheric particles belong to the ﬁne particle fraction as well; they are created by the so-called ‘nucleation process’. The nucleation process in the troposphere is currently not completely understood, and several mechanisms are discussed (Kulmala 2003). One hypothesis about new particle production in the atmosphere assumes that the process is initialised by the formation of sulphuric-acid-containing clusters (thermodynamically stable clusters) in the size range of 1 nm, which grow under suitable conditions, for example, the availability of condensable vapours, into a size range of 3–20 nm, the so-called nucleation mode. If the concentration of condensable vapours is not high enough, the clusters will be rapidly lost by coagulation and no new particles will be formed. However, once formed the nucleation mode particles can continue to grow by uptake of condensable vapours into the Aitken mode (around 20 to 100 nm) and further to particles in the accumulation mode (100-nm range) (Kulmala et al. 2004). Condensable vapours mean low volatile compounds, which are produced during chemical reactions in the atmosphere from volatile precursors. Due to the Kelvin-Effect (smaller droplets have higher vapour pressures), these vapours cannot condense without a condensation nuclei (e.g. a cluster). Condensable vapours might be inorganic like sulphuric acid or organic like low volatile products of the terpene oxidation. These low volatile compounds are not only involved in the formation or growth of secondary particles but they can also condense onto pre-existing particles, leading to increased particle size and mass and to an alteration of the chemical composition. Changes in the chemical composition also alter the physical or physicochemical properties of the particles (light scattering, hygroscopicity, etc.), which affect the impact of aerosols on climate. Figure 9.1 illustrates size modes and sources of atmospheric particles. Processes that act as sinks of atmospheric particles are dependent on the particle sizes. Coarse particles are removed mainly by dry deposition, that is, by sedimentation. Particles in the accumulation mode are eliminated mostly by wet deposition (rainout, washout). The main sink for smaller particles (Aitken and nucleation mode) is by coagulation with other particles.

Organic Aerosols

9.1.3

345

General chemical composition of aerosol particles and aerosol mass in the troposphere

Particles can be characterised by their origin: soot from combustion, mineral dust from levitated crustal material, salt from sea spray, pollen from vegetation and so on. By chemical analysis, the source characteristics of the different particles are replaced by the chemical composition of the whole sample. One approach to characterise the aerosol sample by its chemical composition is to divide the aerosol mass into several compound classes, for example, organic matter (OM), elemental carbon (EC), non-sea-salt (NSS) sulphate, nitrate, ammonia, sea salt, mineral dust (Hueglin et al. 2005; Putaud et al. 2004; Rees et al. 2004; Sellegri et al. 2003; Ward et al. 2004). EC, nitrate and ammonium are deﬁned chemical components whose concentration can be measured. In contrast, NSS-sulphate, sea salt, OM and mineral dust are no single chemical compounds, and, therefore, they have to be calculated from the measured content of selected tracer species. Sea salt can be estimated using the measured concentrations of Na+ and Cl− or Na+ and SO2− 4 and a standard sea water composition. NSS-sulphate is the difference between measured sulphate and calculated fraction of sea salt sulphate. OM can be estimated from measured organic carbon (OC) by multiplying with a certain factor (often 1.4, refer to Section 9.2.3) representing the elemental composition of organic substances present in the atmosphere. Metal species are used as tracers to calculate the mass of particulate minerals (mineral dust) suspended in the atmosphere. Mineral dust (sometimes also called crustal material) consists of oxides, silicates, SiO2 or other minerals containing iron, aluminium, calcium, sodium, potassium and other metal ions. Nitrate and NSS-sulphate are mainly secondary aerosol constituents. Nitrate is formed mostly from anthropogenic NOx -emissions by atmospheric oxidation to nitric acid. Sulphate is formed from SO2 -emissions (predominantly anthropogenic) or emissions of sulphur-containing VOCs, like dimethyl sulphide from marine environments, by oxidation to sulphuric acid (Finlayson-Pitts and Pitts 2000). The predominance of primary aerosol constituents in the coarse and secondary aerosol constituents in the ﬁne-particle fraction can be seen in Figure 9.2 comparing the composition of the ﬁne and coarse fraction at different locations in Europe. The overall mass of the aerosol can be measured directly by weighing. Another method is the calculation of mass from the measured particle size distribution. The size distribution can be used to calculate aerosol volume and, supposing a certain mean density for aerosol particles, aerosol mass can be estimated. Aerosol mass is in the μg/m3 range, depending strongly on the sampling site, meteorological conditions, seasonal and diurnal cycles. The annual mean concentration of PM10 , for example, in locations in Scandinavia, is below 10 μg/m3 , whereas concentrations in urban sites in central Europe can reach mean concentrations of >50 μg/m3 . Also, the aerosol-relative composition strongly depends on location and time.

9.2

Carbonaceous aerosols

The term ‘carbonaceous aerosols’ includes all aerosol constituents that are based on carbon, for example, the variety of different organic compounds, EC, bioaerosols and inorganic constituents. Except inorganic carbon, the characteristics of these different carbonaceous

346

Volatile Organic Compounds in the Atmosphere

Atmospheric concentration (μg/m3)

40 35

unacc.

30

NO3−

25

NH4+ NSS-SO42 −

20

OM

15

BC

10

Sea salt min. dust

5 0 Fine Coarse Natural

Fine Coarse Rural

Fine Coarse Urban

Fine Coarse Kerbside

Figure 9.2 Annual mean composition of ﬁne (PM2.5 ) and coarse (PM2.5 –PM10 ) aerosol fraction at a natural (Sevettijarvi (FIN)), rural (Illmittz (A)), urban (Zuerich (CH)) and kerbside (Barcelona (E)) site. Abbreviations: unacc. = unacccounted mass, NSS = non-sea-salt, OM = organic matter, BC = black carbon, min. dust = mineral dust. Data from http://ies.jrc.cec.eu.int/Download/cc/6.xls, http://ies.jrc.cec.eu.int/Download/cc/5.xls.

fractions will be discussed in the following sections. The concentration of inorganic carbon, essentially as carbonate, is negligible on an average, at least when considering PM2.5 . If inorganic carbon occurs in atmospheric PM, it is limited to the form of carbonate-containing mineral dust, which is part of the coarse mode of PM.

9.2.1

Bioaerosols

Bioaerosols are primary organic aerosols with diameters from ∼10 nm to 100 μm, which are either alive (viruses, bacteria, fungi, algae), carry living organisms or are released from living organisms (pollen, spores, cell debris) (Ariya and Amyot 2004). Roughly, the size of bacteria is around 1 μm, pollen grains are mostly >10 μm and viruses are in the nanometer range. Each of these ‘particles’ is usually itself a complex mixture of various molecules. Bacteria may spread diseases, can act as cloud condensation nuclei and ice nuclei. They were found even at high altitudes in the atmosphere and remote regions. Bacteria and fungi can be suspended from soil or plants by wind and from water surfaces by bubble bursting processes or sea spray. They can also be released by anthropogenic sources such as farming, waste and waste water treatment. Bacteria can even live and grow in atmospheric water droplets like fog (Fuzzi et al. 1997) or even super-cooled cloud droplets (Sattler et al. 2001). The importance of bioaerosols for the atmospheric aerosol content is very unclear. Some authors report bioaerosols as major components, whereas other studies report only insigniﬁcant contribution of bacteria to the atmospheric particulate material. In Amazonian aerosols, the nocturnal increase of coarse size particle mass (PM10 –PM2 ) was assigned to fungi (Graham et al. 2003b). During the wet season, biogenic particles accounted for

Organic Aerosols

347

55–92% of the ﬁne-particle mass and for 65–95% of the coarse-particle mass sampled in the Amazonian basin (Artaxo et al. 1988, 1990). For Russia (Lake Baikal) and Germany (Mainz) (Jaenicke 2005), contributions of primary biological aerosol particles (including plant fragments, pollen, etc.) to total atmospheric particles (>0.2 μm) were reported to be in the range of 20–30%, respectively. On the other hand, studies from the southern ocean report that bacteria represent only about 1% of the total number of particles >0.2 μm (Posfai et al. 2003) or did not even indicate the presence of bacteria in marine or continental samples (Bates et al. 1998). In Hong Kong, the dry mass load of Gram-negative bacteria was estimated to be in the range of 10–100 ng/m3 , which is small compared to the total OM present in the atmosphere (Lee et al. 2004). Nevertheless, bacteria were even found at high altitudes (5.4 km) in the same relative (low) concentration as near to the surface. There is also evidence that microbes can alter the composition of existing aerosol particles (Ariya et al. 2002) using aerosol constituents as nutrients. Dicarboxylic acids, for example, may be transformed into volatile products by airborne microorganisms.

9.2.1.1

Measurements of bacteria and fungi

One approach to measure bacteria and fungi is to count them directly on the collection media (ﬁlter, impactor plates), using light microscopes with or without staining procedures, epiﬂuoresence microscopy (Grifﬁn et al. 2001), electron microscopes or electron probe X-ray microanalysis (EPXMA) (Graham et al. 2003b). The direct counting can be used for the measurement of all types of bioaerosols; staining procedures usually need speciﬁc target molecules, such as proteins. Another approach to measure and identify living cells is to collect them onto culture media and count the forming colonies. Until now, this technique is the most widely used method for monitoring bacteria. However, not all species are able to grow on these media and therefore will not be detected. The identiﬁcation of microbes can be done by their morphology or based on a genetic identiﬁcation using the polymerase chain reaction (PCR) (Grifﬁn et al. 2001). Also, speciﬁc markers can be used to measure microbes in the atmosphere. Applying this approach, living as well as dead microbes will be measured. Lee and co-workers (2004) used 3-hydroxy fatty acids as biomarkers for Gram-negative bacteria in aerosol samples from Hong Kong. The 3-hydroxy fatty acids are constituents of the endotoxins of Gram-negative bacteria. These endotoxins can have strong inﬂammatory properties. Chemically, endotoxins are lipopolysaccharides that are located in the outer membrane of the bacteria. Measured concentrations of 3-hydroxy fatty acids can be used to estimate the endotoxin concentration and, therefore, the mass load of Gram-negative bacteria in the aerosol. Ergosterol is used as a speciﬁc tracer for fungi in various matters (house dust, building materials) and bioaerosols (Pasanen et al. 1999; Saraf et al. 1997; Szponar and Larsson 2001). For Gram-positive bacteria, muramic acid serves as a marker compound. Muramic acid is part of the peptidoglycan structure of the Gram-positive bacteria cell wall (Szponar and Larsson 2001). Fatty acids with carbon atom numbers less than 20 are sometimes also used as markers for the microbial contribution to atmospheric aerosols (Guo et al. 2003; Zheng et al. 2000). However, they are not very speciﬁc, because these substances are also emitted by anthropogenic sources (Rogge et al. 1991). The occurrence of various sugars and sugar alcohols in the atmosphere is also linked to bioaerosols; see also Section 9.2.5.

348

9.2.2

Volatile Organic Compounds in the Atmosphere

EC, black carbon and soot

The term ‘black carbon’ is often used to describe carbonaceous residues produced by incomplete combustion of OM (Fernandes et al. 2003). However, the chemical composition of these residues is not very well deﬁned. Black carbon includes soot and other particulate residues of various combustion processes like biomass burning, coal and oil burning, diesel engines, etc., which can be usually identiﬁed by their black colour. All of these combustion residues are primary aerosol constituents. Black carbon is not equal to EC; however, it might be a good approximation. Black carbon can be measured by optical methods using the absorption of light. Actually, nowadays this light absorption (black colour) is the basic deﬁnition of ‘black carbon’ (Gelencser 2004). Soot is generated in the gas phase (in the ﬂame) from gaseous precursors, whereas carbonaceous residues formed on the surface of solid (biomass) fuels are called chars (charcoals). Soot is a form of appearance of EC, and, as a ﬁrst approximation, is similar to graphite, consisting of layers of hexagonally arranged carbon atoms. Unlike graphite, these layers can be curved, forming spherical bodies. If the order of these layers is very low, carbon is called amorphous; if it is high, it is called graphitic carbon (Wal and Tomasek 2004). Also, fullerenes and carbon nanotubes can be formed under speciﬁc combustion conditions (Bang et al. 2004; Height et al. 2004; Takehara et al. 2005). The size of soot particles depends on the history of their formation (combusted substances, temperature, oxygen content in ﬂame, etc.). Primary soot particles (from diesel engines or oil burning) have diameters between 10 and 50 nm and a spherical onionshell structure; however, they usually agglomerate during the combustion process, forming chain-aggregates with some 100 nm to about 1 μm length (Wentzel et al. 2003). Chars are much bigger than soot particles, having diameters usually in the micrometre range. Although soot and chars consist mostly of carbon they also contain various amounts of hydrogen atoms and other atoms, as well as some extractable OM, e.g. polycyclic aromatic hydrocarbons (PAHs) or alkanes. The OM may be adsorbed onto the soot surface or even enclosed inside the primary soot particles (Fernandes and Brooks 2003). In reality, there is no sharp border between elemental and organic carbon. Starting with an organic molecule such as a low-mass PAH (e.g. naphthalene), the addition of further aromatic rings will lead to polyaromatic systems. Systems consisting, for example, of some thousand carbon atoms might ﬁnally be better described as graphitic layers and thus might be counted as EC instead of OC.

9.2.3

OC and organic matter

EC constitutes only a small fraction of the carbonaceous matter present in atmospheric aerosols. The dominating fraction of carbonaceous aerosols is the so-called organic matter (OM), which includes all organic compounds present in the particle phase. Unfortunately, it is not possible to measure the mass of OM in aerosol samples directly. An indirect method to determine the OM is to measure the so-called OC, which represents all carbon contained in organic compounds. For the estimation of organic mass, the determined OC is multiplied by a certain value that represents the mean ratio of organic mass to OC of the organic substances present in the aerosol. Early estimations often used an OM/OC ratio of 1.4 to calculate OM from measured OC. More recent publications suggest that a value of 1.4 is more likely the lower limit. It was estimated that urban aerosols are best represented by an

Organic Aerosols

349

OM/OC ratio of 1.4 to 1.8, whereas non-urban aerosols usually have higher OM/OC ratios in the range of 1.9–2.3 (Turpin and Lim 2001). The value of the OM/OC ratio depends on the individual molecular composition of the respective aerosols. Organic molecules that consist only of hydrogen and carbon have a low OM/OC ratio, for example, alkanes or PAHs. Such compounds are usually non-polar and are insoluble or only slightly soluble in water. Highly oxygenated compounds have a higher OM/OC ratio. They are often water soluble due to the presence of polar functional groups, like dicarboxylic acids or sugars. Due to the high content of non-polar constituents in urban aerosols, for example, from fossil fuel combustion sources, their OM/OC ratio is rather low. Aerosols produced or inﬂuenced by photochemical activity are usually higher oxidised and, therefore, have a higher OM/OC ratio. The same is true for aerosols from combustion of biomass, which carry oxygen-containing constituents such as levoglucosan (see Section 9.2.4.1). Oxygen is the most important element besides carbon, adding to OM; however, also other elements, such as nitrogen, sulphur, phosphorous, and so on, contribute to OM.

9.2.3.1

The OC/EC tracer method for the estimation of SOA

SOA is formed by the chemical conversation of volatile organic precursors to low volatile compounds and their subsequent condensation. The amount of SOA cannot be measured directly, because there is no possibility to separate SOA constituents from primary organic aerosol constituents. Therefore, several methods were developed to estimate the secondary organic contribution. One possible method is the mathematical modelling of SOA formation in connection with transport and deposition of SOA (Pandis et al. 1992; Strader et al. 1999). Another method is the use of a receptor model to estimate the primary organic aerosol and calculate SOA by subtracting the estimated primary organic aerosol from the measured total organic aerosol (Schauer and Cass 2000; Schauer et al. 1996). However, the mostly used method to estimate SOA is the OC/EC tracer method, which will be described subsequently. Primary organic aerosols are often linked to combustion processes, and these processes yield certain fractions of EC and OC. The ratio of OC to EC depends on several factors; however, on average, a speciﬁc OC/EC ratio, which is characteristic for the individual combustion source, can be estimated. The same counts for other primary organic aerosol sources. A certain location will be inﬂuenced usually by the same sources of primary emissions. However, these sources may change during the year, and the impact of single sources to local aerosols can vary with meteorology, for example, wind direction. If measurements of OC and EC are carried out during a longer time period, typical OC/EC ratios for primary emissions at this location can be obtained. Of course, the measurement of OC/EC ratios typical for primary emissions has to be carried out during periods in which the inﬂuence of primary emissions is dominating. On the other hand, this means that these periods should be inﬂuenced only little by photochemistry (Cabada et al. 2004; Strader et al. 1999; Turpin and Huntzicker 1995). Periods of mainly primary emissions might be identiﬁed using gaseous combustion tracers (e.g. CO, NOx ), whereas ozone can be used as an indicator of photochemical activity. Sometimes, simply the lowest measured OC/EC ratios of a time series are accounted to primary emissions (Castro et al. 1999). Also, primary emissions inventories of OC and EC were used to estimate the primary OC/EC ratio (Gray et al. 1986).

350

Volatile Organic Compounds in the Atmosphere

6 Influenced by SOA production 5

OC (μg/m3)

4 3 2 Lowest OC/EC ratio (= 0.81)

1

Mostly primary emissions

Linear regression: y = 0.71 × +0.41

0 0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

EC (μg/m3) Figure 9.3 Plot of OC-concentrations vs EC-concentrations for measurements in Jülich, Germany.

If the measured OC/EC ratio exceeds the previous determined primary OC/EC ratio ((OC/EC)pri ), contribution of secondary OC to the organic aerosol is indicated. The difference between total OC (OCtot ) and primary OC (OCpri ) is secondary OC (OCsec ). This can be expressed by two simple equations (Turpin et al. 1991): OCpri = EC × (OC/EC)pri

(9.1)

OCsec = OCtot − OCpri

(9.2)

Usually, (OC/EC)pri is calculated as the slope of a linear regression from a plot of the measured OC vs. the measured EC, as indicated in Figure 9.3. The linear regression process normally results a certain value for the intercept, which can be explained by sources of primary OC that emit no EC (e.g. cooking or suspension of biological material). Therefore, Equation 9.1 has to be modiﬁed to (Turpin and Huntzicker 1995): OCpri = EC × (OC/EC)pri + b

(9.3)

where b is the intercept on the OC-axis. Typical values for (OC/EC)pri are in the range of 0.9 in urban areas to more than 3 in rural or remote areas (Table 9.1). Finally, SOA mass has to be calculated from the estimated secondary OC.

9.2.3.2

Concentrations of OC, EC and secondary OC in the ambient atmosphere

Typical mean concentrations of OC in the atmosphere are in the range of 1 μg/m3 in clean areas to >10 μg/m3 in polluted areas (refer to Table 9.2), whereas peak concentrations might reach >50 μg/m3 , for example, in the rainforest during the biomass-burning season (Artaxo et al. 2002). EC concentrations in clean areas are typically below 1 μg/m3 , in polluted areas the EC concentrations might exceed 5 μg/m3 . High primary OC/EC ratios point to a high input of primary OC from non-combustion sources (Table 9.1). As explained above, if

Table 9.1 Mean OC/EC ratios and OC/EC ratios for primary emissions (OC/EC prim.) measured at different locations. Calculated primary OC and secondary (sec.) OC concentration as well as the relative contribution of secondary OC to total OC (secondary OC (%)) are also displayed Location

Site description

Birmingham, UK1

Urban

Areao, Portugal1

Rural/Coastal

Cities, China†2

Oceanic Urban

United States3 , Northeast Central West Mira Loma, United States4 Helsinki, Finland5 Hyytiälä, Finland6 Jülich, Germany6

Time

January 1994 May 1993 May–August 1994 November–December 1993, February–March 1994 May–July 1993 2002 June–August 1999

Continental Rural/Urban Urban Remote Rural/Urban

September 2001–January 2002 July 2000–July 2001 March 2003 June 2002 July 2003

∗ In% of the total OC. † Average of different sites in the cities Hong Kong, Guangzhou, Shenzhen, Zhuhai. 1 Castro et al. (1999). 2 Cao et al. (2004). 3 Yu et al. (2004). 4 Na et al. (2004). 5 Viidanoja et al. (2002). 6 Warnke (2004).

PM

OC/EC

OC/EC prim.

Primary OC (μg/m3 )

Secondary OC (μg/m3 )

10 10 10 10

1.4 3.5 2.9–3.5 2.4–4.0

1.1 1.1 1.5 1.5

3.95 1.65 0.38 0.70

0.63 3.10 0.56 0.71

17 65 56 45

7.3 2.5 2.5 — — — 5.2 — 2.1 1.0 1.8

— 1.1 1.2 1.2 2.4 3.5 3.7 1.1 0.98 0.81 0.81

0.4 — — 0.39 0.47 0.51 6.2 — — — —

1.75 4.9 6.3 1.3 0.9 0.51 4.2 1.68 0.44 0.3 1.5

78 51 47 77 66 48 40 54 60 21 53

10 2.5 10 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5

Secondary OC (%)∗

352

Volatile Organic Compounds in the Atmosphere

Table 9.2 Mean concentrations of OC and EC in the atmospheric ﬁne particle phase (PM2 or PM2.5 ) measured at different locations Location

Description

Lennox, California1 Beijing, China2 Ulan-Bator, Mongolia2 Chongju, Korea3 Conroe, Texas4 Galveston, Texas4 Aldine, Texas4 Budapest, Hungary5

Urban Urban Rural Urban-city Rural Coastal Urban Kerbside (day) Urban (day)

OC (μg/m3 )

EC (μg/m3 )

6.3 12.4 2.3 5.0 2.4 3.1 4.3 6.8∗ 4.1∗

1.7 5.4 0.4 4.4 0.23 0.26 0.57 3.4∗ 0.33∗

∗ Median. 1 Turpin et al. (1991). 2 He et al. (2004b). 3 Lee and Kang (2001). 4 Russell and Allen (2004). 5 Salma et al. (2004).

the mean OC/EC ratios exceed the primary OC/EC ratios, formation of SOA is indicated. However, the measurement results of OC and EC can easily be inﬂuenced by different errors (refer to Section 9.3). Obviously, these errors also inﬂuence source attribution based on OC/EC analysis. Estimated secondary OC is in the range of below 20% to about 80% of the total OC. Its value is strongly dependent on the time of the year and on the origin of the air masses. It is less abundant, for example, in the European winter with low radiation and, therefore, low photochemical activity and simultaneously high primary emissions. Especially high secondary contributions to OC can be found during periods with strong photochemical activity (summer) and simultaneously low input of primary emissions, for example, air masses from the ocean. However, absolute concentration values of OC might be low during these periods. High absolute concentrations of secondary OC (e.g. around 5 μg/m3 in cities in China and Portugal) can be found in polluted areas with high photochemical activity. Typically, mean secondary contribution to OC seems to be around 50%, refer to Table 9.1.

9.2.3.3

C-14 measurements: A method to differentiate between biogenic and anthropogenic carbon

The measurement of the 14 C/12 C ratio of the organic fraction in aerosol samples can be used to estimate the relative contribution of anthropogenic and biogenic sources to the organic aerosol fraction (Currie 2000). 14 C works like a tracer for biogenic carbon. Due to the continuous production of radioactive 14 CO2 in the atmosphere by cosmic radiation, organic substances produced by plants have a certain 14 C/12 C ratio. Over the years, this ratio changes due to the radioactive decay of the 14 C, for example, in fossil fuels such as coal, oil or natural gas. By the measurement of the 14 C/12 C ratio in atmospheric aerosols, the contribution of sources originating from the combustion of 14 C-poor fossil fuel and the contribution of contemporary biogenic sources can be calculated. Obviously, at this point,

Organic Aerosols

353

the anthropogenic contribution to the atmospheric organic aerosol fraction is set equal to fossil fuel combustion, an assumption that is not completely valid (e.g. wood combustion – biomass burning – will be accounted for as biogenic contribution). Therefore, this technique is only effective when the combustion of biomass is not a major contributor to the regional aerosol burden. For these estimations, precise and sensitive measurements of the 14 C/12 C ratio are necessary. The measurements have to be carried out by accelerator mass spectrometry (AMS). These measurements are very complex, expensive and time consuming. Nevertheless, the results are very important since they can be directly used to differentiate between biogenic and anthropogenic contributions to the aerosol total carbon (TC). Results of these measurements indicate a substantial biogenic contribution to the organic aerosol even in urban or anthropogenic inﬂuenced regions. In south-east Texas, 27–73% of the aerosol TC collected at an urban/suburban site in summer 2000 was modern carbon (and thus biogenic). At a rural, forested site, even 44–77% of the TC was of biogenic origin (Lemire et al. 2002). In high alpine snow samples from the eastern Alps about 64% of the non-soluble carbon was of biogenic origin (Weissenbok et al. 2000). Biogenic aerosol contribution in Nashville, Tennessee and Zurich, Switzerland, ranged from 56–80% and from 51–80%, respectively (Lewis et al. 2004; Szidat et al. 2004). In a size resolved sample from Tokyo from April 2002 (Endo et al. 2004), the biogenic contribution ranged from about 40% in the ﬁne-size class (<1.1 μm) to about 60% for coarse particles (>7 μm).

9.2.4

Sources and molecular composition of organic aerosols

The organic fraction of atmospheric particulate material can contain a large number of diverse molecular species. The composition is mainly dependent on the aerosol source with possible modiﬁcations during atmospheric transport. The organic mixture ranges from non-polar hydrocarbons (alkanes) over highly polar and water soluble components, such as short dicarboxylic acids or sugars, to macromolecular organics. Therefore, it is helpful to characterise the different sources (or source types) of organic aerosols in terms of their individual chemical composition. Knowing the source compositions selected source speciﬁc compounds can be used as tracers for the origin of aerosols or to estimate the contribution of the different source types to the measured aerosol.

9.2.4.1

Sources and composition of primary organic aerosols

There is a variety of biogenic and anthropogenic sources of primary organic aerosol constituents. Source strength estimations indicate that globally the input from natural sources is prominent. However, for the aerosol composition on the local or regional scale even weak sources (like cooking or cigarette smoking) can be important (Schauer et al. 1996). The following processes constitute the most important sources of primary organic aerosols that have been characterised: • • • •

Biomass burning Fossil fuel burning Plant abrasion Suspension of soil and dust

354

Volatile Organic Compounds in the Atmosphere

R

O R

O HO

OH

OH Levoglucosan

O CH3O

OH

HO R = H, CH3–C6H13 17(H)-Hopanes

Cholesterol: R = H -Sitosterol: R = ethyl

HO Vanillic acid

Figure 9.4 Tracers utilised for petroleum use (hopanes), biomass burning (levoglucosan, vanillic acid) and cooking (sterols).

• • • •

Suspension and release of bacteria, fungi, viruses, pollen, algae, spores (e.g. fern) Bubble bursting and sea spray Cigarette smoking Cooking.

Trafﬁc is a massive source of particles from fossil fuel burning. The composition of these exhaust-particles depends on the engine-type (gasoline or diesel) and possible treatment of the exhaust (catalyst) (Rogge et al. 1993b; Schauer et al. 2002). Major compound classes are alkanes, alkanoic acids and PAH. Hopanes (see Figure 9.4) and steranes are only minor compounds; nevertheless, they can be used as speciﬁc tracers for emissions of motor vehicles or the general use of petroleum products. Biomass burning is a strong source for atmospheric aerosols, producing about four times more than fossil fuel burning (Kuhlbusch 1998). A main product and a general tracer used for biomass burning is levoglucosan (1,6-anhydro-β-d-glucopyranose; see Figure 9.4), an anhydro-sugar derived from the thermal degradation of cellulose during the combustion process (Simoneit 1999). Anhydro-sugars seem to be the most abundant compounds produced during the combustion process of plant material. Other major compound groups identiﬁed in smoke particles from biomass burning are alkanes, alkenes, alkanoic acids, di- and triterpenoids, monosaccharides, methoxyphenols and PAHs (Simoneit 2002). Some of these constituents derive from thermally altered plant material (like anhydro-sugars); others are unchanged ingredients (like some wax-alkanes). Especially, lignin pyrolysis products, such as vanillic acid, can be used as tracers for certain plant species. Plant abrasion is mainly induced by wind-driven mechanical force, like the rubbing of leaves against each other (Rogge et al. 1993a). Identiﬁed substances in aerosols from plant abrasion (green and dead leaves) are mainly constituents of the epicuticular plant waxes: n-alkanes, n-alkanals, n-alkanols, n-alkanoic acids (fatty acids). These compound groups are not very speciﬁc for biological sources but, due to their biosynthesis, speciﬁc patterns in carbon numbers of plant-derived wax components can be observed. Leaf wax alkanes have a strong odd carbon number predominance with the dominant carbon numbers C29 , C31 and C33 , whereas alkanoic acids, alkanals and alkanols have a predominance of even carbon numbers. Fossil fuel constituents are showing no predominance in carbon numbers. Therefore, it is possible to use these speciﬁc patterns to identify contributions of plants to atmospheric aerosols (Simoneit et al. 1988).

Organic Aerosols

355

Bubble bursting and sea spray are processes important for the production of primary organic aerosols in the marine environment. Surface active organic substances, such as fatty acids (Mochida et al. 2002), are enriched in the oceanic surface layer. Therefore, these substances are transferred into the atmosphere especially easily by bubble bursting and sea spray processes. This surface active organic material may be produced, for example, by phytoplankton (O’Dowd et al. 2004). Also, microorganisms may be transported into the atmosphere by bubble bursting processes. Soil contains several percent of OM (e.g. 5–6% for agricultural soils) that consists of plant litter, microbes, microbial and animal residues, lipids, carbohydrates, peptides, cellulose, lignin and humic material (Simoneit et al. 2004a). Carbohydrates account for about 5–20% of the organic material. Saccharides may be appropriate tracers for soil input into the atmosphere, in particular α- and β-glucose, inositols, sucrose and mycose (trehalose). However, there are also indications that a direct release of bioaerosols could be a main source of sugars and sugar alcohols in the atmosphere. The main constituents identiﬁed in aerosols produced by cooking are alkanoic and alkenoic acids derived from thermal hydrolysis and thermal oxidation of fat and oil. Other identiﬁed compounds are dicarboxylic acids, alkanes, sterols, levoglucosan and PAHs. The detailed composition of the aerosols depends on the food preparation method and the ingredients used (He et al. 2004a; Rogge et al. 1991). Different sterols seem to be appropriate tracers for cooking derived aerosols. While cholesterol (Figure 9.4) indicates use of meat, stigmasterol and β-sitosterol are present in plant oils.

9.2.4.2

Sources and composition of SOAs

SOAs are produced (a) by gas-phase oxidation of VOCs that can either form new particles or condense onto pre-existing particles, (b) by heterogeneous reactions on particle surfaces or (c) by in-cloud processing. Precursors of organic SOA are mostly volatile reactive biogenic (e.g. terpenes) or anthropogenic (e.g. aromatics) hydrocarbons. Products formed can be relative low volatile organics, which convert almost completely to the particle-phase, or semivolatile organics, which partition between the gas and particle phase. This gasparticle partitioning of semivolatile (and also low volatile) compounds can be described by gas-particle-partitioning models introduced by Odum et al. (1996) and Pankow (1994a, 1994b). These models consider the dependence of the concentration of an individual organic compound i in the particle phase from the available absorbing organic aerosol mass (MO ), the partitioning coefﬁcient of compound i and the concentration of i in the gas phase: caer = cgas × Kom × MO

(9.4)

where Kom , partitioning coefﬁcient of i (m3 /μg) (temperature dependent); caer , concentration of compound i in the absorbing organic particle phase (ng/m3 ); cgas , concentration of i in the gas phase (ng/m3 ); MO , concentration of the absorbing organic phase in the aerosol (μg/m3 ). Apart from problems such as source strength of precursors, chemical conversation, atmospheric transport, etc., the temperature dependence of the partitioning coefﬁcient leads to some additional uncertainty in the modelling of SOA (Takekawa et al. 2003). During the ﬁrst years of SOA-research, attention was paid to condensation of non-volatile and partitioning of semi-volatile compounds regarding the formation of particle phase.

356

Volatile Organic Compounds in the Atmosphere

Volatile products formed during the chemical conversation of aerosol precursors seemed to play no role in the formation of particulate matter. In contrast, more recent studies show that volatile carbonylic products formed in the gas-phase oxidation of organics may contribute to SOA mass by the formation of low volatile oligomers. For example, acidcatalysed reactions of aldehydes or ketones on the particle surfaces or inside the particles, like aldol reaction/condensation or acetal formation, are able to form such oligomers. These processes result in increased particle mass and a lower volatility of the particles. Compounds that are able to contribute to particle mass by oligomerisation may be formed from biogenic as well as anthropogenic precursors (Gao et al. 2004; Iinuma et al. 2004; Jang et al. 2002, 2004; Kalberer et al. 2004). There are also evidences for the direct formation of oligomeric products by heterogeneous reactions of unsaturated gas-phase compounds (e.g. isoprene) on particle surfaces (Limbeck et al. 2003). Several groups speculate that these oligomeric products formed from gaseous precursors could represent a substantial fraction of the so-called ‘humic-like substances’ (HULIS) often identiﬁed in atmospheric aerosols. HULIS is a collective term for a group of particle phase compounds, which add to the water soluble organic carbon. However, the exact chemical structures are not known. The chemical properties of the HULIS seem to be similar to humic and fulvic acids, which are known from waters, sediments and soils. Humic and fulvic acids are collective terms for molecules in a large range of molecular weights. These molecules contain multiple functional groups, e.g. carboxylic acid and carbonyl groups, and also aromatic structures. The formation of new particle-phase products from gas-phase constituents during the atmospheric lifetime of aerosols is part of the so-called atmospheric ageing of organic particles, a process that is currently not well characterised. Beside the incorporation of reactive gas-phase species into the organic aerosol fraction by oligomer formation, ageing also includes the degradation or chemical modiﬁcation of particle-phase constituents by atmospheric oxidants. Since these chemical modiﬁcations will result in alterations of the physical (volatility, light absorption, light scattering) and physico-chemical properties (water solubility, activity as cloud condensation nuclei (CCN)) of atmospheric aerosols, which are relevant for the climatic effect of the aerosols, the investigation of these processes has to be addressed in future research on organic aerosols. The incorporation of SOA formation into atmospheric models is not an easy task, since a variety of chemical and physico-chemical processes inﬂuence the SOA particle mass in the ambient atmosphere. A sensitivity analysis of SOA production and transport modelling (Tsigaridis and Kanakidou 2003) showed a factor of about 20 of uncertainty in predicting the SOA production, considering the different inﬂuences of partitioning, ageing and MO , excluding the uncertainties of precursor emissions and individual oxidation pathways. This results in a range for the annual global production of SOA from 2.55 to 47.12 Tg of OM per year. Another study yields annual SOA production of between 15.3 and 24.6 Tg/year using the partition method and bulk yield method, respectively (Lack et al. 2004).

9.2.4.3

Biogenic SOA

Precursors of biogenic SOA in the continental environment are mainly unsaturated hydrocarbons namely (mono-) terpenes, sesquiterpenes and isoprene. The SOA-forming potential of terpenes is well known and was intensively investigated (e.g. Grifﬁn et al. 1999; Hoffmann et al. 1997, 1998; Kavouras et al. 1998; O’Dowd et al. 2002; Went 1960; Yokouchi and

Organic Aerosols

357

OH

HO OH

HO

OH

OH

OH Isoprene

Methyltetrol (erythro- and threo-)

O

2,3-Dihydroxy-4-oxobutanoic acid

Figure 9.5 Isoprene and selected products from its atmospheric oxidation.

Ambe 1985; Yu et al. 1999a; Zhang et al. 1992), whereas isoprene was only most recently found to form low volatile secondary products (Claeys et al. 2004a, 2004b). Known products of the atmospheric isoprene oxidation are polyols and acidic compounds such as 2-methyltetrols and 2,3-dihydroxymethacrylic acid (Figure 9.5). It was estimated that isoprene might add about 2 teragrams of the polyols to the atmospheric SOA. This is a substantial amount, although terpenes may add 10 times more to SOA. The most frequently studied and most important SOA-forming reactions of terpenes are gas-phase oxidations by ozone, OH- and NO3 -radicals. The detailed chemical mechanisms of alkene (terpene) oxidation by atmospheric oxidants can be found in detail elsewhere in this book. Oxidation of terpenes under different conditions (ozone, OH-radicals, photosmog, etc.) generates a variety of oxygenated gas-phase (Calogirou et al. 1999) and particle-phase products, which have been identiﬁed in chamber experiments (e.g. Christoffersen et al. 1998; Glasius et al. 2000; Hoffmann et al. 1998; Jaoui and Kamens 2003a, 2003b, 2003c; Koch et al. 2000; Larsen et al. 2001; Winterhalter et al. 2003; Yu et al. 1999a). Known terpene oxidation products relevant for SOA production mainly contain carbonyl, alcohol and carboxylic acid functions. Products bearing carboxylic acid functions are low volatile and, therefore, are especially interesting for SOA-formation. Figure 9.9 shows some important products from monoterpene oxidation. Recently, it was proposed that peroxides could also represent a major part of the SOA formed by terpene ozonolysis (Bonn et al. 2004), a suggestion that was recently conﬁrmed by chamber studies (Docherty et al. 2005). As mentioned above, also oligomer formation from (semi-) volatile oxygenated terpene oxidation products might contribute to SOA formation from biogenic precursors.

9.2.4.4

Anthropogenic SOA

Of the variety of anthropogenic VOC emissions, aromatic hydrocarbons are believed to be the most important compounds for the formation of SOA. The SOA forming potentials of various aromatic VOCs under photosmog conditions have been studied (e.g. Izumi and Fukuyama 1990; Odum et al. 1996, 1997; Stern et al. 1987; Wang et al. 1992). Detailed mechanisms for the atmospheric oxidation of anthropogenic VOC are given elsewhere in this book. Many particle-phase products have been identiﬁed in these studies (e.g. Bethel et al. 2000; Edney et al. 2001; Fisseha et al. 2004; Forstner et al. 1997a; Jang and Kamens 2001; Kleindienst et al. 2004; Smith et al. 1998, 1999), but only a few of these compounds could be measured in atmospheric aerosols. The products of the oxidation of aromatic hydrocarbons can be divided into three groups: (1) ring retaining aromatic, (2) ring retaining non-aromatic and (3) ring degradation products (Jang and Kamens 2001). Ring retaining

358

Volatile Organic Compounds in the Atmosphere

aromatic products can origin from the OH-addition to the aromatic ring that leads to the formation of OH- and NO2 -substituted rings or from the H-atom abstraction of alkyl substituents (toluene, xylenes, etc.) forming, for example, aromatic aldehydes or carboxylic acids. Ring retaining aromatic products are rather stable whereas ring retaining nonaromatic products containing double bonds can be easily further oxidised by ozone or nitrate radicals due to their high reactivity (Atkinson 2000). They ﬁnally form small multifunctional acids or carbonyls under atmospheric conditions. Therefore, the unsaturated products should be present only at very low concentration levels in atmospheric aerosols in contrast to the partly relative high yields determined in chamber studies. The often-high NOx levels in chamber studies compared to atmospheric conditions further favour the formation of aromatic nitro compounds because of the competitive reaction of NO2 and O2 with the intermediate benzyl radicals (Atkinson and Aschmann 1994; Bethel et al. 2000; Kleindienst et al. 2004). Nevertheless, nitrophenols as aromatic photooxidation products can be found in atmospheric aerosols, though they may also have primary sources such as automobile exhaust (Cecinato et al. 2005; Tremp et al. 1993). Most particle-phase constituents that derive from aromatic hydrocarbon oxidation in the atmosphere (Fisseha et al. 2004; Jang and Kamens 2001; Kleindienst et al. 2004) are ring degradation products: low volatile short carboxylic and dicarboxylic acids, such as succinic, pyruvic, malonic, maleic, methylmaleic, malic, glyoxylic and oxalic acid (Falkovich et al. 2005; Kawamura and Kasukabe 1996; Kawamura et al. 2005; Kerminen et al. 2000; Sempere and Kawamura 2003; Yao et al. 2004), see Figure 9.6 for structures. However, these substances may also be secondary products from other precursors and they can also be emitted by primary sources like automobile exhaust (Kawamura and Kaplan 1987) and biomass burning (Falkovich et al. 2005). Low molecular weight dicarboxylic acids are also suggested to derive from the degradation of higher dicarboxylic acids on particles (Kawamura and Ikushima 1993) and liquid-phase reactions like in-cloud processing (Crahan et al. 2004; Warneck 2003; Yu et al. 2005a). Oxalic acid is frequently found to be the most abundant dicarboxylic acid in aerosols, but also opposite observations exist (Yu et al. 2005b). Since oxalic acid seems to be the last step in the degradation pathway of a variety of gas and particle-phase constituents, it is not surprising that its concentration often dominates the composition of the aerosol particles. Figure 9.7 shows some possible pathways for the formation of oxalic acid from different sources. Also, malonic and succinic acid, which are usually second and third in particle-phase concentrations of dicarboxylic acids, have different primary and secondary sources. Malic acid is sometimes found to be more abundant than succinic acid. More reactive species such as glyoxylic, pyruvic and maleic acid are usually present at lower concentrations and may be seen as intermediate products of the degradation of different substances, including aromatic hydrocarbons. It would be favourable to identify reliable marker substances for the formation of secondary aerosols from aromatic precursors. As outlined above, oxalic acid as well as other shortchain diacids and ketoacids cannot be used as speciﬁc tracers for anthropogenic SOA because of their various other possible sources. Recent investigations (Kleindienst et al. 2004) found that 2,3-dihydroxy-4-oxobutanoic acid and 2,3-dihydroxy-4-oxopentanoic acid might be suitable markers for SOA from toluene or related aromatic compounds. Phthalic acid has been found to derive from degradation of polycyclic aromatic hydrocarbons (Jang and McDow 1997; Kawamura and Ikushima 1993). Measurements of phthalic acid concentrations, its concentration in different particle size ranges (Fine et al. 2004) and comparison with other diacids present in atmospheric aerosol particles (Fraser et al. 2003) conﬁrm the

Organic Aerosols

359

COOH HOOC — COOH Oxalic aicd

HOOC

COOH

HOOC

Malonic acid O

Succinic acid OH

O

COOH

C HC — COOH Glyoxylic acid (wC2)

HOOC

COOH

HOOC Malic acid

Ketomalonic acid (kC3) COOH

O H 3C

O COOH

COOH

C

COOH Fumaric acid

Pyruvic acid

COOH

O HC

4-Oxobutanoic acid (wC4)

COOH

COOH

Phthalic acid

HC

HOOC

COOH

Maleic acid

Methylmaleic acid

O CH

Glyoxal

COOH

H3C

HOC

COOH

3-Oxopropanoic acid (wC3)

H3C

O

O

C

CH

Methylglyoxal

Figure 9.6 Dicarboxylic acids, oxocarboxylic acids and other oxygenated molecular species which can be found in the atmospheric particle phase. Compounds are either emitted directly into the atmosphere or are produced by chemical processes within the atmosphere.

suitability of phthalic acid as a speciﬁc tracer for anthropogenic SOA. Although in principle other compounds also seem to be suitable for tracing anthropogenic secondary aerosol formation, such as benzoic acid and substituted benzoic acids, they are relatively volatile and consequently a substantial portion of them exist in the gas phase.

9.2.5

Atmospheric concentrations of primary and secondary organic aerosols constituents

Until now, the molecular identity of the largest part of OM found in atmospheric aerosols remains unknown. This is illustrated by Figure 9.8 (Kubatova et al. 2002) showing the identiﬁed fraction vs. the unidentiﬁed in the extractable part of OM. In the last column, the most abundant identiﬁed classes of organics are displayed. Different acids (mostly fatty acids) dominate the identiﬁed fraction, but also alkanes are found as important constituents. The composition of the aerosol shown in Figure 9.8 is one possible example. Generally, the composition of aerosol particles is dependent by their sources as well as by the speciﬁc history of the investigated air mass. The concentrations of alkanes are relatively high in urban areas, typically around 100 ng/m3 . Under special conditions concentrations might be considerably higher, for example, up to several thousand ng/m3 in Kuala Lumpur during a pollution event

360

Volatile Organic Compounds in the Atmosphere

Aromatic HCs (toluene, benzene, etc.)

Pyruvic

Glyoxal methylglyoxal

Unsaturated fatty acids

Maleic methylmaleic

Malic

kC3

C2 Oxalic

Fumaric Monocarboxylic

wC3

wC2

C3 Malonic

n-Alkanes aldehydes

wC4 Mid-chain ketoand hydroxyacids

C4 Succinic

Primary sources (combustion processes)

Figure 9.7 Primary and secondary sources for oxalic, malonic and succinic acid. Intermediate and rather reactive products are in broken lines. Abbreviations are explained in Figure 9.6. Modiﬁed ﬁgure after Kawamura and Kasukabe (1996).

caused by air masses from biomass burning in Indonesia (Table 9.3). Usually, the use of fossil fuel is the main source of alkanes in urban areas. In more natural environments, concentrations of alkanes are considerably lower, especially in remote marine areas, where concentrations may be below 1 ng/m3 . In aerosols from remote marine areas, the input of alkanes from marine sources, which are weak in comparison to terrestrial or anthropogenic sources, can be substantial (Sicre and Peltzer 2004). Aldehyde and alcohol concentrations show a similar behaviour as the alkane concentrations. Aldehyde concentrations range from 0.01 in marine to 7 ng/m3 in urban environments; alcohols are also less abundant in remote marine areas (below 1 ng/m3 ) have higher concentrations in urban areas (up to about 100 ng/m3 ) and peak during, pollution events from biomass burning. For example, more than 700 ng/m3 were measured in Kuala Lumpur during pollution events due to air masses transported from tropical forest ﬁres in Indonesia (Fang et al. 1999). Alkanoic acids show concentrations around 10 ng/m3 in remote marine aerosols, from 10 to 100 ng/m3 in terrestrial remote and rural areas and above 100 ng/m3 in urban areas that peak during biomass burning events at several thousand ng/m3 (max. 14 000 ng/m3 ). PAHs are typical for combustion aerosols from fossil fuel especially coal. Concentrations are well below 1 ng/m3 in remote marine environment growing with anthropogenic inﬂuence reaching concentrations of almost 100 ng/m3 in polluted urban/industrial areas.

Organic Aerosols

361

100 EC

Unident.

Percentage

80 Unextr. and/or unelut. OM

60

40

20

Unres. (UCM)

DTs

Acids

Alkanes

OM

Identif.

Fatty Ac. (palm., stearic, oleic)

0.66 μg/m3

0.44 μg/m3

Bases + neutrals Carbon. EEOM

Others Ox. deg. Lign. pyr. PAHs

Resolv.

0 54 μg/m3

17.8 μg/m3

14.5 μg/m3

2.70 μg/m3

0.66 μg/m3

Figure 9.8 Percentage contributions of different carbonaceous and organic compound classes in winter 1998 ‘total’ ﬁlter samples from Gent, Belgium (based on 22 samples from each season). The ﬁrst column represents the contributions of the carbonaceous (OM + EC) and the inorganic aerosol to the particulate mass (PM). The atmospheric concentrations under the columns indicate the average concentration for the sum of the species in the column. From Kubatova et al., 2002 with permission. Abbreviations: EEOM = extractable and elutable organic matter, UCM = unresolved complex mixture, DTs = diterpenoic acids, Lign. pyr. = lingnin pyrolysis products, Ox. deg. = oxidative degradation products.

While these compounds represent mainly the lipid constituents, levoglucosan, sugars and short-chain dicarboxylic acids are major components representing water-soluble organic compounds (WSOC – water-soluble organic carbon) in the atmospheric particle phase. Table 9.4 shows atmospheric concentrations of selected sugars and levoglucosan in different regions. Levoglucosan is a degradation product of cellulose and almost exclusively produced by combustion of plant material. Therefore, it is not surprising that concentrations are low in remote areas (around 1 to about 10 ng/m3 ), higher in urban areas due to the use of wood as fuel (roughly 100 to >1 000 ng/m3 ) and reach highest concentrations in aerosols from massive biomass burning in tropical areas (around 1 000 to >10 000 ng/m3 ). Different sugars also show a wide range of concentrations from less than 1 ng/m3 in marine aerosols to around 1 000 ng/m3 in aerosol samples from Chile. Usually, concentrations of glucose, sucrose and mycose are lower than 100 ng/m3 . The source of sugars is believed to be mainly soil dust including suspended microorganisms (Simoneit et al. 2004a). Other observations in the Amazonian rainforest point to contributions from living plants. Mycose, arabitol and mannitol are well-known constituents of fungal spores, and sucrose, glucose and fructose are known to be present in pollen grains. Consequently, a strong dependence of the sugar concentrations from the daytime was observed with higher glucose, sucrose and fructose concentrations during daytime and higher levels of mycose (trehalose), arabitol and mannitol during the night when a strong release of fungal spores occurred (Graham et al. 2003a).

362

Volatile Organic Compounds in the Atmosphere

Table 9.3 Concentrations (ranges or mean) of lipid compounds measured in aerosol samples from different sites. Concentrations of compound classes sometimes include individual substances with different chain length of carbon atoms. All concentrations were obtained from TSP samples (= total suspended particulate matter), except those reported by Brown et al. (2002). Values are rounded off for better readability Location

South Paciﬁc1 Cichi-jima Island2 Crete, Greece3 Texas4 Petrana, Greece5 Gosnan, Jeju Island, Korea2 Heraklion, Greece3 Kuala Lumpur, Malaysia6 Qingdao, China7 Kuala Lumpur, Malaysia8 Gent, Belgium9

Concentration (ng/m3 )

Site description Alkanes

PAHs

Fatty acids

Remote/Marine Remote/Marine Marine Remote/Continental Rural Rural/Marine

0.02–0.54 0.3–7.0 6.5–26 1.2–15 27 9.6–150

— 0.0–0.08 0.1–2.5 — 0.52 0.05–7.7

— 3.2–7.6 1.0–20 7.0–32 — 23–96

0.05–0.77 4.6–22 3–17 — — 5.0–130

Urban Urban/BMB Urban Urban Urban

65–320 290–9 300 220 26–340 77

20–60 7–46 88 — 11

110–200 320–14 000 650 4 000–7 400 140

17–32 31–740 — 160–720 —

n-Alkanols

1 Sicre and Peltzer (2004). 2 Simoneit et al. (2004b).

3 Kavouras and Stephanou (2002). 4 Brown et al. (2002). 5 Kalaitzoglou et al. (2004). 6 Fang et al. (1999). 7 Guo et al. (2003). 8 BinAbas and Simoneit (1996). 9 Kubatova et al. (2002).

Levoglucosan, sugars and the lipid constituents represent the primary fraction of organic aerosols, whereas short-chain dicarboxylic acids have primary and secondary sources. Oxocarboxylic acids derive mostly from secondary processes, either the oxidation of VOCs or the further oxidation of carbonyls, mono- or dicarboxylic acids. Oxalic acid (and other di- and oxocarboxylic acids) levels are rather low in marine and remote areas, although a secondary production can even be observed in remote arctic regions (Kawamura et al. 2005). Concentrations of oxalic acid are roughly on the order of around 100 ng/m3 in natural or remote regions, as shown in Table 9.5. Urban concentrations are on the order of a few hundred ng/m3 and highest concentrations are observed in biomass burning aerosols (>1 000 ng/m3 ). The same trend can be observed for malonic and succinic acid although concentrations are substantially lower. Pyruvic, glyoxylic and maleic acid are known products of the degradation of aromatic VOCs. Their concentrations are usually lower in remote regions and higher in polluted air masses. The main source of phthalic acid is probably the degradation of (polycyclic) aromatic hydrocarbons as explained above. Products from the oxidation of biogenic VOC (e.g. terpenes) are important contributors to SOAs. Mainly forested regions are inﬂuenced by this group of compounds. The acidic

Organic Aerosols

363

Table 9.4 Concentration of some sugars and levoglucosan in the atmospheric particle phase at different locations. Values are rounded off for better readability Location

Site description

PMa

Gosnan, Jeju Island, Marine/Continental TSP Korea1 North Paciﬁc1 Marine/Remote TSP Rain-forest/Remote Fine Amazonia, Brazil2 Coarse Rondonia, Brazil3 Rain-forest/BMBb Fine Urban TSP Santiago, Chile4 Kuala-Lumpur, Urban/BMB TSP Malaysia4

Concentration (ng/m3 ) Glucose

Sucrose

Mycose

Levoglucosan

11–110

6–440

2.5–30

8–74

0–0.2 4.9–12 21–90 5–18 8–1 700 —

0.2–1.3 7.7–33 <0.04–4.8 1 200–6 900 12–2 500 1 200–33 000

0–4 0–1.8 1.9–16 <0.06–1.9 1.7–120 <0.06–77 14–62 0.8–26 10–2 200 15–3 100 — —

a Sampled particle size range: TSP: total suspended particulate matter, ﬁne: <2.5 μm, coarse: >2.5 μm. b BMB: Biomass burning – aerosols from forest/vegetation ﬁres. 1 Simoneit et al. (2004b). 2 Graham et al. (2003a). 3 Graham et al. (2002). 4 Simoneit et al. (2004a).

products of monoterpene oxidation add to the WSOC and, therefore, they might be important for the formation of CCN. Moreover, the oxidation of terpenes has been observed to be linked to the formation of new particles above forests (O’Dowd et al. 2002). Although SOA from monoterpene oxidation seems to be very important for the global SOA-budget (Chung and Seinfeld 2002; Tsigaridis and Kanakidou 2003), there exist only few measurements of particle-phase constituents derived from monoterpene oxidation. Mostly low volatile products of α- and β-pinene, namely pinic and pinonic acid (see Figure 9.9 for molecular structures), have been measured. Table 9.6 shows a selection of reported concentrations. Measurements of ambient concentrations of products from other important monoterpenes, such as from limonene, 3-carene or sabinene, are even less frequent (see Table 9.7). The concentrations of these terpene oxidation products in the particle phase vary strongly, depending on location, weather conditions, daytime and season. Pinic and pinonic acid concentrations can range from below 1 ng/m3 to about 100 ng/m3 even at the same location. Caric and caronic acid (3-carene products), sabinic acid (sabinene product), ketolimonic and ketolimononic acid (limonene products) are less abundant. However, the relative contribution is strongly dependent on the sampling site. For example, in Finland, higher concentrations of carene products can be found due to the higher carene emissions of conifers in the boreal forest. In Germany, broad-leafed trees are more prevalent, which release high amounts of sabinene into the atmosphere (Dindorf et al. 2005; Hakola et al. 1998). Therefore, in the latter region, higher concentrations of sabinic acid can be found. In winter, when broad-leaved trees show no biological activity, sabinic acid can not even be detected in the Finish forests aerosols. These observations show that terpene oxidation products cannot only be used as biogenic SOA markers but may be also used to differentiate SOA contributions from different plant species, that is, different source regions.

5 Graham et al. (2003a). 6 Graham et al. (2002).

1 Limbeck and Puxbaum (1999). 2 Mochida et al. (2003). 3 Kawamura and Yasui (2005). 4 Kawamura et al. (2005).

Vienna, Austria, urban1 NW-Paciﬁc, marine2 Tokyo, Japan, urban3 Canada, arctic4 Amazonia, natural rainforest5 Amazonia, rainforest (burning season)6

Location and description

340 3.6–430 89–820 6.5–59 8.8–150 51–690

Oxalic acid 240 0.1–53 16–160 1.2–20 4.4–48 7.4–150

Malonic acid 117 0.1–37 13–170 1.6–19 2.1–8 3.5–76

Succinic acid 22 — 5.3–105 <0.01–39 — 1.6–66

Glyoxylic acid 63 — 6.6–140 0.37–2.8 — 1.2–28

Pyruvic acid

Concentration (ng/m3 )

— 0.1–2.5 2.5–31 0.07–1.6 0.19–0.55 0.6–13

Maleic acid

— — 0–99 <0.01–6.5 4.2–24 12–150

Malic acid

18 0.2–20 5–110 <0.01–11 0.24–1.1 1.3–31

Phthalic acid

Table 9.5 Concentrations of different acids in the atmospheric particle phase sampled in distinct areas. Values are mean concentrations or concentration ranges. Sometimes, different size classes were sampled. Values are rounded off for better readability

Organic Aerosols

-Pinene

-Pinene

365

3-Carene

Sabinene

Limonene OH

O COOH

COOH COOH

Pinic acid

O CHO

Pinonaldehyde

Pinonic acid

COOH

COOH COOH

O COOH

Caric acid

O COOH

10-Hydroxypinonic acid COOH COOH

O COOH

COOH

Caronic acid

Sabinic acid

O O Ketolimonic acid Ketolimononic acid

Figure 9.9 Important monoterpenes and some of their products from atmospheric oxidation. Table 9.6 Concentrations (range or mean) of low volatile products from the oxidation of α- and β-pinene in ambient aerosols Location

Concentration (ng/m3 )

Time Pinic acid

Pinonic acid

Norpinonic acid

7.1–98a 1.6–43b 0.13–0.39 1–25.7 —

0.14–38a 0–14b 0.04–0.24 — —

Tabua, Portugal1

August 1996

0.39–83

Nova Scotia, Canada2 Pertouli, Greece1 Amazonia, Brazil3

July 1996 August 1997 July 2001

0.48–0.59 0.4–4.4 1.1

a cis-isomer. b trans-isomer.

1 Kavouras et al. (1999). 2 Yu et al. (1999b). 3 Graham et al. (2003a).

9.3

Analysis of organic aerosols

In the following section, the chemical analysis of organic aerosol constituents will be described. This includes sampling, sample preparation and analysis. It depends on both,

n.m. = not measured. a keto-limonic acid. b keto-limononic acid. 1 Römpp (2003). 2 Warnke et al. (2006).

Jülich, Germany2

Hyytiälä, Finland1 Hohen-preissenberg, Germany1 Hyytiälä, Finland2

Location

August 2001 May 2002 August 2001 March 2003 June/July 2002 July 2003

Time

3–36 0.5–4.5 0.38–4.7 1.1–21 0.43–3.8 0.94–12

Pinic acid 3–33 3–50 0.91–8.2 2–74 1.1–5.7 1.4–79

Pinonic acid 0.05–31 0.03–0.25 0.14–5.1 0.25–8.2 0–1.5 0–2.7

Caric acid 0.05–15 — 0.55–4.3 0.26–23 0–0.62 0.14–7.7

Caronic acid

Concentration (ng/m3 )

0.05–3 0.05–0.85 0.09–1.7 n.d. 0–1.4 0–5.1

Sabinic acid

Table 9.7 Concentrations of low volatile products from the oxidation of different terpenes in ambient aerosol samples

0.2–6.3a 0.05–7.2a n.m. n.m. n.m. 0–3.6b

Limonene product

Organic Aerosols

367

the sample preparation and the method of analysis, which constituents can be measured and sometimes it depends even on the sampling (e.g. sampling substrates). Because of their different chemical properties, it might be necessary for the determination of different compound classes to use different analytical procedures. It is important to know that it depends on the analytical procedure as to which organic compounds can be detected and quantiﬁed. For example, the ﬁrst step of an analytical procedure is often the extraction of ﬁlter samples. Depending on the solvent, different compound classes will be extracted or they will be extracted with different efﬁciencies. According to this, extractable organic compounds are often divided into water soluble (e.g. WSOC) and solvent extractable (e.g. solvent extractable organic matter – SEOM).

9.3.1

Sampling

For chemical analysis of airborne particles, they have to be separated from the surrounding gas phase (the atmosphere). Often, this separation step is realised by particle collection, which at the same time represents an enrichment step of the particulate material. Basic methods for particle collection are ﬁlter sampling and the use of impactors. Fibre ﬁlters, such as quartz ﬁbre ﬁlters, generally collect particles of all sizes, whereas impactors can be used to collect different size fractions. For the characterisation of organic aerosol components, quartz or Teﬂon ﬁlters are often used. For a comprehensive description of physical aspects of aerosol sampling, measurement, etc., refer to Hinds (1999).

9.3.1.1

Impactors

Impactors, for example, the Berner impactor (Berner and Luerzer 1980), use the mass inertia of particles for their separation. The aerosol is accelerated by passing it through deﬁned openings (nozzles). Usually a vacuum is applied at the ‘end’ of the impactor to draw the air inside. Aerosol particles are accelerated together with the air. This jet is guided against a wall, forced to change its direction rapidly (90◦ ). Aerosol particles, which have a high inertia (big particles) cannot follow the air stream and will collide with the wall where they will (ideally) adhere. Particles with a low inertia (small particles) are able to change direction together with the airﬂow and will be transported away. As a result, the smaller particles are separated from the larger ones. Changing the velocity of the airﬂow (usually by reducing the diameter of a nozzle) enables to vary the size of particles being impacted. This is usually done using a series of impactors combined to a cascade impactor (Figure 9.10). Each impactor within this cascade represents one ‘stage’. At the ﬁrst stage, particles with a relative large aerodynamic diameter will be impacted, the aerodynamic diameter of impacted particles will then decrease at each following stage. At the end of the cascade, a ﬁlter can be installed to collect the ﬁnest particles which have passed the last stage. Ideally, all particles larger than a certain aerodynamic diameter will be impacted on one stage, and all smaller particles will pass to the next stage. This certain size is called, for example, cutoff, cutoff size or cutoff diameter. In reality, always some particles larger than the cutoff diameter will enter the next stage, and some particles smaller than the cutoff diameter will be impacted. Therefore, the cutoff for real systems is the aerodynamic diameter with a collection efﬁciency of 50%.

368

Volatile Organic Compounds in the Atmosphere

Aerosol inlet

Impactor stages Nozzle Streamlines

Impaction plate Critical orifice Pump Figure 9.10 Scheme of a Battelle-type impactor. Here, four stages are shown; in reality, usually six stages are used. See text for explanation.

Particles are usually impacted on special substrates, such as foils of aluminium or polymers. These will be removed after a certain period of time and can be analysed afterwards. For organic aerosol analysis, sampling times are between some hours and several days. Impactor sampling suffers from different sampling errors, for example, wall losses and ‘bounce off ’. Particles may bounce from the collection substrate instead of getting stuck if their velocity is high and they are relatively solid. If solid particles hit the impaction surface, only part of their kinetic energy will be used for plastic deformation, and a considerable amount of the energy will be converted elastically to kinetic energy of rebound. To prevent bounce off, collection substrates can be coated with different oils or greases. However, these coatings can heavily interfere with chemical analysis, especially for organic aerosol analysis. In a virtual impactor, the accelerated air is split in two parts, a fast one, which is the main air stream, and a slow one. The fast main stream is forced to change its direction, whereas the slower minor part of the air does not. Smaller particles can follow the main stream of air that was forced to change the direction. The larger particles will continue to move in the previous direction together with the smaller part of the air ﬂow. Particles of both airstreams can be collected on ﬁlters. Virtual impactors have usually only one stage. The beneﬁt of a virtual impactor is the use of conventional ﬁbre ﬁlters that do not suffer from bounce off like the substrates for impaction in normal impactors or interferences of chemical analysis from substrate coating. A special impactor is the so-called steam jet aerosol collector (SJAC). It is not designed for the separation of different-sized particles but for the efﬁcient collection of all particles. After passing a denuder to remove gas-phase constituents, particles grow in supersaturated water vapour to droplets which are afterwards impacted in a cyclone. A cyclone is a special design

Organic Aerosols

369

of an impactor. Here, the aerosol is injected tangential into a cylindrical hole. Due to the geometry, the aerosol begins to rotate around before it leaves the cyclone through an opening at the upper end. Because of the centrifugal forces caused by the rotation larger particles collide with the wall and will be removed from the aerosol. The liquid from the collected droplets in the SJAC can be continuously removed from the lower part of the cyclone. Cyclones have several applications in science and technology (they are also used in some vacuum cleaners). In the SJAC, impacted droplets are much larger than the original aerosol particles. Due to the much larger diameter of the droplets, the SJAC provides high collection efﬁciency and additionally prevents negative artefacts due to the evaporation of semivolatile constituents. This system can be coupled online to, for example, ion chromatography (IC) for the determination of nitrate (Slanina et al. 2001).

9.3.1.2

Preseparators

Impactors with only one stage (also cyclones) are often used to separate coarse particles from the ﬁne particle fraction before ﬁlter sampling. They are called preseparators if used this way, making it possible to sample the smaller particles with diameters below the cutoff size of the preseparator. Preseparators are commercially available for different size classes. There exist different standards cutoff sizes for air monitoring: 1, 2.5 and 10 μm. These size classes are called PM1 , PM2.5 and PM10 . Here, PM means particulate matter and the number in lower case shows the cutoff size. Preseparators are commonly used in ﬁlter-sampling assemblies.

9.3.1.3

Filters

The most commonly method to collect particles is ﬁlter sampling. There are mainly two types of ﬁlters used: ﬁbre ﬁlters (glass, quartz, cellulose, polymers) and porous membrane ﬁlters (cellulose esters and polymers, for example, Teﬂon®). The mechanism of particle ﬁlter sampling from aerosols is not simple and differs from the commonly assumed ‘sieving’ of particles, as it appears in liquid ﬁltration. In aerosol particle sampling, particles much smaller than the pore size of the ﬁlters are retained. There are several mechanisms that take part in deposition of particles onto ﬁbre ﬁlters: • • • • •

Gravitational settling Impaction Interception Diffusion Electrostatic attraction

Gravitational settling is only important for particles larger than a few micrometres. Impaction onto a ﬁbre works like the mechanism in an impactor and is important for particles larger than a few hundred nanometres. Interception occurs if a particle follows a streamline of air passing a ﬁbre. Due to its size, the particle gets contact to the ﬁbre and adheres. Interception is an important collection process for particles slightly smaller than those lost to the ﬁbre by impaction. Diffusion is the main process for the collection of the smallest particles. It becomes the most important for particle diameters smaller than 100 nm, which have a considerable coefﬁcient of diffusion compared to larger particles. The Brownian motion forces small particles to hit a ﬁbre by chance.

370

Volatile Organic Compounds in the Atmosphere

Finally, if ﬁlters or particles are charged, electrostatic attraction can be a very efﬁcient way for particle deposition. It is used in technical applications, for example, for cleaning of exhaust gases. Although porous membrane ﬁlters have a different constitution compared to ﬁbre ﬁlters, the mechanisms of particle deposition are similar. As mentioned above, for the chemical analysis of organic compounds in aerosol particles, quartz ﬁbre and Teﬂon ﬁlters are most widely used. These materials are chemically inert and do not interfere in the chemical analysis of trace compounds. Quartz can be cleaned very easily by heating up to 700◦ C, a treatment that removes all organic contaminants.

9.3.2

Sampling artefacts

Sampling artefacts are errors that occur during sampling due to physical or chemical processes on the sampling substrate (ﬁlters, impactor foils). Sampling artefacts can result in a lower (negative artefact) or higher (positive artefact) mass determined for certain components. Physical artefacts can either be based on the volatilisation of compounds sampled, which would result in a negative artefact, or on adsorption of gas-phase species on active surface sites of the sampling substrate, which would result in a positive artefact. Chemical artefacts base on the formation or loss of compounds by chemical reactions on the sampling substrate, for example, caused by oxidants. Artefacts due to adsorption and evaporation of organics have been investigated in a series of studies (Kirchstetter et al. 2000, 2001; Mader and Pankow 2001; Mader et al. 2003; Muller et al. 2004; Subramanian et al. 2004; Turpin et al. 2000). If particles and the surrounding gas phase are in equilibrium, volatile components will not evaporate if sampled. However, if the equilibrium is disturbed by changing pressure, temperature or gas-phase concentration, for example, which will regularly happen during longer sampling periods, components may evaporate (or condense). Inside ﬁlter systems or impactors, often a pressure drop occurs that can cause evaporative losses. The absorption of gaseous compounds is known to be relevant, especially for quartz ﬁbre ﬁlters. The active surface is able to adsorb compounds in a quantity that might be in the order of the particulate mass. In contrast, Teﬂon ﬁlters do not suffer from adsorption of gaseous compounds because of their non-active surface characteristics. Unfortunately, Teﬂon ﬁlters are not suited in combination with certain analytical methods, for example, OC-measurements, due to their limited temperature resistance. To avoid artefacts due to adsorption or evaporation a variety of sampling techniques have been developed. Most research of sampling artefacts concerned sampling for OC/ECmeasurements, not individual components. Sampling errors without correction might range from a positive artefact of 55% to a negative of up to 80% for OC-measurements, depending on sample composition and sampling conditions. Unfortunately, frequently, the attempt to avoid errors produces new ones. If, for example, the gas phase is removed by a denuder to prevent positive artefacts by adsorption of gaseous compounds, the equilibrium of gas and particle phase is disturbed and semivolatile compounds begin to evaporate, creating a massive negative artefact. A rough estimate of the positive artefact can be done by placing a backup quartz ﬁlter behind the particle collecting quartz ﬁlter (front ﬁlter). The measured OC or component mass on the backup ﬁlter can be accounted as adsorbed gas phase and subtracted from the determined mass of the front ﬁlter for correction. Nevertheless, it might also occur that particle-phase compounds evaporate from the front ﬁlter and

Organic Aerosols

371

condense onto the backup ﬁlter. Then, a negative artefact would be additionally subtracted, making the error even bigger. Newer attempts to eliminate artefacts use combinations of different ﬁlters and denuders in different sampling lines for measurement of the different artefacts (Mader et al. 2003; Subramanian et al. 2004; Turpin et al. 2000).

9.3.3 9.3.3.1

Analytical methods for characterisation of particulate organics Sample treatment

The procedure following the particle collection step depends on the analysis that has to be carried out. The measurement of OC/EC by thermal methods can usually be done without any further treatment of the ﬁlter/impactor substrate. For most other analytical techniques, particulate matter has to be extracted from the collection substrate. The selection of the solvent depends on the chemical characteristics of the analytes of interest. It might be necessary to use several extracting steps using different solvents if the substances of interest cover a wide polarity range. Commonly used solvents are dichloromethane, methanol, water and also mixtures or additives, for example, acids to adjust the pH. Repeated ultrasonic agitation or the soxhlet extraction is applied to optimise the extraction efﬁciency. Another possibility is supercritical ﬂuid extraction (SFE) using CO2 ; however, this method is not widely used for atmospheric aerosol analysis (Forstner et al. 1997b; Shimmo et al. 2004). The further treatment of the extracts likewise depends on the compounds of interest and the selected analytical method. If necessary, solvents can be changed after the extraction step by drying the extracts and resolve them in a suitable solvent. However, relatively volatile analytes might be lost during this treatment by evaporation. The evaporation of the solvent can also be used as enrichment after the extraction process. For gas chromatography (GC) measurements, several enrichment, clean-up and derivatisation methods are available. Derivatisation is required if the components of interest are too little volatile for GC analysis. Derivatisation is sometimes also used for an improvement of the sensitivity. High performance liquid chromatography (HPLC) and IC analysis usually do not require derivatisation. However, the analytes have to be solved in suitable solvents. Extracted components of interest might cover such a wide range of functional groups, polarity or volatility that different procedures have to be applied for sample preparation prior GC analysis. Some authors divide the extracted samples into groups of different polarity by ﬂash chromatography (Alves et al. 2002; Gogou et al. 1998; Stephanou and Stratigakis 1993; Zheng et al. 1997) or thin-layer chromatography (TLC) (Binabas et al. 1995). These fractions are then treated differently according to their chemical properties.

9.3.3.2

Separation and detection methods

Gas chromatography. Gas chromatography–mass spectrometry (GC–MS) is the most widely used method for the separation, identiﬁcation and quantiﬁcation of organic compounds in aerosol samples. It is a powerful, sensitive and very selective analytical method. Occasionally, ﬂame ionisation detectors (FID) are still used for detection (Cheng and Li 2005). They produce lower costs, have a better dynamic range and are often better for quantiﬁcation than MS-detectors; however, they show a poor selectivity. On the other hand, many compound classes present in atmospheric aerosol particles are not accessible

372

Volatile Organic Compounds in the Atmosphere

by a direct GC analysis. Carboxylic acids, low volatile carbonyls, sugars and other polar substances have to be derivatised prior to analysis. There exist a variety of derivatisation techniques. Alcohols, sugars and anhydrosugars can be derivatised to trimethylsilyl ethers by, for example, N , O-bis-(trimethylsilyl)triﬂuoroacetamide (BSTFA) (Simoneit et al. 1999) or N -Methyl-N -trimethylsilyltriﬂuoroacetamide (MSTFA) (Zdrahal et al. 2002). Carbonyl groups can be derivatised using, for example, o-(2,3,4,5,6-pentaﬂuorobenzyl) hydroxy amine (PFBHA) (Yu et al. 1998). Acids can also be treated with BSTFA-forming trimethylsilyl esters (Yu et al. 1998) or with BF3 -methanol to yield methyl esters (Binabas et al. 1995). Diazomethane is a long-known, universal and often-used methylation reagent that is still in use (Kubatova et al. 2000). Liquid chromatography. Besides GC, liquid chromatography (LC) is going to become a widely used method for the analysis of individual compounds in organic aerosols (Anttila et al. 2005; Glasius et al. 1999; Hoffmann et al. 1998; Lee 1995; Römpp 2003; Warnke et al. 2006). LC is especially appropriate for the separation of high-molecular-mass or very polar molecules without derivatisation. The coupling to mass spectrometers is well established, and electrospray ionisation (ESI) is the ionisation method most widely used for organic aerosol characterisation. Atmospheric pressure chemical ionisation (APCI) is also a standard ionisation method, which is particularly suited for more polar and lower molecular weight substances. Ion chromatography. IC is usually used for the separation of inorganic ions. In organic aerosol analysis, it is applied for the analysis of very acidic, short-chain carboxylic or dicarboxylic acids, such as oxalic acid and glyoxylic acid (Jaffrezo et al. 1998; Kerminen et al. 1999; Rohrl and Lammel 2001). Mass spectrometric detection is usually not applied with IC because of the necessity to use buffers that are often not suitable for MS-detection. However, there exist methods for the coupling of IC with mass spectrometric detection (Bauer et al. 1999; Buchberger and Ahrer 1999; Fisseha et al. 2004). Capillary electrophoresis. Electrophoretic separation techniques are also used for the analysis of aerosol constituents. Capillary electrophoresis (CE) has a high separation efﬁciency and can be coupled to mass spectrometers by ESI. However, CE/MS is usually only suited for the separation of analytes that are charged or are at least in equilibrium with a charged form, for example, carboxylic acids. CE/MS has been used in several studies mainly for the analysis of acidic organic compounds (Deng et al. 2003; Havers et al. 1998a; Iinuma and Herrmann 2003; Iinuma et al. 2004; Krivacsy et al. 2000; Neusüß et al. 1999; Rudolph and Stupak 2002). Functional group analysis. Due to the tremendous number of substances present in aerosol samples, some scientists tried not to identify every single component but to identify and quantify compound classes. For a series of atmospherically relevant questions, this information is adequate to give the satisfying answers. Decesari et al. (2000) separated the water extractable organics into groups of molecules having one, two and three or more carboxylic acid functionalities by anion exchange chromatography. This way, they were able to classify up to 82% of WSOC. They also demonstrated the importance of polycarboxylic acids (humic acids) in the water-soluble fraction of the investigated organic aerosol samples. However,

Organic Aerosols

373

there are also uncertainties using this approach, since certain individual substances do not elute within the compound class to which they actually belong (Chang et al. 2005). Especially nuclear magnetic resonance (NMR) can be useful for functional group analysis. Applying 1 H-NMR, it is possible to quantify protons of different functional groups in extracts of aerosol samples, such as OH, CH, CHO and COOH (Decesari et al. 2000; Graham et al. 2002; Havers et al. 1998b). This information can be used to get information about the chemical composition of the main organic components. Besides proton-NMR, solid-state 13 C-NMR of extracted and dried matter was also used for the characterisation of particulate WSOC (Duarte et al. 2005). Finally, infrared spectroscopy (IR) can be applied for functional group analysis. Extracted and dried material is used to prepare KBr pellets (Duarte et al. 2005; Havers et al. 1998b). It is also possible to apply IR measurements directly to collected material on Teﬂon ﬁlters (Maria et al. 2002). OC/EC measurements. OC and EC are usually determined from quartz ﬁbre ﬁlter samples by ‘volatilisation’ of OC and EC using thermal desorption, pyrolysis and combustion processes at different temperatures and/or in different atmospheres (inert/oxidising). Gaseous products are either converted to CO2 , which is measured by spectroscopic techniques (IR), or converted to CH4 , which is measured with an FID. While the determination of TC is a relatively easy task, the challenge is to differentiate between OC and EC. Basically, OC is identiﬁed by its ability of volatilisation/combustion at relative low temperatures compared to EC or its volatilisation/pyrolysis in an inert atmosphere, respectively. However, low volatile substances will partially be converted to EC under combustion or pyrolysis conditions. This process is called ‘charring’. For the correction of the charring, optical methods are used that measure the change in reﬂection or transmission of light due to the blackening of the ﬁlter by the charring process (thermal/optical transmittance, thermal/optical reﬂectance). All carbon measured under EC combustion conditions will be counted as OC until the reﬂection/transmission becomes the initial value. Intercomparisons have shown that different methods may result in different values of measured OC and EC (Birch 1998; Chow et al. 2001, 2004; Gelencser 2004; Huntzicker et al. 1982; Schmid et al. 2001).

9.3.4

Online mass spectrometric methods

The analytical techniques described above are based on a time integrating sampling step (using ﬁlters, impactors, etc.) with the inherent risk of positive and negative artefacts during the multi-step analytical procedure. Moreover, this type of sampling severely limits both spatial and temporal sampling densities. Therefore, analytical techniques to monitor the organic aerosol fraction with a high time resolution are required, especially for the understanding and control of highly dynamic systems as atmospheric aerosols. For certain atmospheric research platforms, such as aircrafts, rapid online systems are even mandatory. Consequently, several real-time analytical methodologies based on mass spectrometry for particle analysis have been developed within the last few years. Laser-based aerosol mass spectrometers allow the real-time analysis of single particles (Johnston 2000; Noble and Prather 2000; Thompson et al. 2000). Using nozzles, capillaries or aerodynamic lenses, particle beams are created and transferred into a vacuum. There,

374

Volatile Organic Compounds in the Atmosphere

Aerodynamic lens system

Ions

to MS

Chopper

Evaporation and lonisation

Aerosol inlet Particle beam

Vacuum stage 1

Vacuum stage 2

Vacuum stage 3

Pumps Figure 9.11 Scheme of the AMS-instrument from Aerodyne. Modiﬁed after Jayne et al. (2000).

the particles are detected by the scattering of a continuous laser beam followed by vaporisation and ionisation by second laser system. It is also possible to use different lasers to separate the ablation/desorption and ionisation step, for example, a combination of IRand ultraviolet (UV) lasers. For ion separation time-of-ﬂight mass spectrometers are used. However, although the organic aerosol fraction gives characteristic ion patterns with laserbased aerosol mass spectrometers, the usually observed strong fragmentation and the strong matrix effects do not allow quantiﬁcation of organic aerosol components. Nevertheless, for the characterisation and identiﬁcation of inorganic aerosol components (e.g. mineral dust, sea salt, metals), these single-particle instruments are valuable tools for aerosol analysis (Reents and Schabel 2001; Sipin et al. 2003). For some speciﬁc applications, for example, for the detection of high-molecular-weight compounds in laboratory-generated SOA, singleparticle instruments appear to be useful tools even for organic aerosol characterisation (Gross et al. 2006). In these cases, the ability to also produce larger ions is probably connected with the presence of a suitable matrix in the investigated particles, which allows a MALDI-like desorption/ionisation processes. Another (non-laser-based) online mass spectrometric technique for aerosol analysis uses thermal desorption (ﬂash evaporation) of the aerosol components, typically followed by electron ionisation of the desorbed components (so-called aerosol mass spectrometers, AMS). At ﬁrst, a particle beam is generated by aerodynamic lenses, and the particles are transferred into the vacuum region of the MS. Before the particles hit a vapouriser, they have to pass a chopper and a drift region, a set-up that generates the size information (see Figure 9.11). The vapourised compounds are than ionised by electron ionisation (EI) and analysed by quadrupole mass spectrometry (Jayne et al. 2000; Zhang et al. 2005b). These systems have proven to be especially useful for the quantitative measurement of major inorganic compounds, such as sulphate, nitrate and ammonium. Also, valuable information about the organic aerosol fraction can be obtained; however, due to the strong fragmentation resulting from EI single species can hardly be identiﬁed. Recently, an algorithm was developed to deconvolve the mass spectra of organic aerosols acquired with these instruments in order to differentiate between hydrocarbon-like and oxygenated organic aerosol (HOA and OOA, respectively) (Zhang et al. 2005a). Currently, developments in this ﬁeld

Organic Aerosols

375

are focussing on soft ionisation techniques and the use of high-resolution, time-of-ﬂight mass spectrometers (Drewnick et al. 2005). However, still a rapid development in ionisation and analyser techniques will keep this type of online aerosol mass spectrometry an active ﬁeld of analytical chemistry in the coming years. Using a similar set-up for the introduction of particles into the vacuum, the particles can also be collected on impactor plates before they are thermally desorbed and ionised. The volatilisation is then done discontinuously after a certain collection time either by heating the collector (Tobias and Ziemann 1999; Tobias et al. 2000) or by the use of infrared lasers followed by soft (Vacuum Ultraviolet) VUV-ionisation (Oktem et al. 2004). Also, atmospheric pressure chemical ionisation (APCI) has been applied for the online analysis of aerosol particles using a modiﬁed APCI-source for LC/MS. Since APCI is a very soft ionisation technique, it can provide valuable information about individual molecular species. Using an ion trap mass spectrometer, pseudo-molecular ions can be characterised by their MSn -spectra. It is a quantitative method that has been used for the characterisation of SOA components in chamber studies (Kuckelmann et al. 2000).

Further reading Atkinson, R. and Arey, J. (2003) Gas-phase tropospheric chemistry of biogenic volatile organic compounds: A review. Atmospheric Environment, 37: S197–219. Jacobson, M.C., Hansson, H.C., Noone, K.J. and Charlson, R.J. (2000) Organic atmospheric aerosols: Review and state of the science. Reviews of Geophysics, 38: 267–94. Kanakidou, M., Seinfeld, J.H., Pandis, S.N., et al. (2005) Organic aerosol and global climate modelling: A review. Atmospheric Chemistry and Physics, 5: 1053–123. Seinfeld, J.H. and Pankow, J.F. (2003) Organic atmospheric particulate material. Annual Review of Physical Chemistry, 54: 121–40.

References Alves, C., Carvalho, A. and Pio, C. (2002) Mass balance of organic carbon fractions in atmospheric aerosols. Journal of Geophysical Research-Atmospheres, 107: D21,8345. Anttila, P., Hyotylainen, T., Heikkila, A., et al. (2005) Determination of organic acids in aerosol particles from a coniferous forest by liquid chromatography-mass spectrometry. Journal of Separation Science, 28: 337–46. Ariya, P.A. and Amyot, M. (2004) New directions: The role of bioaerosols in atmospheric chemistry and physics. Atmospheric Environment, 38: 1231–2. Ariya, P.A., Nepotchatykh, O., Ignatova, O. and Amyot, M. (2002) Microbiological degradation of atmospheric organic compounds. Geophysical Research Letters, 29: 22,2077. Artaxo, P., Maenhaut, W., Storms, H. and Van Grieken, R. (1990) Aerosol characteristics for the Amazon Basin during the wet season. Journal of Geophysical Research, 95: 16971–85. Artaxo, P., Storms, H., Bruynseels, F., Van Grieken, R. and Maenhaut, W. (1988) Composition and sources of aerosols from the Amazonian basin. Journal of Geophysical Research, 93: 1605–15. Artaxo, P., Martins, J.V., Yamasoe, M.A., et al. (2002) Physical and chemical properties of aerosols in the wet and dry seasons in Rondonia, Amazonia. Journal of Geophysical Research-Atmospheres, 107: D20,8081. Atkinson, R. (2000) Atmospheric chemistry of VOCs and NOx . Atmospheric Environment, 34: 2063–101.

376

Volatile Organic Compounds in the Atmosphere

Atkinson, R. and Aschmann, S.M. (1994) Products of the gas-phase reactions of aromatichydrocarbons – Effect of No2 Concentration. International Journal of Chemical Kinetics, 26: 929–44. Bang, J.J., Guerrero, P.A., Lopez, D.A., Murr, L.E. and Esquivel, E.V. (2004) Carbon nanotubes and other fullerene nanocrystals in domestic propane and natural gas combustion streams. Journal of Nanoscience and Nanotechnology, 4: 716–18. Bates, T.S., Kapustin, V.N., Quinn, P.K., et al. (1998) Processes controlling the distribution of aerosol particles in the lower marine boundary layer during the First Aerosol Characterization Experiment (ACE 1). Journal of Geophysical Research-Atmospheres, 103: 16369–83. Bauer, K.H., Knepper, T.P., Maes, A., Schatz, V. and Voihsel, M. (1999) Analysis of polar organic micropollutants in water with ion chromatography electrospray mass spectrometry. Journal of Chromatography A, 837: 117–28. Berner, A. and Luerzer, C. (1980) Mass size distributions of trafﬁc aerosols in Vienna. Journal of Physical Chemistry, 84: 2079–83. Bethel, H.L., Atkinson, R. and Arey, J. (2000) Products of the gas-phase reactions of OH radicals with p-xylene and 1,2,3-and 1,2,4-trimethylbenzene: Effect of NO2 concentration. Journal of Physical Chemistry A, 104: 8922–9. Binabas, M.R. and Simoneit, B.R.T. (1996) Composition of extractable organic matter of air particles from Malaysia: Initial study. Atmospheric Environment, 30: 2779–93. Binabas, M.R., Simoneit, B.R.T., Elias, V., Cabral, J.A. and Cardoso, J.N. (1995) Composition of higher molecular-weight organic-matter in smoke aerosol from biomass combustion in Amazonia. Chemosphere, 30: 995–1015. Birch, M.E. (1998) Analysis of carbonaceous aerosols: Interlaboratory comparison. Analyst, 123: 851–7. Bonn, B., von Kuhlmann, R. and Lawrence, M.G. (2004) High contribution of biogenic hydroperoxides to secondary organic aerosol formation. Geophysical Research Letters, 31: L10108. Brown, S.G., Herckes, P., Ashbaugh, L., Hannigan, M.P., Kreidenweis, S.M. and Collett, J.L. (2002) Characterization of organic aerosol in Big Bend National Park, Texas. Atmospheric Environment, 36: 5807–18. Buchberger, W. and Ahrer, W. (1999) Combination of suppressed and non-suppressed ion chromatography with atmospheric pressure ionization mass spectrometry for the determination of anions. Journal of Chromatography. A, 850: 99–106. Cabada, J.C., Pandis, S.N., Subramanian, R., Robinson, A.L., Polidori, A. and Turpin, B. (2004) Estimating the secondary organic aerosol contribution to PM2.5 using the EC tracer method. Aerosol Science and Technology, 38: 140–55. Calogirou, A., Larsen, B.R. and Kotzias, D. (1999) Gas-phase terpene oxidation products: A review. Atmospheric Environment, 33 : 1423–39. Cao, J.J., Lee, S.C., Ho, K.F., et al. (2004) Spatial and seasonal variations of atmospheric organic carbon and elemental carbon in Pearl River Delta Region, China. Atmospheric Environment, 38: 4447–56. Castro, L.M., Pio, C.A., Harrison, R.M. and Smith, D.J.T. (1999) Carbonaceous aerosol in urban and rural European atmospheres. Estimation of secondary organic carbon concentrations. Atmospheric Environment, 33: 2771–81. Cecinato, A., Di Palo, V., Pomata, D., Sciano, M.C.T. and Possanzini, M. (2005) Measurement of phase-distributed nitrophenols in Rome ambient air. Chemosphere, 59: 679–83. Chang, H., Herckes, P. and Collett, J.L. (2005) On the use of anion exchange chromatography for the characterization of water soluble organic carbon. Geophysical Research Letters, 32(1): L01810. Cheng, Y. and Li, S.M. (2005) Nonderivatization analytical method of fatty acids and cis-pinonic acid and its application in ambient PM2.5 aerosols in the greater Vancouver area in Canada. Environmental Science and Technology, 39: 2239–46.

Organic Aerosols

377

Chow, J.C., Watson, J.G., Crow, D., Lowenthal, D.H. and Merriﬁeld, T. (2001) Comparison of IMPROVE and NIOSH carbon measurements. Aerosol Science and Technology, 34: 23–34. Chow, J.C., Watson, J.G., Chen, L.W.A., Arnott, W.P. and Moosmuller, H. (2004) Equivalence of elemental carbon by thermal/optical reﬂectance and transmittance with different temperature protocols. Environmental Science and Technology, 38: 4414–22. Christoffersen, T.S., Hjorth, J., Horie, O., et al. (1998) cis-pinic acid, a possible precursor for organic aerosol formation from ozonolysis of alpha-pinene. Atmospheric Environment, 32: 1657–61. Chung, S.H. and Seinfeld, J.H. (2002) Global distribution and climate forcing of carbonaceous aerosols. Journal of Geophysical Research-Atmospheres, 107: D19,4407. Claeys, M., Wang, W., Ion, A.C., Kourtchev, I., Gelencser, A. and Maenhaut, W. (2004a) Formation of secondary organic aerosols from isoprene and its gas-phase oxidation products through reaction with hydrogen peroxide. Atmospheric Environment, 38: 4093–8. Claeys, M., Graham, B., Vas, G., et al. (2004b) Formation of secondary organic aerosols through photooxidation of isoprene. Science, 303: 1173–6. Crahan, K.K., Hegg, D., Covert, D.S. and Jonsson, H. (2004) An exploration of aqueous oxalic acid production in the coastal marine atmosphere. Atmospheric Environment, 38: 3757–64. Currie, L.A. (2000) Evolution and multidisciplinary frontiers of C-14 aerosol science. Radiocarbon, 42: 115–26. Decesari, S., Facchini, M.C., Fuzzi, S. and Tagliavini, E. (2000) Characterization of water-soluble organic compounds in atmospheric aerosol: A new approach. Journal of Geophysical ResearchAtmospheres, 105: 1481–9. Deng, Y., Wellons, A., Bolla, D., Krzyaniak, M. and Wylie, H. (2003) Separation and identiﬁcation of photodegradation products of benzoic acid by capillary zone electrophoresis. Journal of Chromatography A, 1013: 191–201. Dindorf, T., Kuhn, U., Ganzeveld, L., et al. (2005) Emission of monoterpenes from European beech (Fagus sylvatica L.) as a function of light and temperature. Biogeosciences Discussions, 2: 137–82. Docherty, K.S., Wu, W., Lim, Y.B. and Ziemann, P.J. (2005) Contributions of organic peroxides to secondary aerosol formed from reactions of monoterpenes with O-3. Environmental Science and Technology, 39: 4049. Dockery, D.W., Pope, C.A., Xu, X., et al. (1993) An association between air pollution and mortality in six US cities. New England Journal of Medicine, 329: 1753–9. Drewnick, F., Hings, S.S., DeCarlo, P., et al. (2005) A new time-of-ﬂight aerosol mass spectrometer (TOF-AMS) – Instrument description and ﬁrst ﬁeld deployment. Aerosol Science and Technology, 39: 637–58. Duarte, R., Pio, C.A. and Duarte, A.C. (2005) Spectroscopic study of the water-soluble organic matter isolated from atmospheric aerosols collected under different atmospheric conditions. Analytica Chimica Acta, 530: 7–14. Edney, E.O., Driscoll, D.J., Weathers, W.S., et al. (2001) Formation of polyketones in irradiated toluene/propylene/NOx /air mixtures. Aerosol Science and Technology, 35: 998–1008. Endo, M., Yamamoto, N., Yoshinaga, J., et al. (2004) C-14 measurement for size-fractionated airborne particulate matters. Atmospheric Environment, 38: 6263–7. Falkovich, A.H., Graber, E.R., Schkolnik, G., Rudich, Y., Maenhaut, W. and Artaxo, P. (2005) Low molecular weight organic acids in aerosol particles from Rondonia, Brazil, during the biomass-burning, transition and wet periods. Atmospheric Chemistry and Physics, 5: 781–97. Fang, M., Zheng, M., Wang, F., To, K.L., Jaafar, A.B. and Tong, S.L. (1999) The solvent-extractable organic compounds in the Indonesia biomass burning aerosols – Characterization studies. Atmospheric Environment, 33: 783–95. Fernandes, M.B. and Brooks, P. (2003) Characterization of carbonaceous combustion residues: II. Nonpolar organic compounds. Chemosphere, 53: 447–58.

378

Volatile Organic Compounds in the Atmosphere

Fernandes, M.B., Skjemstad, J.O., Johnson, B.B., Wells, J.D. and Brooks, P. (2003) Characterization of carbonaceous combustion residues. I. Morphological, elemental and spectroscopic features. Chemosphere, 51: 785–95. Fine, P.M., Chakrabarti, B., Krudysz, M., Schauer, J.J. and Sioutas, C. (2004) Diurnal variations of individual organic compound constituents of ultraﬁne and accumulation mode particulate matter in the Los Angeles basin. Environmental Science and Technology, 38: 1296–304. Finlayson-Pitts, B.J. and Pitts, J.N.J. (2000) Chemistry of the Upper and Lower Atmosphere. Academic Press, San Diego, CA. Fisseha, R., Dommen, J., Sax, M., et al. (2004) Identiﬁcation of organic organic aerosol and the acids in secondary corresponding gas phase from chamber experiments. Analytical Chemistry, 76: 6535–40. Forstner, H.J.L., Flagan, R.C. and Seinfeld, J.H. (1997a) Secondary organic aerosol from the photooxidation of aromatic hydrocarbons: Molecular composition. Environmental Science and Technology, 31: 1345–58. Forstner, H.J.L., Flagan, R.C. and Seinfeld, J.H. (1997b) Molecular speciation of secondary organic aerosol from photooxidation of the higher alkenes: 1-Octene and 1-decene. Atmospheric Environment, 31: 1953–64. Fraser, M.P., Cass, G.R. and Simoneit, B.R.T. (2003) Air quality model evaluation data for organics. 6. C-3-C-24 organic acids. Environmental Science and Technology, 37: 446–53. Fuzzi, S., Mandrioli, P. and Perfetto, A. (1997) Fog droplets – An atmospheric source of secondary biological aerosol particles. Atmospheric Environment, 31: 287–90. Gao, S., Ng, N.L., Keywood, M., et al. (2004) Particle phase acidity and oligomer formation in secondary organic aerosol. Environmental Science and Technology, 38: 6582–9. Gelencser, A. (2004) Carbonaceous Aerosol. Dodrecht, Springer. Glasius, M., Duane, M. and Larsen, B.R. (1999) Determination of polar terpene oxidation products in aerosols by liquid chromatography-ion trap mass spectrometry. Journal of Chromatography A, 833: 121–35. Glasius, M., Lahaniati, M., Calogirou, A., et al. (2000) Carboxylic acids in secondary aerosols from oxidation of cyclic monoterpenes by ozone. Environmental Science and Technology, 34: 1001–10. Gogou, A.I., Apostolaki, M. and Stephanou, E.G. (1998) Determination of organic molecular markers in marine aerosols and sediments: One-step ﬂash chromatography compound class fractionation and capillary gas chromatographic analysis. Journal of Chromatography A, 799: 215–31. Graham, B., Mayol-Bracero, O.L., Guyon, P., et al. (2002) Water-soluble organic compounds in biomass burning aerosols over Amazonia – 1. Characterization by NMR and GC-MS. Journal of Geophysical Research-Atmospheres, 107: D20,8047. Graham, B., Guyon, P., Taylor, P.E., et al. (2003a) Organic compounds present in the natural Amazonian aerosol: Characterization by gas chromatography-mass spectrometry – Art. no. 4766. Journal of Geophysical Research-Atmospheres, 108: D24,4766. Graham, B., Guyon, P., Maenhaut, W., et al. (2003b) Composition and diurnal variability of the natural Amazonian aerosol – Art. no. 4765. Journal of Geophysical Research-Atmospheres, 108: D24,4765. Gray, H.A., Cass, G.R., Huntzicker, J.J., Heyerdahl, E.K. and Rau, J.A. (1986) Characteristics of atmospheric organic and elemental carbon particle concentrations in Los Angeles. Environmental Science and Technology, 20: 580–9. Grifﬁn, D.W., Garrison, V.H., Herman, J.R. and Shinn, E.A. (2001) African desert dust in the Caribbean atmosphere: Microbiology and public health. Aerobiologia, 17: 203–13. Grifﬁn, R.J., Cocker III, D.R., Seinfeld, J.H. and Dabdub, D. (1999) Esitmate of global atmospheric organic aerosol from oxidation of biogenic hydrocarbons. Geophysical Research Letters, 26: 2721–4. Gross, D.S., Galli, M.E., Kalberer, M., et al. (2006) Real-time measurement of oligomeric species in secondary organic aerosol with the aerosol time-of-ﬂight mass spectrometer. Analytical Chemistry, 78: 2130–7.

Organic Aerosols

379

Guo, Z.G., Sheng, L.F., Feng, J.L. and Fang, M. (2003) Seasonal variation of solvent extractable organic compounds in the aerosols in Qingdao, China. Atmospheric Environment, 37: 1825–34. Hakola, H., Rinne, J. and Laurila, T. (1998) The hydrocarbon emission rates of tea-leafed willow (Salix phylicifolia), silver birch (Betula pendula) and European aspen (Populus tremula). Atmospheric Environment, 32: 1825–33. Havers, N., Burba, P., Klockow, D. and Klockow-Beck, A. (1998a) Characterization of humic-like substances in airborne particulate matter by capillary electrophoresis. Chromatographia, 47: 619–24. Havers, N., Burba, P., Lambert, J. and Klockow, D. (1998b) Spectroscopic characterization of humic-like substances in airborne particulate matter. Journal of Atmospheric Chemistry, 29: 45–54. He, L.Y., Hu, M., Huang, X.F., Yu, B.D., Zhang, Y.H. and Liu, D.Q. (2004a) Measurement of emissions of ﬁne particulate organic matter from Chinese cooking. Atmospheric Environment, 38: 6557–64. He, Z., Kim, Y.J., Ogunjobi, K.O., Kim, J.E. and Ryu, S.Y. (2004b) Carbonaceous aerosol characteristics of PM2.5 particles in northeastern Asia in summer 2002. Atmospheric Environment, 38: 1795–800. Height, M.J., Howard, J.B., Tester, J.W. and Sande, J.B.V. (2004) Flame synthesis of single-walled carbon nanotubes. Carbon, 42: 2295–307. Hinds, W.C. (1999) Aerosol Technology. John Wiley & Sons, Inc., New York. Hoffmann, T., Odum, J.R., Bowman, F., et al. (1997) Formation of organic aerosols from the oxidation of biogenic hydrocarbons. Journal of Atmospheric Chemistry, 26: 189–222. Hoffmann, T., Bandur, R., Marggraf, U. and Linscheid, M. (1998) Molecular composition of organic aerosols formed in the alpha-pinene/O-3 reaction: Implications for new particle formation processes. Journal of Geophysical Research, 103: 25569–78. Hueglin, C., Gehrig, R., Baltensperger, U., Gysel, M., Monn, C. and Vonmont, H. (2005) Chemical characterisation of PM2.5, PM10 and coarse particles at urban, near-city and rural sites in Switzerland. Atmospheric Environment, 39: 637–51. Huntzicker, J.J., Johnson, R.L., Shah, J.J. and Cary, R.A. (1982) Analysis of organic and elemental carbon in ambient aerosols by the thermal-optical method. In: G.T. Wolff and R.L. Klimisch (Eds) Particulate Carbon: Atmospheric Life Cycle. Plenum Press, New York, pp. 79–88. Iinuma, Y. and Herrmann, H. (2003) Method development for the analysis of particle phase substituted methoxy phenols and aromatic acids from biomass burning using capillary electrophoresis/electro spray ionization mass spectrometry (CE/ESI-MS). Journal of Chromatography A, 1018: 105–15. Iinuma, Y., Boge, O., Gnauk, T. and Herrmann, H. (2004) Aerosol-chamber study of the alphapinene/O-3 reaction: Inﬂuence of particle acidity on aerosol yields and products. Atmospheric Environment, 38: 761–73. Izumi, K. and Fukuyama, T. (1990) Photochemical aerosol formation from aromatic hydrocarbons in the presence of NOx . Atmospheric Environment, 24A: 1433–41. Jaenicke, R. (2005) Abundance of cellular material and proteins in the atmosphere. Science, 308: 73. Jaffrezo, J.L., Calas, N. and Boucher, M. (1998) Carboxylic acids measurements with ion chromatography. Atmospheric Environment, 32: 2705–08. Jang, M. and McDow, S.R. (1997) Products of benz a anthracene photodegradation in the presence of known organic constituents of atmospheric aerosols. Environmental Science and Technology, 31: 1046–53. Jang, M., Czoschke, N.M. and Northcross, A.L. (2004) Atmospheric organic aerosol production by heterogeneous acid-catalyzed reactions. Chemphyschem, 5: 1647–61. Jang, M.S. and Kamens, R.M. (2001) Characterization of secondary aerosol from the photooxidation of toluene in the presence of NOx and 1-propene. Environmental Science and Technology, 35: 3626–39. Jang, M.S., Czoschke, N.M., Lee, S. and Kamens, R.M. (2002) Heterogeneous atmospheric aerosol production by acid-catalyzed particle-phase reactions. Science, 298: 814–17.

380

Volatile Organic Compounds in the Atmosphere

Jaoui, M. and Kamens, R.M. (2003a) Gaseous and particulate oxidation products analysis of a mixture of alpha-pinene plus beta-pinene/O-3/air in the absence of light and alpha-pinene plus beta-pinene/NOx /air in the presence of natural sunlight. Journal of Atmospheric Chemistry, 44: 259–97. Jaoui, M. and Kamens, R.M. (2003b) Mass balance of gaseous and particulate products from beta-pinene/O-3/air in the absence of light and beta-pinene/NOx /air in the presence of natural sunlight. Journal of Atmospheric Chemistry, 45: 101–41. Jaoui, M. and Kamens, R.M. (2003c) Gas and particulate products distribution from the photooxidation of alpha-humulene in the presence of NOx , natural atmospheric air and sunlight. Journal of Atmospheric Chemistry, 46: 29–54. Jayne, J.T., Leard, D.C., Zhang, X.F., et al. (2000) Development of an aerosol mass spectrometer for size and composition analysis of submicron particles. Aerosol Science and Technology, 33: 49–70. Johnston, M.V. (2000) Sampling and analysis of individual particles by aerosol mass spectrometry. Journal of Mass Spectrometry, 35: 585–95. Kalaitzoglou, M., Terzi, E. and Samara, C. (2004) Patterns and sources of particle-phase aliphatic and polycyclic aromatic hydrocarbons in urban and rural sites of western Greece. Atmospheric Environment, 38: 2545–60. Kalberer, M., Paulsen, D., Sax, M., et al. (2004) Identiﬁcation of polymers as major components of atmospheric organic aerosols. Science, 303: 1659–62. Kavouras, I.G. and Stephanou, E.G. (2002) Particle size distribution of organic primary and secondary aerosol constituents in urban, background marine, and forest atmosphere. Journal of Geophysical Research-Atmospheres, 107: D8,4069. Kavouras, I.G., Mihalopoulos, N. and Stephanou, E.G. (1998) Formation of atmospheric particles from organic acids produced by forests. Nature, 395: 683–6. Kavouras, I.G., Mihalopoulos, N. and Stephanou, E.G. (1999) Secondary organic aerosol formation vs primary organic aerosol emission: In situ evidence for the chemical coupling between monoterpene acidic photooxidation products and new particle formation over forests. Environmental Science and Technology, 33: 1028–37. Kawamura, K. and Kaplan, I.R. (1987) Motor exhaust emissions as a primary source for dicarboxylic acids in Los Angeles ambient air. Environmental Science and Technology, 21: 105–10. Kawamura, K. and Ikushima, K. (1993) Seasonal changes in the distribution of dicarboxylic acids in the urban atmosphere. Environmental Science and Technology, 27: 2227–35. Kawamura, K. and Kasukabe, H. (1996) Source and reaction pathways of dicarboxylic acids, ketoacids and dicarbonyls in artic aerosol: One year of observations. Atmospheric Environment, 30: 1709–22. Kawamura, K. and Yasui, O. (2005) Diurnal changes in the distribution of dicarboxylic acids, ketocarboxylic acids and dicarbonyls in the urban Tokyo atmosphere. Atmospheric Environment, 39: 1945–60. Kawamura, K., Imai, Y. and Barrie, L.A. (2005) Photochemical production and loss of organic acids in high Arctic aerosols during long-range transport and polar sunrise ozone depletion events. Atmospheric Environment, 39: 599–614. Kerminen, V.M., Teinila, K., Hillamo, R. and Makela, T. (1999) Size-segregated chemistry of particulate dicarboxylic acids in the Arctic atmosphere. Atmospheric Environment, 33: 2089–100. Kerminen, V.M., Ojanen, C., Pakkanen, T., Hillamo, R., Aurela, M. and Merilainen, J. (2000) Lowmolecular-weight dicarboxylic acids in an urban and rural atmosphere. Journal of Aerosol Science, 31: 349–62. Kirchstetter, T., Novakov, T., Morales, R. and Rosario, O. (2000) Differences in the volatility of organic aerosols in unpolluted tropical and polluted continental atmospheres. Journal of Geophysical Research-Atmosphere, 105: 26547–54.

Organic Aerosols

381

Kirchstetter, T.W., Corrigan, C.E. and Novakov, T. (2001) Laboratory and ﬁeld investigation of the adsorption of gaseous organic compounds onto quartz ﬁlters. Atmospheric Environment, 35: 1663–71. Kleindienst, T.E., Conver, T.S., McIver, C.D. and Edney, E.O. (2004) Determination of secondary organic aerosol products from the photooxidation of toluene and their implications in ambient PM2.5. Journal of Atmospheric Chemistry, 47: 79–100. Koch, S., Winterhalter, R., Uherek, E., Kolloff, A., Neeb, P. and Moortgat, G.K. (2000) Formation of new particles in the gas-phase ozonolysis of monoterpenes. Atmospheric Environment, 34: 4031–42. Krivacsy, Z., Kiss, G., Varga, B., et al. (2000) Study of humic-like substances in fog and interstitial aerosol by size-exclusion chromatography and capillary electrophoresis. Atmospheric Environment, 34: 4273–81. Kubatova, A., Vermeylen, R., Claeys, M., et al. (2000) Carbonaceous aerosol characterization in the Amazon basin, Brazil: Novel dicarboxylic acids and related compounds. Atmospheric Environment, 34: 5037–51. Kubatova, A., Vermeylen, R., Claeys, M., Cafmeyer, J. and Maenhaut, W. (2002) Organic compounds in urban aerosols from Gent, Belgium: Characterization, sources, and seasonal differences. Journal of Geophysical Research-Atmospheres, 107: D21,8343. Kuckelmann, U., Warscheid, S. and Hoffmann, T. (2000) On-line characterization of organic aerosols formed from biogenic precursors using atmospheric pressure chemical ionization mass spectrometry. Analytical Chemistry, 72: 1905–12. Kuhlbusch, T.A.J. (1998) Black carbon and the carbon cycle. Science, 280: 1903–4. Kulmala, M. (2003) How particles nucleate and grow. Science, 302: 1000–1. Kulmala, M., Vehkamaki, H., Petajda, T., et al. (2004) Formation and growth rates of ultraﬁne atmospheric particles: A review of observations. Journal of Aerosol Science, 35: 143–76. Lack, D.A., Tie, X.X., Boﬁnger, N.D., Wiegand, A.N. and Madronich, S. (2004) Seasonal variability of secondary organic aerosol: A global modeling study. Journal of Geophysical ResearchAtmospheres, 109: D03203. Laden, F., Neas, L.M., Dockery, D.W. and Schwartz, J. (2000) Association of ﬁne particulate matter from different sources with daily mortality in six US cities. Environmental Health Perspectives, 108: 941. Larsen, B.R., Di Bella, D., Glasius, M., Winterhalter, R., Jensen, N.R. and Hjorth, J. (2001) Gasphase OH oxidation of monoterpenes: Gaseous and particulate products. Journal of Atmospheric Chemistry, 38: 231–76. Lee, A.K.Y., Chan, C.K., Fang, M. and Lau, A.P.S. (2004) The 3-hydroxy fatty acids as biomarkers for quantiﬁcation and characterization of endotoxins and Gram-negative bacteria in atmospheric aerosols in Hong Kong. Atmospheric Environment, 38: 6307–17. Lee, H.K. (1995) Recent applications of gas and high-performance liquid-chromatographic techniques to the analysis of polycyclic aromatic-hydrocarbons in airborne-particulates. Journal of Chromatography A, 710: 79–92. Lee, H.S. and Kang, B.W. (2001) Chemical characteristics of principal PM2.5 species in Chongju, South Korea. Atmospheric Environment, 35: 739–46. Lemire, K.R., Allen, D.T., Klouda, G.A. and Lewis, C.W. (2002) Fine particulate matter source attribution for Southeast Texas using C-14/C-13 ratios. Journal of Geophysical Research-Atmospheres. 107: D22,4613. Lewis, C.W., Klouda, G.A. and Ellenson, W.D. (2004) Radiocarbon measurement of the biogenic contribution to summertime PM-2.5 ambient aerosol in Nashville, TN. Atmospheric Environment, 38: 6053–61. Limbeck, A. and Puxbaum, H. (1999) Organic acids in continental background aerosols. Atmospheric Environment, 33: 1847–52.

382

Volatile Organic Compounds in the Atmosphere

Limbeck, A., Kulmala, M. and Puxbaum, H. (2003) Secondary organic aerosol formation in the atmosphere via heterogeneous reaction of gaseous isoprene on acidic particles – Art. no. 1996. Geophysical Research Letters, 30: 19, 1996, doi:10.1029/2003GL017738. Mader, B.T. and Pankow, J.F. (2001) Gas/solid partitioning of semivolatile organic compounds (SOCs) to air ﬁlters. 3. An analysis of gas adsorption artifacts in measurements of atmospheric SOCs and organic carbon (OC) when using Teﬂon membrane ﬁlters and quartz ﬁber ﬁlters. Environmental Science and Technology, 35: 3422–32. Mader, B.T., Schauer, J.J., Seinfeld, J.H., et al. (2003) Sampling methods used for the collection of particle-phase organic and elemental carbon during ACE-Asia. Atmospheric Environment, 37: 1435–49. Mar, T.F., Norris, G.A., Koenig, J.Q. and Larson, T.V. (2000) Associations between air pollution and mortality in Phoenix, 1995–1997. Environmental Health Perspectives, 108: 347. Maria, S.F., Russell, L.M., Turpin, B.J. and Porcja, R.J. (2002) FTIR measurements of functional groups and organic mass in aerosol samples over the Caribbean. Atmospheric Environment, 36: 5185–96. Mochida, M., Kitamori, Y., Kawamura, K., Nojiri, Y. and Suzuki, K. (2002) Fatty acids in the marine atmosphere: Factors governing their concentrations and evaluation of organic ﬁlms on sea-salt particles. Journal of Geophysical Research-Atmospheres, 107: D17,4325. Mochida, M., Kawabata, A., Kawamura, K., Hatsushika, H. and Yamazaki, K. (2003) Seasonal variation and origins of dicarboxylic acids in the marine atmosphere over the western North Paciﬁc. Journal of Geophysical Research-Atmospheres, 108: D6,4193. Muller, K., Spindler, G., Maenhaut, W., et al. (2004) INTERCOMP2000, a campaign to assess the comparability of methods in use in Europe for measuring aerosol composition. Atmospheric Environment, 38: 6459–66. Na, K.S., Sawant, A.A., Song, C. and Cocker, D.R. (2004) Primary and secondary carbonaceous species in the atmosphere of Western Riverside County, California. Atmospheric Environment, 38: 1345–55. Neusüß, C., Brüggemann, E. and Herrmann, H. (1999) Determination of water-soluble organic ions of size-segregated aerosol particles by capillary zone electrophoresis. Proceedings of the EUROTRAC Symposium 1998. Edited by P.M. Borell and P. Borell, WIT Press, Southampton, UK, ISBN 1-85312-743-4, 1: 639–43. Noble, C. and Prather, K. (2000) Real-time single particle mass spectrometry: A historical review of a quarter century of the chemical analysis of aerosols. Mass Spectrometry Review, 19: 248–74. O’Dowd, C.D., Aalto, P., Hameri, K., Kulmala, M. and Hoffmann, T. (2002) Aerosol formation – Atmospheric particles from organic vapours. Nature, 416: 497–8. O’Dowd, C.D., Facchini, M.C., Cavalli, F., et al. (2004) Biogenically driven organic contribution to marine aerosol. Nature, 431: 676–80. Odum, J.R., Hoffmann, T., Bowman, F., Collins, D., Flagan, R.C. and Seinfeld, J.H. (1996) Gas/particle partitioning and secondary organic aerosol yields. Environmental Science and Technology, 30: 2580–5. Odum, J.R., Jungkamp, T.P.W., Grifﬁn, R.C., Flagan, R.C. and Seinfeld, J.H. (1997) The atmospheric aerosol-forming potential of whole gasoline vapor. Science, 276: 96–9. Oktem, B., Tolocka, M.P. and Johnston, M.V. (2004) On-line analysis of organic components in ﬁne and ultraﬁne particles by photoionization aerosol mass spectrometry. Analytical Chemistry, 76: 253–61. Pandis, S.N., Harley, R.A., Cass, G.R. and Seinfeld, J.H. (1992) Secondary organic aerosol formation and transport. Atmospheric Environment, 26A: 2269–82. Pankow, J.F. (1994a) An absorbtion model of gas/particle partitioning of organic compounds in the atmosphere. Atmospheric Environment, 28: 185–8. Pankow, J.F. (1994b) An absorbtion model of the gas/aerosol partitioning involved in the formation of secondary organic aerosol. Atmospheric Environment, 28: 189–93.

Organic Aerosols

383

Pasanen, A.L., Yli-Pietila, K., Pasanen, P., Kalliokoski, P. and Tarhanen, J. (1999) Ergosterol content in various fungal species and biocontaminated building materials. Applied and Environmental Microbiology, 65: 138–42. Posfai, M., Li, J., Anderson, J.R. and Buseck, P.R. (2003) Aerosol bacteria over the southern ocean during ACE-1. Atmospheric Research, 66: 231–240. Putaud, J.P., Raes, F., Van Dingenen, R., et al. (2004) European aerosol phenomenology-2: Chemical characteristics of particulate matter at kerbside, urban, rural and background sites in Europe. Atmospheric Environment, 38: 2579–95. Reents, W. and Schabel, M. (2001) Measurement of individual particle atomic composition by aerosol mass spectrometry. Analytical Chemistry, 73: 5403–14. Rees, S.L., Robinson, A.L., Khlystov, A., Stanier, C.O. and Pandis, S.N. (2004) Mass balance closure and the federal reference method for PM2.5 in Pittsburgh, Pennsylvania. Atmospheric Environment, 38: 3305–18. Rogge, W.F., Hildemann, L.M., Mazurek, M.A., Cass, G.R. and Simoneit, B.R.T. (1991) Sources of ﬁne organic aerosol. 1. Charbroilers and meat cooking operations. Environmental Science and Technology, 25: 1112–25. Rogge, W.F., Hildemann, L.M., Mazurek, M.A., Cass, G.R. and Simoneit, B.R.T. (1993a) Sources of ﬁne organic aerosol .4. Particulate abrasion products from leaf surfaces of urban plants. Environmental Science and Technology, 27: 2700–11. Rogge, W.F., Hildemann, L.M., Mazurek, M.A., Cass, G.R. and Simoneit, B.R.T. (1993b) Sources of ﬁne organic aerosol. 2. Noncatalsyt and catalyst-equipped automobiles and heavy duty trucks. Environmental Science and Technology, 27: 636–51. Rohrl, A. and Lammel, G. (2001) Low molecular weight dicarboxylic acids and glyoxylic acid: Seasonal and air mass characteristics. Environmental Science and Technology, 35: 95–101. Römpp, A. (2003) Analysis of Organic Compounds in Atmospheric Aerosol by Liquid ChromatographyHigh Resolution Mass Spectrometry (LC/MS/MS-TOF): Method Development and Applications. PhD Thesis, University of Mainz, Germany Rudolph, J. and Stupak, J. (2002) Determination of aromatic acids and nitrophenols in atmospheric aerosols by capillary electrophoresis. Journal of Chromatographic Science, 40: 207–13. Russell, M. and Allen, D.T.D.T. (2004) Seasonal and spatial trends in primary and secondary organic carbon concentrations in southeast Texas. Atmospheric Environment, 38: 3225–39. Salma, I., Chi, X.G. and Maenhaut, W. (2004) Elemental and organic carbon in urban canyon and background environments in Budapest, Hungary. Atmospheric Environment, 38: 27–36. Saraf, A., Larsson, L., Burge, H. and Milton, D. (1997) Quantiﬁcation of ergosterol and 3-hydroxy fatty acids in settled house dust by gas chromatography mass spectrometry: Comparison with fungal culture and determination of endotoxin by a Limulus amebocyte lysate assay. Applied and Environmental Microbiology, 63: 2554–9. Sattler, B., Puxbaum, H. and Psenner, R. (2001) Bacterial growth in supercooled cloud droplets. Geophysical Research Letters, 28: 239–42. Schauer, J.J. and Cass, G.R. (2000) Source apportionment of wintertime gas-phase and particle-phase air pollutants using organic compounds as tracers. Environmental Science and Technology, 34: 1821–32. Schauer, J.J., Rogge, W.F., Hildemann, L.M., Mazurek, M.A. and Cass, G.R. (1996) Source apportionment of airborne particulate matter using organic compounds as tracers. Atmospheric Environment, 30: 3837–55. Schauer, J.J., Kleeman, M.J., Cass, G.R. and Simoneit, B.R.T. (2002) Measurement of emissions from air pollution sources. 5. C-1-C-32 organic compounds from gasoline-powered motor vehicles. Environmental Science and Technology, 36: 1169–80. Schmid, H., Laskus, L., Abraham, H.J., et al. (2001) Results of the ‘carbon conference’ international aerosol carbon round robin test stage I. Atmospheric Environment, 35: 2111–21.

384

Volatile Organic Compounds in the Atmosphere

Seinfeld, J.H. and Pandis, S.N. (1998) Atmospheric Chemistry and Physics. John Wiley & Sons, Inc., New York, ISBN 0-471-17815-2. Sellegri, K., Laj, P., Peron, F., et al. (2003) Mass balance of free tropospheric aerosol at the Puy de D(o)over-capme (France) in winter. Journal of Geophysical Research-Atmospheres, 108: D11,4333. Sempere, R. and Kawamura, K. (2003) Trans-hemispheric contribution of C-2-C-10 alpha, omega-dicarboxylic acids, and related polar compounds to water-soluble organic carbon in the western Paciﬁc aerosols in relation to photochemical oxidation reactions. Global Biogeochemical Cycles, 17: 2,1069. Shimmo, M., Anttila, P., Hartonen, K., et al. (2004) Identiﬁcation of organic compounds in atmospheric aerosol particles by on-line supercritical ﬂuid extraction-liquid chromatography-gas chromatography-mass spectrometry. Journal of Chromatography A, 1022: 151–9. Sicre, M.A. and Peltzer, E.T. (2004) Lipid geochemistry of remote aerosols from the southwestern Paciﬁc Ocean sector. Atmospheric Environment, 38: 1615–24. Simoneit, B.R.T. (1999) A review of biomarker compounds as source indicators and tracers for air pollution. Environmental Science and Pollution Research, 6: 159–69. Simoneit, B.R.T. (2002) Biomass burning – A review of organic tracers for smoke from incomplete combustion. Applied Geochemistry, 17: 129–62. Simoneit, B.R.T., Cox, R.E. and Standley, L.J. (1988) Organic matter of the troposphere-IV: Lipids in Harmattan aerosols of Nigeria. Atmospheric Environment, 22: 983–1004. Simoneit, B.R.T., Schauer, J.J., Nolte, C.G., et al. (1999) Levoglucosan, a tracer for cellulose in biomass burning and atmospheric particles. Atmospheric Environment, 33: 173–82. Simoneit, B.R.T., Elias, V.O., Kobayashi, M., et al. (2004a) Sugars – Dominant water-soluble organic compounds in soils and characterization as tracers in atmospheric particulate matter. Environmental Science and Technology, 38 : 5939–49. Simoneit, B.R.T., Kobayashi, M., Mochida, M., et al. (2004b) Composition and major sources of organic compounds of aerosol particulate matter sampled during the ACE-Asia campaign. Journal of Geophysical Research-Atmospheres, 109: D19,D19S10. Sipin, M.F., Guazzotti, S.A. and Prather, K.A. (2003) Recent advances and some remaining challenges in analytical chemistry of the atmosphere. Analytical Chemistry, 75: 2929–40. Slanina, J., ten Brink, H.M., Otjes, R.P., et al. (2001) The continuous analysis of nitrate and ammonium in aerosols by the steam jet aerosol collector (SJAC): Extension and validation of the methodology. Atmospheric Environment, 35: 2319–30. Smith, D.F., McIver, C.D. and Kleindienst, T.E. (1998) Primary product distribution from the reaction of hydroxyl radicals with toluene at ppb NOX mixing ratios. Journal of Atmospheric Chemistry, 30: 209–28. Smith, D.F., Kleindienst, T.E. and McIver, C.D. (1999) Primary product distributions from the reaction of OH with m-, p-xylene, 1,2,4- and 1,3,5-trimethylbenzene. Journal of Atmospheric Chemistry, 34: 339–64. Stephanou, E.G. and Stratigakis, N. (1993) Determination of anthropogenic and biogenic organic compounds on airborne particles: Flash chromatographic fractionation and capillary gas chromatographic analysis. Journal of Chromatography A, 644: 141–51. Stern, J.E., Flagan, R.C., Grosjean, D. and Seinfeld, J.H. (1987) Aerosol formation and growth in atmospheric aromatic hydrocarbon photooxidation. Environmental Science and Technology, 21: 1224–31. Strader, R., Lurmann, F. and Pandis, S.N. (1999) Evaluation of secondary organic aerosol formation in winter. Atmospheric Environment, 33: 4849–63. Subramanian, R., Khlystov, A.Y., Cabada, J.C. and Robinson, A.L. (2004) Positive and negative artifacts in particulate organic carbon measurements with denuded and undenuded sampler conﬁgurations. Aerosol Science and Technology, 38: 27–48.

Organic Aerosols

385

Szidat, S., Jenk, T.M., Gaggeler, H.W., et al. (2004) Radiocarbon (C-14)-deduced biogenic and anthropogenic contributions to organic carbon (OC) of urban aerosols from Zurich, Switzerland. Atmospheric Environment, 38: 4035–44. Szponar, B. and Larsson, L. (2001) Use of mass spectrometry for characterising microbial communities in bioaerosols. Annals of Agricultural and Environmental Medicine, 8: 111–17. Takehara, H., Fujiwara, M., Arikawa, M., Diener, M.D. and Alford, J.M. (2005) Experimental study of industrial scale fullerene production by combustion synthesis. Carbon, 43: 311–19. Takekawa, H., Minoura, H. and Yamazaki, S. (2003) Temperature dependence of secondary organic aerosol formation by photo-oxidation of hydrocarbons. Atmospheric Environment, 37: 3413–24. Thompson, D.S., Schein, M.E. and Murphy, D.M. (2000) Particle analysis by laser mass spectrometry WB-57F instrument overview. Aerosol Science and Technology, 33: 153–69. Tobias, H.J. and Ziemann, P.J. (1999) Compound identiﬁcation in organic aerosols using temperatureprogrammed thermal desorption particle beam mass spectrometry. Analytical Chemistry, 71: 3428–35. Tobias, H.J., Kooiman, P.M., Docherty, K.S. and Ziemann, P.J. (2000) Real-time chemical analysis of organic aerosols using a thermal desorption particle beam mass spectrometer. Aerosol Science and Technology, 33: 170–90. Tremp, J., Mattrel, P., Fingler, S. and Giger, W. (1993) Phenol and nitrophenols as tropospheric pollutants: Emissions from automobile exhausts and phase transfer in the atmosphere. Water Air and Soil Pollution, 68: 113–23. Tsai, F.C., Apte, M.G. and Daisey, J.M. (2000) An exploratory analysis of the relationship between mortality and the chemical composition of airborne particulate matter. Inhalation Toxicology, 12: 121. Tsigaridis, K. and Kanakidou, M. (2003) Global modelling of secondary organic aerosol in the troposphere: A sensitivity analysis. Atmospheric Chemistry and Physics, 3: 1849–69. Turpin, B.J. and Huntzicker, J.J. (1995) Identiﬁcation of secondary organic aerosol episodes and quantiﬁcation of primary and secondary organic aerosol concentrations during SCAQS. Atmospheric Environment, 29: 3527–44. Turpin, B.J. and Lim, H.J. (2001) Species contributions to PM2.5 mass concentrations: Revisiting common assumptions for estimating organic mass. Aerosol Science and Technology, 35: 602–10. Turpin, B.J., Huntzicker, J.J., Larson, S.M. and Cass, G.R. (1991) Los Angeles summer midday particulate carbon: Primary and secondary Aerosol. Environmental Science and Technology, 25: 1788–93. Turpin, B.J., Saxena, P. and Andrews, E. (2000) Measuring and simulating particulate organics in the atmosphere: Problems and prospects. Atmospheric Environment, 34: 2983–3013. Viidanoja, J., Sillanpaa, M., Laakia, J., et al. (2002) Organic and black carbon in PM2.5 and PM10: 1 year of data from an urban site in Helsinki, Finland. Atmospheric Environment, 36: 3183–93. Wal, R.L.V. and Tomasek, A.J. (2004) Soot nanostructure: Dependence upon synthesis conditions. Combustion and Flame, 136: 129–40. Wang, S.C., Paulson, S.F., Grosjean, D., Flagan, R.C. and Seinfeld, J.H. (1992) Aerosol formation and growth in atmospheric organic/NOx systems I. Outdoor smog chamber studies of C7- and C8-hydrocarbons. Atmospheric Environment, 26A: 403–20. Ward, T.J., Hamilton, R.F. and Smith, G.C. (2004) The Missoula, Montana PM2.5 speciation study – Seasonal average concentrations. Atmospheric Environment, 38: 6371–9. Warneck, P. (2003) In-cloud chemistry opens pathway to the formation of oxalic acid in the marine atmosphere. Atmospheric Environment, 37: 2423–7.

386

Volatile Organic Compounds in the Atmosphere

Warnke, J. (2004) Identiﬁcation and Quantiﬁcation of Acidic Marker Compounds for Tropospheric Secondary Organic Aerosol from Biogenic Hydrocarbons by Capillary-HPLC-ESI-MSn . PhD Thesis, University of Mainz, Germany. Warnke, J., Bandur, R. and Hoffmann, T. (2006) Capillary-HPLC-ESI-MS/MS method for the determination of acidic products from the oxidation of monoterpenes in atmospheric aerosol samples. Analytical and Bioanalytical Chemistry, 385: 34–45. Weissenbok, R.H., Currie, L.A., Grollert, C., et al. (2000) Accelerator mass spectrometry analysis of non-soluble carbon in aerosol particles from high alpine snow (Mt. Sonnblich, Austria). Radiocarbon, 42: 285–94. Went, F.W. (1960) Blue hazes in the atmosphere. Nature, 187: 641–3. Wentzel, M., Gorzawski, H., Naumann, K.H., Saathoff, H. and Weinbruch, S. (2003) Transmission electron microscopical and aerosol dynamical characterization of soot aerosols. Journal of Aerosol Science, 34: 1347–70. Winterhalter, R., Van Dingenen, R., Larsen, B.R., Jensen, N.R. and Hjorth, J. (2003) LC-MS analysis of aerosol particles from the oxidation of α-pinene by ozone and OH-radicals. Atmospheric Chemistry and Physics Discussions, 3: 1–39. Yao, X.H., Fang, M., Chan, C.K., Ho, K.F. and Lee, S.C. (2004) Characterization of dicarboxylic acids in PM2.5 in Hong Kong. Atmospheric Environment, 38: 963–70. Yokouchi, Y. and Ambe, Y. (1985) Aerosols formed from the chemical reaction of monoterpenes and ozone. Atmospheric Environment, 19: 1271–6. −COOH, − −OH, Yu, J.Z., Flagan, R.C. and Seinfeld, J.H. (1998) Identiﬁcation of products containing − −C==O in atmospheric oxidation of hydrocarbons. Environmental Science and Technology, and − 32: 2357–70. Yu, J.Z., Cocker, D.R., Grifﬁn, R.J., Flagan, R.C. and Seinfeld, J.H. (1999a) Gas-phase ozone oxidation of monoterpenes: Gaseous and particulate products. Journal of Atmospheric Chemistry, 34: 207–58. Yu, J.Z., Grifﬁn, R.J., Cocker, D.R., Flagan, R.C., Seinfeld, J.H. and Blanchard, P. (1999b) Observation of gaseous and particulate products of monoterpene oxidation in forest atmospheres. Geophysical Research Letters, 26: 1145–48. Yu, J.Z., Huang, X.F., Xu, J.H. and Hu, M. (2005a) When aerosol sulfate goes up, so does oxalate: Implication for the formation mechanisms of oxalate. Environmental Science and Technology, 39: 128–33. Yu, L.E., Shulman, M.L., Kopperud, R. and Hildemann, L.M. (2005b) Characterization of organic compounds collected during Southeastern Aerosol and Visibility Study: Water-soluble organic species. Environmental Science and Technology, 39: 707–15. Yu, S.C., Dennis, R.L., Bhave, P.V. and Eder, B.K. (2004) Primary and secondary organic aerosols over the United States: Estimates on the basis of observed organic carbon (OC) and elemental carbon (EC), and air quality modeled primary OC/EC ratios. Atmospheric Environment, 38: 5257–68. Zdrahal, Z., Oliveira, J., Vermeylen, R., Claeys, M. and Maenhaut, W. (2002) Improved method for quantifying levoglucosan and related monosaccharide anhydrides in atmospheric aerosols and application to samples from urban and tropical locations. Environmental Science and Technology, 36: 747–53. Zhang, Q., Worsnop, D.R., Canagaratna, M.R. and Jimenez, J.L. (2005a) Hydrocarbon-like and oxygenated organic aerosols in Pittsburgh: Insights into sources and processes of organic aerosols. Atmospheric Chemistry and Physics, 5: 3289. Zhang, Q., Canagaratna, M.R., Jayne, J.T., Worsnop, D.R. and Jimenez, J.L. (2005b) Time- and sizeresolved chemical composition of submicron particles in Pittsburgh: Implications for aerosol sources and processes. Journal of Geophysical Research-Atmospheres, 110: D07S09. Zhang, S.H., Shaw, M., Seinfeld, J.H. and Flagan, R.C. (1992) Photochemical aerosol formation from alpha-pinene and beta-pinene. Journal of Geophysical Research, 97: 20717–29.

Organic Aerosols

387

Zheng, M., Wan, T.S.M., Fang, M. and Wang, F. (1997) Characterization of the non-volatile organic compounds in the aerosols of Hong Kong – Identiﬁcation, abundance and origin. Atmospheric Environment, 31: 227–37. Zheng, M., Fang, M., Wang, F. and To, K.L. (2000) Characterization of the solvent extractable organic compounds in PM2.5 aerosols in Hong Kong. Atmospheric Environment, 34: 2691–702.

Volatile Organic Compounds in the Atmosphere Edited by Ralf Koppmann Copyright © 2007 by Blackwell Publishing Ltd

Chapter 10

Gas Chromatography–Isotope Ratio Mass Spectrometry Jochen Rudolph

10.1

Introduction

Although measurement of stable isotope ratios is by now an established tool to gain additional insight into the processes determining the distribution, sources and sinks of important atmospheric trace gases such as carbon dioxide, carbon monoxide, methane, hydrogen, nitrous oxide and ozone (Allan et al. 2001; Brenninkmeijer and Röckmann 1997; Brenninkmeijer et al. 1995; Ciais et al. 1995, 1997; Conny and Currie 1996; Conny et al. 1997; Craig et al. 1988; Francey et al. 1995; Fung et al. 1997; Gerst and Quay 2000, 2001; Gupta et al. 1996; Keeling et al. 1979; Lowe et al. 1994, 1999; Manning 1999; Manning et al. 1997; Mook et al. 1983; Stevens and Wagner 1989) the application of isotope ratio studies to atmospheric volatile organic compounds (VOCs) is a relatively recent development. The ﬁrst paper describing a method for compound-speciﬁc measurements of the stable carbon isotope ratios of VOCs was published in 1997 (Rudolph et al. 1997), but it took several years before more detailed results from applications of this method were published (Rudolph and Czuba 2000; Rudolph et al. 2000; Tsunogai et al. 1999). During the last few years the number of publications studying the various aspects of stable carbon isotope ratios of atmospheric VOCs has been steadily increasing, although in total the number of studies is still quite limited. Goldstein and Shaw (2003) recently reviewed available information on isotope ratios of VOCs and the possibility of using isotope ratio measurements for studying their atmospheric budgets and chemistry. There are several reasons for the relatively small number of applications of isotope ratios in studies of atmospheric VOCs. First, the measurement procedure is quite demanding and the required instrumentation expensive. Second, interpretation of measured atmospheric stable carbon isotope ratios of VOCs requires knowledge of information such as the isotopic composition of VOC emissions and the kinetic isotope effects of gas-phase reactions of VOC, which is not yet available or has only very recently become available. Finally, the conceptual framework and basic understanding necessary for the evaluation of stable carbon isotope ratios of atmospheric VOCs is just emerging. One of the consequences of the still very limited amount of information available for the interpretation of studies of stable isotope ratios of atmospheric VOCs is a presently still rather high uncertainty of conclusions based on such studies. The following chapter will therefore not only give an overview of the state of the art and the fundamental concepts,

Gas Chromatography–Isotope Ratio Mass Spectrometry

389

but also try to identify the most important gaps in the knowledge required to make full use of the potential of compound-speciﬁc stable isotope ratio measurements for studies of atmospheric VOCs. Finally, it should be mentioned that this chapter will mainly concentrate on stable carbon isotope ratios, although, in principle, hydrogen, and in the case of compounds with hetero atoms isotope ratios of other elements, for example chlorine, sulphur or oxygen may be studied. The reason is simply that presently most of the available studies concentrate on carbon. This is predominantly due to experimental reasons, but to some extent also due to the still limited knowledge of the basic framework needed for the interpretation of such measurements.

10.2 10.2.1

Fundamentals of stable isotope ratios of VOCs Deﬁnitions

In principle, substances with identical chemical structure containing different isotopes of one or more of its atoms, the isotopologues, can be treated as individual compounds. However, for a number of reasons, it is customary to use isotope ratios, which is the ratio of the number of atoms of one isotope over that of another. For example, in the case of carbon, the isotope ratio is expressed as the ratio of the number or concentration of 13 C over 12 C atoms (13 C/12 C). Moreover, for experimental and practical reasons, isotope ratios are generally given relative to a reference point and, since changes in isotope ratios are nearly always small, they are usually expressed as per mille values. For carbon, the internationally accepted reference is the Vienna-Peedee Belemnite scale (V-PDB) with an isotope ratio of (13 C/12 C)V-PDB = 0.0112372 (Craig 1957). The so-called delta value (δ 13 C) is deﬁned as follows: (13 C/12 C)sample − (13 C/12 C)V-PDB δ 13 C = × 1000‰ (10.1) (13 C/12 C)V-PDB Rearranging Equation 10.1 allows calculation of the isotope ratio from the delta value: 13 13 13 C C C δ 13 C × 12 = + 12 (10.2) 12 C 1000‰ C V-PDB C V-PDB sample For example, a δ 13 C value of 1‰ in the V-PDB scale corresponds to an isotope ratio of 0.01123. Obviously, the use of the delta notation is far more convenient than the usage of absolute isotope ratios. It has become customary to refer to samples with higher 13 C/12 C values, which corresponds to larger delta values, as ‘heavier’ and to samples with lower 13 C/12 C values as ‘lighter’. For other elements, the usage and deﬁnitions of isotope ratios and the delta convention follow the same principle, although deﬁnitions and values for the reference points are different. For stable hydrogen isotope ratios (D/H), the reference point is Vienna-Standard Mean Ocean Water (V-SMOW) with a D/H ratio of 0.00015576 (Hagemann et al. 1970). When looking at individual compounds with more than one carbon atom, there are potentially several different isotopologues. For a compound with NC C-atoms, hypothetically the number of 12 C-atoms replaced by a 13 C-atom can range from zero to NC , resulting

390

Volatile Organic Compounds in the Atmosphere

potentially in NC + 1 molecules with a different degree of labelling. Furthermore, labelling may occur at different positions of the molecule, which may result in a number of different isotopomers for the individual isotopologues. Fortunately for atmospheric VOCs, and many other isotope applications, this quite complex situation has often only little consequence for practical applications. Since the number of C-atoms in atmospheric VOCs is generally small and the natural isotope abundance of 13 C is only slightly larger than 1%, the probability of having more than one 13 C-atom in any VOC molecule with natural isotope abundance is very low and therefore can generally be ignored without introducing bias. Similarly, isotope effects are nearly always small and, therefore, it is a reasonable ﬁrst-order approximation to assume an equal, random distribution of the 13 C atoms in any given molecule. It should also be noted that currently available measurement techniques for isotope ratios of atmospheric VOC do not allow differentiation between molecules with labelling at different sites. Nevertheless, there is potential additional information in site-speciﬁc isotope fractionation (see, e.g. Section 10.4.7) and therefore the development of such measurement techniques would very likely add additional value to the use of isotope ratio studies of atmospheric VOCs. For atmospheric VOCs, it is thus possible to approximate the ratio of the number or concentration of molecules containing carbon 13 C-atom (13 C-VOC) over those containing only 12 C-atoms (12 C-VOC) by the measured carbon isotope ratio of the studied VOCs multiplied with the number of carbon atoms in one molecule (NC ). 13 C [13 C-VOC] = NC 12 (10.3) [12 C-VOC] C Here, NC is the number of carbon atoms of the molecule.

10.2.2

Isotope effects

Isotopologues and isotopomers have different physical and chemical properties, although the differences are usually only small. This applies to kinetic as well as to equilibrium processes. The atmospheric chemistry of VOCs is, with extremely few exceptions, determined by kinetically limited reactions. The following chapter will, therefore, concentrate on the effect of isotopic labelling on reaction kinetics. The kinetic isotope effects (KIE) of a reaction is generally expressed as ratio of the rate constants for labelled and unlabelled compounds. For example, the rate constant of 13 C labelled benzene with 13 the OH-radical ( C6 HC6 kOH ) is usually given relative to the reaction rate constant of unlabelled 12

benzene ( C6 HC6 kOH ) as the ratio of the two rate constants. C C6 H6 KIEOH

=

12 C C6 H6 kOH 13 C C6 H6 kOH

(10.4)

The convention is to use the ratio of the rate constant for the lighter isotopologue over that of the heavier isotopologue. Since the lighter isotopologue reacts faster in most cases, KIEs larger than unity are called normal isotope effects; if the reaction favours the heavy isotopologue the term ‘inverse isotope effect’ is used. In the formula above, the subscripts and superscripts are used to identify the reacting VOC and the reactant, as well as the isotopes. For simplicity, in this chapter, the indices will not be given if the type of reaction

Gas Chromatography–Isotope Ratio Mass Spectrometry

391

or isotope effect is obvious from the context. For example, Equation 10.4 will be written in the following form. KIE =

12 C 13 C

k k

(10.4a)

Most KIEs for reactions of atmospheric VOCs are small and therefore the ratio of the rate constants is generally close to unity. For the sake of convenience isotope effects are therefore often presented and used in the so-called epsilon notation as per mille values representing the relative difference between the rate constants. For our example of the reaction of benzene with OH, the ratio of the rate constants for the two isotopologues is 1.0075 (Anderson et al. 2004a). The corresponding ε value is 7.5‰ according to the deﬁnition:

12 12 C 13 Ck k − Ck × 1000‰ = 13 − 1 × 1000‰ = (KIE − 1) × 1000‰ εOH = 13 C Ck k (10.5) The use of ε to express isotope effects has not only the advantage of being more convenient, it also allows simple formulations relating isotope ratios in δ notation with the isotope fractionation occurring during chemical reactions (see Section 10.2.3). It should be noted that in publications on isotope ratio studies the isotope fractionation factor α is often used, which is the inverse of the KIE as deﬁned by Equation 10.4. In this chapter, the deﬁnition given by Equation 10.4 will be used, which follows the convention widely applied in publications presenting reaction kinetic studies of isotope labelled compounds. In the following parts of this chapter, the conversion of δ or ε values given in per mille will not be included explicitly, but it is important for any calculation that includes values given in per mille notation to remember that x‰ = x × 10−3 .

10.2.3

Concepts for interpretation of stable isotope ratios of atmospheric VOCs

The stable isotope ratios of atmospheric VOCs are determined by the isotope ratios of the emissions into the atmosphere and the isotope fractionation associated with the atmospheric loss processes. The obvious consequence is that studies of VOC isotope ratios can provide additional information about both sources and processes. In this fundamental aspect, isotope ratios of atmospheric VOCs are not different from other applications of isotope ratio measurements. However, the situation in the atmosphere is very complex; typically, the isotope ratio of VOCs is the result of mixing air masses containing emissions from a variety of sources with different degrees of photochemical processing. Furthermore, atmospheric mixing processes as well as removal rates of VOC have a substantial spatial and temporal variability, which make the determination of representative averages for concentrations and isotope ratios often very difﬁcult. Nevertheless, depending on speciﬁc conditions simpliﬁed concepts can be used to identify the key processes determining VOCs isotope ratios. However, generally the applicability of a speciﬁc concept depends critically on the validity of the underlying assumptions. It is important to realise that one of the consequences of the presently very limited amount of

392

Volatile Organic Compounds in the Atmosphere

information on isotope ratios of atmospheric VOC is that the validity of many assumptions often cannot be veriﬁed beyond reasonable doubt.

10.2.3.1

Two-endpoint mixing

In the absence of reactions of VOCs the atmospheric stable isotope ratio (δAtm ) of a compound with concentration cAtm can be described by a simple relation based on mixing contributions (ci ) from different sources i with individual isotope ratios δi . ci · δ i i ci · δi δAtm = = i (10.6) c c Atm i i In the case of mixing air masses with different VOC concentrations and isotope ratios the contribution ci from a speciﬁc air mass with a VOC concentration of m ci depends on the fraction of the volume, fi = Vi / Vi , contributed by the air mass such that ci = m ci × fi . Therefore for the mixing of air masses with different VOC concentrations and isotope ratios Equation 10.6 can be written as m ci · fi · δi i m ci · fi · δi = i (10.7) δAtm = c · f cAtm im i i Although strictly speaking in the atmosphere the assumption of a total absence of VOC reactions is not realistic, there are many situations for which the changes in VOC concentrations are dominated by mixing processes. Examples are vertical mixing by the break up of the night time inversion layer, the dilution of an urban or industrial plume, or the injection of fresh emissions into a well mixed air mass. In such cases, very often a reasonable description can be based on the mixing of two air masses with concentrations c1 and c2 and isotope ratios δ1 and δ2 . cAtm = c1 · f1 + c2 · f2 = c1 · f1 + c2 · (1 − f1 )

(10.8)

and δAtm =

δ1 · c1 · f1 + δ2 · c2 · f2 δ1 · c1 · f1 + δ2 · c2 · (1 − f1 ) = cAtm cAtm

(10.9)

Combining Equations 10.8 and 10.9 allows describing the isotope ratio as function of concentration. δ1 · c1 · (cAtm − c2 )/(c1 − c2 ) + δ2 · c2 · (1 − ((cAtm − c2 )/(c1 − c2 ))) δAtm = cAtm (10.10) The two endpoints of the function described by Equation 10.10 are given by the concentrations and stable carbon isotope ratios of the individual air masses. Such types of dependencies are therefore often called two endpoint mixing curves. A very simple dependence can be derived for the speciﬁc case that one of the concentrations in one air mass is much larger than in the other (c1 c2 ) and that the contribution of this air mass to the total volume is only small. δAtm = δ1 +

(δ2 − δ1 ) · c2 cAtm

(10.11)

Gas Chromatography–Isotope Ratio Mass Spectrometry

393

−18

13C (‰)

−20 −22 −24 −26 −28 0

1 2 Inverse of mixing ratio (ppb−1)

3

Figure 10.1 Plot of the stable carbon isotope ratio of benzene vs its inverse mixing ratio for measurements at a suburban location in the Greater Toronto Area in summer (squares). The data points and their uncertainties are taken from Rudolph (2002). The three lines represent different two-endpoint mixing models. The solid line describes mixing of two air masses according to Equation 10.11. The concentrations and stable carbon isotope ratio of the two air masses are 5 ppb and 0.3 ppb, and −26.5‰ and −21.5‰, respectively. The intercept is −26.8‰. The two other lines represent linear least square ﬁts to all data (short dashed line) and to data points representing concentrations exceeding 0.5 ppb (long dashed line). The y -axis intercepts are −26.4 ± 0.4‰ and −27.5 ± 0.4‰, respectively.

In this case δ1 can be determined from a plot of δAtm vs the inverse of the concentration. An example for such behaviour is the change of the isotope ratio due to VOC emissions or mixing of highly polluted urban air with background air. This type of plot, which is based on the same principle as plots widely used to eliminate the impact of blank values (Keeling plot) for isotope ratio measurements, can be very useful to determine the isotope ratio of emissions in cases where the background concentration of the studied VOC is not negligible. The extrapolation to concentrations where the background no longer has a visible impact on the isotope ratio (1/c1 = 0) gives an estimate of the average isotope ratio of the emissions. It should be noted that Equation 10.10 also describes a linear dependence between inverse concentration and isotope ratio. This is in contrast to Equation 10.11 with a y-axis intercept which depends on the concentration and stable carbon isotope ratio of both air masses. An example comparing two-endpoint mixing curves with atmospheric observations for benzene is shown in Figure 10.1. It is obvious that to some extent the result of the extrapolation depends on the mixing model used. However, the results of the different extrapolations are all in the range of −27 ± 1‰, which is in very good agreement with the isotope ratio of benzene emissions (Rudolph et al. 2002). Nevertheless, it is important to realise, that, due to the complexities of atmospheric mixing and the inherent assumption of the absence of loss reactions the use of two-endpoint mixing curves in the atmosphere can only be considered a ﬁrst-order approximation.

10.2.3.2

Rayleigh fractionation

The opposite of the simpliﬁed case of mixing without reaction is the idealisation of the complete absence of mixing and dilution. In this case the isotope ratios of an atmospheric

394

Volatile Organic Compounds in the Atmosphere

VOC change only as a consequence of the different reaction rates for different isotopologues. In principle the rate of change of concentration of a light (L c) and heavy VOC (Hc) isotopologue due to reaction with reactants x of concentration cx is described by the following differential equations: dL c L kx · cx = −L c · dt x

(10.12a)

dH c H = −H c · kx · cx dt x

(10.12b)

Here, H kx and L kx are the rate constants for the reaction of reactant x with the heavy and light VOC isotopologue, respectively. The integration for a time interval between zero and t yields the following relation between the initial concentration (0 c) and the concentration at time t (t c): ⎛ ⎞ t L L L ⎝ kx · cx dt ⎠ (10.13a) t c = 0 c · exp − x

and

⎛ H t c

⎝ =H 0 c · exp −

0

⎛ ⎝H kx ·

x

Replacing

Hk x

=

Lk

t

⎞⎞ cx dt ⎠⎠

(10.13b)

0

−1 x · KIEx

(see Equation 10.4), we obtain: ⎞⎞ t H H ⎝ ⎝L kx · KIEx−1 · cx dt ⎠⎠ t c = 0 c · exp − ⎛

⎛

x

(10.13c)

0

The change in isotope ratio can then be derived by combining Equations 10.13a and 10.13c: ⎛ ⎞⎞ ⎛ t Hc Hc t ⎝L kx · (KIEx−1 − 1) · cx dt ⎠⎠ = 0L · exp ⎝− (10.14a) Lc c t 0 x 0

or:

H t Hc t c 0 L −1 ln L kx · (KIEx − 1) · cx dt = ln L − tc 0c x

(10.14b)

0

For atmospheric VOCs, often one loss mechanism, the reaction with the OH-radical, dominates. In this case Equation 10.14 can be simpliﬁed:

H t Hc t c −1 0 L − kOH · (KIEOH − 1) · cOH dt = ln L ln L tc 0c 0

(10.15a)

Gas Chromatography–Isotope Ratio Mass Spectrometry

Since −L kOH

395

t

= ln(Lt c/L0 c):

H Hc Lc t c −1 0 = ln L ln L − 1) + ln tL · (KIEOH c c c t 0 0 0 cOH dt

(10.15b)

Using δ notation to express the isotope ratios a Rayleigh-type equation is obtained, which describes the dependence between concentration change and change in isotope ratio for a closed, homogeneously mixed system.

Lc −1 · (KIEOH ln(1 + t δ) = ln(1 + 0 δ) + ln tL − 1) (10.16a) c 0 or:

tδ

= (1 + 0 δ) ·

Lc t Lc 0

(KIE −1 −1) OH

− 1 = (1 + 0 δ) ·

Lc t Lc 0

−εOH /(1+εOH ) −1

(10.16b)

Here, t δVOC and 0 δVOC are the isotope ratios at time t and t = 0, respectively. For small δ values, ln(1 + δ) ≈ δ and we obtain the simple dependence:

Lc t −1 − 1) (10.16c) · (KIEOH t δ = 0 δ + ln L 0c Since for most cases that are relevant for atmospheric VOCs the isotope ratios and the isotope fractionation effects are small, the concentrations of the light isotopologue can be replaced by the VOC concentration, which, strictly speaking, is the sum of the concentration of all isotopologues, without introducing bias. −1 − 1 in Equation 10.16 has to In case of more than one loss process for the VOC, KIEOH be replaced by a weighted average:

Lc t {fx · (KIEx−1 − 1)} (10.16d) · t δ = 0 δ + ln L 0c x The weighting factors fx are deﬁned such that x fx = 1. t They can be derived from the average concentration of the reactant, av cx = 0 cx dt /t and the rate constant (kx ): fx = (kx · av cx )/( x kx · av cx ) and Equation 10.16c becomes:

Lc {kx · av cx · (KIEx−1 − 1)} t (10.16e) · x t δ = 0 δ + ln L x kx · av cx 0c Since the atmospheric residence time (τx ) of VOC with respect to loss due to reaction with a speciﬁc reactant X is given by τx = (kx · av cx )−1 , Equation 10.16d can also be written as:

−1 Lc Lc ((KIEx−1 − 1)/τx ) t x ((KIEx − 1)/τx ) · = 0 δ + ln tL · x −1 t δ = 0 δ + ln L −1 τLoss 0c 0c x τx (10.16f ) where τLoss = ( x τx−1 )−1 is the VOC lifetime considering all chemical loss processes.

396

Volatile Organic Compounds in the Atmosphere

Based on Equation 10.5, the expression KIE −1 − 1 = (1/(1 + ε)) − 1, which for small values of ε can be approximated by using (1/(1 + ε)) ≈ 1 − ε and then KIE −1 − 1 ≈ −ε can be used to obtain a simple approximation:

Lc (εx /τx ) t (10.17) × x −1 t δ ≈ 0 δ − ln L τLoss 0c

−17.5 10

13C (‰)

−20.0

8 6

−22.5

4 −25.0

2

− 27.5

0 0

0.5

1 1.5 Mixing ratio (ppb)

Photochemical age, 1011OH/cm3 s

For large values of ε or for the interpretation of very high precision isotope data, Equation 10.17 may not be adequate, and Equation 10.16 has to be used. However, compared to the accuracy of presently available atmospheric VOCs isotope ratio data and KIEs for the reaction of atmospheric VOCs the bias due to the approximation KIE −1 − 1 ≈ −ε is small. Although it has been shown that under speciﬁc conditions a Rayleigh-type function may be a reasonable approximation for the dependence between concentration and isotope ratio (Thompson et al. 2003) and has been used to determine kinetic isotope effects for reactions of VOCs from atmospheric observations (Tsunogai et al. 1999), the assumption of no dilution or mixing generally does not provide a reasonable description of the processes determining atmospheric isotope ratios. Yet, a useful application of the Rayleigh dependence for the interpretation of atmospheric VOC isotope ratio measurements is the conceptual differentiation between the impacts of mixing and chemical processing. This is demonstrated in Figure 10.2 using the same data set as in Figure 10.1. For benzene the only atmospheric loss process is reaction with the OH-radical, for which the KIE (ε = 7.5‰) is well known (Anderson et al. 2004a; Rudolph et al. 2000). It is obvious that a Rayleigh-type dependence with ε = 7.5‰ predicts a much stronger change of isotope ratio as function of concentration than observed. However,

2

Figure 10.2 Comparison of observed dependence between mixing ratio and stable carbon isotope ratio of benzene (full circles, same data set as Figure 10.1) and calculated Rayleigh type dependencies. The solid line shows the Rayleigh dependence calculated for the KIE of the benzene+OH reaction determined in the laboratory (ε = 7.5‰), the broken line shows a Rayleigh function with ε = 2.14‰. The secondary y-axis shows the photochemical age corresponding to the stable carbon isotope ratios of benzene calculated according to Equation 10.19.

Gas Chromatography–Isotope Ratio Mass Spectrometry

397

a surprisingly good description of the observations can be achieved with a Rayleigh function using ε = 2.14‰ (standard error of ε is 0.25‰, R 2 = 0.886). Conceptually the dependence between isotope ratio and concentration can be separated into a set of chemical reaction with an overall two processes, isotope effect εLoss = ( x kx · av cx · εx )/( x kx · av cx ) and a time constant τLoss = ( x kx · av cx )−1 , and a dilution process with a time constant τM , which does not result in an isotope fractionation (M ε = 0). In a formalistic approach the Rayleigh-type dependence, which is strictly valid for a closed system only, can be extended to include such a dilution term in εEff : εEff =

[(εLoss /τLoss )]

(10.18)

−1 −1 τLoss + τM

In order to be compatible with εEff = 2.14‰, τM has to be a factor of 2.5 smaller than τLoss ; that is for our example dilution is 2.5 times faster than the chemical loss process. Both the comparison with a Rayleigh-type function and interpretation of the observations as two-endpoint mixing behaviour demonstrate that mixing and dilution processes have a substantial impact on the dependence between isotope ratios and concentrations of VOCs. However, the underlying concepts are different. In one case the change in isotope ratio is due to mixing of air masses with different isotope ratios. In the other, it is explained by the isotope effects associated with the atmospheric VOC loss reactions. Unfortunately, as can be seen from our example, atmospheric observations of isotope ratios and concentrations do not necessarily allow differentiation between the different processes without additional information.

10.2.3.3

Isotope ratios and photochemical age

As shown above, in the atmosphere the interaction between emissions, atmospheric transport and mixing, and chemical reactions result in complex dependencies between VOC concentration and isotope ratio, which often cannot be interpreted unambiguously. However, in the simpliﬁed case of emissions with a uniform isotope ratio, any change in atmospheric isotope ratio can ultimately only be due to the isotope fractionation associated with reactions in the atmosphere. Rudolph and Czuba (2000) have shown that in this case a simple relation between isotope ratio and chemical processing exists. For many VOCs the dominant atmospheric loss process is reaction with the OH-radical and the dependence between isotope ratio and chemical processing then is given by t tδ

= 0 δ + εOH · kOH · t

cOH × dt = 0 δ + εOH · kOH · av cOH · t

(10.19a)

0

Here av cOH = ( 0 cOH × dt )/t is deﬁned as the average concentration of the OH-radical andt is the time between emission (t = 0) of the studied VOC and observation. Rearranged t for 0 cOH × dt = av cOH · t we obtain: av cOH

·t =

t δ − 0δ εOH · kOH

(10.19b)

In the case of mixing different air masses with different extent of photochemical VOC processing, there is no clearly deﬁned dependence between VOC concentration and isotope

398

Volatile Organic Compounds in the Atmosphere

ratio. However, Equation 10.19 can still be applied to atmospheric conditions where complex mixing of air masses occurs. In this case, Equation 10.19 allows calculation of the VOC concentration-weighted average OH-radical concentration: ti i cVOC i=n i cOH i=1 i cVOC 0 i=1

i=n

· dti =

ti i=n i cVOC i=1

cVOC

0

i cOH

· dti =

t δ − 0δ εOH · kOH

(10.19c)

Here, the index i indicates air parcel i and i cVOC is the contribution of air parcel i to the total VOC concentration: cVOC = i=n i=1 i cVOC . Although this describes a rather complex dependence between VOC isotope ratio and OH-radical concentration, it can provide very useful constraints on the possible origin of VOC observed in the atmosphere. t It has become customary to use the term ‘photochemical age’ (Parrish et al. 1992) for 0 cOH · dt = av cOH · t , to differentiate it from the physical age t . The use of isotope ratios to determine photochemical age is based on the same principle as the use of the ratios of VOC concentrations as described by Roberts et al. (1984), Rudolph and Johnen (1990), McKenna et al. (1995, 1997) and others. However, since most KIEs are small, the bias in photochemical ages, which occurs if air masses with different photochemical ages are mixed (McKeen et al. 1990, 1996), is generally negligible for photochemical ages derived from isotope ratios. Rudolph and Czuba (2000) demonstrated that for small KIEs, such as the carbon KIEs for reactions of atmospheric hydrocarbons with the OH-radical, Equation 10.19b gives a valid concentration-weighted average for the photochemical age, even if the studied VOC isotope ratio is the result of mixing of air parcels with an extremely wide range of different photochemical ages. Similarly, it has been shown that for photochemical ages based on hydrogen isotope ratios of atmospheric alkanes, the bias from mixing air masses will be small compared to the uncertainties due to the isotope ratio measurements (Iannone et al. 2004). The reason why isotope ratio-derived photochemical ages are less sensitive to bias from mixing of air masses is the very small error caused by approximating exponential changes in concentration ratios by a linear dependence if the changes are very small. This principle would also apply for atmospheric VOCs with similar atmospheric lifetimes. However, apart from the limited number of VOC pairs with nearly identical reactivity, there are many factors that will result in a small, seemingly random variability of VOC ratios, including the reproducibility of VOC concentration measurements. It thus follows that, for studies of photochemical aging, one of the key advantages of using isotope ratios is the possibility to determine small changes with high accuracy. As an example for the relation between isotope ratios and photochemical age, a photochemical age axis, which is derived by conversion from the isotope ratio scale using Equation 10.19, is included in Figure 10.2. It can be seen that meaningful changes in average photochemical age can be derived from changes in isotope ratios in the per mille range. Measurements of VOC concentrations with a reproducibility that would allow detection of changes in the ratio of VOC concentrations with such accuracy are out of the reach of presently available VOC measurement techniques. There seems to be considerable ambiguity in the interpretation of isotope ratios in terms of mixing or photochemical aging. However, when interpreting photochemical ages derived from isotope ratios, we have to remember that the photochemical age is that of the studied VOC, not that of the studied air mass. Indeed, Rudolph et al. (2000) and Saito et al. (2002) have shown that for the same air mass the photochemical ages of different VOCs can differ

Gas Chromatography–Isotope Ratio Mass Spectrometry

399

substantially. The reason is that, as mentioned above, the average photochemical age is the concentration-weighted mean. This can be shown by comparing a formulation describing the relation between average isotope ratio, av δVOC in an air mass resulting from a mixture of air parcels with different photochemical ages and the average photochemical age. In the case of non-uniform isotope ratios of the VOC emissions, Equation 10.19c. has to be modiﬁed: av

δ=

i=n

fi · (i0 δ + εOH · kOH · i cOH · ti )

(10.20a)

i=1

t For the sake of convenience here, 0 i i cOH dti ≡ i cOH · ti and fi = (i c/c) is used. The average isotope ratio of the emissions and the average photochemical age are then deﬁned as follows: av 0 δ

=

i=n

fi · i0 δ

(10.20b)

i=1

and

− av 0 δ fi · cOH · ti = fi · av cOH · t = εOH · kOH i=1 i=1

i=n i=n 1 · fi · av fi · i0 δ = t δ− εOH · kOH i=n

i=n

i

i=1

av δ t

(10.20c)

i=1

From Equations 10.19 and 10.20, it can be seen that in the case of mixing of air masses with different isotope ratios, additional information is needed to ﬁnd out whether changes in isotope ratios are due to varying impact of sources having different isotope ratios or changes in extent of photochemical processing. Such additional evidence can be derived from comparing photochemical ages for different compounds. In case of photochemical aging, we expect a considerable degree of correlation between the isotope ratios of different compounds. Although it cannot be excluded that variations in the isotope ratio of emissions are also correlated, it is not very likely that such a correlation will follow the dependence predicted by Equation 10.19. An example is shown in Figure 10.3. For hydrocarbons with similar atmospheric residence times, their photochemical ages agree within a limited amount of scatter; for example, the photochemical ages of n-butane and n-pentane are very similar to those of i-pentane, and the photochemical ages derived for propane are very similar to those of benzene. However, in our example, the photochemical ages for propane and benzene are on average nearly a factor of three larger than those of n-butane and the pentanes. The reason is that, as mentioned above, mixing of different air masses results in concentration-weighted average photochemical ages. Since, typically, more reactive compounds are depleted more rapidly than less reactive substances in aged air masses, older air masses obtain more weight in the averaging process for less reactive compounds.

10.2.3.4

VOC budgets and isotope ratios

In principle, the budgets of isotopologues can be set-up identically to conventional mass balances. For 13 C labelled VOCs, the change in average atmospheric concentration,

Volatile Organic Compounds in the Atmosphere

(a)

Photochemical age, OH-radicals/cm3 s

400

4E + 11 3E +11 2E+11 1E +11 n-Butane

0 −1E+11 −1E+11

0

1E+11

2E +11

n-Pentane

3E +11

4E +11

(b)

Photochemical age of benzene, OH-radicals/cm3 s

Photochemical age of i -pentane, OH-radicals/cm3 s 1.1E+12

7E+11

3E+11

−1E+11 −1E+11

3E +11

7E +11

1.1E+12

Photochemical age of propane, OH-radicals/cm3 s Figure 10.3 Comparison of average photochemical age for n-pentane and n-butane with i-pentane (a) and propane with benzene (b). The solid line shows the dependence expected if photochemical ages were identical for all VOC. The photochemical ages were calculated using Equation 10.19. The ambient isotope ratio data were taken from measurements of a diurnal cycle at a suburban location in Toronto reported by Czuba (1999); the isotope ratios of the emissions are taken from Rudolph et al. (2002), the rate constants from Atkinson (1994, 1997) and the KIEs from Anderson et al. (2004a, 2004b).

[13 C-VOC]av , in a closed system is related as follows to the budget terms: d[13 C-VOC]av · VSys = 13 C-VOC Si − 13 C-VOC Lj dt i

(10.21)

j

Here, Si and Lj denote the various sources and loss processes of 13 C-VOC, and VSys is the volume of the considered closed system. Since, in ﬁrst approximation, VOC loss rates are proportional to the VOC concentrations, we can substitute the loss term by a pseudo

Gas Chromatography–Isotope Ratio Mass Spectrometry

401

ﬁrst-order loss expression parameterised by pseudo ﬁrst-order rate constants, kj1st . 13 C-VOC

Lj = 13 C-VOC kj1st · [13 C-VOC]av · VSys

(10.22)

We then obtain d[13 C-VOC]av 13 1st = 13 C-VOC Si − [ C-VOC]av · VSys · 13 C-VOC kj dt i

(10.23)

j

For steady state, (d[13 C-VOC]av )/dt = 0, which results in the simple condition that 1st 13 (10.24a) 13 C-VOC Si = 13 C-VOC Lj = [ C-VOC]av · VSys · 13 C-VOC kj j

i

or 13

[ C-VOC] · VSys

j

Si 13 = i C-VOC 1st j 13 C-VOC kj

(10.24b)

Isotopologue concentrations are typically measured, presented and used in the form of the δ-notation. Therefore, Equation 10.24 has to be modiﬁed into a form that uses isotope ratios and kinetic isotope effects to convert conventional budget terms into isotope budgets. Si · (1 + δi ) 12 (10.24c) [12 C-VOC]av · (1 + δav ) · VSys = i C-VOC 1st j (12 C-VOC kj /(1 + εj )) Here δav is the concentration-weighted average atmospheric isotope ratio, δi the isotope ratio of the sources and εj the isotope effects associated with the loss processes. For small values of ε, (1 + ε)−1 can be approximated by 1 − ε, and Equation 10.24c can be written in a more convenient form. 1st [12 C-VOC]av ·VSys ·(1 + δav )· 12 C-VOC kj ·(1 − εj ) = 12 C-VOC Si ·(1 + δi ) j

i

(10.24d) Both δav and εj are usually much smaller than one and their product can, therefore, be neglected without introducing any signiﬁcant bias. Furthermore, when steady state is achieved for 13 C, sources and sinks for 12 C also will be balanced. 1st 1st δav · [12 C-VOC]av · − [12 C-VOC]av · · εj 12 C-VOC kj 12 C-VOC kj j

1 = 12 C-VOC Si · δi VSys

j

(10.24e)

i

An expression similar to Equation 10.24b can be used to eliminate the average steady-state concentration of 12 C-VOC and after rearranging we obtain: 1st j 12 C-VOC kj · εj i 12 C-VOC Si · δi δav = + (10.24f ) 1st i 12 C-VOC Si j 12 C-VOC kj

402

Volatile Organic Compounds in the Atmosphere

Since the fraction of 13 C in carbon as well as carbon isotope effects for atmospheric VOCs are small, the 12 C source and sink terms in Equation 10.24f can be replaced by the values for total carbon without introducing signiﬁcant bias. 1st j kj · ε j i Si · δi δav = 1st + (10.24g) i Si j kj It has already been mentioned that there are several approximations involved in deriving Equation 10.24g. Although this expression is fully adequate for most atmospheric applications, it is important to understand the magnitude of possible bias originating from these simpliﬁcations. The consequence of approximations such as (1 + x)−1 = 1 − x for small values of x can be understood when looking at one loss process only. For uniform isotope ratios or reactant concentrations the exact solution of the mass balance equation includes a higher-order term: i Si · δi i Si · δi i Si · δi · (1 + ε) = ε + + ·ε (10.24h) δav = ε + S S i i i i i Si Due to the small KIEs for most atmospheric reactions of VOCs, the difference between the two versions of Equation 10.24 is typically less than 1‰ for carbon and less than 10‰ for hydrogen isotope ratios. Similarly, steady-state isotope budget are subject to the same limitations and conditions as conventional budgets. Since it is well established that effectively all VOC loss reactions have a very substantial spatial and temporal variability, any application of Equation 10.24 assumes that the VOC isotope ratios are sufﬁciently uniform in time and space to justify the usage of averages values for the calculation of the budget terms in Equation 10.24. Due to the wide range of atmospheric lifetimes of VOCs the bias resulting from this simpliﬁcation is highly compound speciﬁc. Furthermore, it has to be considered that many VOC have atmospheric lifetimes in the range from several hours to a couple of weeks. Consequently, very often, the variability of atmospheric VOC concentrations is high, making the experimental determination of representative average isotope ratios and concentrations a very challenging task. It has to be considered that most VOCs have a number of sources, but isotope budgets provide only one additional constraint for every isotope considered. Consequently, the value of isotope studies to directly determine the magnitude of emissions should not be overestimated. Nevertheless, isotope budgets can be extremely useful to test the consistency of existing budgets or to help in the identiﬁcation of missing or incorrect items in VOC budgets. To understand the possibilities and limitation of VOC isotope budgets in more detail, we will have a closer look at Equation 10.24. The ﬁrst obvious consideration is that the isotope budget only includes information of the relative contributions of the individual sources and sinks to the budget, whereas considerations of absolute values of VOC emission or removal rates will be based on conventional budget estimates. For simplicity, we therefore introduce weighting factors for the sources (i ws ) and removal processes (j wk ) in the following form: Si i ws ≡

i Si

and

kj1st w ≡ j k 1st j kj

(10.25)

Gas Chromatography–Isotope Ratio Mass Spectrometry

403

13C Methyl chloride (‰)

10

−10

−30

−50

−70 0

1 000

2 000

3 000

4 000

Emission rate missing source (Gg/yr) Figure 10.4 Dependence of the stable carbon isotope ratio of methyl chloride on the emission rate of a source with a stable carbon isotope ratio of −135‰. The stable carbon isotope ratio of the missing source is consistent with the isotope ratio of methyl chloride produced by abiotic decomposition of leaf litter (Keppler et al. 2004). The atmospheric budget is balanced by adjusting the rate of methyl chloride uptake in soil, which is presently only poorly understood. The solid line shows the best estimate calculated according to Equation 10.26. The dotted line indicates the standard error of the best estimate calculated from the uncertainties of the source strengths of the known sources and the error of the stable carbon isotope ratios of the emissions. Emission rates and isotope ratios were taken from Keppler et al. (2005). The horizontal line shows the measured average stable carbon isotope ratio of atmospheric methyl chloride reported by Thompson et al. (2002).

Equation 10.24 then can be written as δav = j wk · εj + i ws · δi j

(10.26a)

i

Separation between sources and sinks with known and unknown strength gives: known unknown known unknown δav = · εj + · εj + · δi + · δi j wL j wL i ws i ws j

j

i

i

(10.26b) It is obvious that, even when considering that the sum of the weighting factors has to be unity, isotope budgets often will allow a range of solutions for the unknown source or sink terms. Nevertheless, for atmospheric VOC budgets, acceptable estimates or limits for loss terms can very often be derived from atmospheric chemistry considerations. An example demonstrating the possibilities and limitations of isotope budgets to constrain unknown sources and sinks is given in Figure 10.4. It is based on the chloromethane budget estimates presented by Keppler et al. (2005). They presented a range of scenarios of sources and sinks for atmospheric chloromethane. Based on these scenarios they concluded that emissions due to abiotic production from leaf litter, which had been studied in the laboratory (Keppler et al. 2004), can be used to obtain consistent 13 C and 12 C budgets for

404

Volatile Organic Compounds in the Atmosphere

chloromethane. Indeed, it can be seen from Figure 10.4 that such a source is fully consistent with our present knowledge of the atmospheric 13 C budget of chloromethane. However, it is also obvious from Figure 10.4 that, due to the present considerable uncertainties in the various contributions to the 13 C chloromethane budget, the magnitude of such a source is poorly constrained.

10.2.3.5

Numerical modelling

As mentioned above, the impact of the isotopic composition of VOC on the chemistry of the atmosphere is extremely small. Therefore the main interest in modelling VOC isotope ratios is the use of isotope ratios as tools to gain additional insight into the dependence between emissions, atmospheric transport processes and chemical reactions. Potentially numerical modelling can provide a very powerful tool to interpret measurements of stable isotope ratios of atmospheric VOCs. However, such model calculations require knowledge of the spatial and temporal distribution of the isotope ratio of the sources as well as of the isotope effects associated with the VOC loss processes. It is, therefore, not surprising that presently there is only one published detailed three-dimensional (3D) model study of the stable carbon isotope ratios of VOCs (Thompson et al. 2003). As a consequence of the limited available information on VOC isotope ratios, this study used uniform stable carbon isotope ratios for all sources and presented little information with respect to the sensitivity of atmospheric stable carbon isotope ratios on the emission rates and stable carbon isotope ratio of the VOC sources. However, as a result of the presently rapidly increasing number of published studies of isotope ratios of VOC sources and KIEs for loss processes, it can be expected that numerical model simulations will become more and more useful in the interpretation of atmospheric studies of VOC isotope ratios. In principle, isotope ratios can be modelled by including the isotopically labelled compounds as separate chemical species and treating them separately, according to their speciﬁc properties. However, for practical applications, there are some speciﬁc considerations, which can be useful to reduce computational effort or simplify interpretation of modelling results and atmospheric observations. A direct consequence of the marginal impact of isotope labelling on reactivity is that the feedback between the model chemistry and the VOC isotope ratios can be neglected. This greatly simpliﬁes the reaction schemes needed to incorporate isotope ratios into models. Furthermore, isotope effects are generally sufﬁciently small to justify the use of a linear dependence between the extent of chemical processing and isotope ratio (see Section 10.2.3.3). This signiﬁcantly reduces the computational effort of conducting sensitivity studies and allows separation of the impact of changes in source composition and chemical processing and transport. This can be derived from Equation 10.19. av t δ

=

i=n ic i=1

c

· i0 δ + εOH ·

i=n ic i=1

c

· kOH · i cOH · ti

(10.27a)

Conceptually, Equation 10.27 can be separated into two terms, one (δs ) representing inﬂuence of the isotope ratios of the VOC sources and the other (δa ) describing the change in

Gas Chromatography–Isotope Ratio Mass Spectrometry

405

isotope ratio due to atmospheric VOC removal reactions. δs =

i=n ic i=1

c

δa = εOH ·

· i0 δ

(10.27b)

i=n ic i=1

c

· kOH · i cOH · ti

(10.27c)

Both terms contain the hydrocarbon concentrations, which are connected to atmospheric transport and chemistry in a complex manner. However, since numerical model calculations of isotope ratios inherently include all information required to apply Equation 10.27, this simple, linear functionality provides an effective way to conduct sensitivity studies and a simple means of identifying possible reasons for discrepancies between model and observation. Equation 10.27 is written in a way, which makes it speciﬁc for reactions with the OH-radical as dominant VOC loss mechanism. In case of other reactions contributing to the atmospheric degradation of VOC, it has to be modiﬁed by adding loss terms to Equation 10.27c.

10.3 10.3.1

Experimental methods Introduction

Presently, the most widely used and most ﬂexible technique to measure isotope ratios of VOCs is gas chromatography (GC) in combination with isotope ratio mass spectrometry (GC–IRMS). This technique was ﬁrst introduced nearly 30 years ago (Matthews and Hayes 1978) and has since then evolved into an established technique, which has been applied successfully in many areas of science. Since GC has been used successfully to measure the concentrations of atmospheric VOC, it seems surprising that the ﬁrst paper describing the application of GC–IRMS to study atmospheric VOCs (Rudolph et al. 1997) was published nearly 20 years after the introduction of GC–IRMS. The main reason for this delay can be found in the low concentrations and the complex mixture of VOCs in the atmosphere. In the early stages of development of GC–IRMS the limited sensitivity of continuous ﬂow IRMS would have required VOC enrichment from air samples of volumes which were beyond the limitations of the techniques available at this time. In the mid-1990s, the sensitivity of continuous ﬂow IRMS had increased to a level where meaningful isotope ratio measurements were possible for samples containing only a few nanograms of carbon. At the same time, GC had become a well-developed and studied technique to measure concentrations of atmospheric VOCs, and the knowledge and experience to concentrate and separate VOC samples containing several nanograms of carbon from air was available. Although a number of experimental details still needed to be resolved at this time, the necessary modiﬁcations of available GC methods for VOC concentration measurements with on-line IRMS presented no major problems. There are also a few studies using derivatisation with 2,4-dinitrophenylhydrazine for measurements of atmospheric carbonyl compounds. This method, which was pioneered for measurements of the concentration of formaldehyde in the atmosphere by

406

Volatile Organic Compounds in the Atmosphere

Lowe et al. (1980), is by now widely used in many variations for measurement of carbonyl concentrations in air (see Sirju and Shepson 1995 and references therein; Vairavamurthy et al. 1992). The use of derivatisation techniques for compound-speciﬁc isotope ratio measurement is not speciﬁc for measurements in the atmosphere, but has been used and studied in detail for applications in different areas (see Rieley 1994 and references therein). In addition to the general difﬁculties encountered when using the derivatisation methods for the measurement of atmospheric VOC concentrations, there are two problems speciﬁc for isotope ratio measurement. These are the possible isotope fractionation associated with the derivatisation process and the isotopic dilution of the sample, if the derivatising reagent contains the element that is the target of the isotope ratio measurement. The ﬁrst of the two problems results in a possible bias, the second in an increase of the uncertainty of the measured isotope ratio. This increase in uncertainty is the consequence of the necessity to correct the measured isotope ratio of the derivative for the contribution from the derivatising group. Nevertheless, recent studies (Wen et al. 2004, 2005) demonstrate that derivatisation with 2,4-dinitrophenylhydrazine and subsequent carbon analysis of the hydrozones by GC–IRMS is a promising technique to study the stable carbon isotope ratios of atmospheric formaldehyde, acetaldehyde and acetone. There is a very large number of measurements of the reaction rate constants of isotopically labelled VOCs (Aschmann and Atkinson 1995; Chiltz et al. 1963; Crawford et al. 2004; Dobis et al. 1994; Droege and Tully 1986a, 1986b, 1987a, 1987b; Farkas et al. 2003; Harry et al. 1999; Hitsuda et al. 2001a, 2001b; Stutz et al. 1997, 1998; Taatjes et al. 1999; Tully et al. 1985, 1986; Wallington et al. 1988). These measurements use established techniques for measurements of absolute or relative rates and the KIE is then determined by calculating the ratios of the rate constants for the labelled and unlabelled compound. Although this is a straightforward procedure, which makes use of standard techniques for reaction rate constant measurements, it has two distinct disadvantages. First, it requires the use of artiﬁcially labelled isotopologues, which can be expensive and difﬁcult to obtain. Second, due to the uncertainty of the individual rate constant measurements, KIEs determined from the ratio of two separately measured rate constants have substantial uncertainties. Typical relative errors are in the range of 5–20%. Thus, this method is only suitable for reactions with large KIEs. Indeed, this method has been most widely used for measurements of hydrogen KIEs of multiply labelled VOCs. Most KIEs for gas-phase reactions of VOCs with natural isotope ratios are in the range of several ten per mille or less (Section 10.4). Determining such small KIEs from the ratio of rate constants measured separately is outside of the currently achievable reproducibility for measurements of individual rate constants. A recently developed method to measure small KIEs is based on the combination of established methods to determine relative rate constants with techniques for isotope ratio measurements. As mentioned above, a very versatile technique for isotope ratio measurements is GC–IRMS. This technique not only allows measurements of a wide range of VOCs, but is also suitable for studies of compounds with natural abundance isotope ratios (Anderson et al. 2003, 2004a, 2004b; Iannone et al. 2003, 2004, 2005; Rudolph et al. 2000). Other studies used FTIR spectroscopy to determine relative rate constants for VOC isotopologues (D’Anna et al. 2003; Enghoff et al. 2003; Feilberg et al. 2004; Gola et al. 2005). However, this technique requires the use of artiﬁcially labelled isotopologues. Many of the individual components used today for VOC isotope ratio studies, such as the reaction chambers for relative rate studies, GC–IRMS instrumentation, GC-techniques

Gas Chromatography–Isotope Ratio Mass Spectrometry

407

for VOC concentration measurements or FTIR instruments and their applications have been described in numerous research papers, review articles and textbooks as well as other chapters of this book. The following paragraphs, therefore, will focus on techniques and considerations speciﬁc for studies of isotope ratios of atmospheric VOCs.

10.3.2

KIE measurement techniques

As mentioned above, a recently developed method for studying the KIEs for gas-phase reactions of VOCs with natural abundance isotope ratios is using relative-rate-type experiments with GC–IRMS as the analytical method of choice. In these experiments, the stable carbon isotope ratios and concentrations of VOCs in a reaction chamber are determined at different stages of the reaction. Typically, the reaction chamber set-up used in such measurements is very similar to that used in conventional relative rate experiments. This method has been used to study the carbon and hydrogen KIEs for the reaction of a wide range of hydrocarbons with OH-radicals, Cl-atoms and ozone (Anderson et al. 2003, 2004a, 2004b; Iannone et al. 2003, 2004, 2005; Rudolph et al. 2000). Obviously, details of the set-up, such as the dimensions and material of the reaction chamber, reactant concentrations and duration of the experiment will depend on the speciﬁc reaction studied. A very versatile set-up has been described and characterised in detail by Anderson et al. (2003). A schematic drawing is shown in Figure 10.5. The 20–30 dm3 reaction chambers, which is made of PTFE foil, is housed in an enclosure that allows temperature control as well as irradiation by ﬂuorescent lights. Most of the published studies were conducted at room temperature and ambient pressure but in a limited

Sample loop

Pump

Vent

Mixing fan Injection port

IRMS He Ref. gas

6-Port sampling valve

Reaction chamber

Cryotrap LN2

UV lights

Vent

He

Heartsplit valve FID

LN2

Water trap LN2 Heated interface

Capillary trap

Vent

Synthetic air He

Gas chromatograph

Combustion/ pyrolysis furnace

Figure 10.5 Schematic drawing of the experimental set-up for KIE measurements (for details see text).

408

Volatile Organic Compounds in the Atmosphere

number of studies the temperature dependence of the KIEs was also investigated (Anderson 2005). The ﬂuorescent lights allow studying VOC reactions with reactants produced by photolytic reactions such as Cl-atoms or OH-radicals. While the reaction proceeds in the reaction chamber, samples of deﬁned volume are taken at regular intervals and analysed for VOC concentrations and isotope ratios. Typically, the reaction is studied until the concentration of the studied VOC is reduced to 20–50% of its initial value. A lower VOC consumption increases the uncertainty resulting from random scatter of concentration and isotope ratio measurements. Furthermore, a signiﬁcantly higher VOC consumption does not necessarily increase the experimental accuracy due to the limited linear range of isotope ratio measurements and the increasing inﬂuence of wall effects. Usually, several GC–IRMS measurements are made with the VOC in the reaction chamber prior to initialising the reaction. This allows checking for VOC losses from the reaction chamber due to leakage or diffusion, as well as testing the reproducibility of the GC–IRMS measurements. Obviously the acceptable threshold for limits of the concentration and isotope ratio variability during these instrument stability tests depend on the speciﬁc reaction studied and the desired level of reproducibility. It has been shown that very reproducible results can be achieved if the reproducibility for concentration measurements is better than 2% and for carbon isotope ratio measurements better than 0.2‰ (Anderson et al. 2003). In hydrogen KIE studies, a reproducibility of the isotope ratio measurements of better than 5‰ could be achieved (Iannone et al. 2003). Often, several VOCs are studied in the same experiment. This allows determining relative rates for the different VOC, which then can be compared to literature values in order to verify that the changes in VOC concentration are indeed due to the chemical reaction studied. However, due to the uncertainty of published rate constants for gas-phase reactions of VOC, this procedure does not allow identiﬁcation of loss processes which contribute in the range of 10–20% or less to the overall decrease in VOC concentrations. The initial VOC concentrations in the reaction chamber are typically in the low to medium ppm range. At this concentration level, accurate isotope ratio measurements require a sample volume in the range of several cm3 . For most published studies sample sizes were in the range of 5–10 cm3 . It should be noted that, although the exact volume of the sample is of minor relevance for the KIE studies, it is important that within a KIE measurement the sample size is kept as constant as possible. This implies that temperature and pressure in the sample loop have to be kept stable. Due to the limited linear range of isotope ratio measurements, concentrations and sample size have to be matched in order to optimise the range available for observing the concentration change during the experiment. In principle, this can be done by changing either the size of the sample loop or the initial VOC concentration for the experiments. For practical reasons, generally, the concentrations are adjusted. Sample volumes in the range of several cm3 are too large to be separated directly on capillary GC-columns. Therefore, typically, a cryogenic focusing step is applied. This is done either in a cold-trap placed between the sample loop and the column or by cooling down the head of the column. The selection of the column and the details of the separation conditions depend on the compounds studied. Usually, the separation presents no problem since the number of compounds to be separated is small. Nevertheless, consideration has to be given to

Gas Chromatography–Isotope Ratio Mass Spectrometry

409

the separation between the reacting VOC and the products formed, especially if several VOCs are studied in the same experiment. Usually, general-purpose, wall-coated capillary columns with low bleed, such as cross-linked methyl or methyl-phenyl polysiloxanes, are used. For the separation of light VOCs also, the use of porous layer open tubular columns, for example, aluminum oxide or Poraplot®-type columns, has been reported. The carrier gas is always helium. Between the end of the column and the IRMS interface a valve system is installed, which allows switching of the column efﬂuent between the interface and a conventional detector, typically a FID. This allows minimising the amount of contaminants transferred from the column to the interface and IRMS, as well as monitoring the chromatographic peaks without using the IRMS. The main component of the detection system is the IRMS and the online interface. The purpose of the interface is to convert the separated compounds into gases suitable for isotope ratio measurements. In the case of carbon isotope ratio measurements, all carbon is oxidised to carbon dioxide; for hydrogen the organic compounds are converted to molecular hydrogen by pyrolysis. The most widely used carbon oxidation interfaces consist of quartz capillaries packed with copper oxide pellets or a ceramic tube with thin wires of copper, nickel or platinum at temperatures in the range of 1 100–1 200 K (Matthew and Hayes 1978). The pyrolysis interface usually consists of a ceramic tube at a temperature slightly exceeding 1 700 K. At this temperature all hydrogen in organic compounds is converted to molecular hydrogen (Burgoyne and Hayes 1998). Since the carrier gas ﬂow rate through the column is usually higher than the optimum ﬂow rate for the ion source, an open split is installed between the online interface and the IRMS, which allows admitting a constant ﬂow of carrier gas into the IRMS ion source. In order to remove water, which would interfere with the mass spectrometric isotope ratio measurements, a cold trap or a Naﬁon® dryer is placed between the interface and the inlet of the IRMS. The hydrogen and carbon isotope ratios are derived by monitoring masses 2 (H2 ) and 3 (HD), or 44 (12 C16 O2 ) and 45 (13 C16 O2 ), respectively. For carbon isotope ratios also, mass 46 (12 C16 O18 O) is monitored in order to allow correction of the mass 45 signal for the contribution from 12 C16 O17 O (Craig 1957; Santrock et al. 1985). The isotope ratios are determined from the ratios of the integrated peak areas for the individual masses. The isotope ratios are always determined relative to a reference gas, which is added to the efﬂuent from the interface for short time periods before, or after the GC runs, or during peak free sections of the chromatograms. It should be noted that, although usually a standard traceable to the internationally accepted reference points is used, this is of secondary importance for KIE measurements since the evaluation of the experiments is based on changes in isotope ratios and does not require knowledge of absolute values. Similarly, the evaluation of KIE experiments only requires knowledge of the relative changes in concentration. The peak areas determined from the traces representing the most abundant hydrogen or carbon isotopes, masses 2 and 44, respectively, are proportional to the concentration of the studied compound. Therefore these peak areas can be used directly to determine relative changes in concentration without calibration. Usually the KIEs are determined from the following linearised dependence between concentrations and isotope ratios, which is equivalent to plots commonly used for evaluation

410

Volatile Organic Compounds in the Atmosphere

0.0 −0.2

ln(12Ct / 12C0)

−0.4 −0.6 −0.8 −1.0 −1.2 −1.4 −1.6 0

0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009

ln{(13Ct +1000)/(13C0+1000)} Figure 10.6 Example for the determination of KIEs from the experimentally determined dependence between concentration and isotope ratio. The data, which represent the reaction of toluene with OH-radicals, are taken from Anderson (2005). The different symbols represent the results from 5 different experiments. The solid line is a linear least square ﬁt to all data (R 2 = 0.977).

of relative rate experiments. [VOC]t δt + 1000 KIE ln ln = [VOC]0 1 − KIE δ0 + 1000

(10.28)

Here, [VOC]0 and [VOC]t are the VOC concentrations at the beginning of the experiment and at time t , respectively. The corresponding isotope ratios are δ0 and δt . The KIE then is calculated from the slope of a linear least square ﬁt: KIE =

slope 1 + slope

(10.29)

An example is given in Figure 10.6 for the carbon KIE of the reaction of toluene with the OH-radical. The slope is −182.8 ± 5.9, from which a KIE of 1.0055 ± 0.0002 is calculated. This corresponds to an ε value of 5.5 ± 0.2‰. It should be noted that the error derived from the linear least square ﬁt mainly reﬂects uncertainty due to random errors of the isotope ratio concentration measurements. Systematic errors, for example bias due to loss processes other than the studied reaction, are not included in this error. It is therefore not surprising that often the reproducibility of KIE experiments is not as good as the error derived from the linear least square ﬁt (Anderson et al. 2003). Nevertheless, measurements of carbon KIEs with reproducibility in the range of 0.03–0.5‰ (Anderson et al. 2004a, 2004b) are possible. For hydrogen KIE measurements the reproducibility is typically between 1‰ and 6‰ (Iannone et al. 2003, 2004). A number of studies have been conducted using FTIR to follow the changes in the concentrations of artiﬁcial VOC isotopologues (D’Anna et al. 2003; Enghoff et al. 2003; Feilberg et al. 2004; Gola et al. 2005). This method has been used to study photolysis of VOCs, as well as reactions with OH and NO3 radicals and Cl-atoms. The advantage of this method is that it allows studies of VOCs that presently cannot be analysed by GC with

Gas Chromatography–Isotope Ratio Mass Spectrometry

411

the accuracy required for isotope ratio studies, for example formaldehyde. Limitations of using FTIR for KIE studies arise from the complexities of FTIR spectra for heavier VOCs. Furthermore, the sensitivity of available FTIR instrumentation is presently not sufﬁcient for accurate isotope ratio measurements in VOCs with natural isotope abundance and thus mixtures containing artiﬁcially labelled VOC isotopologues have to be used. Only for very simple molecules KIEs measured for speciﬁc isotopomers can be directly applied to the mixtures of naturally occurring isotopomers. For complex VOCs, the naturally occurring isotopologues consist of a mixture of compounds labelled at different sites. For these molecules it can be extremely difﬁcult to derive KIEs that are representative for compounds with natural isotope abundance from studies of VOCs with site-speciﬁc labelling (Section 10.4.1). However, measurements of KIEs for VOCs with site-speciﬁc labelling can be extremely valuable to gain deeper insight into the theory of KIEs for gas-phase reactions of organic molecules and reaction mechanisms.

10.3.3

Techniques for measurements of stable carbon isotope ratios of ambient VOCs

Presently GC–IRMS is by far the most widely used, most versatile and most promising technique for compound-speciﬁc measurements of isotope ratios of atmospheric VOCs. There are a number of publications describing GC–IRMS methods and their applications for stable carbon isotope ratio measurements in atmospheric VOCs (Archbold et al. 2005; Bill et al. 2004; Rudolph et al. 1997, 2000, 2002, 2003; Saito et al. 2002; Thompson et al. 2002; Tsunogai et al. 1999). All these methods are very similar to the one published by Rudolph et al. (1997), which in principle consists of a combination of established GC VOC analysis techniques with on-line IRMS. Although the presently available GC–IRMS techniques allow isotope ratio measurements for a number of elements, to the author’s knowledge, for atmospheric VOCs, only stable carbon isotope ratio measurements have been reported to date. A schematic set-up of an instrument for measurement of the stable carbon isotope ratio of VOCs in ambient air is shown in Figure 10.7. The system consists of four basic components, the sample enrichments procedure, the GC separation, the interface and the IRMS. In state of the art systems the control of the different components as well as data acquisition and evaluation is computer controlled and highly automated. As mentioned above, the individual components of the system are derived from existing, well-established measurement techniques. Furthermore, a brief description of measurements of the isotope ratio of VOCs in gas samples has been given in Section 10.3.2. In the following sections (10.3.3.1–10.3.3.3), therefore, only modiﬁcations and problems of speciﬁc importance for the isotope ratio measurements of VOCs in the atmosphere will be discussed. The most important considerations, which determine the details of the methods, are the need to handle gaseous samples, the low concentrations of atmospheric VOC and the large number of VOCs that are present in the atmosphere.

10.3.3.1

Sampling and sample enrichment

In principle sample collection and sample enrichment techniques follow the procedures that are used for concentration measurements. In contrast to concentration measurements,

412

Volatile Organic Compounds in the Atmosphere

He

Vent

Sample In

Vent Pump

CO2Trap I

Vent

H2O Trap I

MFC

IRMS

Cryotrap I

He CO2Trap II H2O Trap II LN2

Heartsplit valve

Vent FID2 FID1

Cryotrap II

Ref. gas He

He

Vent

LN2 Capillary trap

Vent Column II

He He

Column I

Combustion furnace

Gas chromatograph Figure 10.7 Schematic drawing of the experimental set-up for measurements of stable carbon isotope ratios of atmospheric VOC (for details see text).

there is presently no instrumentation, which would be suitable for ﬁeld deployment. Air samples are collected in stainless steel canisters and transferred to the laboratory for analysis. Typically, the samples are pressurised in order to provide the large air volumes required for VOC isotope ratio measurements. The sample volume required for VOC isotope ratio analysis depends on the atmospheric VOC concentration as well as on the desired precision and accuracy of the measurement. The minimum mass of carbon required for any given compound can be derived from the ion yield of the IRMS. State-of-the-art instruments require approximately 2 000 carbon dioxide molecules for each ion registered. Since carbon contains nearly two orders of magnitude less 13 C than 12 C, the impact of random noise on the isotope ratio will be determined by the 13 C signal. Counting statistics require that in order to achieve 0.1‰, reproducibility 108 counts are required. This corresponds to about 2×1011 molecules of 13 CO2 entering the ion source of the IRMS. The equivalent total mass of carbon with natural abundance isotope ratios is approximately 0.4 ng. This estimate does not include the contribution of the mass spectrometer background or the gas chromatographic baseline to the random noise of the signal. Since, typically, only between 10% and 30% of the total sample enter the ion source of the IRMS, the required mass of carbon for each VOC in the sample is in the range of 1–4 ng. Atmospheric mixing ratios of VOC range typically from several ppt to a few ppb. In order to be widely applicable, a method for VOC isotope ratio measurements should therefore be applicable to mixing ratios in the range of 0.1 ppb. For a medium-molecular-weight VOC,

Gas Chromatography–Isotope Ratio Mass Spectrometry

413

for example, benzene, this corresponds to approximately 0.3 ng/dm3 of air at standard temperature and pressure (STP). Therefore, realistically sample volumes in the range of at least 5–15 dm3 (STP) will be required for isotope ratio analysis of ambient VOCs. For measurements of VOCs with mixing ratios signiﬁcantly below 0.1 ppb, sample volumes have to be even larger. For example, Bill et al. (2004) reported a procedure for enrichment of bromomethane, which has an atmospheric mixing ratio of approximately 0.01 ppb, from 1.2 m3 (STP) of air for subsequent GC–IRMS analysis. Similar to procedures used for concentration measurements, the VOC enrichment steps are usually integrated into the overall analytical procedure to minimise the risk of sample loss or contamination during the handling of the sample. Most techniques for measurements of VOC isotope ratios use several steps in the sample enrichment procedure, a consequence of the substantial sample volumes required for accurate isotope ratio measurements. The ﬂow rate for transfer of a gaseous sample to a capillary column is limited to a few cm3 /min. This is not only too small a ﬂow rate for sample volumes in the range of dm3 , but also is too low to allow efﬁcient transfer from a trap designed for sampling at high ﬂow rates. For example the ﬁrst trapping step of the instrument shown in Figure 10.7 consists of a 15-mm-diameter and approximately 30-cm long column packed with glass beads. This allows sampling at a rate of 1–2 dm3 (STP)/min. However, direct transfer of the enriched sample to a capillary column from a trap of this volume would take an hour or more. Adding an additional trap of medium size allows replacing one slow transfer step by two fast steps. The type of packing material used and temperature of the traps depend on the analysed VOCs. Glass beads at temperatures in the range of 100–120 K are suitable for enrichment of high-to-medium volatility VOCs (Czuba 1999; Rudolph et al. 1997, 2000, 2002; Thompson 2003; Thompson et al. 2002; Tsunogai et al. 1999). In a few publications also, the use of organic adsorbents (Bill et al. 2004) or simply unpacked traps (Archbold et al. 2005) has been reported. It should be noted that, in principle, nearly all efﬁcient trapping procedures used for VOC concentration measurements should, with modiﬁcations to accommodate the required large-sample volumes, be suitable for subsequent GC–IRMS analysis. However, the number of publications presently describing use and testing of VOC-trapping systems for GC–IRMS measurements is extremely limited. Depending on the trapping conditions, water vapour and carbon dioxide may also be collected during cryogenic VOC enrichment. The amount of water and carbon dioxide in ambient air samples of several dm3 is far too high for separation on capillary columns and in many cases can even block the traps. Furthermore, the tailing of large carbon dioxide peaks can interfere with the detection of light VOC (Rudolph et al. 1997; Thompson 2003). Therefore most applications include water and carbon dioxide traps prior to the VOC trap. The most widely used trap for water vapour removal consists of an open tube at low temperatures, typically between 245 and 210 K (Czuba 1999; Rudolph et al. 1997; Thompson 2003), but also the use of Naﬁon® dryers (Archbold et al. 2005) or traps packed with magnesium perchlorate (Bill et al. 2004) has been reported. Suitable traps for carbon dioxide removal consist of tubes packed with Carbosorb®, Ascarite® or potassium carbonate (Czuba 1999; Rudolph et al. 1997; Thompson 2003). It has been found that water and carbon dioxide removal is far more efﬁcient when two set of traps are used, one before and one set after the VOC concentration step (Thompson et al. 2003).

414

10.3.3.2

Volatile Organic Compounds in the Atmosphere

Separation and detection

Although for GC–IRMS measurements similar separation conditions as for concentration measurements may be used, there are two speciﬁc factors that have to be considered. First, the efﬁciency of the separation has to be considerably higher for GC–IRMS measurements. Second, the necessary use of a chemical conversion online interface creates a serious potential for peak broadening, which greatly reduces the possibility of using very-high-resolution GC, for example using columns with less than 0.2 mm i.D. The combination of these two factors results in non-trivial separation problems for complex samples such as VOCs in ambient air. The need for effectively complete separation arises from the necessary high accuracy of peak evaluation, which is the obvious consequence of the desired relative uncertainties in the per mille range. Furthermore, there is a difference in the retention times between VOC isotopologues. Usually, the isotopologues containing the heavier isotope have shorter retention times. Although the differences in retention times are very small, this results in substantial uncertainties for the evaluation of not fully resolved peaks. An example demonstrating this problem is given in Figure 10.8. The lower panel shows the chromatographic trace for the mass 44 signal, which is effectively the signal of a carbon-speciﬁc detector. The upper panel shows the ratio of the mass 45 over mass 44 signals. The positive deviations from the baseline value of the 45/44 ratio trace are the result of the shorter retention times of VOCs containing a 13 C atom, which results in an enrichment of 13 C at the front of the peak. Similarly, 13 C is depleted at the end of the peak, usually resulting in a negative deviation of the ratio trace from the baseline. The complete elution of the mass 45 and mass 44 peaks is shown by the return of the 45/44 ratio signal to its baseline value. Consequently the return of the 45/44 ratio trace to its baseline value can be used as the indicator for a complete separation of both the mass 45 and 44 peaks. Based on general guidelines for chromatography, for the example shown in Figure 10.8, the separation of ethane and ethyne is deemed sufﬁcient for concentration measurements. However, isotope ratios determined for these two peaks will have very signiﬁcant uncertainties, as indicated by the absence of a return to the baseline of the 45/44 ratio trace between the two peaks. A useful procedure for testing the quality of the peak evaluation is to apply small changes to the boundaries used for peak evaluation (Rudolph et al. 1997, 2000; Thompson et al. 2002). Although this is a somewhat tedious procedure, it gives very valuable information about the impact of random noise and insufﬁcient peak separation on the quality of the isotope ratio measurements (Section 10.3.3.3). As a consequence of the high demands on the quality of the separation, combinations of two columns and two-dimensional (2D) separations were used in some studies (Czuba 1999; Rudolph et al. 2000, 2002, 2003; Thompson 2003; Thompson et al. 2002). A schematic set-up for a possible two-column combination is included in Figure 10.7. The four-way valve located between the two columns allows selective transfer of VOCs only partially separated on the ﬁrst column to a second column with a different stationary phase for further separation. The second four-way valve allows selection of the efﬂuent from either column for transfer to the IRMS interface or a FID. The obvious disadvantage of such partial 2D separations is the reduction in the number of compounds, which can be analysed in one chromatogram. Nevertheless, using temperature programming with slow heating rates and long columns, measurements of isotope ratios of 20–30 different VOCs have been

45/44

Gas Chromatography–Isotope Ratio Mass Spectrometry

415

1.3 1.28 1.26 1.24 1.22 1.2 1.18 1.16 1.14 1.12 1.1

44 m/z lon count

1 000 800 600

CO2

Carbon tetrachloride

Propane

Ethane

Benzene

Ethyne Ethene

Chloromethane Pentane CFC -113

400

Propene

200 0 3 050

3 250

3 450

3 650

3 850

4 050

4 250

4 450

4 650

4 850

Times (s) Figure 10.8 Example for the GC–IRMS measurement of light VOC in a background air sample. The sample volume is 13.5 dm3 (STP). The separation was performed on a HP1 column. The upper panel shows the ratio of the mass 45 and 44 traces, the lower the intensity of the mass 44 signal in arbitrary units. The mixing ratio of chloromethane is 0.6 ppb, which corresponds to 4 ng of carbon in the air sample.

achieved in a single analysis (Czapiewski et al. 2002; Czuba 1999; Rudolph et al. 2002, 2003; Thompson 2003). An example for a partial 2D separation of VOCs combined with online IRMS analysis is shown in Figure 10.9. Very recently Komatsu et al. (2005) described a procedure that allows selective removal of alkenes from air samples prior to GC–IRMS analysis. They found that passing the concentrated VOC mixture through a cartridge with I2 O5 prior to GC–IRMS analysis removed alkenes from the sample without changing the isotope ratios of light alkanes, chloromethane or bromomethane. This substantially reduced the complexity of the chromatograms and thus increases the reproducibility and the accuracy of the IRMS measurements. The obvious disadvantage of that is this procedure will not allow analysis of VOCs that react with I2 O5 . Several authors (Czapiewski et al. 2002; Czuba 1999; Rudolph et al. 2000, 2002, 2003; Thompson 2003) report on the combination of GC–IRMS with conventional GC–MS. This signiﬁcantly reduces uncertainties in peak identity and also allows verifying the purity of the peaks. However, it is important to realise that this does not allow isotope ratio measurements for individual compounds in the case of overlapping peaks. As mentioned previously, measurements of isotope ratios are made relative to a standard with known composition. For GC–IRMS measurements the reference gas usually is added

416

Volatile Organic Compounds in the Atmosphere

100

%

FID Trace

50

Transfer

Transfer

Transfer

Transfer

0 0

10

20

30

40

50

60

70

80

90

80

90

Time (min) 100

%

IRMS Trace

f

50 c

e

d ab 0 0

10

20

30

40

50

60

70

Time (min) Figure 10.9 Example for the GC–IRMS measurement of light VOC in air using partial 2D separation. The upper panel shows the trace of the FID signal for the ﬁrst column, with the sections of the chromatogram which were transferred indicated by the label ‘transfer’. The lower panel shows the relative (normalised to the maximum signal) mass 44 signal from monitoring the efﬂuent of the second column with the IRMS. The identiﬁed peaks are: a, n-pentane, b, isoprene, c, methacrolein, d, methyl vinyl ketone, e, benzene and f, toluene. The signals at 20, 50, 70 and 85 min are due to the injection of reference gas.

to the efﬂuent from the interface under conditions identical to those during measurement of VOC peaks (Figure 10.9). Ideally, the measurement of the reference gas should be made as close in time to the analysed peak as possible to avoid bias due to instrument drift. In reality, this is not always possible since measurement of the reference gas requires a time window in the chromatogram without peaks. Typically chromatograms of atmospheric VOCs contain a large number of generally densely spaced peaks. 2D separations can easily be tuned to allow reference gas injections without risk of peak overlapping. State-of-the-art GC techniques allow selection of a wide range of ﬂow rates depending on the speciﬁc application. In addition to the availability of capillary columns with a wide range of diameters, there is also the option to operate GC-columns at above-optimum ﬂow rates. For capillary columns, the increase in plate height with increasing ﬂow rate is small in the vicinity of the optimum separation conditions, especially if helium is used as the carrier gas. However, for GC–IRMS analysis of atmospheric VOCs, there are several conditions that limit the range of suitable ﬂow rates for practical reasons. First, if the column ﬂow rate is extremely low, transfer of the sample from the cryogenic traps may take very long or may

Gas Chromatography–Isotope Ratio Mass Spectrometry

417

not be practical at all, even if a two- or three-step sample enrichment procedure is used. Second, due to the ﬁnite volume of the GC–IRMS combustion interfaces, low ﬂow rates may result in signiﬁcant peak broadening. These two considerations suggest that high ﬂow rates are advantageous. However, the helium ﬂow rate, which can be accommodated by IRMS instruments without signiﬁcant loss of performance, is quite low. For most instruments, this value is in the range of 0.5 cm3 /min. The consequence is that for very high carrier gas ﬂow rates only a small fraction of the total sample is available for the IRMS measurement. This would require the use of extremely large sample volumes. Most published procedures use carrier gas ﬂow rates in the range of a few cm3 /min in combination with 0.32 mm i.D. columns. Nevertheless, it is obvious that the optimum ﬂow rate, and therefore also the optimum column diameter, will depend on the VOC concentrations in the sample as well as on the available sample volume and the details of the sample enrichment procedure and the design of the GC–IRMS interface. However, to the author’s knowledge, no systematic study of the relationship between carrier gas ﬂow rate and the performance of VOC isotope ratio measurements has been published.

10.3.3.3

Performance and tests of VOC isotope ratio measurements by GC–IRMS

The available mass of carbon is one of the key factors determining the reproducibility of VOC isotope ratio measurements by GC–IRMS. It has already been shown in the study by Rudolph et al. (1997) that the uncertainties of VOC isotope ratio measurements are highly dependent on the size of the sample. Their results indicate that at least some nanograms of carbon are required for isotope ratio measurements of VOCs by online GC–IRMS if reproducibility of better than 1‰ is desired. Most of the studies of ambient VOCs using GC–IRMS conducted since then found similar performance characteristics (Czapiewski et al. 2002; Rudolph et al. 2000, 2002, 2003; Saito et al. 2002; Tsunogai et al. 1999). However, there are only a very limited number of systematic studies of the dependence between measurement reproducibility and carbon mass for VOC isotope ratio measurements. Figure 10.10 shows the standard deviations obtained from repeat analysis of VOCs in air samples using different sample volumes for a number of VOCs at ppb and sub-ppb concentrations. Although there is considerable scatter in the reproducibility, it is evident that for an isotope ratio measurement with better than 1‰ reproducibility at least 1 ng of carbon is required. If between 1 and 10 ng of carbon is available, the average reproducibility is in the range of 0.6‰, and for more than 30% of the measurements the reproducibility is better than 0.4‰. Czuba (2000) and Thompson (2003) report the reproducibility of repeat measurements of artiﬁcial ppb-level VOC mixtures made during periods of several months. On average, the observed long-term reproducibility was in the range of 1‰. This is somewhat worse than the results of accuracy tests and short-term reproducibility tests (see above). However, Thompson (2003) also observed that the measured isotope ratios showed a temporal trend, most likely due to changes in the test gas mixtures. Although these changes were small, they are sufﬁcient to explain the difference between long- and short-term reproducibilities. The reproducibility of isotope ratio measurements is considerably worse than expected from ion counting statistics alone (solid line in Figure 10.10). This can be explained

418

Volatile Organic Compounds in the Atmosphere

13C Standard deviation (‰)

5.0

4.0

3.0

2.0

1.0

0.0 0.01

0.1

1 10 Mass of carbon (ng)

100

1 000

Figure 10.10 Dependence between the uncertainties of stable carbon isotope ratio measurements of VOC made by GC–IRMS and mass of carbon. The full data points represent the standard deviations derived from repeat measurements. The open points are the reproducibility derived from repeat evaluations of the same analysis using slightly different deﬁnitions of peak boundaries. Data points are taken from Rudolph et al. 1997 (diamonds), Czuba (1999) (triangles) and Thompson (2003) (circles). The solid line is the statistical uncertainty calculated from counting statistics for the mass 45 peak. For this calculation a yield of 1 ion/2000 molecules of carbon dioxide is used and it is assumed that 20% of the sample is introduced into the IRMS ion source.

by baseline noise, including ﬂuctuations or systematic baseline drift due to temperature programming. The inﬂuence of the quality of the GC baseline and separation can be seen from the reproducibility obtained from repeat evaluations of the same chromatograms using marginally changed baseline and peak boundary deﬁnitions. The variability obtained from such repeat evaluations (open symbols in Figure 10.10) is comparable to the reproducibility of repeat measurements of the same sample. A valuable use of this observation is the possibility to derive a ﬁrst-order estimate of the measurement reproducibility from repeat evaluations of GC–IRMS traces. Due to the large sample volumes required for GC–IRMS measurements, the collected samples very often are not sufﬁcient for more than one measurement. There are only very few studies that directly allow the evaluation of the accuracy of GC–IRMS measurement of atmospheric VOCs. Rudolph et al. (1997) report results of GC–IRMS measurements of artiﬁcial mixtures of ppb-level VOCs in air. These mixtures were prepared from bulk substances, which had been analysed by conventional off-line combustion and dual-inlet isotope ratio analysis. This allowed comparing the known isotope ratio of the mixtures with the results of GC–IRMS measurements. The comparison conducted by Rudolph et al. (1997) was limited to ethane and chloromethane. Czuba (2000) used an identical procedure to investigate the accuracy of isotope ratio measurements by GC– IRMS analysis for 14 VOCs in air, including isoprene, several aromatic VOC, alkanes and alkenes. The results of these two studies are shown in Figure 10.11. The agreement between GC–IRMS and off-line results gives no indication for a bias and on average the absolute

Gas Chromatography–Isotope Ratio Mass Spectrometry

419

−20.0

13C GC-IRMS (‰)

−25.0 −30.0 −35.0 −40.0 −45.0 −45

−40

−35

−30

−25

−20

13C Offline (‰) Figure 10.11 Comparison of stable carbon isotope ratios of VOC in ppb level mixtures measured by GC–IRMS with results from off-line analysis of the pure substances. Data are taken from Rudolph et al. 1997 (squares) and Czuba (1999) (diamonds). The error bars show the uncertainty derived from repeat measurements of the artiﬁcial mixtures. The errors of the off-line data are typically in the range of 0.1‰. The broken line represents the linear least square ﬁt δ 13 C GC–IRMS = δ 13 C Ofﬂine × (1.015 ± 0.005) + (0.15 ± 1.5)‰; R 2 = 0.966.

difference is 0.65‰, which is effectively identical to the average reproducibility of repeat measurements of the mixtures (0.7‰). There also have been several studies of individual steps and components of VOC analysis by GC–IRMS. Although these do not directly allow inferences about the overall accuracy, they can be used to estimate the extent of bias or random variability introduced by various steps speciﬁc for analysis of stable isotope ratios of VOCs in the atmosphere. To some extent, concentration-based studies can also be used to draw inferences about possible bias in stable isotope ratio measurements. For example, studies of the efﬁciency of the trapping procedure can provide evidence for the absence of isotope fractionation during this step. If VOC recovery is quantitative, simple isotope budget considerations can be used to rule out any impact on stable isotope ratios. Strictly speaking, recovery tests can never verify that the efﬁciency of the trapping procedures is exactly 100%. However, if the possible loss can only be small, a visible impact of an effectively quantitative step on the stable isotope ratio measured is unlikely, since it would require that an unreasonably large isotope fractionation factor is associated with the studied procedure. Bill et al. (2004) tested a method for concentrating bromomethane at low ppt levels from air samples in the range of 1 m3 and found that the reproducibility of the overall method is approximately 1–2‰ for sample sizes in the range of 5–15 ng of carbon. They also found that the sample enrichment technique had an efﬁciency of (97 ± 8)%. Based on both ﬁndings, it is reasonable to assume that their sampling procedure was free from signiﬁcant bias, although no direct tests of the accuracy of the complete method were performed.

420

Volatile Organic Compounds in the Atmosphere

Archbold et al. (2005) present several tests of a procedure to remove water and carbon dioxide from samples of up to 10 dm3 . They found that including various combinations of a Naﬁon® dryer, a Carbosorb® tarp and an acetone/dry ice cooled water trap did not change the stable carbon isotope ratio of several halocarbons by more than 0.5–1‰. The only exception was chloromethane, for which tests with and without Naﬁon® dryer resulted in isotope ratios differing by 2‰ or more. Similar tests performed by Thompson (2003) using a slightly different enrichment procedure, showed that for chloromethane and 14 hydrocarbons, including isoprene, several aromatic compounds, alkenes and alkanes, changes in isotope ratio due to the trapping step were below the reproducibility of the isotope ratio measurements. Czuba (1999) reported tests of a preconcentration system using a 20-cm long 3/8 i.D. stainless steel tube packed with glass beads as the ﬁrst enrichment step. At liquid argon temperature, varying the ﬂow rate during enrichment between 110 and 340 cm3 /min caused no detectable change in VOC carbon isotope ratio or peak area. Variations of the temperature of the water trap between 253 and 233 K showed that for a wide range of hydrocarbons no dependence between temperature and carbon isotope ratio was found. For heavier VOCs, such as α-pinene, o-xylene, n-decane and n-nonane, a decrease in peak area at water trap temperatures below 243 K was found, but this was not accompanied by a detectable change in isotope ratio.

10.4 10.4.1

Kinetic isotope effects Origin and basics of kinetic isotope effects

Traditionally, measurements of gas-phase kinetic isotope effects of VOCs were motivated by the possibility to gain more detailed insight into reaction mechanisms and, therefore, were primarily studying compounds with either complete or site-speciﬁc labelling. However, due to the recent developments in studies of atmospheric VOC isotope ratios, it has become evident that knowing the KIEs for reactions of VOCs in the atmosphere is essential. This has not only triggered a substantial number of KIE measurements, but also created interest in KIE measurements for VOCs with natural isotope abundances. One of the advantages is that this avoids the often very substantial efforts and costs of synthesising complex molecules with site-speciﬁc labelling. But the main motivation is that KIEs measured for VOCs with natural isotope abundance are directly applicable to the atmosphere. For studies of artiﬁcially labelled compounds, this is only possible in the case of very simple molecules such as halomethanes, formaldehyde or methane. For more complex molecules, the KIE applicable to the atmosphere is the average KIE for all possible singly labelled isotopomers. Although in principle such average KIEs can be derived from studies of artiﬁcially labelled compounds, it is obvious that this may be effectively impossible, or at least impractical, for very complex molecules. Therefore, this section will focus on KIEs for reactions of VOCs with natural abundance isotope ratios and studies with labelled compounds, which are applicable to atmospheric VOCs. The origin of KIEs can best be understood using transition state theory (TST). This theory assumes that the forward and backward reactions between the reactants and the transition state is fast enough that equilibrium thermodynamics can be used to describe

Gas Chromatography–Isotope Ratio Mass Spectrometry

421

the steady-state concentration of the transition state. Based on TST, the bimolecular reaction rate constant k is k = (κT /h) · exp(S ‡ /R) · exp(−H ‡ /RT )

(10.30)

where κ and h are Boltzmann’s and Planck’s constants, respectively. S ‡ and H ‡ are the standard entropy and enthalpy, respectively, for the formation of the transition state from the reactants. For a labelled compound an equivalent expression can be used and it follows that the KIE depends on the differences in standard enthalpies and entropies between unlabelled and labelled reactant and the transition state. KIE = k12 /k13 = exp((S ‡ − SL‡ )/R) · exp(−(H ‡ − HL‡ )/RT )

(10.31)

Here SL‡ and HL‡ are the standard entropy and enthalpy, respectively, for the formation of the transition state from labelled reactants. The difference H ‡ − HL‡ is identical to the differences in zero point energies (E0 ), which can be derived from the fundamental frequencies of the reactants (ν R ) and the transitions state (ν ‡ ). ‡ 1 ‡ R R E0 = E0 − L E0 = h · νi − νj − L νi − L νj 2 i

j

i

j

(10.32) It is obvious from Equation 10.30 that only frequencies that change due to labelling as well as during formation of the transition state will contribute to a difference in zero point energies. Therefore for KIEs of VOC gas-phase reactions generally only a few key frequencies have to be considered. For a harmonic oscillator, the changes in fundamental frequencies due to isotopic √ substitution can be calculated according to the Teller-Redlich theorem: H ν = L ν · mL /mH , where H ν and L ν are the fundamental frequencies for the heavy and light isotopologues, respectively, and mH and mL the masses of the heavy and light isotopes, respectively. In spite of these quite simple dependencies, calculations of the differences in standard enthalpies are generally not trivial due to the difﬁculty of determining the fundamental vibration frequencies for the transition state. For many reactions, the differences in zero point energies determine the KIE. However, in case of very small values for (H ‡ − HL‡ )/RT , the contribution of the entropy term to the KIE may be relevant. Based on fundamental statistical mechanics, the differences in entropy can be calculated from the partition functions. Similar to the contributions to zero point energies, only partition functions that change due to labelling and during formation of the activated complex contribute to the KIE. For some degrees of freedom the calculation is very simple. In other cases exact calculations are not possible due to the lack of detailed information about the transition state. It is beyond the scope of this chapter to present quantum mechanical calculations of the impact of the various degrees of freedom, which contribute to the KIE. An in depth treatment of many of the quantum mechanical aspects of isotope effects has been published by Biegeleisen and Wolfsberg (1958). The simple example of translational degrees of freedom will be used to give some idea about the magnitude of the inﬂuence of entropy terms on the KIEs. In order to be directly applicable to the atmosphere, the isotope

422

Volatile Organic Compounds in the Atmosphere

effects for compounds with natural abundance isotope ratios will be considered. The contribution of one degree of translational freedom to the partition function of molecule is proportional to the square root of the molecular mass. Consequently, the ratio of the one-dimensional (1D) translational partition functions for two isotopologues is the square root of the mass ratio. Based on TST, the change in partition function too has to be considered for both the reactant molecule and the transitions state, and the change in the rate constant due to a changes in molecular mass becomes (H mR · L mT )/(L mR · H mT ) for each translational degree of freedom, where mR and mT are the masses of the reacting VOC molecule and the transition state, respectively. The superscripts H and L denote the light and heavy isotopologues. Thus for the reaction of toluene with the √ OH-radical the change in entropy for one degree of translational freedom is (93 × 110)/(92 × 111) = 1.0009, equivalent to a change in KIE by less then 1‰. This effect is close to the uncertainty of state of the art KIE measurements. However, for the reaction of propene with ozone the corresponding contribution is 6‰, which is well above the uncertainty of current KIE measurements and constitutes a substantial contribution to the experimentally determined KIE of 9.5‰ (Section 10.4.5). It should be noted that the kinetic isotope effects calculated from TST for one degree of translational freedom are similar to the mass dependence of the collision frequency although the concepts behind the two approaches are different. Large isotope effects may result from changes in molecular symmetry. Since isotope labelling reduces the symmetry of symmetric molecules, the loss of symmetry due to the formation of the transition state is potentially lower for labelled molecules than for unlabelled molecules. However, although the VOCs studied to date include a number of highly symmetric molecules such as benzene, ethyne and ethene, no indication for an impact of molecular symmetry on KIEs has been found. In addition to fundamental theoretical approaches, there are some simple, partly empirical, relations between KIEs and chemical structure, which can be used to estimate the magnitude of KIEs, which have not been determined experimentally. For VOCs with similar chemical structure, the KIEs for a speciﬁc type of reaction, for example, hydrogen abstraction by OH radicals can be approximated by an inverse dependence on the number of atoms which can be replaced by the heavy isotope. The reason is that the natural abundance of the heavy carbon or hydrogen isotopes is so small that multiple labelling can be effectively excluded. Therefore, the probability of a reaction occurring at a labelled site is inversely proportional to the number of atoms in the reacting molecule. This is based on the assumption that the distribution of the heavy isotope in the molecule is random, which is a valid ﬁrst approximation for VOCs with natural isotope abundance. An example is shown in Figure 10.12 using the carbon KIEs for reactions of n-alkanes with Cl-atoms. It can be seen that the simple dependence of the KIE on the inverse of the number of carbon atoms gives a reasonable ﬁrst-order approximation. But it is also evident that this simple function underestimates the KIEs for n-alkanes with small carbon numbers and overestimates them for larger carbon numbers. A more detailed semi-empirical method to derive KIE estimates is the use of structurereactivity relationships (SRR). Expanding the SRR concepts developed by Greiner (1970) and Kwok and Atkinson (1995) to include isotopomers allows calculations of rate constants for labelled alkanes. From these rate constants the KIEs can be determined by comparison with conventional rate constants. This concept has been used by Tully et al. (1985, 1986) and

Gas Chromatography–Isotope Ratio Mass Spectrometry

423

Kinetic isotope effect () (‰)

12 10 8 6 4 2 0 0

2

4 6 Number of carbon atoms

8

10

Figure 10.12 Dependence of the KIEs for the reaction of n-alkanes with Cl-atoms on carbon number. The squares represent KIEs measured at room temperature and pressure taken from Anderson (2005). The solid line is a least square ﬁt of the experimental data to an inverse carbon number (NC ) dependence (KIE = 18.9‰/NC , R 2 = 0.974). The triangles are KIEs calculated for n-alkanes from an empirical SRR (for details see text).

Droege and Tully (1986a, 1986b) to derive site-speciﬁc rate constants for the abstraction of hydrogen and deuterium by the OH-radical from rate constant measurements using alkanes with and without deuterium labelling. Later, this concept has been extended to the carbon and hydrogen KIEs for reactions of VOCs with natural isotope abundance (Anderson et al. 2004b; Iannone et al. 2004, 2005). The principle of the isotope SRR can be directly derived from the SRR approach suggested by Kwok and Atkinson (1995) and will be explained here using the carbon KIEs for reaction of alkanes. 12

k=

i=n P

kP · F (Xi ) +

i=1

+

j=nS 12

k=n T

12

kS · F (Xj ) · F (Yj )

j=1 12

kT · F (Xk ) · F (Yk ) · F (Zk )

(10.33)

k=1

Here nP , nS and nT are the number of primary, secondary and tertiary carbon atoms −, − −CH2 − − and >CH− − groups), respectively. The overall rate constant for alkanes (i.e. CH3 − containing only 12 C atoms is 12 k. The group rate constants for reaction at primary, secondary and tertiary 12 C atoms are 12 kP , 12 kS and 12 kT , respectively. The factors F (X ), F (Y ) and F (Z ) reﬂect the dependence of the site-speciﬁc rate constant on the alkyl groups adjacent to the carbon atoms (Kwok and Atkinson 1995). This concept can be easily modiﬁed to include alkanes with 13 C labelling by introducing the corresponding group rate constants 13 kP , 13 kS and 13 kT for reaction at primary, secondary and tertiary 13 C atoms, respectively. The rate constant ktot for a speciﬁc alkane

424

Volatile Organic Compounds in the Atmosphere

isotopomer can then be calculated as follows:

k=

12 i= nP

j=12 nS 12

kP · F (Xi ) +

i=1

+

12

kS · F (Xj ) · F (Yj )

j=1

12 k= nT

12

kT · F (Xk ) · F (Yk ) · F (Zk ) +

13

kP · F (Xi )

i=1

k=1 j=13 nS

+

13 i= nP

13

kS · F (Xj ) · F (Yj ) +

13 k= nT

j=1

13

kT · F (Xk ) · F (Yk ) · F (Zk )

(10.34)

k=1

Here 13 n and 12 n refer to the number of carbon 13 C and 12 C atoms in the n-alkane and the indices P, S and T , as before, to the position at which the alkane is labelled. It should be noted that in principle the F parameters can different for different isotopes, but the presently available data do not justify usage of isotope speciﬁc F parameters. For a speciﬁc artiﬁcially labelled alkane the use of Equation 10.34 is straightforward, since in this case one usually has to deal with only one speciﬁc isotopomer. For compounds with natural abundance isotope ratios the situation is more complex since there usually will be several isotopomers. In principle, the average rate constant for the labelled isotopologue can be determined by calculating the rate constants for all possible isotopomers and averaging these rate constants according to the probability of the occurrence of a speciﬁc isotopomer. Using the probabilities Pi for labelling at any given position I , the following expression is obtained for the average rate constant, kav , of a mixture of isotopomers:

kav =

i=n P i=1

+

j=nS

(1 − Pi ) · 12 kP · F (Xi ) +

k=n T

(1 − Pj ) · 12 kS · F (Xj ) · F (Yj )

j=1

(1 − Pk ) ·

12

kT · F (Xk ) · F (Yk ) · F (Zk ) +

j=1

Pi · 13 kP · F (Xi )

i=1

k=1 j=nS

+

i=n P

Pj · 13 kS · F (Xj ) · F (Yj ) +

k=n T

Pk · 13 kT · F (Xk ) · F (Yk ) · F (Zk ) (10.35)

k=1

As a ﬁrst approximation, it is justiﬁed to assume that the probability of being 13 C labelled is identical for all carbon atoms in the molecule. It has been shown (Anderson et al. 2004b; Iannone et al. 2003) that, for natural abundance isotope ratios, the bias arising from this simpliﬁcation is small compared to the experimental uncertainty of measured KIEs. Furthermore, in the case of natural abundance isotope ratios, for small VOCs the probability of multiple labelling of one molecule is extremely small and, therefore, the probability of labelling at a speciﬁc site a molecule containing a 13 C atom is Pi = 1/NC , where NC is the number of carbon atoms. The rate constant 13 kav for reaction of a 13 C labelled alkane

Gas Chromatography–Isotope Ratio Mass Spectrometry

425

with natural 13 C abundance then can be calculated as follows: 13

kav =

NC − 1 · NC +

k=n T

12

i=n P

kP · F (Xi ) +

i=1

12

kS · F (Xj ) · F (Yj )

j=1

i=n P 1 13 kT · F (Xk ) · F (Yk ) · F (Zk ) + · kP · F (Xi ) NC i=1

k=1

j=nS

+

j=nS 12

13

kS · F (Xj ) · F (Yj ) +

j=1

k=n T

13

kT · F (Xk ) · F (Yk ) · F (Zk )

(10.36a)

k=1

There are numerous ways in which Equation 10.36a can be rearranged, but a very useful procedure is to replace the 13 C group rate constants (13 kP , 13 kS , 13 kT ) by the differences of the group rate constants (kP = 13 kP − 12 kP , kS = 13 kS − 12 kS , kT = 13 kT − 12 kT ). 13

kav =

i=n P

kP · F (Xi ) +

i=1

+

j=nS 12

k=n T k=1

12

kS · F (Xj ) · F (Yj )

j=1

12

1 kT · F (Xk ) · F (Yk ) · F (Zk ) + · NC

j=nS

+

kS · F (Xj ) · F (Yj ) +

j=1

k=n T

i=n P

kP · F (Xi )

i=1

kT · F (Xk ) · F (Yk ) · F (Zk )

(10.36b)

k=1

Combined with Equation 10.33 this can be used to obtain a convenient expression for 13 ktot . 13

kav =

12

+

1 ktot + · NC k=n T

i=n P

j=nS

kP · F (Xi ) +

i=1

kS · F (Xj ) · F (Yj )

j=1

kT · F (Xk ) · F (Yk ) · F (Zk )

(10.36c)

k=1

The KIE can then be calculated from the ratio of Equations 10.33 and 10.36. 12 k tot 13 k av

= 1+

1 12 k tot

+ kT ·

· NC

k=n T k=1

j=nS i=n P · kP · F (Xi ) + kS · F (Xj ) · F (Yj ) i=1

j=1

−1 F (Xk ) · F (Yk ) · F (Zk )

(10.36d)

426

Volatile Organic Compounds in the Atmosphere

In most cases, carbon isotope effects are small compared to the total rate constant and using the approximation 1/(1 + x) ≈ 1 − x, which is valid for small x, ε can be calculated from the following simpliﬁed expression: j=nS i=n P 1 · kP · F (Xi ) + kS · F (Xj ) · F (Yj ) εtot = − 12 ktot · NC i=1

+ kT ·

k=n T

j=1

F (Xk ) · F (Yk ) · F (Zk )

(10.36e)

k=1

By introducing site-speciﬁc ε-values (εP = 12 kP /(13 kP − 1), etc.) Equation 10.36e can be written in a form describing the KIE as a function of group-speciﬁc ε-parameters. i=n P 1 13 · εP · kP · F (Xi ) + εS · 13 kS εtot = 12 ktot · NC i=1

j=nS

×

F (Xj ) · F (Yj ) + εT · 13 kT ·

j=1

k=n T

F (Xk ) · F (Yk ) · F (Zk )

(10.36f )

k=1

Very often, the relative differences between group rate constants for labelled and unlabelled reaction sites are small (12 kP − 13 kP 12 kP , etc.). Equation 10.36 can then be written in a form that allows calculation of εtot from group-speciﬁc ε-values and the probability for reaction at a speciﬁc group. εtot =

1 · [εP · PP + εS · PS + εT · PT ] NC

(10.37)

This uses the following approximation. PP = 13 kP /12 ktot ·

i=n P

F (Xi ) ≈ 12 kP /12 ktot ·

i=1

≈ kP /ktot ·

i=n P

F (Xi )

i=n P

F (Xi )

i=1

(10.38)

i=1

This approximation results in larger bias if applied to group rate constants compared to molecular rate constants since, as can be seen from Equation 10.37, group-speciﬁc ε-parameters are on average by a factor of NC larger than the molecular ε-values. For most carbon KIEs, the bias is still well below the uncertainty of measured rate constants and KIEs. However, generally hydrogen KIEs are substantially higher than carbon KIEs and the simpliﬁcations leading to Equations 10.36e–10.38 may result in signiﬁcant bias. Equation 10.37 not only predicts an inverse dependence of ε on the number of carbon atoms, but also includes a contribution from site-speciﬁc differences in group rate constants. For group-independent ε-values Equation 10.37 gives results identical to the simple 1/NC dependence mentioned above. Similar to the procedure of determining group rate constants from sets of measured rate constants, the differences in group rate constants (k) or ε-values can be derived from sets

Gas Chromatography–Isotope Ratio Mass Spectrometry

427

of measured KIEs using Equation 10.36 or Equation 10.37. Usually published group rate constants are used as group rate constants for unlabelled compounds. The small difference, which arises from the use of natural abundance VOC to determine rate constants, is, due to the low natural abundance of 13 C or 2 H, much smaller than the experimental uncertainties of any measured rate constant. An example for the application of the isotope SRR is included in Figure 10.12. The 13 C group rate constants were determined from the experimental data by minimising the difference between observation and calculations. It is not surprising that the resulting calculated KIEs, which are based on the ﬁt of two parameters (13 kP and 13 kS ), show better agreement with the measurements than the simple 1/NC ﬁt, which uses only one ﬁt parameter. Nevertheless, from an empirical point of view it is obvious that the isotope SRR solves at least one of the major weaknesses of the simple 1/NC dependence, namely the systematic overestimate of KIEs for n-alkanes with four or more carbon atoms. There are also some quantum chemical calculations of KIEs for simple VOCs such as ethane, formaldehyde and acetaldehyde isotopologues (e.g. Beukes et al. 2000; D’Anna et al. 2003; Stutz et al. 1997). In several cases the theoretical calculations agree with experimental results within better than 20%. However, sometimes predictions and observations differ by a factor of 2 or higher.

10.4.2

Hydrogen KIEs for hydrogen abstraction reactions

Since reaction with OH-radicals is the most important atmospheric loss process for most VOCs, it is not surprising that studies of hydrogen KIEs for VOCs concentrated on this reaction. Still, there are also several studies of reactions with the Cl-atom and the NO3 -radical available and a couple of measurements of KIEs for VOC photolysis. The KIEs for the reaction of all possible deuterated isotopologues of methane with the OH-radical and the Cl-atom have been measured (Table 10.1). For comparison the KIEs calculated from an isotope SRR are included in Table 10.1. Since all H-atoms in methane Table 10.1 Hydrogen KIEs for the reaction of deuterated methanes with OHradicals and Cl-atoms Compounds

CH4 /CH3 D CH4 /CH2 D2 CH4 /CHD3 CH4 /CD4

Ratio rate constants OHa Experimental

Calculated from SRR

1.294 (0.018)b 1.81 (0.28)e 3.3 (0.5)e 7.36 (0.88)e

1.28 1.76 2.84 —

Ratio rate constants Cla Experimental

1.456 (0.006)c,d 2.43 (0.01)c 4.73 (0.04)c 14.7 (0.2)c

Calculated from SRR 1.30 1.87 3.32 —

a Values are for room temperature, the number in parenthesis give the error of the KIE.

The calculated values are derived from the KIEs of the completely deuterated isotopomer.

b From Saueressig et al. (2001). c From Feilberg et al. (2005a). d Tyler et al. (2000) reported a slightly different value of 1.474. e From Gierczak et al. (1997).

428

Volatile Organic Compounds in the Atmosphere

Table 10.2 Hydrogen KIEs for the reaction of alkanes with natural abundance isotope ratios and OH-radicals and Cl-atoms at room temperature and atmospheric pressure Compound

n-Butane i-Butane Cyclopentane n-Pentane n-Hexane n-Heptane n-Octane n-Nonane n-Decane

KIE for OH-reactiona,b , ‰

KIE for Cl-reactiona,c , ‰

51.6 (2.1) 97.3 (12.5) 63.2 (5.9) 65.9 (7.0) 52.8 (5.0) 38.9 (7.8) 33.4 (3.1) 29.6 (1.6) 29.0 (5.3)

39.6 (2.7)

28.2 (0.9) 24.6 (1.0) 24.0 (1.2) 17.9 (3.3) 15.1 (0.7) 14.9 (1.8)

a Values in parenthesis give the error of the KIE. b From Iannone et al. (2004). c From Iannone et al. (2005).

are identical, a group-speciﬁc KIE can be directly derived from the KIE of the completely deuterated compound. It can be seen that SRR calculations give reasonable ﬁrst-order estimates. In some cases, the agreement is within the uncertainty expected from experimental errors, but in other cases the difference between estimate and measured KIE is in the range of 15–40%, which is signiﬁcantly larger than the uncertainty of the experimental values. There have been a signiﬁcant number of studies on the reactions of alkanes with the OH-radical and the Cl-atom. Table 10.2 summarises the hydrogen KIEs for these reactions. Only KIEs measured for alkanes with natural abundance isotope ratios are listed. These KIEs are directly applicable to atmospheric conditions. For the reaction of n-alkanes with OH-radicals and Cl-atoms, the KIEs can be described by an inverse dependence on the number of hydrogen atoms (NH ) in the form of εOH = (640 ± 43)‰/NH (Iannone et al. 2004) and εCl = (347 ± 13)‰/NH (Iannone et al. 2005), respectively. A more reﬁned option to estimate KIEs from chemical structure is the use of isotope SRR (see Section 10.4.1). Tully et al. (1985, 1986) and Droege and Tully (1986a, 1986b) derived group-speciﬁc rate constant for deuterium labelling from measured rate constants for OH-radical reactions of alkanes with and without D labelling. They did not consider the impact of different alkyl groups on site-speciﬁc reactivity, which means that it was implicitly assumed that all F -parameters (see Section 10.4.1) are unity. Nevertheless, it was shown by Iannone et al. (2004) that the use of these site-speciﬁc rate constants resulted in very good agreement between calculated and measured KIEs. Only for i-butane and cyclohexane differences larger than the uncertainty of the measured values were observed. This difference may be due to the implicit use of F -parameters with a value of 1, but may also point towards the existence of other problems in extrapolating site-speciﬁc reactivity derived from multiply labelled VOC to VOC with natural isotope abundance. It should be noted that the group-speciﬁc isotope effects determined from studies of fully deuterated alkanes are very consistent. For example, Droege and Tully (1987a) measured KIEs of 2.71 and 2.59

Gas Chromatography–Isotope Ratio Mass Spectrometry

429

Table 10.3 Group speciﬁc hydrogen KIEs for reactions of alkanes with OH-radicals and Cl-atoms Group

Methane Primary carbon Secondary carbon Tertiary carbon

Ratio kH /kD for OH-reaction

Ratio kH /kD for Cl-reaction

7.36a 4.74c 2.61c 1.86e

14.7b 2.54d 1.41d

a From Gierczak et al. (1997). b From Feilberg et al. (2005a). c From Droege and Tully (1986b). d From Iannone et al. (2005). e From Tully et al. (1986).

for fully deuterated cyclopentane and cyclohexane, respectively. Within the experimental uncertainties this is identical to the KIE of 2.61 for deuterated methylene groups derived for aliphatic alkanes (secondary carbon Table 10.3). This suggests that there is no signiﬁcant isotope effect for the F-factors. Nevertheless, this does not eliminate the dependence of KIEs derived from isotope SRR on the values used as F-factors since the weight of the different group speciﬁc KIEs in the SRR calculation depends on the F-factors. However, presently the limited number of measured hydrogen KIEs for alkanes with natural isotope abundance and their experimental uncertainties are not sufﬁcient to allow ﬁrm conclusions on the origin of discrepancies between results from isotope SRR calculations and experimental observations. There are several studies of reactions of artiﬁcially D labelled alkanes with Cl-atoms, which allow estimates of group-speciﬁc KIEs for primary and secondary carbon atoms. The rate constants for the reaction of C2 D6 is very well studied (Chiltz et al. 1963; Dobis et al. 1994 and references therein; Hitsuda et al. 2001a). Thus the value for the C2 H6 /C2 D6 KIE is well established with an average of 2.89 and a 95% conﬁdence interval of 0.15. Using a value of 2.88 and their measurements of the rate constants for reaction of n-C4 H10 and n-C4 D10 with Cl-atoms Stutz et al. (1998) derived a KIE of 1.2 ± 0.2 for hydrogen abstraction from a secondary carbon atom. This fully agrees with the ﬁnding of 1.20 ± 0.06 for the ratio of the rate constants of cyclohexane over that of fully deuterated cyclohexane (Aschmann and Atkinson 1995). Hitsuda et al. (2001b) report a KIE of 1.55 ± 0.12 for C3 H8 /C3 D8 , which is consistent with the KIE of 1.56 predicted by isotope SRR. Iannone et al. (2005) derived site-speciﬁc reactivities from KIEs measured using n-alkanes with natural D abundance. They derived group KIEs of 2.54 and 1.41 for hydrogen abstraction by Cl-atoms from primary and secondary carbon, respectively. They mention that, since the reactions they studied were mainly dominated by reaction at the secondary (methylene group) carbon atoms, the group rate constant they derived for the methyl group has considerable uncertainty. Nevertheless, their values agree very well with the values derived from studies of deuterated alkanes. Although there is no obvious quantitative relation between the degrees of alkyl substitution of the carbon atom and the magnitude of the group KIE, there is strong evidence

430

Volatile Organic Compounds in the Atmosphere

for a systematic qualitative dependence. The presently known group KIEs for reactions with OH-radicals and Cl-atoms increase in the order tertiary < secondary < primary < methane (Table 10.3). This is consistent with the idea that the hydrogen kinetic isotope effect for hydrogen abstraction from carbon increases with increasing bond strength. In addition to the studies of alkanes, there are also numerous studies of the hydrogen KIEs for abstraction of hydrogen from other deuterium labelled VOCs. These measurements were generally made to study reaction mechanisms, and are rarely directly applicable to VOCs with natural abundance isotope ratios. It has already been mentioned that the calculation of KIEs for VOCs with natural abundance isotope ratios from studies of artiﬁcially labelled VOCs is not always possible and often requires detailed studies of VOCs with labelling at different positions. Nevertheless, in a number of simple cases isotope SRR estimates are straightforward and are possible even in the absence of detailed studies of group-speciﬁc KIEs. Table 10.4 lists a number of experimentally determined KIEs for reactions of VOC with artiﬁcial D-labelling. Also included are KIEs estimated for natural abundance isotope ratios using isotope SRR. These estimates are based on a random distribution of the D-atom in the VOCs. For VOCs where all hydrogen atoms are equivalent, such as chloromethane, formaldehyde or acetone, the group-speciﬁc KIE is identical to the KIE of the completely deuterium labelled isotopologue. Similarly, if the reaction of a VOC is dominated by abstraction from only one group of hydrogen atoms, such as reactions of aldehydes, alcohols or formic acid with the OH-radical, the group-speciﬁc KIE is identical to the KIE of the VOCs completely labelled at the reaction sites. Strictly speaking this is a simpliﬁcation, since for VOCs with different H-containing groups very seldom 100% of the overall reaction occurs at the same site. For example, the reactions of methanol and ethanol with the OH-radical proceed primarily via abstraction of hydrogen from the carbon atom with the hydroxyl group. However, 10–15% of the reaction occurs at other hydrogen atoms. Nevertheless, the inﬂuence of these minor reaction pathways is generally small. For the reactions of acetaldehyde with halogen atoms, the OH radical and the NO3 radical and the reaction of ethanol with Cl-atoms the available studies of isotopologues allow determination of group-speciﬁc KIEs −H bonds. The results from calculations using the complete set of group KIEs for all C− differ only marginally from those based solely on the group KIE for the dominant reaction (Table 10.4). Similarly, isotope SRR calculations of the KIEs for reactions of deuterated acetaldehydes using SRR gives results consistent with experimental observations. Droege and Tully (1987b) and Harry et al. (1999) studied the reactions of fully deuterated symmetric n-alkyl ethers with the OH-radical. The average KIEs for the three studied ethers (Table 10.4) is 1.96 with a standard deviation of 0.15. This low variability of the KIE may have two reasons. Either the reaction is always dominated by abstraction from one group only, for example, the ethylene group adjacent to the hydroxyl function, or all group KIEs are effectively identical. In both cases SRR estimates for ethers with natural abundance isotope ratios give identical KIEs. Harry et al. (1999) also report rate constants for reaction of two fully deuterated ethers with Cl-atoms. Similar to the OH-reaction KIEs, these two KIEs are identical within the experimental error of the measurements (Table 10.4). It seems, therefore, plausible to apply an approach identical to that used for the OH-reactions of n-alkyl ethers. Indeed, such treatment gives consistent results (see Table 10.4). However, it has to be considered that for both reaction types the number of available experimental data is very small and

Gas Chromatography–Isotope Ratio Mass Spectrometry

431

Table 10.4 KIEs for the reaction of deuterated oxygen containing VOC and methyl chloride with OH-radicals and Cl-atoms at ambient temperature Compounds

Reactant

Ratio rate constants Experimenta

CH2 O/CD2 O CH2 O/CHDO CH3 OH/CD3 OD CH3 OH/CDH3 O HCOOH/DCOOH HCOOH/DCOOD HCOOH/DHCO2 CH3 Cl/CD3 Cl CH3 Cl/CH2 DCl CH3 CHO/CD3 CDO CH3 CHO/CH3 CDO CH3 CHO/CD3 CHO CH3 CHO/C2 DH3 O C2 H5 OH/C2 D5 OD C2 H5 OH/C2 DH5 O CH3 COCH3 /CD3 COCD3 CH3 COCH3 /CH2 DCOCH3 (C2 H5 )2 O/(C2 D5 )2 O (C2 H5 )2 O/C4 DH9 O (n-C3 H7 )2 O/(n-C3 D7 )2 O (n-C3 H7 )2 O/C6 DH13 O (n-C4 H9 )2 O/(n-C4 D9 )2 O (n-C4 H9 )2 O/C8 DH17 O

OH OH OH OH OH OH/OD OH OH OH OH OH OH OH OH OH OH OH OH OH OH OH OH OH

CH2 O/CD2 O CH2 O/CHDO CH3 Cl/CD3 Cl CH3 Cl/CH2 DCl CH3 CHO/CD3 CDO CH3 CHO/CH3 CDO CH3 CHO/CD3 CHO CH3 CHO/C2 DH3 O C2 H5 OH/C2 H5 OD C2 H5 OH/CD3 CH2 OH C2 H5 OH/CH3 CD2 OH C2 H5 OH/C2 DH5 O (n-C3 H7 )2 O/(n-C3 D7 )2 O (n-C3 H7 )2 O/C6 DH13 O (n-C4 H9 )2 O/(n-C4 D9 )2 O (n-C4 H9 )2 O/C8 DH17 O

Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl

CH2 O/CD2 O CH2 O/CHDO

Br Br

SRR estimatesb

1.62 (0.08)c 1.24 2.67 (0.02)d (1.19) 1.12 (0.03)e 7.02 (1.4)e 3.9 (0.4)f 1.65 (0.08)c 1.42 (0.1)c 1.13 (0.04)c

(1.75) — 1.33 1.70

1.11 (1.08) 2.9 (0.2)d 1.12 6.03 (0.49)d,g 1.16 2.00 (0.15)h 1.049 (1.053) 1.93 (0.10)i 1.038 (1.036) 2.21 (0.18)i 1.031 (1.035) 1.30 (0.14)j 4.91 (007)f 1.32 (0.18)j 1.34 (0.23)j 1.05 (0.025)

1.13 — 1.46 1.47 (1.30) 1.43 (1.30) 1.13 (1.00) 1.065 (1.061)

0.93 (0.17)k 1.21 (0.12)l 2.02 (0.19)k 1.13 (1.09) 1.35 (0.15)i 1.019 (1.019) 1.28 (0.02)i 1.014 (1.012) 7.5 (0.4)j 1.76

432

Volatile Organic Compounds in the Atmosphere

Table 10.4 (Continued) Compounds

Reactant

Ratio rate constants Experimenta

CH3 CHO/CD3 CDO CH3 CHO/CH3 CDO CH3 CHO/CD3 CHO CH3 CHO/C2 DH3 O

Br Br Br Br

3.79 (0.29)j 3.98 (0.26)j 1.04 (0.22)

CH2 O/CD2 O CH2 O/CHDO CH3 CHO/CD3 CHO CH3 CHO/CH3 CDO CH3 CHO/C2 DH3 O

NO3 NO3 NO3 NO3 NO3

2.91 (0.14)c

CH2 O/CD2 O CH2 O/CHDO

hν hν

SRR estimatesb 3.79 (3.80) 3.58 (3.8) 1.015 (1.00) 1.23 (1.23) 1.49

1.19 (0.11)c 1.65 (0.08)b 1.180, (1.177) 1.67 (0.003)m 1.25

a Values in parenthesis give the error of the KIE. b The values in parenthesis give results of SRR calculations based on the simpliﬁed

approach that the reaction is dominated entirely by a single reaction pathway. For details see text. c From D’Anna et al. (2003). d From Wallington et al. (1988). e From Singleton et al. (1988). f From Gola et al. (2005). g Farkas et al. (2003) report a value of 5.33 ± 0.41. h From Droege and Tully (1987b). i From Harry et al. (1999). j From Beukes et al. (2000). k From Crawford et al. (2004). l From Taatjes et al. (1999). m From Feilberg et al. (2005a).

therefore any interpretation based on generalising the observations results in a considerable uncertainty. Overall, the use of the isotope SRR concept to derive hydrogen KIEs for abstraction reactions of VOCs with natural abundance isotope ratios is very promising. However, it has to be remembered that the number of experimental studies of VOCs with natural isotope ratios are very limited. Therefore, with the exception of reactions of alkanes with the OH-radical or the Cl-atom, we presently cannot evaluate the validity of isotope SRR estimates for compounds other than alkanes.

10.4.3

Hydrogen KIEs for reactions of unsaturated VOCs

The available data on hydrogen isotope effects for reactions of unsaturated VOCs is very limited. The main reason for this is that the KIEs are usually small and the experimental uncertainty of the measurement is very often larger than the difference in the rate constants for the different isotopologues.

Gas Chromatography–Isotope Ratio Mass Spectrometry

433

Table 10.5 KIEs for the reaction of deuterated ethenes with OH-radicals and the halogen-atoms Ratio rate constantsa

Compounds

C2 H4 /C2 D4 C2 H4 /C2 H3 D

OH

Cl

Br

1.03 (0.06)b

0.738 (0.024)c 0.853 (0.037)c

0.817 (0.016)c 1.09 (0.013)c

a The values are for room temperature and approximately atmospheric pressure.

Numbers in parenthesis give the error of the KIE.

b From Niki et al. (1978). c From Enghoff et al. (2003).

Iannone et al. (2004) measured the KIEs for the reactions of cyclohexene, 1-heptene, toluene, ethylbenzene and p-xylene with the OH-radical using VOC with natural abundance isotope ratios. The ε-values they report are in the range between 10‰ and 30‰; on average a factor of 3–4 lower that the ε-values they found for reactions of alkanes with the same number of hydrogen atoms. Values for rate constants of the reactions of several deuterated alkenes with the OH-radical and halogen atoms were reported by several authors (Enghoff et al. 2003; Niki et al. 1978; Stutz et al. 1997, 1998). However, with the exception of reactions of ethene isotopologues, all these rate constants were within the uncertainty range of the rate constants for unlabelled alkenes. Therefore these measurements do not provide meaningful KIEs for atmospheric VOC studies for alkenes other than ethane. Nevertheless, they support the ﬁnding of very small hydrogen KIEs by Iannone et al. (2004). Compared to reactions of alkanes, small hydrogen KIEs for reactions of unsaturated compounds can be explained by differences in reaction mechanisms. While alkanes react with the OH-radicals via abstraction of a hydrogen atom, which means that the reaction directly involves breaking of a hydrogen–carbon bond, unsaturated compounds react primarily by addition of OH to the double bond or aromatic ring. Such secondary isotope effects are typically much smaller than primary isotope effects, and it is not surprising that reactions of unsaturated VOCs have signiﬁcantly smaller hydrogen KIEs than saturated compounds. However, it has to be considered that there is often a small contribution from a hydrogen abstraction channel to the total reaction of the OH-radical with unsaturated VOCs. The possible contribution of this channel to the overall KIE is difﬁcult to estimate. The contribution from the abstraction channel is generally not precisely known and the site-speciﬁc KIEs for abstraction from unsaturated VOCs have not been determined. Using a range of 5–10% for the contribution of the abstraction channel and the KIEs observed for abstraction from alkanes, the contribution from abstraction to the overall ε-value is between 3‰ and 10‰. This is lower than the measured KIEs and suggests that there is a small secondary hydrogen isotope effect for the addition of the OH-radical to unsaturated VOCs. As mentioned above, hydrogen KIEs for the reaction of ethene do not ﬁt into the general behaviour found for other alkenes. Table 10.5 summarises the KIEs measured for deuterated ethenes at pressures and temperatures compatible with the conditions in the lower troposphere. In contrast to all other KIEs observed for the gas-phase reactions of VOCs with OH-radicals or halogen atoms, several of these reactions exhibit inverse isotope effects.

434

Volatile Organic Compounds in the Atmosphere

Stutz et al. (1997) studied the reaction of C2 H4 and C2 D4 with Cl-atoms in He at 1 Torr pressure. They found that the reaction of completely deuterated ethene is a factor of 2.7 faster than the reaction of the non-deuterated isotopologue. They explained this ﬁnding with the increased density of the vibrational states in the deuterated compound, which results in a slower decomposition of the excited C2 D4 Cl# adduct compared to the nondeuterated adduct. A similar effect has been reported for the low-pressure third-order rate constant of the reaction of OH with deuterated and non-deuterated ethene (unpublished data of F.P. Tully cited from Atkinson 1989). For pressure regimes exceeding the high-pressure limit the different adduct stabilities have no impact on the overall rate constants. At atmospheric pressure the rate constant for the reaction of ethene with the OH-radical has reached the high-pressure limit, whereas the reactions with Cl and Br atoms are still below the high-pressure limit. The consequence is that at ambient pressure for reactions of ethene with OH no inverse isotope effect is observed, whereas for reactions with halogens inverse KIEs are found (Table 10.5).

10.4.4

Carbon KIEs for hydrogen abstraction reactions

The KIEs for reactions of a substantial number of alkanes with natural abundance isotope ratios have been measured (Anderson 2005; Anderson et al. 2004b; Rudolph et al. 2000). In addition to the most important atmospheric removal process for alkanes, the reaction with OH-radicals, also the reaction with Cl-atoms has been studied. The KIEs are listed in Table 10.6. Also included in Table 10.6 are KIEs determined for artiﬁcially 13 C labelled VOC that contain only one carbon atom. These KIEs are directly applicable to atmospheric VOCs. The Teller-Redlich rule predicts that the relative change in vibration frequencies for any chemical bond resulting from isotope substitution in a molecule is inversely proportional to the square root of the ratio of the isotope masses. Based on this replacing √ a hydrogen atom by = 0.7071, whereas a deuterium atom will change vibration frequencies by a factor of 1/2 √ the change from substituting 12 C by 13 C will only amount to a factor of 12/13 = 0.9608. The consequence of this is that changes in reaction rates due to labelling with 13 C are generally small compared to deuterium labelling. The experimentally determined carbon KIEs for the reactions of alkanes are fully consistent with this expectation. A consequence of these small changes in vibration frequencies is that contributions to isotope effects that depend primarily on the ratios of the molecular masses such as the partition functions for translations and rotations will often be of similar magnitude as the contribution from changes in vibration frequencies. As mentioned previously (Section 10.4.1), translational and rotational degrees of freedom generally contribute only a few per mille or less to the KIEs, decreasing in magnitude with increasing mass of the reacting molecules. This can be compared with the impact of changes in vibration frequencies. Using the frequency −H− −Cl estimates of Donahue (2001) and Donahue et al. (1998) for the newly formed C− −H− −O bonds for reactions of alkanes with the Cl-atom and OH-radical, respectively, and C− the magnitude of contributions from vibrational partition functions and changes in zero point energy can be estimated. The results of such estimates show that these effects are also in the range of a few per mille (Anderson 2005; Anderson et al. 2004b), decreasing with increasing number of carbon atoms of the reacting alkane. The observed KIEs for reactions of alkanes with the OH-radical or the Cl-atom range from 1.5‰ to 11‰ (Table 10.6),

Gas Chromatography–Isotope Ratio Mass Spectrometry

435

Table 10.6 Room temperature carbon kinetic isotope effects for hydrogen abstraction reactions by OH-radicals and Cl-atoms from VOC with natural abundance isotope ratios Compound Methane Ethane Propane n-Butane Methylpropane n-Pentane Methylbutane n-Hexane n-Heptane n-Octane Cyclopentane Cyclohexane Methylcyclopentane Formaldehyde Methylchloride

εOH , ‰a 3.9 (0.4)b 8.57 (1.95)d 5.46 (0.35)d 5.16 (0.67)d 8.45 (1.49)d 3.26 (0.64)d 2.91 (0.43)d 2.20 (0.07)d 1.96 (0.26)d 2.13 (0.39)d 1.84 (0.13)d 4.46 (0.51)d 1.77 (0.53)d −30 (110)f 59 (8)h

εCl , ‰a 62.1 (0.1)c 10.73 (0.20)e 6.44 (0.14)e 3.94 (0.01)e 6.18 (0.18)e 3.22 (0.17)e 1.79 (0.42)e 2.02 (0.4)e 2.06 (0.19)e 1.54 (0.15)e 3.04 (0.09)e 2.30 (0.09)e 2.56 (0.25)e 70 (26)g 70 (10)h

a The values are for room temperature and atmospheric pressure. Numbers in parenthesis give the error of the KIE. b From Saueressig et al. (2001). c From Tyler et al. (2000). d From Anderson et al. (2004b). e From Anderson (2005). f From D’Anna et al. (2003). Niki et al. (1984) report a value of 1.10±0.14. g From Beukes et al. (2000). h From Gola et al. (2005).

which is qualitatively consistent with these estimates. Consequently, the overall KIEs are the result of several contributions of approximately similar magnitude, which presents a rather complex situation. Nevertheless, for the KIEs for the reaction of alkanes, with the exception of methane, a useful ﬁrst-order estimate can be obtained from simple inverse dependencies on carbon number. For the reaction with OH-radicals Anderson et al. (2004b) found that a dependence of the form OH ε = (16.6 ± 1.0)‰ × NC−1 can explain 75% of the observed variability of KIEs for alkane–OH reactions. Similarly the dependence Cl ε = (18.9 ± 1.3)‰ × NC−1 describes 85% of the variance for alkane–Cl KIEs (Anderson 2005). Considering the various factors contributing to the carbon KIEs for alkane reactions, this good agreement with an inverse carbon number dependence is somewhat surprising. However, it has to be considered that isotope effects due to the molecular mass dependence of translational and rotational partition functions will also result in a carbon number dependence; a consequence of the simple relation between molecular mass and carbon number for alkanes. Rudolph et al. (2000) argue that the mass dependence of the collision frequency can explain a signiﬁcant portion of the KIEs observed for reaction of alkanes with

436

Volatile Organic Compounds in the Atmosphere

the OH-radical. Indeed, in a very simple approach, the mass dependence of the collision frequency for reactions of alkanes with 2 to 8 carbon atoms and the OH-radical or Cl-atoms can be approximated by OH ε = (13.5 ± 1.2)‰ × NC−1 (R 2 = 0.948) and Cl ε = (18.9 ± 1.2)‰ × N −1 (R 2 = 0.970), which is in surprisingly good agreement with C the ﬁts to the experimental data. However, a more detailed look at the data reveals some signiﬁcant differences. KIE predictions based on the mass dependence of the collision frequencies alone underestimate the KIEs for alkanes with more than four carbon atoms and also cannot explain the differences in KIEs between alkane isomers. Anderson et al. (2004b) and Anderson (2005) derived group-speciﬁc KIEs for the reactions of alkanes with OH-radicals and Cl-atoms. Although carbon KIEs estimated by isotope SRR agreed better with observations than estimates from a simple inverse carbon number dependence, isotope SRR calculations are less successful for carbon KIE predictions than for hydrogen KIE estimates. To some extent, this can be explained by experimental uncertainties and the smaller number of experimental data. Carbon KIEs are small and thus experimental uncertainties result in larger relative uncertainties of group-speciﬁc KIEs derived from experimental data. Another problem is the larger relative importance of molecular-massdependent isotope effects resulting from changes in translational and rotational partition functions. Although in principle such factors can be incorporated into a SRR, presently the experimental data this would require are not available. There is a measurement of the carbon KIE for reaction of formaldehyde with the NO3 radical (D’Anna et al. 2003). The measured value of ε = −30‰ with a statistical 3σ error of 20‰ suggests that this reaction has an inverse isotope effect. However, the reported error does not include possible systematic bias for the experiment. The reported value may, therefore, also be compatible with a very small normal KIE, similar to other carbon KIEs for hydrogen abstraction reactions. The reaction of formaldehyde with the Br-atoms, which can be derived from the measurements of Beukes et al. (2000), has a normal carbon isotope effect of ε = (100 ± 76)‰. The ε-value for photolysis of formaldehyde by sunlight was measured by Feilberg et al. (2005b) to be (119±8)‰ in July in southern Spain. This KIE is approximately by a factor of ﬁve higher than the typical carbon KIEs for hydrogen abstraction reactions (Table 10.6) and is likely to have a major impact on the carbon isotope ratio of atmospheric formaldehyde. However, the photolytic removal of formaldehyde consists of two different reaction channels with different wavelength dependence. Therefore, it is possible that the KIE for photolysis of formaldehyde depends on the relative wavelength dependence of the photon ﬂux density.

10.4.5

Carbon KIEs for reactions of unsaturated VOC

Table 10.7 lists carbon KIEs for reaction of unsaturated VOCs with natural abundance isotope ratios and OH-radicals, Cl-atoms and ozone. The carbon KIEs for reactions of unsaturated VOCs with OH-radicals and ozone are substantially larger than those for reactions of saturated VOCs with OH-radicals. This can qualitatively be explained by the different types of reaction mechanisms. Most of the reactions of unsaturated VOCs with OH-radicals −O bonds. Although in or ozone are addition reactions resulting in the formation of new C− the transition state these new bonds will only be formed partially, probably in the form of an addition complex to the π-bonds between the carbon atoms, the vibration frequencies

Gas Chromatography–Isotope Ratio Mass Spectrometry

437

Table 10.7 Carbon kinetic isotope effects for reactions of ozone, OH-radicals and Cl-atoms with unsaturated VOC with natural isotope abundance at room temperature and atmospheric pressure Compound Ethene Propene 1-Butene E-2-Butene 1,3-Butadiene 1-Pentene Z-2-Pentene Cyclopentene Isoprene 1-Hexene Cyclohexene 1-Heptene 1-Octene 1-Nonene Methacrolein Methyl vinyl ketone Ethyne Benzene Toluene Ethylbenzene o-Xylene p-Xylene 1,2,4-Trimethylbenzene o-Ethyltoluene

εOH , ‰a

εCl , ‰a

εO3 , ‰a

18.6 (2.9)b 11.7 (0.19)e 7.40 (0.32)e 9.34 (1.98)e 7.55 (0.88)e

5.65 (0.34)c 5.56 (0.18)c 5.93 (1.16)c

18.85 (2.80)d 9.49 (2.47)d 8.7 (0.96)d 8.05 (0.35)d 7.91 (0.35)d 6.72 (0.87)d 7.32 (0.15)d 6.68 (0.73)d 6.05 (0.99)d /8.15 (0.1)g 5.04 (0.69)d 5.64 (0.64)d 4.31 (0.69)d /5.44 (0.09)g 4.75 (0.14)g 4.55 (0.37)g 7.18 (0.28)g 7.77 (0.08)g

4.86 (0.63)c 3.75 (0.14)c 6.94 (0.8)e /6.56 (0.12)f 5.22 (0.41)e

5.88 (0.08)f 6.78 (0.14)f 15.80 (0.61)e 8.13 (0.8)e /7.53 (0.5)h 5.95 (0.28)h 4.34 (0.28)h 4.27 (0.05)h 4.83 (0.05)h 3.18 (0.09)h 4.71 (0.12)h

2.89 (0.31)c 2.17 (0.17)c 1.85 (0.54)c

a The values are for room temperature and atmospheric pressure. Numbers in parenthesis give the error of the KIE. b From Anderson et al. (2004b). c From Anderson (2005). d From Iannone et al. (2003). e From Rudolph et al. (2000). f R. Iannone, R. Koppmann and J. Rudolph, unpublished results. g From Iannone et al. (2005). h From Anderson et al. (2004a).

for these bonds will depend on the mass of the carbon atoms involved. Furthermore, it has to be considered that an addition to unsaturated VOCs involves changes to the bonds of two carbon atoms, which for singly labelled VOCs doubles the probability for the occurrence of a primary isotope effect. The KIEs for reactions of alkenes with the OH-radical and ozone can be described by nearly identical carbon number dependencies. For the reaction with OH radicals, the presently available data can be described by OH ε = (34.6 ± 1.0)‰ × NC−1 with a R 2 of 0.961. For reaction with ozone, the dependence is OH ε = (34.0 ± 1.0)‰ × NC−1 with an R 2 of 0.964. To some extent this extremely good agreement may be fortuitous and indeed recent results (Table 10.7) for reactions of ozone indicate that the KIEs are somewhat higher

438

Volatile Organic Compounds in the Atmosphere

than those reported by Iannone et al. (2003). Nevertheless, the dependence between KIE and carbon number derived from these values alone, OH ε = (42.5 ± 2.9)‰ × NC−1 , still suggests similarities between the carbon KIEs for reactions of alkenes with the OH-radical and ozone. The measured KIEs for the reactions of methacrolein and methyl vinyl ketone are somewhat lower than KIEs predicted from the carbon number dependence of alkene KIEs whereas those for reactions of aromatic compounds are somewhat higher. However, in both cases the number of available measurements is quite limited and not sufﬁcient to prove the existence of a systematic difference. The KIEs for reaction with Cl-atoms are approximately a factor of 2–3 lower than those for reaction with ozone or OH-radicals. For aromatic compounds the reason for this is a difference in reaction mechanisms. In contrast to the reaction with OH-radicals, which predominantly proceeds via addition to the aromatic ring, the reaction with Cl-atoms is dominated by abstraction of a hydrogen atom from the alkyl groups. Indeed, the KIEs measured for reaction of alkylbenzenes with Cl-atoms are very similar in magnitude to KIEs for reaction of Cl-atoms with alkanes. The carbon number dependence of the KIEs for reaction of alkanes with Cl-atoms (Section 10.4.4) predicts a KIE of 2.7‰ and 2.4‰ for alkanes with 7 and 8 C-atoms, respectively. Within the experimental uncertainties, this agrees with the measured KIEs for C-atom reactions of alkylbenzenes (Table 10.7). There is no obvious explanation for the small Cl-KIEs of the alkenes. The small KIE found for ethene may be explained by arguments similar to those used to explain the observed inverse KIE for reaction of deuterated ethene with Cl-atoms. At atmospheric pressure the rate constant for the reaction of ethene with Cl-atoms is below the highpressure limit and, therefore, more efﬁcient adduct stabilisation for 13 C labelled ethene may be the reason for the lower KIE. However, this does not explain the small KIEs measured for the C4 and C5 alkenes. For these alkenes at atmospheric pressure, the high-pressure limit of the reaction rate is reached and, therefore, differences in adduct stability will not have an impact on the KIE. Another possible explanation for the small Cl-alkene KIEs is the contribution of hydrogen abstraction reactions. However, the relative contribution of abstraction is generally small and the KIEs hydrogen abstraction by Cl-atoms is also small. Unless the KIEs for hydrogen abstraction from alkenes are signiﬁcantly larger than for other abstraction reactions, abstraction reactions will only contribute marginally to the experimentally observed KIE. Another factor which has to be considered for reactions of alkenes with Cl-atoms is the very high rate constant, which for C3 and heavier alkenes approaches the collision-controlled regime. For collision-controlled reactions, differences in zero point energies, which result in differences in activation energy, will have little impact on the reaction rate constants and thus the KIE. Although there are possible qualitative explanations of the small carbon KIEs for reactions of alkenes with Cl-atoms, the available experimental data do not present a clear picture, which would allow deriving systematic dependencies between structure and KIE. Another observation, which cannot easily be explained from our current knowledge, is the unexpectedly high KIE for the reaction of 13 C2 H4 (i.e. doubly 13 C labelled ethene) with Cl- and Br-atoms. Enghoff et al. (2003) report values of 1.061 ± 0.023 and 1.117 ± 0.053, respectively. The simple dependence of KIEs on the probability of reactions occurring at labelled sites cannot explain that the ε-values for doubly labelled ethene are

Gas Chromatography–Isotope Ratio Mass Spectrometry

439

approximately a factor of ten higher than for the corresponding reactions of singly labelled ethene.

10.4.6

Temperature dependence of KIEs

The temperature dependence of KIEs is determined by the changes in zero point energies and the vibrational degrees of freedom, but it is independent of the translational or rotational partition functions. The resulting possibility of using the temperature dependence of KIEs for differentiating between the inﬂuences of vibrational and translational or rotational degrees of freedom has triggered a large number of studies. However, the vast majority of these studies were conducted for conditions that are not relevant for reactions of VOCs at atmospheric temperature and pressure. Moreover, the temperature dependence of KIEs is often only small, and many studies only give upper limits for the temperature dependence of KIEs. Nevertheless, the presently available information allows a ﬁrst-order estimate of the importance of the temperature dependence of KIEs for VOCs under atmospheric conditions. Anderson (2005) studied the temperature dependence of carbon KIEs for reaction of −C8 alkanes and alkylbenzenes with natural isotope abundance. The changes of KIEs C6 − observed for the temperature range between 288 and 373 K were either below or very close to the uncertainty of the measurements. For the reactions of alkanes with the OH-radical, the best estimate for the temperature dependence corresponds to a change of 0.005‰/K with a 2σ upper limit of 0.025‰/K. For the reaction of aromatic compounds with OH-radicals the best estimate was 0.007‰/K with a 2σ upper limit of 0.03‰/K. The temperature dependence found for the reactions of alkanes with Cl-atoms averaged around 0.001‰/K with a 2σ upper limit of 0.025‰. Only for the reaction of aromatic compounds with the OH radical at temperatures exceeding 373 K, Anderson (2005) observed signiﬁcant changes of the KIE with temperature. This change in temperature dependence was explained by a change in reaction mechanism. However, this temperature range is not relevant for the atmosphere. The studies of site-speciﬁc hydrogen isotope effects for alkanes with the OH-radical (Droege et al. 1986a, 1986b, 1987a, 1987b; Tully et al. 1985, 1986) include measurements of the temperature dependence of the group rate constants. This allows estimates of the temperature dependence of the hydrogen KIE for alkanes using isotope SRR. For ethane, the calculated hydrogen KIE decreases by 0.2‰/K. For all other alkanes the temperature dependence is less than 0.1‰/K. Taatjes et al. (1999) measured the temperature dependence of the KIEs for the reaction of three ethanol isotopologues with deuterium labelling at different sites for reaction with Cl-atoms. For ethanol with natural abundance isotope ratios isotope SRR estimates give a value of approximately 0.02‰/K for the KIE temperature dependence. In summary, the available information suggests that the temperature dependence of the KIEs for reactions of VOCs is small and for the temperature range relevant for the atmosphere the systematic changes of the KIE with temperature most likely will be smaller than the uncertainties of the KIEs. This is consistent with the general expectation that the temperature dependence of KIEs for reactions of organic compounds with natural isotope abundance is small. However, it should be remembered that the presently available set of

440

Volatile Organic Compounds in the Atmosphere

measurements of KIE temperature dependencies, which is relevant for atmospheric VOC, is very limited.

10.4.7

Isotope ratios of products of VOC reactions

Isotope ratio studies also have the potential to provide additional insight into the atmospheric chemistry of the products of VOC oxidation, including the formation of secondary Particulate Organic Matter (POM). A simple relation between the product’s isotope ratio at time t (t δPROD ), the reactant concentrations and the isotope ratios at time t (Lt cVOC and L t δVOC , respectively) and t = 0 (0 cVOC and 0 δVOC , respectively) can be derived from the principle of mass conservation. 0 cVOC

· 0 δVOC = t cVOC · t δVOC + (0 cVOC − t cVOC ) · t δPROD

(10.39a)

Equation 10.39a can be rearranged to give the product’s isotope delta value as function of the ratio of initial VOC concentration over VOC concentration at time t . 0 δVOC − (t cVOC /0 cVOC ) · t δVOC (10.39b) t δPROD = 1 − (t cVOC /0 cVOC ) It should be noted that, strictly speaking, Equation 10.39b should use the concentration ratios of the light VOC-isotopologue. However, for carbon and hydrogen isotopes the small bias introduced by replacing this ratio with the ratio of the VOC concentrations is generally negligible compared to the uncertainties of VOC concentration measurements. For a closed system, this mass budget can be combined with the dependence between concentration and isotope ratios as described by a Rayleigh type function. For example, the concentration ratio in Equation 10.39b can be expressed as a function of isotope ratios by rearranging Equation 10.16b. t δPROD

=

0 δVOC

−1

−1

− ((1 + t δVOC )/(1 + 0 δVOC ))(KIEOH −1)

· t δVOC

(10.40a)

−1 (KIEOH −1)−1

1 − ((1 + t δVOC )/(1 + 0 δVOC ))

Alternatively, Equations 10.39b and 10.16a can be combined to describe the change of the product’s delta value as function of the extent of precursor processing. t δPROD =

0 δVOC

−1

− (t cVOC /L0 cVOC ) · {(1 + 0 δ) · (t cVOC /0 cVOC )(KIEOH −1) − 1} 1 − (t cVOC /0 cVOC ) −1

(0 δVOC + 1) · (1 − (t cVOC /0 cVOC )(KIEOH ) ) −1 = 1 − (t cVOC /0 cVOC )

(10.40b)

Equation 10.40b predicts a deﬁned relation between the isotope ratio of the products and the change of the precursor’s isotope ratio due to chemical reaction, which can be used for ambient studies of VOC oxidation products. However, Equation 10.40 is strictly valid for an isolated air mass only, in the case of mixing air masses with different concentrations and application of the relation between oxidation product and precursor for the real world will usually require numerical model simulations. A more fundamental problem is due to the fact that Equation 10.40 is based on mass budget considerations and thus valid only for the total of all products. It does not directly allow predictions of the isotope

Gas Chromatography–Isotope Ratio Mass Spectrometry

441

ratios of individual compounds if more than one product is formed by the VOC reactions. Studies in the atmosphere will have to rely on compound-speciﬁc measurements, since it is very unlikely that it will be experimentally feasible to determine the isotope ratio of all possible reaction products. Therefore, it is necessary to understand the processes resulting in isotope fractionation between different products. There are several reasons why the stable isotope ratios of individual products may differ. Site-speciﬁc isotope fractionation, sitespeciﬁc loss of atoms during the formation of secondary products as well as non random distribution of the labelled sites in the reacting molecule can all result in the formation of products with different isotope ratios. A simple example, the atmospheric oxidation of propane by OH-radicals in the presence of NO and the resulting formation of acetone and propionaldehyde, will be used to demonstrate the complex dependence between reactant and product isotope ratios. Propane reacts with OH-radicals via abstraction of hydrogen atoms. This can occur at −) groups. The reaction −CH2 − −CH3 ) and the methylene (− two different sites, the methyl (− at the methylene site results in the formation of acetone. −CH2 − −CH3 + OH ⇒ CH3 − −CH• − −CH3 + H2 O CH3 −

(1a)

This is followed immediately by addition of an oxygen molecule and subsequent reaction of the peroxy radical with NO. −CH• − −CH3 + O2 ⇒ CH3 − −CHO•2 − −CH3 CH3 −

(1b)

−CHO•2 − −CH3 + NO ⇒ CH3 − −CHO• − −CH3 + NO2 CH3 −

(1c)

The abstraction of a hydrogen atom by molecular oxygen from the alkoxy radical results in the formation of a carbonyl compound. −CHO• − −CH3 + O2 ⇒ CH3 − −CO− −CH3 + HO2 CH3 −

(1d)

Following a similar reaction sequence the initial abstraction of a hydrogen atom results in the formation of propionaldehyde. −CH2 − −CH3 + OH ⇒ CH3 − −CH2 − −CH•2 + H2 O CH3 −

(2a)

−CH2 − −CH•2 + O2 ⇒ CH3 − −CH2 − −CH2 O•2 CH3 −

(2b)

−CH2 − −CH2 O•2 + NO ⇒ CH3 − −CH2 − −CH2 O• + NO2 CH3 −

(2c)

−CH2 − −CH2 O• + O2 ⇒ CH3 − −CH2 − −CHO + HO2 CH3 −

(2d)

For the sake of simplicity, minor branching reactions such as formation of peroxides or organic nitrates are neglected. Using this simpliﬁcation, the reaction sequences 1a–1d and 2a–2d can be condensed into a simple reaction describing the rate of product formation. −CH2 − −CH3 + OH ⇒ CH3 − −CO− −CH3 CH3 −

(k2 )

−CH2 − −CH3 + OH ⇒ CH3 − −CH2 − −CHO (k3 ) CH3 −

(2) (3)

For most VOCs with natural abundance isotope ratios there are several possible isotopologues, often with different isotopomers. Here, carbon isotope labelling will be used as an example to demonstrate the substantial increase in reactions that have to be considered.

442

Volatile Organic Compounds in the Atmosphere

There are two isotopologues, 12 C3 H8 and 13 C12 C2 H8 . The latter consists of two isotopomers, 13 CH − 12 −12 CH3 and 12 CH3 − −13 CH2 − −12 CH3 . The probability of having multiple 3 − CH2 − 13 13 −12 CH2 − −12 CH3 and 12 CH3 − −12 CH2 − −13 CH3 labelling with C can be neglected and CH3 − are identical. This results in a set of seven reactions that describe the formation of acetone and propionaldehyde and their isotopologues from propane. 12

−12 CH2 − −12 CH3 + OH ⇒ 12 CH3 − −12 CO− −12 CH3 CH3 −

12

−12 CH2 − −12 CH3 + OH ⇒ 12 CH3 − −12 CH2 − −12 CHO (k3a ) CH3 −

(3a)

13

−12 CH3 −12 CO− −12 CH2 − −12 CH3 + OH ⇒ 13 CH3 − CH3 −

(2b)

13

−12 CH2 − −12 CH3 + OH ⇒ 13 CH3 − −12 CH2 − −12 CHO (k3b ) CH3 −

(3b)

13

−12 CH2 − −12 CH3 + OH ⇒ 12 CH3 − −12 CH2 − −13 CHO (k3c ) CH3 −

(3c)

12

−13 CH2 − −12 CH3 + OH ⇒ 12 CH3 − −13 CO− −12 CH3 CH3 −

(2c)

12

−13 CH2 − −12 CH3 + OH ⇒ 12 CH3 − −13 CH2 − −12 CHO (k3d ) CH3 −

(k2a )

(k2b )

(k2c )

(2a)

(3d)

Acetone and propionaldehyde both react with OH-radicals and therefore their loss reactions also have to be considered. For simplicity, for these reactions only the sum of the products will be considered. Furthermore for small KIEs and as long as the labelling at different sites is close to a random distribution of the 13 C atom the kinetic isotope effects for the loss of acetone and propionaldehyde can be described by the ε-values for the reactions of acetone (OH εAcetone ) and propionaldehyde (OH εAldehyde ) with OH-radicals determined for natural abundance isotope ratios. The set of coupled differential equations describing the reaction system outlined by (2) and (3) can be solved numerically. Today’s desktop computing allows calculation of the concentrations of the different isotopomers with sufﬁcient accuracy to derive numerically precise isotope ratios from these calculations. Figure 10.13 shows the carbon isotope ratios of propane and its oxidation products as a function of the degree of propane reaction. The calculations were performed for propane with an initial δ 13 C value of 0‰ and an equal distribution of the 13 C atoms between the 3 C-atoms of the propane molecule. The rate constants for the reactions of propane, and the two isotopomers of its isotopologue were calculated from the group rate constants of Kwok and Atkinson (1995) and the best estimate for group-speciﬁc KIEs derived by Anderson et al. (2004b). The rate constants for the reactions of carbonyl compounds were taken from Atkinson (1994). The KIEs for the carbonyl reactions were estimated using the inverse carbon number dependence of carbon KIEs derived by Anderson et al. (2004b) for hydrogen abstraction from alkanes. The isotope ratios calculated from the reaction scheme shown above (solid line in Figure 10.13) agree nearly perfectly with the results calculations based on the average KIE of the propane reaction (squares). There is a very small difference when the reaction of propane has run nearly to completion. For 90% completion of the reaction, the difference is 0.003‰ and for 99% completion 0.02‰. These differences are well below the accuracy of any presently available measurement method and they are marginal compared to the change in the propane isotope ratio. As dictated by the mass budget, the isotope ratio of

Gas Chromatography–Isotope Ratio Mass Spectrometry

443

30 Propane Sum all products Acetone Propionaldehyde Propionaldehyde (steady state) Propane from KIE Acetone (no loss) Propionaldehyde (no loss)

13C (‰)

15

0

−15 0

0.2

0.4 0.6 Fraction of propane reacted

0.8

1

Figure 10.13 Carbon isotope ratios of propane and its oxidation products as function of the degree of propane consumption. For details of the calculations see text.

the sum of all products increases gradually with increasing degree of propane reaction. The initial value is exactly one ε lower than the initial δ 13 C of propane and for a complete reaction the isotope ratio of the sum of all products is identical to the initial isotope value of propane. In the absence of further reactions of acetone and propionaldehyde, the isotope ratios of the two compounds run parallel to the isotope ratios of all products (triangles and diamonds). In this case, the difference in isotope ratios is simply determined by the different site-speciﬁc KIEs. The changes in isotope ratio due to acetone and propionaldehyde loss reactions do not only depend on the KIEs of the loss and formation reactions, but also on the ratio of the rate constants for these reactions. In the example presented here (Figure 10.13) the loss reaction for acetone is slower than the reaction of the precursor propane. Consequently the KIE of the loss reaction for acetone has little impact on the isotope ratio of acetone if less than 50% of propane reacted. Only for very substantial propane processing the reaction of acetone has a visible impact on its isotope ratio. In contrast to this, the loss of propionaldehyde, which is faster than the oxidation of propane, has a strong inﬂuence on its isotope ratio. The isotope ratio of propionaldehyde changes rapidly from the δ 13 C of propionaldehyde formed by the reaction of propane to a value determined by the steady state between formation and loss. In steady state the isotope ratio of propionaldehyde follows the changes of the isotope ratio of the propane precursor. This difference of 6.3‰ is determined by the steady state between propane reactions resulting in propionaldehyde formation and the loss reaction for propionaldehyde. For this simple reaction scheme determining the steady state of propionaldehyde the dependence between the isotope ratio of the precursor VOC (propane) and the product isotope ratio can be described by the following relation. 13 C

[

C3 H6 O]

[

C3 H6 O]

12 C

13

=

KIELoss [ C C3 H8 ] · KIEForm [12 C C3 H8 ]

(10.41a)

444

Volatile Organic Compounds in the Atmosphere 13

12

Here, [ C C3 H6 O] and [ C C3 H6 O] are the concentrations of the 13 C labelled and unlabelled 13 12 propionaldehyde isotopologues. Similarly, [ C C3 H8 ] and [ C C3 H8 ] denote the concen13 trations of the C labelled and unlabelled propane isotopologues. KIEForm is the rate constant-weighted average of the KIEs for all reactions of propane resulting in the formation of propionaldehyde, and KIELoss is the KIE for the propionaldehyde loss reaction. The concentration ratios of the isotopologues can be expressed in the form of the δ-values of propionaldehyde (δC3 H6 O ) and propane (δC3 H8 ): δC3 H6 O =

KIE Loss · (δC3 H8 + 1) − 1 KIE Form

(10.41b)

It should be noted that Equation 10.41b does not include the number of carbon atoms of the precursor and product since it is identical for propionaldehyde and propane. In cases where product and precursor have different numbers of carbon atoms the carbon number dependence of the conversion of concentration ratios into isotope ratios (Equation 10.3) has to be considered. Although Equation 10.41b shows that the difference between δC3 H6 O and δC3 H8 will change depending on the isotope ratio, this dependence is small since usually (δC3 H8 ) 1. Replacing the KIE with (1 + ε), Equation 10.41b can be written as: 1 + εLoss · (δC3 H8 + 1) − 1 1 + εForm 1 = · (1 + δC3 H8 + εLoss + δC3 H8 · εLoss ) − 1 1 + εForm

δC3 H6 O =

(10.41c)

Approximating (1 + ε)−1 ≈ 1 − ε gives the following dependence: δC3 H6 O = δC3 H8 + εLoss − εForm + εLoss · δC3 H8 − δC3 H8 × εForm − εForm · εLoss − εForm · δC3 H8 · εLoss

(10.41d)

In many cases, including the carbon isotope effects for VOC reactions, the terms that are higher order in δ and ε can be neglected and a good ﬁrst-order steady-state estimate of the difference in δ-values between reactant and product can be derived from the difference in ε values between the loss and formation reactions. δC3 H6 O ≈ δC3 H8 + εLoss − εForm

(10.41e)

The hydrogen isotope ratios of propane and its oxidation products can be calculated following a procedure identical to that used for carbon isotope ratios. The results are substantially different from what might be expected intuitively (Figure 10.14). The rate constants for the reactions of the deuterium labelled propane isotopomers were calculated from the group rate constants for deuterium labelled and unlabelled alkanes determined by Droege and Tully (1986b). For the reaction of acetone with OH-radicals the hydrogen KIEs derived from isotope SRR considerations (Table 10.4) were used. Similar to the calculations made for carbon isotope ratios, the initial δD value for propane was set to 0‰ and a random distribution of the deuterium atom between the hydrogen atoms of propane was assumed. The difference between propane stable hydrogen isotope ratios derived from calculations using the reactions of the individual propane isotopomers and the results of calculations using the average KIE for the reaction of propane with OH-radicals is more pronounced than for stable carbon isotope ratios. The reason for the difference in hydrogen isotope

Gas Chromatography–Isotope Ratio Mass Spectrometry

445

400 Propane Propane from KIE Propane methylene group Propane methyl group Acetone Acetone (no loss) Sum all products

D (‰)

300

200

100

0

−100 0

0.2

0.4 0.6 Fraction of propane reacted

0.8

1

Figure 10.14 Hydrogen isotope ratios of propane and its oxidation products as function of the degree of propane consumption. For details of the calculations see text.

ratios calculated by the different procedures is the different rate of change of δD in the methyl and methylene groups (Figure 10.14). Due to the preferential reaction of propane at the methylene group and the differences in group-speciﬁc KIEs, the change of δD with increasing propane processing is far more pronounced for the methylene group than for the methyl groups. The consequence is that the effective KIE for the reaction of propane changes. The impact of this difference is highly dependent on the degree of propane processing. For a propane turnover of 20% the difference in δD is 0.15‰, which is approximately a factor of 100 below the change of the propane isotope ratio. For 50% processing the difference is 1.5‰, which is still small compared to the nearly 60‰ change in propane δD. For a 90% propane turnover the difference has increased to approximately 20‰, which amounts to nearly one tenths of the change in δD of propane. However, the most surprising observation is that the δD value of acetone formed is identical to the initial isotope ratio of propane. This can be explained by the loss of the methylene hydrogen atoms during the formation of acetone. Only oxidation of the propane isotopomer with the deuterium atom located at the methyl group (CH2 DCH2 CH3 ) results in the formation of deuterium labelled acetone. The reaction of the propane isotopomer labelled at the methylene group (CH3 CDHCH3 ) does not result in acetone containing a deuterium atom but in the formation of non-deuterated acetone and deuterated inorganic compounds. Therefore the formation of deuterium labelled acetone occurs only via a reaction that is not subject to a primary KIE. The formation of deuterated inorganic compounds explains that the isotope ratio of the total products is signiﬁcantly lower than that of acetone, the predominant organic product of atmospheric propane oxidation. There is only a small change in the isotope ratio of acetone as the reaction progresses. This change is primarily due to the isotope fractionation associated with the acetone removal. As shown by the existence of a minor change in the acetone isotope ratio in the absence of the acetone loss reaction (Figure 10.14), the change in the δD values of the propane methyl groups results in a change of the acetone isotope ratio. The hydrogen atoms of the propane

446

Volatile Organic Compounds in the Atmosphere

methyl group are conserved during the formation of acetone and since acetone contains no other hydrogen atoms the isotope ratio of the methyl groups directly determines the isotope ratio of acetone. For the example given in Figure 10.14 this has only a minor impact on the isotope ratio of acetone since most of the acetone formation happens before there are substantial changes in the δD values of the methyl group hydrogen. However, if for whatever reason propane emitted into the atmosphere has been subjected to site-speciﬁc enrichment or depletion of deuterium, the δD value of acetone formed by the atmospheric oxidation of propane will reﬂect the isotope ratio of the methyl groups and not of the propane molecule. It is important to realise that the potential for a large impact of site-speciﬁc isotope enrichment or depletion on the isotope ratio of a reaction product depends on the sitespeciﬁc loss of one or more atoms in the formation reaction. For carbon the formation of acetone from the oxidation of propane is not associated with the loss of a carbon atom. In this case, the impact of site-speciﬁc labelling is very small. For example, site-speciﬁc enrichment of the methyl carbon relative to the methylene carbon by 10‰ will change the carbon isotope ratio of the formed acetone by only 0.02‰. This extremely small effect is the result of the product of two small effects, the site-speciﬁc enrichment and the difference in group-speciﬁc KIEs. As demonstrated by the simpliﬁed examples given above, it is in principle possible to derive the stable isotope ratios of the products of atmospheric VOC oxidation from sitespeciﬁc rate constants and KIEs. However, due to the complexity of the reaction pathways and of the various processes determining the isotope ratios of products of atmospheric oxidation of VOCs with natural abundance isotope ratios, such predictions will have signiﬁcant uncertainties. Unfortunately there are extremely few laboratory studies of the isotope ratios of products of VOC oxidation using VOCs with natural abundance stable isotope ratios. Irei et al. (2006) studied the stable carbon isotope ratio of secondary POM formed in a ﬂow reactor generated in a ﬂow reactor by OH-radical induced oxidation of toluene. They found a substantial carbon isotope fractionation in the range of −6‰ between parent toluene and formed particulate matter. This fractionation was very close to values that were calculated using the KIE for the reaction of toluene with OH-radicals as the only source of isotope fractionation. Based on mass budget considerations it was concluded that the carbon isotope fractionation between particle and gas phase is only 0.6 ± 0.3‰. Particulate matter from gas-phase VOCs consists of a signiﬁcant number of compounds, which are formed via several reaction channels, each of which consists of several steps, often including branching reactions. It is, therefore, somewhat surprising that there is only little isotope fractionation between gas and particle phase. There are two possible explanations for this. One possibility is that the isotope ratio of each of the compounds forming particulate matter is determined by the isotope ratio of toluene and the step initiating the reaction chains, the reaction with OH-radicals. The other possibility is that, either due to mere coincidence or due to some unknown principle in the formation mechanism, the average fractionation is identical to the KIE for reaction of toluene with OH-radicals. This question could be answered unambiguously by compound-speciﬁc isotope ratio analysis of the particulate matter. However, the results of such measurements are presently not available. The ﬁnding of a deﬁned dependence between the isotope ratio of the VOC precursor and the formed secondary particulate matter opens the possibility to use isotope ratio

Gas Chromatography–Isotope Ratio Mass Spectrometry

447

measurements to test the applicability of the results of laboratory studies to the atmosphere. Combining measurements of the isotope ratios of precursors and secondary particulate matter will allow comparing predictions based on laboratory studies with ambient observations. Obviously, this will require compound-speciﬁc measurements since atmospheric POM is complex in its composition and derived from a variety of sources.

10.4.8

Atmospheric removal of VOC by deposition

For most VOCs, deposition is not an important loss process. However there are some exceptions such as formaldehyde, peroxyacetyl nitrate and several other oxygenated VOCs. Recently it has been suggested that deposition is an important sink for halomethanes (Keppler et al. 2005). However, there is no direct information on the dependence of deposition velocities on isotopic labelling of VOCs. The process of trace gas deposition on surfaces consists of several steps, ranging from turbulent transport in the atmosphere, to molecular diffusion in the surface layer, and ultimately some chemical conversion process. Turbulent diffusion does not result in isotope fractionation and isotope fractionation due to molecular diffusion will generally be small due to the relatively small difference in molecular mass due to labelling of VOCs. However, it has been shown that loss processes due to degradation by microorganisms can have very substantial isotope effects (Bill et al. 2002a; Gray et al. 2002; Mancini et al. 2003; Miller et al. 2001, 2004; Slater et al. 2001; Ward et al. 2000). In some studies, it has been assumed that the isotope fractionation observed in laboratory studies for the degradation mechanisms is identical to the isotope fractionation of VOC removal by deposition (Harper et al. 2001, 2003; Keppler et al. 2005). However, considering the complexity of deposition processes and the lack of studies of the isotope fractionation during VOC deposition such a simpliﬁcation seems premature. Especially in the case of VOCs with soil sources (see also Section 10.5.1) the overall impact of loss, production and transfer processes is very difﬁcult to predict in the absence of experimental studies of the isotope fractionation of the actual deposition process.

10.5

Stable isotope ratios of atmospheric VOC and their sources

It has already been mentioned that our present knowledge of the stable isotope ratios of atmospheric VOCs and their sources is very limited. The presently available data are limited to measurements of stable carbon isotope ratios. Effectively no studies of hydrogen isotope ratios of atmospheric VOCs have been published, although the hydrogen KIE measurements suggest that for many VOCs very substantial variations in isotope ratios can be expected. The following part of this section therefore will concentrate on VOC carbon isotope ratios. The boundary between source studies and ambient measurements is not always clearly deﬁned. It is a well-established procedure to use ambient measurements made under conditions which are dominated by local emissions from speciﬁc sources or groups of sources to determine the characteristic patterns of these sources. Since isotope ratios do not directly depend on concentrations, this method is very useful for studies of isotope ratios of emissions. Indeed, many studies of isotope ratios of VOC emissions are based on this method.

448

Volatile Organic Compounds in the Atmosphere

The main advantage is that atmospheric mixing helps to average out variations between individual sources. There are also a few studies of the isotope ratio of VOC emissions under controlled laboratory conditions. There are several possibilities to combine the available isotope ratio studies of ambient VOC into separate groups. Here, the combination of VOCs into one group will be based on the different concepts which can be used to interpret VOC isotope ratios. For halogenated VOCs such as chloromethane, bromomethane and chloroﬂuorocarbons the atmospheric life time is about 1 year or longer. Consequently, the atmospheric variability of the concentrations and isotope ratios of these compounds is small and it is possible to determine global or hemispheric average values for the concentrations and isotope ratios. These averages can be used to derive constraints on the trace gas budgets. Non-methane hydrocarbons (NMHC) are substantially more reactive than most halocarbons and there atmospheric concentrations and isotope ratios are determined by emissions and atmospheric processes on local and regional scales. The third group of VOCs consists of compounds which can be formed by chemical reactions in the atmosphere. The most important example is formaldehyde. For this group of VOCs, the interpretation of measured isotope ratios is complex (see Section 10.4.7) and the available information on the isotope ratios of secondary VOCs is too limited to allow quantitative interpretation of atmospheric observations.

10.5.1

Halomethanes and chloroﬂuorocarbons

The result of the presently available measurements of carbon isotope ratios of atmospheric halomethanes and chloroﬂuorocarbons (CFC) are summarised in Table 10.8. From the long atmospheric residence times of the halocarbons, it can be expected that the variability of Table 10.8 Stable carbon isotope ratio measurements of atmospheric halocarbons VOC Chloromethaneb Chloromethanec Chloromethaned Chloromethanee Bromomethanef Bromomethanec CFC-12c CFC-11c CFC113b CFC113c

Data points

Rangea (δ 13 C, ‰)

Type of location

78 25 17 2 7 5 25 22 38 12

−36.2 (2.6, 0.3) −39.1 (2.5, 0.5) −44 to −30 −43.5, −37.4 −43.1 (1.5, 0.9) −33.6 (4.4, 2.2) −37.0 (6.4, 1.3) −29.4 (4.1, 0.9) −23.3 (9.6, 2.6) −28.1 (4.3, 1.3)

Marine and continental background Belfast, UK Marine, western North Paciﬁc Coastal, New Zealand Urban, Berkeley, California (USA) Urban, Belfast, UK Urban, Belfast, UK Urban, Belfast, UK Marine and continental background Urban, Belfast, UK

a The number in brackets give the standard deviation and the standard error of the mean. If no statistical

evaluation is available the range of observations is given. If less than three data points are available the individual values are listed. b From Thompson et al. (2002). c From Archbold et al. (2005). d From Tsunogai et al. (1999). e From Rudolph et al. (1997). f From Bill et al. (2004).

Gas Chromatography–Isotope Ratio Mass Spectrometry

449

the carbon isotope ratio is only small. For chloromethane, most of the observations give values in the range of −35‰ to −40‰. Nevertheless, there are some indications that stable carbon isotope ratios in urban areas are slightly higher than in background air. The δ 13 C values from measurements of background air (Thompson et al. 2002) differ by 2.9‰ from the results of measurements in Belfast (Archbold et al. 2005). The standard error of this difference is only 0.6‰ and the 99% conﬁdence intervals for these measurement series are 0.8‰ and 1.3‰, respectively. Similarly, Thompson et al. (2002) and Tsunogai et al. (1999) reported −34.3‰ to −46.9‰ with an average of −39.2‰ for measurements in urban areas. This is slightly lower than the average of the background isotope ratios reported by Thompson et al. (2002), but the number of measurements is only small and the spatial coverage too limited to allow generalisation of this ﬁnding. In any case, the results indicate that the difference is only small, in the range of a few per mille or less. For CFC-113 and bromomethane, the two published sets of measurement differ by 4.8‰ and 10.5‰, respectively (Table 10.8). However, the 99% conﬁdence intervals for the two CFC-113 measurement series are 4.1‰ and 3.2‰, respectively. It can therefore not be ruled out that the differences between the two CFC-113 data sets are due to the considerable spread of the measured data. However, for bromomethane, the 99% conﬁdence intervals are 1.5‰ and 5.1‰. This strongly points towards a substantial systematic difference between the two measurement series. Considering the long atmospheric residence time of bromomethane and the uniform distribution of the bromomethane mixing ratios, a systematic seasonal or spatial variability seem unlikely as explanation for the observed difference. More plausible is the possibility of local or regional emissions of bromomethane, but also systematic measurement bias cannot be excluded. Table 10.9 presents an overview of the stable carbon isotope ratios of halomethane sources. For CFC, the available information on isotope ratios of sources is limited to the analysis of samples of laboratory grade substances or mixtures (Archbold et al. 2005; Thompson et al. 2002). The atmospheric concentration of CFC is the result of venting CFC produced by various suppliers over a period of several decades into the atmosphere. The relation between the isotope ratios of these few samples and the average of atmospheric emissions is therefore highly uncertain. This problem is evident from the two available measurements of CFC-113 samples. Thompson et al. (2002) found a value of −31.3‰ for a commercial CFC-113 sample, whereas Archbold et al. (2005) report a value of −27.5‰ for CFC-113 for an artiﬁcial CFC mixture. It should also be noted that in some cases the given isotope ratios do not necessarily represent the isotope ratio of the emissions into the atmosphere. Miller et al. (2004) report that bromomethane and chloromethane degradation in soils results in isotope fractionation. The KIEs for bacterial degradation was measured to be 69 ± 9‰ and 49 ± 3‰, respectively. For chemical loss of bromomethane, ε values of 59 ± 7 were reported. Bill et al. (2002a) report a mean fractionation of 12‰ for the uptake of bromomethane by soils. Uptake by soils is not only a potentially very important loss process for atmospheric bromomethane, but will also occur during agricultural soil fumigation. Bromomethane emitted into the atmosphere as a consequence of soil fumigation will therefore not have the isotope ratio of industrially produced bromomethane, but also reﬂect the impact of isotope fractionation due to soil uptake. Similarly, for halomethanes produced in soils by bacteria the isotope ratio of the emissions will depend on the isotope ratio of the produced halomethanes as well as the isotope fractionation occurring during any subsequent loss in the soil. The consequence is

450

Volatile Organic Compounds in the Atmosphere

Table 10.9 Stable carbon isotope ratios of sources of atmospheric halomethanes δ 13 C Chloromethanea , ‰

Type of source Biomass burning Salt marsh emissions Production by higher plants Production by fungi Industrially produced Production from dried biomass at 225◦ C Production from dried biomass at 40−60◦ C VOC in ocean water

−51.7 (5.7)b ; −45.1c ; −56.6 (0.9)d (−44 to −62)e −62.6 (5.1)f ; −63.4 (1.0)g −42.6 (0.4)g ; (−38 to − 26)h −46.9g ; −61.9g ; −41.9c ; −30.8i −89.7 (3.4)k

δ 13 C Bromomethanea , ‰ −48.6 (2.7)d −43 (−14 to −62)e

−39.6i ; −54.4 (−49.5 to −62.3)j

−128.6 (4.2)k −38 (4)l

a Numbers in parenthesis give the error of mean; if the available information is insufﬁcient for a statistical

evaluation the range of data is given; if less than three data points are available the individual values are listed.

b From Czapiewski et al. (2002). c From Rudolph et al. (1997). d From Komatsu et al. (2005). e From Bill et al. (2002b). f From Harper et al. (2003). g From Harper et al. (2001).

h From Kalin et al. (2001). i Isotope ratio measured in artiﬁcial mixture, from Archbold et al. (2005). j From McCauley et al. (2001). k From Keppler et al. (2004). l From Komatsu et al. (2004).

that in these instances an estimate of the isotope ratios of the emissions requires knowledge of the isotope ratio of the production and the KIE of the loss process as well as that of the ratio of formation and emission rates. However, this information is not available. Similar difﬁculty arises with respect to the isotope ratio of oceanic emissions. For most halocarbons, including chloromethane and bromomethane, the oceanic concentration is close to equilibrium with the atmosphere. In this case, the net emission rate is given by the difference of two ﬂuxes, namely the uptake by oceans and the ﬂux from the oceans. Although for any given compound this can formally be described by a net ﬂux, it is not correct to assign the isotope ratio of ocean water to the net ﬂux from the ocean. To determine the impact of a bidirectional exchange process on the isotope budget, the two ﬂuxes have to be considered separately. The observation that the stable carbon isotope ratios of chloromethane in oceans and in the atmosphere are similar (Tables 10.8 and 10.9) agrees with the concept that the ocean and atmosphere are close to equilibrium. The consequence is that the net effect of the oceans on the isotope ratio of atmospheric chloromethane will depend on the difference between the two isotope ratios. Although this difference is small, it is presently not very well constrained due to the limited number of available measurements. Our present knowledge of the halomethane sources and sinks does not allow setting up a balanced budget. The most important atmospheric loss process identiﬁed for chloromethane and bromomethane is reaction with OH-radicals. For chloromethane, there

Gas Chromatography–Isotope Ratio Mass Spectrometry

451

is one measurement of the KIE for this reaction (Table 10.6), which indicates that this reaction results in enrichment of 13 C in atmospheric chloromethane by on average 59‰ relative to the emissions. The KIE for microbial degradation in soils is approximately 10‰ lower (Miller et al. 2004; see above), but within the uncertainty of the presently available data this difference is not signiﬁcant. There are some minor loss processes such as transport into the stratosphere which may contribute to the isotope budget, but are unlikely to have a major impact. Similarly, exchange with oceans is unlikely to have a major impact; the limited available data suggest that the isotope ratios of chloromethane in the atmosphere and in ocean water are very similar (Tables 10.8 and 10.9). Based on this, an average isotope ratio in the range of −90‰ to −95‰ for all chloromethane sources can be derived. This value is signiﬁcantly lower than the isotope ratios of the known sources of atmospheric chloromethane (Table 10.9). Presently, the only known potential process, which may result in emission of chloromethane with an isotope ratio lower than −90‰, is the production from dry biomass (Table 10.9). It has been suggested that this points towards a substantial role of the degradation of biomass as a source of atmospheric chloromethane (Keppler et al. 2005). However, it should be remembered that the carbon isotope budget of chloromethane has very substantial uncertainties and that there presently is no direct evidence for signiﬁcant emissions of chloromethane from decaying plant matter. For bromomethane, the KIE for the atmospheric removal by OH-radicals is not known and, therefore, it is not possible to deduce an average isotope ratio for bromomethane emissions from its atmospheric carbon isotope ratio.

10.5.2

Non-methane hydrocarbons

The results of ambient measurements of the stable carbon isotope ratios of atmospheric NMHC are compiled in Tables 10.10 and 10.11 and gives an overview over the carbon isotope ratios of NMHC sources. Although the number of available data is still limited, the results of these studies present a reasonably consistent picture. Generally isotope ratios of ambient measurements are either similar to the isotope ratios of NMHC sources or enriched in 13 C. This can be explained by the KIEs of the atmospheric NMHC loss processes (see Section 10.2.3.3). Rudolph et al. (2000) derived estimates of photochemical ages from NMHC isotope ratio measurements in a suburban area and Saito et al. (2002) used isotope ratios measurements for light alkanes combined with back trajectories to identify the regions of origin for these NMHC. A slightly different possibility to use isotope ratio measurements was suggested by Rudolph et al. (2003). They demonstrated that the combination of concentration and carbon isotope ratio measurements of isoprene can be used to differentiate between the existence of small local sources and advection of air containing isoprene from sources upwind of the observation site. Although most of the available measurements of NMHC isotope ratios have not been subject to a detailed quantitative analysis, most of the observed variations can be explained by fractionation due to atmospheric reactions. Rudolph et al. (2002) showed that NMHC in urban areas have stable carbon isotope ratios close to those of emissions. This is consistent with the concept that substantial changes in atmospheric NMHC isotope ratios are predominantly due to isotope fractionation associated with loss reactions. This can be used to obtain ﬁrst-order estimates of isotope ratios for urban VOC emissions. For alkanes

452

Volatile Organic Compounds in the Atmosphere

Table 10.10 Stable carbon isotope ratio measurements of atmospheric non-methane hydrocarbons VOC

Ethane

Ethyne

−C5 Alkanes C3 −

>C5 Alkanes

−C3 Alkenes C2 −

Number of data points

Rangea (δ 13 C, ‰)

3 2 6 23 8 41 14 32

−26.3 (3.9, 2.8)b −26.5c ; −27.7c −22 to −25.5c −19 to −26d −21.9 (3.5, 1.3)e −24 (3.4, 0.5)e −25.2 (2.5, 0.7)e −27.8 (2.1, 0.4)e

9 5 10

−26.9 (1.4, 0.5)e −27.14 (1.7, 0.8)e −25.5 to −28c

2 9 41 14 23

−6.5c ; −3.5c −10.7 (9.1, 3.2)e −5 (6.2, 1.2)e −1.4 (6.1, 1.7)e −13.6 (4.3, 0.9)e

10 11 5

−3.5 to −19.5c −8.3 (8.0, 2.5)e −12.3 (1.8, 0.9)e

4 10 18 45 9 42 33

−21.1 (9.7, 5.6)b −24.5 to −28.5c −25 to −30c −16 to−37d −26.9 (3.3, 1.2)e −27.7 (2.6, 0.4)e −26.1 (2.4, 0.4)e

40 45 29 4

−24 to −31c −27.5 (1.6, 0.3)f −27.6 (5.3, 1.0)e −27.2 (0.9, 0.6)e

21

−23.3 (4.4, 1.0)e

36 16

−26.8 (5.0, 0.8)e −25.5 (6.5, 1.6)e

4 55

−15 to −20.5c −14.2 (10.5,1.4)e

12

−12.5 to − 25.5c

Type of location

Coastal, New Zealand Coastal, Japan Marine, western North Paciﬁc Marine, western North Paciﬁc Baring Head, New Zealand Alert, Canadian Arctic Continental background, Canada Airplane samples, continental boundary layer, Houston, Texas (USA) Urban, Canada Urban, New Zealand Urban, Japan Coastal, Japan Baring Head, New Zealand Alert, Canadian Arctic Continental background, Canada Airplane samples, continental boundary layer, Houston, Texas (USA) Urban, Japan Urban, Canada Urban, New Zealand Coastal, New Zealand Coastal, Japan Marine, western North Paciﬁc Marine, western North Paciﬁc Baring Head, New Zealand Alert, Canadian Arctic Airplane samples, Propane only, continental boundary layer, Houston, Texas (USA) Urban, Japan Urban, Toronto, Canada Urban, Canada Urban, New Zealand Airplane samples, continental boundary layer, Houston, Texas (USA) Urban, Canada Urban, New Zealand Coastal, Japan Airplane samples, continental boundary layer, Houston, Texas (USA) Urban, Japan

Gas Chromatography–Isotope Ratio Mass Spectrometry

453

Table 10.10 (Continued) Number of data points

Rangea (δ 13 C, ‰)

16 21 8

−21.6 (4.0, 1.0)f −22.4 (6.0, 1.3)e −22.0 (1.0, 0.4)e

Urban, Toronto, Canada Urban, Canada Urban, New Zealand

8 46 14 31

−27.3 (1.9, 0.7)e −25.1 (3.4, 0.5)e −23 (3.1, 0.9)e −23.9 (2.8, 0.5)e

16 13 6

−24.6 (2.3, 0.6)f −25.2 (2.7, 0.8)e −24.0 (0.9, 0.4)e

Baring Head, New Zealand Alert, Canadian Arctic Continental background, Canada Airplane samples, continental boundary layer, Houston, Texas (USA) Urban, Toronto, Canada Urban, Canada Urban, New Zealand

Toluene

27 13 6

−25.9 (1.7, 0.3)g −25.2 (4.3, 1.3)e −24.4 (1.3, 0.6)e

Urban, Toronto, Canada Urban, Canada Urban, New Zealand

>C7 Alkylbenzenes

97 15 22

−25.7 (2.6, 0.3)g −25.7 (6.3, 1.7)e −24.0 (2.4, 0.5)e

Urban, Toronto, Canada Urban, Canada Urban, New Zealand

Isoprene

5 3 5 19

−28.8 (1.5, 0.8)e −32.4 (3.7, 2.6)e −20.0 (4.0, 2.0)e −23.3 (4.2, 1.0)e

Continental background, Canadah Continental background, Canadai Lower Fraser Valley, Canada (summer) Airplane samples, continental boundary layer, Houston, Texas (USA)

VOC

Benzene

Type of location

a The number in brackets give the standard deviation and the standard error of the mean. If no statistical evaluation is

available the range of observations is given. If less than three data points are available the individual values are listed.

b From Rudolph et al. (1997). c From Tsunogai et al. (1999). d From Saito et al. (2002). e From Thompson (2003). f From Rudolph et al. (2000). g From Rudolph et al. (2002).

h May to August, average isoprene mixing ratio 490 ppt. i September and October, isoprene mixing ratio 50 ppt.

the isotope ratios observed in urban areas fall into a narrow range (Table 10.10), which is identical to the range of isotope ratios found in studies of urban alkane sources. The number of studies is still too limited to allow any general conclusions. The available data include measurements from the United States, Canada, New Zealand and Japan, but there is effectively no information for Europe, continental Asia, Africa and South America. Nevertheless, the isotope ratio found in studies of urban alkane sources is fully compatible with the average carbon isotope ratio of crude oil (Yeh and Epstein 1981) and gasoline (Smallwood et al. 2002).

454

Volatile Organic Compounds in the Atmosphere

Table 10.11 Stable carbon isotope ratios of sources of atmospheric non-methane hydrocarbons VOC or group of VOC

δ 13 Ca , ‰

Type of source

Ethane

Biomass burning Street tunnel

−28.7b ; −27.9c ; −27.6c −31.3b ; −25.8 (1.3)d ; −29.6 (1.0)e

Ethyne

Biomass burning Street tunnel

−13.0b ; 3.7 (1.2)f −16.8 (2.1)d ; −10.2 (1.2)e

−C5 Alkanes C3 −

Biomass burning Street tunnel Underground garage Gas station Reﬁnery

−27.6 (0.9)b ; −27.4 (0.4)c ; −26.3 (0.5)f −25.9 (1.1)b ; −25.8 (0.4)d ; −28.1 (0.5)g −30.1 (0.3)g −28.1 (0.5)g −28.9 (0.5)g

>C5 Alkanes

Street tunnel Underground garage Gas station Reﬁnery

−27.8 (1.5)d ; −26.7 (0.7)e ; −27.3 (0.1)g −27.6 (0.3)g −27.1 (0.2)g −27.0 (0.1)g

−C3 Alkenes C2 −

Biomass burning Street tunnel Underground garage

−24.2, −25.2b ; −22.5 (0.7)f −25.9, −23.1b ; −21.3 (0.5)d ; −24.0 (0.9)g −25.0 (0.5)g

>C3 Alkenes

Biomass burning Street tunnel Underground garage Gas station

−20.1 (8.4)b ; −24.3 (0.6)f −22.2b ; −24.7 (0.5)g −23.8 (0.7)g −27.1 (0.9)g

Benzene

Biomass burning Street tunnel Underground garage Gas station Reﬁnery

−26.0 (1.0)f −24.5 (0.6)d ; −26.2 (0.8)e ; −26.5 (0.4)g −27.6 (0.3)g −28.8 (0.3)g −28.6 (0.1)g

Toluene

Biomass burning Street tunnel Underground garage Gas station Reﬁnery

−26.5 (0.9)f −26.1 (0.6)d ; −27.2 (0.3)e ; −27.5 (0.4)g −27.1 (0.4)g −27.1 (0.8)g −28.4 (1.7)g

>C7 Alkylbenzenes

Biomass burning Street tunnel Underground garage Gas station

−25.7 (0.5)f −25.3 (0.5)d ; −25.5 (0.7)e ; −27.1 (0.6)g −27.6 (0.1)g −26.9 (0.3)g

Styrene

Biomass burning

−23.3 (1.0)f

Isoprene

Vegetation (velvet bean)

−27.7 (0.4)h

a Numbers in parenthesis give the error of mean; if the available information is insufﬁcient for a statistical evaluation the range of data is given; if less than three data points are available the individual values are listed. b From Rudolph et al. (1997). c From Komatsu et al. (2005). d Samples collected in summer, downtown Toronto (Canada), from Thompson (2003). e Samples collected in winter, downtown Toronto (Canada), from Thompson (2003). f From Czapiewski et al. (2002). g Measurements in the area of Toronto (Canada), from Rudolph et al. (2002). h From Rudolph et al. (2003).

Gas Chromatography–Isotope Ratio Mass Spectrometry

455

For aromatic compounds and alkenes in urban air 13 C is, compared to alkanes, generally slightly enriched. This most likely reﬂects the higher KIEs for their atmospheric reactions (Table 10.7). There is also some indication that alkenes in emissions from incomplete combustion sources are slightly heavier than alkanes or aromatic compounds. This difference is only small, but it can be found for transportation related emissions as well as for emissions from biomass burning. Overall, the presently available data suggest that the carbon isotope ratios of emitted NMHC are generally very close to that of the burnt fuel. There is some evidence for the existence of a systematic carbon isotope fractionation between fuel and emitted NMHC for incomplete combustion processes but these effects are small and in most cases the variations are comparable in magnitude to the uncertainty of the measurements (Czapiewski et al. 2002; Rudolph et al. 2003). Similar to the ﬁndings of little carbon isotope fractionation effects for combustion processes, evaporation of VOCs only results in a small enrichment of 13 C in the gas phase (Harrington et al. 1999; Irei et al. 2006; Smallwood et al. 2002). Typically, the VOCs in the gas phase are a fraction of a per mille heavier than in the liquid phase. Most of the reported fractionation effects are in the range of 0.2–0.5‰. However, the actual impact of this effect on the isotope ratio of atmospheric VOC depends on the type of evaporation process. For partial evaporation such as evaporative losses from storage containers or during fuelling evaporation will cause a small isotope fractionation. However, or in the case of complete evaporation, for example from gasoline spills, there is overall no isotope fractionation. There are an extremely large number of studies of isotope ratios of total crude oil as well as of individual NMHC in crude oil or natural gas. A detailed review of this extremely extensive body of information is beyond the scope of this chapter. In general, crude oil has stable carbon isotope ratios in the range of −2333‰ with most of the values between −25‰ and −31‰ (see, for example, Yeh and Epstein 1981). An effectively identical range of carbon isotope ratios was found by Smallwood et al. (2002) for a substantial number of individual compounds in 19 gasoline samples from different areas of the United States. Harrington et al. (1999) determined the stable carbon isotope ratio of 44 samples of commercially available light alkyl benzenes. The results ranged from −23.9‰ to −29.4‰ with an average of −27.1‰. The 10‰ and 90‰ were −28.9‰ and −25.4‰, respectively. Overall the observed isotope ratios of most anthropogenic NMHC sources fall into a tight range, which is consistent with the range of isotope ratios of the parent materials. Somewhat different isotope ratios are expected for emissions of light alkanes due to leakage of natural gas. Although for ethane and heavier alkanes the depletion of 13 C relative to crude oil is far less pronounced than for methane, ethane in natural gas is, depending on the origin of the natural gas, generally a few per mille lighter than the average of crude oil. Similar to methane, the isotope ratio of NMHC in natural gas depends on the origin of the natural gas; however, the variability is by far less pronounced for ethane and very small for propane and heavier alkanes. The available stable carbon isotope ratio measurements in urban areas (Table 10.10) show no evidence for a systematic difference between ethane and other alkanes. A possible explanation is that the presently available studies are conducted for areas, where the inﬂuence of natural gas loss is small compared to other fossil fuel derived emissions such as automotive exhaust. Furthermore, the background concentration of ethane is substantial and enriched in 13 C relative to ethane over urban areas. Thus the contribution of heavier ethane from background air may compensate for the inﬂuence of urban ethane emissions depleted in 13 C.

456

Volatile Organic Compounds in the Atmosphere

A large variability is found for the isotope ratios of ethyne. The presently available data show a very large variability for emissions as well as for ambient ethyne. Moreover all observations indicate that 13 C in ethyne is substantially enriched relative to fossil fuel. The reasons for this unusual behaviour of ethyne are not known. The isotope ratios of isoprene emitted by vegetation have been studied by several groups. The ﬁrst study was conducted by Sharkey et al. (1991) who report that isoprene emitted from oak leaves was depleted by 2.8 ± 1.1‰ relative to photosynthetically ﬁxed carbon. Rudolph et al. (2003) studied the carbon isotope ratio and its temperature and light dependence for isoprene emitted from velvet bean (Mucana pruriens) at ambient carbon dioxide concentrations and isotope ratios. The emitted isoprene was 2.6 ± 0.9‰ lighter than the leaf carbon. The observed light and temperature dependence of the isotope ratio was small. Funk et al. (2004) found that isoprene emitted from Populus deltoids was between 2.5‰ and 3.2‰ lower in 13 C than photosynthetically ﬁxed carbon in the absence of heat or water stress. For plants under severe heat- or water-stress the fractionation effect increased to −8.5‰ and −9.3‰. Affek and Yakir (2003) studied the carbon isotope fractionation of isoprene emitted from myrtle (Myrtus communis), buckthorn (Rhamnus alaterus) and velvet bean (Mucana pruriens). They report that isoprene was depleted by 4–11‰ relative to photosynthetically ﬁxed carbon. This is signiﬁcantly larger than the 13 C depletion found by other groups in the absence of stress. Based on these results the carbon isotope ratios of isoprene emitted from vegetation in the absence of stress can be expected to range from −27‰ to −30‰. Under conditions of severe stress the isotope ratio may decrease to approximately −35‰. The measured carbon isotope ratios of isoprene in the atmosphere (Table 10.11) are often considerably heavier than the emissions, which is consistent with the carbon KIE for the reaction of isoprene with the OH-radical or ozone (Table 10.7) and the short atmospheric residence time of isoprene. In most cases, the carbon in atmospheric NMHC becomes heavier with increasing distance of the observation site from major NMHC sources. However, there are some exceptions. Several of the NMHC carbon isotope ratios measured at background sites such as Baring Head or Alert are very close to those of emissions. This can be explained by the combination of two factors. For remote locations the contribution to NMHC concentrations from long-range transport is very small for reactive NMHC. Therefore local or regional emissions can play a dominating role, even if the emission rates are only small. Furthermore, the currently applied measurements techniques for NMHC isotope ratios are not sufﬁcient to reliably determine isotope ratios for the lower end of the range of NMHC concentrations at remote concentrations. The consequence is that for NMHC with very low concentrations in background air isotope ratio measurements at remote locations will be biased towards samples inﬂuenced by local or regional sources.

10.5.3

Oxygenated VOC

Table 10.12 summarises measurements of the stable carbon isotope ratios of oxygenated VOCs in the atmosphere. Most of the studies targeted identiﬁcation of the sources, sometimes using measurements of 14 C as indicator for a biogenic origin of the VOCs. Glasius et al. (2000) concluded from their 14 C measurements that formic acid found in the atmosphere

Gas Chromatography–Isotope Ratio Mass Spectrometry

457

Table 10.12 Stable carbon isotope ratio measurements of atmospheric oxygenated VOC Number of data points

Rangea δ 13 C, ‰

Formaldehydeb

2

−26.69; −28.51

Formaldehydeb Formaldehydec Formaldehydec Formaldehyded Acetaldehydeb

2 6 3 7 2

−16.68; −18.99 −17 −28.3 −22.5 (0.8) −29.2; −29.28

Acetaldehydeb Formic acide Formic acidf

2 4 52

−21; −21 −27.6 (0.4) −18‰ to −25‰

VOC

Formic acidf Acetic Acidg

5 5

−30.1 (0.6) −20.5 (0.7)

Location of study

Industrial, near petrochemical factory Ghuangzhou China Bus station in Ghuangzhou China Mountain, continental US Clean air site, coastal NZ Coastal, Nova Scotia, Canada Industrial, near petrochemical factory Ghuangzhou China Bus station in Ghuangzhou China Rural, Denmark Several sites continental and coastal United States Rainwater Los Angeles, California (USA) Rainwater Los Angeles, California (USA)

a The number in brackets give the standard deviation and the standard error of the mean. If no statistical evaluation is

available the range of observations is given. If less than three data points are available the individual values are listed.

b From Wen et al. (2005). c From Johnson and Dawson (1990). d From Tanner et al. (1996). e From Glasius et al. (2000). f From Johnson and Dawson (1993).

g From Sagukawa and Kaplan (1995).

in a rural area of Denmark is predominantly derived from biological material. This is consistent with the observation of a narrow range of δ 13 C values with an average value of −27.6‰. This also suggests that for these conditions isotope fractionation due to the formation processes or reactions in the atmosphere is marginal. A nearly identical type of study was conducted by Johnson and Dawson (1993) for the continental and coastal United States. Similar to the ﬁndings of Glasius et al. (2000) the measured 14 C content strongly suggested a predominantly biogenic origin of gas-phase formic acid. However, the corresponding δ 13 C values range from −18‰ to −25‰, indicating the existence of a process resulting in 13 C enrichment. Qualitatively this is compatible with isotope fractionation associated with atmospheric loss processes. However, presently neither the KIEs for atmospheric loss processes nor the carbon isotope ratio of formic acid formed by biological processes are known. The available stable carbon isotope ratio measurements for formaldehyde and acetaldehyde show a considerable variability. The δ 13 C values range from −29‰ to −17‰. Sources of formaldehyde are the atmospheric oxidation of VOC and methane as well as emissions from industrial sources and incomplete combustion. It has been argued that 13 C in atmospheric formaldehyde is enriched relative to the emissions due to isotope fractionation associated with formaldehyde photolysis (Johnson and Dawson 1990). Qualitatively this can indeed explain the observations. Photolysis is one of the most important atmospheric removal formaldehyde processes and its KIE is substantial (Section 10.4.4). Unfortunately,

458

Volatile Organic Compounds in the Atmosphere

the necessary information for quantitative considerations is not available. The KIE for another very important atmospheric loss process, the reaction with OH-radicals, is only poorly known (Table 10.6), and the isotope ratios of formaldehyde emissions and formation processes have not been measured.

10.6

Conclusions

There is a substantial set of data available for sources, sinks, atmospheric reactions and measured ambient isotope ratios of atmospheric VOCs. It is, therefore, somewhat unexpected that the quantitative interpretation of atmospheric measurements VOC isotope ratios is still seriously limited by the lack of data for comparison. For those who remember the early days of studies of VOC concentrations, this will sound familiar. It took about 20–30 years to obtain a reasonable level of basis for the interpretation of atmospheric VOC concentrations in general and, as can be seen from other chapters in this book, there are still very signiﬁcant gaps. One of the problems is the very large number of different chemical substances summarised under the heading of VOCs. The consequence is that the available information on VOCs is spread over a large number of different compounds, with often only a limited amount of knowledge for individual compounds. It is easy to predict that once the necessary basic information is available, isotope ratio measurements will belong to the established tools for studying atmospheric VOC chemistry. However, the number of research groups studying VOC stable isotope ratios is presently small, although there has been considerable increase during the past few years. Consequently, we have to accept that developing a broad, general basis for studies of atmospheric VOC isotope ratio measurements will take some time. The present lack of quantitative information makes it very speculative to predict the accuracy, and therefore also the usefulness of constraints which can be derived from VOC isotope ratio studies. Nevertheless, it is possible to identify several areas where it is extremely likely that isotope ratio studies will be particularly valuable in the near future or where their usefulness has already been demonstrated. The possibility to use measurements of NMHC isotope ratios for studies of atmospheric processing and identiﬁcation of the origin of NMHC has been shown in a small number of studies. The main impediment for a wider application is the substantial experimental effort required for such measurements. However, the continuing progress in GC–IRMS instrumentation and the decreasing costs of such instrumentation make it likely that such measurements will become a routine tool in the near future. Due to their important role for stratospheric ozone halogenated VOCs belong to the most important organic compounds in the atmosphere. Therefore, additional tools to constrain the budget, such as isotope ratio studies, can be extremely valuable. Indeed, there already is substantial progress towards establishing the isotope ratio budgets for chloromethane and bromomethane. Although these isotope budgets still have gaps and substantial uncertainties, the missing pieces of information can be identiﬁed, and the necessary experimental studies are well within the reach of existing methods. It is somewhat surprising that there has been little effort towards studying the isotope ratio budgets for other halogenated VOCs. Especially for 1,1,1-trichloroethane (methyl chloroform) and the hydrochloroﬂuorocarbons (HCFC), isotope ratio studies have the potential

Gas Chromatography–Isotope Ratio Mass Spectrometry

459

of providing valuable constraints for their budgets. For these compounds, removal in the troposphere by reaction with OH-radicals is an important sink. This can and has been used to derive estimates for the average tropospheric OH-radical concentration. One of the obvious problems is that discrepancies in the mass budget may be explained by errors in the emission rates as well as changes in the OH-radical concentration. Isotope ratio budgets can supply the additional constraints needed to differentiate between these possibilities. The experimental techniques for measurements of the isotope ratios for halogenated VOCs in the atmosphere are available, and, due to the long atmospheric lifetime of these compounds, the number of atmospheric measurements needed to derive accurate averages for their isotope ratios is within reasonable limits. However, accurate measurements of the KIEs for the dominant tropospheric loss reaction of these compounds, the reaction with OH-radicals, will present some experimental challenges due to the low reactivity of these compounds. Nevertheless, such measurements are within the reach of existing KIE measurement techniques. There is one area where it can be expected that isotope ratio studies will have a number of valuable applications, although presently there is little information available. The chemistry of the oxidation products of atmospheric VOCs is extremely complex. Therefore, any additional constraints, including isotope ratio studies, will be highly welcome. Possible applications range from studies of formaldehyde to methacrolein and methyl vinyl ketone and secondary POM. Inherent in the complexity of the atmospheric chemistry of secondary VOCs and POM is the need for laboratory studies of isotope fractionation effects associated with the formation and loss reactions and measurement techniques for their isotope ratios in the atmosphere. Both tasks are experimentally challenging. Nevertheless, it has been demonstrated that such measurements are possible. Kawamura and Watanabe (2004) described a method for compound-speciﬁc carbon isotope ratio measurements of dicarboxylic acids and ketocarboxylic acids in atmospheric POM and very recently ambient and laboratory studies of methacrolein and methyl vinyl ketone formed by the oxidation of isoprene have been conducted (R. Iannone, R. Koppmann and J. Rudolph, unpublished results). The examples for applications of VOC isotope ratio measurements given above should not be considered as a complete list of all possibilities. With more detailed quantitative information it is very likely that there will be additional applications. Especially further development of experimental and interpretative tools have the potential of leading to a variety of new and valuable applications of VOC isotope ratio studies. An obvious development in experimental techniques is measurement of VOC isotope ratios for other elements but carbon. Methods for the measurement of VOC hydrogen isotope ratios are within the range of existing state of the art methods. With very few exceptions, all atmospheric VOCs contain hydrogen. The value of combining measurements of isotope ratios for different elements in one compound has been demonstrated in numerous applications in various ﬁelds of science. There can be little doubt that such an approach will be equally valuable for studies of atmospheric VOCs. Finally, it should be remembered that the diagnostic value of VOC stable isotope ratio measurements depends on the conceptual and numerical tools available for interpretation. Numerical model simulations belong to the most versatile and advanced methods to evaluate atmospheric observations and thus have a large potential for interpreting observations of VOC isotope ratios. Similarly, comparison between VOC isotope ratio measurements

460

Volatile Organic Compounds in the Atmosphere

and observations can play an important role for model testing and validation. Presently, there are very few modelling studies of VOC isotope ratios, but it has already been shown that VOC isotope ratio predictions can be added to models without changing the model chemistry. This opens a wide range of possibilities to use VOC isotope ratios as diagnostic tracers for the performance of numerical models.

References Affek, H.P. and Yakir, D. (2003) Natural abundance carbon isotope composition of isoprene reﬂects incomplete coupling between isoprene synthesis and photosynthetic carbon ﬂow. Plant Physiology, 131: 1727–6. Allan, W., Manning, M.R., Lassey, K.R., et al. (2001) Modeling the variation of δ 13 C in atmospheric methane: Phase ellipses and the kinetic isotope effect. Global Biogeochemical Cycles, 15: 467–81. Anderson, R.S. (2005) Carbon Kinetic Isotope Effects in the Gas-Phase Reactions of Nonmethane Hydrocarbons with Hydroxyl Radicals and Chlorine Atoms. PhD Thesis, York University, Toronto, Canada. Anderson, R.S., Czuba, E., Ernst, D., et al. (2003) Method for measuring carbon kinetic isotope effects of gas-phase reactions of light hydrocarbons with the hydroxyl radical. Journal of Physical Chemistry, 107: 6191–9. Anderson, R.S., Iannone, R., Thompson, A.E., et al. (2004a) Carbon kinetic isotope effects in the gasphase reactions of aromatic hydrocarbons with the OH radical at 296 ± 4 K. Geophysical Research Letters, 31. Anderson, R.S., Iannone, R., Thompson, A.E., et al. (2004b) Carbon kinetic isotope effects in the gasphase reactions of light alkanes and ethene with the OH radical at 296 ± 4 K. Journal of Physical Chemistry A, 108(52), 11537–44 . Archbold, M.E., Redeker, K.R., Davis, S., et al. (2005) A method for carbon stable isotope analysis of methyl halides and chloroﬂuorocarbons at pptv concentrations. Rapid Communications in Mass Spectrometry, 19: 337–42. Aschmann, S.M. and Atkinson, R. (1995) Rate constants for the gas-phase reactions of alkanes with Cl-atoms at 296 ± 2 K. International Journal of Chemical Kinetics, 27: 613–22. Atkinson, R. (1989) Kinetics and mechanisms of the gas-phase reactions of the hydroxyl radical with organic compounds. Journal of Physical and Chemical Reference Data, 1: 1–246. Atkinson, R. (1994) Gas phase tropospheric chemistry of organic compounds. Journal of Physical Chem. Ref. Data Monogr, 2: 11–216. Atkinson, R. (1997) Gas phase tropospheric chemistry of organic compounds. 1. Alkanes and Alkenes. Journal of Physical Chemistry Reference, 26: 215–19. Beukes, J.A., D’Anna, B., Bakken, V., et al. (2000) Experimental and theoretical study of the F, Cl and Br reactions with formaldehyde and acetaldehyde. Journal of Physical Chemistry Chemical Physics, 2: 4049–60. Biegeleisen, J. and Wolfsberg, M. (1958) Theoretical and experimental aspect of isotope effects in chemical kinetics. In: I. Progogine (Ed.) Advances in Chemical Physics. Interscience Publishers, Inc., NY, pp. 15–76. Bill, M., Miller, L.G. and Goldstein, A.H. (2002a) Carbon isotope fractionation of methyl bromide during agricultural fumigation. Biogeochemistry, 60: 181–90. Bill, M., Rhew, R.C., Weiss, R.F., et al. (2002b) Carbon isotope ratios of methyl bromide and methyl chloride emitted from coastal salt marshes. Geophysical Research Letters, 29: 1045–8. Bill, M., Conrad, M.E. and Goldstein, A.H. (2004) Stable carbon isotope composition of atmospheric methyl bromide. Geophysical Research Letters, 31: L0410.

Gas Chromatography–Isotope Ratio Mass Spectrometry

461

Brenninkmeijer, C.A.M., Lowe, D.C., Manning, M.R., et al. (1995) The 13 C, 14 C, and 18 O istopic composition of CO, CH4 , and CO2 in the higher southern latitudes lower stratosphere. Journal of Geophysical Research, 100: 26163–72. Brenninkmeijer, C.A.M. and Röckmann, T.T. (1997) Principal factors determining the 18 O/16 O ratio of atmospheric CO as derived from observations in the southern hemispheric troposphere and lowermost stratosphere. Journal of Geophysical Research, 102: 25477–85. Burgoyne, T.W. and Hayes, J.M. (1998) Quantitative production of H2 by pyrolysis of gas chromatographic efﬂuents. Analytical Chemistry, 70: 5136–41. Chiltz, G., Eckling, R., Goldﬁnger, P., et al. (1963) Kinetic isotope effect in photochlorination of H2 , CH4 , CHCl3 , and C2 H6 . Journal of Physical Chemistry, 38 (5): 1053–61. Ciais, P., Tans, P.P., Trolier, M., et al. (1995) A large northern hemisphere terrestrial CO2 sink indicated by the 13 C/12 C ratio of atmospheric CO2 . Science, 269: 1098–102. Ciais, P., Tans, P.P., Denning, A.S., et al. (1997) A three-dimensional synthesis study of δ 18 O in atmospheric CO2 , simulations with the TM2 transport model. Journal of Geophysical Research, 102: 5873–83. Conny, J.M. and Currie, L.A. (1996) The isotopic characterization of methane, non-methane hydrocarbons and formaldehyde in the troposphere. Atmospheric Environment, 30: 621–38. Conny, J.M., Verkouteren, R.M. and Currie, L.A. (1997) Carbon 13 composition of tropospheric CO in Brazil: A model scenario during the biomass burn season. Journal of Geophysical Research, 102: 10683–93. Craig, H. (1957) Isotopic standards for carbon and oxygen and correction factors for mass-spectrometric analysis of carbon dioxide. Geochimica et Cosmochimica Acta, 12: 133–49. Craig, H., Chou, C.C., Welhan, J.A., et al. (1988) The isotopic composition of methane in polar ice cores. Science, 242: 1535–9. Crawford, M.A., Li, Z., Heuerman, H.A., et al. (2004) A kinetic and product study of reaction of chlorine atom with CH3 CH2 OD. International Journal of Chemical Kinetics, 36: 584–90. Czapiewski, K.V., Czuba, E., Huang, L., et al. (2002) Isotopic composition of non-methane hydrocarbons in emissions from biomass burning. Journal of Atmospheric Chemistry, 43: 45–60. Czuba, E. (1999) Development of a Technique to Study Stable Carbon Isotope Composition of NMHCs in Ambient Air; MSc. Thesis, York University, Toronto. D’Anna, B., Bakken, V., Are Beukes, J., et al. (2003) Experimental and theoretical studies of gas phase NO3 and OH radical reactions with formaldehyde, acetaldehyde and their isotopomers. Journal of Physical Chemistry Chemical Physics, 5 (9): 1790–805. Dobis, O., Benson, S.W. and Mitchell, T.J. (1994) Kinetics of the reactions of C1 atoms with C2 D6 and C2 D5 and the disproportionation of 2C2 D5 . The deuterium isotope effect. Journal of Physical Chemistry, 98: 12284–93. Donahue, N.M. (2001) Revisiting the Hammond postulate: The role of reactant and product ionic states in regulating barrier heights, locations, and transition state frequencies. Journal of Physical Chemistry A, 105: 1489–97. Donahue, N.M., Anderson, J.G. and Demerjian, K.L. (1998) New rate constants for ten OH-Alkane reactions from 300–400 K: An assessment of accuracy. Journal of Physical Chemistry A, 102: 3121–6. Droege, A.T. and Tully, F.P. (1986a) Hydrogen atom abstraction from alkanes by hydroxyl. 5. n-Butane. Journal of Physical Chemistry, 90: 5937–41. Droege, A.T. and Tully, F.P. (1986b) Hydrogen-atom abstraction from alkanes by hydroxyl. 3. propane. Journal of Physical Chemistry, 90: 1949–54. Droege, A.T. and Tully, F.P. (1987a) Hydrogen-atom abstraction from alkanes by OH. 6. cyclopentane and cyclohexane. Journal of Physical Chemistry, 91: 1222–5. Droege, A.T. and Tully, F.P. (1987b) Kinetics of the reaction of the hydroxyl radical with dimethyl ether and diethyl ether. International Journal of Chemical Kinetics, 19: 251–9.

462

Volatile Organic Compounds in the Atmosphere

Enghoff, M.B., Von Hessberg, P., Nielsen, C.J., et al. (2003) The kinetic isotope effects in the reactions of four ethene isotopologues with chlorine and bromine atoms. Journal of Physical Chemistry A, 107 (39): 7667–70. Farkas, E., Szilágyi, I., Dóbé, S., et al. (2003) Kinetic isotope effect in the reaction of OH radical with acetone-d6. Reaction Kinetics and Catalysis Letters, 80: 351–8. Feilberg, K.L., Johnson, M.S. and Nielsen, C.J. (2004) Relative reaction rates of HCHO, HCDO, DCDO, H13 CHO, and HCH18 O with OH, Cl, Br, and NO3 radicals. Journal of Physical Chemistry A, 108 (36): 7393–8. Feilberg, K.L., Grifﬁth, D.W.T., Johnson, M.S., et al. (2005a) The 13 C and D kinetic isotope effects in the reaction of CH4 with Cl. International Journal of Chemical Kinetics, 37 (2): 110–18. Feilberg, K.L., D’Anna, B., Johnson, M.S., et al. (2005b) Relative tropospheric photolysis rates of HCHO, H13 CHO, HCH18 O, and DCDO measured at the European photoreactor facility. Journal of Physical Chemistry A, 109: 8314–19. Francey, R.J., Tans, P.P., Allison, C.E., et al. (1995) Changes in oceanic and terrestrial carbon uptake since 1982. Nature, 373: 326–30. Funk, J.L., Mak, J.E. and Lerdau, M.T. (2004) Stress-induced changes in carbon sources for isoprene production in populus deltoides. Plant, Cell and Environment, 24: 747–55. Fung, I., Field, C.B., Berry, J.A., et al. (1997) Carbon 13 exchanges between the atmosphere and biosphere. Global Biogenochemical Cycles, 11 (4): 507–33. Gerst, S. and Quay, P. (2000) The deuterium content of atmospheric molecular hydrogen: Method and initial measurements. Journal of Geophysical Research, 105 (D21): 26433–45. Gerst, S. and Quay, P. (2001) Deuterium component of the global molecular hydrogen cycle. Journal of Geophysical Research, 106 (D5): 5021–31. Gierczak, T., Talukdar, R.K., Herndon, S., et al. (1997) Rate coefﬁcients for the reactions of hydroxyl radicals with methane and deuterated methanes. Journal of Physical Chemistry A, 101 (17): 3125–34. Glasius, M., Wessel, S., Christensen, C.S., et al. (2000) Sources to formic acid studied by carbon isotopic analysis and air mass characterization. Atmospheric Environment, 34: 2471–9. Gola, A.A., D’Anna, B., Feilberg, K.L., et al. (2005) Kinetic isotope effects in the gas phase reactions of OH and Cl with CH3 Cl, CD3 Cl, and 13 CH3 Cl. Atmospheric Chemistry and Physics Discussion, 5: 3873–98. Goldstein, A.H. and Shaw, S.L. (2003) Isotopes of volatile organic compounds: An emerging approach for studying atmospheric budgets and chemistry. Chemical Reviews, 102: 5025–48. Gray, J.R., Lacrampe-Couloume, G., Gandhi, D., et al. (2002) Carbon and hydrogen isotopic fractionation during biodegradation of methyl tert-butyl ether. Environmental Science and Technology, 36: 1931–8. Greiner, N.R. (1970) Comparison of kinetics of alkane H-atom abstraction by methyl and hydroxyl radicals. Journal of Chemical Physics, 53: 1285–7. Gupta, M., Tyler, S. and Cicerone, R. (1996) Modeling atmospheric δ 13 CH4 and the causes of recent changes in atmospheric CH4 amounts. Journal of Geophysical Research, 101: 22923–32. Hagemann, R., Nief, G. and Roth, E. (1970) Absolute isotopic scale for deuterium analysis of natural waters. Absolute D/H ratio for SMOW. Tellus, 22: 712–15. Harrington, R.R., Poulson, S.R., Drever, J.I., et al. (1999) Carbon isotope systematics of monoaromatic hydrocarbons: Vaporization and adsorption experiments. Organic Geochemistry, 30: 765–75. Harper, D.B., Kalin, R.M., Hamilton, J.T.G., et al. (2001) Carbon isotope ratios of chloromethane of biological origin: Potential tool in determining biological emissions. Environmental Science and Technology, 35: 3616–19. Harper, D.B., Hamilton, J.T.G., Ducrocq, V., et al. (2003) The distinctive isotopic signature of plantderived chloromethane: Possible application in constraining the atmospheric chloromethane budget. Chemosphere, 52: 433–6.

Gas Chromatography–Isotope Ratio Mass Spectrometry

463

Harry, C., Arey, J. and Atkinson, R. (1999) Rate constants for the reactions of OH radicals and Cl atoms with di-n-propyl ether and di-n-butyl ether and their deuterated analogs. International Journal of Chemical Kinetics, 31: 425–31. Hitsuda, K., Takahashi, K., Matsumi, Y., et al. (2001a) Kinetics of the reactions of Cl(2 P1/2 ) and Cl(2 P3/2 ) atoms with C2 H6 , C2 D6 , CH3 F, C2 H5 F, and CH3 CF3 at 298 K. Journal of Physical Chemistry A, 105: 5131–6. Hitsuda, K., Takahashi, K., Matsumi, Y., et al. (2001b) Kinetics of the reactions of Cl(2 P1/2 ) and Cl(2 P3/2 ) atoms with C3 H8 , C3 D8 , n-C4 H10 , and i-C4 H10 at 298 K. Chemical Physics Letters, 346: 16–22. Iannone, R., Anderson, R.S., Rudolph, J., et al. (2003) The carbon kinetic isotope effects of ozonealkene reactions in the gas-phase and the impact of ozone reactions on the stable carbon isotope ratios of alkenes in the atmosphere. Journal of Geophysical Letters, 30: 1684. Iannone, R., Anderson, R.S., Vogel, A., et al. (2004) Laboratory studies of the hydrogen kinetic isotope effects (KIES) of the reaction of non-methane hydrocarbons with the OH radical in the gas phase. Journal of Atmospheric Chemistry, 47: 192–208. Iannone, R., Anderson, R.S., Vogel, A., et al. (2005) The hydrogen kinetic isotope effects of the reactions of n-alkanes with chlorine atoms in the gas phase. Journal of Atmospheric Chemistry, 50: 121–38. Irei, S., Huang, L., Collin, F., et al. (2006) Flow reactor studies of the stable carbon isotope composition of secondary particulate organic matter generated by OH-radical induced reactions of toluene. Atmospheric Environment, 40: 5857–67. Johnson, B.J. and Dawson, A.G. (1990) Collection of formaldehyde from clean air for carbon isotopic analysis. Environmental Science and Technology, 24: 898–902. Johnson, B.J. and Dawson, G.A. (1993) A preliminary study of the carbon-isotopic content of ambient formic acid and two selected sources: Automobile exhaust and formicine ants. Journal of Atmospheric Chemistry, 17 : 123–40. Kalin, R.M., Hamilton, J.T.G., Harper, D.B., et al. (2001) Continuous ﬂow stable isotope method for study of fractionation of δ 13 C fractionation during halomethane production an degradation. Rapid Communications in Mass Spectrometry, 15: 357–63. Kawamura, K. and Watanabe, T. (2004) Determination of stable carbon isotopic compositions of low molecular weight dicarboxylic acids and ketocarboxylic acids in atmospheric aerosol and snow samples. Analytical Chemistry, 76: 5762–8. Keeling, C.D., Mook, W.G. and Tans, P.P. (1979) Recent trends in the 13 C/12 C ratio of atmospheric carbon dioxide. Nature, 277: 121–3. Keppler, F., Kalin, R.M., Harper, D.B., et al. (2004) Carbon isotope anomaly in the major plant C1 pool and its global biogeochemical implications. Biogeosciences, 1: 123–31. Keppler, F., Harper, D.B., Röckmann, T., et al. (2005) New insight into the atmospheric chloromethane budget gained using stable carbon isotope ratios. Atmospheric Chemistry and Physics Discussion, 5: 2403–11. Komatsu, D.D., Tsunogai, U., Yamaguchi, J., et al. (2004) The budget of atmospheric methyl chloride using stable carbon isotopic mass-balance, Poster A51C-0784, AGU Fall Meeting, San Francisco, 2004. Cited from Keppler et al. (2005). Komatsu, D.K., Tsunogai, U., Yamaguchi, J., et al. (2005) A selective unsaturated hydrocarbon subtraction technique for stable carbon isotopic analysis of atmospheric methyl chloride, methyl bromide, and C2 –C5 saturated hydrocarbons using continuous-ﬂow isotope ratio mass spectrometry. Rapid Communications in Mass Spectrometry, 19: 477–83. Kwok, E.S.C. and Atkinson, R. (1995) Estimation of hydroxyl radical reaction rate constants for gas-phase organic compounds using a structure-reactivity relationship: An update. Atmospheric Environment, 29: 1685–95. Lowe, D.C., Schmidt, U. and Ehhalt, D.H. (1980) A new technique for measuring atmospheric formaldehyde [CH2 O]. Geophysical Research Letters, 7: 825–8.

464

Volatile Organic Compounds in the Atmosphere

Lowe, D.C., Brenninkmeijer, C.A.M., Brailsford, G.W., et al. (1994) Concentration and 13 C records of atmospheric methane in New Zealand and Antarctica: Evidence for changes in methane sources. Journal of Geophysical Research, 99: 16913–25. Lowe, D.C., Allan, W., Manning, M.R., et al. (1999) Shipboard determinations of the distribution of 13 C in atmospheric methane in the paciﬁc. Journal of Geophysical Research, 104: 26125–35. Mancini, S.A., Ulrich, A.C., Lacrampe-Couloume, G., et al. (2003) Carbon and hydrogen isotopic fractionation during anaerobic biodegradation of benzene. Applied and Environmental Microbiology, 69: 191–8. Manning, M.R. (1999) Characteristic modes of isotopic variations in atmospheric chemistry. Geophysical Research Letters, 26: 1263–6. Manning, M.R., Brenninkmeijer, C.A.M. and Allan, W. (1997) Atmospheric carbon monoxide budget of the southern hemisphere: Implications of 13 C/12 C measurements. Journal of Geophysical Research, 102: 10673–82. Matthews, D.E. and Hayes, A.J. (1978) Isotope ratio monitoring gas chromatography-mass spectrometry. Analytical Chemistry, 50: 1465–73. McCauley, S.E., Goldstein, A.H. and DePaolo, D.J. (2001) An isotopic approach for understanding the CH3 Br budget of the atmosphere. Proceedings of National Academy of Science of USA, 96: 10006–9. McKeen, S.A., Trainer, M., Hsie, E.-Y., et al. (1990) On the indirect determination of atmospheric OHradical concentrations from reactive hydrocarbon measurements. Journal of Geophysical Research, 95: 7493–500. McKeen, S.A., Liu, S.C., Hsie, E.-Y., et al. (1996) Hydrocarbon ratios during PEM-WEST A: A model perspective. Journal of Geophysical Research, 101: 2087–109. McKenna, D.S. (1997) Analytic solution of reaction diffusion equations and implications for the concept of an air parcel. Journal of Geophysical Research, 102: 19719–13725. McKenna, D.S., Hord, C.J. and Kent, J.M. (1995) Hydroxyl radical concentrations and kuwait oil ﬁre emission rates for March 1991. Journal of Geophysical Research, 100: 26005–26. Miller, L.G., Kalin, R.M., McCauley, S.E., et al. (2001) Large carbon isotope fractionation associated with oxidation of methyl halides by methylotrophic bacteria. Proceedings of the National Academy of Science of USA, 98: 5833–7. Miller, L.G., Warner, K.L., Baesman, S.M., et al. (2004) Degradation of methyl bromide and methyl chloride in soil microcosms: Use of stable C isotope fractionation and stable isotope probing to identify reactions and the responsible microorganisms. Geochemica Cosmochemica Acta, 68: 3271–83. Mook, W.G., Koopmans, M., Carter, A.F., et al. (1983) Seasonal latitudinal, and secular variations in the abundance and the isotope ratios of atmospheric carbon dioxide. Journal of Geophysical Research, 88: 10915–33. Niki, H., Maker, P.D., Savage, C.M., et al. (1978) Relative rate constants for the reaction of hydroxyl radical with aldehydes. Journal of Physical Chemistry, 82: 132–4. Niki, H., Maker, P.D., Savage, C.M., et al. (1984) Fourier transform infrared study of the kinetics and mechanism for the reaction of hydroxyl radical with formaldehyde. Journal of Physical Chemistry, 88: 5342–4. Parrish, D.D., Hahn, C.J., Williams, E.J., et al. (1992) Indications of photochemical histories of paciﬁc air masses from measurements of atmospheric trace species at Point Arena, California. Journal of Geophysical Research, 97: 15883–901. Rieley, G. (1994) Derivatization of organic compounds prior to gas chromatographic-combustionisotope ratio mass spectrometric analysis: Identiﬁcation of isotope fractionation processes. Analyst, 119: 915–19. Roberts, J.M., Fehsenfeld, F.C., Liu, S.C., et al. (1984) Measurements of aromatic hydrocarbon ratios and NOx concentrations in the rural troposphere: Observation of air mass photochemical aging and NOx removal. Atmospheric Environment, 18: 2421–32.

Gas Chromatography–Isotope Ratio Mass Spectrometry

465

Rudolph, J. (2002) Stable carbon ratio measurements: A new tool to understand atmospheric processing of volatile organic compounds. Proceedings of the NATO Advanced Research Workshop on Global Atmospheric Change and its impact on regional air quality, Irkutsk, Russian Federation, 21–27 August 2001. In: I. Barnes (Ed.) Global Atmospheric Change and Its Impact on Regional Air Quality, NATO Science Series, IV. Earth and Environmental Sciences – Vol. 16. Kluwer Academic Publishers, Dordrecht. Rudolph, J. and Czuba, E. (2000) On the use of isotopic composition measurements of volatile organic compounds to determine the ‘photochemical age’ of an air mass. Geophysical Research Letters, 27: 3865–8. Rudolph, J. and Johnen, F.J. (1990) Measurements of light atmospheric hydrocarbons over the Atlantic in regions of low biological activity. Journal of Geophysical Research, 95: 20583–91. Rudolph, J., Lowe, D.C., Martin, R.J. and Clarkson, T.S. (1997) A novel method for compound speciﬁc determination of C13 in volatile organic compounds at ppt levels in ambient air. Geophysical Research Letters, 24: 659–62. Rudolph, J., Czuba, E. and Huang, L. (2000) The stable carbon isotope fractionation for reactions of selected hydrocarbons with OH-radicals and its relevance for atmospheric chemistry. Journal of Geophysical Research, 10: 29329–46. Rudolph, J., Czuba, E., Norman, A.L., et al. (2002) Stable carbon composition of nonmethane hydrocarbons in emissions from transportation related sources and atmospheric observations in an urban atmosphere. Atmospheric Environment, 36: 1173–81. Rudolph, J., Anderson, R.S., Czapiewski, K.V., et al. (2003) The stable carbon isotope ratio of biogenic emissions of isoprene and the potential use of stable isotope ratio measurements to study photochemical processing of isoprene in the atmosphere. Journal of Atmospheric Chemistry, 44: 39–55. Saito, T., Tsunogai, U., Kawamura, K., et al. (2002) Stable carbon isotopic compositions of light hydrocarbons over the western north paciﬁc and implication for their photochemical ages. Journal of Geophysical Research, 107: ACH 2/1–ACH 2/9. Sakugawa, H. and Kaplan, R.I. (1995) Stable carbon isotope ratio measurements of atmospheric organic acids in Los Angeles, California. Geophysical Research Letters, 22: 1509–12. Santrock, J., Studley, S.A. and Hayes, J.M. (1985) Isotopic analyses based on the mass spectrum of carbon dioxide. Analytical Chemistry, 57: 1444–8. Saueressig, G., Crowley, J.N., Bergamaschi, P., et al. (2001) Carbon 13 and D kinetic isotope effects in the reactions of CH4 with O(1 D) and OH: New laboratory measurements and their implications for the isotopic composition of stratospheric methane. Journal of Geophysical Research, 106: 23127–38. Sharkey, T.D., Loreto, F., Delwiche, C.F., et al. (1991) Fractionation of carbon isotopes during biogenesis of atmospheric isoprene. Plant Physiology, 97: 463–6. Singleton, D.L., Paraskevopoulos, G., Irwin, R.S., et al. (1988) Rate and mechanism of the reaction of hydroxyl radicals with formic and deuterated formic acids. Journal of American Chemical Society, 110: 7786–90. Sirju, A.-P. and Shepson, P.B. (1995) Laboratory and ﬁeld investigation of the DNPH cartridge technique for the measurement of atmospheric carbonyl compounds. Environmental Science Technology, 29 (2): 384–92. Slater, G.F., Lollar, B.S., Sleep, B.E., et al. (2001) Variability in carbon isotopic fractionation during biodegradation of chlorinated ethenes: Implications for ﬁeld applications. Environmental Science Technology, 35: 901–7. Smallwood, B.J., Philp, R.P. and Allen, J.D. (2002) Stable carbon isotopic composition of gasolines determined by isotope ratio gas chromatography mass spectrometry. Organic Geochemistry, 33: 149–9. Stevens, C.M. and Wagner, A.F. (1989) The role of isotope fractionation in atmospheric chemistry. Zeitschrift für Naturforschung, 44a: 376–84.

466

Volatile Organic Compounds in the Atmosphere

Stutz, J., Ezell, M.J. and Finlayson-Pitts, B.J. (1997) Inverse kinetic isotope effect in the reaction of atomic chlorine with C2 H4 and C2 D4 . Journal of Physical Chemistry A, 101: 9187–90. Stutz, J., Ezell, M.J., Ezell, A.A., et al. (1998) Rate constants and kinetic isotope effects in the reactions of atomic chlorine with n-butane and simple alkenes at room temperature. Journal of Physical Chemistry A, 102: 8510–19. Taatjes, C.A., Christensen, L.K., Hurley, M.D., et al. (1999) Absolute and site-speciﬁc abstraction rate coefﬁcients for reactions of Cl with CH3 CH2 OH, CH3 CD2 OH, and CD3 CH2 OH between 295 and 600 K. Journal of Physical Chemistry A, 103: 9805–14. Tanner, R.L., Ziclinska, B., Uberna, E., et al. (1996) Concentration of carbonyl compounds and the carbon isotope ratios of formaldehyde at a coastal site in nova scotia during the NARE summer intensive. Journal of Geophysical Research, 101: 28961–70. Thompson, A., Rudolph, J., Rohrer, F., et al. (2003) Concentration and stable carbon isotopic composition of ethane and benzene using a global 3D isotope inclusive chemical tracer model. Journal of Geophysical Research, 108: D13. Thompson, A.E. (2003) Stable Carbon Isotope Ratios of Nonmethane Hydrocarbons and Halocarbons in the Atmosphere. PhD. Thesis, York University, Toronto. Thompson, A.E., Anderson, R.S., Rudolph, J., et al. (2002) Stable carbon isotope signatures of background tropospheric chloromethane and CFC113. Biogeochemistry, 60: 191–211. −C5 hydrocarbons Tsunogai, U., Yoshida, N. and Gamo, T. (1999) Carbon isotopic composition of C2 − and methyl chloride in urban, coastal, and maritime atmospheres over the western north paciﬁc. Journal of Geophysical Research, 104: 16033–9. Tully, F.P., Koszykowski, M.L. and Binkley, J.S. (1985) Hydrogen-atom abstraction from alkanes by hydroxyl radical. I. Proceedings of the 20th Neopentane and Neooctane. Symposium (International) on Combustion, 715–21. Tully, F.P., Goldsmith, J.E.M. and Droege, A.T. (1986) Hydrogen atom abstraction from alkanes by hydroxyl. 4. Isobutane. Journal of Physical Chemistry, 90: 5932–7. Tyler, S.C., Ajie, H.O., Rice, A.L., et al. (2000) Experimentally determined kinetic isotope effects in the reaction of CH4 with Cl: Implications for atmospheric CH4 . Geophysical Research Letters, 27: 1715–18. Vairavamurthy, A., Roberts, J. and Newman, L. (1992) Methods for determination of low-molecularweight carbonyl-compounds in the atmosphere. Atmospheric Environment A, 26: 1965–93. Wallington, T.J., Dagaut, P. and Kurylo, M.J. (1988) Correlation between gas-phase and solutionphase reactivities of hydroxyl radicals toward saturated organic compounds. Journal of Physical Chemistry, 92: 5024–8. Ward, J.A.M., Ahad, J.M.E., Lacrampe-Couloume, G., et al. (2000) Hydrogen isotope fractionation during methanogenic degradation of toluene: Potential for direct veriﬁcation of bioremediation. Environmental Science and Technology, 34: 4577–81. Wen, S., Feng, Y., Wang, X., et al. (2004) Development of a compound-speciﬁc isotope analysis method for acetone via 2,4-dinitrophenylhydrazine derivatization. Rapid Communications in Mass Spectrometry, 18: 2669–72. Wen, S., Feng, Y., Yu, Y., et al. (2005) Development of a compound-speciﬁc isotope analysis method for atmospheric formaldehyde and acetaldehyde. Environmental Science and Technology, 39: 6202–7. Yeh, H.-W. and Epstein, S. (1981) Hydrogen and carbon isotopes of petroleum and related matter. Geochimica and Cosmochimica Acta, 45: 753–62.

Volatile Organic Compounds in the Atmosphere Edited by Ralf Koppmann Copyright © 2007 by Blackwell Publishing Ltd

Chapter 11

Comprehensive Two-Dimensional Gas Chromatography Jacqueline F. Hamilton and Alastair C. Lewis

11.1

Introduction

One of the greatest difﬁculties in analysing the organic content of the atmosphere is the sheer number of individual species present. The range of organic compounds in the atmosphere is vast, with emission sources that are complex and varied, being both anthropogenic and biogenic in origin. In urban areas, the troposphere is dominated by petrochemical emissions, ranging in size from 2 to upwards of 30 carbon atoms, and their oxidation products formed through reactions with OH, O3 and NO3 . As molecular weight increases, the number of possible compounds increases exponentially. An example of this is the monoaromatic compounds; with one carbon substituent there is only one possible isomer, toluene. However, with ﬁve carbon substituents the number of possible isomers rises to 90 and to ∼450 when there are six carbons on the ring. In areas inﬂuenced by biogenic emissions, the number and type of compounds emitted are dependent on factors such as plant type, temperature, season and light conditions. Gas chromatography (GC) coupled with ﬂame ionisation detection (FID) or mass spectrometry are the most common methods for the analysis of complex organic mixtures, such as those found in atmospheric samples, and have been described in the previous chapter. However, the separation power of GC is limited, and even with extremely long, narrow bore columns, the total number of species that can be isolated is considerably lower than the vast range of volatile organic compounds (VOCs) present in the atmosphere. A recent development in complex mixture analysis has been the introduction of comprehensive two-dimensional gas chromatography (GC × GC), which was pioneered by J. Phillips in the early 1990s (Liu and Phillips 1991; Phillips and Xu 1995), but not applied routinely to atmospheric samples until 2000 (Lewis et al. 2000). GC × GC is a hyphenated chromatographic technique, involving the coupling of two GC columns, providing greater resolution of organic mixtures than ever before. It has found use in a number of branches of atmospheric science, including ﬁeld measurements, smog chamber studies and aerosol analysis. In this chapter, a description of the fundamentals of comprehensive two-dimensional gas chromatography will be presented, followed by a number of examples of its use in the analysis of VOCs in the atmosphere.

468

Volatile Organic Compounds in the Atmosphere

Primary column separation Second column separation Figure 11.1 Peak proﬁles of primary and secondary peaks in comprehensive GC.

11.2

Fundamentals of comprehensive gas chromatography

In comprehensive two-dimensional gas chromatography, generally abbreviated as GC × GC, two columns of different selectivity are coupled together via a mid-point modulator. The modulator sends discrete bundles or ‘plugs’ of eluent from the ﬁrst column to the second, effectively re-injecting small partially separated aliquots of the original sample. The two columns have different stationary phases and should separate chromatographically according to different analyte properties. One of the most common column combinations is a non-polar primary column, which separates on the basis of volatility and a polar secondary column, which separates on the basis of polarity. The modulator must pre-concentrate a bundle of eluent and then launch it onto the second column for GC × GC to function; otherwise, the peak capacity of the system is simply the sum of the two columns. The second column’s chromatographic separation must also be short so that the analysis is complete before the next bundle of eluent is launched from the modulator. This produces a series of short second-dimension chromatograms. Peak widths of analytes eluting from the ﬁrst column should be at least three to four times the modulation frequency. Thus, each analyte peak eluting from the ﬁrst dimension will be present in at least three secondary chromatograms. Figure 11.1 shows the relationship between the peak proﬁle of an analyte as it elutes from the primary and secondary columns. In GC × GC, it is essential that the eluent is injected onto the second column as a narrow band. This most obviously aids resolution on the second column, but can also greatly improve the signal/noise ratio and, thus, sensitivity of the technique. Peaks reaching the detector are much narrower and much higher than the primary peak due to full mass conservation, as can be seen in Figure 11.1. A recent study (Lee et al. 2001) has shown that signal enhancement using GC × GC can be modelled to be up to a factor 70 over one-dimensional separations, albeit neglecting the inﬂuence of additional sources of noise arising from data capture at high speeds.

11.2.1

Peak capacity

Unlike heart-cut GC, where a small section of the eluent is transferred to a second column, in GC × GC the entire eluent is transferred in discrete bundles to the second column. The overall resolving power of a chromatographic column can be described in terms of peak capacity, n. This is the maximum number of component peaks that can be theoretically resolved on a given column. The peak capacity can be related to the number of theoretical plates (N ) as shown in Equation 11.1 (Giddings 1990 and references therein).

Comprehensive Two-Dimensional Gas Chromatography

469

A chromatographic column, of length, L, can be separated into a series of identical segments called theoretical plates, N . The process of partitioning can be thought of as a series of absorption and desorption events. Each interaction with the stationary phase can be thought of as a ‘plate’ for separation. n ≈ 1 + N 1/2

(11.1)

This equation assumes that the peaks will be evenly spaced throughout the chromatogram. In real samples, however, this is not true, and the actual peak capacity is much lower than that calculated using Equation 11.1. The statistical model of overlap (Davis and Giddings 1983) takes into account the random way in which component peaks fall over the separation space. This model showed that in chromatograms lacking uniform spacing, resolution was reduced due to overlapping peaks. The number of single-component peaks was actually no more than 18% of the potential peak capacity. In heart-cut GC, the maximum peak capacity is given by the sum of the two columns’ respective peak capacity, as shown below: n1 + n2 = ntotal

(11.2)

where n1 and n2 are the peak capacities of columns one and two, respectively, and ntotal is the peak capacity of the entire system. In GC × GC, the theoretical peak capacity for orthogonal separations approaches that of the product of the two columns respective peak capacities (Liu et al. 1995). n1 × n2 ≈ ntotal

(11.3)

In GC × GC, there is generally a degree of retention correlation between dimensions (there is an element of volatility selectivity in all GC), and the actual peak capacity is somewhat lower than the theoretical value. However, using two columns with only modest peak capacities can yield very high total peak capacities due to this product relation, which is simply not possible in heart-cut GC. If the primary and secondary column peak capacities are 50 and 20, respectively, typical values for this type of separation, the capacity of the system is 1 000. This is already well beyond the limits of conventional GC. Optimising the system can improve this ﬁgure dramatically, giving peak capacities in the tens of thousands (Dalluge et al. 2002).

11.2.2

Orthogonality

One of the key features of GC × GC is the orthogonality of the system. The separations performed on the two columns must be independent of each other to ensure maximum use of the two-dimensional separation plane. The retention of an analyte on one column must not be related to its retention on the other column. In GC × GC, the primary column separation is long, and the temperature in the oven is ramped throughout the run. However, the second column separation is very fast in comparison, on the order of a few seconds. Using conventional ramping rates, for example 1–3◦ C/min, the second column separation can essentially be considered as isothermal. Where both columns are based on similar stationary phases, that is, a 100% dimethyl-polysiloxane primary column and a 50% phenyl dimethyl-polysiloxane secondary column, this difference is of utmost importance. When

470

Volatile Organic Compounds in the Atmosphere

Signal A B

5.0

0

2.5 30 60

Primary retention time (s)

0.0

Secondary retention time (s)

Figure 11.2 Relationship between one-dimensional, two-dimensional and the contour plot in GC × GC.

columns are properly tuned, the volatility part of the second column’s retention mechanism is cancelled out as the entire second column separation occurs at the same temperature. This allows a polarity-based secondary column separation.

11.2.3

Visualisation

GC × GC produces a series of very short second-dimension chromatograms. These can be reconstructed and visualised as a contour plot where the retention time on columns one and two represent the X- and Y-axis, respectively, and the peak height is represented as a coloured contour. The process of producing a GC × GC chromatogram is shown in Figure 11.2. Computer software is used to convert a single-stream one-dimensional chromatogram, into a two-dimensional matrix, which can be viewed using a visualisation software package, a number of which are now commercially available.

11.2.4

Ordered chromatograms

One of the most useful features of GC × GC is the production of ordered chromatograms. GC × GC can create very high peak capacities; however, the effectiveness of this increased peak capacity is highly dependant on the distribution of component peaks. If compounds with similar structures are ordered within the separation space, then it is more likely that the effective peak capacity will be high (Davis and Giddings 1983). Giddings (1990) introduced the idea of sample dimensionality to describe the complexity of a sample with respect to predicting order/disorder in multidimensional separations. A series of n-alkanes require only one chemical variable to sufﬁciently specify each member, for example, volatility. Giddings demonstrated that the degree of order within a chromatogram was determined by the sample dimensionality and the separation system dimensionality. The system dimensionality refers to the number of separation mechanisms, with different selectivity used in the system. Thus, the dimensionality in conventional GC and GC × GC is one and two, respectively. In the case of complex hydrocarbon mixtures, the dimensionality is also two, and thus the system and sample dimensionality match. Ledford et al. (1996) showed the analysis of

Comprehensive Two-Dimensional Gas Chromatography

471

Retention time 2 (s)

0.5 1.0

2

1

1.5 3 2.0 2.5

4

3.0 3.5 0

10

20

30

40

50

Retention time 1 (min)

Figure 11.3 Structured GC × GC chromatogram of gasoline. Red square is expanded in lower section. 1, alkanes; 2, alkenes and cycloalkanes; 3, monoaromatics; 4, polyaromatics. This image appears in full colour in the plate section that follows page 268 as Plate 7.

petroleum products by GC × GC produced structured chromatograms, where groups of similar analytes formed bands across the separation plane. This allows maximum use of the retention space, which in turn, maximises effective peak capacity. This structured nature is shown in Figure 11.3, for the analysis of a gasoline sample. As retention time on column one increases, a corresponding decrease in analyte volatility is seen. Thus, within a speciﬁc group, for example, the alkanes in group 1, moving from left to right of the chromatogram generally indicates an increasing carbon number. This is mirrored in the Y-axis, where an increasing retention time corresponds to an increase in polarity of the separated analytes, thus moving from the alkanes (group 1) to the polyaromatics (group 4). This greatly aids peak identiﬁcation, with compounds arranged according to their chemical and structural properties. The red square in Figure 11.3 is expanded in the lower section and shows a common feature of structured GC × GC chromatograms. The ‘roof tile’ structure is apparent where each tile corresponds to an isobaric group, in this case monoaromatics isomers with C4 , C5 , C6 and C7 substitution. These visual representations can simplify pattern matching within a series of samples, in comparison to one-dimensional chromatograms.

11.3

Modulators

The key component of the GC × GC system is the modulator, which delivers discrete bundles of eluent from the primary to the secondary column. The modulator must concentrate a fraction and launch it onto the second column in a narrow band to ensure maximum resolution on the second column. A number of possible modulation devices have

472

Volatile Organic Compounds in the Atmosphere

been introduced, including thermal methods, both heated and cryogenic and valve-based modulators. Almost all types of GC × GC modulators have been used in the analysis of VOCs in the atmosphere, with the choice highly dependent on the analytes of interest.

11.3.1

Thermal modulators

The original modulator introduced by Liu and Phillips (1991), the thermal desorption modulator, incorporated a piece of capillary tubing coated with thin layers of gold or aluminium. The modulator is resistively heated using a pulse of electrical current. The analytes are trapped on a thick ﬁlm of stationary phase in the modulator tube and are subsequently desorbed when the modulator is heated. The analytes are then swept onto the secondary column by the carrier gas. This modulator was found to be unreliable as the application of the thin layer of metallic paint can be non-uniform and can cause local overheating and a temperature gradient within the modulator. Electrically heated modulators were eventually replaced by mechanical movement modulators with two different approaches, both being highly successful. The ﬁrst is based on rotating mechanical heater, termed ‘the sweeper’, and the second based on cryogenic cooling. Heated modulators have not found usage in atmospheric analysis as a result of the high maintenance required, for example, the repeated heating cycle can cause degradation of the stationary phase. The heater must also be approximately 100◦ C above the analyte boiling point, producing a volatility range restriction. The second type of thermal mechanical modulator involves the use of a cryogenic trap and was pioneered by Marriott and Kinghorn (1997) for the analysis of essential oils, and its use extended to atmospheric samples in 2000 (Lewis et al. 2000). The chromatographic column, usually the second dimension, ﬁts through the centre of a small chamber or trap cooled using CO2 . The trap is physically moved back and forth along the column. The efﬂuent from the ﬁrst column is condensed as it enters the region cooled by the trap and accumulates into a small band of solutes. The cryogenic trap then moves position, and the oven heats up the short length of column, which was under the trap. The plug of solutes rapidly heats up, vaporises and is swept onto the second column by the carrier gas. The trap can be moved by a number of means including pneumatic, electrical or stepper motor. The most modern design is the Longitudinally Modulated Cryogenic System (LMCS) (Kinghorn et al. 2000), which incorporates a fast-acting solenoid valve and a pneumatic ram, or stepper motor, mounted on a stand. Rubber spacers are placed between the mounting plate and the GC oven chassis to minimise vibrations. Liquid CO2 is used as the coolant, and dry nitrogen or air purge is used to prevent ice build up. A dual-stage jet modulator has also been introduced, where two jets of cryogen are sprayed onto the capillary column, to trap, focus and release analytes (Beens et al. 2001; Ledford and Billesbach 2000). The release of analytes can be achieved using the ambient oven temperature alone or by the use of hot air jets. Original jet modulators used liquid CO2 as the cryogen, but recent improvements allow the use of liquid N2 or N2 gas that has been cooled by passing through tubing within a liquid N2 Dewar. A secondary oven may also be housed within the GC oven, to hold the second column. It is usually held at a higher temperature than the main oven to ensure no thermal cold spots. The sequential trapping mechanism of the cryojet modulator is shown in Figure 11.4. The

Comprehensive Two-Dimensional Gas Chromatography

473

Column 1: Non-polar phase Column 2: Polar phase Heater 1 Off and Heater 2 Off

Heaters Off

1

4

1

2

1

2 Analytes partially resolved on Column 1

Cryogenic Cooling On

1

2

Stage 1 returns to thermal trapping mode and the analytes are trapped on Stage 2 of the thermal modulator Cryo 1 On and Cryo 2 On 1

2

Heater 1 Off and Heater 2 On

2

1

2

1

2

5 Analytes trapped on Stage 1 of the thermal modulator

1

2

1

2

Analytes released to Column 2

Cryo 1 On and Cryo 2 Off Heater 1 On and Heater 2 Off

3

1

2

1

2

Heater 1 Off and Heater 2 Off

6 Analytes released to Stage 2 of the thermal modulator

Cryo 1 Off and Cryo 2 On

1

2

Analytes separate on Column 2 next bands enter the thermal modulator Cryo 1 On and Cryo 2 On 1

2

Figure 11.4 Modulation using LECO Thermal Quad Jet Modulator. Figure reproduced courtesy of LECO Corporation, St. Joseph, MI. Licensed by Zeox.

jets use dual stage trapping to ensure minimum bandwidth and breakthrough of volatile analytes. Two-stage modulation can also be achieved using a single jet and a delay loop (http://www.zoex.com/TechnicalNote_word2.doc). The column passes the cold jet and is looped around and passes under the cold jet again. Analytes are trapped as they enter the cooled region as previously. When the hot jets are activated, the analytes are released and travel through the loop, usually 0.6–1 m. The cold jets are then reactivated and the analytes trapped a second time, as they pass through the jets. Cryogenic modulations allow the range of volatilities amenable to GC × GC to be increased, compared to heated thermal methods. The majority of atmospheric applications of GC × GC have involved the use of cryogenic modulators, particularly in aerosol analysis. However, there are limitations, the most important of which relates to ﬁeld measurements, where the use of a cryogen is avoided wherever possible. In addition, highly volatile and polar compounds are prone to breakthrough from the trap, causing very wide peaks and a severe loss of chromatographic efﬁciency.

11.3.2

Valve modulators

In valve modulation, narrow pulses of eluent are diverted from the primary column to the secondary column using a switching valve. Opiteck et al. (1998) have demonstrated the use of valves in two-dimensional liquid chromatography, and valves are regularly used in

474

Volatile Organic Compounds in the Atmosphere

heart-cut GC. The ﬁrst valve-based GC × GC system, incorporated a multi-port diaphragm valve, housed within the GC oven (Bruckner et al. 1998). Most of the primary eluent was vented to waste and small bundles transferred to the second column at a high frequency (2 Hz). This method of valve-based modulation was successful and, although not truly comprehensive, as most of the eluent is vented, can be considered as such. The second valve-based GC × GC system was introduced by Seeley et al. (2000) and incorporated a six-port diaphragm valve with sample loop, housed within a secondary oven, as the modulation device. In the sample position, primary eluent ﬂowed through the sample loop, while a secondary carrier gas ﬂow is simultaneously used to run the second column chromatography. In the load position, the secondary gas supply pushes the loop contents on the second column, and the primary eluent is vented. Seeley et al. used a high secondary column ﬂow rate relative to the ﬁrst (at least 20 times higher) to ensure both maximum transfer of analytes between columns and minimum injected plug thickness. The valve was in the sample position for 0.8 s and in the launch position for 0.2 s, and, thus, approximately 80% of the sample was transferred. Valves have found only very limited use in GC×GC, but have some signiﬁcant advantages over thermal methods for atmospheric monitoring (Hamilton et al. 2003a). Valve modulators are independent of temperature, and, when the valve is housed within the GC oven, analytes are transferred from column one to column two at the ambient oven temperature. This is of particular importance when very high volatility or polar compounds are being investigated, such as those found in gas-phase atmospheric samples. In cryogenic modulation, breakthrough can occur in the high-volatility region of the analysis as the temperature is not sufﬁciently low to trap analytes and, in general, species with less than ﬁve carbons are not effectively trapped. The valve, however, suffers no such breakthrough problems and is independent of analyte volatility. In addition, the valve requires no cryogen, making it much more suited to ﬁeld studies than cryogenic modulation. The main disadvantage of the valve system is the loss of sensitivity compared to thermal modulation methods, when venting most of the sample to waste. However, a recent study has shown that, provided the ﬁrst-dimension peak is adequately sampled, valve modulation retains the peak height of the one-dimensional peak and, in some cases, may provide some degree of peak amplitude enhancement (Hamilton et al. 2003a).

11.4

Detectors

In GC × GC, the peaks produced during the secondary separation are inherently narrow, in the order of tens to hundreds of milliseconds. To properly deﬁne a peak, at least ten data points should be collected across the peak, with between 20 and 50 being optimum. Therefore, the detection system must be capable of collecting data in the tens to hundred hertz range. The most common detector used in GC × GC is the FID, which is ideal for hydrocarbon analysis (Phillips and Beens 1999), giving a universal response for hydrocarbons. The disadvantage of the FID is the lack of structural information given, although this is partially overcome by the structured nature of the chromatograms. Electron capture detectors have also been used in GC × GC for the selective detection of polychloro-biphenyl compounds (PCBs) in environmental matrices, but have yet to be used in atmospheric samples (Korytar et al. 2002).

Comprehensive Two-Dimensional Gas Chromatography

475

The coupling of GC×GC to a mass spectrometer forms an extremely powerful technique, combining the improved resolution of the GC×GC with mass spectral information obtained across the peak. Early attempts used a quadrupole mass spectrometer (Frysinger and Gaines 1999), but the mass spectral acquisition rate was insufﬁcient. A conventional quadrupole usually operates at a scan rate of 1 Hz but can be operated up to about 20 Hz although usually in a single ion mode or with reduced sensitivity. For a 1-s-wide peak, the quadrupole would give very poor peak deﬁnition and produce triangular-shaped peaks. In order to obtain faster acquisition rates than are possible using a quadrupole or ion trap, GC × GC can be coupled to a time-of-ﬂight mass spectrometer (TOFMS). A TOFMS can be operated at up to 500 mass scans per second, each scan being the sum of ten pulses. The TOFMS would be able to fully deﬁne such a peak without a great loss in sensitivity. The mass analyser in a TOFMS is a cylindrical ﬂight tube. Ions are accelerated by a negatively charged plate and achieve the same kinetic energy. The ions present will be of variable mass, and, thus, the speed imparted to them will also vary. The ions travel down the ﬂight tube at different speeds and are separated according to their m/z ratio. Although coupling of GC×GC to TOFMS has been reported, this is very much a research lab tool rather than ﬁeldwork tool at present. The fast mass spectral acquisition rates of the TOFMS allow adequate peak coverage across the narrow peaks inherent to GC × GC (van Deursen et al. 2000). One of the most interesting new applications of GC×GC-TOFMS is the analysis of organic aerosol (OA) content. GC × GC-TOFMS has been used to separate over 10 000 individual analytes in an urban PM10 sample, a resolution not possible using any current methodologies (Hamilton et al. 2004; Welthagen et al. 2003).

11.5

Examples of GC × GC use in atmospheric samples

In the early days of GC × GC, the primary uses were in the analysis of complex oil samples, including both petrochemical and essentials oils. The high resolving power and structured nature of GC × GC allowed previously impossible separation of complex samples in a short single analysis. In 2000, GC × GC was used for the ﬁrst time in atmospheric analysis (Lewis et al. 2000). Lewis et al. used a Peltier-cooled multi-bed absorbent trap to pre-concentrate hydrocarbon species in the urban atmosphere of Melbourne, Australia. The trapped analytes were desorbed at high temperature into a GC × GC–FID system, with LMCS cryogenic modulation. The primary column was a 100% dimethyl polysiloxane column and the second a 50% phenyl dimethyl polysiloxane column, to provide a volatility- and polarity-based separation. Figure 11.5 shows the GC×GC chromatogram obtained, alongside a concurrent one-dimensional chromatogram. It is clear that there are signiﬁcantly more peaks isolated in the GC × GC chromatogram, and, on close inspection, more than 110 peaks are visible, compared to only 20–30 using GC. Large numbers of peaks exist in the higher-boilingpoint ranges above C6 , which are isolated in the GC × GC chromatogram only. In the latter regions of the single column separation, no single component is present at concentrations sufﬁcient to be raised above the baseline. This baseline is, in fact, not a true baseline, but the constant co-elutions of many analytes at low concentrations. The impact of this ‘missing’ carbon in the atmosphere, which was previously undetected, was investigated further by Hamilton and Lewis (2003), with particular emphasis on the monoaromatic fraction. Aromatic hydrocarbons are released into the atmosphere

476

Volatile Organic Compounds in the Atmosphere

1

2

3

4

40 H

G

30

F 25 E 20 D 15 C B

Retention time on BP-1 column (min)

35

10

A 5

30 20 10 Flame ionization detector response (pA)

1 2 3 4 Retention time on BP-50 columns (s)

Figure 11.5 Comprehensive and one-dimensional separations of VOCs in urban air. A, benzene; B, heptane; C, toluene; D, xylenes; E, C3 -benzenes; F, C4 -benzenes; G, C5 -benzenes; H, naphthalene; 1, aliphatic band; 2, carbonyl band; 3, monoaromatic band; 4, bi-aromatic band. Reproduced with permission of Nature Publishing Group, from: Lewis et al. (2000). A larger pool of ozone-forming carbon compounds in urban atmospheres. Nature, 405: 778–81. This image appears in full colour in the plate section which follows page 268 as Plate 8.

through automotive emissions and industrial and domestic solvent use and have been estimated to represent as much as 20% of anthropogenic non-methane hydrocarbons emissions. The ozone production potential for multi-substituted aromatics is high, increasing with the degree of ring substitution. A recent model calculation predicts that simple aromatics are responsible for about 40% of ozone formation in Northern Europe (Derwent et al. 1996). Previous studies of VOC composition in the atmosphere using GC–MS and GC–FID have provided detailed inventories, and typically monoaromatics with up to three carbons substituents are routinely reported. More detailed inventories have been reported by Grosjean et al. (1998), Sin et al. (2001) and Yassaa et al. (2001) (143 VOCs 26 monoaromatics, 150 VOCs 22 monoaromatics and 190 VOCs 17 monoaromatics respectively) using GC-quadrupole MS. All of these studies can be considered state-ofthe-art, and covered a wide range of not only hydrocarbon but also halocarbon and other

Comprehensive Two-Dimensional Gas Chromatography

477

gases. The upper limit in these analyses for reporting speciated alkyl substitution has been between 3 and 11 C4 isomers (all three studies), 1–2 C5 isomers (Grosjean et al. and Sin et al.) and 1 C6 isomer by Sin et al. The GC × GC chromatogram obtained in Figure 11.5, indicated that these onedimensional GC studies would isolate only a small fraction of the monoaromatic fraction of the summer time urban atmosphere of Melbourne. Hamilton and Lewis (2003) used a cryogenic and a valve modulator to investigate the monoaromatic content of both atmospheric samples and gasoline vapours, the major emission source in urban areas. Analysis of air collected at a roadside location in the United Kingdom identiﬁed a total of 147 monoaromatic species, and a sample chromatogram is shown in Figure 11.6. The enhanced resolution is very clear, and, in the higher boiling regions that are successively expanded at higher gain, in Figure 11.6, there are many hundreds of components isolated.

400 300 200 100 0

0

10

20

30

40

50

60

70

80

90

100

Second column RT/s

3

1 2 3

5 6

A

B

C

D 4

8 9

H

11

Second column RT/s

5

5

10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 First retention time/min

3 4 5D

E

G

F

7 8 9 50

55

61 66 72 Retention time/min

77

82

66

75

84

Figure 11.6 Comparison of single column (upper) and GC × GC separations (lower) of a Leeds urban air sample. Areas of the full chromatogram are successively extracted at higher gain to illustrate increasing isomeric complexity at higher boiling points. GC × GC chromatograms are annotated with start of individual Cx isomer band (running right to left) where A = C2 , B = C3 , C = C4 , D = C5 , E = C6 , F = C7 , G = C8 , H = naphthalene. Chemical banding assignments: 1, aliphatic; 2, oleﬁns; 3, oxygenates; 4, monoaromatics; and 5, polyaromatics. Reproduced with permission of Elsevier, from: Hamilton et al. (2003). Mono aromatic complexity Atmospheric Environment, 37(5): 589–602. This image appears in full colour in the plate section which follows page 268 as Plate 9.

A

B

A

B

Increasing polarity

Volatile Organic Compounds in the Atmosphere

C

Increasing polarity

478

C

Increasing boiling point Figure 11.7 Comparison of GC × GC chromatograms of gasoline vapours and urban air. Upper: Leeds urban air chromatogram. Lower: Gasoline vapours at 20◦ C chromatogram. A, C3 -substituted monoaromatics; B, C4 -substituted monoaromatics; C, C5 -substituted monoaromatics.Reproduced with permission of Elsevier, from: Hamilton et al. (2003) Mono aromatic complexity. Atmospheric Environment, 37(5): 589–602. This image appears in full colour in the plate section which follows page 268 as Plate 10.

An advantage of the 3D contour plot is that they are much more amenable to visual interpretation than the analogous two-dimensional versions. Pattern matching or source ﬁngerprinting using this technique is greatly enhanced, since it is analyte chemical characteristics rather than sheer abundance that drive visual identiﬁcations. Figure 11.7 illustrates this concept, showing a comparison of the GC × GC chromatograms obtained for 1 ml gasoline vapours at 20◦ C and 1 l urban air. Even though different modulation devices have been used (valve and LMCS, respectively), an identical set of separation conditions were employed, and as such the two chromatograms are directly comparable. There are obviously some differences between the two chromatograms, both in terms of content and peak shape. The air chromatogram has additional carbonyls and oleﬁns visible, and the peak shape appears poorer due to working nearer the detection limit of the instrument. In some aromatic regions, the two chromatograms are remarkably similar, with almost identical distributions of C3 –C5 aromatics compounds. Although this is only a qualitative visual matching between source and receptor, it is clear that at this location urban air has a signiﬁcant input from evaporative/unburnt fuel sources. Overall, the air chromatogram is more complex than the gasoline vapour, particularly in terms of the number of compounds in the slightly polar or oxygenated regions. While the C3 aromatics and C4 monoaromatics are clearly identiﬁable in both gasoline and air chromatograms and are present in similar relative abundances (when normalised to toluene), as substitution increases, the relative proportion of C5 and greater aromatics in air diminishes signiﬁcantly when compared to fuel vapour.

Comprehensive Two-Dimensional Gas Chromatography

479

It was calculated that the effect of these low-concentration aromatics (i.e. <5 pptv/isomer) combined could contribute to a signiﬁcant ozone forming potential. For example, the concentration of the sum of C4 -substituted aromatics found in Leeds urban air on this day was calculated to produce a theoretical production rate of 1.1 ppbv O3 /h, similar in magnitude to production from highly reactive species such as isobutene. The use of GC × GC in atmospheric VOC analysis was extended to ﬁeld measurements by Xu et al. (2003b). The in situ measurements were carried out during the Mediterranean Intensive Oxidant Study (MINOS) project, based in Finokalia, Crete, in August 2001. Air samples were collected using an online absorbent at 10◦ C (to avoid trapping of water at lower temperatures) at a ﬂow rate of 50 ml/min for 60–80 min, giving sample sizes of 3 or 4 l. GC × GC analysis used a dual-jet modulator, where both hot and cold jets were nitrogen gas from a Dewar, the cold jet gas being cooled by passing it through copper tubing coiled inside a cryogenic trap (LN2 ) and the hot jets heated by a heater at the tube outlet. A DB-5 (5% phenyl dimethyl polysiloxane) column and a highly polar Carbowax (polyethylene glycol) column were used as the ﬁrst and second columns, respectively. Peak identiﬁcations were achieved using GC × GC-TOFMS of absorbent tube samples back in the laboratory and calibrations were achieved using standard gas mixtures (Xu et al. 2003a). About 650 individual peaks were identiﬁed, and, using published retention indices, the identity of 235 analytes could be conﬁrmed. A wide range of functionalities were found, including alkanes, alkenes, aromatics, alcohols, aldehydes, ketones, esters, nitriles and halogenated hydrocarbons. Measurements of anthropogenic VOCs covered the period of 2–21 August 2001, at a resolution of about 2 h. Calibrations were performed every ﬁve days and blanks every three days. The Finokalia station is a remote site but can have signiﬁcant anthropogenic input through transportation of photo-chemically aged air, such as from the Greater Athens area. Mixing ratios of hydrocarbons were in the pptv to sub-ppbv range, with signiﬁcant day-to-day and diurnal variations. Figure 11.8, shows the time proﬁle for the sum of n-alkanes and aromatics during MINOS. Results from analysis using a simpliﬁed box model indicate that the variability in observed hydrocarbons was dependent on both emissions source characteristics and strength and the degree of chemical processing (i.e. via reaction with OH). One of the main advantages of GC×GC in the MINOS study was the ability to isolate lowconcentration species from the complex background, which would have been hidden by coelution with high-concentration analytes using one-dimensional GC. This key characteristic of GC × GC was vital during the investigation of the reaction products of alkyl-benzenes in a smog chamber (Hamilton et al. 2003b). Smog chamber studies can provide invaluable data regarding a compound’s reactivity and reaction mechanism. As shown previously, aromatic hydrocarbons are important ozone precursors and their reaction products can form secondary OAs. The complex reaction mechanisms of monoaromatics are currently poorly understood, even though they have been extensively studied for many years. Toluene is one of the most abundant anthropogenic hydrocarbons released to the environment and its photo-oxidation was studied during the EXACT project (Effects of the oX idation of Aromatic Compounds in the T roposphere) in the summers of 2001 and 2002. A series of reactions were carried out at the European Photoreactor (EUPHORE), a large-volume outdoor reaction chamber in Valencia, Spain. A GC × GC system was developed, which utilises valve modulation and independent separations as a function of both volatility and polarity (Hamilton et al. 2003b). The gas-phase reaction products of toluene are both volatile and highly polar, and so the valve provides the only feasible modulation device.

480

Volatile Organic Compounds in the Atmosphere

100

P2

700

P4

60

600 500 400 300

40

200 20 100 0 06.08.2001

11.08.2001 16.08.2001 Time (UTC)

Aromatics and acetonitrile (pptv)

ΣC8– C11 n-alkanes ΣC7– C10 Aromatics Acetonitrile

80 Alkanes (pptv)

P3

0 21.08.2001

−C11 n-alkanes and C7 − −C10 aromatics during the MINOS camFigure 11.8 Summed mixing ratios of C8 − paign. For the purpose of comparison average mixing ratios of acetonitrile from the proton transfer reaction mass spectrometry (PTRMS) measurements (Salisbury et al. 2003) are also shown. Due to interferences, 1,2,4-trimethylbenzene is not included in the mixing ratio of aromatics. Reproduced with permission of the European Geosciences Union (EGU), from Xu et al. (2003b) GC × GC measurements of C7 − C11 aromatic and n-alkane hydrocarbons on Crete, in air from Eastern Europe during the MINOS campaign. Atmospheric Chemistry and Physics, 3: 1461–75.

Two-litre chamber air samples were cryofocused, with a sampling frequency of 30 min, allowing the evolution of species to be followed over oxidation periods of 3–6 h. Figure 11.9 shows a comparison of a typical standard GC × GC chromatogram and a chamber sample. A number of carbonyl reaction products were identiﬁed, and the time proﬁle for one of the toluene photo-oxidation reactions is shown in Figure 11.10. It was possible to separate a number of trace-level carbonyl species (i.e. <5 ppbv) from the extremely large toluene peak that had a starting concentration of around 500 ppbv. Using one-dimensional chromatography, these trace components peaks would be hidden under the toluene peak and difﬁcult to identify without some sort of sample preparation, such as derivatisation. Comparison of FTIR and GC × GC for the measurement of the parent aromatics generally showed good agreement. Comparison of the concentrations observed by GC × GC to concentration–time proﬁles simulated using the Master Chemical Mechanism, MCMv3, were made. The MCM is a near-explicit chemical mechanism developed to describe, in detail, the tropospheric degradation of emitted VOC, and is suitable for use over a wide range of atmospheric conditions (http://mcm.leeds.ac.uk/MCM). Results obtained during EXACT by GC × GC indicate that this mechanism signiﬁcantly over-predicts the concentrations of many product compounds and highlights the uncertainties that exist in our understanding of the atmospheric oxidation of aromatics. Aerosols play an important role in the Earth’s atmosphere by scattering or absorbing solar radiation (Pilinis et al. 1995; Twomey 1991), acting as cloud condensation nuclei (Novakov and Corrigan 1996; Novakov and Penner 1993) and are associated with damaging effects on human health (Dockery et al. 1993; Samet et al. 2000; Wichmann and Peters 2000).

Second retention time (s)

Comprehensive Two-Dimensional Gas Chromatography

0.75 1.00

1

1.25 1.50

6

4

2

7

12 9

10

13

11

15

3

1.75

14

8

2.00

5

2.25 7.5

Second retention time (s)

481

1.25 1.50 1.75 2.00 2.25 2.50 2.75 3.00 3.25 3.50 3.75

10.0

12.5

15.0 17.5 20.0 First retention time (min)

7

22.5

25.0

22.5

25.0

13 12 16

Toluene 7.5

10.0

12.5

15.0 17.5 20.0 First retention time (min)

Figure 11.9 Chromatograms for a typical standard mixture and a chamber sample during the photooxidation of toluene. Peak identiﬁcations are given in Table 11.1. Reproduced with permission of the EGU, from Hamilton et al. (2003b) Measurements of photo-oxidation products from the reaction of a series of alkyl-benzenes with hydroxyl radicals during EXACT using comprehensive gas chromotography. Atmospheric Chemistry and Physics, 3: 1999–2014. This image appears in full colour in the plate section which follows page 268 as Plate 11.

The composition of OA is even more complex than the gas-phase organic content. OA can be both primary (i.e. diesel particles) and secondary (partitioning of oxidised gasphase species). Thus, particles can have both a wide range of organic functionalities and volatilities and are closely linked to gas-phase VOCs. The analysis of this type of sample can prove extremely difﬁcult, even with the large separating power of GC × GC. Coupling to a TOFMS can add an additional separation mechanism, in this case, the m/z ratio. PM2.5 particulate samples were collected daily in Augsburg, Germany (Welthagen et al. 2003) onto quartz ﬁlters at a ﬂow rate of 1 m3 /h. Filters were cut into small pieces and analysed using direct thermal desorption. In a typical sample, more than 15 000 peaks could be detected. Using a mass spectrometer as a detector allowed structural information and peak identiﬁcations to be obtained, which would have required numerous standards if an FID was used. In addition, complex sample preparation steps, such as solvent extractions and derivatisations, which are required to reduce sample complexity when using low-resolution techniques, can be removed. The organic fraction of OA was released using direct thermal desorption, allowing a single-shot inventory of all the GC-amenable species in one run. The data sets obtained in a GC × GC-TOFMS analysis of OA are enormously complex. Welthagen et al. (2003) developed a preliminary classiﬁcation procedure to group sets of compounds and reduce the overall complexity of data analysis. They generated a bubble plot where each peak is represented by a bubble scaled according to the peak area, as shown in

482

Volatile Organic Compounds in the Atmosphere

Table 11.1 Peak identiﬁcations of EXACT standard mixture chromatogram 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

2,3-Epoxy-butanal Z-Butenedial E-Butenedial (3H)-Furan-2-one Maleic anhydride 2-Furaldehyde α-Angelicalactone (5H)-Furan-2-one γ -Angelicalactone β-Angelicalactone Z-hexene-2,5-dione Benzaldehyde p-Methyl-benzoquinone Toluene-oxide 6-Oxo-2,4-heptadienal Citraconic anhydride

Figure 11.11(a). The bubbles were coloured according to a set of predeﬁned rules, to separate them into group classiﬁcations, Figure 11.11(b). This coloured bubble plot provides a visual tool to recognise pattern changes at a glance. An investigation of oxygenated volatile organic compounds (OVOC) contained within urban aerosol has been performed using GC × GC-TOFMS (Hamilton et al. 2004). In a typical sample, around 130 OVOCs were identiﬁed, based on retention behaviour and spectral match. In excess of 100 other oxygenated species were also observed, but lack of mass spectral library or pure components prevented positive identiﬁcation. The advantage of using a TOFMS for detection in GC × GC is the ability to create single ion GC × GC chromatograms, adding an additional level of system dimensionality. Figure 11.12 shows the extraction power of single-ion contour plots, allowing speciﬁc groups, isomers or even, in some cases, single compounds to be extracted out of the complex aerosol background. Many of the carbonyl species observed could be mechanistically linked to gas-phase aromatic hydrocarbon oxidation, and there was good agreement in terms of speciation between the urban samples analysed and those degradation products observed in smog chamber experiments of aromatic oxidation. The presence of partially oxidised species such as linear chain aldehydes and ketones and cyclic products such as furanones suggested that species generated relatively early in the oxidative process may undergo gas-to-particle partitioning despite their relatively high volatility.

11.6

Conclusions

GC×GC is a relatively new technique and has so far had only very limited use in the analysis of VOCs in atmospheric samples. As the need to fully understand the complex composition of our atmosphere increases, high-resolution techniques such as GC × GC will provide an

Comprehensive Two-Dimensional Gas Chromatography

(a)

483

(b) 10 Benzaldehyde

Toluene

600 Concentration (ppb)

Concentration (ppb)

8 6 4 2 0

10

11

400 300 200 Chamber opened 100

Chamber opened 9

GC × GC FTIR

500

12

13

14

15

Time at end of sampling (UTC)

9

10

11

12

13

14

15

16

17

Time at end of sampling (UTC)

(c)

Concentration (ppb)

2.4

Maleic anhydride Angelicalactone Citraconic anhydride Methyl benzoquinone

2.0 1.6 1.2 0.8 0.4 0.0

Chamber opened 9

10 11 12 13 14 Time at end of sampling (UTC)

15

Figure 11.10 Experimentally observed concentration–time proﬁles for parent compound and intermediates in the toluene high NOx experiment (25 September 2001). Error limits on the concentrations of intermediates are ±50%. Toluene, benzaldehyde, angelicalactone and methyl benzoquinone are quantiﬁed by liquid standard calibration, and maleic anhydride and citraconic anhydride peak areas are presented to show their time dependent behaviour. The concentration of the parent compound obtained by GC × GC is compared to FTIR measurements and shows reasonable agreement. Reproduced with permission of the EGU, from Hamilton et al. (2003b) Measurements of photo-oxidation products from the reaction of a series of alkyl-benzenes with hydroxyl radicals during EXACT using comprehensive gas chromotography. Atmospheric Chemistry and Physics, 3: 1999–2014.

important advancement. GC × GC has already identiﬁed that there are considerably more VOCs in the atmosphere than previously indentiﬁed. The combined effect of these compounds can be great, inﬂuencing ozone production, atmospheric reactivity and secondary aerosol formation. It has been used under a wide range of conditions, from lab and smog chamber studies to ﬁeld measurements. GC × GC is still in its infancy, and there are many possibilities for future development. For example, the technology to produce miniaturised GC systems using chemical and photo etching to manufacture micro-fabricated columns onto silica monolith structures has been around for a number of years, but extension into real applications has been limited by detectors. Field measurements of atmospheric VOCs would beneﬁt greatly

484

Volatile Organic Compounds in the Atmosphere

Polarity second dimension (s)

(a)

1060 GC × GC-TOFMS peaks from Augsburg aerosol sample (All peaks shown) 4 3.5 3 2.5 2 1.5 1 1 000

1 500

2 000

2 500

3 000

3 500

Volatility first dimension (s)

Polarity second dimension (s)

(b)

1060 GC × GC-TOFMS peaks from Augsburg aerosol sample coloured (Assigned peaks shown in colour) 4 3.5 3 2.5 2 1.5 1 1 000

1 500

2 000

2 500

3 000

3 500

Volatility first dimension (s) Figure 11.11 (a) shows the bubble plot of the entire peak apices used in the study for grouping. (b) indicates the different groups identiﬁed on (a). Colour assignments: Burgundy, alkanes; Lilac, alkenes and cycloalkanes; Pink, n-alkane acids; Blue, alkyl substituted aromatics; Brown, polar benzenes; Green, hydrated naphthalenes and alkenyl benzenes; Red, naphthalenes and alkylated naphthalenes. Reproduced with permission of Elsevier from Welthagen et al. (2003) Search criteria and rules for comprehensive twodimensional gas chromatography-time-of-ﬂight mass spectrometry analysis of airborne particulate matter. Journal of Chromatography A, 1019: 233–49. This image appears in full colour in the plate section which follows page 268 as Plate 12.

from miniaturised GC systems, where both size and power requirements are substantially reduced. A valuable by-product of comprehensive methods has been the ability to directly introduce heterogeneous samples such as aerosols to the analytical system without sample preparation or clean up. Labour time savings are an obvious beneﬁt, but importantly the reduction in amount of material required for analysis improves both measurement resolution and detection limits. As high-complexity methods reach maturity, however, it is likely that it will be data interpretation that becomes the rate-determining step in the analytical procedure. This may require adopting data-handling technologies associated with image processing to ﬁnd ways in which useful and timely information can be extracted from chemical methods

Comprehensive Two-Dimensional Gas Chromatography

485

Masses: 91

(a)

2nd Dimension time (s) 1 2 3

4

Groups

Alkyl-benzenes

0

Chloro-alkanes

6

506

1006 1506 1st Dimension time (s)

2006

4

Masses: 134

(b)

2nd Dimension time (s) 1 2 3

Isomers

0

C4-mono aromatics 6

506

1006 1506 1st Dimension time (s)

2006

4

Masses: 128

(c)

2nd Dimension time (s) 1 2 3

Single compound

0

Naphthalene 6

506

1006 1506 1st Dimension time (s)

2006

Figure 11.12 Extraction of groups, isomers and single compounds using single ion GC × GC chromatograms. (a) Groups (m/z 91 Da), (b) isomers (m/z 134 Da) and (c) single compound (m/z 128 Da). This image appears in full colour in the plate section which follows page 268 as Plate 13.

with such high levels of detail. The use of chemometric techniques, such as analysis of variance (ANOVA), based feature selection, principle component analysis and parallel factor analysis, will be key to decoding complex GC × GC chromatograms, for both compound identiﬁcation and to determine signiﬁcant differences between samples.

486

Volatile Organic Compounds in the Atmosphere

Further reading Ong, R.C.Y. and Marriott, P.J. (2002) A review of basic concepts in comprehensive two-dimensional gas chromatography. Journal of Chromatographic Science, 40: 276–91. Phillips, J.B. and Beens, J. (1999) Comprehensive two-dimensional gas chromatography: A hyphenated method with strong coupling between the two dimensions. Journal of Chromatography A, 856 (1–2): 331–47. Mondello, L., Lewis, A.C. and Bartle, K.D. (2001) Multidimensional Chromatography. John Wiley and Sons, Chichester.

References Beens, J., Adahchour, M., Vreuls, R.J.J., van Altena, K. and Brinkman, U.A.T. (2001) Simple, nonmoving modulation interface for comprehensive two-dimensional gas chromatography. Journal of Chromatography A, 919: 127–32. Bruckner, C.A., Prazen, B.J. and Synovec, R.E. (1998) Comprehensive two dimensional high-speed gas chromatography with chemometric analysis. Analytical Chemistry, 70: 2796–804. Dalluge, J., van Stee, L.L.P., Xu, X.B., et al. (2002) Unravelling the composition of very complex samples by comprehensive gas chromatography coupled to time-of-ﬂight mass spectrometry – Cigarette smoke. Journal of Chromatography A, 974: 169–84. Davis, J.M. and Giddings, J.C. (1983) Statistical-theory of component overlap in multicomponent chromatograms. Analytical Chemistry, 55: 418–24. Derwent, R.G., Jenkin, M.E., and Saunders, S.M. (1996) Photochemical ozone creation potentials for a large number of reactive hydrocarbons under European conditions. Atmospheric Environment, 30: 181–99. van Deursen, M., Beens, J., Reijenga, J., Lipman, P., Cramers, C. and Blomberg, J. (2000) Grouptype identiﬁcation of oil samples using comprehensive two-dimensional gas chromatography coupled to a time-of-ﬂight mass spectrometer (GC × GC–TOF). Hrc – Journal of High Resolution Chromatography, 23: 507–10. Dockery, D.W., Pope, C.A., Xu, X.P., et al. (1993) An association between air-pollution and mortality in 6 United States cities. New England Journal of Medicine, 329: 1753–9. Frysinger, G.S. and Gaines, R.B. (1999). Comprehensive two-dimensional gas chromatography with mass spectrometric detection (GC × GC/MS) applied to the analysis of petroleum. Hrc – Journal of High Resolution Chromatography, 22: 251–5. Giddings, J.C. (1990) Multidimensional Chromatography. Marcel Dekker, New York. Grosjean, E., Rasmussen, R.A. and Grosjean, D. (1998) Ambient levels of gas phase pollutants in Porto Alegre, Brazil. Atmospheric Environment, 32: 3371–9. Hamilton, J.F. and Lewis, A.C. (2003) Monoaromatic complexity in urban air and gasoline assessed using comprehensive GC and fast GC–TOF/MS. Atmospheric Environment, 37: 589–602. Hamilton, J.F., Lewis, A.C. and Bartle, K.D. (2003a) Peak amplitude and resolution in comprehensive gas chromatography using valve modulation. Journal of Separation Science, 26: 578–84. Hamilton, J.F., Lewis, A.C., Bloss, C., et al. (2003b) Measurements of photo-oxidation products from the reaction of a series of alkyl-benzenes with hydroxyl radicals during EXACT using comprehensive gas chromatography. Atmospheric Chemistry and Physics, 3: 1999–2014. Hamilton, J.F., Webb, P.J., Lewis, A.C., Hopkins, J.R., Smith, S. and Davy, P. (2004) Partially oxidised organic components in urban aerosol using GC × GC–TOF/MS. Atmospheric Chemistry and Physics, 4: 1279–90.

Comprehensive Two-Dimensional Gas Chromatography

487

Kinghorn, R.M., Marriott, P.J. and Dawes, P.A. (2000) Design and implementation of comprehensive gas chromatography with cryogenic modulation. Hrc – Journal of High Resolution Chromatography, 23: 245–52. Korytar, P., Leonards, P.E.G., de Boer, J. and Brinkman, U.A.T. (2002) High-resolution separation of polychlorinated biphenyls by comprehensive two-dimensional gas chromatography. Journal of Chromatography A, 958: 203–18. Ledford, E.B. and Billesbach, C. (2000) Jet-cooled thermal modulator for comprehensive multidimensional gas chromatography. Hrc – Journal of High Resolution Chromatography, 23: 202–4. Ledford, E.B., Phillips, J.B., Xu, J.Z., Gaines, R.B. and Blomberg, J. (1996) Ordered chromatograms: A powerful methodology in gas chromatography. American Laboratory, 28: 22–23. Lee, A.L., Bartle, K.D. and Lewis, A.C. (2001) A model of peak amplitude enhancement in orthogonal two-dimensional gas chromatography. Analytical Chemistry, 73: 1330–5. Lewis, A.C., Carslaw, N., Marriott, P.J., et al. (2000) A larger pool of ozone-forming carbon compounds in urban atmospheres. Nature, 405: 778–81. Liu, Z.Y. and Phillips, J.B. (1991) Comprehensive 2-dimensional gas-chromatography using an on-column thermal modulator interface. Journal of Chromatographic Science, 29: 227–31. Liu, Z.Y., Patterson, D.G. and Lee, M.L. (1995) Geometric approach to factor-analysis for the estimation of orthogonality and practical peak-capacity in comprehensive 2-dimensional separations. Analytical Chemistry, 67: 3840–5. Marriott, P.J. and Kinghorn, R.M. (1997) Longitudinally modulated cryogenic system. A generally applicable approach to solute trapping and mobilization in gas chromatography. Analytical Chemistry, 69: 2582–8. Novakov, T. and Corrigan, C.E. (1996) Cloud condensation nucleus activity of the organic component of biomass smoke particles. Geophysical Research Letters, 23: 2141–4. Novakov, T. and Penner, J.E. (1993) Large contribution of organic aerosols to cloud-condensationnuclei concentrations. Nature, 365: 823–26. Opiteck, G.J., Ramirez, S.M., Jorgenson, J.W. and Moseley, M.A. (1998) Comprehensive twodimensional high-performance liquid chromatography for the isolation of overexpressed proteins and proteome mapping. Analytical Biochemistry, 258: 349–61. Phillips, J.B. and Beens, J. (1999) Comprehensive two-dimensional gas chromatography: A hyphenated method with strong coupling between the two dimensions. Journal of Chromatography A, 856: 331–47. Phillips, J.B. and Xu, J.Z. (1995) Comprehensive multidimensional gas-chromatography. Journal of Chromatography A, 703: 327–34. Pilinis, C., Pandis, S.N. and Seinfeld, J.H. (1995) Sensitivity of direct climate forcing by atmospheric aerosols to Aerosol-size and composition. Journal of Geophysical Research-Atmospheres, 100: 18739–54. Salisbury, G., Williams, J., Holzinger, R., et al. (2003) Ground-based PTR-MS measurement of reactive organic compounds during the MINOS campaign in Crete, July-August 2001, Atmospheric Chemistry and Physics, 3: 925–40. Samet, J.M., Dominici, F., Curriero, F.C., Coursac, I. and Zeger, S.L. (2000) Fine particulate air pollution and mortality in 20 US cities, 1987–1994. New England Journal of Medicine, 343: 1742–9. Seeley, J.V., Kramp, F. and Hicks, C.J. (2000) Comprehensive two-dimensional gas chromatography via differential ﬂow modulation. Analytical Chemistry, 72: 4346–52. Sin, D.W.M., Wong, Y.C., Sham, W.C. and Wang, D. (2001) Development of an analytical technique −C12 volatile organic compounds in Summa((R)) canisters by and stability evaluation of 143 C3 − gas chromatography-mass spectrometry. Analyst, 126: 310–21. Twomey, S. (1991) Aerosols, clouds and radiation. Atmospheric Environment Part A–General Topics, 25: 2435–42.

488

Volatile Organic Compounds in the Atmosphere

Welthagen, W., Schnelle-Kreis, J. and Zimmermann, R. (2003) Search criteria and rules for comprehensive two-dimensional gas chromatography-time-of-ﬂight mass spectrometry analysis of airborne particulate matter. Journal of Chromatography A, 1019: 233–49. Wichmann, H.E. and Peters, A. (2000) Epidemiological evidence of the effects of ultraﬁne particle exposure. Philosophical Transactions of the Royal Society of London Series A–Mathematical Physical and Engineering Sciences, 358: 2751–68. Xu, X., van Stee, L.L.P., Williams, J., et al. (2003a) Comprehensive two-dimensional gas chromatography (GC × GC) measurements of volatile organic compounds in the atmosphere. Atmospheric Chemistry and Physics, 3: 665–82. −C11 aromatic Xu, X., Williams, J., Plass-Dulmer, C., et al. (2003b) GC × GC measurements of C7 − and n-alkane hydrocarbons on Crete, in air from Eastern Europe during the MINOS campaign. Atmospheric Chemistry and Physics, 3: 1461–75. Yassaa, N., Meklati, B.Y., Brancaleoni, E., Frattoni, M. and Ciccioli, P. (2001) Polar and non-polar volatile organic compounds (VOCs) in urban Algiers and saharian sites of Algeria. Atmospheric Environment, 35: 787–801.

Volatile Organic Compounds in the Atmosphere Edited by Ralf Koppmann Copyright © 2007 by Blackwell Publishing Ltd

Index

accumulation mode, 344 accuracy, 157 of isotope ratio measurements, 412, 418, 442–3, 458 acetaldehyde, 58, 85, 89, 93, 95, 107, 130, 133, 141–2, 146, 147, 148, 250 compensation point, 153 acetal formation, 14, 356 acetic acid, 137, 141–2 acetone, 12–13, 85, 89, 93, 96, 129, 136–7, 138, 142, 146, 147, 151, 155, 230, 305, 441, 442 global air-to-sea ﬂux, 155 global budget, 155 global distribution, 103, 105, 107 global emission rate, 138 global oceanic uptake, 147 global source, 155 global source strength, 138 acetophenone, 330 acetyl coenzyme A, 141 acid catalysed reactions, 14 acids, 42, 234, 371 adduct stability, 438 adsorbent, 63, 64, 159, 209, 330 AED, see atomic emission detector aerodynamic diameter, 367 aerosol, 5, 15 aerosol mass, 14, 345 aerosol mass spectrometer, 373, 374 aerosol source, 353 AGAGE, 69, 188, 208 age of air mass, 59, 323 agricultural plants, 3, 6, 88, 89, 139, 140, 301, 302, 303 aircraft measurements, 8, 133, 136, 244, 247, 323 Aitken mode, 344 alcohol, 15, 58, 65, 133–6, 139, 152 alcohol concentrations, 360 alcohol dehydrogenase, 141

alcohol fuel program, 138 alcoholic fermentation, 141 aldehyde concentration, 360 aldehydes, 57, 58, 131, 133 aldol condensation, 14, 356 aldol reaction, 356 algae, 184, 346 aliphatic hydrocarbons, 159 alkanes, 5, 6, 10, 33, 42, 56, 283–4, 349, 354, 422–4, 428–30, 433, 434–6, 439, 451, 453, 455 alkanoic acids, 306, 308, 324, 329, 360 alkanols, 354 alkyl nitrates, 109, 270, 273, 274, 276–80, 283, 284, 285 Alloxysta vitrix, 302 Amazonian rain forest, 16, 97, 137, 153, 346–7 ambient measurements, 249–55, 275, 276, 280, 451 ammonia, 68, 157 AMS, see aerosol mass spectrometer analysis of organic aerosols, 365–75 anhydro-sugar, 354 animal- and plant-type biological detectors, 9 animal waste, 138 anion exchange chromatography, 372 annual global production of SOA, 356 Antarctic ﬁrn air, 183 anthropogenic carbon, 352–3 anthropogenic contribution, 3, 4, 185, 352–3 anthropogenic emissions, 5, 36, 43, 54–5, 138–9 emission inventory, 11, 17, 34–7, 97–103, 154–5, 204 anthropogenic NMHC sources, 11, 33, 54, 455 anthropogenic SOA, 357–9 anthropogenic sources, 33–4, 37–42, 137–8, 244, 321 anthropogenic VOCs, 33, 138 APAN, 223, 244

490

Index

APCI, see atmospheric Pressure Chemical Ionisation aphids, 302 arenes, 304 aromatic compounds, 144, 438, 439, 455 aromatic hydrocarbons, 57, 357–8, 475–6, 479 aromatic VOCs, 43, 49, 53, 57, 60, 362 Arrhenius parameters for the thermal decomposition of PANs, 232 artefacts, 91, 293, 369, 370–71 Artemisia californica, 302 asphalt blowing, 3 atmospheric lifetimes, 6, 55, 59, 103, 151, 175, 183, 184, 185, 192 atmospheric loss, 231–3, 396 atmospheric NMHC loss, 451 atmospheric oxidation of VOC, 57, 446, 457 atmospheric pressure chemical ionisation, 372, 375 atmospheric pressure mass spectrometry, 331 atmospheric removal, 6, 285, 447, 457 atmospheric residence time, 395 atomic emission detector, 9, 67 automobile exhaust, 4, 39, 42, 43, 137, 296–8, 304, 306, 358 average photochemical age, 398, 399 bacteria, 186, 346, 347 bacterial degradation, 449 bacterial metabolism, 146 basal emission factors, 100, 144 basal emission rate, 95, 98, 100–101 Battelle impactor, 368 benzaldehyde, 300, 319 benzene, 39, 43, 47, 51, 304, 390, 393 benzoic acid, 306, 359 benzylhydroxyl oximes, 332 Berner impactor, 367 bimolecular reaction rate constant, 421 bioaerosols, 346–7 biofuel combustion, 34, 41, 43 biogenic contribution to organic aerosol, 353 biogenic SOA, 113, 356–7 biogenic SOA markers, 363 biogenic sources, 5, 300–302 biological uptake, 6 biomarkers for bacteria, 347 biomass burning, 41–2, 138–9, 248, 255, 294–6, 353, 354 black carbon, 348 black carbon aerosol, 19 blue haze, 14, 342, 343 boreal forest, 363 box model, 191, 202, 207, 208, 479

branching ratio, 109, 271, 272, 273–4 break-through, 473, 474 break-through volume, 207 broad leafed trees, 363 bromine compounds, 173–4, 184–6, 196, 198–9 bromine explosion, 198, 199 bromoform, 186 bromomethane, 413, 419, 449 bubble bursting, 346, 355 bulk yield method, 356 2,3-butanedione, 300 butanone, 300 2-butanone, 143 calibration, 67, 68–70, 237–43 calibration methods, 223, 242, 256, 333 C5 - and C6 -alcohols, 136 capillary electrophoresis, 372 capillary GC, 60, 66, 239 capillary GC with electron capture detection, 238–9, 240, 275 carbon bond mechanism, 11 carbon dioxide traps, 413 carbon isotope ratios, 388, 393, 396, 403, 404, 406, 407, 409, 411–20, 442, 446, 448–9, 455, 456, 457 carbon KIE, 398, 423, 434–9 carbon tetrachloride, 183 carbonyl, 309–23, 328–9, 406 carbonyl compounds, 147, 299–300, 302, 303, 327, 329–32, 441 carbonyl compounds from gas-phase oxidation, 303–6, 307 carbonyl nitrates, 284, 285 carbonyls from ozone-surface reactions, 308–9 Carbosieve S-III, 159, 330 Carboxen 569, 159 carboxylic acids, 137, 148, 152, 153, 323–4, 332–3 carboxylic acids from gas-phase oxidation, 306, 308 carboxylic anhydride, 223 3-carene, 306, 363 caric acid, 363, 365 caronic acid, 363, 365 β-caryophyllene, 84, 85, 86, 281 cascade impactor, 367 catalytic converter systems, 38, 39, 51 catalytic cracking, 3 C4 –C11 carbonyls, 293, 294, 298, 309, 319, 321, 327 cell debris, 346 cell degradation, 140 CFC, see chloroﬂuorocarbon

Index

CFC-113, 8, 180, 449 CH4 , see methane charcoal combustion, 294–6 chemical analysis of organic aerosol constituents, 365–7 chemical artefacts, 370 chemical composition, 194, 344, 345, 373 chemical drying, 158 chemical ionisation mass spectrometry, 10, 158, 160, 237, 238, 241 chemical lifetimes, 7–8 chemical loss, 86, 150–52 chemical mass balance, 44 chemical processing, 55, 359, 396 chemical structure, 55, 103, 389 chemical transport models, 16, 45, 252, 254 chemiluminescence, 116, 242 chlorine, 13, 179, 182–4, 198–9, 203–4 chlorine chemistry, 198 chloroﬂuorocarbons, 8, 180, 193–7, 448–51 chloroform, 182–3 chloromethane, 403–4, 420, 450–51 cholesterol, 354 chromatographic techniques, 61 cigarette smoking, 299–300, 354 CIMS, see chemical ionisation mass spectrometry 1,8-cineole, 144 citrus plant species, 301 climate, 2, 93, 176 climate effects, 13–14 cloud condensation nuclei, 15, 199, 329, 346, 480 Cl-radicals, 55, 56, 59, 174 CO2 , 13, 17, 87, 96, 176, 472 coagulation, 14, 344 coal chemistry, 40–41, 138 coal production, 3 coarse particles, 343, 344 coffee-roasting, 138 coking, 3 cold start emissions, 39 cold trap, 159 combustion of fossil fuels, 3, 33–4, 43, 321 combustion of plant material, 354, 361 combustion of synthetic materials, 138 compensation point, 6, 153 complete evaporation, 455 composition of aerosol particles, 345, 359 composition of primary organic aerosols, 353–5 compound-speciﬁc isotope ratio measurement, 406 comprehensive chromatography, 467 concentration of organic carbon, 346

491

concentration of primary and secondary organic aerosol constituents, 359–65 conifers, 91–2 cooking, 298–9, 329, 354, 355 counting statistics, 412 Criegee biradicals, 109, 148 Criegee intermediates, 57, 306, 309 crotonaldehyde, 319 crude oil production, 3 crustal material, 345 cryogenic preconcentration, 158 cyclohexanone, 300, 319 cyclone, 368–9 dead plant material, 146 decanal, 140, 144, 301 decarboxylation, 142 degradation in soils, 192, 449, 451 degradation of lignin, 140 denuders coated with o-benzylhydroxyl ammonium chloride, 332 deposition velocities, 153, 236–7 derivatisation, 65, 371, 405–6 detectors biological detector, 9 electron capture detector, 67, 173, 211, 238 ﬂame ionisation detector, 67, 158, 371, 467 deuterated methanes, 427–8 dicarboxylic acids, 137, 347, 358, 459 dienes, 113 differential optical absorption spectroscopy, 157 diffusion, 8, 63, 333, 369 dimethyl allyl diphosphate, 87, 143 2,4-dimethylphenylhydrazine, 157 dimethyl polysiloxane, 469, 475 dimethyl sulphide, 5, 13 dinitrates, 284, 285 2,4-dinitrophenylhydrazine, 157, 330, 331, 405–6 direct effect, 14, 342 direct thermal desorption, 481 diterpenoids, 354 diurnal cycle, 47, 48, 105, 107 DMAPP, see dimethyl allyl diphosphate DMS, see dimethyl sulphide DNPH, see 2,4-dinitrophenylhydrazine DNPH-coated cartridges, 331, 332 DNPH-derivatives, 331 DNPH on Tenax, 331 DOAS, see Differential Optical Absorption Spectroscopy dry deposition, 5, 152–4, 234–7, 285, 344 dual-jet modulator, 479

492

Index

dual-stage jet modulator, 472 dust, 353, 361 dynamometer test measurements, 38, 39 Earth’s radiation budget, 342 EC, see elemental carbon EC concentration, 350–52 ECD, see Electron Capture Detector eddy covariance, 116 EDGAR, see Emission Database for Global Atmospheric Research EECl, see effective equivalent chlorine EESC, see effective equivalent stratospheric chlorine effective equivalent chlorine, 203 effective equivalent stratospheric chlorine, 204 effective quantum yields of high molecular weight carbonyls, 326 effects of lightning, 254 electron capture detector, 67, 173, 211, 238 electron-spray Fourier-transform ion-cyclotron-mass-spectrometry, 328 electrospray ionisation, 372 electrostatic attractions, 370 elemental carbon, 348 EMEP stations, 321 Emission Database for Global Atmospheric Research, 11, 34, 41, 154 emission estimates, 36, 40, 43–5, 206 emission factors, 100, 115, 144 emission inventories, 11, 17, 34–7, 97–103, 154–5, 204 emission ratios, 18, 139, 206 emissions from agricultural plants, 85, 135, 136, 139, 140, 142, 144, 152–3 anthropogenic, 5, 36, 43, 54–5, 138–9 biogenic, 5, 17, 85–6, 103, 140 from forests, 42, 97, 105, 139, 319 from gasoline passenger cars, 3, 4 from grassland, 42, 139, 300 from trees, 91–2, 96, 139, 142, 363 enclosure measurements, 115 engraver insect, 302 enrichment, 201, 207, 371, 420 ENVISAT, 133 epsilon notation, 391 ESI, see electrospray ionisation ESI-FTCIP, see electron-spray Fourier-transform ion-cyclotron-mass-spectrometry esters, 41, 86, 139 ETBE, see ethyl-tert -butyl ether ethane, 42, 54, 56, 253–4, 439, 455 ethanol, 41, 52, 65, 89, 96, 136, 141–2, 146

ethylbenzene, 304, 433 ethyl-tert -butyl ether, 138 ethyne, 43, 54 eucalyptol, see 1,8-cineole EUPHORE, see European Photoreactor European Photoreactor, 479 evaporation of VOC, 38, 42, 455 evaporative emissions, 3 exchange times, 7 exhaust emissions, 38, 39, 296–8, 306, 320 extraction, 62, 367, 371, 482

fatty acids, 347 ﬁbre ﬁlters, 367, 369 FID, see ﬂame ionisation detector ﬁlters, 18, 212, 369–70 ﬁlter samples, 361, 367, 369 ﬁne particles, 343 ﬂame ionisation detection, 10, 467 ﬂame ionisation detector, 158, 371 ﬂash chromatography, 371 ﬂux measurements, 115 foliar density, 98, 100 food cooking, 298–9, 329 formaldehyde, 16, 58, 85, 140–41, 147, 148, 153, 155, 157–8, 436, 457 global distribution, 133 latitudinal distribution, 133 seasonal distribution, 133 formation of atmospheric particles, 343 formic acid, 140–41, 456–7 formyl radical, 229, 324 fossil fuel, 3, 42, 43, 352 fossil fuel burning, 353, 354 fossil fuel use, 34, 37–41, 54, 360 Fourier Transform Infrared Spectroscopy, 139, 157, 213–14, 273–4, 406 free fatty acids, 143 free troposphere, 12, 129, 133, 174–5 freezing of water, 158 fructose, 361 FTIR, see Fourier Infrared Transform Spectroscopy fuel additives, 43, 138 fugitive emissions, 3, 39, 41 fullerenes, 348 functional group analysis, 372–3 fundamental vibration frequencies, 421 fungal spores, 361 fungi, 346, 347 fulvic acids, 356 furaldehydes, 294 furanones, 304, 482

Index

gas chromatography, 158, 210–11, 238, 270, 371–2 gas chromatography – atomic emission detector, 9, 67, 158 gas chromatography – isotope ratio mass spectrometry, 388 gas chromatography – mass spectrometry, 68, 158, 211–13, 292, 330, 331, 371 gas chromatography – reduction gas detector, 158 gasoline, 43, 471, 478 gasoline-fuelled cars, 3, 4, 39, 296 gas-phase oxidation, 16, 303–8 of organic compounds, 6, 356 gas-to-particle partitioning, 16 GC, see gas chromatography GC-AED, see gas chromatography – atomic emission detector GC × GC-TOFMS, 475, 479, 481–2 GC pre-separation, 9, 223, 242–3 GC-IRMS, see gas chromatography – isotope ratio mass spectrometry GC-IRMS combustion interfaces, 417 GC-MS, see gas chromatography – mass spectrometry GC-RGD, see gas chromatography – reduction gas detector GEIA, see Global Emissions Inventory Activity generation of standard mixtures of VOC, 333 geranyl acetone (GA), 300 global air-to-sea ﬂux, 155 global annual emission rate anthropogenic VOC, 138 global budget, 5, 10, 95, 182 of acetone, 155 of OVOC, 138, 139, 147 role of the oceans, 147 Global Emissions Inventory Activity, 11, 34 global impact of organic ozone, 13 global network of measurements, 203 global tropospheric burden, 192 global warming potential, 176–9 glucose, 361 glycoladehyde, 152 glyoxal, 147, 150 global distribution, 149 glyoxylic acid, 372 gradient elution, 331 graphitic carbon adsorbents, 329, 330, 332, 348 grassland emission, 300 gravitational settling, 369 greenhouse gases, 2, 9, 13, 173, 221 green leaf volatiles, 143 group rate constants, 423, 426–7, 429, 439, 442, 444

493

group speciﬁc hydrogen KIEs, 429 group speciﬁc KIEs, 427–8, 429, 430, 436, 442, 445, 446 group speciﬁc rate constant, 428 growth rates, 188 GWP, see global warming potential H abstraction, 57, 303, 304, 308, 327 halocarbons, 13, 173–4, 175, 179, 192, 203, 204, 205, 210, 448, 450 halogenated chloroﬂuorocarbons, 197 halogenated VOC, 173, 448, 459 halomethanes, 448–51 halomethane sinks, 447, 450 halomethane sources, 449, 450 halons, 174, 185–6, 191, 194 Hantzsch reaction, 157–8 HCFCs, see hydrochloroﬂuorocarbons heavy-duty vehicle emissions, 296, 298 helium ionisation detectors (HID), 67, 158 hemiacetal, 14 Henry’s law constants, 90, 273, 285 Heptanal, 130, 301, 321, 323 2-heptanone, 319 3-heptanone, 319 heterogeneous reactions, 112, 113, 343, 355, 356 2-hexanone, 319 (Z)-3-hexen-1-ol, 140 (E)-2-hexenal, 140, 300, 302 (Z)-3-hexenal, 140 (E)-2-hexenol, 302 (Z)-3-hexenol, 136 (E)-2-hexenyl acetate, 302 (Z)-3-hexenylacetate, 140 HFCs, see hydroﬂuorocarbons higher molecular weight aldehydes, 130, 133, 147, 159 high-molecular-weight alkanals, 309, 323 high-molecular-weight alkanoic acids, 306, 323, 329, 332 high-molecular-weight carbonyls, 292 high-molecular-weight carboxylic acids, 137, 292 high performance liquid chromatography, 60, 66, 331, 371 high pressure limit, 272, 434 high watersoluble oligomers, 328 hopanes, 354 household products, 300 HOx radicals, 12, 82, 105, 113, 149–50 HPLC, see high performance liquid chromatography human health, 17–18, 129, 343, 480 human nose, 9

494

Index

humic acids, 356 humic like substances, 356 HULIS, see humic like substances hydration, 14 hydrochloroﬂuorocarbons, 13–14, 180–81, 197, 458–9 hydroﬂuorocarbons, 181, 191, 197 hydrogen abstraction channel, 433 hydrogen abstraction reactions, 427–32, 434–6, 438 hydrogen bond formation, 273 hydrogen KIEs, 406, 407, 427–32, 432–4, 439, 444, 447 hydrolysis, 235, 242, 355 hydrophilic organic species, 14 hydroxycresols, 304 hydroxyl radical, 13, 55, 179 hydroxy nitrates, 276, 280, 284, 285 δ-hydroxy nitrates, 285 hyphenated techniques, 68, 158 IC, see ion chromatography ice cores, 14, 54 impaction, 368, 369 impactor, 367–9, 369, 375 in-cloud processing, 355, 358 incomplete combustion, 42, 43, 348, 455, 457 indirect effect, 14, 342 indoor pollution, 18, 319–20 indoor sources, 299–300 industrial development, 19 industrial emissions, 43, 51, 183 industrial processes, 34, 41 infrared absorption constants for PAN, 226, 227 infrared spectroscopy, 9–10, 226, 373 injured plants, 302 insects, 9, 140, 146, 302 in situ measurements, 138, 479 intercomparison, 69–70, 160, 373 intercontinental pollution, 8 intercontinental transport, 55 Intergovernmental Panel on Climate Chance, 17 interhemispheric gradients, 192 international pollution export, 18 inventories, 34–7, 206 inverse isotope effect, 390, 433–4 iodine compounds, 186–7, 199–201 ion chromatography, 372 ion-trap sources, 331 IPCC, see Intergovernmental Panel on Climate Chance IRMS, see isotope ratio mass spectrometry IRMS interface, 409, 414 IR spectroscopy, 222, 226, 238, 242

isoprene, 55, 82, 86–8, 89, 105, 111, 113, 144, 244, 253, 256, 273–4, 305 isoprene, oceanic emissions, 92 isoprene carbon isotope ratios, 451, 456 isoprene emissions, 18, 83, 88, 92, 93, 96, 97, 102 isoprene lifetimes, 103 isoprene nitrates, 274, 275, 276, 280–81 isoprene synthesis, 87, 102 isoprenoids, 83, 87, 300, 302 isotope budgets, 402–3, 458 isotope effects, 390–91, 397, 404, 433–4, 436 isotope fractionation, 391, 395, 397, 406, 441, 446, 447, 449, 455, 457 isotope ratio mass spectrometry, 409, 412, 417 isotope ratio of VOC emissions, 447–8 isotope ratios, 397–9, 404 isotope ratios for urban VOC emissions, 451 isotope ratios of products of VOC reactions, 440–47 isotopically labelled VOC, 406 isotopologue concentrations, 394, 401, 440 isotopologues, 390, 406, 411, 414, 424, 430, 442 isotopomers, 390, 411, 424, 442, 444, 445 keto-enolic equilibrium of carbonyls, 328 ketolimonic acid, 363 ketolimononic acid, 363 ketones, 136–7, 151 KIE, see kinetic isotope effect KIE measurement techniques, 407–11, 459 kinetic isotope effect, 390, 396, 420–27 land use, 11, 19 laser-induced ﬂuorescence, 241, 276, 281 latitudinal gradient, 182, 280 leaf alcohols, 140 leaf esters, 140 levoglucosan, 361, 362 LIF, see laser-induced ﬂuorescence lifetimes of VOCs, 58–9 light dependency, 93–5, 98, 101, 141, 301 light-duty vehicles, 296, 298 lightning, 254 limonene, 301, 363 linalool, 103, 140, 144 linear range, 408 lipoxygenase, 143 lipoxygenase activity, 130, 136 lipoxygenase products, 136, 143 liquid chromatography, 61, 372 liquid extraction, 329, 331 liquid fossil fuel production, 3 LMCS cryogenic modulation, 475 long-lived semi-volatiles, 8

Index

Los Angeles, 45, 47, 51, 100, 137, 222, 243, 302, 306 lower stratosphere, 136, 160, 186 low volatile organic compounds, 344 LOX, see lipoxygenase luminol chemiluminescence, 239, 240 lumped species method, 11 lyase enzymes, 143 MACR, see methacrolein Mainz Isoprene Mechanism, 12 maleic acid, 358, 362 malic acid, 358 malonic acid, 358, 362 man-made sources, 293–300 marine source of alkyl nitrates, 277 markers for SOA, 358, 363 mass budget, 440, 442–3, 446, 459 mass conservation, 44, 440, 468 mass spectrometer, 63, 211, 213, 330, 372, 481 mass spectrometry, 68, 211, 330, 331 Master Chemical Mechanism, 12, 253, 480 maximum incremental reactivity, 60 MBO, see methyl butenol MCM, see Master Chemical Mechanism measurement technique, 114–16, 155–60, 210 measurements of VOC isotope ratios, 413, 458, 459 meat cooking, 299 mechanical wounding, 91 medium-duty diesel truck, 296 membrane ﬁlters, 370 mending of proteins, 140 menthone, 300 MEP, see methyl-erythritol phosphate MEP pathway, 87, 88, 89, 102 methacrolein, 109, 250, 438, 459 methane, 129, 427, 435 methanol, 89, 107, 112, 133, 135, 145 global source, 154 global source strength, 138, 154 methanol formation, 140 methoxyphenols, 354 methyl benzoquinone, 304 methyl bromide, 185 2-methyl-butanal, 302 3-methyl butanal, 300 2-methyl-3-buten-2-ol, 89, 94–5, 97, 103, 136, 139, 143 methyl chloride, 182 methyl chloroform, 6 4-methyl-3-cycloxene-1-one, 305 methylene chloride, 183 methyl-erythritol phosphate, 87, 88, 89, 102

495

methyl ethyl ketone (MEK), 136, 298, 300, 320, 321 3-methyl furan, 148 6-methyl-5-heptene-2-one (6-MHO), 294, 300, 301, 302, 319, 321 methyl hydroperoxide, 151 methyl iodide, 186–7 2-methyl pentanal, 302 4-methyl-2-pentanone, 300 methyl salicylate, 144 methyl-tert -butyl ether, 40, 43, 138 methylvinyl ketone, 109, 136, 148, 298, 305, 319, 320, 321 mevalonic acid, 87 Michelson interferometer, 18 microorganisms, 186, 355, 361, 447 mid-point modulator, 468 mineral dust, 345, 346 miniaturised GC, 483–4 MIM, see Mainz Isoprene Mechanism MIR, see maximum incremental reactivity mixture of isotopomers, 424 mobile sources, 38 modelling of ambient measurements, 34, 249–55 modelling SOA production, 281, 356 modelling VOC isotope ratios, 404 modern carbon, 353 modulation, 468, 471–2, 473 modulators, 468, 471–4 molecular composition of organic aerosols, 353–9 molecular symmetry, 422 Montreal Protocol, 173, 174, 181, 184, 185, 188, 202–3, 207 monoaromatic compounds, 467 monocarboxylic acids, 293, 298, 306 monoterpene emissions, 86, 88, 93, 96, 98, 105 monoterpene oxidation, 305, 321, 327–8, 362–3 monoterpenes, 68, 83, 103, 105, 113, 301 MPAN, see peroxymethacrylic nitric anhydride MS, see mass spectrometry MS-MS techniques, 331 MTBE, see methyl-tert -butyl ether multi-dimensional gas chromatography, 10, 114 multidimensional separations, 470 multifunctional organic nitrates, 280–82 MVA, see mevalonic acid MVA pathway, 87 MVK, see methylvinyl ketone myrcene, 306 myrtenal, 144 NAFION, 159 n-alkenals, 294

496

Index

nanotubes, 348 natural abundance isotope ratios, 407, 412, 421–2, 424, 428, 430, 432, 433, 441, 446 natural gas, 42, 47, 455 natural gas emissions, 41, 51 n-butyric acid, 137 NDSC, see Network for Detection of Stratospheric Change negative artefacts, 369, 373 negative ion chemical ionisation, 240, 275 Network for Detection of Stratospheric Change, 187 NICI, see negative ion chemical ionisation nitrate, 345 nitrate radical, 149, 152, 358 nitrophenols, 358 NMHC in natural gas, 455 NMR, see nuclear magnetic resonance NO, 55, 56, 58, 109, 242, 271, 303, 330, 332 N2 O, 14 NO2 , 55, 56, 109, 239, 241, 303 NOAA/GMD, 188, 214 nonanal, 130, 280, 300, 319, 332, 333 non-methane hydrocarbons (NMHC), 11, 54, 448, 451–6, 458 non-sea-salt (NSS) sulphate, 345 nopinone, 319, 328 normal isotope effects, 390 norpinonic acid, 109, 306 Norrish type I aldehydes, 324 Norrish type II aldehydes, 324 NOx , 13, 107–8, 112, 221, 247, 254, 255–6, 269, 358 nuclear magnetic resonance, 231, 373 nucleation, 113, 344 nucleation mode, 344 numerical modelling, 10, 11, 249, 404–5, 459–60

OC, see organic carbon ocean emissions, 16, 183 OC/EC-measurements, 370, 371, 373 OC/EC ratio, 349–50 OC/EC tracer method, 349–50 ocimene, 306 o-cresol, 304 octadecanoid pathway, 143 octanal, 130, 300, 301 1-octene, 304 odd carbon number, 354 ODE, see ozone depletion events ODP, see ozone depletion potential OH addition, 109, 284, 357–8

OH radical, 57, 58, 112, 151, 155, 303–6, 327, 329, 398, 423, 427, 428, 430, 433–9, 441–2, 446, 459 OH reactivity, 86, 281, 327 OH ratios, 103, 391 oil palm, 17 oligomers, 60, 328–9, 356 OM, see organic matter OM/OC ratio, 348–9 on-line aerosol mass spectrometry, 375 on-line mass spectrometry, 373–5 optical absorption, 9 ordered chromatograms, 470–71 organic aerosol mass, 355 organic aerosols, 342 organic carbon, 129, 345–6, 348–53, 370 organic hydroperoxides, 152 organic matter, 186, 345, 348–53 organic nitrates, 13, 112, 269, 303 organic nitrogen, 269–70, 282–3 organic peroxy radicals, 269, 270–71, 280 organic solvents, 41, 331 organic species, 3, 6, 7, 9, 10, 12–15, 16–19, 82 orthogonality, 469–70 OVOCs, see oxygenated volatile organic compounds oxalic acid, 137, 358, 362 oxidation interface, 409 oxidation of terpenes, 343, 357, 363 oxidation of toluene, 446 oxidation pathways, 129, 137, 147, 306 oxocarboxylic acids, 362 4-oxopentanal (4-OPA), 309, 319, 321, 332 oxygenated biogenic VOCs, 82, 85, 112 oxygenated monoterpenes, 144 oxygenated volatile organic compounds, 33–4, 40, 43, 57–8, 129, 482 oxy radicals, 303 ozone, 12, 82, 103, 130–31, 145–6, 252, 330, 349 ozone control, 12, 16 ozone depleting Cl cycle, 195 ozone depletion events, 185 ozone depletion potential, 176 ozone deposition, 86 ozone formation, 52, 55, 60, 103, 324, 327 ozone hole, 13, 343 ozone layer, 173, 180, 195 ozone production potential, 111, 292, 326, 476 ozone scrubbers, 330–31, 332 ozone-surface reactions, 308–9 ozonolysis, 148, 301, 304, 306, 329, 330 ozonolysis of lipids, 301, 308–9 ozonolysis of unsaturated phospholipids, 309

Index

PAH, see polycyclic aromatic hydrocarbons paint industry, 41 PAN, see peroxy acetyl nitrate PAN/NOy ratio PANs in regionally impacted air masses, 244, 247–8 PANs in remote environments, 246 PAR, see photosynthetic active radiation particles, 14–15, 201, 298, 346, 367, 368, 369, 481 particulate matter, 327–8, 344, 346, 369, 371, 446–7 partitioning coefﬁcient, 355 partition method, 356 pathogen attack, 143, 146, 302 PCBs, see polychloro-biphenyl compounds peak capacity, 62, 468–9, 470 pectin demethylation, 140 pentanal, 300, 320 2-pentanone, 300 3-pentanone, 143 1-penten-3-ol, 136, 143 1-penten-3-one, 143 perﬂuorocarbons, 174, 181–2, 191 perﬂuorophenylhydrazine, 331–2 perﬂuorosulfonic acid, 159 peroxide-bicyclic route, 304 peroxides, 13, 14, 56, 109, 152, 223, 328, 357 peroxyacetic acid (PAA), 223, 241 peroxyacetic nitric anhydride, 62, 222, 270 peroxy acetyl nitrate, 6, 10, 13, 58, 129, 221, 270, 308, 327 peroxybenzoic nitric anhydride, 223 peroxycarboxylic acid, 223 peroxycarboxylic nitric anhydride, 221, 222–3, 231 peroxyisobutyric nitric anhydride, 223, 327 peroxymethacrylic nitric anhydride, 222, 223, 234, 244, 245, 249–50, 255, 327 peroxy nitrite, 270, 271–3, 286 peroxy radicals, 56, 57–8, 109, 152, 241, 269, 270, 271, 273–4, 280 persistent organic pollutants, 8, 15 petrochemical products, 3, 244, 467, 475 petrochemistry, 138 PFCs, see perﬂuorocarbons PFPH, see perﬂuorophenylhydrazine phenol, 330 phenolic pathway, 304 phenology, 93, 96–7 phenyl propanoid pathway, 144 photochemical age, 270, 397–9, 400, 451 photochemical clock, 277 photochemical losses, 191, 192 photochemical models, 16, 252–3, 255, 256

497

photochemical OVOC formation, 147–9 photochemical ozone creation potential, 60, 253 photochemical processing, 55, 129, 148–9, 179, 333, 391, 399 photochemical production, 8, 14, 53, 242, 244, 247, 248, 292, 303–9, 320, 321, 323 photochemical reactions, 1, 59, 107, 187 photodissociation of high molecular weight carbonyls, 326 photoionisation detector, 67, 114 photolysis, 6, 55, 58, 105, 129, 149, 150–52, 179, 192, 194, 199–200, 201, 230, 233, 242, 252, 283, 284–5, 309, 410, 457 photolysis of formaldehyde, 148–9, 436 photosynthetic active radiation, 93, 141, 142, 145, 321 phthalic acid, 358–9, 362 physical artefacts, 370 physical properties, 226 PiBN, see peroxyisobutyric nitric anhydride PID, see photoionisation detector α-pinene, 90, 109, 111, 274, 301, 306, 328, 365, 420 β-pinene, 301, 306, 365 pine plantations, 17, 95, 139, 153 pinic acid, 109, 363 pinonaldehyde, 306, 319 pinonic acid, 109, 306, 363 Pinus pinea, 144, 301, 306 Pinus sylvestris, 144 plant abrasion, 354 plant wax, 354 PM, see particulate matter PM2.5, 346, 352, 369, 481 polar compounds, 150, 292, 293, 296, 473, 474 polar sunrise, 198, 277 polar VOC, 292, 293, 294, 300, 302, 303, 330 pollen, 345, 346, 347, 361 polluted areas, 17, 48–50, 55, 136, 350, 352 pollution events, 8, 203, 359–60 polyaromatic hydrocarbons, 8, 15, 349, 354, 360 polychlorinated biphenyls, 8, 15, 474 polychlorinated dibenzo-p-dioxins, 8 polychloro-biphenyl compounds, 8 polycyclic aromatic hydrocarbons, 348, 358, 362 polymerisation, 14, 223 polyols, 113, 357 Ponderosa pine, 86, 94, 95, 136, 139 POCP, see Photochemical Ozone Creation Potential POPs, see Persistent Organic Pollutants porous layer open tubular columns, 409 porous polymers, 329, 330 positive artefacts, 370 positive matrix factorization, 44–5

498

Index

preconcentration, 158, 159, 275, 329, 333, 420 precursor, 9, 14, 15, 16, 42, 53, 58, 87, 89, 113, 133, 137, 155, 237, 244, 246, 249, 269, 276, 304, 306, 308, 328, 342, 355, 356, 479 precursors of biogenic SOA, 356–7 preseparators, 369 primary OC, 349–52 primary organic aerosol, 346, 349, 353–5 primary organic aerosols in the marine environment, 355 primary ozonide, 148, 303 primary particles, 343, 359–65 primary sources, 41, 107, 111, 114, 129, 137, 138, 358, 362 principal component analysis (PCA), 44 production rate, 182, 186, 237, 244–5, 276, 283, 304 products from its atmospheric oxidation, 357 products of VOC oxidation, 440, 446 propanol, 143, 146 propionaldehyde, 133, 441, 442, 443–4 propionic acid, 137 proton transfer reaction mass spectrometry, 70, 114, 116, 138–9, 158, 241, 286, 293, 332 proton transfer reactions, 10 pseudo ﬁrst-order loss, 400–401 PTR-MS, see proton transfer reaction mass spectrometry pyrolysis interface, 409 pyrolysis products, 354 pyruvic acid, 89, 137 quadrupole mass spectrometer, 68, 69, 475 quality assurance, 68–70 quartz ﬁbre, 370 quartz ﬁbre ﬁlter, 367, 370, 373 Quercus ilex, 144, 301, 306 Quercus petraea, 300 RACM, see Regional Atmospheric Chemistry Model radiative properties, 176, 343 radiocarbon data, 321 RADM, see Regional Acid Deposition Model rate constant, 232, 234, 390, 391, 406, 422–7, 428, 429, 434, 438 Rayleigh dependence, 396 Rayleigh fractionation, 393–7 Rayleigh type equation, 395 Rayleigh type function, 396, 397, 440 reactions with radical species, 233–4 reaction yields, 305, 308

reactivity, 11, 16, 42, 83, 103, 105, 107, 281, 304, 358, 404, 459, 479 reference gas injections, 416 reﬁning, 98, 138, 205–7 Regional Acid Deposition Model, 11, 252 regional air quality modelling, 97, 252 Regional Atmospheric Chemistry Model, 11–12, 252 regional scale advection, 8 relative rate, 256, 406, 407, 408, 409–10 relative retention, 330 relaxed eddy accumulation (REA), 114, 115–16 remote sites, 34, 62, 131, 256, 309, 323, 479 remote troposphere, 137, 221, 230, 254 removal of water, 63–4, 158–9, 409, 420 removal process, 5, 6, 7, 129, 152, 155, 183, 184, 193, 434 respiratory diseases, 129, 342 reversed-phase liquid chromatography, 331 ring degradation products, 358 ring retaining, 57, 357–8 roof tile structure, 471 rural areas, 51, 83, 131, 132, 133, 135–6, 137, 256 rural environments, 109, 130 sabinene, 301, 306, 363 sabinic acid, 363 saccharides, 347, 355 Salvia mellifera, 302 sample dimensionality, 470 sample enrichment, 411–13, 417, 419 sample pre-concentration, 64–5, 66 sample preparations, 64, 365–7, 371, 480 sample treatment, 371 sampling, 62–3, 67, 114–16, 155–160, 207–10, 239, 255, 294, 329–33, 365, 367–70, 373, 411–13 sampling artefacts, 293, 330, 332, 370–71 sampling errors, 368, 370 saturated aldehydes, 140, 144 saturated linear ketones, 294 SCIAMACHY, 133, 149 sea salt, 179, 199, 200, 201, 345 sea spray, 346, 355 seasonal cycle, 49, 54, 96 secondary carbonyl compounds, 147 secondary OC, 350–52 secondary organic aerosol, 60, 82, 83, 86, 108, 109–10, 112–13, 116, 342, 349–50, 352, 355–65, 374, 375 secondary organic aerosol formation, 113, 281, 327–9, 349, 356, 357 secondary particles, 343, 344 secondary sources, 129, 137, 358, 360, 362

Index

selective desorption, 159 selective ions, 330 selective preconcentration, 158 selective sampling, 158 semi permeable membranes, 158, 159 semi volatile organics, 1, 66, 355 senescence of plant tissues, 140 SEOM, see solvent extractable organic matter separation, 10, 15, 61–2, 65–7, 68, 69, 158, 159, 210, 239, 243, 367, 371–3, 403, 404, 408–9, 414, 468, 469–70, 475 sesquiterpenes, 83–5, 88, 89, 96, 103, 113, 114, 281 sewage treatment, 138 SFE, see supercritical ﬂuid extraction short-chain carboxylic acids, 372 short-chain dicarboxylic acids, 361, 362, 372 signal enhancement, 468 signalling compounds, 302 sinks of atmospheric particles, 343–4 site speciﬁc enrichment, 446 site speciﬁc rate constants, 423, 428 site speciﬁc reactivities, 428, 429 size fractions, 367 size modes of atmospheric particles, 344 SJAC, see steam jet aerosol collector smog chamber, 113, 306, 479, 482, 483 SMOW, see standard mean ocean water SO2 , 13, 142, 345 SOA, see secondary organic aerosol SOA production, 281, 356, 357 SOA formation, 60, 113, 281, 327–9, 349, 355–9 SOGE, 187, 206, 212, 213 soil, 33, 91, 93, 146, 153, 182, 183, 355, 447 soil uptake, 449 solvent evaporation, 41, 42, 137, 371 solvent extractable organic matter, 367 soot, 15, 298, 348 sources of anthropogenic VOCs, 33–45 sources of primary organic aerosol, 353 source studies, 447 spectroscopic data, 226–9 spores, 346, 361 SRR, see structure-reactivity relationships stable carbon isotope ratios of atmospheric NMHC, 451, 454 stable carbon isotope ratios of oxygenated VOC, 457 stable carbon isotope ratios of sources, 450, 454 stable isotope ratios of atmospheric VOC, 388, 391–405, 447–58 stainless steel canisters, 62, 209, 412 standard emission rates, 4, 138, 295, 296, 297 standard mean ocean water, 448, 450, 451 stationary emissions, 40

499

steady-state, 176, 250, 401, 402, 443, 444 steam jet aerosol collector, 368–9 steranes, 354 sterols, 355 stomatal conductance, 90, 96, 141 stomates, 90, 237 stratospheric ozone depletion, 184, 194, 198, 201 structure–reactivity relationship, 55 substituents, 273, 274, 467, 476 suburban sites, 49, 53, 319, 320, 321, 324 succinic acid, 358, 362 sucrose, 355, 361 sugar, 347, 349, 354, 355, 361, 362, 372 sugar alcohols, 347, 355 sugar concentrations, 361, 363 sulcatone, 302 supercritical ﬂuid extraction, 371 surrogate species method, 11 suspension, 153, 343, 350, 353, 354 sweetgum, 17 synthetic preparation, 222–6 system dimensionality, 482 TAME, see tert-amyl-methyl ether TAN, see total alkyl nitrates TBA, see tert-butyl alcohol TC, see total carbon TDLAS, see Tunable Diode Laser Absorption Spectroscopy Teﬂon ﬁlters, 367, 370, 373 temperature dependence, 88, 94, 95–6, 145, 233, 271–3, 285, 301, 355, 408 temperature dependence of KIEs, 439–40 TENAX, 63, 64, 159, 275, 329, 330, 331 tetrachloroethene, 184 terpene oxidation products, 327, 357, 363 terpene ozonolysis, 357 terpenes, 3, 9, 16, 17, 18, 55, 60, 88, 89, 94, 95, 105, 109, 281, 343, 344, 355, 356, 357, 362–3 terpenoid emissions, 83, 88, 91 terpenoids, 82, 83, 85, 86, 88, 89, 92, 281, 301 terpinolene, 305, 306 terpinolenonaldehyde, 306 terrestrial emissions, 93, 182, 186 terrestrial vegetation, 89, 91, 92, 139–46, 149, 300–302, 343 tert-amyl-methyl ether, 138 tert-butyl alcohol, 138 theoretical plates, 65, 468, 469 thermal decomposition, 231, 232, 233–4, 235, 240, 241, 245, 250, 276, 281 thermal desorption, 66, 159, 329, 331, 373, 374, 481

500

Index

thermal modulators, 472–3 thermotolerance, 88 thin layer chromatography, 371 three-dimensional chemistry and transport models, 10, 12 three-member transition state, 271 time-of-ﬂight mass spectrometer, 374–5, 475 tobacco smoke, 138, 300 TOF, see time-of-ﬂight mass spectrometer tolualdehydes, 319 toluene, 43, 51, 358, 446, 479, 480 total alkyl nitrates, 327 total carbon, 15, 353, 373 total OC, 352 total organic bromine (CBry ), 203 total organic chorine (CCly ), 203 tracer for anthropogenic SOA, 358–9 tracer for certain plant species, 354 tracer for fungi, 347 tracer-ratio techniques, 43 tracers, 2, 245, 354, 355 trafﬁc, 47, 48, 52, 354 transition state, 271, 272, 421, 436 transition state theory, 420–21, 422 translational degrees of freedom, 421 transportation related emissions, 455 trend, 18, 187–92, 243, 246, 319–23, 362 trichloroethene, 184 triterpenoids, 354 tropospheric chlorine, 174, 202 tropospheric ozone, 17, 107, 112, 192, 198, 269 tropospheric ozone formation, 37, 55, 56, 59–60, 116 TROPOZ II, 133 TST, see transition state theory tunable diode laser absorption spectroscopy, 157 tunnel measurements, 39 two-dimensional separations, 469 two-endpoint mixing, 392–3 two-endpoint mixing curves, 393

undecanal, 319 unsaturated VOC, 432–4, 436–9 upper troposphere, 12, 16, 82, 129, 136, 151, 229, 249, 254, 256 uptake by plants, 153–4 urban alkane sources, 453 urban areas, 39, 41, 43, 92, 137, 238, 244, 319–20, 359–60, 361, 449, 467, 477 urban environments, 37, 42, 44, 45, 52, 129 urban sites, 131, 319–20 UV photolysis, 303, 309, 321, 323 UV-visible absorption, 228–9, 331

valve modulation, 473, 479 valve modulator, 473–4, 477 vanillic acid, 354 varnish industry, 138 vegetation patterns, 17, 19 vegetation species, 83, 100, 300, 302 vehicle exhaust, 42, 43, 45 vehicular exhaust emission, 296–8, 304, 306, 319–20 verbenone, 144 vertical distribution, 54, 107 very reactive BVOC (VR-BVOC), 86, 89, 95–6, 103, 107, 281 vibration frequencies, 421, 434 vibrational degrees of freedom, 439 vibrational partition functions, 434 Vienna-Peedee Belemnite scale (V-PDB), 389 virtual impactor, 368 viruses, 346 VOC budgets, 399–404 VOC degradation pathways, 55–9 VOC deposition, 447 VOC isotope ratios, 402, 404, 406–7, 448, 460 VOC life-time, 54, 58–9, 448 VOC oxidation, 109, 114, 440 VOCs in global background regions, 53-5 VOCs in rural and continental background areas, 52–3 VOCs in urban areas, 50–52 volatile organic compounds (VOC), 11, 42–3, 65–7, 292, 294, 330, 408, 414, 417, 422, 430, 446, 455 wall coated capillary columns, 409 wall losses, 368 waste incineration, 138, 182 waste management, 42 water soluble organic carbon, 356, 361, 362–3, 373 water soluble organic compounds, 361 water trap, 420 wet and dry deposition, 6, 109, 129, 152–3, 154, 234–7, 285 wet chemical methods, 157, 223 wet deposition, 149, 150, 152–4, 234, 237, 283, 285, 344 wound-VOC, 91, 143 WSOC, see Water Soluble Organic Carbon xylem ﬂow, 141 xylenes, 43 zero point energy, 421, 434, 439

Unit: Gg NMVOC =0 0.0–0.1 0.1–1 1–2

2–10 10–50 50–100 100–2000

Plate 1 Gridded VOC emissions from anthropogenic sources in the year 2000 from EDGAR 32FT2000 (With permission from EDGAR 2005).

Volatile Organic Compounds in the Atmosphere Edited by Ralf Koppmann Copyright © 2007 by Blackwell Publishing Ltd

Global total: 1.8e + 11 kg (min. = 60.9, max. = 1.1e + 009)

60

HCHO (molecule/cm2)

Latitude

30

2.0 × 1016 1.6 × 1016 1.2 × 1016 8.0 × 1015 4.0 × 1015 0.0

0

−30 −60 −180 −150 −120 −90

−60

−30 0 30 Longitude

60

90

120

150

180

Plate 2 The global distribution of formaldehyde derived from SCIAMACHY measurements from August 2004–July 2005 (with permission from F. Wittrock et al. 2006). 60

Glyoxal (molecule/cm2)

Latitude

30

1.2 × 1015 1.0 × 1015 8.0 × 1014 6.0 × 1014 4.0 × 1014 2.0 × 1014

0

−30 −60 −180 −150 −120 −90

−60

−30 0 30 Longitude

60

90

120

150

180

Plate 3 The global distribution of glyoxal (CHOCHO) for the year 2005 derived from SCIAMACHY measurements. Glyoxal is an oxidation product of VOCs and thus an indicator for photochemical processing of polluted air masses (with permission from Wittrock et al. 2006).

CFC -12 Year 2000

<10−3%

<10−2% <3×10−2% <6×10−2% <10−1% <3×10−1%

<1%

Plate 4 Gridded distribution of CFC-12 emissions for Europe in year 2000. Percentages refer to the global total emission (134.1 Gg) Reproduced with permission from A. McCulloch, Final Report of the EU FP5 project “System for Observation of Halogenated Greenhouse Gases in Europe (SOGE)”. Project EVK2-2000-00674.

Latitude (°N)

55

50

45

−70 −65 −60 Longitude (°W)

−75

−55

0

100 200 300 400 500 600 (PAN), pptv

0

1

2 3 4 5 6 7 Pressure altitude (km)

8

Plate 5 Measurements of PAN made on the 11 September 1997 ﬂight during NARE 97. PAN values are coded by colour, and altitude is coded by marker size as indicated by the scale. 10

8 hrs

8 7 6 5 4

4 hrs

3 2

(PAN)/(Propanal)

2 hrs 1 8 7 6 5 4

1 hr

3

0.5 hrs

2

0.1

0.25 hrs

8 7 6 5 4 3 2 3

4

5

6 7 8 9

0.1

0

25

2

3

4

5

6 7 8 9

(PAN)/(Ethanal) 50 75 100 (Ozone), ppbv

2

3

1

125

Plate 6 The ratio [PPN]/[propanal] vs [PAN]/[ethanal]. The individual points are from the NEAQS 2002 study for samples taken near the coast of New England, colour coded by the co-measured O3 mixing ratios. Also shown are the data from the Nashville 1999 Study (squares), and TexAQS 2000 Study (circles) plotted in 8 bins, with error bars denoting the extent (maximum and minimum) of each bin on the horizontal scale, and the standard deviation of each bin on the vertical scale. The solid diamond is the point derived from the steady state calculation of Sillman et al. (1990).

Retention time 2 (s)

0.5 1.0

2

1

1.5 3 2.0 2.5

4

3.0 3.5 0

10

20

30

40

50

Retention time 1 (min)

Plate 7 Structured GC × GC chromatogram of gasoline. Red square is expanded in lower section. 1, alkanes; 2, alkenes and cycloalkanes; 3, monoaromatics; 4, polyaromatics.

1

2

3

4

40 H

G

30

F 25 E 20 D 15 C B

Retention time on BP-1 column (min)

35

10

A 5

30 20 10 Flame ionization detector response (pA)

1 2 3 4 Retention time on BP-50 columns (s)

Plate 8 Comprehensive and one-dimensional separations of VOCs in urban air. A, benzene; B, heptane; C, toluene; D, xylenes; E, C3 -benzenes; F, C4 -benzenes; G, C5 -benzenes; H, naphthalene; 1, aliphatic band; 2, carbonyl band; 3, monoaromatic band; 4, bi-aromatic band. Reproduced with permission of Nature Publishing Group, from: Lewis et al. (2000). A larger pool of ozone-forming carbon compounds in urban atmospheres. Nature, 405: 778–81.

400 300

A

200 100 0

C

B 0

10

20

30

40

50

60

70

80

90

100

Second column RT/s

3

1 2 3

5 6

A

B

C

D 4

8 9

H

11 5

5

10 15 20 25 30 35 40 45 50 55 60 65 70 75 80

Second column RT/s

First retention time/min 3 4 5D

E

G

F

7 8 9 50

55

61 66 72 Retention time/min

77

82

66

75

84

Plate 9 Comparison of single column (upper) and GC×GC separations (lower) of a Leeds urban air sample. Areas of the full chromatogram are successively extracted at higher gain to illustrate increasing isomeric complexity at higher boiling points. GC × GC chromatograms are annotated with start of individual Cx isomer band (running right to left) where A = C2 , B = C3 , C = C4 , D = C5 , E = C6 , F = C7 , G = C8 , H = naphthalene. Chemical banding assignments: 1, aliphatic; 2, oleﬁns; 3, oxygenates; 4, monoaromatics; and 5, polyaromatics. Reproduced with permission of Elsevier, from: Hamilton et al. (2003). Mono aromatic complexity Atmospheric Environment, 37(5): 589–602.

A

Increasing polarity

C

B

B

Increasing polarity

A

C

Increasing boiling point

Second retention time (s)

Plate 10 Comparison of GC × GC chromatograms of gasoline vapours and urban air. Upper: Leeds urban air chromatogram. Lower: Gasoline vapours at 20◦ C chromatogram. A, C3 -substituted monoaromatics; B, C4 -substituted monoaromatics; C, C5 -substituted monoaromatics. Reproduced with permission of Elsevier, from: Hamilton et al. (2003) Mono aromatic complexity. Atmospheric Environment, 37(5): 589–602.

0.75 1.00

1

1.25

2

1.50

7

12 9

10

13

11

15

3

1.75

14

8

2.00

5

2.25 7.5

Second retention time (s)

6

4

1.25 1.50 1.75 2.00 2.25 2.50 2.75 3.00 3.25 3.50 3.75

10.0

12.5

15.0 17.5 20.0 First retention time (min)

7

22.5

25.0

22.5

25.0

13 12 16

Toluene 7.5

10.0

12.5

15.0 17.5 20.0 First retention time (min)

Plate 11 Chromatograms for a typical standard mixture and a chamber sample during the photo-oxidation of toluene. Peak identiﬁcations are given in Table 11.1. Reproduced with permission of the EGU, from Hamilton et al. (2003b) Measurements of photo-oxidation products from the reaction of a series of alkylbenzenes with hydroxyl radicals during EXACT using comprehensive gas chromotography. Atmospheric Chemistry and Physics, 3: 1999–2014.

Polarity second dimension (s)

(a)

1060 GC × GC-TOFMS peaks from Augsburg aerosol sample (All peaks shown) 4 3.5 3 2.5 2 1.5 1 1 000

1 500

2 000

2 500

3 000

3 500

Volatility first dimension (s)

Polarity second dimension (s)

(b)

1060 GC × GC-TOFMS peaks from Augsburg aerosol sample coloured (Assigned peaks shown in colour) 4 3.5 3 2.5 2 1.5 1 1 000

1 500

2 000

2 500

3 000

3 500

Volatility first dimension (s) Plate 12 (a) shows the bubble plot of the entire peak apices used in the study for grouping. (b) indicates the different groups identiﬁed on (a). Colour assignments: Burgundy, alkanes; Lilac, alkenes and cycloalkanes; Pink, n-alkane acids; Blue, alkyl substituted aromatics; Brown, polar benzenes; Green, hydrated naphtalenes and alkenyl benzenes; Red, naphthalenes and alkylated naphthalenes. Reproduced with permission of Elsevier from Welthagen et al. (2003) Search criteria and rules for comprehensive twodimensional gas chromatography-time-of-ﬂight mass spectrometry analysis of airborne particulate matter. Journal of Chromatography A, 1019: 233–49.

4 2nd Dimension time (s) 1 2 3 0 4 2nd Dimension time (s) 1 2 3 0 4 2nd Dimension time (s) 1 2 3 6

6

6

506

506

Single compound

506

Isomers

Groups

1006 1506 1st Dimension time (s)

Naphthalene

Masses: 128

1006 1506 1st Dimension time (s)

2006

2006

2006

(c)

(b)

Chloro-alkanes

C4-mono aromatics

Masses: 134

(a)

Alkyl-benzenes

1006 1506 1st Dimension time (s)

Masses: 91

Plate 13 Extraction of groups, isomers and single compounds using single ion GC × GC chromatograms. (a) Groups (m/z 91 Da), (b) isomers (m/z 134 Da) and (c) single compound (m/z 172 Da).

0