Handbook of Physical Testing of Paper

An aged god reading a Maya codexÐa screenwise-folded book. Detail from a carved vase of pre-Columbian age.

Handbook of Physical Testing of Paper Volume 2 Second Edition, Revised and Expanded edited by

Jens Borch IBM Corporation Boulder, Colorado

M. Bruce Lyne International Paper Tuxedo, New York

Richard E. Mark Empire State Paper Research Institute State University of New York College of Environmental Science and Forestry Syracuse, New York

Charles C. Habeger, Jr. Institute of Paper Science and Technology Atlanta, Georgia Associate Editor

Koji Murakami Faculty of Agriculture Kyoto University Kyoto, Japan

Marcel Dekker, Inc.

New York • Basel

TM

Copyright © 2001 by Marcel Dekker, Inc. All Rights Reserved.

The ®rst edition of this book was published as Handbook of Physical and Mechanical Testing of Paper and Paperboard, R. E. Mark, ed., Marcel Dekker, Inc., 1984. Cover illustration: Environmental scanning electron micrograph of an uncoated commercial boardÐ35# High Performance (HP) linerboard by Solvay Paperboard, Syracuse, New York. This product has a nominal basis weight of 171 g/m2 . It is made from a 100% recycled furnish drawn almost exclusively from OCC (Old Corrugated Container) stock. Arrow (horizontal) indicates machine direction; bar scale below it shows a 200 mm length, which yields a magni®cation of  170X. This picture has the same magni®cation as that on the cover of Vol. 1, illustrating the greater coarseness and degree of collapse of these recycled softwood ®bers compared with papermaking bark ®bers. Micrograph courtesy of Dr. Susan E. Anagnost, N.C. Brown Center for Ultrastructure Studies, SUNY College of Environmental Science and Forestry, Syracuse, New York. ISBN: 0-8247-0499-1 This book is printed on acid-free paper. Headquarters Marcel Dekker, Inc. 270 Madison Avenue, New York, NY 10016 tel: 212-696-9000; fax: 212-685-4540 Eastern Hemisphere Distribution Marcel Dekker AG Hutgasse 4, Postfach 812, CH-4001 Basel, Switzerland tel: 41-61-261-8482; fax: 41-61-261-8896 World Wide Web http://www.dekker.com The publisher offers discounts on this book when ordered in bulk quantities. For more information, write to Special Sales/Professional Marketing at the headquarters address above. Copyright # 2002 by Marcel Dekker, Inc. All Rights Reserved. Neither this book nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, micro®lming, and recording, or by any information storage and retrieval system, without permission in writing from the publisher. Current printing (last digit): 10 9 8 7 6 5 4 3 2 1 PRINTED IN THE UNITED STATES OF AMERICA

PREFACE TO THE SECOND EDITION

In the mid-1990s, an agreement was concluded between the four editors and the publisher to produce a second edition of the Handbook of Physical and Mechanical Testing of Paper and Paperboard. The new version has a shorter title, Handbook of Physical Testing of Paper; but it is a substantial revision, modernization, and expansion of the previous work. This edition contains 30 chapters (16 in Vol. 1, 14 in Vol. 2), one-third of which are entirely new. The other 20 chapters are more speci®cally ``revised and expanded'' from the previous edition. However, some of them have been totally rewritten to re¯ect the perspectives of the author(s) as well as to modernize the technical content. We hope that this edition will be at least as useful in today's world as the ®rst edition has been over the past years. We have learned that the creation of a technical book today involves vastly different problems from those encountered while preparing the ®rst edition. But the positive experiencesÐwith our authors, with the publisher's excellent professional staff, with Associate Editor Koji Murakami, with readers and critics, and with supportive employersÐfar outweigh the negatives. For these factors, we are most grateful. Jens Borch M. Bruce Lyne Richard E. Mark Charles C. Habeger, Jr.

iii

PREFACE TO THE FIRST EDITION

We have tried to assemble here as many of the modern aspects of properties testing in the paper and paperboard ®eld as possible. One of our objectives is to enable those concerned with planning, specifying, and evaluating the physical and mechanical testing of these materials to take advantage of the many advances and improvements that have taken place in recent years. We also feel that it is important and useful to codify the excellent pioneering work that has provided the essential base for these advances, which has been included to the maximum extent possible. As with their associates in other aspects of the broad and dynamic paper industry, the pioneers in paper testing have come from many diverse areas of the world; the authors who have contributed to these volumes live on four continents and represent a spectrum of industrial, government, and educational institutions, bringing their own special insights to the areas covered in their respective chapters. It goes without saying that the very nature of the subject matter covered here has wideranging applicability; our intention has been to treat the topics in our various chapters in a holistic manner. We wish to give some justly deserved praise to the contributors to this handbook and their af®liated organizationsÐacademic, industrial, and governmental. Creation of a multiauthored book of this type requires more than professional expertise. It also requires a great deal of diligence, perseverance, and willingness to accept criticism, as well as laborious revision and adjustment to the coverage of related topics in other chapters. In short, it is hard work that calls for dedication to purpose. An editor fortunate enough to have become associated with this group of contributors is lucky indeed, for they have been exceptional in all these aspects, as has Associate Editor Koji Murakami. The organizations of which each contributor is a part must also be praised, for without their cooperation and support, even the most dedicated contributor would ®nd it dif®cult if not impossible to do all that is required. Finally, we are all deeply indebted to the management and staff of Marcel Dekker, Inc., who have combined rigorous publishing professionalism with genuinely compassionate accommodation to the severe personal traumas that accompanied the preparation of these volumes. Thank you, all. v

vi

Preface to the First Edition

ACKNOWLEDGMENTS The history of pre-Columbian paper and books in the New World is little known. Through the generous contributions of several leading companies in the paper industry, it has been possible to include color illustrations of several aspects of paper and papermaking in early Mesoamerica. We are indebted to these donors: Abitibi-Price, Inc., Crown Zellerbach Corporation, International Paper Company, Mead Corporation Foundation, St. Regis Paper Company, Scott Paper Company, Westvaco Corporation, and Weyerhaeuser Company. During the ®nal preparation days of this book, all of us associated with this work were saddened to learn of the untimely death of our esteemed colleague, Dr. Fred Sha®zadeh. We are proud to dedicate this volume to his memory. Richard E. Mark

CONTRIBUTORS

Ernst L. Back

Feedback Consulting, E&E Back KB, LidingoÈ, Sweden

Gary A. Baum

The Institute of Paper Chemistry, Appleton, Wisconsin

Raymond G. Bayer Einar Bùhmer Jens Borchy

Cogito Consulting A.S., Oslo, Norway IBM Corporation, Boulder, Colorado

John L. de Yong Les Groom

Tribology Consultant, Vestal, New York

Consultant Physicist, Victoria, Australia

Southern Research Station, U.S. Forest Service, Pineville, Louisiana

Lucille D. Hahny Technical Center, Champion International Corporation, West Nyack, New York Per-AÊke Johansson Sweden Sueo Kawabataz M. Bruce Lyne

STFI, Swedish Pulp and Paper Research Institute, Stockholm,

Kyoto University, Kyoto, Japan International Paper, Tuxedo, New York

Current af®liation:  Institute of Paper Science and Technology, Atlanta, Georgia. y Retired. z Professor Emeritus, Kyoto University.

vii

viii

Contributors

Patricia A. Moss The University of Manchester Institute of Science and Technology, Manchester, England Koji Murakamiy

Graduate School of Agriculture, Kyoto University, Kyoto, Japan

Gary M. Scott Empire State Paper Research Institute, State University of New York College of Environmental Science and Forestry, Syracuse, New York Sami Simulaz Finland Pia WaÊgberg

The Finnish Pulp and Paper Research Institute (KCL), Espoo, SCA Packaging Research, Sundsvall, Sweden

Tatsuo Yamauchi Japan

Graduate School of Agriculture, Kyoto University, Kyoto,

Current af®liation:  Oy KeskuslaboratorioÐCentrallaboratorium Ab (KCL), Espoo, Finland. y Professor Emeritus, Kyoto University. zOkmetic Oyj, Vantaa, Finland.

CONTENTS

Preface to the Second Edition Preface to the First Edition Contributors Contents of Volume 1 Paper, Books, and Paper Testing: The Origins of Paper (and Books) in the New World Part 1.

iii v vii xi xiii

Testing in the Laboratory and on the Paper Machine

1.

The Paper and Board Testing Laboratory Lucille D. Hahn

2.

On-Line Testing of Paper Gary M. Scott

13

3.

Conditioned Test Atmospheres John L. de Yong

47

Part 2.

Appearance and Structure Analysis

4.

Optical and Appearance Properties Jens Borch

5.

Microscopy Patricia A. Moss and Les Groom

Part 3.

1

95 149

Interactions with Liquids and Gases

6.

Porosity and Gas Permeability Tatsuo Yamauchi and Koji Murakami

267

7.

Wetting and the Penetration of Liquids into Paper M. Bruce Lyne

303

ix

x

Contents

Part 4. Electrical and Thermal Interactions 8.

Electrical Properties: I. Theory Gary A. Baum

333

9.

Electrical Properties: II. Applications and Measurement Methods Sami Simula

361

10.

Thermal Properties Einar Bùhmer

389

Part 5. Physical and Surface Properties 11.

Characterization of Paper Surfaces Using Optical Pro®lometry Pia WaÊgberg and Per-AÊke Johansson

429

12.

Paper Friction Ernst L. Back

451

13.

Paper Abrasivity Raymond G. Bayer

477

14.

Testing the Tactile Properties of Tissue and Nonwovens Sueo Kawabata

505

Index

531

CONTENTS OF VOLUME 1

Part 1.

Theory and Test for Mechanical Parameters

1.

Models for Describing the Elastic, Viscoelastic, and Inelastic Mechanical Behavior of Paper and Board Richard W. Perkins, Jr.

2.

Viscoelastic Properties Lennart SalmeÂn and Roger Hagen

3.

Dimensional Stability and Environmental Effects on Paper Properties Tetsu Uesaka

4.

Moisture-Accelerated Creep Henry W. Haslach, Jr.

5.

Bending Stiffness, with Special Reference to Paperboard Christer Fellers and Leif A. Carlsson

6.

Ultrasonic Determinations of Paper Stiffness Parameters Charles C. Habeger, Jr.

7.

Deformation and Failure Behavior of Paper Curt A. Bronkhorst and Keith A. Bennett

8.

Fracture of Paper M. T. Kortschot

9.

Edgewise Compression Strength of Paper Christer Fellers and Benjamin C. Donner

10.

Residual Stresses in Paper and Board John Frederick Waterhouse

xi

xii

Contents of Volume 1

Part 2. Special SituationsÐMechanical Properties of Products 11.

Corrugated Board Rob Steadman

12.

Physical and Mechanical Properties of Towel and Tissue M. K. Ramasubramanian

Part 3. Structural ParametersÐFibers, Bonds, and Paper 13.

Fiber Structure Leena Paavilainen

14.

Mechanical Properties of Fibers Richard E. Mark

15.

Determination of Fiber±Fiber Bond Properties Tetsu Uesaka, Elias Retulainen, Leena Paavilainen, Richard E. Mark, and D. Steven Keller

16.

Structure and Structural Anisotropy Tsutomu Naito

PAPER, BOOKS, AND PAPER TESTING: THE ORIGINS OF PAPER (AND BOOKS) IN THE NEW WORLD Much has been written about the ``humanity'' of paper: its in¯uence on civilization as its geographical and cultural distribution has widened; the increasing diversity in its use; and its central role in providing a convenient vehicle for the acquisition, storage, and dissemination of both tangible goods (as in packaging), and these same functions applied to intangible areas, such as communication and the in®nitely diverse applications of human knowledge. It is altogether in keeping with this humanistic, universal view of paper and civilization that we have acknowledged in our jacket design and elsewhere in these volumes, the role that paper played in the waxing and waning of human culture in several ancient societies not often associated with paper or papermaking. The origin of paper as invented by T'sai Lun in China is a well-known story. The art of making paper also arose independently in Mesoamerica sometime before A.D. 660, and perhaps many centuries before that. As with their counterparts in the orient, Maya, Toltec, Aztec, and Zapotec papermakers utilized the bark ®bers of trees of the Moraceae family. They devised their own techniques for ®ber separation, washing, beating, felting, couching, sizing, drying, hot pressing, coating, and converting. They developed pigments, dyes, inks, and glues. Some of their writing was done in codicesÐhandwritten booksЮtted with covers of leather, jaguar skin, or wood, the latter often studded with decorative stones. From the few codices that survived the conquest, one senses that Mesoamerica may have been on the verge of printing when the conquistadores arrived, for the Mayas and Aztecs were already using wood, clay, and metal stamps for decorative stampings of ceramics and weavings. Paper was of tremendous importance in pre-Columbian Mesoamerica. It was used in many rituals. It was a substantial article of commerce and served as a vehicle to sustain and enlarge commerce generally. The Maya originators of paper built librariesÐguarded stone buildingsÐto house their records, documents, and sacred books, which were consulted before decisions were made in matters ranging from crop planting to war. The Aztecs, who improved on the techniques of papermaking that the Maya had started, employed paper for a much broader range of records keeping, including land surveys, engineering plans, tax rolls, and tribute lists, and xiii

xiv

The Origins of Paper

they had even begun to use the medium for communication when the ships of CorteÂs arrived at their shores. How important a role was assigned to paper in the governance of the Aztec domain can be deduced from the tribute lists that were housed in the libraries at Tenochtitlan. Areas of Mesoamerica that were under Aztec control were required to contribute substantial quantities of foodstuffs, spices, fabrics, blankets, skins, garments, shields, incense, jade, metals, ®rewood, hewn timbers, and other products of value, both natural and manufactured. The tribute lists were, of course, on paper. A handful of these tribute lists, such as those in the Codex Mendoza, have been preserved. From the lists in that codex we learn that two areas of present-day southern Mexico were required to contribute an amount of paper that seems impressive when we recall that each step of the operation, from the cutting of the trees to the carrying of the bundles of paper to the Aztec capital, was done by manual labor. The tribute towns had to deliver 16,000 resmas twice yearlyÐan annual contribution amounting to 480,000 sheets. The Aztecs did not allow any paper or other tributes to be toted to their storehouses without prior approval of the ruler's inspectors. In order to carry out their responsibilities, the inspectors had standards of quality to verify; evidently some form of quality control testing was conducted in the papermaking towns, probably by both fabricators and inspectors. Paper also played a role in communication in the empire that CorteÂs encountered in 1519Ða situation he used as an aid, after minor skirmishing, to bring about the empire's downfall. The shrewd Spaniard noted that Indian artists were sketching scenes of the ships, the military deployment of the soldiers and horses, and the equipment that he had landed. What was being painted on the pads, he reasoned, would be the messages delivered to the capital. He then arranged for a mounted drill and demonstration ®ring of his cannon, which was duly recorded by the artists. As CorteÂs expected, the renditions of the scene on the beach caused great consternation in Tenochtitlan, for the Aztecs had no knowledge of horsemanship, gunpowder, or ®rearms, and could not be sure if they were dealing with men or gods. To act on either assumption was fraught with uncertainty and danger; while Moctezuma III temporized and, with lavish gifts, tried to induce the Europeans to leave, CorteÂs learned where all the disaffections and weaknesses, civil and military, lay in the restive, tribute-burdened satrapies. The rest is history. The Aztec and all the other Indian nations of Mesoamerica were conquered militarily. Following that conquest came a spiritual assault on their customs and traditions, and a progressive disintegration of these once-proud societies ensued. It does not seem to be an overstatement to say that the loss of their paper records played a signi®cant role in that disintegration. The chroniclers of the time duly recorded the extreme anguish of the Indians as the contents of their libraries were burned. So complete was the destruction of the paper documents of the Maya ordered by Bishop Diego de Landa in 1561, that only four of the Maya codices have survived, and each of these relics has damaged and/or missing pages. Destruction of the Aztec libraries was not as total, and about 500 Aztec books were collected by a handful of individuals who recognized their value, monetary or otherwise. However, the man responsible for collecting most of them, the Chevalier Lorenzo Boturini, was shortly imprisoned on religious grounds. Partly ``torn, pillaged, and dispersed''

The Origins of Paper

xv

and partly stored with little care in a damp location for many years thereafter, the remains of Boturini's collection were ®nally auctioned off in 1804. At that time they were examined by the eminent scientist Alexander von Humboldt, who declared them to be so deteriorated that ``there exists at present only an eighth part of the hieroglyphic manuscripts taken from the Italian traveler.'' After succeeding vicissitudes and misadventures, there remain today less than 20 of the Aztec codices. Not a single Toltec or Zapotec book is known to have survived. And so the accumulated centuries of writings by the Mesoamerican peoplesÐ the paper recordÐhas largely been lost and forgotten. Gone are their maps and hieroglyphic charts, their tribute lists and tax rolls, their herbals, their land surveys, the plans for their canals, roads, buildings, and other engineering and architectural works, the records of their wars and migrations, their histories, genealogies, charts of the constellations, calendars and almanacs, and other known writings on agriculture and crops, ®shing, soils, astronomy and astrology, mythology, disease and medicine, mathematics, commerce, songs and chants, cosmogony (including the legendary History of Heaven and Earth of the Toltecs), and especially prophecy, religion, and sacred themes. On the front inside cover, we have reproduced a few pages of the most famous relict Maya book, the Dresden Codex. Although giving recognition to the Amerindian paper-making societies constitutes only a minor (but we hope interesting and thought-provoking) part of these volumes, we have made one use of the number system of the Maya in the chapter texts. Their vigesimal system (see rear inside cover design) lends itself rather ideally to the problem of setting off itemizations (for example, lists of principles to follow in test design) from the rest of the textual material in a given section or subsection. We think it ful®lls a real need in the publishing ®eld, especially as regards technical books. Any comments on this usage will be welcomed by the editors. Richard E. Mark

1 THE PAPER AND BOARD TESTING LABORATORY LUCILLE D. HAHN Champion International Corporation West Nyack, New York

I. Introduction II. A History of Testing in the Paper Industry

2

III. The Testing Laboratory A. Laboratory Type B. Outside In¯uences C. Inside In¯uences D. Precise and Accurate Data E. Sampling F. The Importance of a Quality System G. In Review

3 3 4 5 5 6 6 7

IV. Inspection, Measuring, and Test Equipment

8

V. Laboratory Environment VI. Laboratory Automation VII. Interlaboratory Reference Systems References



2

9 9 10 11

Retired. 1

2

I.

Hahn

INTRODUCTION

The original author of this chapter, Helen Schuierer, began by stating, Innumerable papers and books have been written on the theories and procedures to characterize and de®ne the physical, optical, and surface properties of paper and board. The study of wood, pulp, and the art of papermaking are without question essential for the progress and advancement of paper technology. To verify the literature, prove the technology, and qualify the industry's products, testing equipment has and will continue to be the means to numerically evaluate not only the properties but also the performance of paper and board.

What appears to be missing from the literature is the importance of identifying the laboratory that will be performing the testing. Is it a mill, is it a research center, or is it a contract lab? Will the data be used to verify or qualify the paper and/or board properties? What part does quality play? Are ballpark results good enough, or are precision and accuracy statements necessary? What about the conditions under which the product is tested? Is there a need to de®ne and then control the temperature and humidity to tight conditions? There was a time when the important issue was time; i.e., how fast can you turn this information around? Precision and accuracy were good things to have but not if they took too much time. Today, timeliness is still important. However, timeliness without the consideration of quality and cost is foolhardy. Which is more important? That depends on the customer you're dealing with. For some customers, how fast you can get the information to them may be their measure of your success. For others, how economically you can perform the testing may be the key. Still others may demand the most precise and accurate job you can do, no matter how much it costs or how long it takes. The importance of quality is recognized throughout the industry. However, quality is de®ned in many different ways depending on where you look. I choose to de®ne quality in terms of consistently meeting the needs of your customers while taking into consideration their speci®cations for timeliness and cost. Testing should not be thought of as just a job of ``number crunching.'' Successful testing is an important link in the overall papermaking process. This is true whether the quality tests are performed in a testing laboratory or right on the machine. Taking into consideration the training of the tester, the precision of the instrument used, and the accuracy achieved, testing is the indicator of whether or not the product meets the speci®ed properties of the paper or board tested.

II.

A HISTORY OF TESTING IN THE PAPER INDUSTRY

Little information is documented on the origin of paper test instruments, although the list of books pertaining to the history of papermaking is extensive. Before the advent of testing equipment, paper in general was simply made to match a previously accepted sample. Papermakers relied upon their senses of sight and touch. In essence, they were the ®rst test instruments. They used their visual perception for such properties as color, formation, opacity, and brightness. Their sense of touch was used to determine the smoothness, thickness, burst, and tearing strength. To some degree,

The Paper and Board Testing Laboratory

3

subjective evaluation is still used, but it is in¯uenced by the physiological and psychological factors of the person evaluating the paper. I doubt whether we will ever completely eliminate these human interventions, but no one can deny the value of test instruments that are unbiased and that numerically qualify or quantify a property of paper or board. Just after the beginning of the twentieth century, some enterprising companies introduced the use of test instruments for quality control. This practice was not widespread, but it was a beginning. At approximately the same time, a concerned group of pulp and paper producers realized the need for a common language within the industry. As a result of their efforts, The Technical Association of the Pulp and Paper Industry (TAPPI) was founded. TAPPI published its ®rst test methods in 1917. Within ®ve short years, the ®rst edition of the TAPPI testing methods appeared in book form. These methods went on to become universally accepted and recognized. With the input of experts throughout the industry, the TAPPI methods have been reviewed and updated on a regular basis ever since [14]. In addition, TAPPI encourages submission of new methods that can be used to de®ne and characterize a particular property of pulp, paper, or board. Each of the methods that is reviewed and adopted relates to one or more particular test instruments used to verify or qualify a speci®c property. Of all the instruments in use today, the Mullen, or bursting strength, tester is one of the oldest and most prevalent in the paper industry. It was invented by John Mullen, a papermaker of some renown. He made his ®rst burst strength tester while employed as a superintendent of the Crocker Manufacturing Company, now one of the Crocker-Burband mills. The ®rst Mullen tester was sold to the Parson Paper Co. in 1887 and is now in the Ford Museum. The basic principles of the Mullen burst testers used in the paper industry have not changed. There have been modi®cations, the most signi®cant being in the acquisition of data. Now in use are sensing systems with digital readouts that are compatible with a microprocessor or direct interface with a computer.

III.

THE TESTING LABORATORY

A.

Laboratory Type

Generally, paper and board testing laboratories may be classi®ed into research, mill, customer, and academic laboratories. Research Laboratories Research laboratories must have the capability of testing any product manufactured by their company and its competitors. The more diversi®ed the company's products, the greater the need for more varied equipment. Investigation and development of new test instruments for recommendation to the divisions of the company are other important duties of such a lab. This particular phase of work can result in increased productivity and more accurate de®nition of properties related to the end-use performance of products. It is also in such labs that 

The Technical Association of the Pulp and Paper Industry, P.O. Box 105113, Atlanta, GA 30348-5113.

4

Hahn

test results aid in the prediction of new methods, in improvement of existing products, and as a tool for problem solving. Mill Laboratories Mill laboratories are used to monitor production and to assure the quality of the product before it is shipped to the customer. Over the past few years, there have been differences of opinion concerning the future of this type of lab. One school of thought is that the labs will continue as before, whereas the other, stronger opinion is that paper will eventually be fully quali®ed on-line and the off-line labs will be used for veri®cation of on-line results and possibly product development. In the meantime, however, the mill labs continue to ensure that speci®cations are met, whether mandated by the mill, the customer, the industry, or the government. Customer Laboratories Realistically, customer labs are not interested in how the paper or board was made. The customer is concerned with how the product performs [3]. The testing done here is to ensure that product speci®cations are met. Academic Laboratories Academic labs are instrumental in proving or disproving the existing and/or proposed principles and theories of the physical characteristics of paper and board. The academic lab may also be called upon as a neutral party to help solve a problem that is either common to the industry or con®ned to a single company. New test instruments have been researched thoroughly for the industry in labs such as these. B.

Outside In¯uences

Before looking at internal conditioning requirements, it is important to consider those outside factors that could adversely affect the outcome of the test performed in the testing laboratory. For example, is the lab located in an area where it is subjected to a great deal of vibration? Vibration is a real enemy of good testing. An instrument that is subjected to constant vibration cannot be relied upon to hold its calibrated state. Drifting can often be attributed to excessive vibration in the lab. Damping elements are usually suggested to reduce or eliminate the problem. These elements include multilayered ®ber-impregnated pads, leveling feet mounted on sponge rubber, and balance tables made of marble slabs to diminish ¯oor vibrations. One laboratory with excess vibration solved its bench stability problem by attaching the bench as a cantilever to the building's reinforced concrete pillars [13]. And what about electrical stability? Can the electric lines that are feeding current to the lab be relied upon to remain consistent without spikes, interruptions, or brownouts, all of which can also throw an instrument out of calibration? If not, it may be necessary to install a line stabilizer to eliminate the possibility of adverse electrical in¯uence. Do any of the instruments in use require the use of water or air? If so, then it is equally essential that the water and/or air used be clean and free of any type of contaminants that could affect the instrument. Filters, strategically placed for the greatest effect, should be used and replaced at regularly scheduled intervals. Contaminants occur at all locations, including research centers and universities, and no matter how clean you think your air supply might be, a power failure,

The Paper and Board Testing Laboratory

5

mechanical breakdown in lab controls, or repair on the building's main air lines will introduce some or all of the contaminants present in a mill or plant atmosphere.

C.

Inside In¯uences

Where is the lab located? Is it in a research center or in a mill? If it's in a mill, where in the mill is it? Is it a stand-alone lab away from the machine ¯oor, or is it just a little room located very near the machine? In other words, if the quality of testing is indeed paramount, is it possible to control the conditions under which testing is performed? Changes in temperature and humidity, even slight changes, can often affect the test results. For example, an increase in temperature without an increase in humidity cause tensile and burst characteristics to increase. An increase in humidity can cause tear strength and even optical parameters such as gloss to change [8]. It seems clear that even in the most adverse conditions the control of temperature and humidity is the ®rst step toward achieving the precision and accuracy required in the production of quality testing. Chapter 3 of this book, which describes the conditioned test atmospheres environment, should receive the undivided attention of anyone responsible for a paper-testing facility. Maintaining TAPPI conditions (23  1 C; 50  2% relative humidity) is by far the most exacting, trying, and frustrating requirement in an ef®cient, reliable test laboratory. There have been times when this requirement appeared well under control and then situations occurred that required changes in the room's system. The following are examples of changed situations. The updating of manually operated equipment with automated modern equipment and the addition of new instruments to better characterize paper and board. An increase in the number of technicians in the test laboratory and/or the number of outside personnel using the laboratory facilities. Renovation and/or expansion.

D.

Precise and Accurate Data

Webster describes a test as ``a means of examination, trial, or proof.'' The ``means'' mentioned by Webster can be de®ned as the instrumentation and the procedure used to attain the desired test results. However, using the proper instrument and following a proper procedure is not enough to ensure that the information obtained is of the quality desired. Two areas go hand in hand in ensuring successful testing of any process or product: precision and accuracy. Precision deals primarily with the instrument in use and its ability to reproduce test results over and over again. It is de®ned as the agreement between numerical values of two or more measurements that have been made in an identical fashion. Data may be precise and reproducible but yet be very inaccurate because of methodology, technique, calibration, or instrument differences. It is in the area of accuracy that the testers make their greatest contribution to successful testing. The term ``accuracy'' denotes the nearness of a measurement to its

6

Hahn

accepted value and involves a comparison to a true or accepted value. Accuracy can be achieved only if technique, calibration, and methodology are all correct [8]. To show the importance of these two factors in any laboratory, whether paper or board, consider the following example. Testing has been performed on an instrument that has not been properly calibrated. The proper procedure is being used, and the technique of the tester is not in question. The resulting data will be precise; however, they will not be a true measure of the property in questionÐthey will not be accurate. Another scenario: The instrument has been maintained properly and the calibration veri®ed; however, the proper procedure and/or technique has not been used to obtain the data. Again, the data will be precise but not accurate. E.

Sampling

Successful testing is also dependent on having a suitable sample for testing. Samples that are the wrong size, dirty, creased, or wrinkled cause errors in testing. When test samples are gathered, they must be of the proper size and cut properly, with the machine direction clearly identi®ed. They must be clean, free of wrinkles, and delivered to the laboratory in a timely manner. If any of these factors is not controlled, errors can be expected in the data produced. If the lab that is to perform the testing is located remote from the paper machine, add to these factors the importance of packaging the samples in a way that will ensure their integrity in transport. F.

The Importance of a Quality System

For many years, the importance of controlling and maintaining the quality of material produced in the laboratory (whether paper or board) has been gaining in acceptance throughout the paper industry. Those wanting to make a public commitment to quality certi®ed their testing to a guideline provided by the National Voluntary Laboratory Accreditation Program (NVLAP) or a similar program provided by the American Association for Laboratory Accreditation (A2 LA). Around 1993 both NVLAP and A2 LA adopted a program based on a document provided by the International Organization for Standardization (ISO), ISO/IEC Guide 25 [9]. ISO is a worldwide federation of national standards bodies, at present comprising 107 members, one in each of 107 countries. Their mission is to promote the development of standardization and related activities throughout the world. Their work results in international agreements that are published as international standards [11]. In 1987, ISO adopted a series of standards identi®ed as the ISO 9000 Standards Compendium [10]. The series of standards include the ISO 9001, ISO 9002, and ISO 9003, all of which can be used as a model for the installation of a quality management system. The ISO/IEC Guide 25 is similar to ISO 9003, and it can be used speci®cally in a laboratory environment. Additional information can be obtained from the American Society for Quality Control (ASQC).y ASQC is the administra

ISO Central Secretariat, 1 rue de VarembeÂ, Case postale 56, CH-1211 GeneÁve 20, Switzerland. y American Society for Quality Control, P.O. Box 3005, Milwaukee, WI 53201-3005.

The Paper and Board Testing Laboratory

7

tor of the United States Technical Advisory Group (TAG) to ISO/TC176. The ISO Technical Committee 176 (ISO/TC176) is made up of members from around the world. Members of this committee are the persons with ultimate responsibility for reviewing and updating all of the ISO 9000 and related documents. Registration to the universally recognized ISO 9000 standard began in Europe. On October 15, 1991, Westvaco Corp. registered the ®rst paper-related site in the United States [6]. Since then over 400 sites related to paper or allied industries have been registered in the United States alone. The total number of sites, covering all industries, is in the thousands. I wouldn't want to give the impression that installing and registering your quality system is the only way to assure quality in a laboratory. When it comes to installing a system, the standard to use and/or the decision to register should not be the deciding factor. The important factor is that you make a sincere and long-term commitment to producing quality. Once the commitment is made, either Guide 25 or ISO 9000 can be used as the model for upgrading your current system or designing a new system. Through the conscientious use of a good quality system, the maintenance and calibration of test instruments, the control of room conditions to desired speci®cations, the capability of the testers performing the tests, the consistency of technique, and the procedures in use are all ensured. Further, through the use of corrective action and an internal audit process, which can be modeled after the requirements described in the ISO 9000 standards, maintenance and control of the system itself are ensured [10]. This aids in ensuring continuous improvement and customer satisfaction.

G.

In Review

Many areas require consideration during the design and equipping of a good paper or board testing laboratory (Table 1). Yes, the instruments to be used are of primary importance. However, so are the lab conditions, test capabilities, maintenance and calibrations, and the quality system that de®nes the process used to ensure that only precise and accurate data are produced [2].

Table 1 Questions to Ask When Equipping a ``Good'' Testing Laboratory Are the conditions under which the testing is performed de®ned and maintained? Is the lab situated in an area where it will not be subjected to undue (line) noise, vibration, or contamination that could adversely affect the test instruments? Are the personnel who are performing the tests trained by an authorized trainer, and is the training documented? Are the tests performed according to documented procedures to ensure consistency? Are the instruments maintained and calibrated using a formally de®ned procedure and on a regular schedule? Are the resulting actions recorded? Are corrective actions and internal audit functions part of normal routine for the laboratory? Is there evidence of continuous improvement?

8

IV.

Hahn

INSPECTION, MEASURING, AND TEST EQUIPMENT

The preceding sections were written with the assumption that we are dealing with an already existing laboratory. But what of the lab that is just in the process of being built? Once the location of the lab has been established and the outside and inside in¯uences on good testing have been controlled, it is time to decide what type of inspection, measuring, and/or test equipment is required to perform the testing. However, even this is not a simple matter. There are still some questions that need to be answered. For example, is the lab going to qualify the product? Or will the lab be used to verify the results obtained through the on-line instrumentation? A few years ago there was very little debate about the lab's purpose. The lab de®nitely quali®ed the paper, and the on-line instrumentation, where it existed at all, did little more than generate a whole lot of data that no one bothered to look at. Although it was true that some of the on-line instruments were equipped with actuators that caused changes to be made to the process, very few if any of the data produced were used for anything. Today there appears to be a change occurring. More and more mills are looking at the data produced on-line as a ready source of real-time information that can be used to qualify the product. This does not diminish the importance of the testing laboratory, but it does somewhat change its function. I believe that in the not too distant future the testing laboratory will be used almost exclusively for the veri®cation of on-line instrumentation and for the development of new paper and board products. Only time will tell if I am right. For example, let us assume that the purpose of the lab in question is quali®cation of the product. The results obtained will be used to ensure that the speci®cations set down by the customer are met. Of course, this does not prevent the use of the lab as a veri®cation source as well. The next step is to determine the type of product produced. Is it board or paper? If it's paper, is it newsprint, ®ne paper, kraft, specialty grade, etc.? Is it bleached or natural? So many questionsÐbut they are all important in determining the type of instrumentation that will be required to determine the product's essential characteristics. What tests are critical to your process? Are they those that determine the optical properties such as brightness, opacity, or gloss? Or are they strength properties such as tensile, tear, or burst strength? What about internal properties such as internal bond strength, sizing, or water vapor transmission? And let's not forget about the surface properties such as smoothness, gloss and scuff/abrasion. Over the past few years, the list of possible tests has grown and grown [1] As technology advanced and new grades came on the scene, the capability to measure newly identi®ed properties became a reality. Old instruments have been upgraded or replaced by more precise electronic, state-of-the-art versions. When a new property was identi®ed, an instrument manufacturer was quick to develop and bring to market the right instrument to perform the testing required. This makes a complete listing of instrumentation used in the paper and board laboratory less meaningful for reference purposes. However, some of the paper and board characteristics can be found in this handbook. Individual chapters should be consulted to determine the proper test method and instrumentation to be used. For other test procedures for which standardization exists, the responsible organization,

The Paper and Board Testing Laboratory

9

e.g., TAPPI or ISO, issues yearly updates of the methods used [11,14]. Each method clearly de®nes the type of instrument in use, the properties being tested, the instrument calibration and maintenance necessary, and even an indication of the precision and accuracy that can be obtained through their use. The reality is that the important tests, those that are critical to your operation, are those that will best de®ne the physical, optical, and chemical properties of your product, and only you can decide what they are. V.

LABORATORY ENVIRONMENT

The type of laboratory environment should also be dictated by need. There are many manufacturers of laboratory furniture and other laboratory equipment, who are ready, willing, and able to help you select the items you will need to equip your lab. There are, however, a few considerations that you should be aware of before you begin the process. Enough room should be delegated to each instrument to ensure adequacy of air displacement, sample preparation, and, if appropriate, the placement of a computer interface device. The tester's position (sitting or standing) should be taken into consideration when determining bench height, depth, and length. Placement of the laboratory benches should allow for adequate aisle space. This is partially dictated by the requirements of the U.S. Occupational Safety and Health Administration (OSHA). However, a good rule of thumb is to allow suf®cient room for someone to pass down the aisle while a tester is seated at the bench. Ergonomic considerations, for the health and safety of the people working in the lab, must be taken into consideration. Suf®cient storage for maintaining sample stability must be provided. Noise, whether audible or contained within the electrical lines, must be controlled for the comfort and safety of the testers and instruments. Lighting ®xtures should be selected on the basis of room size, ceiling height, bench placement, and test instruments. Any light ®xtures chosen should reduce or eliminate the possibility of glare, which causes eye fatigue and discomfort for the testers. VI.

LABORATORY AUTOMATION

The next question to answer is whether or not the laboratory should be automated. In this day of computerization, this is no longer a question, automation is a given. There was a time when, to move a lab out of the dark ages, all that was required was to upgrade the instruments from manual to electronic. Today, all instruments come with some kind of internal computer interface capability. The bene®ts to automation are obvious: Increased productivity Fast turnaround of data and information

10

Hahn

Networking of information to other locations Instant statistical evaluations, etc. The real question is whether or not the automation should be accomplished through in-house capability or through the purchase of a commercially available laboratory automation system. Once again, the decision is yours. No matter what your need, from the automation of a single instrument to a completely robotized lab, there is a system manufacturer out there to help you.

VII.

INTERLABORATORY REFERENCE SYSTEMS

Prior to 1969, the resolution of problems attributed to test results, test instruments, and test methods among different test laboratories was possible but the solutions were often time-consuming and costly. In 1969, The National Bureau of Standards (now designated the National Institute of Standards and Technology) and the Technical Association of the Pulp and Paper industry (TAPPI) developed an interlaboratory program for paper and paperboard testing [12,14]. Since 1971, Collaborative Testing Services, Inc. (CTS) has operated the Collaborative Reference Program for Paper and Paperboard with technical guidance from TAPPI. Both the Process and Product Quality and Container Board divisions of TAPPI support a Collaborative Testing standing committee whose function is to oversee the programs expansion and to make general recommendations concerning the program. With more than 400 organizations around the world participating in these tests, this program has become one of the largest of its kind. This allows laboratories to compare the performance of their testing with that of other participating laboratories and provides a realistic picture of the state of paper testing for TAPPI. The global program is designed to demonstrate real-world lab performance and to assist the participating members in achieving and maintaining quality assurance objectives. The Collaborative Reference Program is not and does not claim to be a universal cure for all the problems in the area of physical testing of paper and board products, but, if used properly, it can aid in the satisfactory resolution of many problems. For example, each report that is issued by CTS contains the Laboratory Mean, the Grand Mean (lab means of all participants), the BetweenLab Standard Deviation, and the Comparative Performance Value (CPV) [5]. By assuming the Grand Mean to be zero, the CPV can be charted and used to determine long-term performance and trend analysis of the test(s) in question. This information can then be used, for example, to determine whether complete overhaul or replacement of a test instrument is justi®ed [4]. In Europe, the interlaboratory reference system is that of the European Confederation of Pulp, Paper and Board Industries (CEPAC). Recognized institutions in England, France, Germany, Italy, and the Netherlands serve as coordinating laboratories. Each location is responsible for determining the provisional values on samples for speci®c test methods. The samples are then distributed to the partici

Collaborative Testing Services, Inc., P.O. Box 1049, Herndon, VA 22070.

The Paper and Board Testing Laboratory

11

pants. Analyses and reports are issued [7]. Similar systems are in effect elsewhere within the paper and board testing community. An interlaboratory reference program provides documentation of your abilities to test accurately in comparison with the other participants in the program. The information can help assure your manufacturing divisions, suppliers, and clients of your capabilities in maintaining an ef®cient testing laboratory. As with any worthwhile endeavor, the success of this type of program is dependent upon everyone involved, as its name implies.

ACKNOWLEDGMENT This chapter is a revision and update of the original chapters, ``Equipping the Testing Laboratory'' and ``Interlaboratory Reference Systems,''written by Helen R. Schuierer.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.

Bichard, D. W. (1996). Physical testing: A fresh look at old habits. Tappi J. 79(10):199± 202. Church F. (1971). The operations of an instrument department. Paper Technol. 12(2):T52±T65. Clarke, B., and Minshall, H. (1971). Raw materials testing from the customer's point of view. Paper Technol. 12(4):T131±T140. Collaborative Testing Services (1996). Pro®ciency & Interlaboratory Testing. Paper and Paperboard Report, Collaborative Testing Services, Inc., Herndon, VA. Czyryca, C. J. (1997). Interlaboratory testing: A quality assurance tool for the real world. Tappi J. 80(9):97±101. Ferguson, K. H. (1992). Westvaco (Corp.) Covington (VA) mill. First in U.S. to attain certi®cation under ISO 9002. Pulp Paper 66(2):76±81. Franke, W. (1977). CEPAC European test instrument calibration service. Operation and results to data. Wochenbl. Papierfabr. 105(20):834±837. Hahn, L. D. (1990). Testing Guidebook. TAPPI Press, Atlanta, GA. International Organization for Standardization. (1990). ISO/IEC Guide 25. General Requirements for the Competence of Calibration and Testing Laboratories. ISO Central Secretariat, Geneva, Switzerland. International Organization for Standardization. (1994). ISO 9000 Quality Management, ISO Standards Compendium (ISO Forum Library). 5th ed. ISO Central Secretariat, Geneva, Switzerland. International Organization for Standardization (1995). Paper TechnologyÐPaper and Board. Selective list of international standards 1995-08-07. ISO Central Secretariat, Geneva, Switzerland. National Bureau of Standards (1979). NBS Collaborative Reference Programs 1979±80. National Bureau of Standards, Washington, DC. Stanton, P., and Oxley, S. (1979). A simple, inexpensive balance ``bench.'' Lab. Pract. 28(9):934. Technical Association of the Pulp and Paper Industry (1996). TAPPI Test Methods 1996±1997. TAPPI Press, Atlanta, GA.

2 ON-LINE TESTING OF PAPER GARY M. SCOTT Empire State Paper Research Institute State University of New York College of Environmental Science and Forestry Syracuse, New York

I. Introduction A. History of On-Line Sensors B. Philosophy of On-Line Testing C. Overview of the Chapter II. Principles of On-Line Testing A. Comparison to Laboratory Testing B. Characteristics of On-Line Testing C. Static versus Dynamic Measurements D. Calibration of On-Line Sensors III. On-Line Sensors A. Structural Measurements B. Optical Measurements C. Strength Tests D. Paper Defect Detection E. Manufacturers References

I.

13 14 14 15 16 16 19 22 22 25 25 38 40 42 43 43

INTRODUCTION

All testing of paper is ultimately done for the same purpose: To ensure that the paper meets the end-use requirements of the product. This testing, which can be both offline and on-line, provides the papermaker with the information needed to operate the machine and keep quality acceptable. At the beginning of the twentieth century, papermaking was an art, and testing depending mainly on the skill of the paper13

14

Scott

maker. He would feel, tear, shake, and fold the sheet and hold it up to the light to determine its physical properties. Based on his assessment, he would make changes in the operation of the paper machine. Gloss, color, and formation would be determined by visual evaluation. Papermakers soon developed laboratory tests to measure what they had previously done with their hands. The tests would still tear, fold, and observe the sheet but with the objectivity of a standardized test. The papermakers still used their skill and knowledge to control the paper machine based on these measurements. However, these tests took time, and there was still a delay between testing the paper and taking corrective action to ®x any problems. It was thus desired to bring the test onto the machine itself. With this began the on-line testing of paper [29]. A.

History of On-Line Sensors

On-line sensors have been in development over the past three or four decades. A great deal of literature on their development has been published in the various trade and research journals. This chapter does not attempt to cover the entire history of sensor development in the paper industry, only to mention some of the key developments. References to historical articles can be found in Ref. 6. The earliest on-line sensors were developed in the 1950s. The ®rst to be developed were stationary beta gauge sensors that measured an area about a foot in the cross direction (CD) and 2 in. in the machine direction (MD), with measurements being averaged over several minutes to eliminate short-term variations. With these earliest sensors, CD variations could not be measured and only long-term trends in the MD were detectable [17]. Moisture sensors were soon to follow to basis weight sensors. These earliest gauges were based on both conductivity and radio-frequency (RF) methods. It was not long before the sensors were being mounted on tracked frames so that the entire width of the sheet could be measured. The 1960s brought caliper and formation testing on-line. The development of on-line optical property sensors began in the 1960s and continues to the present day [9,10,16,32,33,46,47,49]. Developments in the past decade include the ability to measure paper strength properties on-line using nondestructive techniques. The early beta gauges used for measuring basis weight were used only for process control and not for quality assurance. That is, although the changing signal could be used to indicate what control actions were needed, relating the measurement to the actual basis weight required careful calibration. This is a characteristic of most on-line measurements: They do not directly measure the property in question but rather some other property correlated with the desired property [29]. Researchers continually make improvements in the ability of sensors to measure key paper properties and to relate these values to the end-use properties important to the customer. B.

Philosophy of On-Line Testing

On-line testing and control of the paper machine is becoming especially important for many reasons. As machines become wider and faster, the potential for making offspeci®cation paper increases. This leads to the need both for tighter control of the machine and for specialty machines and quicker grade transitions [30]. At the same

On-Line Testing of Paper

15

time, customers demand greater uniformity and better quality paper. Printers, for example, expect the paper to have uniform caliper and surface smoothness. The infrequent testing that can reasonably be done in a laboratory is not suf®cient to meet these needs. On-line sensors, coupled with feedback control, continuously test the paper and automatically adjust the machine to produce paper to the customer's speci®cations. Several advantages of using on-line sensors are summarized in Table 1. On-line sensors allow papermakers to make a better quality product at less expense. The capital costs of installing on-line systems are quickly paid back in improved productivity. In addition, on-line testing, although not eliminating the need for laboratory testing, reduces the labor requirements for papermaking. As on-line systems become more stable, there will be less reliance on off-line laboratory testing. However, there will still be a need for off-line testing according to strict standards to ensure that the paper machine is making the correct product [39]. With the continuing improvements in sensors, machine operators are coming to expect more. The more timely the information that is received, the quicker the response to changes in the paper machine and the less off-speci®cation paper produced. In some cases, where sensor technology is mature, it is even possible to ``ship by the gauge'' and forgo regular laboratory testing of paper (e.g., basis weight and moisture) [41]. The heart of on-line control is the sensor, and control of the process can be no more accurate than the quality of the sensor. Thus, much research in the last decade has focused on improving the accuracy, resolution, and speed of the measurements [42]. On-line testing is growing rapidly with the development of new on-line sensors and increased computing power to analyze the resulting data. The Technical Association of the Pulp and Paper Industry (TAPPI) is an industry association with committees and subcommittees that deal with the many aspects of papermaking from harvesting to converting. The Paper Machine Sensors Subcommittee of the Process Control Committee is part of the Process Control, Electrical and Information Division. This subcommittee includes members from many of the major suppliers of sensors and other process control equipment [41]. C.

Overview of the Chapter

The remainder of this chapter is divided into two main sections. Section II discusses the use of on-line sensors in general, covering such topics as their calibration and Table 1

Advantages of On-Line Testing of Paper

Faster process feedback Greater test standardization Increased paper production rates Improved paper quality Improved uniformity of paper quality Reduced off-speci®cation material Faster grade changes Reduced material costs Lower laboratory testing costs

16

Scott

standardization, their comparison to laboratory testing, and interpretation of the data. Section III discusses in detail the various sensors that are used in the paper industry. Both optical and physical properties are discussed, and a list of sensor manufacturers is provided.

II.

PRINCIPLES OF ON-LINE TESTING

Many mills are moving toward measuring paper properties on-line rather than in the laboratory. Papermakers use on-line testing for a variety of reasons, and they see several advantages to its use. The characteristics of on-line testing also have several disadvantages or at least differences from laboratory testing. It is these advantages and disadvantages that need to be carefully considered when using on-line testing. The signi®cant differences can include the frequency of testing, sample preparation, testing conditions, or even what property is being measured. As discussed in Section I, mills are turning toward greater use of on-line testing for many reasons. However, some level of laboratory testing will still be needed to verify the proper operation of the sensors, especially for those not directly measuring the paper property. For example, using ultrasonic measurements to infer the strength of paper requires careful calibration with laboratory measurements. On-line testing is extremely important for process control. Basis weight, moisture, and caliper are the most commonly measured properties for CD and MD control. Recent developments have expanded this list to include smoothness, gloss, and coat weight [42]. For process control applications, it is not necessary to have the sensor precisely calibrated, because control actions are calculated as deviations from a setpoint. As long as the machine is meeting the grade speci®cations, the control need only maintain the paper property at its current measured value. This section ®rst compares the characteristics of laboratory testing to those of on-line testing, highlighting some of the key differences. Then some of the characteristics of on-line testing and the different manner in which measurements are made are described. An important characteristic of on-line testing is the ability to measure both CD and MD variations. Measurements can be made in both static and dynamic modes, which is important in the calibration of on-line sensors.

A.

Comparison to Laboratory Testing

Many technical associations publish standards for the testing of paper. In the United States, the most used standards are the test methods published by TAPPI [44]. These methods were developed to provide standardized methods of testing paper so that results could be compared between laboratories and independently veri®ed. These standards are very strict in the description of the test methods, including the preparation of the sample, the environment in which the testing is to be done, and the number of samples needed for each measurement. On-line testing, which is carried out on the machine with a moving web, differs signi®cantly from these laboratory methods.

On-Line Testing of Paper

17

Testing Frequency and Delay One of the ®rst differences between on-line and laboratory testing is that a greater area of the paper can be tested by an on-machine sensor than by laboratory testing. For laboratory testing, a single sample taken from the reel is assumed to be representative of the entire reel. From a machine operational standpoint, it is often not practical to take more than one sample per reel. Online tests measure the entire reel and present information on the average property as well as on variation within the reel. The greater immediacy of the information allows corrective action to be taken sooner, thus bringing the machine back into speci®cation [6,13]. Table 2 details the frequency and the testing delay for various testing methodologies. Laboratory testing can usually be done only infrequently with a signi®cant delay from the time the sample is taken until results are available. At the other extreme, a single-point on-line sensor (one that does not scan the sheet) can give very frequent measurements (potentially more than one per second) with only minimal delay. The major drawback to single-point sensors is that since they do not scan the sheet they are not measuring a representative sample of the sheet in the cross direction. By continuously moving across the sheet, a scanning sensor can calculate an average value for the measured property. Information regarding the average property across the sheet is limited by the scanner's velocity as it crosses the sheet. Thus, changes in the property in the machine direction are available every 30±120 s (depending on the time for one scan) and are delayed by the same amount as the sensor waits for the entire scan to be completed before reporting the results. The scanning sensor also provides information about the changes in the sheet in the cross direction. Of course, the scanner path is actually a diagonal path, as the sheet is moving beneath the scanner. The implications of this path are discussed in further detail in Section II.B. The difference from one measurement to the next represents changes in both the MD and CD as the scanner moves and the sheet moves. To get good representative data for the CD pro®le, several scans, sometimes up to 10±15, must be averaged. Thus the frequency is lower and the delay is longer for CD information, even though the data are coming from the same scanner. The frequency of CD information can be improved to the level of the MD information by using a moving window of scans to be averaged, with the oldest scan in the window being replaced by the most recently completed scan. However, the delay is still present as the information being presented is, on average, several scan cycles old. Recent sensor development has concentrated on full-width or full-sheet measurements. Scanning measurements, although covering the full width of the sheet, cover less than 1% of the total area of the sheet and thus still represent a sampling Table 2 Testing Frequency with Different Testing Protocols Type of testing Laboratory testing Single-point sensor Scanning sensor (MD) Scanning sensor (CD) Full-width sensor (MD/CD)

Frequency (Time between samples) 1h 1±10 s 30±120 s 1±5 min 1±30 s

Delay 5±60 min 1±10 s 30±240 s 1±10 min 1±60 s

18

Scott

process [3]. Full-width measurements can be thought of as scanning with an in®nite scanner speed. In full-width measurements, the entire width of the sheet is measured almost simultaneously, giving a true CD pro®le. The limitation then becomes that of how often the measurement can be made in the machine direction. In many cases, because the entire sheet is being measured, the limitation becomes a matter of the speed at which the large amounts of data from the sensor can be processed and displayed. Testing Environment Laboratory testing is done under strictly controlled conditions to eliminate the effect of the environment on the measurement. Paper, being a hygroscopic material, is very sensitive to ambient moisture. On-line sensors are not in a controlled environment. Thus there are many factors that can in¯uence their accuracy. Under laboratory conditions, the physical properties of paper are measured under constant humidity after the sheets are conditioned to the testing atmosphere. On-line, this conditioning cannot be done, and the measurement becomes a function of the ambient relative humidity and temperature. Dirt and sensor contamination are a major concern. The sensor is often exposed to a harsh uncontrolled environment. Dust and dirt can settle on it, thus causing errors, especially in sensors that involve interaction with light. As dirt builds up on it, the sensor will begin to read high as the dirt and dust cause additional scattering. If this sensor is being used for process controlÐfor example, a beta gauge to control basis weightÐthese high readings will cause the actual basis weight of the sheet to drop as the controller compensates for the high readings. Standardization cycles, in which the sensor is taken off the sheet and a standard is inserted periodically to calibrate the sensor, can help compensate for this contamination. Other material can also affect the performance of sensors. Condensation can greatly affect both nuclear and infrared sensors [25]. Some sheet components can also greatly affect the measurement. For example, ®llers and carbon black can change the correlations for nuclear and infrared sensors. Mills, especially recycle mills, that have changing feedstocks can have signi®cant dif®culties keeping the sensors updated. Static electricity is very common around the dry end of a paper machine. Static electricity is generated by the rapidly moving web in contact with rolls and other paper machine equipment in much the same way as a Van de Graaff generator. As with all electronic equipment, the reading of on-line sensors can be corrupted by static electricity. Beyond just corrupting data, a severe discharge of static electricity can destroy the sensor circuits. Static eliminators need to be a key part of all on-line testing installations [25]. These can be as simple as grounding wire in contact with the sheet or take the more complicated form of grounded metal brushes discharging the static buildup on the sheet just before the sensor. Statistics of Laboratory Testing Researchers have understood the statistical analysis of laboratory testing for many years [21]. The variability of a test measurement can be divided into several components. In general, the variation can be partitioned into the variation between laboratories, V…) (when more than one laboratory is being used), and the variation within the laboratory. The withinlaboratory variation can be further partitioned into variation related to the number of replicates that are measured, V…†, and the irreducible or random variation, V…†. Replication can reduce the magnitude of V…†, and standardization between labora-

On-Line Testing of Paper

19

tories can reduce V…†. However, there still remains a certain amount of variation inherent in the process. The number of replicates speci®ed in the TAPPI test methods reduces the replication variation V…† to be on the order of the random variation, V…†. Thus, further reduction in the test method variation would require procedural changes to reduce V…†. On-line testing, because it is automated, is able to deal easily with many samples. However, this can reduce only one of the three components of the variation previously discussed. In addition, variation in the cross and machine directions can also be analyzed using the information obtained from scanning and full width sensors. This analysis is discussed in more detail in the following section under the heading ``Real-Based versus Scan-Based Measurements.''

B.

Characteristics of On-Line Testing

From the foregoing discussion, we see that on-line testing is signi®cantly different from laboratory testing. The number of measurements that can be taken is greatly increased, and with the newest sensors the full sheet can be inspected. Direct vs. Inferred Measurements The on-line sensor is often not measuring the same property as the laboratory test method but rather inferring that property from some other measurement. On-line sensors often depend on correlations with laboratory testing to relate their values to standard tests. For example, in the laboratory, basis weight is measured gravimetrically by measuring the mass of a known area of paper. Beta ray adsorption is the most common basis weight measurement. Reel-Based vs. Scan-Based Measurements Before the advent of on-line sensors, the paper from the machine was tested only at the end of each reel, and the skill of the operators maintained the proper operation of the machine between these tests. Furthermore, the results of the test often had a certain amount of delay involved as the samples had to be taken, sent to the laboratory, and tested and the results then relayed back to the operators. The most common method for measuring both MD and CD variations in a property uses a scanning-based system. In this type of system, a single sensor traverses back and forth across the web. Thus, sequential measurements from the sensor represents a combination of changes in both the machine and cross directions as both the scanner and the sheet are moving. Figure 1 shows the effect of scanning speed on the path of the sensor on the sheet. If the sensor were to move in®nitely quickly back and forth across the sheet, true CD pro®les could be obtained (scanning pattern 1). At the other extreme, if the sensor is ®xed at one location across the width, true MD measurements can be obtained, but only at that particular point (scanning pattern 4). Because in®nite scanning is not possible, scanning patterns such as patterns 2 and 3 result. The return of the sensor to its starting point causes a zigzag scanning pattern to be seen. The maximum practical scanning speed of approximately 25 m/min coupled with machine speeds on the order of 500± 1500 m/min results in a scanning angle across the sheet of 0.2±1:0 with respect to the machine direction. Thus for each unit of CD movement by the sensor, the sheet will move one or two orders of magnitude further in the machine direction [6].

20

Scott

Fig. 1 Scanning patterns on a sheet for traversing sensors. The actual path of the sensor depends on the speed of the sensor relative to the speed of the moving web.

Since the sequential measurements of the sensor represent changes in both the machine and cross directions, the control computer must statistically analyze these data to give both MD and CD pro®les of the sheet. In practice, the average value from the sensor from one scan to the next is used to characterize the MD changes. Averaging the readings at the same sheet CD position over several scans gives the CD pro®le for the sheet. Thus, the response time for MD changes can be three to 10 times faster than for CD changes, depending on the number of scans (Table 2). The data from a scanning sensor can be rigorously analyzed, as has been done by Spitz [43] and subsequently by Brewster [6]. Let us say that data have been gathered from a scanning sensor with the cross direction divided into n ``slices,'' with a total of m scans taken. Thus the point xij is the measurement for the jth position of the ith scan. From this we can de®ne the scan average as x i ˆ

n X jˆ1

xij

for each of the m scans. The average value at a particular CD position j is given by y j ˆ

m X iˆ1

xij

The average over the entire data set is given by x ˆ

m X iˆ1

x i ˆ

n X jˆ1

yj

Assuming that the variations in the measurements are random, as is the case for all variance analysis, we can divide the variation into two components: vbs , the between-scan variation, and vws , the within-scan variation. The between-scan variation consists primarily of MD variation, which is in phase across the sheet and is given as vbs ˆ

m 1X …x m iˆ1 i

x †2

On-Line Testing of Paper

21

The within-scan variation, given as vws ˆ

m X n 1 X x mn iˆ1 jˆ1 ij

x i


2

consists of four different types of variation: vcd : vmdi : vmdo : vr :

CD variation MD variation in phase across the sheet MD variation out of phase across the sheet Random or residual variation

Unfortunately, it is not possible to distinguish between all these sources of variation, because some are confounded with each other owing to the geometry of the scanning process. However, if we combine the last three sources of variation into vr , a residual variance, it is possible to distinguish between this combination and the CD variation, vcd . The CD variation can be calculated as vcd ˆ

n X m y n…m 1† jˆ1 j


2

x

1 m

1

vws

when m and n are again the number of scans and the number of CD measurements, respectively. Note that for a large number of scans, this reduces to vcd ˆ

n 1X …y n jˆ1 j

2 x†

The total variance, vt , including both the between-scan variance and the within-scan variance, is given by vt ˆ

m X n 1 X x mn iˆ1 jˆ1 ij

x


2

The residual (or random) variance is given by vr ˆ vt

vbs

vcd

These variances, calculated from the equations given above, can be statistically analyzed using standard statistical techniques. Speci®cally, an F statistic can be calculated for the between-scan variation (MD variation: F ˆ vbs =vr ; CD variation: F ˆ vcd =vr ) and the statistical signi®cance of the variations calculated. These types of calculations, especially the magnitude of the residual variance, indicate the degree of randomness in the system and the expectations for control for the process. Systems with a large residual variation cannot be controlled to the same degree as systems with less variation. In fact, a large residual variation may make both the MD and CD variations dif®cult to identify statistically. A method of statistical analysis of the MD and CD measurement using a spreadsheet is given by Deodhar [14]. CD Resolution On-line sensors take many more measurements than are possible with conventional laboratory techniques. Most systems divide the width of the sheet into discrete segments, with a measured value for each segment (as was done in the analysis in the previous section). Several factors determine the appropriate resolution

22

Scott

for a particular application. First, for those sensors that depend on radiation sources such as beta gauges, increasing the resolution increases the amount of noise due to the random nature of decay. The measurement must be averaged over a certain amount of time to reduce this noise. For beta gauges measuring basis weight, the sampling rate can only be 0.1±1 Hz [12]. The CD resolution is determined on the basis of the traversing speed of the sensor, but it can be on the order of 10 cm. Optical and camera-based measurements depend on the number of cameras across the web and the speed at which the machine is running. Typically, the CD resolution limit is around 0.2±0.4 mm and the MD resolution is approximately 0.6 mm [20]. A practical limitation to the CD resolution is the number of actuators that are available to control the measured property in the cross direction. Quite often, this determines the number of ``slices''that are reported by the on-line system. In this way, each measurement can be paired with the corresponding actuator at that position. Typically, the actuators for CD control are spaced on the order of 7.5 cm apart. Scan-Based vs. Full-Width Measurements One of the latest developments in on-line testing has been the development of full-web testing. Traveling scanners, while providing information in both the cross and machine directions, can not provide an instantaneous cross section of the sheet. Rather, the scanner provides information that is a convolution of both CD and MD variations that has to be computationally separated for control purposes. The development of full-width scans by several manufacturers allows much better CD control than the scanning systems. Shorter delays in deconvoluting the CD and MD information vastly improve the control, especially at the wet end of the paper machine [30]. C.

Static versus Dynamic Measurements

One of the major objectives of on-line testing is to reduce the delay between the testing of the paper and the control action taken to correct deviations. Laboratory testing of reel samples for the purpose of quality assurance is a static type of measurement. Although the measurements may be very accurate and performed according to strict standards, they are of minimal use in the minute-to-minute operation of the paper machine. On-line testing brings the product measurements to the machine. The laboratory and the measurements become part of the machine, changing the operation of the machine to make a better product. Thus the measurement becomes dynamic [11]. This difference in measurement is important in the calibration of online sensors. D.

Calibration of On-Line Sensors

Calibration of the sensors is an important part of the use of on-line measurements. Table 3 summarizes some of the reasons that make this calibration necessary [6]. Since the sensors are often not directly measuring the property of interest, care must be taken in calibrating them if they are to truly correlate with laboratory measurements. Two methods of calibrating the sensor are possible. The ®rst uses prepared samples of a known value for the property being measured. This is called static testing, indicating that the sample is not moving through the sensor at the

On-Line Testing of Paper

23

Table 3 Reasons for Sensor Calibration To conform to customer requirements, who probably use the standard test methods for their quality control To assist in optimal machine operation To reduce the need for routine off-line testing

time of calibration. Dynamic calibration, on the other hand, involves sampling the sheet after it has moved through the sensor and making measurements on these samples. These two methods are discussed below. There are many possible sources of error in sensor calibration. Knowledge of these particular sources can help the operator troubleshoot the sensor when trouble is found. Some of these sources of error, though not all, are listed in Table 4.

Table 4

Sources of Calibration Errors for On-Line Sensors

Sensor Change in sheet characteristics or composition Change in environment: temperature, dust, pitch, condensation Failure of sensor source, optics, electronics, or mechanics Scanner and heads Beam de¯ection and misalignment Head damage and misalignment Change in belt tension giving biased readings on alternate scans Circuitry and software Failure of accuracy enhancers Detection of other signals Stray voltage Software bugs Static calibration Constants incorrectly determined or entered Constants not properly entered into permanent storage Accuracy enhancement features Incorrect installation Component failure Incorrect multipoint standardization Incorrect temperature measurements Incorrect assignment of constants to grades Dynamic calibration Constants incorrectly determined or entered Constants not properly entered into permanent storage Incorrect handling of paper samples Misalignment of CD samples with on-line testing points Missynchronization of MD samples with measurement time Operator error Entry of incorrect grade code Incorrect calibration of sensor Source: Ref. 6.

24

Scott

Static Calibration In static calibration, prepared samples are used to measure the response of the sensor. These standards are then inserted into the sensor while it is parked off the moving web. The sensor values are then adjusted (tuned) to read the standard value. Standards covering the range of operation for the sensor are needed to tune and verify that it is operating properly. Static tests verify the mechanical operation of the sensor, as different routines are used during static testing than during normal operation. This type of test provides a fast method of ensuring the proper operation of the sensor and should be one of the ®rst tests performed if the operation of the sensor is in question. For routine maintenance of the sensor, a multipoint calibration should be performed monthly. Single-point validation, preferably near the current operating point, should be done three times a week. Static testing is useful for troubleshooting, allowing problems to be isolated more easily under more controlled conditions [13]. Dynamic Calibration The second method of calibration is dynamic calibration. In this method, samples of the paper being produced on the machine are taken and tested according to standard laboratory test methods. The sensor reading and the laboratory test results are compared, and the necessary changes are made to the sensor tuning. This dynamic method of sensor calibration has several advantages. First, the calibration is being done on the paper that the machine is producing, not on some prepared calibration sample that may be different from the grade of paper being produced. Second, during the calibration the sheet is moving at machine speed through the sensor and not held stationary in the sensor as in the ®rst method. Third, dynamic testing veri®es the operation of the entire sensor system, including the software and hardware. It will also bring to light any effect that the environment is having on the sensor performance. Other errors that may be detected by dynamic testing and not static testing include problems induced by the movement of the scanner and possible alignment problems [13]. With this dynamic method, care must be taken in the handling of the samples to ensure that the laboratory test is measuring the same area of the sheet and that the property of the sheet being measured has not changed between the sampling time and the testing time. The sample must also be representative of the sheet that was measured by the sensor. High frequency machine direction variations can make dynamic calibration dif®cult. In many cases, dynamic testing tends to be more effective but also more time-consuming. However, it usually warrants the extra effort [6,25]. Pro®le Testing Pro®le testing of a scanning system or a full-web system can only be done dynamically. A sample across the full width of the reel is taken, with the corresponding readings taken from the computer of the scanning sensor. In the case of moisture pro®ling, extreme care must be exercised to ensure that the sample is sealed immediately to prevent the moisture from changing prior to the laboratory testing. The test strip is divided into samples corresponding to the CD positions of the on-line sensor. In the case of basis weight, it is better that the oven dry basis weight be considered in order to reduce the errors that will occur due to the moisture content of the paper equilibrating. Figure 2 shows the presentation of the data from a dynamic pro®le test for a basis weight sensor. Two observations can be made from this calibration. First, the sensor is reading well within the tolerances normally seen

On-Line Testing of Paper

25

Fig. 2 Plot of pro®le data from the sensor compared to laboratory measurements. The laboratory measurements agree within experimental error and show the same trend across the sheet.

for this type of sensor. Second, the laboratory measurements con®rm the on-line sensor's determination of high basis weight at the front of the sheet and low basis weight near the back edge [5,13].

III.

ON-LINE SENSORS

The sensors used in the paper industry for on-line testing operate on many different principles. All sensors require proper maintenance and calibration to be used effectively. The various sensors available, discussed below, are divided into three groups, those for structural measurements, optical properties, and strength properties. A ®nal section discusses the latest developments in full-sheet inspection. A.

Structural Measurements

Basis Weight The mainstay of basis weight measurement has been beta gauge transmission methods. Figure 3 illustrates the general con®guration of such a sensor. Beta gauge measurements depend on the uniform attenuation of the transmitted radiation detected by an ion chamber. This attenuation, which is nearly independent of furnish or pulp type, can be expressed as I ˆ exp‰ tŠ I0

26

Fig. 3

Scott

Schematic of a basic beta gauge detector for on-line basis weight measurements.

where I is the transmitted radiation intensity, I0 is the incident radiation intensity,  is the mass attenuation coef®cient,  is the density of the paper, and t is the caliper of the sheet. Thus, knowing the degree of attenuation of the radiation source and having reasonable numbers for the mass attenuation coef®cient, the basis weight (t) can be calculated. The relationship given above is the familiar exponential attenuation of Beer's law, which holds for a monoenergetic source of beta rays. However, because the source is not monoenergetic, there are some minor deviations from this form, which are greatest with lighter sheets [6,31]. During their development, many re®nements were made to basis weight sensors. Early sensors had to be calibrated for each particular basis weight being measured. This was due primarily to the fact that the measurement was sensitive to the position of the sheet between the transmitter and the receiver. The position of the sheet, or ``¯utter,'' and the misalignment of the two heads of the sensor were signi®cant dif®culties with the beta gauge technology. Developments in the 1950s and 1960s led to better compensation for both of these effects, providing application over broader ranges of basis weights and making the sensors more robust. Further geometric changes in the 1970s improved the robustness of the sensors toward changes in furnish [31]. Beta detectors have several advantages that make them particularly suited for use in paper mills. First, the emitter and detector do not contact the sheet, thus reducing interference with the operation of the paper machine. Proper design of the detector makes it quite rugged and insensitive to small misalignments. This is important in the harsh environment in the mill. Compensators can be built in to adjust the calibration of the detector due to the buildup of dirt on the detector and due to the air gap between the paper and the detector. Changes in pressure and temperature change the absorption properties of the air, thus requiring a correction based on continuous measurements of the air [6,25].

On-Line Testing of Paper

27

Beta detectors do have several disadvantages. They may be sensitive to the presence of ash, especially heavier components such as titanium dioxide. Although these effects are relatively small, they can account for about a 0.3% error for a 1% change in ash content. The random nature of nuclear decay also means that measurements must be accumulated over a certain amount of time in order to reduce the amount of variation caused by the decay process. The coef®cient of variation, cv , is related to the source strength S and the averaging time t by the relationship r 1 cv / St Thus, increasing the source strength will reduce the need for long ``exposure'' times. However, this still places a limitation on the resolution that can be obtained with beta detectors (see ``CD Resolution'' in Section II.B) [6]. Beta detectors for basis weight can be calibrated both statically and dynamically. In static calibration, standard samples made of Mylar are used to calibrate the detector. Mylar is very stable in a variety of environments and needs protection only from dirt and physical damage. Ideally, samples of the same type as those used by the initial installer should be used for periodic calibration. In static calibration of a basis weight sensor, the scanner is taken off-line, the samples are inserted one at a time, and the measurements are read from the computer. The results of such a calibration can be reported as a deviation as a function of basis weight (Fig. 4) or as a plot of measured basis weight as a function of the basis weight of the standard sample (Fig. 5). In the case of the second method, a straight-

Fig. 4 Plot of sensor deviations as a function of basis weight during static calibration of a sensor. Note that the sensor tends to read relatively higher at the higher basis weights.

28

Scott

Fig. 5 Plot showing the use of a straight-line calibration of a basis weight sensor with laboratory measurements.

line (or other function) correlation can be ®t to the data, which are then programmed into the computer [13]. Dynamic calibration is best accomplished by integrating the basis weight readings over an entire reel of paper. The resulting reel, whose width is known and whose length can be calculated from the machine speed and the time it takes to build, is weighed and the basis weight is calculated. A correction to the calibration constants can then be calculated [6]. Optical methods can also be used for basis weight measurements, which addresses one of the limitations of beta gauges. As discussed previously, beta gauges are limited to their rate of scanning across the sheet. With the typical scan taking 30 s or more, the rate of response for both CD and MD control is limited. A series of sensors across the sheet would help reduce the response time, but this would be prohibitively expensive and dif®cult to maintain. An optical sensor can provide a nearly instantaneous measurement of basis weight across the entire sheet and provide short-term CD pro®les. However, an optical sensor responds to the opacity of the sheet, which will change in response to composition changes in addition to changes in basis weight even more so than the beta gauge. Thus the optical sensor needs to be calibrated frequentlyÐseveral times a day, in factÐwith the more stable beta gauge sensors. Thus, in concert, the two types of measurements can give timely data that are stable over long time periods [15]. Basis weight can also be measured using infrared light and CCD cameras. The light intensity of the full width of the sheet is scanned at a rate of 60,000±120,000 times per minute. The resolution of the system is such that in addition to reporting the basis weight, information regarding the formation of the sheet can also be

On-Line Testing of Paper

29

reported. Also, unlike scanning sensors, the entire sheet is being measured; thus a true representation of the formation is obtained [1]. Moisture Content Moisture measurements have depended on three main technologies: electromagnetic radiation (radio frequencies), infrared measurements, and microwave technologies. Both transmission and re¯ective methods of moisture measurement have been developed. The re¯ective methods do not require the synchronization of two moving heads, one above and one below the web. However, re¯ective methods can be affected by changes in the moisture pro®le through the thickness of the sheet. Sensors operating on re¯ective principles measure only the moisture content of the surface of the sheet, which can be signi®cantly different from the bulk sheet moisture, especially if that side of the sheet was recently in contact with a hot dryer cylinder. Infrared absorption causes vibrational resonance of water molecules around 1010 Hz. In this case, the measurement is not a transmission measurement but rather a re¯ective measurement, having both the transmitter and detector on the same side of the sheet (see Fig.6). In infrared detection, typically two wavelengths are used. One wavelength (1:95 m) is strongly absorbed by water, whereas the reference wavelength (1:8 m) is nearly independent of sheet moisture (see Fig. 7). The moisture content is determined as a function of these two signals. Because of the reference signal used, the detector automatically adjusts for changes in source intensity and

Fig. 6 Schematic of a basic infrared moisture sensor that depends on re¯ected measurements. The ®lter wheel alternately measures the re¯ected infrared radiation at two different wavelengths.

30

Scott

Fig. 7 Spectrum of the infrared absorption of wet and dry paper showing the two wavelengths that are used in infrared moisture sensors.

detector gain shifts [31]. This relative attenuation at the two wavelengths gives the amount of moisture per unit area of the paper. To correctly determine the moisture content, the data from the IR detector must be combined with the measurements of the basis weight detector. Thus, a properly operating basis weight detector is necessary for proper determination of the moisture content of the paper [6]. Because the moisture is determined by comparing the absorption at two different wavelengths, either two measurements must be taken with a single detector using narrow bandpass ®lters or two detectors must be used, one dedicated to each wavelength. There are advantages and disadvantages to each method. Using one detector, the sensor would have a rotating disk with two ®lters mounted that alternately ®lter the beam (Fig. 8). Because the sheet will move, however slightly, during the time between the reference and measurement signals, the basis weight of the sheet may change, which will cause measurement errors when the relative intensities are compared. The two-detector method, typically using a beam splitter, does not suffer from this drawback. However, both detectors must be calibrated frequently to prevent the drift of one relative to the other [6]. Radio-frequency sensors tend to be much more rugged than infrared sensors. In radio-frequency measurements, a contacting con®guration is used with a driver, detector, and grounded shield electrodes. Interfacial polarization mechanisms dominate the coupling through the sheet. Also radio-frequency measurements tend to have a wider range of moisture contents at which they can operate than the infrared measurements [31]. In microwave measurements, two resonances are established in the cavity of the detector. One resonance mode is coupled with the sheet, and the other is

On-Line Testing of Paper

31

Fig. 8 Schematic of a basic infrared moisture sensor that depends on transmitted measurements. The ®lter wheel alternately measures the infrared radiation at two different wavelengths.

independent. Since moisture in the sheet alters the paper-coupled resonance, the amount of moisture can be determined. Again, like infrared measurements, the measurement is compared to a reference, providing compensation for drift of the sensor [31]. Many moisture sensors tend to be somewhat grade-dependent. This is not usually a problem on commodity machines that produce the same grade of paper from day to day. However, on specialty machines, standards for several of the grades will probably need to be made and the calibration done for each grade. In this way, functional relationships between the scanner measurement and the actual moisture can be stored in the sensor software and used as appropriate. Grouping of grades that produce similar calibrations can reduce the amount of work needed for periodic recalibrations [13]. Carbon black, which is a common component of newsprint ink, greatly affects the absorption of IR. This can cause signi®cant measurement errors in furnishes that contain this material, requiring special compensation [6]. As with basis weight sensors, moisture sensors can be calibrated either statically or dynamically. Static calibration requires the preparation of samples of known moisture content that are sealed in IR-transparent envelopes. Careful preparation of these samples is very important. Samples of the appropriate grade of paper are cut, dried in an oven, and weighed. Samples are then placed in desiccators containing saturated solutions of various salts that produce known humidities. After conditioning, the samples are sealed in the preweighed IR-transparent envelope, reweighed,

32

Scott

and the moisture content determined. Typically, a four-point calibration over the normal operating range of the sensor is suf®cient. Dynamic calibration requires careful sampling to prevent the loss or gain of moisture during sampling. The samples must be quickly sealed in preweighed plastic bags and should be taken close to the sensor, although this may not always be possible. With dynamic calibration, the sensor can be calibrated to account for the moisture loss (or gain) between the location of the sensor and the reel where the moisture is actually desired to be known [6,13]. A detailed procedure for the calibration of on-line moisture sensors is given in the TAPPI technical information papers [34]. Caliper The caliper of a sheet depends on the measurement instrument because paper is compressible. However, its measurement is similar to the corresponding laboratory measurement. The caliper can vary at different locations on the sheet, just as any paper property varies in both the cross and machine directions on many scales. To accurately report caliper, the pressure in loading the testing device and the size of the probe, which both have an effect on the measurement, must be reported. In general, the greater the pressure, the lower the caliper of the sheet as it is compressed. Larger probe sizes typically result in larger calipers as a larger probe is more likely to press onto a high spot in the paper. As Fig. 9 shows, the pressure used in online sensors is not the same as that used in laboratory testing. In fact, the laboratory standards worldwide differ, with TAPPI (T 411 om-89) specifying a measurement pressure of 50 kPa whereas the standard in Europe uses 100 kPa.

Fig. 9 The caliper of paper as a function of the measurement pressure. Note the different pressures used in on-line applications compared to the standard laboratory measurements.

On-Line Testing of Paper

33

Caliper is still a dif®cult property to measure on-line because of the necessity of sheet contact through the sensor. Maintaining good contact with the sheet while preventing wear on the equipment is still an engineering challenge [25]. This close contact with the sheet must be maintained in order to achieve accurate readings of caliper from the sensor. However, this can prove dif®cult considering the ¯ocs, holes, and other defects that the sensor must contend with. Care must also be taken to prevent excessive buildup of ®bers and other materials on the sensors, because they can cause erroneous readings. Caliper measurements have taken several different forms. The most common is the air bearing/proximity sensor devices. To accurately measure the caliper, the air bearing is used to accurately position one side of the sheet and the distance from the proximity sensor to the other side of the sheet is measured, thus determining the caliper. Earlier designs required a light constant pressure on a backing plate, which could be a source of machine breaks. The air bearing has allowed for high accuracy measurements of sheets [31]. Calibration can be done statically using specially prepared Mylar standards of a known caliper. Dynamic calibration is necessary in order to have the measurements correspond to the laboratory values. As discussed above, the measurement pressures may not be the same for the on-line sensor as for the laboratory, thus requiring calibration. To dynamically calibrate the sensor, periodic samples are taken, tested in the laboratory, and compared to the on-line sensor reading. Any necessary adjustments are made on the basis of this comparison [6]. A calibration procedure is given in the TAPPI technical information sheets [35]. Formation Formation is a measure of the degree of ¯occulation in the sheet. It is often qualitatively determined by holding the sheet up to the light and observing the pattern of light transmission. The size of the ¯ocs when determining the formation of a sheet can range from 0.1 to 100 mm. Formation can also be thought of as the local variations in basis weight of the sheet that cause these differences in light transmittance. Two on-line methods of formation detection have been developed that depend on the relative transmission of light through the sheet [48]. The ®rst method compares the intensity of light passing through the sheet over a small area (approximately 1 mm2 ) with that of light passing through a larger, concentric area (approximately 900 mm2 ). In this way, the local basis weight (measured by the attenuation of light) is compared to the average basis weight over a larger area. The difference between the two is an indication of the formation of the sheet. The second method uses only a single detector that measures the light transmittance in a small (approximately 1 mm2 ) area. The changes in this signal as the sheet passes through the sensor allow the variations in basis weight to be determined. The rate of change in the light intensity gives an indication of the size of the ¯oc, which is determined by the rapidity of the pulse, the machine speed, and the scanner speed. In this way, the average ¯oc size can be determined along with a distribution of ¯oc size represented as a power spectrum [6]. The subsequent analysis of these values is the most critical step in the determination of formation with an on-line sensor. One such method was developed using laser light transmission through the sheet [37] but should be applicable for any transmission method. This method de®nes a look-through index, E0 , which is de®ned

34

Scott

as the root mean square of the deviations of the transmission divided by the average transmission value. That is, „  2 dt  RMS …1=T† T ‰F…t† FŠ V‰F…t† FŠ 0 ˆ E0 ˆ „t VF …1=T† 0 F…t†dt where F…t† is the transmission value at a particular point, V is the sheet velocity, T is the time lapse, and F is the mean value for the transmission. Thus the lower the value of E0 the less the variation in the transmission and the better the formation of the sheet. As with all scanning methods, the above two methods sample only a very small section of the sheet. Optical methods using digital cameras or charge-coupled devices (CCDs) can scan much larger sections of the sheet as they pass under the sensor. In fact, as computer processing speeds improve, full-sheet scanning becomes possible. Using these devices, formation can be measured by digitizing a larger sample of the paper and analyzing the light transmittance using image analysis techniques. Techniques using infrared light together with CCD cameras have also been developed for the measurement of formation [1]. Fiber Orientation Fiber orientation is important in the strength of paper. It is well understood that the ®ber orientation (typically in the machine direction of the paper) correlates with the ratio of the MD to CD strength of the paper. Some work has been done on the use of microwave attenuation for the measurement of ®ber orientation. Although this technique is described as a laboratory method, it could be easily incorporated on-line. In this technique, the ratio of the attenuation of the microwaves when the electric ®eld is aligned in the machine direction is compared to that when it is aligned in the cross direction. This ratio gives information regarding the anisotropy of the sheet. Because the microwaves pass completely through the sheet, it gives information on the average orientation through the sheet [19]. Schacanski and Dodson [38] developed an image analysis technique for measuring the anistropy of a sheet. The method uses two-dimensional basis weight images captured by -radiography or light transmission methods. The method has a potential for on-line measurements of anisotropy for monitoring and control. With this technique, the local variations could be monitored across the machine both online and off-line. Porosity Porosity measurements on-line essentially mimic the laboratory methods of measurement. The systems measure the time it takes to pull a ®xed volume of air through the sheet under a constant vacuum. The major difference between the on-line test and the laboratory test is the need to maintain a constant seal with a moving sheet. This is done by using specially designed measurement shoes that have two areas under vacuum (see Fig. 10). The outer ring area of vacuum serves to seal the moving sheet to the porosity-measuring shoe. The actual measurement is then made through the opening within this circular seal. Different shoes are available for sheets of varying porosity. Because the measuring shoe is in contact with the sheet during its measurement cycle, its surface must be carefully designed. It must be extremely smooth and almost ``frictionless'' to prevent marking the sheet or causing sheet breaks. Also, because the sheet could potentially be moving

On-Line Testing of Paper

35

Fig. 10 Sensor head vacuum patterns for measuring porosity on-line. The smaller head would be used for more porous papers.

at over 1000 m/min across the shoe, the shoe must be extremely resistant to abrasive wear by the sheet [26]. The on-line sensor has a very important advantage over the laboratory testing device. Porosity can be highly variable within a sheet. This requires, as with many paper property tests, that multiple measurements be made over the sheet and the results averaged. In the case of the on-line sensor, the sheet is moving under the sensor, thus automatically providing an average value over several hundred meters of paper, depending on the speed of the machine. Thus, with a single measurement cycle, the on-line tester can give a much more representative measurement than the corresponding laboratory test. The porosity sensor does not have to be mounted on a traversing scanner. Most of the operating variables that would affect the porosity of a sheet of paper are located well before the headbox of the paper machine. On-line measurements of porosity can be used to control the amount of re®ning for wet-end chemistry evaluations and for vacuum control in the forming section of the paper machine. Using an on-line porosity measurement in this way allows more rapid feedback than is achievable with laboratory testing. It must be remembered that even though the on-line sensor is very close to the laboratory apparatus measuring porosity, a program of regular calibration is still necessary. Static calibration can be done off-line with samples of a known porosity. It is best that these standards be of very uniform formation to reduce the effect of the exact placement of the sheet on both the on-line sensor shoe and the corresponding laboratory tester. The same considerations must be made for dynamic testing. A suf®cient number of tests must be done in the laboratory to ensure that an accurate representation is achieved. Furthermore, for ®xed (nonscanning) sensors, the samples should be taken from the same CD position as the sensor. Smoothness Smoothness is most easily measured on-line by employing optical methods. Two different measurement techniques are used that differ by the angle of

36

Scott

incidence of the light beam. In the ®rst method, a narrow beam of light is directed perpendicularly onto the sheet. Detectors placed at shallow angles to the paper on the same side of the web detect the scattered light. A ¯at sheet scatters light equally to the two detectors, which are placed symmetrically about the point of incidence. A rough spot, which will have some angle relative to the ¯at sheet,will preferentially scatter light to one detector (Fig. 11). The amplitude of the difference in the intensity at the two detectors, which varies as the sheet passes beneath the sensor, indicates the roughness of the sheet [6]. The second method uses a light beam that contacts the paper at a low angle of incidence. A light detector array opposite the incident point measures the amount by which the light beam is scattered (Fig. 12). A smooth sheet activates a minimum number of closely-spaced sensors in the array, whereas the greater scattering of the rougher sheet activates a number of widely-spaced sensors [6,45]. Calibrations with laboratory measurements are extremely important in the case of on-line smoothness measurements, because the measurement techniques of the laboratory and of on-line sensors are completely different. The laboratory method for smoothness depends on measurements of air leakage around the sealed edge of the seat, whereas the on-line methods depend on light scattering. Fortunately, there is a good correlation between the two methods. Two sensors, one on each side of the sheet, are needed to measure the smoothness of both sides of the sheet.

Fig. 11 Smoothness measurement of paper using perpendicular incident light. The smooth sheet re¯ects more evenly than the rough sheet to the two opposing detectors.

On-Line Testing of Paper

37

Fig. 12 Smoothness measurement of paper using low-angle incident light. The rough sheet activates a greater number of sensors in the linear array

Web Flatness (Curl) Flatness (or the lack of curl) in paper is an important property for many end uses of the paper, including many printing and converting operations. Flatness is the ability of paper to lie in close contact with a ¯at surface while not under tension. Deviations from a ¯at sheet are often caused by local differences in the sheet, especially in terms of its stiffness and elongation. These variations are not immediately apparent, because the sheet is kept under tension within the machine and in the reel. To measure ¯atness on-line, the cross machine direction tension, which is related to the stresses in the paper, is measured instead. The paper is passed over a roll consisting of a number of coaxial friction-free rotors approximately 10 cm wide. The load on each of the rotor sections, which is related to the tension in the web, is measured. The ¯atness is then related to the difference in the tension on a particular rotor compared to the average tension across the web. In relating this to the ¯atness of the sheet while it is not under tension, it is assumed that the modulus of elasticity is not appreciably different across the sheet. Note that the tension is being measured simultaneously across the entire web, and thus a true cross-machine direction measurement is being made [50]. Coating Weight Coating weight is often measured as the difference between two scans, one before and one after the coating process. Beta absorption can be used before and after coating to measure the differential basis weight. Since modern beta detectors are fairly insensitive to sheet composition, the difference in basis weight represents the applied coating. Ash measurements based on low energy X-ray mea-

38

Scott

surements can also be used in such a differential method. Again, two sensors are needed, one before and one after the coating operation. In practice, because the basis weight measurements are less sensitive to the sheet composition, they are more accurate in long-term average coat weight measurements. However, X-ray measurements of ash are about ®ve times as sensitive, making them better for short-term variations and CD pro®les of coat weight [6]. Another method used to measure coat weight is beta-ray backscatter. In this method, beta rays impinge the sheet at a low angle of incidence, and the amount of scatter is measured. This scatter was found to vary linearly with the coat weight. However, this method is very sensitive to the formulation of the coating, and the sensor must be recalibrated at each grade change. The method is also limited to a coat weight of approximately 10 g=m2 [36]. Ash Ash or ®ller content measurements are typically made with low energy X-ray attenuation. In this case, the attenuation mechanism is absorption of the X-rays by ®ller particles, which causes ionization. Like all materials, ®llers such as clay, calcium carbonate, and titanium dioxide absorb different frequencies of X-rays to different degrees. Although the attenuation curve for clay is a smooth function of the X-ray energy over the range of energies typically used, the curve for TiO2 is sharply discontinuous at approximately 5 keV. This discontinuity can be used to develop a measurement that is independent of the relative composition between the two ®llers. By providing a proportion of the spectrum above this discontinuity, the average attenuation coef®cient of the two components is the same and the measurement becomes independent of the composition. If we let xl be the fraction of the spectrum below the discontinuity, given that the coef®cient of TiO2 is half that of clay below the discontinuity and 4 times that of clay above the discontinuity, TiO2 ˆ xl 0:5clay ‡ …1

xl †4:0clay

Setting this equal to clay , we ®nd that xl ˆ 0:86 or that 86% of the radiation below the discontinuity allows the overall coef®cients of the two ®llers to be equal, thus making the measurement independent of the proportion of the two. Of course, to make use of this property requires that the X-ray source be extremely stable and reliable. As with the other sensors, everything between the emitter and the detector absorbs or scatters the X-rays [31]. Ash detectors are calibrated in the same manner as basis weight sensors. Standard Mylar sheets of known ash contents can be used to statically calibrate the sensor. Samples and careful laboratory measurements are used to calibrate the sensor dynamically. B.

Optical Measurements

Brightness and Color On-line measurements of optical properties, although they use many of the same principles as laboratory testing methods, differ in signi®cant ways. Brightness and color, according to laboratory methods, should be measured with the sheet backed by an in®nite stack of sheets of the same type. In the case of on-line testing, this is clearly dif®cult, and the brightness of only a single sheet can be

On-Line Testing of Paper

39

measured. These single-sheet measurements thus need to be correlated with the in®nite-sheet laboratory methods [6]. Two methods of color measurement are commonly used in on-line sensors. The ®rst uses the tristimulus approach, which is the traditional method of color measurement. In this method, the responses of the paper to the three primary colors are measured and combined to produce values indicating the sheet color. However, there are dif®culties with this method because the entire spectrum is being represented by only three parameters [22]. A more accurate approach is to measure the entire spectrum. For color measurements, the light beam is dispersed into the spectral band, forming a continuous color spectrum over a portion of the sheet. This is unlike laboratory spectrophotometers, which scan the spectrum over time. A set of diodes, each detecting a certain wavelength of light, respond to give the spectrum of the re¯ected light. This spectrum can then be transformed into the various measurements that are commonly reported for color, such as L, a, b values where L represents the intensity of the color and the a and b values indicate the hue along red-green and yellow-blue axes, respectively [6,22]. Some development work has been done to address the issue of single-sheet measurements of on-line sensors compared to the in®nite sheets used in laboratory testing. This method primarily addresses the concern that a change in opacity can change the color when only one sheet is being measured. Of course, the sensitivity of the sensor to this type of error increases as the basis weight decreases. A new concept for a color sensor would have it located above the reel and measure the color of the paper on the reel. As the reel is built, a light projector and receiver track its surface to measure the color. Two important points must be dealt with in a system such as this. Some fairly sophisticated tracking software is needed to maintain the proper angle of incidence and re¯ectance from the surface of the reel. Additionally, the distance from the surface to the sensor is continuously changing as the reel is built. Because there is a fairly large distance between the sample and the sensor (approximately 2 m), there is a greater chance of interference by dust and other contaminants. Although still in the development stage, this method is useful for obtaining the a and b values of the color measurement; work on the L value still needs some re®nement [4]. As with all sensors, calibration is extremely important for on-line color and brightness measurements. Static calibration is performed using standard samples while the sensor is off the sheet. The standards are placed within the sensor, and the reading is adjusted to give the proper value. In dynamic calibration, sheet samples are taken and tested according to the standard laboratory measurements. With dynamic calibration, the difference between the single-sheet and in®nite-sheet measurements can be taken into account. Opacity On-line measurements are probably the simplest of the on-line measurements and the type that most closely resembles the corresponding laboratory test (Chapter 4). Light transmission, using ®lters that correspond to human vision, can be measured with a sensor on one side of the sheet and a light source on the other. Standardization can be easily done to correspond to TAPPI Standard T 425 by the use of standard samples in a static manner or by measuring sheet samples while the machine is running. [6].

40

Scott

Gloss Gloss measurements on-line are performed quite similarly to the standard laboratory methods (Chapter 4). Gloss, the specular re¯ectance of the sheet, is measured by the light intensity of a beam, with the angles of incidence and re¯ection being 75 as is done in TAPPI Standard T 480. Again, as with smoothness, a sensor is needed on each side of the sheet in order to measure both sides. The sensor is commonly calibrated by inserting prepared standards of a known gloss into the sensor. Dynamic calibration can also be done by carefully sampling and comparing the results to laboratory measurements [6]. C.

Strength Tests

Laboratory tests for the strength properties of paper are commonly destructive tests. That is, the paper is damaged or otherwise rendered unsuitable for further testing or use. The complexity of the tests, which involve sample preparation, cutting, and placement in the testing apparatus, usually limits the frequency of these tests to once per reel. Because it is impossible to perform such destructive tests on-line, it is necessary to ®nd some property of the paper that is related to the strength that can be measured on-line. Most on-line strength measurements actually measure properties of the sheet related to the elastic modulus, which is then correlated to the strength of the paper. The methods differ mainly in the manner in which the modulus is measured. Ultrasonic Measurement of Elastic Modulus The various engineering elastic constants such as the Young modulus, shear modulus, and Poisson ratio, can be determined from ultrasonic measurements in paper. These values are calculated from measurements of the speed of ultrasonic waves in different directions of the paper. Many of the end-use tests, especially the tensile test, have been correlated to these values [6]. The velocity of longitudinal waves in paper, VL , is related to the Young modulus, E, and the density of the sheet, , by the equation p VL ˆ E= Likewise, the velocity of transverse or shear waves, VS , is related to the modulus of rigidity, G, as p VS ˆ G= Thus it is necessary only to be able to measure the time elapsed between a sonic transmitter and a receiver a known distance apart to calculate the velocity of sound in paper. With a single transmitter and two receivers placed perpendicularly to it, both the CD and MD velocities can be measured. Given the relationships above, the moduli can be calculated and a correlation with the strength properties determined [2,18,23,40]. Mechanical Measurements of Elastic Modulus The extensional stiffnesses in the machine and cross direction can be calculated using mechanical methods. The drawback to these methods is that there must be contact with the sheet and the sheet must be de¯ected slightly. The force of the de¯ection in various directions is measured by using a segmented ring sensor. A proximity sensor measures the amount of

On-Line Testing of Paper

41

de¯ection caused by a spherical wheel in contact with the sheet. With this information, the stiffnesses can be calculated. These measurements have been correlated with tensile tests and STFI compression tests for particular grades [6]. The Young modulus can also be related to the web tension and drying stress of paper. By the placement a web tension device between the dryers, the modulus and, through correlations, the strength properties can be calculated. Using the relationship of Nissan that states that the modulus is proportional to the cube root of the number of hydrogen bonds, the equation can be written !1=3 0 Keb
v Eˆk P1=3 wt where K and b 0 are parameters related to the hydrogen bonding,
v is the velocity differential between the dryer's sections, w is the sheet width, t is the sheet thickness,  is the average amount of energy required to form a single hydrogen bond, P is the measured tension in the web, and k is a proportionality constant. Under normal operations,
v and the rest of the parameters remain constant or can be easily measured (e.g., t and w), and the elastic modulus is proportional to the cube root of the tension P. Note that this method of on-line monitoring of the elastic modulus gives a composite modulus over the entire width of the sheet [7,8]. Holographic Interference Measurements of Elastic Modulus The propagation of waves through a sheet of paper can also be detected with holographic interferometry. In this method, bending waves are initiated and detected by a pulsed laser to determine the propagation rate. The modulus of elasticity can be found as ! 3 …1 2 †r4 Eˆ 2 4 h3 t2 where  is the basis weight of the sheet,  is the Poisson ratio, r is the radial distance from the impact center, t is the time after impact, and h is the thickness of the sheet. This expression can be more simply expressed as the bending stiffness D, eliminating the need for the Poisson ratio: ! 1 r4 Dˆ 162 t2 Both of these values, E and D, can be related to the strength properties of the sheet [27,28]. Holographic interferometry also has the potential to detect anisotropy and defects in the sheet. In a material that has uniform properties in all directions, the interference pattern will be perfectly circular. Noncircular or elliptical patterns indicate anisotropy in the paper, and the strength properties in both the machine and cross-machine directions can be estimated. Additionally, defects in the paper, such as holes and areas of locally higher basis weight, will become evident in the interference patterns produced. Using these techniques for these purposes, however, requires more sophisticated analysis of the interference patterns produced [27,28]. Linting and Dusting (Surface Strength) Linting and dusting can be a signi®cant problem related to the printability of a sheet of paper. A practical test for linting is to

42

Scott

run the paper through an offset lithographic printing press and observe the buildup of lint on the printing blanket. A three-stage on-line tester for linting ®rst loosens weakly bound ®bers and particles by using a positive pressure on the sheet. The movement of the sheet generates a negative pressure in the second stage and carries the particles to the third stage, where they are vacuumed from the sheet onto a ®lter. The pressure drop across the ®lter is monitored until a set pressure drop is achieved. At this point, the measurement stops and the time elapsed is reported as being inversely proportional to the linting of the sheet [24,51]. Calibration of Strength Sensors Many of the methods used to measure the mechanical properties of paper on-line depend on correlations between different properties. Thus dynamic calibration is extremely important in the use of these sensors. Care must be taken to get a suf®cient number of representative samples that can be tested and compare the results to those of laboratory testing.

D.

Paper Defect Detection

In some paper applications, such as the production of photographic paper, the sheet must be nearly perfect in appearance and uniformity in both the machine and cross directions. Defects that will cause rejection of the paper include holes, light spots, dark spots, wrinkles, scratches, streaks, calender cuts, calender darkening, and translucent areas. Because such defects can occur anywhere in the sheet, traditional scanners, which cover only a small portion of the sheet (less than 1%), are unacceptable for this application except for chronic problems of suf®cient duration. Thus, in most cases, some sort of full-sheet defect detection that inspects all of the sheet is necessary. Holes in the sheet can be detected in several ways. Paper, being a good insulator, prevents the completion of a circuit between two conductors such as a metal backing roll and wire brushes. A hole in the sheet will allow the completion of the circuit and set off an alarm. Changes in visible, ultraviolet, and infrared light transmission can also indicate the presence of faults in the sheet. These analog detectors measure a change in the intensity caused by a fault such as a hole or other defect. In the inspection of coated paper, fast scanning lasers can essentially inspect nearly the entire sheet. Detectors on the opposite side of the sheet from the laser can detect holes and thin spots, whereas detectors on the same side of the sheet can detect wrinkles and scratches due to the scattering effect these would have on the laser light [6]. Finally, full-web inspection can be accomplished by videocameras and CCD cameras. These devices create an image of the sheet that is analyzed for defects. Inspecting the entire sheet at this level of detail involves a great deal of computational power that requires powerful computers. Some systems incorporate some of the computational power in the camera devices themselves, thus carrying out a prescreening of the images acquired by the camera. In this way, only those images that show some deviation from normal sheet images are transmitted to the main computer for further identi®cation and processing. Image analysis software has reached the point of being able to distinguish between the various surface and sheet defects and give the appropriate alarm.

On-Line Testing of Paper

43

Table 5 Manufacturers of On-Line Sensors for the Paper Industry ABB (Accuray) Avatron Ltd. BTG Inc. Cognex Corporation CyberMetrics, Inc. Dantec Measurement Technology Inc. Davy Fisher-Rosemount Industry Solutions Foxboro Company Honeywell-Measurex Impact Systems, Inc. Infrared Engineering

E.

Kaptra Inc. KDY Assoc., Inc. Lorentzen & Wettre USA Inc. M/K Systems, Inc. Moisture Register Products NDC Infrared Engineering Papertech Inc. Rockwell Automation RYECO, Inc. Systronics, Inc. Thermoelectron Valmet

Manufacturers

There are many manufacturers of on-line sensors for the pulp and paper industry (see Table 5). Some specialize in equipment for that industry whereas others are more general and serve several industries. Many have a full range of sensors for many of the important properties of paper. More information on particular product lines can be obtained from the companies directly, from the Internet, or from the various trade journals and directories.

REFERENCES 1.

Anonymous. (1997). Accuracy hyperscan for monitoring of grammage and formation during paper making. Wochenbl. Papierfabr. 125(23-24):1192. 2. Baum, G. A., and Habeger, C. C. (1980). On-line measurement of paper mechanical properties. Tappi J. 63(7):63±66. 3. Bennett, M. B. (1998). Focusing on the future of controlling the process. Tappi J. 81(3):64±68. 4. Bradford, R. A., Bonham, J. S., Flowers, A. G., Gardner, J. L., and Shaw, J. E. (1991). An on-line sensor for the measurement of paper color. Appita 44(2):139±143. 5. Brewster, D. B. (1983). Calibration of on-machine caliper gauges for reel building. Pulp Paper Can. 84(8):T180±T184. 6. Brewster, D. B. (1993). Xii. Papermaking area. In: Mill-Wide Process Control & Information Systems, Vol. 10, (Brewster, D. B. and Kocurek, M. J., eds.) TAPPI Press, Atlanta, pp. 179±236. 7. Byrd, V. L. (1988). On-line device helps to control strength properties. Paper Technol. 30(1):31±33. 8. Chase, L., Goss, J., and Anderson, L. (1989). On-line sensor for measuring strength properties. Tappi J. 72(12):89±96. 9. Christie, J. S. (1973). On-machine measurements of the chromatic aspects of appearance. Tappi J. 60(2):119±121. 10. Cook, A. J. (1973). On-line appearance measurementÐA status report. Tappi J. 56(2):55±58. 11. Crocker-McAlister, L., and Kopkin, B. (1997). ``Dynamic lab'' vs. ``static lab'' promises pro®tability for papermakers. Tappi J. 80(8):65±67.

44

Scott

12. Cutshall, K., Xviii. (1991). Cross direction control. In: Paper Machine Operations, Vol. 7, (Thorp, B. A. and Kocurek, M. J., eds.) TAPPI Press, Atlanta, GA, pp. 472±506. 13. Dennebaum, E. T. (1987). Verifying the accuracy of scanning sensors. Tappi J. 70(6):65±68. 14. Deodhar, S. (1988). Statistical analysis of cross-machine variations in basis weight. Tappi J. 71(10):223±224. 15. Francis, K., Stenbak, M., and Kleinsmith, C. (1989). The stationary sensor for determining oven-dry weight at the wet end. Tappi J. 72(4):101±105. 16. Gill, J. P. (1969). Continuous on-machine color control of paper. Tappi J.52(2):232±236. 17. Goodsole, C. F. (1988). Integrated paper machine controlÐpast, present, future. In: Proceedings of the 1988 Engineering Conference. TAPPI Press, Atlanta, GA, pp. 569± 576. 18. Habeger, C. C., and Baum, G. A. (1986). On-line measurement of paper mechanical properties. Tappi J. 69(6):106±111. 19. Habeger, C. C., and Baum, G. A. (1987). The use of microwave attenuation as a measure of ®ber orientation anisotropy. Tappi J. 70(2):109±113. 20. Jaaskelainen, M. (1994). Lights, cameras, action ®nd defects. PIMA 76(12):56±57. 21. Lashof, T. W. (1962). APPA-TAPPI reference material program. I. Interlaboratory investigation of TAPPI Standard T 411 m-49, Internal Tearing Resistance of Paper. Tappi J. 45(8):656±663. 22. Long, J. (1988). Sensor advances broaden online measure of papermaking variables. Pulp Paper 62(3):179±183. 23. Lu, M. T. (1975). On-line measurement of strength characteristics of a moving sheet. Tappi J. 58(6):80±81. 24. Mason, R. J. H. (1993). New on-line lint test uses air pressure to simulate a printing press nip. J Pump Paper Sci. 19(1):J40±J44. 25. Mercer, P. (1997). Take control of your sensors. Pulp Paper Int. 39(9):75±79. 26. Nettamo, K. (1998). Online porosity measurement system offers cost-effective quality control. Pulp Paper 72(6):91±94. 27. Olofsson, K., and KyoÈsti, A. (1994). Stiffness and stiffness variation in paper measured by laser-generated and laser-recorded bending waves. J Pulp Paper Sci. 20(11):J328± J333. 28. Olofsson, K., Molin, N.-E., and Kyosti, A. (1991). A new method to detect anisotropy and local variations in paper. Tappi J. 74(3):195±200. 29. Parker, H. V. (1982). What's happening in off-line and on-line test equipment. Tappi J. 65(7):27±31. 30. Patrick, K. L. (1998). Future paper machines to run faster with advanced calendering control. Pulp Paper 72(7):95±101. 31. Pfeifer, R. J. (1981). A review of the development of on-line paper machine sensors. Pulp Paper 55(2):68±73. 32. Popson, J. (1974). On-line measurement of paper color, brightness, and opacity. Paper Trade J. 158(31):24±27. 33. Presgrave, J. E. (1975). On-line control of ¯uorescent white papers. Paper Technol. Ind. 16(1):34±43. 34. Quality Management Committee. (1997). On-line moisture veri®cation/calibration of Infrared moisture sensors. TAPPI Press, Atlanta, GA, pp. 906±914. 35. Quality Management Committee. (1997). On-line veri®cation of caliper sensors. TAPPI Press, Atlanta, GA, pp. 915±923. 36. Rutledge, W. C. (1989). On-line coat weight measurement. Pulp Paper Can. 90(9):102± 107. 37. Sabater, J, Kerneis, J. C., Bauduin, S., and Kiernan, P. (1989). New online sensor technology approaches commercial application. Pulp Paper 63(8):81±83.

On-Line Testing of Paper

45

38. Schacanski, J., and Dodson, C. T. J. (1996). Texture image analysis for paper anistropy and its variability. Appita 49(2):100±104. 39. Schuierer, H. (1992). I. Pulp and paper testing. G. Test facilities and equipment. 3. Automated paper testing. In: Mill Control & Control Systems, Vol. 9, (Kouris, M. and Kocurek, M. J., eds.) TAPPI Press, Atlanta, pp. 37±45. 40. Senko, E., and Thorpe, J. (1985). On-line ultrasonic measurement of sheet modulus. Tappi J. 68(2):95±99. 41. Shapiro, S. I., and Shearin, R. H. (1995). On-line sensors: Exploring the trends. Tappi J. 78(9):83±84. 42. Shead, R. P. (1992). Process control in the 90s. Paper 217(6):36±37. 43. Spitz, D. A. (1990). Control of variations in a moving web with a scanning gauge. Tappi J. 73(1):133±137. 44. TAPPI. (1998). TAPPI Test Methods. TAPPI Press, Atlanta, GA. 45. Vahey, D. (1984). Measuring smoothness on line with an optical-based sensor. PIMA 66(1):41±42. 46. Van Brimer, R. H., and Howard, R. C. (1967). A noncontact contrast-ratio opacimeter for on-machine measurements. Tappi J. 50(2):65A±70A. 47. Ward, J. W. (1969). Production experience with a continuous color monitor. Tappi J. 52(2):239±244. 48. Waterhouse, J. F. (1996). On-line formation measurements and paper quality. In: Proceedings of the 1986 Papermakers Conference. TAPPI Press, Atlanta, GA, pp. 275± 288. 49. Wickstrom, W. A., and Horner, M. (1970). Closed-loop color control for printing papers. Tappi J. 53(5):784±791. 50. Wilson, A. (1995). The production bene®ts from on-machine measurement of web ¯atness. Das Papier 49(10A):V123-V126. 51. Wood, J. R., Grondin, M., and Shallhorn, P. M. (1995). Evaluation of the MacMillan Bloedel on-line lint test. Pulp Paper Can. 96(9):T318±T323.

3 CONDITIONED TEST ATMOSPHERES JOHN L. DE YONG Consultant Physicist Victoria, Australia

I. Introduction II. Relative Humidity A. Theory and Terminology B. Standard Atmospheres

48 49 49 52

III. Measurement of Relative Humidity A. General Comments B. Miscellaneous Methods C. Hygroscopic Sensors D. Thermodynamic Methods

52 52 53 54 55

IV. Control of Relative Humidity

60

V. Humidity Controlled Enclosures A. Design and Performance of Paper Testing Rooms B. Controlled Environment Cabinets VI. Calibration of Sensors VII. Tables and Graphs A. Vapor Pressure Table B. Dew Point Tables C. Psychrometric Tables and Graphs D. Saturated Salt Solutions Appendix 1: Symbols

64 64 67 68 69 69 69 78 79 87

The author was previously af®liated with CSIRO, Division of Forestry and Forest Products, Clayton, Victoria, Australia. Correspondence can be sent to: Dr. John L. de Yong, 76 Warrigal Road, Surrey Hills, Victoria 3127, Australia. 47

48

I.

de Yong

Appendix 2: Programs A. HP-15C Program for Relative Humidity from Dry Bulb and Dew Point Temperatures B. HP-15C Program for Relative Humidity from Wet and Dry Bulb Temperatures C. GW-Basic Program for Relative Humidity from Dry Bulb and Dew Point Temperatures D. GW-Basic Program for Relative Humidity from Dry Bulb Temperature and Wet Bulb Depression

89

91

References

91

89 90 90

INTRODUCTION

With the progress of civilization has come a continual development of technology, and the manufacture of a wide range of materials, for frequent common use. These may be produced from both natural and arti®cial sources, and for useful application, they must be tailored to suit such criteria as economy of manufacture, performance satisfaction, and possibly recycling. This requires the physical testing of materials to meet appropriate speci®cations, in which the testing environment must be considered. Paper, in its various forms, is certainly such a product in every day use for writing, printing, and packaging, as well as specialized papers for archival preservation, or security paper for currency applications, and a range of tissues for medical and personal care. It may be noted that for some uses, plastic sheet is becoming increasingly competitive with paper. For such a wide range of products, the physical properties of particular types and grades of paper must be precisely measured for comparison with required quality standards, and to ensure satisfactory performance. As paper is usually tested in air, which is a mixture of gases and contaminants, the effects of these components on each paper product must be considered. Dry air is a mixture of gases, namely nitrogen, about 78%, oxygen, about 21%, and argon, carbon dioxide, neon, helium, krypton, hydrogen, and xenon accounting for about 1%. However, air is seldom really dry, so that in all practical circumstances we are dealing with a mixture of dry air, whose constituents are effectively constant, and water vapor, the amount of which is quite variable. Apart from treatment to eliminate gaseous and/or particulate contaminants, the air in which paper should be tested may be controlled with respect to pressure, temperature, and moisture content. As a prerequisite to discussing measurement and control of conditioned test atmospheres, the relative importance of each of these three properties will be considered. Papers and boards, being hydrogen-bonded assemblages of cellulose ®bers, can be signi®cantly affected by water vapor in the air surrounding the test specimen. Pressure and temperature have very little direct effect, being mainly involved through their use in the estimating the moisture content of the air by indirect measuring techniques. Atmospheric pressure may be measured with an aneroid barometer, mercury barometer, or one of a number of electric pressure transducers, together with associated electronic circuitry to provide either analog or digital readout. Whereas it is

Conditioned Test Atmospheres

49

not really necessary for the de®nitive measurement of controlled atmospheres for paper testing, a continuous record of pressure could be obtained from an electronic system or with a barograph, that is, an aneroid barometer with a pen attachment. Because of its in¯uence on the accuracy of relative humidity measurement, temperature must be controlled and possibly measured to an accuracy of at least 0:1 C. This may be done with liquid-in-glass thermometers (the liquid usually being mercury to meet accuracy requirements) or with electrical resistance sensors. The required temperature range is rather low for thermocouple applications, but temperature-dependent resistive elements such as platinum resistors are quite suitable. Thermistors may also be used, but their nonlinear resistance versus temperature characteristic may present problems. Associated electronic circuitry can, as stated earlier, provide analog or digital indication and outputs for continuous recording, which may be very useful. The study of the amount of water vapor in the atmosphere forms a speci®c branch of physics, known as hygrometry. A variety of techniques have been developed to meet a range of applications of atmospheric moisture measurement, such as meteorological data acquisition, health and comfort air-conditioning, and precise environmental control for the processing, storage, and testing of products. As a result, the moisture content of air may be reported in a number of different and confusing terms. This chapter will explain these terms, consider the measurement ranges and accuracy required, and the various methods of measuring and controlling relative humidity for testing. An air-conditioning test room is a primary requirement for a pulp and paper research laboratory, as well as an aid to quality control in a paper mill, and possibly in a high quality printing operation. The expert knowledge and understanding of the measurement and control of relative humidity required to build and maintain a testing room at a performance level de®ned by Industry Standards is not always appreciated. The emphasis on relative humidity in this chapter is an attempt to correct this situation.

II.

RELATIVE HUMIDITY

A.

Theory and Terminology

Atmospheric moisture content is often loosely and regrettably referred to as humidity; because this quantity can be expressed in terms of volumes, masses, moles, or pressures, a confusion of de®nitions exists. Engineers de®ne humidity as the mass of water vapor per unit mass of dry air, whereas meteorologists consider this a mixing ratio and use the term ``relative humidity'' to indicate degree of saturation. The terms ``absolute humidity,'' ``speci®c humidity,'' and ``humidity ratio'' are also used, so some basic de®nitions must be initially established. The most fundamental method of measuring the moisture content of the air is the gravimetric method, which involves weighing the moisture of water vapor and weighing the dry air. The ratio of these two quantities is de®ned as the mixing ratio r, r ˆ mv =ma

…1†

where mv is the mass of water vapor in the given sample of moist air and ma is the mass of dry air in the same sample.

50

de Yong

Fig. 1

Apparatus for gravimetric determination of water vapor in air.

In practice, rather than attempting to weigh the air, a measured volume of air is aspirated through a dryer trainer as shown in Fig. 1. The increase in mass of the dryer train then corresponds to the mass of water vapor mv in a volume V of dry air. The ratio of these quantities is de®ned as absolute humidity dv such that dv ˆ mv =V

…2†

This quantity, ``absolute humidity,'' is sometimes also referred to as vapor concentration or vapor density. Equations (1) and (2) are related in terms of the partial density of dry air ; thus dv ˆ r. Another humidity ratio, speci®c humidity q, is sometimes encountered. This is de®ned as q ˆ mv =mma

…3†

where mma is the mass of moist air in the sample, i.e., mma ˆ ma ‡ mv

…4†

Whereas these forms of reporting humidity have the advantage of being clearly de®nitive, the time taken for their direct measurement would be at least 30 min, so they are not suitable for following short-term variations in moisture content. They are thus dif®cult to apply to the monitoring and control of the atmosphere in paper testing rooms and cabinets. Hence, other techniques have been developed for more rapid assessment of the quantity of water vapor in air. These usually express moisture content as relative humidity by using the concept of a ratio between the actual moisture content of the sample and the saturation moisture content of the same sample. To express relative humidity in terms of the basic quantities already described, it will be useful to de®ne some additional terms. The mole fraction of water vapor in moist air, xv , is de®ned as the ratio of the number of moles of water vapor to the total number of moles of dry air and water vapor in the sample. Thus,

Conditioned Test Atmospheres

xv ˆ

mv =Mv ma =Ma ‡ mv =Mv

51

…5†

where Mv is the molecular weight of water vapor, i.e., 18.01534 based on the C12 scale, and Ma is the apparent molecular weight of clean dry air based on a standard distribution of constituents, i.e., 28.9645 based on the C12 scale. Using these numerical values and Eq. (1), we obtain r …6† xv ˆ 0:62198 ‡ r The relative humidity of moist air with respect to water at a given pressure p and temperature T may now be de®ned by the relationship   xv Uw ˆ 100% …7† xvw p;T where xv is the mole fraction of water vapor in the given sample of moist air at pressure p and temperature T, and xvw is the mole fraction of water vapor in air that is saturated with respect to water at pressure p and temperature T. As the partial pressure of a component in an air mixture is proportional to its mole fraction, we may express the effective vapor pressure of unsaturated moist air, e, as e ˆ xv p

…8†

and similarly the saturation vapor pressure of moist air, ew , as ew ˆ xvw p Substituting Eqs. (6) and (9) into Eq. (7) yields   r p Uw ˆ 100% 0:62198 ‡ r ew p;T

…9†

…10†

which allows calculation of relative humidity from gravimetric determination of the mixing ratio. Another expression that is sometimes used is ``degree of saturation.'' This is de®ned as the ratio between the mixing ratio of a given sample at pressure p and temperature T and the saturation mixing ratio at the same pressure and temperature, i.e.,   r Dw ˆ 100% …11† rw p;T Relative humidity may also be expressed in terms of vapor pressures. Substitution of Eqs. (8) and (9) into Eq. (7) yields   e uw ˆ 100% …12† ew p;T Detailed derivations of the above equations are given in Harrison [18], with which the symbols used here correspond. Equation (12) shows that relative humidity is directly related to the partial pressure of water vapor in air, and so an ideal instrument for measuring relative

52

de Yong

humidity would be a linear pressure gauge speci®c to water. For monitoring a conditioned test atmosphere, it should have a rapid response to small changes of water vapor. Such a direct measuring instrument would be very dif®cult to build, and consequently sensors using some humidity-related property have been developed, but their stability may be questionable. The techniques of dew-point hygrometry, and wet and dry bulb psychrometry offer more reliable methods of measurement. These various methods will be discussed in Section III. B.

Standard Atmospheres

There have been three standard atmospheres used in the testing of paper and board. Their de®nition, in terms of temperature and relative humidity, has formed the basis of at least 10 recognized industry standards in different countries. These standard atmospheres are speci®ed as (a) Temperature 23  1 C, relative humidity 50  2% (b) Temperature 20  1 C, relative humidity 65  2% (c) Temperature 27  1 C, relative humidity 65  2% The atmosphere designated (a) is the preferred International Standards Organization standard, ISO 187:1990, toward which most of the major paper-producing countries have either changed or are in the process of changing. It has been in use in the United States as TAPPI T 402 om-93 and in Canada, South Africa, and Australia and New Zealand have adopted the ISO Standard, as Appita p 415s-98, which now includes the Standard Methods for Determining Relative Humidity. While the Scandinavian countries have, in the past, used both (a) and (b) because of their markets in North America, they have also been moving to the ISO Standard exclusively, together with the United Kingdom and other European countries. Most countries are expected to adopt the ISO Standard within the next few years, so any attempt to list countries and relevant standards would be somewhat futile. Atmospheres (b) and (c) may still be used in special circumstances, e.g., standard (c) in tropical situations. III.

MEASUREMENT OF RELATIVE HUMIDITY

A.

General Comments

The role of humidity in the human environment has led to a long history of instrument development associated with humidity measurement. A result of this development is a wide choice of measurement techniques. The selection of the best method for a particular application is usually made by carefully de®ning the measurement requirements and then assessing the degree of their ful®llment associated with the various methods. In paper testing rooms, where the atmospheric conditions are speci®ed in terms of standard atmospheres, the measurement requirements may be described as narrowspan, midrange humidity measurements, to a fairly high accuracy. Whereas most methods may be used, the uncertainties of measurement arising from errors inherent in the sensors are quite signi®cant in view of the accuracy required. These errors are mainly due to sampling problems and contamination. Sampling errors result from failure to locate or install the sensor so that it can detect a representative sample of

Conditioned Test Atmospheres

53

the atmosphere being measured. In this context, the proximity of the human operator, whose breath can saturate 0:001 m3 of air at 20 C, is quite signi®cant. It is therefore important to follow the manufacturer's instructions for the operation of each type of sensor. Contamination reduces accuracy and increases response time. Some sensors are permanently damaged by contamination, whereas others can be cleaned and reactivated. Some idea of the comparative uncertainties of various types of sensors, over the range of measurements in paper testing rooms, may be gained from the following list, based on data by Wexler [31]: Sensor

Uncertainty (parts per 1000)

Gravimetric train Condensation-type dew point Aspirated psychrometer Typical electrical resistance element Rolled and treated hair Saturated salt (LiCl) dew point

2 5 10±20 20±50 30±50 60

A number of other factors such as physical size, convenience of operation, provision of recording facilities, calibration requirements, reliability, and cost must also be considered. Some of the advantages and disadvantages of various methods are outlined in this chapter. For this purpose they may be divided into three broad groups, e.g., . Miscellaneous methods. .. Hygroscopic sensors. ... Thermodynamic methods.

B.

Miscellaneous Methods

These methods include such techniques as infrared absorption, attenuation of microwaves, changes in refractive index, and coated ®ber-optic sensors. Quartz crystal oscillators have been described, where moisture sensitive coatings are applied using a low-cost photo-lithographic etching technique. A paper by Morten et al. [25] claims a linear humidity versus frequency response from an acetate coated, PZT based thick ®lm, on an alumina diaphragm. A microwave interferometer system has been described by Carullo, Ferrero, and Parris [5] and a microwave differential cavity resonator by Rouleau et al. [28]. In another approach by Strembicke et al. [30], the mass changes due to the adsorption of water vapor on a porous thin ®lm deposited on a CMOS micro-machined cantilever create humidity dependent frequency shifts. These methods are often costly, and their development is usually associated with some speci®c application. The gravimetric method does not readily ®t into the other categories. It is best used for high accuracy determination, such as calibration work. Sampling may take some time, and hence its response must be considered as slow. Since 1997, a number of papers on instrumentation and measurement of relative humidity have been presented at international conferences. Rather than attempt to select and discuss appropriate references, it would seem preferable to indicate two recent conferences for personal perusal to suit individual

54

de Yong

interest, e.g., Metrology Society of AustraliaÐ2nd Biennial Conference, Victoria, Australia (1997) and 3rd International Symposium on Humidity and Moisture, N.P.L. Teddington, London, April (1998). It would seem futile to include a large selection of current references, since, methods can become rapidly superseded by later developments, and more recent references can be readily accessed on the Internet. C.

Hygroscopic Sensors

The sensors included in this group listed may produce color changes, dimensional changes in materials, or electrical changes in various elements. Color changes are dif®cult to assess quantitatively and cannot be used for recording and control. Dimensional changes occur in materials such as goldbeater's skin, cotton, hair, and nylon, and instruments using these elements are usually cheap and reliable; however, all exhibit varying degrees of hysteresis, cyclical effects, and long-term drift. They must therefore be calibrated at fairly frequent intervals. By laminating paper to a thin metal backing, the dimensional instability of paper with respect to moisture is used to produce an operation similar to that of a bimetallic strip and hence a rather simple hygrometer whose performance is, however, subject to the problems listed. Electrochemical sensors are more rugged and compact and are better suited to remote readout of relative humidity than hair-type devices. They depend on the adsorption of water to change the electrical resistance or capacitance of the sensor. Many types are available, and they are relatively inexpensive but some cannot be repaired if contaminated; all must be calibrated fairly frequently. They include Dunmore sensors, in which a wire grid is bi®lar-wound on an insulating substrate that is coated with a lithium chloride solution of controlled concentration. Because the salt is hygroscopic, the resistance of the sensor indicates the water vapor content of its surroundings over a narrow range. The Pope cell uses a similar deposited bi®lar conductive grid on a substrate of sulfonated polystyrene, whose surface resistivity is a function of the relative humidity of its environment. The Pope cell covers a wide range of relative humidity, but its resistance versus moisture content characteristic is nonlinear. To prevent polarization of these sensors, they are usually used in ACexcited Wheatstone bridges. One capacitance-type sensor uses aluminum and gold electrodes separated by aluminum oxide. Another type has an electrode etched on glass, the surface of which is then coated with an active polymer layer, and another electrode is evaporated onto the polymer surface. These types of sensors use radio frequency (RF) excitation to detect variations of capacitance with moisture content. Some general articles on these methods include those by Wiederhold [33,34], Crosby [6], Leivers and Letherman [23], and the National Physical Laboratory (NPL) Notes on Applied Science. No. 4, [24]. With the rapid progress of microelectronics, most hygroscopic sensors developed before 1990 are of mainly historic interest. A number of units based on the moisture sensitivity of polyamide (PI) ®lm have been described, e.g., an IC compatible sensor using PI ®lm fabricated on a silicon substrate by Glenn and Scheutz [14], a digital hygrometer consisting of a PI ®lm spin coated on a Si substrate, with a onechip microcomputer for capacitance to frequency conversion by Shibita et al. [29], a parallel plate capacitor with a PI ®lm dielectric either with permitivity linearly

Conditioned Test Atmospheres

55

related to relative humidity, Denton et al. [7], or as a switched capacitor with an integrated driving circuit based on a charge sharing technique Denton et al. [8]. The humidity sensitivity of sputtered TiO2 ®lms is described by Bearzotti et al. [3], and following a brief review of ceramic sensors, the principle, fabrication, and use of SI doped -hematite (-Fe2 O3 ) sintered compacts is described by Pelino et al. [27]. Many of the above type of sensors may be used in control instrumentation, but their use in monitoring the close tolerance levels of relative humidity required for Standard Paper Testing Atmospheres may be limited by their drift and hysteresis. This is discussed further in the section of Calibration of Sensors. D.

Thermodynamic Methods

The thermodynamic methods may be divided into two subgroups, i.e., dew point hygrometers and wet and dry bulb psychrometers. Both of these may be described as primary methods, because (when used in well-constructed and carefully maintained instruments) their accuracy depends on the calibration of their temperature sensors, of which a number of reliable and stable units are now available. Dew Point Hygrometers In a dew point hygrometer, a polished metal surface, exposed to a stream of air to be measured, is cooled until a temperature is reached at which dew is deposited on the surface. This temperature is called the dew point, and the relative humidity can be calculated from it. In practice, the surface may be cooled by means of an acetone and dry ice mixture, cryogenic ¯uids, mechanical refrigeration, or thermoelectric effects. Optical detection of condensate on the surface is used in both manual and automatic instruments. This detection is sometimes improved by using polarized light incident upon the surface at the Brewster angle. Whereas the dew point method has signi®cant fundamental limitations, such a condensation-type instrument is an accurate basic measuring device with a broad range and excellent reliability. However, it is complex and expensive, and maintenance of the cleanliness of the polished surface is extremely important. Whereas some instruments indicate only dew point temperature, others also measure dry bulb temperature and provide a direct reading of relative humidity to an accuracy of a few percent. Considering the theory of dew point hygrometry, an air±water vapor mixture is cooled out of contact with liquid water so that while the mixing ratio remains constant the percentage saturation increases as the temperature decreases. The temperature at which this percentage saturation reaches 100% and moisture condenses on the cooled surface is then the dew point temperature. This relation may be stated in terms of the mixing ratio r: r ˆ rwd

…13†

where rwd is the saturation mixing ratio at the dew point temperature td , and both r and rwd relate to the same pressure p. Equation (13) may be written in terms of mole fractions by using Eq. (6), so that xv ˆ xvwd

…14†

where xvwd is the mole fraction of water vapor in moist air that is saturated with respect to liquid water at pressure p and temperature td .

56

as

de Yong

Further, using the relationship expressed in Eq. (8), Eq. (14) may be rewritten e ˆ ewd

…15†

where ewd is the saturation vapor pressure at the dew point temperature. Substituting Eq. (15) into the familiar expression for relative humidity as given in Eq. (12) gives Uw ˆ

ewd  100% ew

…16†

The vapor pressures ewd and ew can be obtained from Table 1 and the relative humidity calculated. Alternatively, Tables 2, 3, and 4, each for one of the standard atmospheres, may be used. Wet and Dry Bulb Psychrometers The use of wet and dry bulb psychrometers has long been a popular method for monitoring relative humidity, mainly because of their simplicity and low cost. A wet and dry bulb psychrometer consists of a pair of matched temperature sensors, one of which is maintained dry (referred to as the dry bulb) and the other covered with a sleeve of woven material, usually rayon or cotton, that is kept wet by wicking action (referred as the wet bulb). The moisture content of the air surrounding the sensors determines the rate of evaporation of water from the wet sleeve and hence the decrease in temperature of the wet bulb. The temperature difference between the dry bulb temperature t and the wet bulb temperature t 0 is known as the wet bulb depression, and its maximum instantaneous value is used together with the corresponding dry bulb temperature to determine relative humidity from tables or graphs based on values calculated using the semiempirical psychometric equation. These instruments may depend on random air movement around the sensors, in which case they are termed unventilated. Such instruments are less accurate than those that have forced air circulation of the appropriate constant velocity across the sensors, which are described as ventilated or aspirated. A number of methods are used for temperature measurement, for example, (1) alcohol thermometers or mercury-in-glass thermometers for direct reading, (2) bimetallic elements or ®lled temperature sensors with mechanical linkages for chart recording, and (3) electrical resistance elements, thermistors, germanium diodes, or thermocouples to produce electrical output signals for either direct indication or chart recording. When electronic techniques are used, the two temperature signals can be combined to yield a direct output for relative humidity, thus removing the necessity to use graphs or tables. Where the range of measurements is limited, such as in the paper testing room, these tables may be approximated to an overall accuracy of better than 0.05% by a simple linear equation. This method has been used by de Yong [9] in the design of an electronic psychrometer. This technique has been further developed by de Yong [10] so that, for a range covering all paper testing atmospheresÐrelative humidity 45±70% and temperature 15±30 CÐan equation of the form Uw ˆ

t ‡ t 0 ‡  0t ‡ 0t 0 ‡ 1

…17†

Conditioned Test Atmospheres

57

for which ,  0 , , 0 , and are constants, may be used to yield a relative humidity output to an accuracy of 0:2%. A technique of piecewise linearization of the psychrometric chart by Bhuyan and Bhuyan [4] is shown to provide for on-line monitoring of relative humidity using thermal sensors. Additional relevant papers include one in which a miniature psychrometer using microthermocouples with the wet one coated with boron nitride as a wicking material is described and its advantages assessed by Eisner and Martonen [11], another by Kalogiros and Helmis [21] that presents a simple method for correcting the time series of a wet temperature sensor using a ®rst-order linear approximation of the heat transfer equation to yield fast response humidity measurements, and an approach by Abdul Wahab et al. [1], where the electromotive force (emf) generated by the temperature difference between wet and dry bulb thermocouple junctions is ampli®ed, converted to a digital signal, and passed to an IBM microcomputer using a programmable peripheral interface (PPI). When such psychrometers are ®tted with good quality calibrated thermometers and are designed for correct aerodynamic and thermodynamic performance, careful attention to the cleanliness of the wet wick is all that is required to maintain the rated accuracy. The Assman psychrometer, which is sometimes speci®ed in testing room standards, is fairly satisfactory except it has no provision for producing a continuous record and its response time is often inadequate. It is an aspirated instrument, but its pattern of air movement along the wet wick is somewhat dubious for accurate work. In general, the wet bulb, being a source of moisture, prevents the use of wet and dry bulb psychrometers in small, closed volumes such as conditioned instrument cabinets. Large errors can occur if the wet bulb is improperly ®tted or becomes contaminated. On the other hand, the simplicity of such a device allows easy and cheap repair. If the method of humidi®cation of the test atmosphere produces a fallout of dissolved salts, the wick can become contaminated; it is essential that a routine of wick cleaning and/or replacement be established. To obtain relative humidity from wet bulb and dry bulb temperatures, use is made of the following relationship, recommended by Ferrel [12], which states that the partial pressure of water vapor e can be calculated as e ˆ ew0

ap…t

t 0†

…18†

where t = dry bulb temperature t 0 = wet bulb temperature ew0 = saturation vapor pressure of water at the wet bulb temperature a = psychrometric coef®cient, which was originally given an empirical value of 0:66  10 3 for temperature in  C or 0:37  10 3 for temperature in  F Equation (18) is sometimes written with an additional term, i.e., e ˆ ew0

ap…t

t 0 †…1 ‡ bt 0 †

…19†

where the coef®cient b has the value 1:15  10 3 for t 0 in  C or 0:64  10 3 for t 0 in  F. However, as the 1 ‡ bt 0 factor contributes about 1% to the value of the second term over the range of interest, it may be omitted; thus, if Eq. (18) is used to substitute for e in Eq. (12), one obtains

58

de Yong

Uw ˆ

ew0 ap…t t 0 † ew  100%

…20†

Equation (20) can be used for direct calculation of relative humidity from dry bulb and wet bulb temperatures, or from dry bulb temperature and wet bulb depression, using a calculator or computer (see Appendix 2 for programs). Alternatively, tables or graphs based on such calculations may be convenient. For the standard atmospheres for paper testing, these are presented as Tables 5±7 and Figs. 6±8 in the last section of this chapter. Considerable work has been done on the signi®cance of the psychrometric coef®cient a, and the application of thermodynamic theory has led to equations in which a can be identi®ed with known physical constants. One approach, reviewed by Harrison (19), is based on a balance between the enthalpy of the moist air at ambient temperature plus the enthalpy of the water added at the wet wick and the resulting enthalpy of saturated air at the web bulb temperature. This approach leads to the following expression for the mixing ratio of the sample: Cpa hf ;0 …t t 0 †

r ˆ rw0

…21†

where rw0 = saturation mixing ratio at the wet bulb temperature t 0 Cpa = speci®c heat of dry air at constant pressure hf ;0 = latent heat of vaporization of water at 0 C When Eq. (21) is rewritten in terms of vapor pressures, we obtain= e ˆ ew0

Cpa p …t …Mv =Ma †hf ;0

t 0†

…22†

Substituting the expression for e in Eq. (22) into Eq. (12) gives Uw ˆ

ew0

Cpa p …t …Mv =Ma †hf ;0 ew

t 0†

 100%

…23†

Comparison of Eqs. (20) and (23) yields the following expression for the value of a: aˆ

Cpa ˆ 0:646  10 …Mv =Ma †hf ;0

3

…24†

where =1:00524  10 3 J=kg  C Cpa Mv =Ma = 0.62198 =2:5016  106 J=kg hf ;0 As an introduction to their compilation of psychrometric tables and charts, Zimmerman and Lavine [37] summarize the convective transfer theory. In this theory, the rate of heat transfer from the air to the wetted surface by convection and conduction is equated with the rate of loss of heat from the wet wick, calculated as the product of the rate of evaporation and the latent heat of water at the wet bulb temperature. This approach yields the equation

Conditioned Test Atmospheres

ew0

eˆ

0:3895p …t hf ;t 0

59

t 0†

…25†

where 0.3895 is an experimentally determined value at 1 atm for the ratio between the ®lm coef®cient of diffusion of water from the wet wick and the air±®lm heat transfer coef®cient for an air±water vapor system and hf ;t 0 is the latent heat of vaporization of water at the wet bulb temperature. Using a value of 588.6 cal/g for hf ;t 0 at 16 C, Eq. (25) becomes ew0

e ˆ 0:662  10 3 p…t

t 0†

…26†

hence yield a value of 0:662  10 3 for a From such theoretical and empirical considerations of a, its value is known to be strongly dependent on the velocity of air passing over the wet bulb. As a result, a number of different value of a appear in the literature. Investigations by Wylie [35] showed that the behavior of an appropriately designed wet element can be accurately predicted, and the generally accepted values of a are about 8% too high. In this work, the coef®cient a is expressed as the product of three factors, a ˆ ac Fr F

…27†

where Fr is concerned with the radiative heat transfer process and F with the effect of the molecular kinetic evaporation resistance at the water surface. These factors are slightly greater than unity. The factor ac is the hypothetical value of a for a purely convective system; hence, ac ˆ hc =kc hf ;t 0

…28†

where hc and kc are the surface heat transfer and mass transfer coef®cients, respectively, and hf ;t 0 is the latent heat of vaporization at the water surface. Using this treatment, the value of the psychrometric coef®cient a depends on element diameter, air speed, temperature, atmospheric pressure, and the water content of the air stream, so that it should be considered variable and speci®cally determined for each particular instrument con®guration. However, it appears that for air velocities greater than 3 m/s, the wet bulb depression becomes relatively independent of the ventilation rate and wet bulb geometry. In this context, Wylie and Lalas [36] gave considerable attention to specifying the essentials of a psychrometer that allows accurate determination of the psychrometric coef®cient and hence con®dent use of Eq. (20). A brief description of the resulting design requirements particular to an instrument for use in paper testing rooms is as follows. Both wet and dry bulb elements should be in the form of thin-walled stainless steel tubes with an overall diameter of 4:5  0:3 mm. They should be subjected to a transverse air stream having an effective velocity of 4.5 m/s across the wet bulb element. They are best arranged with their axes horizontal in a horizontally ¯owing air stream, with the wet bulb element offset from and downstream of the dry bulb element. To avoid errors in temperature measurement, these cylinders should be of suf®cient length to be considered effectively endless, that is, the measured temperature should be uniform for a few centimeters on either side of the actual sensing position. Mercury thermometers or electric temperature sensors can be used to measure both wet bulb and dry bulb temperatures to an accuracy of 0:05 C or better. It

60

de Yong

should be noted that this corresponds to an estimated error of 0:071 C in wet bulb depression; it can be observed from Tables 5±7 (Section VII) that this level of accuracy leads to an uncertainty of 0:5% RH. The wet bulb element should have as smooth and clean a water surface as possible. It is inevitable that some of the ®bers of the woven wick will project through the water surface, but with suitable materials the effect should be negligible. The wet bulb element can be considered adequately wet when it presents a glistening appearance under strong illumination. The surface will then be effectively black to the incident thermal radiation corresponding to the temperature of the air stream. A capillary feed of water to the wick is quite satisfactory provided that there is no more than 4 cm length of free wick between the end of the cylindrical element and the supply water surface. The water surface itself should be no more than 2 cm below the axis of the wet bulb element. Both the cleanliness of the wet wick and the purity of the water supply are extremely important. The more detailed discussion of wet and dry bulb psychrometers than of other instruments is considered appropriate because of their suitability and hence extensive use in monitoring test atmospheres for paper and board. IV.

CONTROL OF RELATIVE HUMIDITY

Whereas the actual moisture content of the test atmosphere is the prime concern in paper testing, its determination, as speci®ed in terms of relative humidity, requires control of both temperature and relative humidity. The control of these quantities may be either interrelated or completely separate, depending on the method used to achieve the required atmospheric moisture content. In either case, the control of temperature can follow the well-established practices of control system design with appropriate feedback. Simple on±off control may be used, but a closer tolerance will usually be achieved if proportional control is employed. The techniques of temperature sensing and control are continuously being updated through improvements in such items as thermistors, semiconductor diodes, and integrated circuits, so current literature will be the best source of speci®c details. The control of relative humidity will depend on which of the three broadly de®ned methods is used for humidi®cation and dehumidi®cation. These methods may be considered as . Addition or removal of water vapor .. The two-temperature or dew point saturation and reheat method ... The two-pressure method, i.e., saturation at a high pressure and transfer to a test chamber at lower pressure Because the third method involves pressure vessels, its practicality is limited to small volumes such as conditioning and calibrating cabinets. Because no appreciable thermal mass is involved, this method has the advantage of providing rapid changes of humidity over a wide range. Papers by Folland and Sparks [13] and Hasegawa and Little [20] describe two-pressure generators for calibrating hygrometers. Most commercially available packaged units operate according to the ®rst method, and this approach has become fairly well established for paper testing rooms. Its basic con®guration as an on±off system is shown in Fig. 2. Each of the

Fig. 2 Control method 1, on±off control.

Conditioned Test Atmospheres 61

62

de Yong

required functionsÐheating, cooling, humidi®cation, and dehumidi®cationÐis usually controlled separately by a temperature sensor and a humidity sensor. Although humidifying and dehumidifying operations can produce changes in temperature, and changes in temperature themselves affect humidity, the control mechanisms may not be interconnected, and undesirable cycling effects may result, particularly if close control is attempted. Such a system is usually improved by replacing some of the on±off operations by either multistage or proportional control. If heating is electrical, a number of individual elements may be switched on in succession, or continuous variation of power may be applied using silicon-controlled recti®ers. When steam or hot water is used for heating, the circuitry can be arranged to drive a motorized valve to give proportional control. Similarly, proportional control of cooling may be achieved by using a chilled water supply instead of on±off refrigeration. If chilled water is not available on-line, a small local refrigeration unit with brine storage would provide the same facility but at some additional cost. However, the improvement in temperature control should be considered, as well as the possible duplicate use of chilled water in the dehumidi®er coils, as shown in Fig. 3. Relatively simple temperature control systems appear to provide satisfactory results within the limits set by the standards, and it is usually the close humidity control that is more dif®cult to achieve. Humidi®cation is most frequently attempted by injecting atomized water droplets, or steam, into the air circulation system. For dehumidi®cation, the air is passed over the evaporator coils of a refrigeration unit so that, as the air temperature is lowered, the dew point is reached and condensation occurs on the surfaces of the coil assembly. This condensate is then removed by a drainage system. The reduction in air temperature associated with condensation must be corrected by reheating; in the interests of ef®ciency the design may include the use of the condenser coils of the refrigerator to supply most of the required heat. These operations to control humidity are performed in essentially an on±off mode. The on±off feature, together with some thermal lag in the system, results in a time delay between the initiation of an operation and its effect on the system. When this delay is added to the time taken for the air to circulate from the humidi®er or the refrigerator coils, both of which are usually in the duct work, and the humidity sensor, which may be in the room or preferably in the return air stream, an unstable control situation results. This is evidenced by ¯uctuations of increasing amplitude that often exceed reasonable control limits. To prevent this excessive cycling, the control band must be widened, often beyond the desired limits. The input of a normally controlled quantity is sometimes maintained constant, and the opposite quantity is proportionally controlled against this to produce the required conditionÐfor example, constant refrigeration with proportioned control of heating to maintain the desired temperature. This may be combined with continuous dehumidi®cation from the same refrigeration unit and a controlled steam generator to provide humidi®cation as required. Whereas such a system has good performance characteristics, it is quite wasteful of energy and hence expensive to operate. Precise control within a speci®c tolerance such as 2% RH requires extremely good equipment and would be impractical with many installations. Any deterioration of the sensors or the system may result in testing room perfor-

Fig. 3

Control method 1, part proportional control.

Conditioned Test Atmospheres 63

64

de Yong

mance falling outside the required limits. In situations where the required humidity is close to external ambient conditions, a long-term drift followed by a rapid control correction will give rise to an uneven, sawtooth pro®le of the humidity level. The moisture content of paper in such a situation might be expected to differ from that of paper subjected to a regular short-term humidity cycle, with a consequent difference in test results. This undesirable cyclical performance may be overcome by rapid additions of moisture to bring the dehumidifying control into more frequent operation, but the problem remains basically one of the limitations of on±off control. The two-temperature method, although regarded as expensive to operate because a large cooling capacity must be provided, is nonetheless becoming more widely accepted owing to its better and more reliable performance. Its basic arrangement is shown in Fig. 4. Air is withdrawn from the room and cooled to a dew point temperature corresponding to the required relative humidity at the room temperature as indicated in Fig. 5. It is then spray-saturated at this dew point temperature, reheated to the required room temperature, and returned to the testing room. Varying amounts of fresh air may be admitted as desired. Dew point control may also be applied via the saturator water temperature. Recirculation of excess spray water with makeup from the chilled water supply, under control of the saturation temperature sensor and associated control elements, may be used to improve ef®ciency. Because all control operations are based on two temperature measurements, they can be proportionally controlled with less control bandwidth than independent on±off control operation. When using mechanical refrigeration for humidity control, the evaporator coil is used for cooling and the condenser coil is used for heating as appropriate, with the air ¯ows over the coils regulated by ¯apper valves and bypass ducts. Details of the thermal considerations of such a system are beyond the scope of this brief outline and are best treated as a specialist area of refrigeration engineering. V.

HUMIDITY CONTROLLED ENCLOSURES

A.

Design and Performance of Paper Testing Rooms

For those who become involved in planning, specifying, and operating an air-conditioned testing room but must leave the design and construction to architects and contractors, the most important fact to bear in mind is that a number of basic requirements should be clearly established. This will prevent subsequent problems of unsatisfactory room performance that may otherwise arise in the commissioning, acceptance, and long-term use of the facility. Architects tend to consider such a project in terms of building design, in which they are undisputed experts. They may have experience with the general air-conditioning of buildings, but they may not fully appreciate the special close tolerance requirements, particularly with respect to relative humidity (RH), associated with paper testing rooms. Such a facility must be considered quite distinct from an of®ce block or even a computer installation. It is not unusual to see a speci®cation of 5% RH associated with 2 C temperature control, whereas 0:5 C is a more realistic tolerance for such a level of relative humidity control. If possible, all design speci®cations should be approved by the user before being issued for tender. Those who

Fig. 4

Control method 2, proportional control.

Conditioned Test Atmospheres 65

66

de Yong

Fig. 5

Relative humidity curves and related data for the standard atmospheres.

will be using the room should have some further input, via discussion with the architect and the air-conditioning contractors, to clearly establish the importance of close tolerance performance. Every attempt should be made to isolate the room from ambient conditions. Thus: 1. 2.

3.

4.

5.

If possible, either the room should have no external walls or its external wall area should be minimized. All walls should be extremely well insulated against both heat and moisture transfer. This is of the utmost importance for any external walls. The thermal insulation should be protected by a moisture-vapor barrier on both sides to prevent vapor condensation within the insulation. Windows should be avoided, but, if they are included, double glazing is a minimum requirement. Windows on external walls are especially undesirable, and if such walls receive incident sunlight, it may be impossible to control the room atmosphere within the narrowly prescribed limits. All internal surfaces, i.e., walls, ¯oor, and ceiling, should be made impervious to water vapor transfer by the use of appropriate surface coatings. It should be noted in this context that most building materials are quite porous, and uncoated areas behind cupboards or benches place an unnecessary load on the control system. A double-door entrance airlock is always desirable to lessen disturbances to room conditions when personnel enter or exit. Where this may con¯ict with safety requirements, a separate exit door ®tted with an approved edge seal and openable only from the inside may be necessary. A small positive internal air pressure should be maintained. The philosophy of providing a control system with a large capacity to overcome internally or externally

Conditioned Test Atmospheres

67

induced variations can lead to excessive correction and undesirable oscillatory performance. Some aspects of the overall health and safety regulations that must be satis®ed are as follows: 1. 2. 3. 4. 5.

The air-moving equipment must have ®lters to remove dust particles and paper ®ber lint, and no chemical fumes should ever be present in the room. The room size should provide for an air space of 30±60 m3 per person. There should be 15±30 changes of room air per hour, with about 8% fresh air. The level of illumination for the work areas should be as recommended for ``close work.'' The noise level within the room from all sources should not exceed 65 dB on a continuous basis.

The air circulation pattern in the room is extremely important. With the fairly common system of an outlet duct suspended from the ceiling and a return air duct along one or more walls near ¯oor level, the location of registers in these ducts must be carefully planned. After workbenches and ®ttings have been installed, the outlet or inlet ¯ow rate of each register should be adjusted for both magnitude and direction to provide, as nearly as possible, uniform air movement around the room and also the optimum penetration of corners and dead spots. Although the air¯ow pattern may be checked with an anemometer, a simple technique of suspending a number of small pieces of facial tissue in a grid pattern from the ceiling with appropriate lengths of cotton thread provides a good indication of the general air movement in the room. B.

Controlled Environment Cabinets

Although the major requirement for a standard test atmosphere is associated with the routine evaluation of paper properties, in terms of either quality control or comparative assessment of research projects, nonstandard atmospheres are often useful for specialist applications or research work. Conditions of higher or lower or cyclical relative humidity may be required. Such conditions would be dif®cult to provide in a typical paper testing room with a volume of 100±200 m3 . In such circumstances, environmental cabinets with volumes of 1±2 m3 are available. The same techniques of heating, refrigeration, condensation, and evaporation are generally used to produce the desired atmosphere. Unfortunately, much of the equipment required for these techniques cannot be reduced in proportion to the working volume, so such facilities can be quite large and expensive. Some years ago, while still at CSIRO, the author devised a small and simpler cabinet for hygroexpansivity experiments. This cabinet was constructed from Perspex and had a small tangential fan to circulate air within the unit. Although reasonably made, the cabinet was not hermetically sealed, so a constant supply of either wet or dry air was injected into it to either raise or lower the humidity between

68

de Yong

controlled limits, at a rate of about 5% RH per minute. The humidity level was measured with a PCRC-11 sulfonated polystyrene sensor, which, with associated circuitry, provided a corresponding voltage output. This signal was fed through a data logger and HP interface to an HP-97 programmable calculator, from which output signals were obtained when either the high or low limits were exceeded. These signals operated solenoid valves on a compressed air line to direct the air¯ow either through a water saturator column or a heatless regenerative dryer using Union Carbide molecular sieve adsorbent. By reducing the range between the high and low limits, it was also possible to maintain a constant humidity within a few percent. This unit was operated in an air-conditioned laboratory where temperature control was not required. If independent temperature control is required, small thermoelectric heat pump modules could be build into the cabinet.

VI.

CALIBRATION OF SENSORS

Although a number of references are made in this chapter to the calibration of sensors for moisture content measurement, some repetition may be useful under this speci®c heading. A paper by Parkes [26] gives a brief but useful review of calibration and standards in general. A more recent presentation by Hardy [17] describes an automated approach to precise humidity generation using an HP Series 300 computer and an HP 3852 data acquisition/control system. As discussed in Section III.B, the gravimetric method of calibration is the most accurate and also the most demanding in terms of time and equipment. It is therefore not readily available to many users of humidity monitoring equipment. Hygroscopic sensors, discussed in Sections III.B and III.C, such as resistance and capacitance types, are usually calibrated at ®xed points by placing them in an enclosure with a saturated aqueous solution of a binary salt in a shallow tray. Provided the enclosure is well sealed, the air above the salt solution will reach an equilibrium moisture content, the value of which is obtainable from tables such as those published by Greenspan [16]. A selection of values appropriate to the range of paper testing atmospheres is given in Table 8 (Section VII). With some salts, a signi®cant dependence on temperature must be considered. It should also be noted that the stated accuracies of the relative humidity above such solutions is only about 1%, so the calibration accuracy of this method is similarly limited. Another method to obtain an accurately known humidity in a small enclosure has been developed in Australia, at CSIRO Division of Applied Physics, in the form of a low cost, rapid response controlled humidity generator. It uses an existing technique of mixing measured ¯ows of wet and dry air that up until now presented a problem of accurately controlling the two ¯ows. This has been overcome by using critical ¯ow nozzles (CFNs), where the ¯ow reaches the speed of sound in the throat of a ``sonic nozzle'' and therefore is unaffected by downstream pressure. By using diverter valves to direct ¯ows through a set of matched CFNs, the desired multiples of humidity can be accurately set. Both dew point and psychrometric instruments, if well designed and constructed and correctly maintained according to manufacturer's speci®cations, require only periodic calibration of their temperature sensors to ensure their speci®ed accu-

Conditioned Test Atmospheres

69

racy. The required temperature calibration to 0:1 C is not beyond the capability of many laboratories. VII. A.

TABLES AND GRAPHS Vapor Pressure Table

Table 1 gives values of the saturation vapor pressure of pure water vapor over liquid water, in millibars, for temperatures of 0±40 C in 0:1 C increments. It is a reproduction of Table 7 of Wexler [32] with values converted to millibars and truncated to three decimal places. When this table is used for relative humidity calculations, the value of saturation vapor pressure required should be strictly that of saturated water vapor mixed with air. The difference between these two situations was investigated by Goff and Gratch [15], and their data indicate that a correction of 0.5% would be required over a temperature range of 0±30 C and a pressure range of 900±1100 mbar. However, this order of correction has a negligible effect on the calculation of dew point tables (Tables 2±4) and adds less than 0.1% to the relative humidity value in Tables 5±7. It has thus been ignored. With the increased use of personal computers and small hand-held programmable calculators, the direct calculation of vapor pressure from temperature may be preferred to dependence on tabular data. Wexler's Table 7 contains saturated vapor pressure values derived from a thermodynamically based eight-term equation. Comparison of these values with experimental data over a temperature range of 12±29 C, covering standard atmospheres (a), (b), and (c) (see Section II.B), shows agreement with 43±60 ppm. In discussing the simpli®cation of this equation by reducing the number of terms, Wexler provides two empirical equations, of ®ve and four terms, respectively. Similar comparison between eight- and four-term calculated values yields agreement within 580±670 ppm. Wexler's four-term equation is of the form ln…ew † ˆ g1 T

1

‡ g2 ‡ g3 T ‡ g4 T 2

…29†

where T is in kelvin and g1 , g2 , g3 , and g4 are constants. Programs for Hewlett Packard HP-15C and an IBM-compatible PC, using Eq. (29) as a subroutine, are listed in Appendix 2 and discussed further in the next two subsections. B.

Dew Point Tables

Tables 2±4 give values of relative humidity corresponding to dry bulb temperature and dew point temperature, over ranges corresponding to the standard atmospheres for paper testing. They were calculated using Eq. (16). These tables were originally prepared using an HP-97 programmable calculator and a seven-term vapor pressure versus temperature equation attributed to Keyes [22] to allow repetitive programmed calculation of each table value without the need for manual entry of vapor pressure data. The accuracy of this method was compared with selected values calculated using vapor pressure table ®gures, and agreement was exact to the accuracy quoted. It should be noted that these tables indicate that an error of 0:1 C in either dry bulb temperature or dew point temperature corresponds (text continues on p. 78)

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Temp. … C†

6.112 6.571 7.059 7.580 8.135 8.725 9.352 10.019 10.728 11.481 12.279 13.127 14.025 14.977 15.986 17.053 18.183 19.379 20.641

0.0 6.157 6.618 7.110 7.634 8.192 8.786 9.417 10.088 10.801 11.558 12.362 13.214 14.118 15.075 16.090 17.163 18.299 19.501 20.771

0.1

Table 1 Vapor Pressure of Water (mbar)

6.202 6.666 7.161 7.688 8.250 8.847 9.482 10.158 10.875 11.637 12.445 13.302 14.211 15.174 16.194 17.274 18.417 19.625 20.902

0.2 6.247 6.714 7.212 7.743 8.308 8.909 9.548 10.227 10.949 11.715 12.528 13.391 14.305 15.274 16.300 17.385 18.534 19.749 21.034

0.3 6.292 6.762 7.264 7.798 8.366 8.971 9.614 10.298 11.024 11.795 12.612 13.480 14.399 15.374 16.405 17.498 18.653 19.875 21.166

0.4 6.338 6.811 7.316 7.853 8.425 9.034 9.681 10.368 11.099 11.874 12.697 13.569 14.494 15.474 16.512 17.610 18.772 20.001 21.299

0.5 6.384 6.850 7.368 7.909 8.484 9.097 9.748 10.439 11.174 11.954 12.782 13.660 14.590 15.575 16.619 17.723 18.892 20.127 21.433

0.6

6.430 6.910 7.421 7.965 8.544 9.160 9.815 10.511 11.250 12.035 12.867 13.750 14.686 15.677 16.727 17.837 19.012 20.255 21.568

0.7

6.477 6.959 7.473 8.021 8.604 9.224 9.883 10.583 11.327 12.116 12.953 13.841 14.782 15.779 16.835 17.952 19.133 20.383 21.703

0.8

6.524 7.009 7.527 8.078 8.664 9.288 9.951 10.655 11.403 12.197 13.040 13.933 14.880 15.882 16.944 18.067 19.255 20.511 21.839

0.9

70 de Yong

Source: Ref. 32.

19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40

21.976 23.385 24.874 26.444 28.101 29.847 31.687 33.626 35.667 37.815 40.075 42.452 44.950 47.575 50.332 53.227 56.265 59.451 62.793 66.297 69.968 73.813

22.113 23.531 25.027 26.606 28.271 30.027 31.877 33.825 35.877 38.036 40.308 42.696 45.207 47.845 50.615 53.524 56.576 59.778 63.136 66.656 70.344 74.207

22.252 23.677 25.181 26.769 28.443 30.207 32.067 34.026 36.088 38.258 40.541 42.942 45.465 48.116 50.900 53.823 56.890 60.107 63.481 66.017 70.722 74.602

22.391 23.824 25.336 26.932 28.615 30.389 32.259 34.227 36.300 38.481 40.776 43.189 45.724 48.389 51.186 54.123 57.205 60.437 63.827 67.380 71.102 75.001

22.530 23.971 25.492 27.096 28.788 30.571 32.451 34.430 36.513 38.706 41.012 43.436 45.985 48.662 51.473 54.424 57.521 60.769 64.175 67.744 71.484 75.401

22.671 24.120 25.649 27.261 28.962 30.755 32.644 34.633 36.728 38.931 41.249 43.686 46.247 48.937 51.762 54.727 57.839 61.103 64.524 68.111 71.868 75.803

22.812 24.269 25.806 27.428 29.137 30.940 32.839 34.838 36.943 39.158 41.487 43.936 46.510 49.213 52.052 55.032 58.158 61.438 64.875 68.479 72.253 76.206

22.954 24.419 25.964 27.595 29.313 31.125 33.034 35.044 37.159 39.385 41.727 44.188 46.774 49.491 52.344 55.338 58.479 61.774 65.228 68.848 72.640 76.612

23.097 24.569 26.123 27.762 29.490 31.312 33.230 35.251 37.377 39.614 41.967 44.441 47.040 49.770 52.637 55.645 58.802 62.112 65.583 69.220 73.029 77.019

23.241 24.721 26.283 27.931 29.668 31.499 33.428 35.458 37.596 39.844 42.209 44.695 47.307 50.050 52.931 66.954 59.126 62.452 65.939 69.593 73.420 77.428

Conditioned Test Atmospheres 71

11.4 54.2 53.9 53.5 53.2 52.9 52.6 52.2 51.9 51.6 51.3 51.0 50.7 50.4 50.0 49.7 49.4 49.1 48.8

Dry bulb ( C)

21.0 21.1 21.2 21.3 21.4 21.5 21.6 21.7 21.8 21.9 22.0 22.1 22.2 22.3 22.4 22.5 22.6 22.7

54.5 54.2 53.9 53.6 53.2 52.9 52.6 52.3 51.9 51.6 51.3 51.0 50.7 50.4 50.1 49.8 49.5 49.2

11.5 54.9 54.6 54.2 53.9 53.6 53.3 52.9 52.6 52.3 52.0 51.6 51.3 51.0 50.7 50.4 50.1 49.8 49.5

11.6 55.3 54.9 54.6 54.3 53.9 53.6 53.3 53.0 52.6 52.3 52.0 51.7 51.4 51.1 50.7 50.4 50.1 49.8

11.7

Table 2 Relative Humidity (%), Standard Atmosphere (a)a

55.6 55.3 55.0 54.6 54.3 54.0 53.6 53.3 53.0 52.7 52.3 52.0 51.7 51.4 51.1 50.8 50.5 50.2

11.8 56.0 55.7 55.3 55.0 54.7 54.3 54.0 53.7 53.3 53.0 52.7 52.4 52.0 51.7 51.4 51.1 50.8 50.5

11.9 56.4 56.0 55.7 55.4 55.0 54.7 54.3 54.0 53.7 53.4 53.0 52.7 52.4 52.1 51.8 51.4 51.1 50.8

12.0

Dew point ( C)

56.8 56.4 56.1 55.7 55.4 55.0 54.7 54.4 54.0 53.7 53.4 53.1 52.7 52.4 52.1 51.8 51.5 51.2

12.1

57.1 56.8 56.4 56.1 55.7 55.4 55.1 54.7 54.4 54.1 53.7 53.4 53.1 52.8 52.4 52.1 51.8 51.5

12.2

57.5 57.2 56.8 56.5 56.1 55.8 55.4 55.1 54.8 54.4 54.1 53.8 53.4 53.1 52.8 52.5 52.2 51.8

12.3

57.9 57.5 57.2 56.8 56.5 56.1 55.8 55.5 55.1 54.8 54.4 54.1 53.8 53.5 53.1 52.8 52.5 52.2

12.4

58.3 57.9 57.6 57.2 56.9 56.5 56.2 55.8 55.5 55.1 54.8 54.5 54.1 53.8 53.5 53.2 52.8 52.5

12.5

58.7 58.3 57.9 57.6 57.2 56.9 56.5 56.2 55.8 55.5 55.2 54.8 54.5 54.2 53.8 53.5 53.2 52.9

12.6

72 de Yong

a

See Section II.B.

22.8 22.9 23.0 23.1 23.2 23.3 23.4 23.5 23.6 23.7 23.8 23.9 24.0 24.1 24.2 24.3 24.4 24.5 24.6 24.7 24.8 24.9 25.0

48.5 48.3 48.0 47.7 47.4 47.1 46.8 46.5 46.3 46.0 45.7 45.4 45.2 44.9 44.6 44.4 44.1 43.8 43.6 43.3 43.0 42.8 42.5

48.9 48.6 48.3 48.0 47.7 47.4 47.1 46.8 46.6 46.3 46.0 45.7 45.5 45.2 44.9 44.6 44.4 44.1 43.9 43.6 43.3 43.1 42.8

49.2 48.9 48.6 48.3 48.0 47.7 47.4 47.2 46.9 46.6 46.3 46.0 45.8 45.5 45.2 44.9 44.7 44.4 44.1 43.9 43.6 43.4 43.1

49.5 49.2 48.9 48.6 48.3 48.0 47.8 47.5 47.2 46.9 46.6 46.3 46.1 45.8 45.5 45.2 45.0 44.7 44.4 44.2 43.9 43.6 43.4

49.9 49.6 49.3 49.0 48.7 48.4 48.1 47.8 47.5 47.2 46.9 46.6 46.4 46.1 45.8 45.5 45.3 45.0 44.7 44.5 44.2 43.9 43.7

50.2 49.9 49.6 49.3 49.0 48.7 48.4 48.1 47.8 47.5 47.2 47.0 46.7 46.4 46.1 45.8 45.6 45.3 45.0 44.8 44.5 44.2 44.0

50.5 50.2 49.9 49.6 49.3 49.0 48.7 48.4 48.1 47.8 47.6 47.3 47.0 46.7 46.4 46.1 45.9 45.6 45.3 45.1 44.8 44.5 44.3

50.8 50.5 50.2 49.9 49.6 49.3 49.0 48.7 48.4 48.2 47.9 47.6 47.3 47.0 46.7 46.5 46.2 45.9 45.6 45.4 45.1 44.8 44.5

51.2 50.9 50.6 50.3 50.0 49.7 49.4 49.1 48.8 48.5 48.2 47.9 47.6 47.3 47.0 46.8 46.5 46.2 45.9 45.7 45.4 45.1 44.8

51.5 51.2 50.9 50.6 50.3 50.0 49.7 49.4 49.1 48.8 48.5 48.2 47.9 47.6 47.4 47.1 46.8 46.5 46.2 46.0 45.7 45.4 45.1

51.9 51.5 51.2 50.9 50.6 50.3 50.0 49.7 49.4 49.1 48.8 48.5 48.2 48.0 47.7 47.4 47.1 46.8 46.5 46.3 46.0 45.7 45.4

52.2 51.9 51.6 51.3 51.0 50.6 50.3 50.0 49.7 49.4 49.1 48.8 48.6 48.3 48.0 47.7 47.4 47.1 46.8 46.6 46.3 46.0 45.7

52.5 52.2 51.9 51.6 51.3 51.0 50.7 50.4 50.1 49.8 49.5 49.2 48.9 48.6 48.3 48.0 47.7 47.4 47.2 46.9 46.6 46.3 46.0

Conditioned Test Atmospheres 73

12.6 70.7 70.2 69.8 69.4 68.9 68.5 68.1 67.6 67.2 66.8 66.4 66.0 65.6 65.2 64.8 64.4 64.0 63.6 63.2

18.0 18.1 18.2 18.3 18.4 18.5 18.6 18.7 18.8 18.9 19.0 19.1 19.2 19.3 19.4 19.5 19.6 19.7 19.8

71.1 70.7 70.3 69.8 69.4 68.9 68.5 68.1 67.7 67.2 66.8 66.4 66.0 65.6 65.2 64.8 64.4 64.0 63.6

12.7 71.6 71.2 70.7 70.3 69.8 69.4 69.0 68.5 68.1 67.7 67.3 66.8 66.4 66.0 65.6 65.2 64.8 64.4 64.0

12.8 72.1 71.6 71.2 70.7 70.3 69.9 69.4 69.0 68.6 68.1 67.7 67.3 66.9 66.5 66.0 65.6 65.2 64.8 64.4

12.9

Relative Humidity (%) Standard Atmosphere (b)a

Dry bulb ( C)

Table 3

72.6 72.1 71.7 71.2 70.8 70.3 69.9 69.4 69.0 68.6 68.1 67.7 67.3 66.9 66.5 66.1 65.6 65.2 64.8

13.0 73.0 72.6 72.1 71.7 71.2 70.8 70.3 69.9 69.5 69.0 68.6 68.2 67.7 67.3 66.9 66.5 66.1 65.7 65.3

13.1 73.5 73.1 72.6 72.1 71.7 71.2 70.8 70.4 69.9 69.5 69.0 68.6 68.2 67.6 67.3 66.9 66.5 66.1 65.7

13.2

Dew point ( C)

74.0 73.5 73.1 72.6 72.2 71.7 71.3 70.8 70.4 69.9 69.5 69.1 68.6 68.2 67.8 67.4 66.9 66.5 66.1

13.3

74.5 74.0 73.5 73.1 72.6 72.2 71.7 71.3 70.8 70.4 70.0 69.5 69.1 68.7 68.2 67.8 67.4 67.0 66.6

13.4

75.0 74.5 74.0 73.6 73.1 72.6 72.2 71.7 71.3 70.9 70.4 70.0 69.5 69.1 68.7 68.3 67.8 67.4 67.0

13.5

75.5 75.0 74.5 74.0 73.6 73.1 72.7 72.2 71.8 71.3 70.9 70.4 70.0 69.6 69.1 68.7 68.3 67.8 67.4

13.6

75.9 75.5 75.0 74.5 74.1 73.6 73.1 72.7 72.2 71.8 71.3 70.9 70.4 70.0 69.6 69.1 68.7 68.3 67.9

13.7

76.4 76.0 75.5 75.0 74.5 74.1 73.6 73.2 72.7 72.2 71.8 71.4 70.9 70.5 70.0 69.6 69.2 68.7 68.3

13.8

74 de Yong

a

See Section II.B.

19.9 20.0 20.1 20.2 20.3 20.4 20.5 20.6 20.7 20.8 20.9 21.0 21.1 21.2 21.3 21.4 21.5 21.6 21.7 21.8 21.9 22.0

62.8 62.4 62.0 61.6 61.2 60.9 60.5 60.1 59.7 59.4 59.0 58.7 58.3 57.9 57.6 57.2 56.9 56.5 56.2 55.8 55.5 55.2

63.2 62.8 62.4 62.0 61.6 61.3 60.9 60.5 60.1 59.8 59.4 59.0 58.7 58.3 58.0 57.6 57.3 56.9 56.6 56.2 55.9 55.5

63.6 63.2 62.8 62.4 62.0 61.7 61.3 60.9 60.5 60.2 59.8 59.4 59.1 58.7 58.3 58.0 57.6 57.3 56.9 56.6 56.2 55.9

64.0 63.6 63.2 62.8 62.5 62.1 61.7 61.3 60.9 60.6 60.2 59.8 59.4 59.1 58.7 58.4 58.0 57.7 57.3 57.0 56.6 56.3

64.4 64.0 63.6 63.3 62.9 62.5 62.1 61.7 61.3 61.0 60.6 60.2 59.8 59.5 59.1 58.7 58.4 58.0 57.7 57.3 57.0 56.6

64.9 64.5 64.1 63.7 63.3 62.9 62.5 62.1 61.7 61.4 61.0 60.6 60.2 59.9 59.5 59.1 58.8 58.4 58.1 57.7 57.4 57.0

65.3 64.9 64.5 64.1 63.7 63.3 62.9 62.5 62.1 61.8 61.4 61.0 60.6 60.3 59.9 59.5 59.2 58.8 58.4 58.1 57.7 57.4

65.7 65.3 64.9 64.5 64.1 63.7 63.3 62.9 62.5 62.2 61.8 61.4 61.0 60.7 60.3 59.9 59.5 59.2 58.8 58.5 58.1 57.8

66.1 65.7 65.3 64.9 64.5 64.1 63.7 63.3 63.0 62.6 62.2 61.8 61.4 61.0 60.7 60.3 59.9 59.6 59.2 58.8 58.5 58.1

66.6 66.2 65.8 65.4 64.9 64.5 64.2 63.8 63.4 63.0 62.6 62.2 61.8 61.4 61.1 60.7 60.3 60.0 59.6 59.2 58.9 58.5

67.0 66.6 66.2 65.8 65.4 65.0 64.6 64.2 63.8 63.4 63.0 62.6 62.2 61.8 61.5 61.1 60.7 60.4 60.0 59.6 59.3 58.9

67.5 67.0 66.6 66.2 65.8 65.4 65.0 64.6 64.2 63.8 63.4 63.0 62.6 62.3 61.9 61.5 61.1 60.7 60.4 60.0 59.6 59.3

67.9 67.5 67.1 66.6 66.2 65.8 65.4 65.0 64.6 64.2 63.8 63.4 63.0 62.7 62.3 61.9 61.5 61.1 60.8 60.4 60.0 59.7

Conditioned Test Atmospheres 75

19.2 70.2 69.8 69.4 69.0 68.6 68.2 67.8 67.4 67.0 66.6 66.2 65.8 65.4 65.0 64.6 64.2 63.9 63.5 63.1

Dry bulb ( C)

25.0 25.1 25.2 25.3 25.4 25.5 25.6 25.7 25.8 25.9 26.0 26.1 26.2 26.3 26.4 26.5 26.6 26.7 26.8

70.7 70.2 69.8 69.4 69.0 68.6 68.2 67.8 67.4 67.0 66.6 66.2 65.8 65.4 65.0 64.6 64.3 63.9 63.5

19.3 71.1 70.7 70.3 69.8 69.4 69.0 68.6 68.2 67.8 67.4 67.0 66.6 66.2 65.8 65.4 65.1 64.7 64.3 63.9

19.4 71.5 71.1 70.7 70.3 69.9 69.4 69.0 68.6 68.2 67.8 67.4 67.0 66.6 66.2 65.8 65.5 65.1 64.7 64.3

19.5

Table 4 Relative Humidity (%) Standard Atmosphere (c)a

72.0 71.6 71.1 70.7 70.3 69.9 69.5 69.1 68.6 68.2 67.8 67.4 67.0 66.6 66.3 65.9 65.5 65.1 64.7

19.6 72.4 72.0 71.6 71.2 70.7 70.3 69.9 69.5 69.1 68.7 68.3 67.9 67.5 67.1 66.7 66.3 65.9 65.5 65.1

19.7 72.9 72.5 72.0 71.6 71.2 70.8 70.3 69.9 69.5 69.1 68.7 68.3 67.9 67.5 67.1 66.7 66.3 65.9 65.5

19.8

Dew point ( C)

73.3 72.9 72.5 72.0 71.6 71.2 70.8 70.4 69.9 69.5 69.1 68.7 68.3 67.9 67.5 67.1 66.7 66.3 65.9

19.9

73.8 73.4 72.9 72.5 72.1 71.6 71.2 70.8 70.4 70.0 69.5 69.1 68.7 68.3 67.9 67.5 67.1 66.7 66.3

20.0

74.3 73.8 73.4 72.9 72.5 72.1 71.7 71.2 70.8 70.4 70.0 69.6 69.2 68.7 68.3 67.9 67.5 67.1 66.7

20.1

74.7 74.3 73.8 73.4 73.0 72.5 72.1 71.7 71.2 70.8 70.4 70.0 69.6 69.2 68.8 68.4 68.0 67.6 67.2

20.2

75.2 74.7 74.3 73.8 73.4 73.0 72.5 72.1 71.7 71.3 70.8 70.4 70.0 69.6 69.2 68.8 68.4 68.0 67.6

20.3

75.6 75.2 74.8 74.3 73.9 73.4 73.0 72.6 72.1 71.7 71.3 70.9 70.4 70.0 69.6 69.2 68.8 68.4 68.0

20.4

76 de Yong

a

See Section II.B.

26.9 27.0 27.1 27.2 27.3 27.4 27.5 27.6 27.7 27.8 27.9 28.0 28.1 28.2 28.3 28.4 28.5 28.6 28.7 28.8 28.9 29.0

62.8 62.4 62.0 61.7 61.3 60.9 60.6 60.2 59.9 59.5 59.2 58.8 58.5 58.2 57.8 57.5 57.2 56.8 56.5 56.2 55.8 55.5

63.1 62.8 62.4 62.0 61.7 61.3 61.0 60.6 60.3 59.9 59.6 59.2 58.9 58.5 58.2 57.8 57.5 57.2 56.8 56.5 56.2 55.9

63.5 63.2 62.8 62.4 62.1 61.7 61.3 61.0 60.6 60.3 59.9 59.6 59.2 58.9 58.5 58.2 57.9 57.5 57.2 56.9 56.5 56.2

63.9 63.6 63.2 62.8 62.5 62.1 61.7 61.4 61.0 60.7 60.3 59.9 59.6 59.3 58.9 58.6 58.2 57.9 57.6 57.2 56.9 56.6

64.3 64.0 63.6 63.2 62.8 62.5 62.1 61.7 61.4 61.0 60.7 60.3 60.0 59.6 59.3 58.9 58.6 58.3 57.9 57.6 57.3 56.9

64.7 64.4 64.0 63.6 63.2 62.9 62.5 62.1 61.8 61.4 61.1 60.7 60.3 60.0 59.6 59.3 59.0 58.6 58.3 57.9 57.6 57.3

65.1 64.8 64.4 64.0 63.6 63.3 62.9 62.5 62.2 61.8 61.4 61.1 60.7 60.4 60.0 59.7 59.3 59.0 58.6 58.3 58.0 57.6

65.5 65.2 64.8 64.4 64.0 63.6 63.3 62.9 62.5 62.2 61.8 61.5 61.1 60.7 60.4 60.0 59.7 59.3 59.0 58.7 58.3 58.0

65.9 65.6 65.2 64.8 64.4 64.0 63.7 63.3 62.9 62.6 62.2 61.8 61.5 61.1 60.8 60.4 60.1 59.7 59.4 59.0 58.7 58.4

66.4 66.0 65.6 65.2 64.8 64.4 64.1 63.7 63.3 63.0 62.6 62.2 61.9 61.5 61.1 60.8 60.4 60.1 59.7 59.4 59.1 58.7

66.8 66.4 66.0 65.6 65.2 64.8 64.5 64.1 63.7 63.3 63.0 62.6 62.2 61.9 61.5 61.2 60.8 60.5 60.1 59.8 59.4 59.1

67.2 66.8 66.4 66.0 65.6 65.2 64.9 64.5 64.1 63.7 63.4 63.0 62.6 62.3 61.9 61.5 61.2 60.8 60.5 60.1 59.8 59.4

67.6 67.2 66.8 66.4 66.0 65.6 65.3 64.9 64.5 64.1 63.8 63.4 63.0 62.7 62.3 61.9 61.6 61.2 60.9 60.5 60.2 59.8

Conditioned Test Atmospheres 77

78

de Yong

to an error of 0.3±0.5% in relative humidity. This fact should be recognized when accuracy of dew point instruments is quoted. Use of the simpler four-term Eq. (29) for vapor pressure, together with rounding effects, results in about 7% of the calculated relative humidity values being 0.1% higher than the printed tables. However, the ef®ciency of memory usage and program operation was considered to justify using Eq. (29) with Eq. (16) in the dew point programs listed in Appendix 2. C.

Psychrometric Tables and Graphs

Tables 5±7 give values of relative humidity corresponding to dry bulb temperature and wet bulb depression over ranges corresponding to the standard atmospheres for paper testing. They were calculated using Eq. (20). Again a programmable calculator was used as outlined by Alvin [2], using the vapor pressure±temperature relationship of Keyes [22] to eliminate the need for manual entry of vapor pressure values. The accuracy of this method was compared as for the dew point table, and values checked within 0.02% for relative humidity. In using the psychrometric Eq. (20), ®xed values must be assigned to a and p, and for calculation of the tables these were taken to be a ˆ 6:565  10 4 = C, p ˆ 1013:25 mbar, whence ap ˆ 0:65652 mbar= C. As variations in a and/or p will be effective as variations in ap, the results of percentage changes in ap have been investigated. Considering an atmospheric pressure range of 980±1030 mbar, i.e., 3:3% to ‡1:6%, the corresponding effect on table values will be Table 5:

0:7 to ‡ 0:3

Table 6:

0:6 to ‡ 0:3

Table 7:

0:5 to ‡ 0:2

As discussed previously, changes in a may also occur between different instrument geometries, but these should be less than the listed variations in atmospheric pressure. Work by Wylie [35] leads to a revised value of 6:26  10 4 for a for the geometry described by de Yong [9]. It should be noted from the tables that a change of 0:1 C in dry bulb temperature corresponds to a change of 0.1% or less in relative humidity, whereas a change of 0:1 C in wet bulb depression corresponds to a change of 0.6±0.9% relative humidity. Where it is preferred to use wet and dry bulb temperatures directly, the graphs in Figs. 6±8 are the most convenient form of presentation. They represent largely the same data as the corresponding tables, and hence the same accuracy comments apply. Psychrometric table calculation based on the use of the simpler four-term Eq. (29) for vapor pressure, with Eq. (20), results in up to 50% of the calculated values being 0.1% lower than those in the printed tables. The uncertainty of relative humidity calculations using Eq. (20) is related to the accuracy of the value of ap, where a depends on the geometry of a particular instrument and p is the barometric pressure, which may vary with the weather and with the altitude of the measurement location. Weather changes can account for a variation of 2:5% in barometric pressure, and this results in a change of 0:4% in the calculated relative humidity. A change in elevation of 300 ft above sea level reduces barometric pressure by 11.3 mbar, or

Conditioned Test Atmospheres

79

1.1%, with a corresponding change of ‡0:2% in RH. The effect of a temperature measurement error of 0:1 C is at least 0.5% RH. In view of these ®gures, a difference of 0.1% is considered acceptable, and hence the psychrometric programs listed in Appendix 2 are appropriate.

D.

Saturated Salt Solutions

Hygrometer calibration and material conditioning in small cabinets require accurate humidity control of the working space. This can be achieved either by using a humidity generator, which can be complex and expensive, or by equilibration of the closed system with a chemical system that produces the required equilibrium vapor pressure. The latter method provides ®xed humidity levels in a relatively inexpensive and simple way. The chemical systems used for this purpose may be aqueous solutions of sulfuric acid, or glycerin, or saturated salt solutions. The aqueous solutions provide humidity adjustment by changing concentrations. Accordingly, these concentrations must be accurately measured and controlled, and this control requirement may present problems where there are sources or sinks of moisture within the closed space. However, because the concentration of saturated salt solution is known and ®xed as a function of temperature, it does not have to be measured, so saturated aqueous solutions of binary salts, where the solute is nonvolatile, are a very useful means of humidity control. Further, with an excess of solute, the solution will remain saturated with reasonable sinks of moisture within the controlled space. Because each saturated salt solution provides one level of relative humidity at a desired temperature, a number of salt solutions must be used. A summary of equilibrium relative humidities of saturated salt solutions was compiled by Greenspan [16], from which the data in Table 8, relevant to standard atmospheres for paper testing, have been extracted. These table values may be used for ®xed-point calibration of various direct-reading humidity sensors.

ACKNOWLEDGMENTS As much of the original First Edition text remains, the assistance of CSIRO colleagues, Dr. R. Wylie, for extensive and helpful discussions on Relative Humidity, and Dr. H. Higgins for constructive reading of the original manuscript, continues to be acknowledged.

6.1 51.8 51.9 52.0 52.2 52.3 52.4 52.5 52.6 52.7 52.9 53.0 53.1 53.2 53.3 53.4 53.5 53.6 53.7 53.9

Dry bulb ( C)

21.0 21.1 21.2 21.3 21.4 21.5 21.6 21.7 21.8 21.9 22.0 22.1 22.2 22.3 22.4 22.5 22.6 22.7 22.8

51.1 51.2 51.3 51.5 51.6 51.7 51.8 51.9 52.0 52.2 52.3 52.4 52.5 52.6 52.7 52.8 53.0 53.1 53.2

6.2 50.4 50.5 50.6 50.7 50.9 51.0 51.1 51.2 51.4 51.5 51.6 51.7 51.8 51.9 52.1 52.2 52.3 52.4 52.5

6.3 49.7 49.8 49.9 50.1 50.2 40.3 50.4 50.6 50.7 50.8 40.9 51.0 51.1 51.3 51.4 51.5 51.6 51.7 51.8

6.4

Table 5 Relative Humidity (%) Standard Atmosphere (a)

49.0 49.1 49.2 49.4 49.5 49.6 49.7 49.9 50.0 50.1 50.2 50.4 50.5 50.6 50.7 50.8 50.9 51.1 51.2

6.5 48.3 48.4 48.5 48.7 48.8 48.9 49.1 49.2 49.3 49.4 49.6 49.7 49.8 49.9 50.0 50.2 50.3 50.4 50.5

6.6 47.6 47.7 47.9 48.0 48.1 48.2 48.4 48.5 48.6 48.8 48.9 49.0 49.1 49.2 49.3 49.5 49.6 49.7 49.8

6.7 46.9 47.0 47.2 47.3 47.4 47.6 47.7 47.8 48.0 48.1 48.2 48.3 48.5 48.6 48.7 48.8 48.9 49.1 49.2

6.8

Wet bulb depression ( C)

46.2 46.4 46.5 46.6 46.8 46.9 47.0 47.1 47.3 47.4 47.5 47.7 47.8 47.9 48.0 48.2 48.3 48.4 48.5

6.9

45.5 45.7 45.8 45.9 46.1 46.2 46.3 46.5 46.6 46.7 46.9 47.0 47.1 47.3 47.4 47.5 47.6 47.8 47.9

7.0

44.8 45.0 45.1 45.3 45.4 45.5 45.7 45.8 45.9 46.1 46.2 46.3 46.5 46.6 46.7 46.8 47.0 47.1 47.2

7.1

44.2 44.3 44.4 44.6 44.7 44.9 45.0 45.1 45.3 45.4 45.5 45.7 45.8 45.9 46.1 46.2 46.3 46.4 46.6

7.2

43.5 43.6 43.8 43.9 44.1 44.2 44.3 44.5 44.6 44.7 44.9 45.0 45.1 45.3 45.4 45.5 45.7 45.8 45.9

7.3

80 de Yong

22.9 23.0 23.1 23.2 23.3 23.4 23.5 23.6 23.7 23.8 23.9 24.0 24.1 24.2 24.3 24.4 24.5 24.6 24.7 24.8 24.9 25.0

54.0 54.1 54.2 54.3 54.4 54.5 54.6 54.7 54.8 54.9 55.0 55.1 55.2 55.3 55.4 55.5 55.6 55.7 55.8 55.9 56.0 56.1

53.3 53.4 53.5 53.6 53.7 53.8 53.9 54.0 54.2 54.3 54.4 54.5 54.6 54.7 54.8 54.9 55.0 55.1 55.2 55.3 55.4 55.5

52.6 52.7 52.8 53.0 53.1 53.2 53.3 53.4 53.5 53.6 53.7 53.8 53.9 54.0 54.1 54.2 54.3 54.4 54.5 54.6 54.7 54.8

52.0 52.1 52.2 52.3 52.4 52.5 52.6 52.7 52.8 52.9 53.1 53.2 53.3 53.4 53.5 53.6 53.7 53.8 53.9 54.0 54.1 54.2

51.3 51.4 51.5 51.6 51.7 51.9 52.0 52.1 52.2 52.3 52.4 52.5 52.6 52.7 52.8 52.9 53.0 53.1 53.3 53.4 53.5 53.6

50.6 50.7 50.9 51.0 51.1 51.2 51.3 51.4 51.5 51.6 51.8 51.9 52.0 52.1 52.2 52.3 52.4 52.5 52.6 52.7 52.8 52.9

50.0 50.1 50.2 50.3 50.4 50.5 50.7 50.8 50.9 51.0 51.1 51.2 51.3 51.4 51.6 51.7 51.8 51.9 52.0 52.1 52.2 52.3

49.3 49.4 49.5 49.7 49.8 49.9 50.0 50.1 50.2 50.4 50.5 50.6 50.7 50.8 50.9 51.0 51.1 51.2 51.3 51.5 51.6 51.7

48.6 48.8 48.9 49.0 49.1 49.3 49.4 49.5 49.6 49.7 49.8 49.9 50.0 50.2 50.3 50.4 50.5 50.6 50.7 50.8 50.9 51.1

48.0 48.1 48.2 48.4 48.5 48.6 48.7 48.8 49.0 49.1 49.2 49.3 49.4 49.5 49.7 49.8 49.9 50.0 50.1 50.2 50.3 50.4

47.4 47.5 47.6 47.7 47.8 48.0 48.1 48.2 48.3 48.4 48.6 48.7 48.8 48.9 49.0 49.1 49.3 49.4 49.5 49.6 49.7 49.8

46.7 46.8 47.0 47.1 47.2 47.3 47.4 47.6 47.7 47.8 47.9 48.0 48.2 48.3 48.4 48.5 48.6 48.7 48.9 49.0 49.1 49.2

46.1 46.2 46.3 46.4 46.5 46.7 46.8 46.9 47.1 47.2 47.3 47.4 47.5 47.6 47.8 47.9 48.0 48.1 48.2 48.4 48.5 48.6

Conditioned Test Atmospheres 81

3.5 68.7 68.8 68.9 69.0 69.0 69.1 69.2 69.3 69.4 69.5 69.5 69.6 69.7 69.8 69.9 69.9 70.0 70.1 70.2

Dry bulb ( C)

18.0 18.1 18.2 18.3 18.4 18.5 18.6 18.7 18.8 18.9 19.0 19.1 19.2 19.3 19.4 19.5 19.6 19.7 19.8

67.9 68.0 68.0 68.1 68.2 68.3 68.4 68.5 68.6 68.6 68.7 68.8 68.9 69.0 69.0 69.1 69.2 69.3 69.4

3.6 67.0 67.1 67.2 67.3 67.4 67.5 67.6 67.6 67.7 67.8 67.9 68.0 68.1 68.2 68.2 68.3 68.4 68.5 68.6

3.7 66.2 66.3 66.4 66.5 66.6 66.7 66.7 66.8 66.9 67.0 67.1 67.2 67.3 67.4 67.4 67.5 67.6 67.7 67.8

3.8

Table 6 Relative Humidity (%) Standard Atmosphere (b)

65.4 65.5 65.6 65.7 65.7 65.8 65.9 66.0 66.1 66.2 66.3 66.4 66.5 66.6 66.6 66.7 66.8 66.9 67.0

3.9 64.5 64.6 64.7 64.8 64.9 65.0 65.1 65.2 65.3 65.4 65.5 65.6 65.7 65.8 65.8 65.9 66.0 66.1 66.2

4.0 63.7 63.8 63.9 64.0 64.1 64.2 64.3 64.4 64.5 64.6 64.7 64.8 64.9 65.0 65.1 65.1 65.2 65.3 65.4

4.1 62.9 63.0 63.1 63.2 63.3 63.4 63.5 63.6 63.7 63.8 63.9 64.0 64.1 64.2 64.3 64.4 64.4 64.5 64.6

4.2

Wet bulb depression ( C)

62.1 62.2 62.3 62.4 62.5 62.6 62.7 62.8 62.9 63.0 63.1 63.2 63.3 63.4 63.5 63.6 63.7 63.8 63.9

4.3

61.3 61.4 61.5 61.6 61.7 61.8 61.9 62.0 62.1 62.2 62.3 62.4 62.5 62.6 62.7 62.8 62.9 63.0 63.1

4.4

60.5 60.6 60.7 60.8 60.9 61.0 61.1 61.2 61.3 61.4 61.5 61.6 61.7 61.8 61.9 62.0 62.1 62.2 62.3

4.5

59.6 59.8 59.9 60.0 60.1 60.2 60.3 60.4 60.5 60.6 60.7 60.8 60.9 61.0 61.1 61.2 61.3 61.4 61.5

4.6

58.8 59.0 59.1 59.2 59.3 59.4 59.5 59.6 59.7 59.8 59.9 60.0 60.1 60.2 60.4 60.5 60.6 60.7 60.8

4.7

82 de Yong

19.9 20.0 20.1 20.2 20.3 20.4 20.5 20.6 20.7 20.8 20.9 21.0 21.1 21.2 21.3 21.4 21.5 21.6 21.7 21.8 21.9 22.0

70.2 70.3 70.4 70.5 70.5 70.6 70.7 70.8 70.8 70.9 71.0 71.0 71.1 71.2 71.3 71.3 71.4 71.5 71.5 61.6 71.7 71.7

69.4 69.5 69.6 69.7 69.7 69.8 69.9 70.0 70.0 70.1 70.2 70.3 70.3 70.4 70.5 70.6 70.6 70.7 70.8 70.9 70.9 71.0

68.6 68.7 68.8 68.9 69.0 69.0 69.1 69.2 69.3 69.3 69.4 69.5 69.6 69.6 69.7 69.8 69.9 69.9 70.0 70.1 70.2 70.2

67.9 67.9 68.0 68.1 68.2 68.3 68.3 68.4 68.5 68.6 68.7 68.7 68.8 68.9 69.0 69.0 69.1 69.2 69.3 69.3 69.4 69.5

67.1 67.2 67.2 67.3 67.4 67.5 67.6 67.6 67.7 67.8 67.9 68.0 68.0 68.1 68.2 68.3 68.4 68.4 68.5 68.6 68.7 68.7

66.3 66.4 66.5 66.5 66.6 66.7 66.8 66.9 67.0 67.0 67.1 67.2 67.3 67.4 67.4 67.5 67.6 67.7 67.8 67.8 67.9 68.0

65.5 65.6 65.7 65.8 65.8 65.9 66.0 66.1 66.2 66.3 66.4 66.4 66.5 66.6 66.7 66.8 66.8 66.9 67.0 67.1 67.2 67.2

64.7 64.8 64.9 65.0 65.1 65.2 65.3 65.3 65.4 65.5 65.6 65.7 65.8 65.9 65.9 66.0 66.1 66.2 66.3 66.3 66.4 66.5

63.9 64.0 64.1 64.2 64.3 64.4 64.5 64.6 64.7 64.8 64.8 64.9 65.0 65.1 65.2 65.3 65.4 65.4 65.5 65.6 65.7 65.8

63.2 63.3 63.4 63.5 63.5 63.6 63.7 63.8 63.9 64.0 64.1 64.2 64.3 64.4 64.4 64.5 64.6 64.7 64.8 64.9 65.0 65.0

62.4 62.5 62.6 62.7 62.8 62.9 63.0 63.1 63.2 63.2 63.3 63.4 63.5 63.6 63.7 63.8 63.9 64.0 64.0 64.1 64.2 64.3

61.6 61.7 61.8 61.9 62.0 62.1 62.2 62.3 62.4 62.5 62.6 62.7 62.8 62.9 63.0 63.0 63.1 63.2 63.3 63.4 63.5 63.6

60.9 61.0 61.1 61.2 61.3 61.4 61.5 61.6 61.7 61.7 61.8 61.9 62.0 62.1 62.2 62.3 62.4 63.5 62.6 62.7 62.8 62.9

Conditioned Test Atmospheres 83

4.4 67.3 67.4 67.5 67.6 67.6 67.7 67.8 67.8 67.9 68.0 68.0 68.1 68.2 68.2 68.3 68.4 68.4 68.5 68.6

Dry bulb ( C)

25.0 25.1 25.2 25.3 25.4 25.5 25.6 25.7 25.8 25.9 26.0 26.1 26.2 26.3 26.4 26.5 26.6 26.7 26.8

66.7 66.7 66.8 66.9 67.0 67.0 67.1 67.2 67.2 67.3 67.4 67.4 67.5 67.6 67.6 67.7 67.8 67.8 67.9

4.5 66.0 66.1 66.1 66.2 66.3 66.4 66.4 66.5 66.6 66.6 66.7 66.8 66.8 66.9 67.0 67.1 67.1 67.2 67.3

4.6 65.3 65.4 65.5 65.5 65.6 65.7 65.8 65.8 65.9 66.0 66.0 66.1 66.2 66.3 66.3 66.4 66.5 66.5 66.6

4.7

Table 7 Relative Humidity (%) Standard Atmosphere (c)

64.6 64.7 64.8 64.9 64.9 65.0 65.1 65.2 65.2 65.3 65.4 65.5 65.5 65.6 65.7 65.7 65.8 65.9 66.0

4.8 64.0 64.0 64.1 64.2 64.3 64.4 64.4 64.5 64.6 64.7 64.7 65.8 64.9 64.9 65.0 65.1 65.2 65.2 65.3

4.9 63.3 63.4 63.5 63.5 63.6 63.7 63.8 63.8 63.9 64.0 64.1 64.2 64.2 64.3 64.4 64.4 64.5 64.6 64.7

5.0 62.6 62.7 62.8 62.9 63.0 63.0 63.1 63.2 63.3 63.3 63.4 63.5 63.6 63.7 63.7 63.8 63.9 64.0 64.0

5.1

Wet bulb depression ( C)

62.0 62.1 62.1 62.2 62.3 62.4 62.5 62.5 62.6 62.7 62.8 62.9 62.9 63.0 63.1 63.2 63.2 63.3 63.4

5.2

61.3 61.4 61.5 61.6 61.6 61.7 61.8 61.9 62.0 62.0 62.1 62.2 62.3 62.4 62.4 62.5 62.6 62.7 62.8

5.3

60.6 60.7 60.8 60.9 61.0 61.1 61.2 61.2 61.3 61.4 61.5 61.6 61.6 61.7 61.8 61.9 62.0 62.0 62.1

5.4

60.0 60.1 60.2 60.3 60.3 60.4 60.5 60.6 60.7 60.8 60.8 60.9 61.0 61.1 61.2 61.3 61.3 61.4 61.5

5.5

59.3 59.4 59.5 59.6 59.7 59.8 59.9 59.9 60.0 60.1 60.2 60.3 60.4 60.5 60.5 60.6 60.7 60.8 60.9

5.6

84 de Yong

26.9 27.0 27.1 27.2 27.3 27.4 27.5 27.6 27.7 27.8 27.9 28.0 28.1 28.2 28.3 28.4 28.5 28.6 28.7 28.8 28.9 29.0

68.6 68.7 68.8 68.8 68.9 68.9 69.0 69.1 69.1 69.2 69.2 69.3 69.4 69.4 69.5 69.5 69.6 69.7 69.7 69.8 69.8 69.9

68.0 68.0 68.1 68.2 68.2 68.3 68.4 68.4 68.5 68.5 68.6 68.7 68.7 68.8 68.9 68.9 69.0 69.0 69.1 69.2 69.2 69.3

67.3 67.4 67.5 67.5 67.6 67.6 67.7 67.8 67.8 67.9 68.0 68.0 68.1 68.2 68.2 68.3 68.3 68.4 68.5 68.5 68.6 68.6

66.7 66.7 66.8 66.9 66.9 67.0 67.1 67.1 67.2 67.3 67.3 67.4 67.5 67.5 67.6 67.7 67.7 67.8 67.8 67.9 68.0 68.0

66.0 66.1 66.2 66.2 66.3 66.4 66.4 66.5 66.6 66.6 66.7 66.8 66.8 66.9 67.0 67.0 67.1 67.2 67.2 67.3 67.3 67.4

65.4 65.5 65.5 65.6 65.7 65.7 65.8 65.9 65.9 66.0 66.1 66.1 66.2 66.3 66.3 66.4 66.5 66.5 66.6 66.7 66.7 66.8

64.7 64.8 64.9 65.0 65.0 65.1 65.2 65.2 65.3 65.4 65.4 65.5 65.6 65.6 65.7 65.8 65.8 65.9 66.0 66.0 66.1 66.2

64.1 64.2 64.2 64.3 64.4 64.5 64.5 64.6 64.7 64.7 64.8 64.9 65.0 65.0 65.1 65.2 65.2 65.3 65.4 65.4 65.5 65.6

63.5 63.5 63.6 63.7 63.8 63.8 63.9 64.0 64.0 64.1 64.2 64.3 64.3 64.4 64.5 64.5 64.6 64.7 64.7 64.8 64.9 65.0

62.8 62.9 63.0 63.1 63.1 63.2 63.3 63.4 63.4 63.5 63.6 63.6 63.7 63.8 63.9 63.9 64.0 64.1 64.1 64.2 64.3 64.3

62.2 62.3 62.4 62.4 62.5 62.6 62.7 62.7 62.8 62.9 63.0 63.0 63.1 63.2 63.2 63.3 63.4 63.5 63.5 63.6 63.7 63.7

61.6 61.6 61.7 61.8 61.9 62.0 62.0 62.1 62.2 62.3 62.3 62.4 62.5 62.6 62.6 62.7 62.8 62.9 62.9 63.0 63.1 63.1

60.9 61.0 61.1 61.2 61.3 61.3 61.4 61.5 61.6 61.6 61.7 61.8 61.9 61.9 62.0 62.1 62.2 62.2 62.3 62.4 62.5 62.5

Conditioned Test Atmospheres 85

Table 8

Equilibrium Relative Humidity of Saturated Salt Solutions Relative humidity (%)

Salt Potassium carbonate Magnesium nitrate Sodium bromide Potassium iodide

Fig. 6



20 C

23 C

27 C

43:2  0:3 54:4  0:2 59:1  0:4 69:9  0:3

43:2  0:4 53:5  0:2 58:2  0:4 69:3  0:2

43:2  0:4 52:3  0:2 57:0  0:4 68:5  0:2

Psychrometric chart for standard atmosphere (a). (See section II.B.)

Fig. 7 Psychrometric chart for standard atmosphere (b). (See Section II.B.) 86

Conditioned Test Atmospheres

Fig. 8

87

Psychrometric chart for standard atmosphere (c). (see Section II.B.)

APPENDIX 1: a ac b Cpa dv Dw e ew ew0 ewd Fr ; F g1 ; g2 ; g3 ; g4 hc hf ;0 hf ;t 0 kc mv ma mma Mv Ma p q r rw

SYMBOLS

psychrometric coef®cient convective psychrometric coef®cient coef®cient in Eq. (19) speci®c heat of dry air at constant pressure absolute humidity degree of saturation effective vapor pressure of unsaturated moist air saturation vapor pressure of moist air saturation vapor pressure of moist air at the wet bulb temperature saturation vapor pressure of moist air at the dew point temperature factors in Eq. (27) constants in Eq. (29) convective surface heat transfer coef®cient latent heat of vaporization of water at 0 C latent heat of vaporization of water at the wet bulb temperature convective mass transfer coef®cient mass of water vapor mass of dry air mass of moist air molecular weight of water vapor apparent molecular weight of clean dry air atmospheric pressure speci®c humidity mixing ratio saturation mixing ratio

88

rw0 rwd t t0 td T Uw V xv xvw xvwd ;  0 ; 0



de Yong

saturation mixing ratio at the wet bulb temperature saturation mixing ratio at the dew point temperature dry bulb temperature wet bulb temperature dew point temperature absolute temperature relative humidity of moist air volume of dry air mole fraction of water vapor in moist air mole fraction of water vapor in air that is saturated with respect to water mole fraction of water vapor in air that is saturated with respect to water, at the dew point temperature constants in Eq. (17) constants in Eq. (17) constant in Eq. (17) partial density of dry air

Conditioned Test Atmospheres

APPENDIX A.

2:

89

PROGRAMS

HP-15C Program for Relative Humidity from Dry Bulb and Dew Point Temperatures Key in : Dry Bulb t, ENTER, Dew Point td f D

Keystrokes

Display

f LBL D STO 0 x„y GSB 1 STO 1 RCL 0 GSB 1 RCL 1  EEX 2 x R/S f LBL 1 2 7 3
1 6 ‡ RCL 8 x„y  g LST x RCL 10 x„y x LST x g x2 RCL 11 x RCL 9 ‡ ‡ ‡ ex g RTN

001-42, 002003004005006007008009010011012013014-42, 015016017018019020021022023024025026027028029030031032033034035036037038-

Operation 21, 44 32 44 45 32 45

21,

45 43 45 43 43 45 45

43

14 0 34 1 1 0 1 1 10 26 2 20 31 1 2 7 3 48 1 6 40 8 34 10 36 .0 34 20 36 11 .1 20 9 40 40 40 12 32

td into R0 Reg.Exch. Calc.ew ew into R1 td from R0 Calc.ewd ew from R1 ewd =ew 102 Uw % End t

Memory Usage R 0 ˆ td R 1 ˆ ew

Lines 014 to 038 Vapor Pressure Subroutine

Memory Storage Calc.T R8 ˆ 6:3536311  103 A from R8 Reg.Exch. R9 ˆ 3:40492603  10 A/T R10 ˆ 1:9509874  10 2 Recover T R11 ˆ 1:2811805  10 5 C from R10 R13 ˆ 6:68745  10 Reg.Exch. CT Recover T Note: R2 to R7 are used T2 for Statistic Functions. D from R11 D  T2 B from R9 B ‡ D  T2 B ‡ C  T ‡ D  T2 A=T ‡ B ‡ C  T ‡ D  T2 Calc.ew Return

90

de Yong

B.

HP-15C Program for Relative Humidity from Wet and Dry Bulb Temperatures Key In : Dry Bulb t, ENTER, Wet Bulb t 0 , f E

Keystrokes

Display

f LBL E STO 1 x„y STO 0 GSB 1 STO 12 RCL 1 GSB 1 RCL 0 RCL 1

039-42, 040041042043044045046047048049050051052053054055056057058-

RCL 13 x RCL 12  EEX 2 x R/S

C. 10 20 30 40 50 60 70 80 90 100 110 120 130

Operation 21, 44 44 32 42 45 32 45 45 45 45

15 1 34 0 1 .2 1 1 0 1 30 .3 20 30 .2 10 26 2 20 31

t 0 into R1 Reg.Exch. t into R0 Calc.ew ew into R12 t 0 from R1 Calc.ew0 t from R0 t 0 from R1 t t0 ap from R13 ap(t t 0 ) ew0 -ap(t t 0 † ew from R12 Uw

Memory Usage R0 ˆ t R1 ˆ t 0 R12 ˆ ew

Vapor Pressure Subroutine Lines 014 to 038

102 Uw % End

GW-Basic Program for Relative Humidity from Dry Bulb and Dew Point Temperatures INPUT "Dry Bulb ",X:INPUT "Dew Point ",Y T = Y + 273.16 A = 6.3536311#*10^3 : B = 3.404926034#*10 C = 1.9509874#*10^ 2} : D = 1.2811805#*10^ 5 GOSUB 110 L = J : T = X + 273.16 GOSUB 110 M = (CINT ( (L/J)*1000) )/10 PRINT "Relative Humidity "; M END I = A/T + B + C*T + D*(T^2) J = (EXP (I) )/100 RETURN

Conditioned Test Atmospheres

D. 10 20 30 40 50 60 70 80 90 100 110 120 130

91

GW-Basic Program for Relative Humidity from Dry Bulb Temperature and Wet Bulb Depression INPUT "Dry Bulb ",.X:INPUT "Wet Bulb Dep. ",Y T = (X - Y) + 273.16 : K = .668745 A = 6.3535311*10^3 : B = 3.404926034*10 C = 1.9509874*10^ 2 : D = 1.2811805*10^ 5 GOSUB 110 L = J - K*Y : T = X + 273.16 GOSUB 110 M = (CINT ( (L/J)*1000) )/10 PRINT "Relative Humidity ";M END I = A/T + B + C*T + D*(T^2) J = (EXP (I) )/100 RETURN

REFERENCES 1.

Abdul Wahab, M. D., Badros, M. Z., and Ramli, A. R. (1995). Simple technique for measuring air humidity using an IBM microcomputer. Proc. Inst. Mech. Eng. Part E 209(E2):145±150. 2. Alvin, K. E. (1976). Relative humidity from psychrometric data. Svensk Papperstidn. 79(1):18±19. 3. Bearzotti, A., Bianco, A., Montesperelli, G., and Traversa, E. (1994). Humidity sensitivity of sputtered TiO2 thin ®lms. Sensors Actuators B (Chem) B19(1-3):525±528. 4. Bhuyan, M., and Bhuyan, R. (1995). An on-line method for monitoring relative humidity using thermal sensors. Proc. IEEE/IAS Int. Conf. on Industrial Automation and Control, p. 711. 5. Carullo, A., Ferrero, A., and Parvis, M. (1999). A Microwave System for Relative Humidity Measurement, IEEE Instrumentation and Measurement Technology Conference Proceedings, Venice, Italy, May 1999, 1:124±129. 6. Crosby, A. M. (1978). PsychrometricsÐor the measurement of humidity. Heating Vent. Eng. 52(605):24±27. 7. Denton, D. D., Ho, C. M., and He, S. G. (1990). A solid-state relative humidity measurement system. IEEE Trans. Instrum. Measure. 39(3):508±511. 8. Denton, D. D., Jaafar, M. A. S., and Ralston, A. R. K. (1992). The long term reliability of a switched-capacitor relative humidity sensor system. 1992 IEEE Int. Symp. on Circuits and Systems 4:1840±1843. 9. de Yong, J. L. (1975). A precision, narrow range electronic psychrometer. IICA '75 Conf. Proc., Institute of Instrumentation and Control, Australia, pp. 3-1 to 3-5. 10. de Yong, J. L. (1982). Relative humidity and its measurement. Appita 35(6):483±490. 11. Eisner, A. D., and Martonen, T. B. (1989). Design and development of a micro-thermocouple sensor for determining temperature and relative humidity patterns within an airstream. Trans. ASME J. Biomech. Eng. 111(4):283±297. 12. Ferrel, W. (1886). Report on psychrometric tables for use in the signal service. Ann. Rep., Chief Signal Of®cer (of the Army to the Secretary of War for the year 1886), Washington, D.C., Appendix 24, pp 233±259.

92

de Yong

13. Folland, C. K., and Sparks, W. R. (1975). A two-pressure humidity generator for calibrating electrical hygrometers used in meteorology. J. Phys. E. Sci. Instrum. 9:112±116. 14. Glenn, M. C., and Schuetz, J. A. (1985). An IC compatible polymer humidity sensor. Digest of Tech. Papers, TRANSDUCERS '85. 1985 International Conference on SolidState Sensors and Actuators, pp. 217±220. 15. Goff, J. A., and Gratch, S. (1945). Thermodynamic properties of moist air. Trans. Am. Soc. Heating Vent. Eng. 51:125±164. 16. Greenspan, L. (1977). Humidity ®xed points of binary saturated aqueous solutions, J. Res. Natl. Bur. Stand. 81A(1):89±96. 17. Hardy, B. (1988). Precise humidity generation: An automated approach. Natl. Conf. of Standards Labs. 1988 Workshop and Symposium Technical Presentations: Competitiveness in a World Market, pp. 6/1±22. 18. Harrison, L. P. (1965). Fundamental concepts and de®nitions relating to humidity. In: Fundamentals and Standards in Humidity and Moisture, Vol. 1. A. Wexler, ed. Reinhold, New York, pp. 3±69. 19. Harrison, L. P. (1965). Some fundamental considerations regarding psychrometry. In: Fundamentals and Standards in Humidity and Moisture, Vol. 1. A. Wexler, ed. Reinhold, New York, pp. 71±103. 20. Hasegawa, S., and Little J. W. (1977). The NBS two-pressure humidity generator, Mark 2. J. Res. Natl. Bur. Stand. 81A(1): 81±88. 21. Kalogiros, J. A., and Helmis, C. G. (1993). Fast-response humidity measurements with the psychrometric method. J. Appl. Meteorol. 32(9):1499±1507. 22. Keyes, F. G. (1947). The thermodynamic properties of water substance, 0 C to 150 C, Part VI. J. Chem. Phys. 15:602±612. 23. Leivers, M. F., and Letherman, K. M. (1978). The extension characteristics of some materials used for humidity sensors. Build. Ser. Eng. 45:205±210. 24. Ministry of Technology, U.K. (1970). Measurement of humidity. In: National Physical Laboratory Notes on Applied Science, No. 4. 25. Morten, B., De Cicco, G., and Prudenziati, M. (1993). A thick-®lm resonant sensor for humidity measurements. Sensors Actuators A (Phys.) A37±A38:337±342. 26. Parkes, K. G. (1979). Humidity and dew point. Meas. Control 12(2):73±77. 27. Pelino, M., Cantalini, C., and Faccio, M. (1994). Principles and applications of ceramic humidity sensors. Active Passive Electron. Compon. 16(2):69±87. 28. Rouleau, J. F., Goyette, J., Bose, T. K., Frechette, M. F. (1999). Investigation of a microwave differential cavity resonator device for the measurement of humidity in gases. Rev. Sci. Instrum. 70(9):3590±3594. 29. Shibata, H., Ito, M., Asakura, M., and Watanabe, K. (1995). A digital hygrometer using a capacitance-to-frequency converter. Proc. 1995 IEEE Instrumentation and Measurement Technology ConferenceÐIMTC'95, pp. 100±106. 30. Strembicke, D., Robinson, A. M., Vermeulen, F. E., Seto, M., and Brown, K. B. (1999). Humidity measurement using resonating CMOS microcantileverstructures, IEEE Canadian Conference on Electrical and Computer Engineering, Edmonton, Alberta, May 1999 3:1658±1661. 31. Wexler, A. (1970). Measurement of humidity in the free atmosphere near the surface of the earth. Metereol. Monograph 11(33). Am. Meteorological Soc., pp. 262±282. 32. Wexler, A. (1976). Vapor pressure formulation for water in range 0 to 100 C. A revision. J. Res. Natl. Bur. Stand. 80A(5±6):775±785. 33. Wiederhold, P. R. (1975). Humidity measurements, Part 1. Psychrometers and percent R.H. sensors. Instrum. Technol. 22(6):31±37. 34. Wiederhold, P. R. (1975). Humidity measurements, Part 2. Hygrometery. Instrum. Technol. 22(8):45±50.

Conditioned Test Atmospheres

93

35. Wylie, R. G. (1979). Psychrometric wet elements as a basis for precise physico-chemical measurements. J. Res. Natl. Bur. Stand. 84(2):161±177. 36. Wylie, R. G., and Lalas, T. (1981). Detailed determination of the psychrometer coef®cient for the wet cylinder in a transverse air stream and an analysis of accuracy. CSIRO Australia, Div. Appl. Phys. Tech. Paper No. 7. 37. Zimmerman, O. T., and Lavine, I. (1960). Psychrometric Tables and Charts. 2nd ed. Industrial Research Service, Inc., Dover, NH.

4 OPTICAL AND APPEARANCE PROPERTIES JENS BORCH IBM Corporation Boulder, Colorado

I. Introduction II. The Evolution of Paper Optics

96 96

III. Principles of Measurement A. Light±Sheet Interactions B. Routine Re¯ectance Methods C. Colorimetry D. Whiteness, Yellowness, and Fluorescence E. Gloss and Surface Texture Perception F. Light Transmittance Methods

97 97 99 103 110 112 115

IV. Instrumentation Characteristics A. Instrument Types B. Design Principles C. Calibration

118 118 123 125

V. Interpretation of Optical Measurements A. Kubelka±Munk Theory B. Particulate Scattering Approaches C. Interpretation of Gloss D. Formation Theory References



127 127 133 137 138 140

Retired. 95

96

I.

Borch

INTRODUCTION

In all applications for which paper is used to print on or write on, the appearance of the paper contributes to the image quality that is obtainable. Thus, whereas the paper base must be suitable for the reproduction process that is used, the appearance characteristics will ultimately determine whether the document is acceptable to the customer. Color is created in paper and paperboard for decorative or functional reasons. Gloss or shininess is often produced to make the marketing of a product more effective. Various testing methods have been developed to measure the appearance characteristics of paper and paper board. In order to comprehend the principles behind optical testing and to apply the measurement techniques effectively it is necessary to understand the manner in which paper interacts with light. This chapter examines the principles behind classical optical and appearance measurements of paper and paperboard. It also describes the proper choice of instrumentation and the optical theories that are available for the interpretation of optical measurements. II.

THE EVOLUTION OF PAPER OPTICS

The measurement and description of paper optical properties date back to the start of the 1900s. At that time agreed-on de®nitions of paper optical properties and procedures for describing paper optics were not available, and methods were strongly dependent on instrument design. Harrison [58] described paper opacity measurements that led to the acceptance of standard procedures and better de®ned instrumentation in the United states and Great Britain in the early 1930s. Due to this and other work, re¯ectance methods (de®nition of optical properties based on re¯ected light) gained more use than transmittance methods (de®nition of optical properties based on transmitted light) for paper optics measurements. The early re¯ectometers with their then state-of-the-art illumination systems, light detection units, etc., required considerable attention to calibration, measurement procedure, and maintenance. Van den Akker [187] reviewed aspects of instrumentation design including studies carried out by the Institute of Paper Chemistry. Instrumentation superior with respect to light geometry and sensitivity (e.g., the Elrepho colorimeter) [38] stimulated further development of national standards and the ®rst proposals for international standardization for both opacity and brightness. The international methods are now issued by the International Organization for Standardization (ISO) (see Section III.B). With some exceptions, development of gloss- and color-measuring instrumentation has been less speci®c for the paper industry because similar measurements were done in related areas (e.g., paint and plastics) [74]. Goniophotometric measurements of gloss [188,193] demonstrated the complexity of paper and board gloss depending on surface structure and added coatings. In colorimetry, the design of both colorimeters and spectrophotometers has progressed enormously since they were ®rst applied in paper optics. Well-de®ned spectral properties, silicon diode detection, and computer interfaces provide both standard paper re¯ectance characteristics and more recently introduced additional optical properties (e.g., whiteness and ¯uorescence) (Section III.D). Similarly, modern instrumentation technology has

Optical and Appearance Properties

97

made possible the automated collection and sampling of re¯ected or transmitted radiation from small sheet areas in image analysis or formation testing. In addition to the development of instrumentation, early work in paper optics focused on the interrelationship between re¯ectance and transmittance measurements and the calculation of optical properties for similar sheet structures without additional measurements (for example, the effect of sheet thickness or basis weight on re¯ectance values). This need was alleviated when Steele [166] demonstrated the use of the Kubelka±Munk theory (see Section V.A) in paper optics and when basis weight rather than sheet thickness was introduced into the computation procedures [187]. Kubelka±Munk calculations are now routinely carried out for conversion of optical paper measurements and prediction of sheet characteristics based on physical changes in sheet weight, scattering ability, or absorption power (Section V.A). More recent work has focused on the desire to characterize the effect of sheet morphology (for example, in¯uence of ®ber nature and additives such as ®llers and pigments) on optical properties using particulate scattering approaches. To this end, layer calculations, Monte Carlo simulations, or both, have been applied to re¯ectance measurements. Although these approaches are more complex than the Kubelka±Munk approach and do not exactly describe the light interactions in a paper sheet, they deal with structural sheet elements and should better predict optical properties on the basis of morphology changes [36] (Section V.B).

III.

PRINCIPLES OF MEASUREMENT

A.

Light±Sheet Interactions

The Physics of Light Visible light is radiation in the wavelength range 400± 700 nm, which is perceived by the human eye at the wavelength-dependent luminosity (lightness) response shown in Fig. 1. Light interactions with matter follow the physical laws valid for the electromagnetic spectrum extending from short gamma rays and X-rays to long radio waves. In paper and paperboard, matter consists mainly of cellulosic ®bers, inorganic ®ller and pigment particles, and air voids of

Fig. 1

Luminosity response of the human eye to the visible spectrum.

98

Borch

certain sizes and shapes. Fibers and ®ber fragments are generally larger than the wavelengths of visible light. Inorganic ®ller and pigment particles and air voids often will be smaller than this. From a fundamental viewpoint, the ratio of particle size to wavelength is functional in determining how the individual particles interact with light. The larger particles that usually predominate in any sheet structure will surface refract and re¯ect light according to the laws of Fresnel; the smaller particles interact with light in the more complex manner described by Rayleigh and Mie. Light absorptance, which creates sheet color, is mainly due to the larger particles. The refractive index and the absorption coef®cient determine the energy scattered and absorbed by the particle. Re¯ectance, Transmittance, and Absorptance Due to the multitude, complexity, and close proximity of light-interacting particles in a sheet of paper, it is customary to consider the effect rather than the exact nature of light±sheet interactions when carrying out routine paper optics measurements. Light interactions with sheet constituents have only recently been evaluated in some detail (Section V.B). The radiation is either re¯ected from, transmitted through, or absorbed in the sheet structure, as shown in Fig. 2. Re¯ectance is the ratio of re¯ected light intensity to incident light intensity. Transmittance and absorptance are the ratios of transmitted and absorbed light intensities, respectively, to incident light intensity. Generally, re¯ectance is the light ¯ux that is measured in paper optics. Transmittance measurements are carried out for rather speci®c purposes or where the accuracy of transmittance measurements exceeds that using re¯ectance measurements (Section III.F). Absorptance is not susceptible to direct photometric measurements but may be calculated as an absorption coef®cient or power using re¯ectance measurements interpreted through, for example, the Kubelka±Munk theory (Section V.A) or a particulate scattering approach (Section V.B). Routine re¯ectance measurements such as those described in the following pages will not show the exact manner in which light interacts with paper constituents. For example, a sheet made of many layers of bleached pulp ®bers diffuses the incident light to such a degree that details of particle scattering are lost. The re¯ected light may approach the light distribution from re¯ectance standards such as magne-

Fig. 2

Light±sheet interactions.

Optical and Appearance Properties

99

sium oxide, barium sulfate, and the ``perfect'' diffuser that re¯ects light of equal intensity in any direction. Such paper is opaque, in contrast to transparent, which describes, for example, cellulosic sheets such as parchment paper and cellophane sheets in which the gross ®ber structure is not present. Similarly, paper sheets made of fully bleached pulp ®bers are white and do not show the light absorption characteristics of transparent cellulose, which make cellulosic ®lms appear yellowish. B.

Routine Re¯ectance Methods

Opacity and Brightness Instrumental assessment of visual paper sheet appearance has been performed for nearly a century. Routine measurements are usually those of opacity and brightness, for which standard methods have been issued by national pulp and paper associations or materials testing organizations such as the Technical Pulp and Paper Association (TAPPI) or the American Society for Testing and Materials (ASTM) in the United States. In the area of paper optics, testing methods have been greatly in¯uenced by procedures devised by the International Commission on Illumination (CIE). As for other paper testing procedures, the International Organization for Standardization (ISO) has undertaken the task of integrating different procedures into more universally acceptable methods. Test methods for opacity, brightness, and other optical paper and board characteristics are listed in Tables 1 and 2. Unfortunately, even the most basic re¯ectance measurements such as opacity and brightness may re¯ect varying measurement conditions and instrumentation design. Therefore, the meaning of paper re¯ectance terms such as re¯ectance, brightness, opacity, and so on is nonspeci®c unless de®nitions are provided through standard methods or other procedures. For these terms to be universally applicable, the measurement procedures should rely only on spectral and geometric light characteristics and should not be governed by instrument design. Similarly, the size, shape and orientation of the illuminated paper specimen must be accounted for when quantitative values are obtained. All re¯ectance measurements are relative, that is, the re¯ectance value is obtained relative to that of a standard sample as described later in this chapter (Section IV.C). Light Geometry CIE speci®es four geometries for light re¯ectance measurements (Fig. 3). Of these, diffuse illuminationÐnear normal viewing (d=0 in Fig. 3) is the Table 1

Standard Methods for Opacity and Brightness

International Organization for Standardization United States Japan Canada Scandinavia Germany

Opacity

Brightness

ISO 2471 TAPPI T 425 TAPPI T 519 ASTM D 589 JIS P-8138 CPPA E.2 SCAN-P 8 DIN 53 146

ISO 2470 TAPPI T 452 TAPPI T 525 ASTM D 985 JIS P-8123 CPPA E.1 SCAN-P 3 DIN 53 145

100

Borch

Table 2 International Standards and TAPPI Test Methods and Technical Information Sheets (TIS) Referred to in the Text International Organization for Standardization International Standard ISO 2469 (1994). Paper, board and pulpsÐMeasurement of diffuse re¯ectance factor International Standard ISO 2470 (1977). Paper and boardÐMeasurement of diffuse blue re¯ectance factor (ISO brightness) International Standard ISO 2471 (1977). Paper and boardÐDetermination of opacity (paper backing)ÐDiffuse re¯ectance method International Standard ISO/DIS 8254-1. Paper and boardÐMeasurement of specular glossÐPart 1: 75 degree gloss Technical Association of the Pulp and Paper Industry TAPPI T 425 om-91. Opacity of paper (15 /diffuse illuminant A, 89% re¯ectance backing and paper backing. TAPPI T 442 om-95. Spectral re¯ectance factor, transmittance and color of paper and pulp (polychromatic illumination) TAPPI T 452 om-92. Brightness of pulp, paper, and paperboard (directional re¯ectance at 457 nm) TAPPI T 480 om-92. Specular gloss of paper and paperboard at 75 degrees TAPPI T 515 om-94. Visual grading and color matching of paper TAPPI 519 om-96. Diffuse opacity of paper (d=0 paper backing) TAPPI T 524 om-94. Color of paper and paperboard (45 =0 geometry) TAPPI T 525 om-92. Diffuse brightness of pulp (d=0 ) TAPPI T 527 om-94. Color of paper and paperboard (d=0 geometry) TAPPI T 560 pm-96. CIE whiteness and tint of paper and paperboard (using d=0 , diffuse illumination and normal viewing) TAPPI T 562 pm-96. CIE whiteness and tint of paper and paperboard (using 45 =0 directional illumination and normal viewing) TAPPI T 653 pm-90. Specular gloss of paper and paperboard at 20 degrees TAPPI TIS 0804-01 (1990). Light sources for evaluating papers including those containing ¯uorescent agents TAPPI TIS 0804-03 (1987). Interrelation of re¯ectance, R0 ; re¯ectivity, R1 , TAPPI opacity, C0:89 ; scattering, s; and absorption, k TAPPI TIS 0804-04 (1988). The determination of instrumental color differences TAPPI TIS 0804-05 (1987). Indices for whiteness, yellowness, blue re¯ectance factor, and luminous (green) re¯ectance factor

geometry agreed upon for routine paper re¯ectance measurements by ISO and is therefore often used in the paper industry on a worldwide basis. However, 45 illuminationÐnormal viewing (45 =0 in Fig. 3) and exceptions from CIE geometries are also common, especially in the standards of TAPPI in the United States (Table 2). These geometries will not produce identical re¯ectance values for any sheet structure. A change from normal to diffuse incident light makes both transmitted and light distributions more diffuse. Light reversalÐthat is, interchange of detector and light source (45 =0 $ 0 =45 and d=0 $ 0 =d in Fig. 3)Ðproduces identical results only when the irradiated sample is a perfect diffuser, a requirement that may be approximated but not ful®lled for paper and paperboard. For uncoated sheets,

Optical and Appearance Properties

Fig. 3

101

CIE-speci®ed light geometries for re¯ectance measurements.

the size and orientation of the ®bers in the sheet surface may affect the light distributions [15,52,89,90]. Similarly, a collimated incident light beam (as in the 45 =0 and 0 =45 geometries) introduces the possibility of directionality in the re¯ected light [90]. That is, surface formation and anisotropy due to ®ber orientation may produce ¯uctuations in measured light when the sheet is rotated under ®xed incident light, a characteristic that may be put to use for speci®c instrumentation designs [192]. The feasibility of further treatment of re¯ectance and transmittance measurements by theories like that of Kubelka and Munk (Section V.A) depends on the manner in which the data are obtained. Kubelka±Munk calculations require diffuse light geometry [86]. The stepwise calculation of re¯ectance and transmittance characteristics in paper structures simulated by layered models (Section V.B) does not require a speci®c radiation geometry though the optical properties will depend on the nature of it [103]. ISO Standardization The issue of ISO paper standards is an effort to create international agreement on optical paper measurements [26]. The optical standards ISO 2469±2471 may be used for routine re¯ectance characterization (Table 2). ISO 2469 [22] speci®es the equipment and procedures for measuring the diffuse re¯ectance (re¯ectance factor) of pulp, paper and board. ISO 2470 and 2471 de®ne the determination of brightness and opacity under diffuse illumination in the following way where all re¯ectance measurements should be given in percentage units. Diffuse Blue Re¯ectance Factor (ISO Brightness) (R1 ) This brightness measurement is the re¯ectance of an opaque pile of sheets relative to that of a perfect re¯ecting diffuser (ideal uniform diffuser of no absorptance or transmittance) at an effective wavelength of 457 nm (blue part of the wavelength spectrum as shown in Fig. 4) and as determined with an instrument employing diffuse illumination and normal viewing. In practice, the measurement of R1 means that there is no measurable change in re¯ectance when the thickness of the sheet structure being measured for re¯ectance is doubled. Measurement of ISO Brightness requires the elim-

102

Borch

Fig. 4

Brightness: The re¯ectance of blue light. (From Ref. 145.)

ination of gloss components in the re¯ected light ¯ux by the use of a gloss trap [26,173]. Opacity (Paper Backing) …R0 =R1 † Opacity is the ratio between the amount of light re¯ected from a single sheet with a black (total light absorbing) backing, R0 (re¯ectance factor), to the amount of light re¯ected by the same paper backed by an opaque (not light-transmitting) pile of the same paper, R1 (intrinsic re¯ectance factor). Using the diffuse re¯ectance method, spectral light characteristics should be adjusted according to conditions speci®ed by CIE standard illuminant C (Section III.C). The most notable difference between ISO 2470 and 2471 is the difference in the spectral characteristics under which the re¯ectance measurements are carried out. Brightness measurements are obtained under bluish spectral characteristics whereas opacity measurements are carried out under greenish characteristics. For colorimeters this is accomplished by employing appropriate illumination and ®lters in the light paths (Section IV.A). Other Methods Alternative procedures, including those of TAPPI (Table 2), re¯ect changes in light geometry, paper backing, or illumination. For example, alternatives may be speci®ed as follows. Directional Blue Re¯ectance Factor (TAPPI Brightness) …R1 † This brightness measurement is similar to ISO Brightness except that it is determined with an instrument employing 45 =0 light geometry. Using TAPPI T 452 (Table 2) this measurement is sometimes referred to as TAPPI Brightness. Owing to the difference in light geometry, there is no simple relationship between the ISO and TAPPI Brightness scales. Opacity (White Backing) …R0 =RW † In contrast to opacity (paper backing), opacity (white backing) requires only one sheet of paper. It is the ratio between the amount of light re¯ected from a single sheet with a black backing, R0 , to the amount of light re¯ected by the same sheet of paper backed by a standard white, RW , where the white standard has an absolute re¯ectance of W. This measurement has been termed contrast ratio when performed according to TAPPI T 425 and W ˆ 89% (Table 2). TAPPI T 425 de®nes the procedures used in the United States for both

Optical and Appearance Properties

103

opacities (paper and white backing) and speci®es 15 =d light geometry for illuminant A (Section III.C). In addition, TAPPI speci®es methods that employ the same optical geometry as those of ISO for both brightness and opacity determinations. TAPPI T 525 (Table 2) describes the measurement of the brightness of handsheets or machine-dried sheets of pulp similar to ISO 2470. TAPPI T 519 (``printing opacity'') (Table 2) is an alternative to TAPPI T 425 for measuring paper opacity (white backing) that is similar to ISO 2471 and other national methods employing d=0 light geometry and the spectral characteristics of the ISO method. There is no doubt that the development of ISO standards in the 1960s created uncertainty as to how optical paper properties should be speci®ed [64,184]. Paper technologists who had been using perfectly reliable but different characterization methods were not about to change their procedures. This reluctance was intensi®ed in cases where a change in method produced lower values, as, for example, when the perfect diffuser rather than magnesium oxide was applied as the reference standard for a speci®c re¯ectance procedure [173]. C.

Colorimetry

Spectral Re¯ectance The majority of optical measurements that are routinely performed on paper and paperboard are re¯ectance measurements like those described above. That is, a limited number of well-de®ned re¯ectance values are used to describe the sheet appearance. In contrast, colorimetry is the quantitative measurement of spectral re¯ectance variation through a broader range of the visible spectrum. The need for more complete paper optics characteristics is most pronounced when paper is required to exhibit speci®c spectral re¯ectance properties similar to those that are routinely demanded, for example, for paints in the paint industry. White writing and printing papers are required to have a pleasing appearance as de®ned by certain color characteristics that cannot be speci®ed solely by standard re¯ectance values (Section III.D). Similarly, paper and paperboard for packaging are often required to exhibit quite speci®c color characteristics. These requirements, in addition to advances in instrumentation design, have increased the use of color measurements both in the mill and in the laboratory. The CIE Standard System The basic operation in colorimetry is the determination of the tristimulus values XYZ in the CIE standard system [40,84,111,172]. The tristimulus values are calculated through the following equations: …  X ˆ k P x…† d …1† 

…

 d Y ˆ k P y…†

…2†

…  Z ˆ k P z…† d

…3†

where

104

Borch

P d k    x…†; y…†; z…†

ˆ spectral power function ˆ a normalizing factor ˆ color matching functions for a ``standard'' observer

These functions are shown in Fig. 5. The color matching functions for the standard observer are based upon the color matches of a human eye of normal color vision.  Therefore, the y…† response is the luminosity response shown in Fig. 1 that de®nes   lightness (color intensity). The additional x…† and z…† responses produce the amount of the two extra primary colors required to produce for the standard observer the color of the spectral stimulus of unit radiance. The concepts of primary colors, color stimuli, and their effect on the human eye (color vision) are discussed in more detail by Judd and Wyszecki [84]. The ``standard observer'' was adapted in 1931 by CIE based on experiments during which observers obtained color matches at a 2 viewing angle. A supplementary observer was introduced in 1964 for a viewing angle of 4 and up. The latter is characterized by a slightly higher peak and re¯ects visual color matching of ®elds of large angular subtense. The object color is now derived by replacing P d with R H d, where R is the spectral re¯ectance of the object. The equation system is normalized by specifying kˆ

100  d H y…†

…4†

but other more appropriate constants may be chosen [84]. The equation system is fully de®ned by the spectral object re¯ectance R and the spectral distribution H of the light illuminating the paper. The appropriate summations are often performed by

Fig. 5

Color matching functions for 1931 CIE standard observer. (From Ref. 84.)

Optical and Appearance Properties

105

minicomputers attached directly to the photometric instrumentation. Figure 6 shows examples of spectral light re¯ectance variations for white and colored bond papers measured relative to a barium sulfate standard. The XYZ coordinates (luminance or re¯ectance factors) were calculated for the spectral distribution of the CIE C illuminant described below. Chromaticity Coordinates The description of color measurements in chromaticity coordinates or trichromatic coef®cients is merely an expression of XYZ values in the form x ˆ X=…X ‡ Y ‡ Z†

…5†

y ˆ Y=…X ‡ Y ‡ Z†

…6†

z ˆ Z=…X ‡ Y ‡ Z†

…7†

where y may be plotted against x as shown in Fig. 7 for the sheets analyzed in Fig. 6. Hue, the attribute of color perception by means of which the paper is judged to be red, yellow, and so on, is described quantitatively as the dominant wavelength, that is, the wavelength that if added to the illumination will match a given color.

Fig. 6 Spectral object re¯ectance variations relative to Eastman white re¯ectance standard for white and colored papers (Xerox 4024 DP and DP colors).

106

Fig. 7

Borch

Chromaticity coordinates for the sheets described in Fig. 6.

Saturation is the quality of color sensation by which the paper shows different purities of any one dominant wavelength. It is measured as excitation purityÐthe percent departure from a neutral gray of the same lightness. Both dominant wavelength and saturation purity can be determined using a chromaticity diagram [84] or by means of calculation aids [114,171]. TAPPI T 524 and T 527 (Table 2) describe the measurement of color characteristics under 45 =0 and d=0 light geometry, respectively. Illuminants In 1931 CIE de®ned standard illuminants to represent incandescent light, sunlight, and daylight (illuminants A, B, and C in Fig. 8). Their positions in the chromaticity diagram are given in Fig. 9. The variation in X, Y coordinates in which the illuminants are situated in Fig. 9 de®nes their color temperature, that is, the temperature to which a blackbody must be heated to produce the same color stimuli. For paper color, the energy function H used to de®ne tristimulus coordinates is generally that of illuminant C which has also been speci®ed for ISO re¯ectance measurements. However, none of the illuminants given above is adequate to simulate the short wavelength spectral distributions of different daylight conditions. CIE has therefore added an additional series known as D illuminants, based on spectrophotometric measurements at different locations (Fig. 10) [11,84]. A variety of these and other illuminants are increasingly being used for both colorimetric and routine

Optical and Appearance Properties

107

Fig. 8

Relative spectral power distributions of CIE illuminants A, B, and C. (From Ref. 84.)

Fig. 9

CIE 1931 chromaticity diagram. (From Ref. 84.)

108

Fig. 10

Borch

Relative power distributions of CIE D illuminants. (From Ref. 84.)

re¯ectance characterization of paper and board [22,24]. TAPPI TIS 0804-01 (Table 2) provides the color temperatures of a number of illuminants presently used in the United States and elsewhere. In practice, a range of illuminants is generally created by a limited number of lamps using different ®lters. This can greatly expand the range of visual color impressions that a single sheet of paper may create. Metameric matches occur when the same X; Y; Z coordinates can be calculated for different P variations. This is possible for either the same R at different H values (same object color illuminated by different illuminants) or the same H for different R values (different object colors under the same illumination). Metamerism is discussed in detail by Judd and Wyszecki [84]. Visual Color Matching Visual color matching is still being used in the paper industry (TAPPI T 515) (Table 2), though quantitative measurements are becoming more widespread for routine characterization of unprinted colored and white paper (Section III.D). The human eye is quite sensitive to changes in hue. Lightness affects the eye less. The term strength is used to indicate color difference between two samples by specifying the concentration difference of the dyes in the color formulations [176]. Strength differences in single dye concentrations create color differences within single color areas (pink in red, canary in yellow, etc.). For general color matching, the Munsell system is the most widely used system that does not rely on instrumental measurements [124]. A number of manufacturers produce color charts that simulate the Munsell standards and that can be visually matched to paper products under well-de®ned illumination conditions. Huey [65,66] reviewed the practices for color matching in industries where color is assessed visually. Methods are available for converting the CIE luminance factor Y to the Munsell value [119]. Color Scales Numerical color scale calculations are often easier to relate to visual color perceptions than tristimulus values. In the paper industry, opponent color

Optical and Appearance Properties

109

systems like the CIE 1976 L a b color space are now widely used for color characterization (Fig. 11) [82,161]. This is one of two color scales proposed by CIE [5]. The L a b color space, that is suitable for small color differences [95], is generally available for direct readout in modern color-measuring instruments. The L u v space applies to larger color differences or when color differences result from additively mixed lights [5,95]. Judd and Wyszecki [84] and Hunter [73] describe the application of various color spaces, and calculator programs are available for color space calculations [115,194] for instruments not equipped with appropriate microprocessors. Any color space can be mathematically derived from the more universally accepted X; Y; Z values under the illumination and observer conditions that pertain to the color space. For example, L a b coordinates are derived in the following manner: L ˆ 116…Y=Yil †1=3 16   a ˆ 500 …X=Xil †1=3 …Y=Yil †1=3   b ˆ 200 …Y=Yil †1=3 …Z=Zil †1=3

…8† …9† …10†

where Xil , Yil , Zil are the tristimulus values for the standard illuminant used …Yil ˆ 100). The preference for L a b over XYZ is due to the user's easier identi®cation of visual shades with the L a b color space. Thus, positive a values indicate red, negative indicate green. Similarly, b shows yellow when positive and blue when negative. The zero levels of both a and b indicate grayness that increases to white when L increases (see also ``Fluorescence'' in the following section). Color spaces have been graduated further to provide meaningful scales for the visual perception of color changes [118]. Color Difference A change in color from one point to another in a color space like that shown in Fig. 11 is simply described as the distance
E  between the points:  1=2 …11†
E  ˆ …
L †2 ‡ …
a †2 ‡ …
b †2 where
L ,
a , and
b are differences in coordinates. Similarly, the Friele±MacAdams±Chickering color difference formula (FMC II color difference) expresses color differences in terms of a single number [31]. A

Fig. 11

CIE 1976 (L a b ) color space.

110

Borch

computer program is available for the calculation of FMC II color differences in terms of X, Y, Z coordinates for the two colors constituting the color pair to be evaluated [16]. These and other approaches for measuring color differences are described by TAPPI (TIS 0804-04) (Table 2). A color difference expressed by a single number will never adequately describe the three-dimensional shift in color coordinates. However, color differences like those described above are useful when the performance characteristics of colorimeters are evaluated and compared with each other [13,14]. Small color differences are especially applicable to the measurement of metamerism (metamerism index) and to the description of changes in whiteness as described in the following section. Hunter and Harold [74] present a comprehensive review of color scales and color difference systems. The development of color scales and their relation to visual color perception by the Optical Society of American is described by Nickerson [126]. Hunt [69] gives concise de®nitions of CIE colorimetry procedures. D.

Whiteness, Yellowness, and Fluorescence

Whiteness Based on colorimetric measurements, several whiteness de®nitions have been introduced for industries in which this appearance characteristic is important. Generally, these de®ne a whiteness index, that is, a single number obtained from re¯ectance measurements. More recently, CIE introduced formulas for whiteness and tint (color of lightly colored samples) that are also preferred by TAPPI (TAPPI T 560 and T 562) (Table 2). WCIE ˆ Y ‡ 800…xil TCIE ˆ 1000…xil

x† ‡ 1700…yil

x† ‡ 650…yil

y†

y†

…12† …13†

where WCIE and TCIE quantify whiteness and tint, respectively, Y is the luminositydependent tristimulus factor (luminance factor), and x and y are chromaticity coordinates for the sample illuminated by illuminant D65 with chromaticity coordinates xil and yil . Formulas are also available for calculating whiteness and tint from measured L a b values [44]. Although the calculation procedure is straightforward for CIE whiteness and tint, their application requires careful consideration of instrumentation characteristics and, in particular, consideration of ¯uorescence in illuminated and re¯ected light ¯uxes [140] as discussed in the following. The problems encountered in applying these measurements using different instrumentation designs are described by Griesser [48], and it is likely that further standardization efforts will be carried out [23,24]. Yellowness The yellowing of paper due to light exposure can also be quanti®ed by colorimetry. As for whiteness, a number of speci®c procedures and formulas have been proposed for the measurement of yellowness or a yellowness index. TAPPI has issued technical information sheets (TIS 0804-05) (Table 2) that describe the measurement of both whiteness and yellowness indices. Fluorescence Fluorescence is the ability of a substance to absorb light at one wavelength and emit it at a higher wavelength. In the paper industry, optical bright-

Optical and Appearance Properties

111

eners are often added to improve the appearance properties of printing and writing papers [169]. Owing to the nature of the optical brighteners added, the ¯uorescent whites shift the color coordinates toward the blue area of the color spaces (for example, negative b values in the L a b space). Ideally, a complete spectral characterization of a ¯uorescent material requires that spectral re¯ectances be obtained at each individual wavelength as discussed by Judd and Wyszecki [84]. The incoming light generates re¯ected and ¯uorescent radiation over a wavelength band that is wider than the single wavelength of irradiation. Consequently, it is necessary to scan re¯ected radiation spectrophotometrically and separate it into the different spectral components. The procedure must be repeated for all wavelengths smaller than that for which maximum ¯uorescence is emitted. This can be time consuming and is feasible only with the use of spectrophotometric instrumentation and procedures designed for that purpose (bispectral ¯uorescent colorimetry) [196]. Using conventional spectrophotometric instrumentation with well-de®ned illumination sources paper ¯uorescence measurements can be obtained by using ``reversed'' optics [37,183]. The spectrum is obtained by scanning re¯ected instead of incoming light with the monochromator. A comparison with the spectrum obtained for nonreversed optics permits the determination of non¯uorescent re¯ected light by locating the lowest re¯ectance at each wavelength [63]. Numerical methods for describing whiteness and ¯uorescence are closely linked to the design of the instrumentation that is required to measure them. For example, the addition of ¯uorescent dyestuff to bleached pulp can be quanti®ed by inserting an ultraviolet ®lter between light source and paper sample in a brightness meter [53]. Suitable combinations of colorimetric measurements can be applied when the ¯uorescent component is added by use of the proper light source and ®lters (TAPPI T 452, Appendix C) (Table 2) [106,147]. Two methods for whiteness evaluation of ¯uorescent paper are described by Grum and Patek [50]. Similarly, methods have been given for predicting ¯uorescence measurements under changing illumination without experimental measurements [49]. More recently, Griesser [47,48] described ¯uorescence measurements in paper optics and the advantages of using a speci®c illumination device (Gaertner-Griesser UV adjustment device). Visual Ranking Anyone who has been involved in evaluating the shade of white bond papers will realize the dif®culties involved in getting a consensus on whiteness among a number of sheet candidates showing slightly different hue [43]. They can all be accurately characterized through their tristimulus values, yet the human eye does not perceive color as an absolute physical sensation. The evaluation of each individual sheet is affected by the color of the sheets to which it is being compared and by the nature of the illuminants. Corte [35,36] discusses the psychological uncertainty in whiteness measurements and recommends the design of instrumentation that simulates psychological sensation in color and other appearance measuring techniques. Yet recent studies do show promise for de®ning current instrumental methods for the subjective assessment of paper ranking. For example, in that of Jordan and O'Neill [83] the CIE whiteness [Eq. (12)] measured on colorimeters using quartzhalogen lamps produced excellent correlations with visual ranking under of®ce light.

112

Borch

This study also demonstrates the necessity of adjusting the illumination characteristics to those under which visual rankings are carried out. Popson et al. [144] discuss the optical characterization of color removal for recycled pulp and paper products where brightness measurements are inadequate. E.

Gloss and Surface Texture Perception

Specular Re¯ectance Gloss, like color, is psychophysical. The physical effect is one for which re¯ected light intensity is directional (Fig. 12). Gloss, or directional (mainly specular) light re¯ectance, is created in the uppermost surface layer of the sheet structure (thickness less than half the wavelength of the incident light). Consequently, gloss depends strongly on the smoothness of the paper surface, as discussed by Van den Akker and Sears [187,188] and demonstrated experimentally for clay-coated sheets by Gate et al. [46]. Because gloss is directional, it is attractive to consider its physical basis as the enhancement of diffuse re¯ectance by Fresnel re¯ection from the sheet surface [59]. Fresnel re¯ected light ¯ux is dependent upon the refractive index ratio n at the interface and the angle of incidence i, according to the formula " # IR 1 sin2 …i r† tan2 …i r† ‡ ˆ …14† I0 2 sin2 …i ‡ r† tan2 …i ‡ r† where IR is the intensity of the re¯ected light beam and I0 is the intensity of the incoming light beam. Further, i and r (angle of refraction) are related such that   1 sin i …15† r ˆ sin n In contrast, a perfect re¯ecting diffuser provides re¯ected light ¯ux that is independent of the angle of re¯ection. A combination of Fresnel re¯ected and diffusely re¯ected light will yield light distributions of the type shown in Fig. 12, with an

Fig. 12 Directional light re¯ectance (gloss). Re¯ected light intensity is enhanced in the direction of mirror re¯ection ( 45 ). (From Ref. 84.)

Optical and Appearance Properties

113

increase in the specular component at higher angles of incidence due to the nature of the equations shown above. Gloss Characterization Because gloss is directional and its measurement generally includes the diffusely scattered light components, it is necessary to achieve a measuring geometry that is meaningful for the surface that is being analyzed. Hunter [70] distinguishes ®ve different kinds of surface gloss as shown in Table 3. The gloss perception that is applicable to the majority of paper and board products is of a lowTable 3

Types of Gloss and Their Measurement

Kind of glossiness Specular

Sheen

Contrast

Distinctness of image

Absence of bloom

Source: From Ref. 84.

Gloss factor measurement Incident ¯ux  (at 60 † ˆ reflected ¯ux from perfect mirror v (at 60 † ˆ reflected ¯ux from sample Gloss factor ˆ v =1 Incident ¯ux 1 (at 85 † ˆ reflected ¯ux from perfect mirror v (at 85 † ˆ reflected ¯ux from sample Gloss factor ˆ v =1 Incident ¯ux 1 (at 60 ); v1 (at 60 † ˆ reflected ¯ux from sample v2 (at 0 † ˆ reflected ¯ux from sample Gloss factor ˆ v1 =v2 Incident ¯ux 1 (at i); v (at j† ˆ reflected ¯ux from sample Angle of view ( j) differs by a few minutes of arc from angle of mirror re¯ection ( j) Gloss factor ˆ rate of change of v with angle of incidence (i) Incident ¯ux 1 (at i) v2 (at j† ˆ reflected ¯ux from sample Angle of view ( j) differs from the angle of mirror re¯ection ( i) by a few degrees Gloss factor ˆ v1 =v2

114

Borch

to-moderate gloss (shininess at grazing angles) that requires a relatively large receptor angle (75 in TAPPI T 480) (Table 2). Papers of higher gloss such as waxed and cast-coated sheets are measured at a much smaller receptor angle (20 in TAPPI T 653) (Table 2). Figure 13 describes numerical gloss values obtained using a 20 or 75 receptor angle for perceived visual gloss ratings. It is apparent that for midrange gloss values a nearly linear relationship exists between 75 TAPPI gloss values and visual gloss perception. Contrast gloss or luster (Table 3) is a measure of the ratio of specularly re¯ected light to diffusely re¯ected light. Relying on the effect of polarization, the ®rst commercially available glossmeter, the Ingersoll Glarimeter, was designed to measure contrast gloss [76]. Initially polarized light (light waves oscillating in a speci®c direction) retains its directionality when it is Fresnel re¯ected. In contrast, diffusely re¯ected light cannot maintain directionality owing to the multiple light interaction events beneath the sheet surface. The same effect has been used in the design of instruments suitable for determining the effect of the paper surface in high quality printing applications at high print gloss levels [25,99,146] and in gloss anisotropy of unprinted paper at lower gloss levels [90]. A novel contrast gloss method is described by Nishiwaki [128]. Based on instrumentation design of Fukushima [58], re¯ected light distributions are measured on a goniophotometer in which the sample is rotated and the illuminator and receptor axes are ®xed. Diffusely and specularly re¯ected light components are evaluated by plotting the logarithmic intensity against the

Fig. 13 Numerical gloss values for visual gloss perceptions at 20 and 75 measuring angles. (From Ref. 145.) 

An ISO procedure (ISO/DIS 8254-1) (Table 2) is being drafted for 75 gloss.

Optical and Appearance Properties

115

square of the rotation angle. Distinctness-of-image gloss is obtained by comparing specular re¯ectance to the re¯ectance obtained when the angle is changed slightly from that of the specular direction (Table 3) [73,125,187]. This property and the effect of angular resolution on gloss measurements have been discussed in more detail by Van den Akker [187] based on earlier goniophotometric studies of coated paper and board [188]. Surface Texture All routine re¯ectance measurements require an illuminated paper surface area that is suf®ciently large to eliminate gross ¯uctuations in re¯ectance values from different spots on the same sheet structure. In contrast, if the area measured for re¯ectance is limited in size, re¯ectance will vary owing to surface unevenness that shows up in scanning the paper surface. This effect can simulate the appearance phenomenon of mottle (optical unevenness) of both printed and nonprinted sheets [146]. The re¯ectometer measures the re¯ectance of a small paper area situated within a larger one that is similar in size to the area commonly measured by routine instrumentation. The re¯ectance of the large area is simultaneously measured and compared with the re¯ectance from the smaller, spotlike area, which will ¯uctuate due to optical unevenness [100]. Speckle is the nonuniform appearance of printed matter that is caused by nonprinted depressions in the paper surface. Poulter [146] reviewed the measurement of print unevenness. A mathematical model that accounts for both mottle and speckle of solid prints was proposed by Wahren and Bryntse [189]. Oittinen [134] reviewed the role of paper as information carrier, including the effects of unprinted paper characteristics. As will be shown in Section IV.A, improved methods for characterizing both printed and unprinted paper structure on a micro scale rely on instrumentation that can be designed speci®cally for the measurement of re¯ectance and transmittance from areas of microdimensions. Data analysis methods using optical character recognition (OCR) systems and newer printing methods such as nonimpact inkjet devices impose speci®c demands on microscale paper surface structure and appearance. Consequently, the use of equipment and techniques for measuring and analyzing variations in microscale re¯ectance has increased greatly [45,80,142,195]. The appearance of surface ®nish can be as important as the more general re¯ectance characteristics. There exists here a papermaking terminology that is nonspeci®c and based on visual perception and ®nishing techniques. Uncoated printing papers may be machine ®nished or increased in smoothness to levels of English, writing, and glazed ®nish. A grainy, rough texture is antique ®nish, which, with increased ®neness, becomes eggshell and vellum ®nish. Cockle ®nished papers are desirable in business paper manufacture. There are no speci®c guidelines for paper ®nish terms, and well-de®ned instrumental methods are lacking for the measurement of these appearance characteristics. Emerton et al. [39] present a selection of photomicrographs of printing paper surfaces. More recently, attempts have been made to correlate the visual appearance of papers of different random texture through image analysis [6]. F.

Light Transmittance Methods

Transmittance Versus Re¯ectance The number and variety of re¯ectometers that are in use today demonstrate the extensive use of re¯ectance measurements for

116

Borch

routine control of the appearance properties of paper and board. Nevertheless, this does not mean that in principle optical appearance characteristics can be analyzed only by using re¯ectometry. When photometric instrumentation for optical paper characterization was still in its infancy, transmittance methods were shown to give reliable measurements of some paper appearance properties such as contrast ratio and printing opacity (Section III.B) [110]. Harrison [58] presents an extensive review of early opacity measurement methods. Properly done, the measurement of the transmitted light ¯ux can present advantages over the measurement of re¯ected ¯ux. Re¯ectance and transmittance are complementary, always adding up to 100% minus absorbed light (Fig. 2, Section III.A). Therefore, variations in transmittance may be greater than variations in re¯ectance thereby increasing the measurement precision for transmittance analysis. Together with a properly applied light scattering theory like that of Kubelka and Munk (Section V.A), transmittance measurements can be taken advantage of in the measurement of sheet opacity and other optical sheet characteristics, both in the laboratory and in the mill (on-line appearance control) [57,86,105,120,163]. Unfortunately, light transmitted through a partially opaque paper sheet does not lend itself to measurement as easily as does light re¯ected from the same structure with a suitable backing. The spatial distribution of transmitted light is less well de®ned and more sensitive to incoming light geometry, especially for collimated incoming light ¯ux. Changes in the formation of the sheet sometimes introduce obscure changes in transmitted light distributions, as discussed below in relation to formation testing [133]. Therefore, transmittance characterization should be limited to instances where good correlation between measurement and sheet properties can be established [163] or for specialized purposes (transparency and formation). In general, re¯ectance methods using well-designed instrumentation are more desirable for appearance characterization, including sheet opacity. Transparency Transparency ratio is the ratio of near regular transmittance to near hemispherical transmittance. This measurement characterizes glassine papers, for which transmitted light intensity distribution becomes directional, similar to that for gloss re¯ected light shown in Fig. 12. Regular transmittance is the light transmittance normal to the sheet surface (Fig. 14). Hemispherical light transmittance is a measure of the total light transmittance obtained by placing a light-collecting sphere behind the sample (Fig. 14). The measurement re¯ects the clarity with which an object placed at a distance from the sheet is visually perceived through the sheet structure [73]. Formation As de®ned by Norman and Wahren [133], ``Formation is the manner in which the ®bers are distributed, disposed, and intermixed to constitute the paper.'' It affects paper appearance and also in¯uences the strength properties of paper. For example, semitransparent paper products will show a cloudy or blotchy appearance if the sheet constituents are not suf®ciently evenly distributed in the paper structure. The psychological aspect of paper formation is dif®cult to characterize satisfactorily [55]. Consequently, instrumentation for formation measurements has varied in design principle and has often been of an experimental nature [133]. Most methods depend on the scanning of paper sheets with visible light and the detection of light transmitted through microdomains of the sheet structure. The scanning area is simi-

Optical and Appearance Properties

Fig. 14

117

Measurement of regularly and diffusely (hemispherically) transmitted light.

lar in size to that of the smaller area used for the measurement of optical unevenness by re¯ectance comparison as described earlier in this chapter (0.1±1 mm beam diameter). The light intensity is monitored continuously and either is plotted as a function of paper position, thereby providing a ``map'' of paper basis weight variation [27], or is treated electronically to give a measure of paper ¯oc frequency [132]. Because the radiation beam size is so small, the distribution of light transmitted through the paper becomes very sensitive to both the nature of the sheet structure and the sheet thickness. Figure 15 shows transmitted light intensities obtained for sheets of bleached and unbleached pulps [133]. In spite of lower basis weight, the light is diffused more by the bleached sample. Also, reducing the Fresnel index ratio

Fig. 15 Transmitted light intensity distributions for sheets of (upper) unbleached and (lower) bleached pulps. (From Ref. 133.)

118

Borch

by wetting the ®bers affects the bleached structure more than the light absorbing structure. To some degree, the in¯uence of composition and scattering in the sheet structure can be eliminated by using beta-ray transmission in a scanning or stationary mode. In the stationary mode, beta radiographs are recorded on X-ray ®lms and these are then scanned in the formation tester by means of visible light [87,133].

IV.

INSTRUMENTATION CHARACTERISTICS

A.

Instrument Types

Classi®cation Photometers for measurements of paper appearance are generally spectrophotometers, colorimeters, glossmeters, goniophotometers, or microdensitometers. Image analysis systems are being used more and more for spatial comparison of small-area optical images, especially for the evaluation of formation and gloss. These instruments are not speci®c to the paper industry, and design and cost often re¯ect other applications. Modern instruments are gaining signi®cant improvements in design due to microprocessor technology, ®ber optics and photodetector application, etc., and the following descriptions are intended to describe instrumentation principles rather than state-of-the-art design. Scott et al. [161] discuss some of the instruments that are widely used by the paper industry today. Spectrophotometers Figure 16 shows a typical spectrophotometer. A double prism provides monochromatic light that is split into two beams that alternately irradiate the sample and a surface for which the re¯ectance is known (the standard). Reversible optics allow the recording of ¯uorescence in the reversed mode. When the complete recording of light re¯ectance through the visible range is not needed, routine re¯ectance measurements may be performed on the handier colorimeters

Fig. 16

Diano Match-Scan color spectrophotometer.

Optical and Appearance Properties

119

(Fig. 17) or abridged spectrophotometers where the monochromator is replaced by narrowband ®lters. Because of its spectral characteristics, the spectrophotometer usually surpasses the ®lter colorimeter with regard to color measurement accuracy, and it is often the standard with which less sophisticated instrumentation is compared [12,13]. TAPPI T 442 (Table 2) describes instrumental procedures for measuring re¯ectance and transmittance characteristics when a spectrophotometer is available. In this case, measurements are carried out with ``reversed'' optics as described in Section III.D. Colorimeters In the ®lter colorimeter, different spectral characteristics are achieved through the placement of ®lters in front of the photodetectors, as shown in Fig. 17 [38]. This arrangement permits more freedom with regard to illumination and vastly shortens the time required to obtain routine re¯ectance measurements. A great number of instruments have been designed that rely on both 45 =0 and on d=0 geometry as, for example, in Fig. 17. Diffuse illumination requires the use of traps and baf¯es for the elimination of gloss and incident light on the sample, which come directly from the light source. In modern colorimeter design, advantage is taken of illumination and detection via ®ber optics, silicon photodiode light sensing technology, and automatic electronic measures to ensure light and calibration stability. Often these are ``hybrid'' instruments, re¯ecting more a spectrophotometer than a classical ®lter colorimeter design [22,187]. Though mainly intended for colorimeteric characterization of a number of materials including paper, they may also provide calculations of the routine paper re¯ectances described in Section III.B through their built-in microprocessor. Glossmeters A glossmeter is a type of photometer in which the illumination and photodetector systems are at ®xed angles to each other. The light geometry is strictly

Fig. 17 Measuring principle of the Zeiss Elrepho colorimeter. Ph1 , photocell; Ph2 photocell; F, ®lter; GK, neutral wedge; S, swing-in standard; A, sample; MB, measuring diaphragm; N, balancing instrument; PK, photometer sphere; V, ampli®er.

120

Borch

speci®ed as in that for TAPPI T 480 (Table 2) shown in Fig. 18. Because glossmeters are designed to measure physical attributes (Section III.E) rather than spectral ones, their design requires more attention to light uniformity and geometry than that of colorimeters employing diffusing spheres [56]. Goniophotometers The main characteristic of the goniophotometer is the continuous angular adjustment of the photodetector system relative to ®xed or adjustable radiation geometries, as shown in Fig. 19 [72]. Here, both sample holder and lamp arm are rotatable to enable any combination of incident and re¯ected light angles to be created in the horizontal plane. Other con®gurations are possible for both re¯ected and transmitted light geometries, additional azimuthal adjustments, and measurement of spectral light distributions (spectrogoniophotometer) [72,73]. The more complex systems are precision instruments that demand considerable attention to optical alignment and measurement procedures. Photometers, like glossmeters that require light measurement at ®xed speci®c angles, sometimes permit adjustment to account for various gloss types [125]. In this case, therefore, they are goniophotometers, where freedom in angular measurement is more restricted. Both glossmeters and goniophotometers require the accurate de®nition of both incident and re¯ected light distributions, especially if high angular resolution is required. Because of their complex design and the time-consuming measurement procedure, goniophotometers such as the one diagrammed in Fig. 19 are mostly applied in the research laboratory [193]. Routine measurements of directional light are mainly obtained using ®xed geometries (glossmeters). Microdensitometers A microdensitometer measures re¯ected or transmitted light from areas of the sheet structure that are suf®ciently small for the testing of surface unevenness or formation. The STFI formation tester shown in Fig. 20 is a research design, but it eliminates a number of dif®culties encountered with similar commercially available testers [133]. The incoming light is diffused by illumination from inside a Te¯on cylinder, and light transmitted through the sample can be further diffused by placing a diffusing mat over the test specimen. Re¯ectance attachments allow the collection of re¯ected light from scanned micro areas, a capability also found in instrumentation speci®cally designed for measurement of surface unevenness. The microdensitomer is unique among paper appearance testing instruments in that re¯ectance or transmittance values are collected and compared from different parts of the same sheet structure. As can be seen in Fig. 20, this is accomplished by the automatic revolution of the cylinder and axial motion of the photodetector system. Image Analysis Systems As in formation testing, in image analysis a sheet area is scanned and subsequently evaluated for speci®c characteristics. Electronic software has been developed that is intended to simulate these parameters including that of formation [10,180]. Image analysis systems have proven useful in gloss imaging [104,112,162] and wire-mark analysis [75,174]. Additionally, these systems are obvious candidates for particle count and print quality analysis [68].

Fig. 18 75 glossmeter according to TAPPI T 480 (see Table 2).

Optical and Appearance Properties 121

122

Fig. 19

Borch

Goniophotometer. (From Ref. 72.)

Fig. 20 Formation tester. 1. Driving belt and screw for axial motion of measuring equipment and lamp; 2. Te¯on cylinder; 3. Photomultiplier; 4. Aperture; 5. Lens; 6. Driving belt for rotating the cylinder; 7. Track-hold unit; 8. Translucent ring; 9. Lamp; 10. Cooling air nozzle; 11. Cooling air intake. (From Ref. 133.)

Optical and Appearance Properties

123

Most systems are multipurpose instruments featuring custom design of both hardware and software depending on capability requirements and cost [68]. An example is shown in detail (see Fig. 46, Volume 1, Chapter 16). B.

Design Principles

Optical Geometry The issue of proper re¯ectometer design has been a controversial one [37,168,183,184]. It has rightly been pointed out that the stability of the diffuse re¯ecting coating in the instruments employing diffuse illumination is subject to deterioration. Such instruments therefore introduce an extra variability that can be avoided by using instrumentation that employs 45 =0 light geometry [37,184]. As discussed above, different design geometries will not provide identical measurements of any re¯ectance type for any paper product, and the choice of design depends upon the use of the measurements (mill comparisons versus Kubelka±Munk applications, etc.). For color measurements that require sheet stacking (R1 values), instrumentation geometry becomes less important, and measurement agreement between different instrument types has in some cases been shown to be better than for instruments of similar types [14]. More recently, color measurements of L a b characteristics of Pantone color samples showed measurable differences for different geometries [8]. Glossmeters measure directional re¯ectance and do not, therefore, employ diffusing spheres. However, because all measured light will in practice show some angular distribution (Fig. 12), the maximum light acceptance angle and light distribution of the photodetector system will determine instrument response [56,71]. Colorimeters of modular design allow re¯ectance measurements under different viewing and illumination geometries and sometimes also permit gloss measurements. Equipped with multiple sensors (Fig. 21), glossmeters allow the measurement of gloss and can measure more complex image attributes (e.g., distinctness-of-image gloss) without time consuming goniophotometry [85]. Spectral Characteristics The meaning of any light measurement relies strongly on how well its spectral characteristics are generated by light source, photodetector, and possible ®lters. In particular, the increasing use of ¯uorescent brighteners in ®ne paper manufacture has made it imperative to employ in re¯ectance instrumentation light sources that contain a suitable amount of ultraviolet radiation [42,106,113]. Older instruments that illuminate mainly in the visible range of the spectrum are less suitable than similar types that emit suf®cient ultraviolet light such as the xenon lamp. The switch to UV lamps has increased the problem of maintaining instrument stability and has not eliminated spectral differences when the performance characteristics of different instruments are compared. Figure 22 shows the chromaticity coordinates obtained for the same ¯uorescent sample using ®ve different instruments, all equipped with xenon lamps [42]. The values vary signi®cantly owing to differences in spectral ¯uorescence response. A method using white metameric sample pairs has been devised for testing light sources and natural daylight for relative UV content [4]. Light sources, photodetectors, and possibly ®lters and diffuse sphere coatings are all limited in regard to long-term stability. To some degree, instrument drift is counteracted by the use of microprocessor technology in the more modern instruments. The microprocessor controls the instruments, allows the rapid collection of

124

Borch

Fig. 21 Multi-sensor glossmeter that measures (a) directional re¯ectance at 20 or (b) re¯ected light just outside the 20 direction. (From Ref. 85.)

Fig. 22 Chromaticity coordinates for a ¯uorescent sample as obtained from ®ve different instruments all equipped with xenon lamps. (From Ref. 42.)

Optical and Appearance Properties

125

re¯ectance data at predetermined wavelengths, and also has storage capability for reference and calibration data. Digital output microcomputers automatically derive speci®c colorimetric calculations when programmed to do so. On-Line Control Radiation sources are ideally suited for on-line control, because radiation of suitable intensity will not interfere with the paper structure. For example, the monitoring of moisture and calliper using beta rays, IR radiation, and microwave techniques on the moving paper web are often seen in the mill environment (Chapter 2). Although some of the test equipment for optical properties designed for the control and testing laboratory may be modi®ed for on-line use [32,143,147,191], inmill color control may be better achieved with speci®c instrumentation designed for the mill environment using high intensity illumination and multiple sensors that measure the entire spectral range simultaneously. This allows the spectral evaluation within the time frame required for automatic color adjustments at machine speed (closed loop system) [98,117]. C.

Calibration

Re¯ectometers All commercially available re¯ectometers provide relative re¯ectance values, that is, the measurement is obtained relative to the re¯ectance of a standard sample. The reference standard must be identi®ed if the measurement is to be meaningful. ISO speci®es the following reference standards (ISO 2469) (Table 2): ISO Reference Standard of Level I (IR 1) IR 1 is the perfect re¯ecting diffuser according to CIE 45-20-195. This ideal uniform diffuser is de®ned to have a re¯ectance equal to 1. Until the ISO standard procedures were agreed on, magnesium oxide was the primary standard against which re¯ectance values were compared in re¯ectometry [84]. Unfortunately, samples of magnesium oxide were not suf®ciently stable for extended use, nor were they easy to prepare in a reproducible manner. Therefore, ISO adopted an imaginary primary standard [26,173]. ISO Reference Standard of Level 2 (IR 2) IR 2 is a standard whose re¯ectance factor has been determined by a standardizing laboratory in relation to IR 1. These standards are used by the authorized laboratories for the calibration of their reference instrument. Barium sulfate is a suitable IR 2 standard. Figure 23 shows the absolute re¯ectance of barium sulfate using methods and instrumentation speci®cally designed for absolute rather than comparative light measurements [51]. ISO Reference Standard of Level 3 (IR 3) The IR 3 standard is one whose re¯ectance factor has been determined by an authorized laboratory in relation to IR 2. These standards are used by working laboratories for the calibration of their instruments. They are often paper sheets that have been characterized using the reference instrument at the issuing laboratory. They require proper storage to maintain their optical characteristics during any extended period of use. Dark storage at low temperature and humidity minimizes the effect of aging [181].

126

Borch

Fig. 23 Absolute re¯ectance of Eastman white re¯ectance standard (barium sulfate). (From Ref. 51.)

A list of present standardizing and authorized laboratories is given in Table 4. Re¯ectance standards are also available from instrument makers and laboratories engaged in optics and paper research. Lately, ¯uorescent paper calibration standards have become available from some of the laboratories [113]. Glossmeters, Goniophotometers, and Microdenistomers In contrast to diffuse re¯ectance measurements measured relative to the perfect diffuser or calibrated Table 4

Standard Issuing ISO Laboratories

Standardizing laboratories National Institute for Science and Technology Gaithersburg, MD National Research Council of Canada Ottawa, ON, Canada Physikalisch-Technische Bundesanstalt Braunschweig, Germany Authorized laboratories Bundesanstalt fuÈr Materialprufung Berlin, Germany Centre technique de l'industrie des papiers, cartons et celluloses Grenoble, France Finnish Pulp and Paper Research Institute Helsinki, Finland Pulp and Paper Research Institute of Canada Pointe-Claire, Canada Swedish Pulp and Paper Research Institute Stockholm, Sweden The Technidyne Optical Laboratory New Albany, IN

Optical and Appearance Properties

127

light-diffusing standards, gloss values like those shown in Fig. 13 are generally obtained relative to perfect mirror re¯ections or black glass standards. For example, the theoretical specular gloss standard of TAPPI T 480 (Table 2) is assigned 384.4 gloss units using a mirror of 100 gloss units with a glass standard of refractive index equal to 1.540 for the sodium D lines (589±590 nm). The scanning type instruments (goniophotometers, microdensitomers, etc.) are generally required to measure variation in intensity with angle or sample position. Various standards are employed to ensure instrument stability comparable to that of a spectrophotometer, for which the measurement procedure is time consuming. Hunter and Harold [74] describe standardization of appearance-measuring instruments in detail.

V.

INTERPRETATION OF OPTICAL MEASUREMENTS

A.

Kubelka±Munk Theory

Principle The industry relies on the Kubelka±Munk theory for detailed analyses of re¯ectance measurements [62,94,106]. The theory allows the interconversion of light re¯ectance and transmittance data. Furthermore, it permits the derivation of parameters that are independent of sheet caliper and that can be added together for the purpose of evaluating the optical effect of discrete papermaking components. Details of Kubelka±Munk calculations are given in nearly any textbook dealing with color and re¯ectance. Their application to paper optics in general has been described by Stenius [167], Van den Akker [182,186], and Casey [30]. Homogeneous Sheets The Kubelka±Munk scattering theory is a phenomenological approach to the description of light interactions in a paper sheet. Each increment in basis weight dW absorbs, transmits, and backscatters parts of the incident light (Fig. 24). The change in light transmittance due to scattering is characterized by the scattering coef®cient s or scattering power s dW. The change due to absorption is characterized by the absorption coef®cient k or absorption power k dW. For homogeneous sheets (all elements dW are identical through the crossdirectional sheet structure), differential equations valid for the element dW can be integrated to yield the following relationships for the re¯ectance R0 and the transmittance T of the sheet:

Fig. 24

Kubelka±Munk scattering element.

128

Borch

R0 ˆ

sinh …bsW † a sinh …bsW † ‡ b cosh …bsW†

…16†

Tˆ

b a sinh …bsW † ‡ b cosh …bsW†

…17†

R1 ˆ R0 in cases where W is so large that T ˆ 0, and sinh …bsW† and cosh …bsW† are hyperbolic functions of bsW [92]. The sheet scattering power sW can be calculated from measurements of R0 and R1 by using Eq. (16), which can be transformed from its hyperbolic dependence to the more convenient form sW ˆ

1 1 R0 R1 ln 2b 1 R0 =R1

…18†

The sheet absorption power kW is obtained from measurement of R and calculated values of sW, because k kW …1 R1 †2 ˆ ˆ s sW 2R1

…19†

Further evaluation of sheet scattering and absorption powers in terms of the values of sW and kW for the individual components of multicomponent scattering structures is feasible if the following additives may be assumed: sW ˆ sa Wa ‡ sb Wb ‡ sc Wc ‡




…20†

kW ˆ ka Wa ‡ kb Wb ‡ kc Wc ‡




…21†

where sa , ka , and Wa are coef®cients and weight of component a; where sb , k, and Wb are those of components b; and so on [1,61,67]. The additive equations are valid for homogeneous mixtures (pulp and coating mixtures, pigment-loaded sheet structures, etc.). Further optical interaction must be accounted for by including extra interaction terms in the equation system [186]. Gross heterogeneity in the cross-machine direction of the sheet structure necessitates the calculation procedures for heterogeneous sheets or a summation procedure like that for layer calculations, both shown below. In principle, the light interactions for the Kubelka±Munk element shown in Fig. 24 are similar to those shown for the whole sheet in Fig. 2. Kubelka±Munk calculations are quanti®cations of the re¯ectance, transmittance, and absorptance processes in the sheet elements. The integration procedure produces quantitative relationships that are valid for changes in light ¯uxes through the total sheet cross section. Note that the Kubelka±Munk coef®cients do not account for scattering particle size, refractive index, or absorption coef®cient in their derivation. The speci®c coef®cients s and k are merely expressions of scattering and absorption powers per basis weight unit. The hyperbolic nature of the re¯ectance and transmittance equations Eqs. (16) and (17), was not apparent when the theory was originally applied to paper but was derived later by Kubelka [92]. Meanwhile, extensive graphic aids had been developed to provide the speed and convenience desired by users who were not familiar with or

Optical and Appearance Properties

129

interested in the mathematical detail (TAPPI TIS 0804-03) (Table 2). Today's pocket calculators provide both suf®cient programmability and the more specialized mathematical functions that are necessary to calculate any of the numerous relationships presented in Kubelka's second paper [77,123,135,139,170]. A summary of the equations used for paper optics calculations is presented by Robinson [150]. Recently, the exactness of the Kubelka±Munk derivations was questioned [107]. Viewed in the context of general theories of radiation transport, the theory can be shown to overestimate k by a factor of 2. Whereas this affects the numerical value of k, it does not change the nature of the re¯ectance and transmittance equations for paper optics calculations. Heterogeneous Sheets Paper coatings seldom have the same scattering and absorption coef®cients as the underlying base stock. Therefore, optical calculations for coated paper sheets are more complex than those for uncoated sheets. The calculations of the re¯ectance and transmittance of multilayered structures were also considered by Kubelka [93]. His approach was applied to coated papers by Clark and Ramsay [33] and to coated board by Ramsay [148]. The principle of light interactions in two-layer structures is shown in Fig. 25. The incoming light beam interacts consecutively with both upper layer (coating) and lower layer (base stock). The re¯ectance (R0 †1;2 and transmittance T1;2 of the composite structure are given through the following equations:

Fig. 25

Light interactions in a two-layered structure.

130

Borch


…R0 †1;2 ˆ R01 ‡ T12 R02 1 ‡ R01 R02 ‡ R201 R202 ‡ R301 R302 ‡


ˆ R01 ‡

T12 R02 1 R01 R02

…22†

T1;2 ˆ T1 T2 1 ‡ R01 R02 ‡ R201 R202 ‡ R301 R302 ‡


ˆ

1




T1 T2 R01 R02

…23†

where R01 , T1 , and R02 , T2 characterize re¯ectance and transmittance of coating and base stock, respectively. In practice, it is not necessary to calculate transmittance to obtain …R0 †1;2 because it can be shown that …R0 †1;2 ˆ

…R01

R11 †=R11 …R02 R11 †

R11 …R02 1=R11 †e2s1 W1 b1 …R02 1=R11 †e2s1 W1 b1

…24†

where b1

is defined according to Eq. (18) using reflectance factor R11 of the coating

s1 ˆ scattering coefficient of the coating W1 ˆ weight of applied coating Equation (24) [33,150] is similar to the equation used when homogeneous sheets are characterized on the basis of re¯ectance measurements over ``white'' and ``black.'' Robinson [150] describes methods suitable for characterizing both one-sided and two-sided coated stock. Pauler [141] presents calculation examples for multilayered sheets relevant to papermaking procedures. Van den Akker [186] describes aspects of coating optics in more detail. Error Sources The Kubelka±Munk approach is relatively simple both in principle and in application compared to other scattering theories, such as the Mie theory (Section V.B). Nevertheless, its application to paper optics has not always been successful. There are two main reasons why the mathematics may fail in describing appearance properties adequately: . The sheet structure does not interact with light ``as intended'' due to gross heterogeneity in the Kubelka±Munk de®ned layers. .. Re¯ectance values cannot be measured with suf®cient accuracy for calculation purposes. Migration of pigment particles is a well-known phenomenon in both uncoated paper sheets and paper coatings. Sheet re¯ectance may change, because the scattering power of added pigment is different from that anticipated on the basis of its scattering coef®cient and the amount added. Similarly, pigment particles will produce optical effects that depend on whether they are added into or onto the sheet structure. Filler pigments are often effective light scatterers because of their good dispersion in the bulk sheet structure. Coating pigment particles can aggregate into less effective, fewer and larger aggregates ± especially when pigment and binder

Optical and Appearance Properties

131

concentrations are changed [61]. Nevertheless, theoretically, a high brightness pigment (one with high scattering coef®cient and low absorption coef®cient) will produce maximum re¯ectance when applied in coating layers rather than added to the base stock, assuming that particle aggregation does not alter its scattering ability (Fig. 26) [18,20]. In view of the differences in re¯ectance shown in Fig. 26 it becomes understandable that the application of Kubelka±Munk calculations to optical sheet characterization should always take into account physical sheet structure. Gross heterogeneity in the distribution of scattering particles requires modi®ed calculation procedures as demonstrated theoretically by Van den Akker [185]. Paper re¯ectance measurements [41,130] and diffuse re¯ectance spectroscopy of powders consisting of particles similar in size to paper ®bers [88] indicate that Kubelka±Munk calculations are not suitable for samples with high light absorbing characteristics. The increase in k values over the wavelength range of light absorption produces a decrease in s values, as shown in Fig. 27 for data obtained by Nordman et al. [130]. Consequently, re¯ectance at high absorption levels cannot be accurately predicted solely on the basis of measurements of s at low absorption levels [155] (see Sheet Simulation in Section V.B). Dissatisfaction with predicted re¯ectance values versus those measured experimentally and the calculation work involved in obtaining predicted values ± especially for coated sheets, have in some instances led to calculation shortcuts. For example, Luey [108,164] postulated an opacity factor concept that may lead to better re¯ectance predictions for coated boxboard. Bauer [9] proposed that the light scattering coef®cients of paper coatings can in some instances be determined by the method valid for ®lled papers. Use of this proposal would make the two curves shown in Fig 26 coincide [20]. Robinson and Linke [151] have attempted to account for pigment dispersion characteristics when predicting coating opacities. At high binder levels, the pigment ``®lm'' on the paper surface becomes continuous and it is then necessary to account for Fresnel re¯ected light distributions as de®ned by the refractive index change at the coating surface [Eqs. (14) and (15) in Section III.E] [154].

Fig. 26 Re¯ectance of coated sheet compared to loaded sheet for a high brightness additive. (From Ref. 20.)

132

Borch

Fig. 27 Variations in scattering and absorption coef®cients for sheets made from red-dyed pulp. (From Ref. 130.)

The application of the Kubelka±Munk theory requires that at least two re¯ectance measurements be obtained for the purpose of determining both scattering and absorption coef®cients. It is necessary to minimize the error in both re¯ectance value measurements to obtain the necessary accuracy in calculated coef®cient values. This makes it dif®cult to determine the scattering coef®cient of heavy and dark paper and board products with suf®cient accuracy using re¯ectance measurements only. Therefore, Knox and Wahren [86,105] present instrumentation and a procedure relying on both re¯ectance and transmittance measurements via the Kubelka± Munk theory for that purpose. Similarly, for coated sheets, the substrate re¯ectance R02 must be suf®ciently different from the re¯ectance factor R11 to ensure reasonable accuracy in calculated coating scattering coef®cient values [18,165]. The effect of substrate re¯ectance can be ignored when a translucent substrate is used for successive coating applications [165]. High Brightness Sheets For uncoated sheets of high re¯ectance factor (R1 ), the necessity of obtaining two suf®ciently different re¯ectance measurements for calculations of s can be avoided by using the relationship [84] sW ˆ

R0 …1 R0 †

…25†

This much faster procedure is also useful for the analysis of high brightness coating pigments applied on light absorbing substrates (black glass technique) [179]. Here, the absorption coef®cient is now known and will, in addition to the coating weight, determine how much the scattering coef®cient is lowered relative to that obtained using exact analysis (Fig. 28) [18]. Neglecting the absorption coef®cient also implies that the scattering power sW is, in fact, the ratio between sheet re¯ectance and transmittance (T ˆ 1 R0 ; sW ˆ r0 =T). Lathrop [97] pointed out that this relationship may be useful for lightweight sheets of low absorption. Unfortunately, the

Optical and Appearance Properties

133

Fig. 28 Decrease in scattering coef®cient using black glass technique for increased levels of coating weight and absorption coef®cient. (From Ref. 18.)

transmitted light is dif®cult to measure accurately, as discussed earlier (Section III.F). B.

Particulate Scattering Approaches

Layer Calculations The lack of physical reality inherent in the Kubelka±Munk approach has led to re¯ectance analysis procedures that are more closely related to the manner in which light interacts with granular or ®brous structures [34,88]. Layerlike models for light interactions with granular powders have been proposed by Johnson [78] and Melamed [121]. For paper sheets, the approach of considering consecutive light interactions between layers (Stokes calculations) [136,175] has been demonstrated for two layers (see Fig. 25). The concept can be extended to any number of layers (Fig. 29). The layers may be different or identical; they may consist, for example, of separate paper sheets [159,160]. The principle is to interpret the optical properties of the stack in terms of single layer properties, that is, R0 ˆ Rn ˆ R1 ‡

Rn 1 T12 …1 R1 Rn 1 †

…26†

where n ˆ 2; 3; . . . when carrying out summations similar to those for Eq. (22). More recently both Scallan [156] and Olf [137,138] discussed the interrelationship between the Kubelka±Munk theory and Stokes calculations.

Fig. 29

Light interactions in n-layered structure. (From Ref. 157.)

134

Borch

Small Particle Scattering The scattering from single spherical particles of submicrometer size was rigorously calculated by Mie [17,122]. As shown in Fig. 30, the scattered intensity is strongest in the forward direction, around which it is centrosymmetrically distributed for spheres [153]. Computer programs are available for calculating the intensity distributions from spheres, coated spheres and cylinders [17]. Cylinder scattering is asymmetrical, as is the scattering from larger paper and textile ®bers [109]. This asymmetry has created efforts to develop light transmission methods as a means of deriving ®ber orientation in paper sheets, as described in Vol. 1, Chapter 16. The size range of paper and board constituents is such that scattering and absorption by small particles, that is, particle sizes comparable to the wavelength of irradiation, should be considered in particulate scattering approaches. This is particularly so for sheets heavily ®lled or coated with submicrometer sized pigments. Thus, based on an approach followed by Ross [152] for pigmented paints, Borch and Lepoutre [19] showed that the scattering coef®cient of paper coated with spherical plastic pigments is strongly dependent on both the size and the refractive index in the submicrometer range (Fig. 31). The particle size for maximum scattering is larger than half the radiation wavelength, in contrast to what is usually found experimentally for pigments of higher refractive index [116,186]. Similar calculations for coated spheres have been carried out to predict the effects of using coated pigments versus uncoated pigments in paper coating applications [79]. An alternative approach is to consider the scattering arising from microvoids between the closely packed spheres [2]. This approach may be advantageous for coatings where the void structure is affected by changes in packing of the pigment particles [2,3]. Sheet Simulation Because particulate scattering approaches do require a speci®c scattering structure it is tempting to use those for interpreting paper re¯ectance measurements in terms of physical scattering structure. Such analyses have been carried out by Scallan and Borch [21,157,158] and LeskelaÈ [101,102]. Attempts to correlate the Kubelka±Munk scattering coef®cient s with ®ber size or speci®c sheet surface area were described by Rennel [149]. Lacking a physical reality of the Kubelka±Munk scattering element (Fig. 24), such relationships are correlation variations mainly. In contrast, Scallan and Borch [157] assumed that a

Fig. 30

Mie scattering from submicrometer sphere. (From Ref. 153.)

Optical and Appearance Properties

135

Fig. 31 Calculated variation of the scattering coef®cient with spherical pigment size and refractive index ( ˆ 457 nm). (From Ref. 19.)

single sheet could be further subdivided into layers for which Eq. (26) would be valid and T1 ˆ

…1 …1

R1 ˆ r ‡

r†2 f r2 f 2 † r…1 …1

r†2 f 2 r2 f 2 †

…27† …28†

by summing up the incident light transmitted and re¯ected for the single layer shown in Fig. 32 where r is the average fraction of light re¯ected at each air/cellulose or cellulose/air surface and f is the average fraction transmitted at each passage through the layer of an average thickness that can be calculated from the speci®c surface area of the sheet, A0 . Consequently, the layer thickness is the cell wall thickness for a ``sheet'' made of uncollapsed and unbonded ®bers for which maximum re¯ectance would be obtained at any basis weight [158]. Layer calculations for single sheets are attractive in that the concept of layering often simulates what is seen in a cross-sectional view of a paper structure consisting

Fig. 32

Light interaction in a single layer. (From Ref. 157.)

136

Borch

of cellulosic ®bers only. The scattering structure constitutes well-de®ned micrometersized paper/void interfaces fairly well aligned with the sheet surfaces. Re¯ectance is due mainly to Fresnel scattering at relatively few interfaces. For sheets of no absorption (f ˆ 1), the Kubelka±Munk scattering coef®cient s may be calculated from the relationship [157] sˆ

 r  A 1 r 0

…29†

The effect of papermaking operations or ®ber coarseness on optical measurements is readily predicted [60,158]. Changes in mass density over the irradiated area can be modeled (Section V.D), and r and f can be adjusted to more closely simulate the scattering and absorption processes in the individual layers. For example, measurements of the re¯ectance levels of cellophane sheets dyed at different levels (decrease in f ) show that the sheet re¯ectivity (r in Eq. 28) also decreases at higher dye levels as shown in Fig. 33. This, at least qualitatively, explains the decrease in s values with increase in k values shown in Fig. 27 [91]. LeskelaÈ [101,102] extended layer calculations to consider variation in scattering particles as in sheets containing pigments or smaller submicrometer particles. His approach is to generate a layer structure as in Fig. 34 that consists of dissimilar layers of probability of bonding to each other. The optical properties are then calculated by simulating a Monte Carlo process for the number of particles, their order, and their packing. So far, their packing (number of voids) has shown agreement with the number of voids measured through image analysis on cross sections similar to that shown in Fig. 34 for paper sheets containing different particles (®bers, ®llers, etc.) [102]. Similar Monte Carlo simulations have been presented by Carlsson et al. [28,29]. These approaches show promise for de®ning the effect of changes in structure and composition on optical properties of paper consisting of more complex paper constituents.

Fig. 33 Re¯ectivity variation for a cellophane sheet dyed with increasing levels of dye. (From Koukoulas and Jordan [91].)

Optical and Appearance Properties

Fig. 34

C.

137

The concept of a sheet structure consisting of dissimilar layers. (From Ref. 102.)

Interpretation of Gloss

Specular Versus Diffuse Re¯ectance The interpretation of gloss and the in¯uence of gloss on other appearance properties are among the most severe challenges that face the papermaker within the area of appearance properties. The types of gloss described in Table 3 are based on visual perception rather than corresponding physically measurable properties [70]. Any sheet structure, including uncoated plain paper, will surface re¯ect light, and there is no accurate method by which surface-re¯ected light can be completely distinguished from light that has been internally re¯ected. Nevertheless, considerable attempts have been made to quantify paper re¯ectance in terms of specular and diffuse light components (elementary mirror concept) [7,59]. Barkas [7] originally proposed that surfaces showing low gloss can be modeled as containing small elementary facets that may be set at any angle to the mean surface. They re¯ect the light either diffusely or specularly according to Fresnel's theories [for example, Eqs. (14) and (15) in Section III.E]. At low gloss levels, adequate agreement with re¯ectance measurements is attained. Angular increases in peak re¯ectance position can be explained as being caused by random tilts in Fresnel re¯ecting surface parts that produce excess light at higher angles [52,84]. However, further analysis at high gloss levels produces irrational conclusions with regard to diffuse and specular light fractions as shown by Kurita et al. [96], who employed goniophotometry, and Tanaka [177,178], who used polarized light analysis at three re¯ectance angles. Therefore, this approach has had limited use, as discussed in detail by Harrison [59]. More recently, Koukoulas and Jordan [90] described gloss anisotropy for machine-made papers. For most papers studied, the gloss was biased in the machine direction for near normal and oblique (70 ) incident angles and in the cross-machine direction for angles between 40 and 70 . This was attributed to a larger angular distribution of re¯ected light from edges of ®bers aligned in the machine direction.

138

Borch

Surface Roughness Gate et al. [46] demonstrated that surface roughness in the submicrometer range correlates with TAPPI gloss. Light is surface-re¯ected in a manner similar to the scattering from submicrometer-size particles for which Fresnel re¯ection does not occur because of the relationship between size and wavelength of irradiation. For example, the theories applicable to scattering of IR radiation and microwaves and radiowaves from rough surfaces predict that re¯ectance in the specular direction is the sum of specular and diffuse components for which the roughness on a wavelength scale must be suf®ciently small to create a measurable specular re¯ectance [190]. This demonstrates why glossy papers are usually those that are coated or supercalendered to a roughness level appreciably below that of the ®ber structure itself. However, it is also apparent from the consideration of surface facet orientation described above that roughness size alone may not adequately characterize gloss levels. Thus, the use of image analysis techniques for describing both facet angle orientation and roughness indicates a more complex relationship between surface roughness and gloss [104]. Effect on Color Because routine color measurements are obtained with instrumentation of ®xed optical design, paper color measurements seldom re¯ect the appreciable variation in surface color with viewing angle for both glossy and matte sheets. Radiation diffusely scattered from the internal sheet structure re¯ects the object color of the sheet. Light specularly re¯ected from the sheet surface re¯ects the color of the light source [52,84,127]. The color shift with viewing angle has been described by Nishiwaki [129] and Gunji et al. [52,127]. For both colored glossy and colored dull papers, the tristimulus values increase for increasing incidence receiving angle [52]. In addition, this study illustrates a number of paper re¯ectance properties described earlier in this chapter: Light reversal does not produce identical color values; angle of maximum re¯ectance is not necessarily that of incidence; and paper smoothness has a profound effect on paper re¯ectance. Present color characterization is based on separate, well-de®ned measurements using concise instrumentation design, as shown above. In practical paper applications, illumination and viewing geometries are often not well controlled; there is a need for more sophisticated instrumentation design that is capable of simulating the visual combinations of specular and diffuse re¯ectance in a more universal manner. D.

Formation Theory

Paper Structure The effect of sheet formation is that re¯ected and transmitted light ¯uctuates from the irradiated sheet area because of ¯uctuations in paper structure (basis weight variation). Neither the Kubelka±Munk theory nor the layer approach as originally introduced by Scallan and Borch account for formation. The Kubelka±Munk scattering structure is a continuum (Fig. 23). The layer model does not show variation in basis weight because it contains n layers at any surface point (Fig. 28). Sheet ¯occulation, as described by Corte [36], is not accounted for. Both Borch and Scallan [21] and Jordan [81] considered the effect of formation on sheet opacity. A layer model exhibiting variation in basis weight is readily visualized by varying the number of layers within the irradiated surface area and summing up the re¯ectances from different surface points [21]. This is shown in Fig. 35. Sheets similar to those described in Fig. 27 were measured for re¯ectance, and the results

Optical and Appearance Properties

139

Fig. 35 Variations in opacity and re¯ectance factor for sheets made from red-dyed pulp. Increased absorption produces re¯ectance change as predicted by layer calculations modi®ed for sheet formation (shifts along constant layer thickness curves). (From Ref. 21.)

were compared to the theoretically de®ned variations in opacity and re¯ectance factor for a model of randomly distributed layers of constant re¯ectivity. It is apparent that agreement between calculated and measured values could now be obtained due to formation. However, this approach assumed diffuse re¯ectance unaffected by absorptance in contrast to the re¯ectivity decrease measured in Fig. 33 [91]. Wavelength Spectra Norman and Wahren [132] developed the concept of wavelength spectra to formation testing. Here, the sheet is considered to consist of ¯ocs of different sizes and weights. The spectral density represents the intensity of various ¯oc sizes (Fig. 36). For a given ¯oc in the sheet, its wavelength is equal to twice its geometrical size, and measured wavelength spectra can be compared to calculated distributions for a given scanning size [54]. An extensive account of the concept and interpretation of wavelength spectra and their applicability to formation measurements has been presented in connection with a general review of formation testing [131,133]. Komppa [87] discusses practical aspects of optical formation measurements and their limitation.

ACKNOWLEDGMENT I would like to express my thanks and appreciation to S. J. Popson, who offered valuable comments and criticism during the revision.

140

Borch

Fig. 36 Frequency and wavelength spectra concepts applied to paper formation. (From Ref. 54.)

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8.

9. 10. 11. 12. 13. 14. 15. 16. 17.

Adrian, A. P. (1944). New procedure for evaluating the opacifying properties of pigments. Tech. Assoc. Papers 27:465±471. AlincÆe, B., and Lepoutre, P. (1980). Light-scattering of coatings formed from polystyrene pigment particles. J. Colloid Interface Sci. 76(1):182±187. AlincÆe, B., and Lepoutre, P. (1980). Porosity and optical properties of clay coatings. J. Colloid Interface Sci. 76(2):439±444. Anders, C., and Ganz, E. (1979). Metameric white samples for testing the relative UV content of light sources and of natural daylight. J. Appl. Opt. 18(7):1067±1072. Anon. (1974). Proposal for study of color spaces and color difference equations. J. Opt. Soc. Am. 64(6):896±897. Arney, J. S., Morris, J., and Gammell, J. R. (1992). Image analysis and the characterization of the texture of paper. J. Imaging Sci. Technol. 36(1):99±103. Barkas, W. W. (1939). Analysis of light scattered from a surface of low gloss into its specular and diffuse components. Proc. Phys. Soc. (Lond.) 51:274±292. Bassemir, R. W., Castillo, G. D., DiBernardo, A. S., DiPiazza, J. T., Hade, M. E., Rybny, C. B., Saylor, S. D., and Zawacki, W. F. (1996). Color measurements of prints of varying hue and gloss using spectrophotometers with three different geometries. Tappi J. 79(4):149±156. Bauer, R. W. (1973). A simple method for calculation of light scattering coef®cients for coated papers and paper coatings. Tappi 56(11):113±116. Bernie, J.-P., and Murray Douglas, W. J. (1996). Local grammage distribution and formation of paper by light transmission image analysis. Tappi J. 79(1):193±202. Billmeyer, F. W., Jr. (1968). The signi®cance of recent CIE recommendations for color measurement. Color Eng. (Jan.±Feb.):34±38. Billmeyer, F. W., Jr. (1969). Current American practice in color measurement. Appl. Opt. 8(4):737±750. Billmeyer, F. W., Jr. (1969). Comparative performance of color-measuring instruments. Appl. Opt. 8(4):775±783. Billmeyer, F. W., Jr., Campbell, E. D., and Marcus, R. T. (1974). Comparative performance of color-measuring instruments: Second report. Appl. Opt. 13(6):1510± 1518. Billmeyer, F. W., Jr. and Marcus, R. T. (1969). Effect of illuminating and viewing geometry on the color coordinates of samples with various surface textures. Appl. Opt. 8(4):763±768. Billmeyer, F. W., Jr. and Smith, R. (1967). Optimized equations for MacAdam color differences. Color Eng. (Nov.±Dec.):28±29. Bohren, C. F., and Huffman, D. R. (1983). Absorption and Scattering of Light by Small Particles. Wiley, New York.

Optical and Appearance Properties 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38.

39. 40.

141

Borch, J. (1978). Optical properties of paper coatings. Paper presented at the 1978 Summer Conference in Polymer Science and Technology, New Paltz, NY, June 26±28. Borch, J., and Lepoutre, P. (1978). Light re¯ectance of spherical pigments in paper coatings. A comparison with theory. Tappi 61(2):45±48. Borch, J., and Scallan, A. M. (1975). A comparison between the optical properties of coated and loaded paper sheets. Tappi 58(2):135. Borch, J., and Scallan, A. M. (1976). An interpretation of paper re¯ectance based upon morphology. The effect of mass distribution. Tappi 59(10):102±105. Bristow, J. A. (1995). Optical properties of pulp and paper ± New standardization proposals. Tappi J. 78(12):45±47. Bristow, J. A. (1999). ISO brightness ± A more complete de®nition. Tappi J. 82(10):54±56. Bristow, J. A., and Karipidis, C. (1999). ISO brightness of ¯uorescent papers and indoor whiteness ± proposal for illuminant. Tappi J. 82(1):183±193. Bryntse, G., and Norman, B. (1976). A method to measure variations in surface and diffuse re¯ectance of printed and unprinted paper samples. Tappi 59(4):102±106. Budde, W., and Chapman, S. (1975). Measurement of brightness and opacity according to ISO-standards. CPPA Trans. Tech. Sect. 1(2):61±64. Burkhard, G., Wrist, P. E., and Mounce, G. R. (1960). A formation tester which graphically records paper structure. Pulp Paper Mag. Can. 61(6):T319±T333. Carlsson, J., Hellentin, P., Malmqvist, L., Persson, W., and Wahlstrom, C.-G. (1995). Time-resolved studies of light propagation in paper. Appl. Opt. 34(9):1528±1535. Carlsson, J., Persson, W., Hellentin, P., and Malmqvist, L. (1995). The propagation of light in paper: Modelling and Monte Carlo simulations. 1995 International Paper Physics Conference, Niagara Falls, pp. 83±86. Casey, J. P. (1961). Pulp and Paper Chemistry and Chemical Technology, Paper Testing and Converting. Vol. 3. 2nd ed. Interscience, New York, pp. 1362±1434. Chickering, K. D. (1967). Optimization of the MacAdam-modi®ed 1965 Friele colordifference formula. J. Opt. Soc. Am. 57(4):537±541. Christie, J. S. (1977). On-machine measurement of the chromatic aspects of appearance. Tappi 60(2):119±122. Clark, H. B., and Ramsay, H. L. (1965). Predicting optical properties of coated papers. Tappi 48(11):609±612. Companion, A. L. (1965). Theory and applications of diffuse re¯ectance spectroscopy. In Developments in Applied Spectroscopy, vol. 4. E. N. Davis, ed. Plenum, New York, pp. 221±234. Corte, H. (1976). Perception of the optical properties of paper. In The Fundamental Properties of Paper Related to Its Uses. F. Bolam, ed. British Paper and Board Industry Federation, London, pp. 626±658. Corte, H. (1982). The structure of paper. In The Structure and Physical Properties of Paper. H. F. Rance, ed. Elsevier, Amsterdam, pp. 262±276. Dearth, L. R. (1963). Numerical evaluation of color employing suitable instrumentation. Tappi 46(10):146A±151A. Dearth, L. R., Shillcox, W. M., and Van den Akker, J. A. (1960). Instrumentation studies. LXXXV. A study of photoelectric instruments for the measurement of color: Re¯ectance and transmittance. XII. The Elrepho (photoelectric re¯ectance photometer). Tappi 43(2):230A±239A. Emerton, H. W., Page, D. H., and Hale, W. H. (1962). Structure of papers as seen in their surfaces. In: The Formation and Structure of Paper. F. Bolam, ed. British Paper and Board Industry Federation, London, pp. 53±99. Foote, W. J. (1939). An investigation of the optical scattering and absorption coef®cients of dyed handsheets and the application of the I.C.I. system of color speci®cation to these handsheets. Paper Trade J. 108(10):TS125±TS132.

142

Borch

41. Foote, W. J. (1939). An investigation of the fundamental scattering and absorption coef®cients of dyed handsheets. Paper Trade J. 109(25):TS333±TS340. 42. Ganz, E. (1976). Whiteness: Photometric speci®cation and colorimetric evaluation. Appl. Opt. 15(9):2039±2058. 43. Ganz, E. (1979). Whiteness perception: Individual differences and common trends. J. Appl. Opt. 18(17):2963±2970. 44. Gantz, E., and Pauli, H. K. A. (1995). Whiteness and tint formulas of the Commission Internationale de l'Eclairage: Approximations in the L*a*b* color space. Appl. Opt. 34(16):2998±2999. 45. Gartaganis, P. A., Heintze, H. U., and Gordon, R. W. (1976). From printograph to videoscanner. Tappi 59(12):113±117. 46. Gate, L., Windle, W., and Hine, H. (1973). The relationship between gloss and surface microtexture of coatings. Tappi 56(3):61±65. 47. Griesser, R. (1993). Whiteness not brightness: A new way of measuring and controlling production of paper. Appita 46(6):439±444. 48. Griesser, R. (1996). CIE whiteness and tint: possible improvements. Appita 49(2):105± 112. 49. Grum, F., and Costa, L. (1977). Color evaluation by ¯uorescence measurement without the need for multiple illumination sources. Tappi 60(8):119±121. 50. Grum, F., and Patek, J. M. (1965). Evaluation of whiteness using relative spectral radiance measurements. Tappi 48(6):357±362. 51. Grum, F., and Wightman, T. E. (1977). Absolute re¯ectance of Eastman white re¯ectance standard. Appl. Opt. 16(11):2775±2776. 52. Gunji, T., Nihira, K., and Tsuboi, T. (1973). Goniophotometric color measurement for papers. J. Soc. Fiber Sci. Technol. (Jpn.) (Sen-i Gakkaishi) 28(11):462±469 (in Japanese). 53. Haddad, S. F. (1967). Measurement of optical brightener in paper. Tappi 50(12):90A± 91A. 54. Haglund, L., Norman, B., and Wahren, D. (1974). Mass distribution in random sheets: Theoretical evaluation and comparison with real sheets. Svensk Papperstidn. 77(10):362±370. 55. Hall, R. J. (1976). Instrumental quanti®cation of subjective formation. In The Fundamental Properties of Paper Related to Its Uses. F. Bolam, ed. British Paper and Board Industry Federation, London, pp. 662±677. 56. Hammond, H. K., and Hsia, J. J. (1975). Evaluation of instrument tolerances for 75 gloss. Tappi 58(11):143±144. 57. Hardacker, K. W. (1989). An optical transmission meter for paperboard-weight papers. Tappi J. 72(12):168±172. 58. Harrison, V. G. W. (1940). Measurement of opacity of paper: A theoretical survey. Proc. Tech. Sect. British Paper and Board Makers Assoc. 21:67±173. 59. Harrison, V. G. W. (1962). Optical properties of paper. In The Formation and Structure of Paper. F. Bolam, ed. British Paper and Board Industry Federation, London, pp. 467± 485. 60. Hasuike, M., Shingai, M., Murakami, K., and Imamura, R. (1979). Evaluation of the optical properties of thin- and thick-walled ®ber sheets. J. Jpn. Wood Res. Soc. (Mokuzai Gakkaishi) 25(2):132±138 (in Japanese). 61. Hemstock, G. A. (1962). The effect of clays upon the optical properties of paper. Tappi 45(2):158A±159A. 62. Hillend, W. J. (1966). Opacity problems in printing papers: Kubelka±Munk theory gives good, quick answers. Tappi 49(7):41A±47A. 63. Hoban, R. F. (1974). Fluorescence measurements by the two-mode method. Tappi 57(7):97±100.

Optical and Appearance Properties 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90.

143

Hopkins, L. F. (1969). The standardization of color measurement of paper. Printing Technol. 13(3):114±122. Huey, S. J. (1972). Scienti®c approach to visual color comparison. J. Paint Technol. 44(573):83±89. Huey, S. J. (1972). Standard practices for visual examination of small color differences. J. Color Appearance 1(4):24±26. Hughes, D. A. (1962). Methods of obtaining the optical properties of papers containing titanium dioxide and mixtures of titanium dioxide and other ®llers. Tappi 45(2):159A± 163A. Humphrey, K. (1995). Image Analysis. Pira Int. Leatherhead, England. Hunt, R. G. W. (1991). Measuring Colour. 2nd ed. Ellis Horwood, Chichester, England. Hunter, R. S. (1952). Gloss evaluation of materials. ASTM Bull. 186:48±55. Hunter, R. S. (1958). Standardization of test for specular gloss of paper at 75 TAPPI Standard T 480 m-51. Tappi 41(8):385±396. Hunter, R. S. (1968). High resolution goniophotometer and its use to measure appearance properties and light-scattering phenomena. In: Modern Aspects of Re¯ectance Spectroscopy. W. Wendlandt, ed. Plenum, New York, pp. 226±241. Hunter, R. S. (1975). The Measurement of Appearance. Wiley, New York. Hunter, R. S., and Harold, R. W. (1987). The Measurement of Appearance, 2nd ed. Wiley, New York. I'Anson, S. J., and Kropholler, H. W. (1991). Enhancing visibility of wire-mark by image analysis. J. Pulp Paper Sci. 17(1):J22±J26. Ingersoll, L. R. (1921). The glarimeter. J. Opt. Soc. Am. 5(3):213±217. Johnson, J. W. D. (1979). Newsprint properties: Opacity, percent saturation, dominant wavelength, and scattering coef®cient. Tappi 62(12):115±116. Johnson, P. D. (1952). Absolute optical absorption from diffuse re¯ectance. J. Opt. Soc. Am. 42(12):978±981. Johnson, R. W., Thiele, E. S., and French, R. H. (1997). Light scattering ef®ciency of white pigments: An analysis of model core-shell pigment vs. optimized rutile TiO2 . Tappi J. 80(11):233±238. Jones, C. L., and Budinger, A. B. (1965). Equipment and techniques for measurement and characterization of paper re¯ectance variation. Tappi 48(1):54±59. Jordan, B. D. (1985). Predicting the effect of formation on opacity and scattering coef®cient. J. Pulp Paper Sci. 11(2):J56±J59. Jordan, B. D., and O'Neill, M. A. (1987). The case for switching to CIE L*a*b color description for paper. Pulp Paper Can. 88(10):T382±T386. Jordan, B. D., and O'Neill, M. A. (1991). The whiteness of paper ± Colorimetry and visual ranking. Tappi J. 74(5):93±101. Judd, D. B., and Wyszecki, G. (1975). Color in Business, Science, and Industry, 3rd ed. Wiley, New York. Kigle-Boeckler, G. (1996). Measuring gloss and re¯ection properties of surfaces. Tappi J. 79(9):194±198. Knox, J. M., and Wahren, D. (1984). Determination of light scattering coef®cient of dark and heavy sheets. Tappi J. 67(8):82±85. Komppa, A. (1988). Measurement of formation. Paperi Puu (Paper Timber) 70(3):243± 250. KortuÈm, G. (1969). Re¯ectance Spectroscopy. Springer-Verlag, Berlin Koukoulas, A. A., and Jordan, B. D. (1994). Understanding gloss anisotropy in machine-made papers. J. Pulp Paper Sci. 20(2):J55±J60. Koukoulas, A. A., and Jordan, B. D. (1994). MD/CD re¯ectance anisotropy in machine-made papers observed with the 45/0 re¯ectance geometry. J. Pulp Paper Sci. 20(3):J77±J82.

144

Borch

91. Koukoulas, A. A., and Jordan, B. D. (1997). Effect of strong absorption on the Kubelka-Munk scattering coef®cient. J. Pulp Paper Sci. 23(5):J224±J231. 92. Kubelka, P. (1948). New contributions to the optics of intensely light-scattering materials. Part I. J. Opt. Soc. Am. 38(5):448±457, 1067. 93. Kubelka, P. (1954). New contributions to the optics of intensely light-scattering materials. Part II: Non-homogeneous layers. J. Opt. Soc. Am. 44(4):330±335. 94. Kubelka, P., and Munk, F. (1931). A contribution to the optics of colorant layers. Z. Tech. Physik 12(11a):593±601 (in German). 95. Kuehni, R. G. (1976). Color-tolerance data and the tentative CIE 1976 L*a*b* formula. J. Opt. Soc. Am. 66(5):497±500. 96. Kurita, T., Yano, H., Nara, J., and Hasunuma, H. (1955). Barkas' analysis for the light scattered by surface of paper. J. Appl. Phys. (Jpn.) 24(8):318±325 (in Japanese). 97. Lathrop, A. L. (1964). A Kubelka-Munk calculation chart for light-scattering materials. Tappi 47(12):789±791. 98. Laverie, H. (1985). On-line color measurement improvements increase papermaking applications. Paper Trade J. 169(4):53±54. 99. Leekley, R. M., Denzer, C. W., and Tyler, R. F. (1970). Measurement of surface re¯ection from papers and prints. Tappi 53(4):615±621. 100. Leekley, R. M., and Tyler, R. F. (1975). The measurement of optical unevenness. Tappi 58(3):124±127. 101. LeskelaÈ, M. (1993). A model for the optical properties of paper. Part 1. The theory. Paperi Puu (Paper Timber) 75(8):683±688. 102. LeskelaÈ, M. (1994). A model for the optical properties of paper. Part 2. The simulation model. Paperi Puu (Paper Timber) 76(1±2):67±73. 103. LeskelaÈ, M. J. (1995). Optical calculations for multilayer papers. Tappi J. 78(10):167±172. 104. Lipshitz, H., Bridger, M., and Derman, G. (1990). On the relationship between topography and gloss. Tappi J. 73(10):237±245. 105. Lodzinski, F. P. (1973). Experience with a transmittance-type online opacimeter for monitoring and controlling opacity. Tappi 56(2):78±82. 106. Loof, H. (1967). De®nition, measurement and valuation of the white color of paper. Das Papier 21(6):297±309 (in German). 107. Loyalka, S. K., and Riggs, C. A. (1995). Inverse problem in diffuse re¯ectance spectroscopy: Accuracy of the Kubelka-Munk equations. Appl. Spectrosc. 49(8):1107±1110. 108. Luey, A. T. (1971). Predicting the brightness of clay-coated boxboard using the opacity factor. Tappi 54(2):252±254. 109. Lynch, L. J., and Thomas, N. (1971). Optical diffraction pro®les of single ®bers. Textile Res. J. 41:568±572. 110. Maas, O. (1936). Opacity measurements. Pulp Paper Mag. Can. 37(11):689±695. 111. MacAdam, D. L. (1955). Color measurement and whiteness. Tappi 38(2):78±87. 112. MacGregor, M. A., and Johansson, P.-AÊ. (1990). Submillimeter gloss variations in coated paper. Part 1: The gloss imaging equipment and analytical techniques. Tappi J. 73(12):161±168. 113. Malthouse, D. D., and Popson, S. J. (1995). A new way of measuring the effect of ¯uorescent whitening agents. Appita 48(1):56±58. 114. Mangin, P. J., and Lyne, M. B. (1979). Color analysis of newsprint in the dominant wavelength range 570±585 nm. Tappi 62(5):94±96. 115. Mangin, P. J., and Lyne, M. B. (1979). Color analysis of newsprint: Hunter L a b and L*a*b* (CIELAB). Program for SR-52 or TI-59 calculators. Tappi 62(8):129±130. 116. Marchetti, F. R., and Willets, W. R. (1960). The optical ef®ciency of titanium pigments in paper coatings. Tappi 43(3):273±276. 117. Marchi, R. (1986). Latest developments allow online color and shade measure, control. Pulp Paper 171(11):124±127.

Optical and Appearance Properties 118. 119. 120. 121. 122. 123. 124. 125. 126. 127. 128. 129. 130.

131. 132. 133. 134. 135. 136. 137. 138. 139. 140.

145

Marsh, R. S. (1965). Instrumental color speci®cations as they pertain to the paper manufacturer. Tappi 48(2):59A. McCamy, C. S. (1992). Munsell value as explicit functions of CIE luminance factor. Color Res. Appl. 17(3):205±207. McDonald, J. D., Farrell, W. R., Stevens, R. K., Hussain, S. M., and Roche, A. A. (1986). Using an on-line light transmittance gauge to identify the source of grammage variations. J. Pulp Paper Sci. 12(1):J1±J9. Melamed, N. T. (1963). Optical proeprties of powders. 1. Optical absorption coef®cients and the absolute value of the diffuse re¯ectance. 2. Properties of luminescent powders. J. Appl. Phys. 34(3):560±570. Mie, G. (1908). Contributions to the optics of diffuse media. Ann. Phys. 25:377±445 (in German). Mosher, C. (1978). Optical properties of paper. Calculated from single-sheet re¯ectance data R0 , R, and Rg . Tappi 61(10):123±125. Munsell, A. H. (1905). A Color Notation. 11th ed. Ellis, Boston, and Munsell Color Co., Baltimore, 1961. Nara, J. (1957). Gloss meter by ``distinctness-of-image.'' J. Appl. Phys. (Jpn.) 26(9):452±454 (in Japanese). Nickerson, D. (1977). History of the OSA committee on uniform color scales. Optics News (Winter):8±17. Nihira, K., Gunji, T.,a nd Tsuboi, T. (1976). Study of goniophotometric curve by color measurement for papers. J. Soc. Fiber Sci. Technol. (Jpn.) (Sen-i Gukkaishi) 32(1):T23±T31 (in Japanese). Nishiwaki, J. (1957). Gloss of machine glazed papers. J. Phys. Soc. (Jpn.) 12(1):53±57. Nishiwaki, J. (1959). Effect of color on visual gloss. (Part 1 and 2). J. Appl. Phys. (Jpn.) 28(5):267±276. Nordman, L., Aaltonen, P., and Makkonen, T. (1966). Relationships between mechanical and optical properties of paper affected by web consolidation. In Consolidation of the Paper Web. F. Bolam, ed. The British Paper and Board Industry Federation, London, pp. 909±927. Norman, B. (1986). The formation of paper sheets. In: Paper Structure and Properties. J. A. Bristow and P. Kolseth, eds. Marcel Dekker, Inc., New York, Chapter 6. Norman, B., and Wahren, D. (1972). A comprehensive method for the description of mass distribution in sheets and ¯occulation and turbulence in suspensions. Svensk Papperstidn. 75(20):807±818. Norman, B., and Wahren, D. (1976). Mass distribution and sheet properties of paper. In: The Fundamental Properties of Paper Related to Its Uses. F. Bolam, ed. British Paper and Board Industry Federation, London, pp. 7±70. Oittinen, P. (1993). Characterization of paper as an information carrier. In: Products of Papermaking. C. F. Baker, ed. Pira Int., Leatherhead, UK, pp. 311±349. Olf, H. G. (1980). Printing opacity as a function of basis weight. Tappi 63(9): 149±150. Olf, H. G. (1988). Stokes's pile of plates revisited. J. Opt. Soc. Am. A5(10): 1620±1625. Olf, H. G. (1989). Correspondences between the Kubelka±Munk and the Stokes model of strongly light-scattering materials. Part 1: Theory. Tappi J. 72(5):222±227. Olf, H. G. (1989). Correspondence between the Kubelka±Munk and the Stokes model of strongly light-scattering materials. Part II: Implications. Tappi J. 72(7):159±163. Ostromecki, R. R. (1978). Optical properties of paper. Calculated from multisheet TAPPI brightness and opacity. Tappi 61(10):123±125. Parkes, D. (1989). Instrumental methods for evaluating whiteness, brightness, and ¯uorescence. Tappi J. 69(9):95±100.

146

Borch

141. Pauler, N. (1986). Opacity and re¯ectivity of multilayer structures. In: Paper Structure and Properties. J. A. Bristow and P. Kolseth, eds. Marcel Dekker, Inc., New York, Chapter 10. 142. Peltz, G., Heiszler, L., and Hilke, E. (1973). Optical properties of optical-characterrecognition paper for electronic data processing. Das Papier 27(6):217±224 (in German). 143. Popson, S. J. (1974). On-line measurement of paper color, brightness, and opacity. Paper Trade J. 158(31):24±27. 144. Popson, S. J., Malthouse, D. D., and Robertson, P. C. (1997). Applying brightness, whiteness and color measurements to color removal. Tappi J. 80(9):137±147. 145. Popson, S. J., Malthouse, D. D., Crawford, T. B., Popson, M. T., and Crawford, P. M. (1996). Measurement and Control of the Optical Properties of Paper. Technidyne Corp., New Albany, IN. 146. Poulter, S. R. C. (1968). Measurement of print unevenness. Tappi 51(8):87A±91A. 147. Presgrave, J. E. (1975). On-line control of ¯uorescent white papers. Paper Technol. 16(1):34±38, 42±43. 148. Ramsay, H. L. (1966). Simpli®ed calculation for predicting optical properties of coated board. Tappi 49(12):116A±118A. 149. Rennel, J. (1969). Opacity in relation to strength properties of pulps. III. Light-scattering coef®cient of sheets of model ®bers. Tappi 52(10):1943±1947. 150. Robinson, J. V. (1975). A summary of re¯ectance equations for applications of the Kubelka±Munk theory to optical properties of paper. Tappi 58(10):152±153. 151. Robinson, J. V., and Linke, E. G. (1963). Theory of the opacity of ®lms of coating pigment and adhesive: A method for calculating the opacity of coatings. Tappi 46(6):384±390. 152. Ross, W. D. (1971). Theoretical computation of light scattering power: Comparison between TiO2 and air bubbles. J. Paint Technol. 43(563):50±66. 153. Ross, W. D. (1974). Theoretical light-scattering power of TiO2 and microvoids. Ind. Eng. Chem., Prod. Res. Dev. 13(1):45±49. 154. Ruckdeschel, F. R. (1979). Light scattering and spectral properties of white cast-coated enamel paper. Tappi 62(1):61±64. 155. Rundlof, M., and Bristow, J. A. (1997). A note concerning the interaction between light scattering and light absorption in the application of the Kubelka-Munk equations. J. Pulp Paper Sci. 23(5):J220±J223. 156. Scallan, A. M. (1985). An alternative approach to the Kubelka-Munk theory. J. Pulp Paper Sci. 11(3):J80±J84. 157. Scallan, A. M., and Borch, J. (1972). An interpretation of paper re¯ectance based upon morphology. I. Initial considerations. Tappi 55(4):583±588. 158. Scallan, A. M., and Borch, J. (1976). Fundamental parameters affecting the opacity and brightness of uncoated paper. In: The Fundamental Properties of Paper Related to Its Uses. F. Bolam, ed. British Paper and Board Industry Federation, London, pp. 152± 163. 159. Schmidt, G. (1959). Transparency and re¯ectance of paper. Theory and method of measurement. Das Papier 13(7/8):141±149 (in German). 160. Schmidt, G. (1960). The paper stack theory. A simpli®ed description of opacity and transparency of paper. Das Papier 14(10):445±452 (in German). 161. Scott, W. E., Dearth, L. R., and Jordan, B. (1992). Optical properties of paper. In: Pulp and Paper Manufacture. Vol. 9, Mill Control & Control Systems: Quality & Testing, Environmental, Corrosion, Electrical. M. Kouris, ed. TAPPI Press, Atlanta, pp. 152±191. 162. Shiratori, N., Ishimura, H., Yoshidomi, N., Taniguchi, M., and Nakahara, Y. (1993). Quantitative evaluation of submillimeter gloss variations in coated papers. Tappi J. 76(10):89±97.

Optical and Appearance Properties 163. 164. 165. 166. 167. 168. 169. 170. 171. 172. 173. 174. 175. 176. 177. 178. 179. 180. 181. 182. 183. 184. 185. 186. 187.

147

Springer, G. (1971). A light transmission type on-line opacity meter. Tappi 54(3):411±412. Starr, R. E., and Young, R. H. (1975). An improvement in the determination of R1 and scattering coef®cients for paper, pigments, and coatings. Tappi 58(5):75±78. Starr, R. E., and Young, R. H. (1978). Paper coating formulations. A study of limitations involved in the determination and use of the Kubelka±Munk constants. Tappi 61(6):78±80. Steele, F. A. (1935). The optical characteristics of paper. 1. The mathematical relationships between basis weight, re¯ectance, contrast ratio, and other optical properties. Paper Trade J. 100(12):37±42. Stenius, AÊ. S. (1951). The application of the Kubelka-Munk theory to the diffuse re¯ection of light from paper. I and II. A critical study. Svensk Papperstidn. 54(19):663±709. Stenius, AÊ. S. (1965). SCAN-test brightness measuring system. Tappi 48(12):45A±52A. Stenius, AÊ. S. (1972). Fluorescence in paper. J. Color Appearance 1(6):8±10. Stenius, AÊ. S. (1979). Multi- or single-sheet light scattering and absorption coef®cients. Tappi 62(1):89±91. Stenius, AÊ. S. (1979). Tristimulus values and chromaticity coordinates from spectral re¯ectance values. Program for HP-97 calculator. Tappi 62(11):123. Stenius, AÊ. S., Kyrklund, B., and LoraÊs, V. (1966). Color: Its graphical representation, its measurement and evaluation. A review of the three CIE color co-ordinate systems. Svensk Papperstidn. 69(5):150±158. Stenius, AÊ. S., Rydberg, J., and SoÈderhjelm, L. (1975). ISO brightness: The new brightness value. Svensk Papperstidn. 78(11):403±408. Stewart, D., Scharf, R. A., and Arney, J. S. (1995). Techniques for digital image capture of watermarks. J. Imaging Sci. Technol. 39(3):261±267. Stokes, G. G. (1862). On the intensity of the light re¯ected from or transmitted through a pile of plates. Proc. Roy. Soc. (London) 11:454±556. Sundstrom, F. O., and Stearns, E. I. (1950). Practical art of color matching on paper. Paper Mill News (July 1):12±15. Tanaka, S. (1956). Measurement of re¯ection characteristics of paper using polarized light. J. Appl. Phys. (Jpn.) 25(5):207±213 (in Japanese). Tanaka, S. (1958). Measurement of re¯ection characteristics of paper using polarized light (II). J. Appl. Phys. (Jpn.) 27(10):600±604 (in Japanese). Trader, C. D. (1971). Laboratory studies relating coating structure and coating performance. Tappi 54(10):1709±1713. Trepanier, R. J. (1989). User-friendly system analyzes paper formation, dirt speck content, and solid-print nonuniformity. Tappi J. 72(12):153±157. Trosset, S. W., Jr. (1966). A method of storage for maintaining brightness and color of white standard samples. Tappi 49(4):61A±65A. Van den Akker, J. A. (1963). Theory of the optical properties. In: The Bleaching of Pulp. W. H. Rapson, ed. TAPPI Monograph No. 27. TAPPI Press, Atlanta, GA, pp. 17±39. Van den Akker, J. A. (1965). Developments in spectrophotometry and papermaking. Tappi 48(2):57A. Van den Akker, J. A. (1965). Standard brightness, color, and spectrophotometry with emphasis on recent information. Tappi 48(12):57A±62A. Van den Akker, J. A. (1968). Theory of some of the discrepancies observed in application of the Kubelka-Munk equations to particulate systems. In: Modern Aspects of Re¯ectance Spectroscopy. W. Wendlandt, ed. Plenum, New York, pp. 27±46. Van den Akker, J. A. (1977). Optical aspects of coating pigments. In: Physical Chemistry of Pigments in Paper Coatings. C. L. Garey, ed. TAPPI Press, Atlanta, GA. Van den Akker, J. A. (1982). Optical Properties of Paper. In: The Structure and Physical Properties of Paper. H. F. Rance, ed. Elsevier, Amsterdam, pp. 127±174.

148

Borch

188. Van den Akker, J. A., and Sears, G. R. (1964). Gloss of coated paper and paperboard. Tappi 47(11):179A±182A. 189. Wahren, D., and Bryntse, G. (1976). An improved model of the re¯ectance properties of uneven solid prints. In: The Fundamental Properties of Paper Related to Its Uses. F. Bolam, ed. British Paper and Board Industry Federation, London, pp. 616±621. 190. Warren, C. A., and Peel, J. D. (1973). The scattering of infrared radiation as a method of measuring paper roughness: Part 1. Paper Technol. 14(4):91±97. 191. Wickstrom, W. A., and Horner, M. (1970). Closed-loop color control for printing paper. Tappi 53(5):784±791. 192. Williams, D. J. (1970). The measurement of ®bre orientation and its relation to directional strength properties. Appita 24(3):196±200. 193. Wink, W. A., Delevanti, C. H., Jr., and Van den Akker, J. A. (1953). Instrumentation studies. LXXVII. Study on Gloss. I. A goniophotometric study of high-gloss papers. Tappi 36(12):163A±172A. 194. Young, R. H., and Drexel, R. J. (1980). Another approach to CIELAB coordinates. Tappi 63(7):121±122. 195. Yule, J. A. C., Howe, D. J., and Altman, J. H. (1967). The effect of the spread-function of paper on halftone reproduction. Tappi 50(7):337±344. 196. Zwinkels, J. C., Gignac, D. S., Nevius, M., Powell, I., and Bewsher, A. (1997). Design and testing of a two-monochromator reference spectro¯uorimeter for high-accuracy total radiance factor measurement. Appl. Opt. 36(4):892±902.

5 MICROSCOPY PATRICIA A. MOSS* The University of Manchester Institute of Science and Technology Manchester, England LES GROOM Southern Research Station U.S. Forest Service Pineville, Louisiana

I.



Introduction

150

II.

A Brief History of the Microscopy of Wood and Paper

151

III.

Microscopical Techniques for Pulp and Paper Research A. Optical Microscopes B. Electron Microscopes C. Scanning Probe Microscopes

152 152 159 167

IV.

The A. B. C. D. E. F. G.

Confocal Laser Scanning Microscope Basic Principles of the CLSM Imaging Modes Specimen Preparation Techniques Fluorescence Quanti®cation Image Acquisition and Processing Artifacts of CLSM Imaging Applications of the CLSM to Pulp and Paper Research

168 168 174 175 178 178 180 185

V.

The A. B. C. D. E. F.

Low Temperature Scanning Electron Microscope The Cryosystem Specimen Preparation Comparative Evaluation of Preparation Techniques for LTSEM Advantages and Disadvantages of Cryopreparation Techniques Artifacts of Cryopreparation Applications of the LTSEM to Pulp and Paper Research

203 204 206 208 208 209 211

Current af®liation: Oy KeskuslaboratorioÐCentrallaboratorium Ab (KCL), Espoo, Finland. 149

150

Moss and Groom

VI.

VII.

I.

The A. B. C. D.

Atomic Force Microscope Principles and Operation of Atomic Force Microscopy Operating Modes Environmental Conditions Applications of the AFM to Pulp and Paper Research

216 217 220 223 224

Conclusions

247

Abbreviations

247

References

248

INTRODUCTION

Microscopy is the study and interpretation of images produced by a microscope. ``Interpretation'' is the key word, because the microscope enables one to see structures that are too small, or too close together, to be resolved by the unaided eye. (The human eye cannot separate two points or lines that are closer together than 0.1 mm.) It is important to remember that microscopy is not simply a matter of magni®cation and making objects larger, but of resolving features that may not have been seen before. Much microscopical analysis is subjective. Images cannot be interpreted intuitively on the basis of nonmicroscopical experience but require highly skilled microscopists with knowledge and practical experience of the materials they are examining and an ``eye'' for their subject. Correct interpretation can be achieved only when one has a thorough understanding of all the factors that in¯uence the ®nal image. These factors include instrumental effects, specimen preparation techniques, and microscope±specimen interactions. Three-dimensional objects projected onto a two-dimensional image, such as is the case with scanning electronmicrographs, can be particularly confusing and dif®cult to interpret. Images of surface features can be illusory, and the observation, recorded by the seventeenth century microscopist Robert Hooke [128], that ``it is exceedingly dif®cult in some objects to distinguish between a prominence and a depression, between a shadow and a black stain, or a re¯ection and a whiteness in a color'' still holds true today. The aim of microscopy is to study material in a condition as near as possible to its natural state. This, therefore, means a minimum of specimen processing because each treatment that is applied runs the risk of introducing artifacts. As Stone and Scallan [309] pointed out, ``The structural studies of water-swollen ®bers should always be performed, if possible, while the ®bers are still saturated with water.'' This chapter discusses the applications of microscopy to pulp and paper research. Conventional optical and electron microscopical techniques are covered only brie¯y, because these are well documented elsewhere. Novel forms of microscopyÐconfocal laser scanning microscopy (CLSM), low temperature scanning electron microscopy (LTSEM), and atomic force microscopy (AFM)Ðare presented in detail. As yet, there are no of®cial or standard methods pertaining to their use.

Microscopy

II.

151

A BRIEF HISTORY OF THE MICROSCOPY OF WOOD AND PAPER

Robert Hooke was one of the ®rst people to use a microscope to examine the structure of wood cells. He published his drawings in Micrographia in 1665 [128]. His observations raised many questions concerning the ultrastructure and functions of wood cells, but little advance was made until the second half of the nineteenth century, when the polarizing microscope was developed. This made possible the examination of crystalline structures and bundles of parallel ®laments, and Carl von NaÈgeli laid the foundations of the physical chemistry of colloidal matter, which must also include the cell walls of wood. Other advances included the development of ultraviolet and ¯uorescence microscopical techniques. KoÈhler developed the ®rst ultraviolet microscope in 1904, and Reichert and Heimstadt demonstrated the ®rst practical ¯uorescence microscope in 1911. Both instruments have proved to be very useful tools for studying plant material, which generally exhibits relatively intense primary ¯uorescence in the ultraviolet (UV) and blue spectra. The discovery early in the twentieth century of electrons and an understanding of their behavior led to the development of electron microscopes and the opening up of a whole new ®eld. High energy particles, which have wavelike properties analogous to those of light waves could be generated, so the wavelength of light was no longer a limiting factor to the resolution that could be achieved with a microscope. Ernst Ruska [280] produced the ®rst electron micrographs of surfaces. The development of the transmission electron microscope presented enormous scope for ultrastructural research because of its very high resolution. The micro®brillar structure of cell walls could be clearly seen, and measurements could be made of the ®brillar units. The scanning electron microscope, on the other hand, allowed larger specimens to be examined, but its main advantage was its enormous depth of ®eld, enabling three-dimensional viewing of specimens. In 1959 the ®rst SEM to be used speci®cally for solving problems associated with papermaking was installed at the Pulp and Paper Research Institute of Canada in Montreal [301]. Nowadays the SEM is widely used in paper science research and is a standard piece of equipment in most research centers. The ¯ying spot microscope, described by Young and Roberts in 1951 [371], was the ®rst example of a sequential imaging system. The confocal scanning microscope was invented by Minsky [211], who ®led a patent for it in 1957, but further development took a long time. Forerunners of the modern confocal laser scanning microscope (CLSM) used UV or white light, but the later development of the laser, giving a powerful source of monochromatic radiation that could be concentrated into one spot, provided greater versatility. Commercial production of the CLSM in the late 1980s led to a renaissance in optical microscopy, particularly in ¯uorescence microscopy. Its use quickly became a widely established technique in cell biology and medical research and is now indispensable for studying the threedimensional structure of cells as well as dynamic processes in living cells. Nanko and Ohsawa [224] were the ®rst to report on the use of a scanning laser microscope (SLM) for examination of paper, using it to observe changes that occurred in wet paper webs as they dried. The CLSM has become a recognized part of paper science research, with more and more centers acquiring their own microscope or gaining access to one.

152

Moss and Groom

A new family of scanning probe microscopes was developed in the 1980s that offered the possibility of atomic resolution. These include the scanning tunneling microscope (STM) invented by Binnig and Rohrer in 1982 [21, 22] and the atomic force microscope (AFM) subsequently developed by Binnig and coworkers in 1985 [20]. III.

MICROSCOPICAL TECHNIQUES FOR PULP AND PAPER RESEARCH

Microscopes are powerful research tools, and both optical and electron microscopes have played an important role in elucidating mechanisms of papermaking processes. Direct observations are very important to gaining an understanding of the structure of ®bersÐboth of their behavior in aqueous solutions and of their properties in the resultant paper. Direct visual methods need to be complemented with carefully planned measurements of physical properties carried out simultaneously so that causal relationships can be established. Microscopy needs to be applied systematically, with quantitative analysis of images undertaken. To this end most microscopes can be linked directly to an image analysis system and images can be processed and measured interactively. Image quality for all microscopy depends on optimizing the setup and adjustment of the microscope and on good specimen preparation. Good image quality results in high contrast, high resolution images that are free of artifacts. Finally, calibration of magni®cation is essential for interpretation. Diffraction gratings and latex spheres are commonly used for calibration purposes, but Watts and Emerton [346] recommend imaging the periodic structure of a diatom skeleton. This is particularly suitable for electron microscopy, because, being silica, the diatom skeleton is not subject to shrinkage. So, what do these different methods of microscopy have to offer? A.

Optical Microscopes

Optical microscopy has always been, and will continue to be, an important part of paper science research. It has been much used for qualitative analysis but also has a large range of quantitative applications. However, the resolution of optical microscopes (around 0.2 mm) is poor compared to that of electron microscopes (down to a few nanometers). There are a plethora of books on optical microscopy, but a good overview of the different types of optical microscopes is given by Rollins and deGruy [277] in a book that was written primarily for analysis of cotton cellulose but also makes reference to papermaking ®bers. Much of the important work in the elucidation of ®ber structure was done using optical microscopes, a great deal of it pioneered by plant anatomists. Many phenomena relating to the structure of wood and ®bers were reported early on by skillful microscopists but did not receive recognition until much later. One such phenomenon was ®brillation. Strachan [310] was the ®rst to observe ®brils on the surfaces of beaten ®bers, but his observations could not be con®rmed until Clark [45] produced micrographs of disintegrated ®bers that had been silvered. Optical microscopes can be used with incident or transmitted light or a combination of these depending on the specimen. Specimens may be living or dead, wet or dry. Optical microscopy using transmitted light requires samples to be relatively transparent and thin, and if they are not thin enough, then sections need to be

Microscopy

153

embedded and cut with a microtome. Embedding procedures normally entail dehydration, involving the use of a solvent, followed by in®ltration and polymerization of the embedding medium, usually a resin. The Stereoscopic Microscope Stereoscopic microscopy allows specimens to be viewed in three dimensions. Using incident light with a stereomicroscope at low power gives good depth of ®eld, making it possible to obtain a great deal of information about surface structures of fairly large specimens. The stereomicroscope is often used for preliminary examination of specimens. Compound Microscope The compound microscope is the ``conventional'' light microscope, a standard piece of equipment in paper technology laboratories and the classical tool for ®ber and pulp analysis. Isenberg [136] gives a comprehensive account of the optical system of the compound microscope and lists its common uses in pulp and paper mill laboratories where routine standardized methods are employed for quality control. It is used for identi®cation of wood [323,324] and nonwood [41,322] ®bers. Photographic atlases have been compiled for identi®cation of papermaking ®bers [54,250], although for purposes of basic wood ®ber identi®cation it would be hard to better the ®ne light micrographs in Ilvessalo-PfaÈf¯i's Fiber Atlas [134]. The compound microscope is also used for ®ber analysis of paper and board, determination of ®ber coarseness and weight factor, analysis of ®llers and coating particles, and analysis of foreign particulate matter in pulp, all according to the appropriate standards [320,321,325,326]. Jordan and Page [155] discuss the application of image analysis for measurement of ®ber length, width, and coarseness and the determination of weighted average ®ber length using automated and manual procedures. They also describe ways of de®ning and measuring ®ber curl. Weidner [348] designed a special cell with through¯ow of air from a constant humidity apparatus that ®tted onto the microscope stage. The top of the cell was covered with a glass cover slip so that changes in ®ber length and width with varying relative humidity could be measured. Giertz [101] characterized ®ber and ®nes fractions of various mechanical and chemimechanical pulps and showed some interesting features in his light micrographs. Pelton et al. [254] and Luukko et al. [194] measured the size distribution of ®nes particles in mechanical pulps. Samples of ®nes material were dyed and mounted on glass slides, and their images were captured by video camera for image analysis. McCool and Taylor [206] describe the techniques for quantitative analysis of ink particles and stickies that they used to determine the ef®ciency of deinking processes in waste paper recycling. Measurement of contact ratio, the area of ®bers bonded to a glass slide (measured under incident light) as a ratio of the total area of ®bers (measured in transmitted light), was ®rst proposed by Clarke [47] as a method for obtaining an indication of the wet plasticity of ®bers. Sloane [300] used this method to study the effect of recycling on the bonding potential of softwood kraft pulps and provided details of slide preparation and image analysis procedure. Kallmes and Eckbert [157] felt that relative bonded area (RBA) was the parameter that best de®ned the structure of paper and applied it to pulp and paper evaluation. They found that direct measurements of RBA using optical microscopy did not correlate well with indirect methods such as N2 adsorption, which measures uncollapsed lumen area in addition

154

Moss and Groom

to external ®ber surface, so they used a lumen-corrected adsorption technique to take into account the contribution made by uncollapsed lumina. Investigations into wet ®ber ¯exibility have also been undertaken with the light microscope. Various workers have developed techniques for measuring wet ®ber ¯exibility and conformability. Tam Doo and Kerekes [318] derived an index of ¯exibility by measuring the de¯ection of selected ®bers in a ¯ow of water. Their experimental setup included a microscope for measuring the diameter of each ®ber before it was de¯ected. Mohlin [213] showed that the conformability of a wet ®ber draped over a glass ®ber on a microscope slide could be assessed by measuring the length of the ®ber that was not in contact with the glass, and Steadman [305,306] formed a thin ®ber network over a series of very ®ne steel wires laid parallel across glass slides. This allowed ¯exibility measurements to be made of a larger number of ®bers. Bucher [35] examined the structural organization of wood ®bers. Fibers were macerated, and the use of a differential metachromatic stain enabled the primary and tertiary walls to be distinguished from the secondary wall. Helix angles in the tertiary walls were measured for several conifer species. Many workers have studied the effects of beating on the development of ®brillation. McIntosh [207] dried down well-beaten ®bers onto microscope slides and then shadow-coated them with silver to enhance the ®brillar structures. These effects of beating can be studied qualitatively and quantitatively. Laamanen et al. [179] obtained a ®brillation index for ®bers dried down onto microscope slides. A differential staining technique was developed by Simons [296] to evaluate the degree of ®brillation. Intact ®bers stain blue, whereas ®brillated ®bers turn orange. This technique was used by Blanchette et al. [23] as a rapid screening tool for evaluating the effectiveness of various species and strains of fungi used to ®brillate ®bers in biomechanical pulping. Similarly, Jackson et al. [138] used Simons stain for characterizing surfaces of ®bers after treatment with cellulase and/or hemicellulase. Microscopical analysis strongly suggested that cell wall fragments or ®brils were being removed from surfaces by hydrolysis. Jackson et al. [139] also adapted a gold-labeling technique normally used for transmission electron microscopy (see Section IV.E) to see where the cellulases bind to ®bers. They enhanced the enzyme±gold complexes with silver so that they were large enough to be resolved using a light microscope. Lackner et al. [180], in their study of wood rot, also employed the silver staining technique to enhance immunogoldlabeled ligninases for optical microscopy. Most paper sheets are too opaque for examination in transmitted light. Surface features can be studied using oblique lighting or vertical illumination methods, but this can be done only at low magni®cation. For high magni®cation work, surface contrast must be greatly enhanced in order to reveal ®ne structural details, and this can be achieved by metal shadowing or surface replication techniques. Williams and Wykoff [358] ®rst proposed the use of metal shadowing for examining surface structures with the electron microscope, and this technique was subsequently adapted for optical microscopy [359]. It entails directional deposition of a thin layer of metal by evaporation under high vacuum onto the surface of the specimen. Raised features intercept the beam and give rise to metal-free regions on the leeward side. Paper sheets that are suf®ciently translucent can be prepared in this way and produce a negative image when viewed in transmitted light (the raised features appearing dark and the metal-free depressions light), but a positive image can be obtained by photo-

Microscopy

155

graphic reversal. Alternatively, replicas of paper or wood surfaces can be made. These were also initially made of metal for electron microscopy, but the replica method was extended to the ®eld of light microscopy, and for this purpose they are usually made of transparent plastic material. Plastic replicas are then metal shadowed and can be viewed in transmitted or incident light. These techniques for studying wood ®bers and paper were developed and employed largely by Emerton [79], Emerton et al. [81], and Page and Emerton [239]. Page and Sargent [240] used the carbon replica method to study the ®ne structure in contact regions of ®bers that had been dried onto glass or a steel drying plate and then removed. They examined the same bond areas using both light microscopy and electron microscopy and showed clearly the areas where the ®ber surfaces had formed close contact with the substrate. Emerton's classic work [78,79] is illustrated by many examples of replica techniques used by various workers, and Emerton et al. [80] presented an atlas of paper surfacesÐlight micrographs of metal replicas of a wide range of papers viewed in oblique lighting. Optical microscopy has been much used for examining cross sections of ®bers and sheets. This usually necessitates the use of embedding techniques and mechanical sectioning. The method used for cross-sectioning paper depends on the purpose of the study; Graff and Schlosser [105] review various embedding procedures and their appropriateness for various types of material. Nissan [228] recommended that sedimentation studies of papermaking ®bers should be accompanied by direct visual observation of the dimensions and shapes of ®bers. Single ®bers, placed in a sedimentation cell and timed as they fell through a known distance, were then placed, in a wet state, on a microscope slide for measurement of length and width. They were then dipped in paraf®n wax and microtomed so that ®ber wall thickness could be measured from cross sections. Page et al. [241] described in detail their methods for obtaining sections thin enough for examination in transmitted light but thick enough to maintain structural integrity and presented some examples of cross sections of various types of papers. Leopold and McIntosh [185] measured the tensile strength of individual alkali-extracted ®bers. The ®bers were then embedded in resin, and thin sections were cut as close as possible to the break in order to measure the crosssectional area of the ®bers. They showed a correlation between reduction in ®ber strength with the removal of xylan-based hemicellulose and also a correlation between decrease in strength and cross-sectional area with an increase in the concentration of the alkali used for extraction. Robertson and Mason [273] examined cross sections of embedded pulp ®bers to study the effect of deligni®cation on ®ber collapse by comparing cross sections of spruce ®bers from high and low yield pulps when wet, and then after sheet forming and drying. The cross-sectional shapes and transverse dimensions of individual ®bers have also been examined for changes in structure arising from beating processes, much of it by Kibblewhite and coworkers [165±168]. Kibblewhite and Shelbourne [169] found that the combination of length and width/thickness ratio of kraft ®bers could be used to predict apparent density. Szikla and Paulapuro [317] developed a method using image analysis for measuring z-direction changes in density distribution during wet pressing. They tinted pulp with methylene blue during re®ning in order to avoid the damage that can ensue from post-staining thin cross sections of embedded samples, and they also tested handsheets made with and without methylene blue to con®rm that the staining

156

Moss and Groom

treatment did not affect ®ber bonding. Mardon et al. [202] investigated changes in paper structure during calendering by examining paper sections at various stages of the process. Bergh and Thomin [19] investigated paper±liquid interactions and penetration of sizing and coating mixtures. Such techniques have been used largely for analyzing the distribution of ®llers in sheets, coating thickness, and pro®le analysis, details of which are given by Elton and coworkers [77,99], and also for examining the effects of coating on the surface roughness of base papers [186,297]. Marton and Crosby [204] cut cross sections of embedded laminating papers in order to study the penetration of dyed resins into the paper sheets and examine the reaction between the resins and the base paper. Quackenbush [259] developed a technique for making precise measurements of individual cells on the metering roll of an offset rotogravure coater. An impression was taken of a small area of the roll on a thin piece of plastic such as cellulose acetate or cellulose nitrate. This was then sandwiched between two blocks of wax for microtoming. Measurements made from the cross sections made it possible to determine the correct coating weight for print roll engraving. This technique could also be applied to studying the corrosion of metal surfaces and the wear and tear of printing plates and of the cutting edges of re®ning equipment. Phase Contrast Microscope Phase contrast microscopy makes visible the differences in retardation of light waves as they pass through a specimen, the amount of retardation being proportional to the thickness of the object and the difference between its refractive index and that of the surrounding medium. A phase plate placed in the microscope converts the differences in refractive index into differences of contrast in the image, allowing better observation of transparent specimens. Phase contrast microscopy has been especially useful in biological and medical research because it obviates the need for staining living cells. Robertson [274] used it as a diagnostic tool to identify microbial populations to facilitate the appropriate selection and placement of biocides in papermaking systems. Asunmaa and Marteny [5] used phase contrast microscopy to reveal the manner in which the outer cell wall layer peels off ®bers, and Crosby and Mark [57] employed phase contrast microscopy in conjunction with near ultraviolet illumination to determine ®bril angle in the S2 wall layer of pulp ®bers. Their technique overcame some of the problems encountered in using polarizing microscopy for measurement of ®bril angle, which necessitates careful preparation of ®bers to expose a single wall layer. Generally, phase contrast microscopy does not appear to have been used much for paper science. There are some references to its use in textile research (see Rollins and deGruy [277]). Interference Microscope Like the phase contrast microscope, the interference microscope makes visible the relative retardation of light through a specimen, but it is primarily designed to measure the differences in retardation of light rays passing through the object and those passing through a reference area in a mounting medium of known refractive index. Knowing the thickness of the specimen and the refractive index of the surrounding medium, it is then possible to determine the concentration of solids and the dry mass of the specimen. Thus it is generally used for quantitative studies rather than the observation of specimens. Lange [183] described methods for determining mass per unit area of lignin and carbohydrates in different regions of the

Microscopy

157

cell wall. Page et al. [238] used interference microscopy to measure the compacted thickness of single ®bers, in order to determine ®ber stress after straining. Howarth et al. [131] describe the use of a modi®ed interference microscope to measure adhesive distribution in pigment coatings on paper. Polarizing Microscope The polarizing microscope measures and analyzes the effects of a beam of polarized light passing through a transparent specimen, thus determining the orientation and crystalline characteristics of the structure of the specimen. Polarization of the light is achieved by the use of Nicol prisms or polaroid sheets. Two polarizing devices (a polarizer and analyzer) are used. When a polarizing object, such as a crystal or highly oriented ®ber, is placed on the microscope stage with the polarizers in crossed position …908†, the parallel-polarized light striking the specimen is repolarized in the two perpendicular directions, and the waves, on striking the analyzer, are repolarized. The polarizing microscope is used principally for determining orientation by measurements of refractive index and birefringence. Many materials polarize light to some extent, and evaluation of this characteristic allows correlation of the optical properties of a material with its physical behavior. Bailey and Kerr [9,10] were pioneers in elucidating the cell wall structure of ®bers. Using polarizing microscopy they showed the concentric layering of the secondary wall of ligni®ed wood ®bers and observed the helical orientation of ®brils in cell wall layers. On the basis of observations and measurements made using polarizing microscopy and X-ray spectrometry, Wardrop and Preston [344] proposed a model of the wall structure of tracheids and ®bers. Kallmes [156] used polarizing and metallurgical microscopes to examine the structure of wall layers unraveled by prolonged disintegration. His ®ne micrographs illustrate the ®brillar structure peculiar to the primary, S1, and S2 wall layers, and wall thicknesses was measured using the polarizing microscope. Alexander et al. [2] used polarized light to measure ®bril angle and to determine changes in ®bril angle caused by beating and wet pressing, and Leney [184] describes in great detail the technique he employed for measuring ®bril angle. Measurements of ®bril angle are obtained by rotating the cell wall in a beam of plane polarized light to the major extinction position, which appears black so nothing can be seen. The drawback with this technique is that it requires careful preparation of ®bers so that only a single wall layer is exposed because polarized light passing through the two walls of a whole ®ber will effectively cancel. However, Page [234] overcame this problem by introducing mercury into the lumina of ®bers, which blocked out the other half of the cell wall. Recently, Ye and SundstroÈm [369] described a new technique for determining the ®bril angle of the S2 wall layer, based on microscopic transmission ellipsometry using a conventional polarizing microscope equipped with a charge-coupled device (CCD) camera or video camera. Whole ®bers, wet or dry, can be examined without any special preparation. Page and De GraÃce [237] used polarized light to observe delamination of ®ber cell walls resulting from beating and re®ning. Pulps differ in their ability to swell during beating, and some delaminate more than others. Page and De GraÃce showed that sul®te pulp ®bers split into 10 or more lamellae, whereas sulfate pulp ®bers split into only two to six lamellae. McIntosh [207] used polarized light to increase de®nition when examining delamination in cross sections of wet ®bers that had been set in gelatin, frozen, and then sectioned on a microtome with a freezing attachment.

158

Moss and Groom

A great deal of knowledge of ®ber bonding has accrued from the use of the polarizing microscope, much of it from the work of Emerton, Page, and coworkers. Page [233] described a new technique of sandwiching dry ®bers between two strips of Scotch tape and using a combination of transmitted and polarized light to study the collapse behavior of ®bers. Emerton and Watts [82] proposed the use of polarized vertical illumination as a means of distinguishing the primary wall (which shows bright under crossed polarized devices) from the secondary wall (which shows dark) to study the effects of beating on the removal of the primary wall in spruce tracheids. Page [232] adopted this technique to make direct observations of areas of optical contact between ®bers in a paper sheet (where areas of optical contact appear black), and Page and Tydeman [242] made a survey of ®ber±®ber bonding in sheets made of various furnishes, both wood and non-wood. Page et al. [244] examined the size and shape of ®ber±®ber bonds, looked at their frequency of occurrence in paper sheets, and investigated the effects of beating on these parameters. They also made a quantitative examination of loss of bonded area when ®bers in a sheet of paper were strained to rupture [245]. McIntosh and Leopold [208] determined the bonding strength of individual ®bers. The area of ®bers bonded either to cellulose ®lm or to shives was ®rst measured in polarized light, and then tension was applied to determine the load required to release the ®bers. Page and Tydeman [243] also used polarized vertical illumination to examine bond formation in paper during drying. The web was formed in situ, and a modi®ed porous plate apparatus allowed controlled removal of water so that each image was recorded at a known moisture content of the web. Yang et al. [368] evolved a technique for making direct measurement of the bonded area of a ®ber in a sheet. This entailed using an embedding technique developed by Quackenbush [260] and cutting numerous serial sections. The sections were examined using both phase contrast and transmitted polarized light, and many ®bers could be followed through a series of sections without any dif®culty. Bonded area, aspect ratios, moments of inertia, and bonding state probabilities of the ®bers were measured using a point mode digitizer interfaced with a computerized data processing system. Polarized light has also been used for lignin studies, and Coppick and Fowler [50] developed a silver staining techniques for woody tissue that differentiates degrees of ligni®cation in the cell wall by color reaction. Jordan and O'Neill [154] describe a method of detecting mechanical pulp ®nes particles composed of lignin and hemicellulose, which are not birefringent and therefore are not detected by ®ber length analyzers such as the Kajaani FS series. Quackenbush [261] describes techniques developed to study ¯aws in coated papers and presents a series of case studies to illustrate the usefulness of the polarizing microscope as an analytical tool for the identi®cation of contaminants in coated papers. Ultraviolet and Fluorescence Microscopes Ways of using shorter wavelengths of light were developed in an attempt to increase the resolution of optical microscopes. A ¯uorescence microscope is just a standard instrument that has been modi®ed so that the wavelengths needed to induce ¯uorescence can be focused on the specimen and the emitted ¯uorescence observed. This is achieved by the use of ®lters. An excitation ®lter is placed between the light source and the specimen, and a barrier ®lter is placed between the specimen and the viewer. KoÈhler's original observation of

Microscopy

159

¯uorescence was with UV light, so early workers thought it was always necessary to use wavelengths of <350 nm, but ¯uorescence microscopy commonly employs nearUV and short-wavelength visible light, because most ¯uorochromes that occur naturally in plant tissues ¯uoresce strongly in these wavelengths. The most commonly used light source for general-purpose ¯uorescence work is a mercury vapor lamp. Fluorescence microscopy has featured largely in the study of living plant tissues and fossilized plant remains, and ¯uorochromes have been employed to trace the movement of cytoplasm and the transpiration stream in plants [231]. It is not necessary to embed tissues and cut thin sections for ¯uorescence microscopy if epi-illumination is used instead of light transmission. Fresh material needs no more than a deft cut with a sharp razor and immersion in a ¯uorescent stain. This enables rapid screening of large amounts of material. The main advantage of using ultraviolet light is that it is strongly absorbed by unstained biological material, producing good contrast in the image. The UV microscope has not been useful for cellulose, but it has been frequently used for the study of lignin. Frey-Wyssling [94] used UV optics and ¯uorescence for quantitative evaluations of the content and distribution of lignin in the cell wall, and Fergus and coworkers [87,288] carried out similar studies using UV light. Goring [104] measured the distribution of lignin in spruce and birch wood and showed the progressive removal of lignin during chemical pulping. Fukazawa and Imagawa [96] used both UV and ¯uorescence microscopy to study the decay and deligni®cation of wood, using acridine orange stain in conjunction with ¯uorescence microscopy for qualitative histochemical identi®cation of cell wall components. Boutelje and Eriksson [26] used UV microscopy to determine the concentration of lignin in fragments of a spruce thermomechanical pulp (TMP) and used an interference microscope to obtain the thickness measurements. Wardrop [343] used a simple piece of apparatus to draw or force liquids through small blocks of wood. A silver staining reagent [50] was added to the pulping medium passing through the specimens. Ultraviolet microscopy of cross sections showed that the pits provided the main path of penetration in both heartwood and sapwood, even when the pits were aspirated. Confocal Laser Scanning Microscopes Confocal laser scanning microscopes (CLSMs) are optical microscopes that use an intense laser beam, focused to a small point, as a light source. The beam is scanned across the specimen in raster fashion. Unlike images produced by a conventional optical microscope, which contain both blurred and sharp areas, confocal laser scanning microscopic images show only structures that are in focus, and out-of-focus areas are not seen. The confocal scanning microscope makes possible applications that have been dif®cult or impossible to perform with conventional optical microscopes, such as the nondestructive sectioning of specimens in their natural state. The CLSM is described in detail in Section IV.

B.

Electron Microscopes

The advent of electron microscopy (EM) presented greater scope for structural research into ®bers. A good general review of electron microscopy is given by Rollins et al. [276], and Parham [248] discusses the advantages and disadvantages

160

Moss and Groom

of transmission electron microscopy (TEM) and scanning electron microscopy (SEM) and their appropriateness to different research applications. In electron microscopes a beam of electrons is used instead of light. The electrons are generated by passing a current through a very thin tungsten wire that is heated to  25008C, at which temperature electrons are emitted from the hot metal. Electrons are accelerated from the ®lament by applying a high negative voltage to the cathode assembly. The greater the voltage, the greater the energy of the electrons and the higher the resolution that can be achieved. The electron gun consists of a cathode and anode assembly, and the electrons pass through a small hole in the center. The electron beam is focused by electromagnetic lenses. All this takes place in a vacuum column, because air molecules will de¯ect the beam. For the same reason specimens must be dry so the electrons are not scattered by water molecules. The drawback with all conventional electron microscopy is that the presence of water prevents direct examination of pulp stock and wet sheets, so water must ®rst be removed from specimens, but in a way that maintains the wet structure. Dehydration of material may be necessary, and this may cause shrinkage, distortion, and collapse of specimens to varying degrees. Embedding techniques necessary for TEM work may cause further structural changes [251], and mechanical sectioning may also cause damage or deformation. Finally, the deposition of a thin layer of metal during sputter-coating of samples for scanning electron microscopy can disproportionately increase the dimensions of very ®ne structures [37], although this can be advantageous in enhancing very ®ne ®brillar material, which is otherwise dif®cult to detect. Watts and Emerton [346] discuss some of the problems associated with interpreting electron micrographs arising both from the preparation techniques, such as embedding and sectioning, and from the conditions of the instrument itself, such as beam damage. Sectioning can give rise to several types of artifacts. Imperfections in the cutting edge of the knife show up as lines scored across the specimen, and knife chatter may give rise to periodic waviness in specimen thickness. Compression in the cutting direction may distort the shape of structures in the specimen and may even cause cracking or delamination of structures, and the subsequent mounting of ultrathin sections may result in wrinkling and creasing of the specimen. Studies undertaken by Davies [61±63] are typical of many that were made to investigate ®ber±®ber bonding in sheets and the effects of beating on ®ber conformability and bonding. However they also serve to illustrate some of the defects resulting from ultrathin sectioning outlined above. Transmission Electron Microscope Conventional transmission electron microscopy (TEM) has certain limitations, because the electron beam can penetrate only very thin specimens. The preparation techniques are tedious, usually entailing dehydration, impregnation with liquid resin or plastic, polymerization, sectioning, and staining. Only ultrathin sections (<100 nm) can be examined, and the size of the specimen is limited to the size of the supporting grid (typically 3 mm diameter). Sections require staining to increase contrast; salts of heavy metals (e.g., lead citrate, uranyl acetate) do this by creating electron-opaque regions in the specimen. Furthermore, each stage in the preparation of specimens for TEM examination may induce certain physical changes, resulting in shrinkage and/or distortion of a specimen. Some of the problems associated with dehydration can be overcome by freezing the specimen and then embedding it by freeze substitution.

Microscopy

161

However, certain suitable specimens can be mounted directly onto a sample grid. Fibers can be disintegrated into small fragments, either in a laboratory blender or by ultrasonic vibration. Alternatively, thin replicas of the surfaces can be prepared (see details in Section III.A). Replica techniques were developed as a means of studying the ultrastructure of wood and ®ber surfaces. CoÃte et al. [55] review the techniques for surface replication of samples for examination in the electron microscope and give a detailed description of the technique for making direct carbon replicas of wood and paper samples. This method is recommended for its simplicity and high reliability in representing ®ne details. Comer et al. [49] present examples of replicas of papers coated with various pigments commonly used in coatings. The excellent structural details of particle size, shape, and orientation help to explain properties such as gloss and ink receptivity. Despite the limitations imposed by the TEM on the study of papermaking ®bers, its great advantage lies in the high resolution that can be achievedÐdown to 0.2 nm with modern instruments. The earliest reported instance of the use of a TEM for examining pulp was by Sears and Kregel [289], who discussed its application to problems associated with pulp and papermaking. The TEM has been invaluable in characterizing the primary units of native cellulose and showed for the ®rst time the micro®brillar structure of cellulose ®brils [53,86,124,125,263]. Examination of the morphology of cellulose synthesizing bacteria has also helped to shed further light on the structure of cellulose [65]. As with all microscopy, correct interpretation is essential, and Hanna [117] pinpoints some of the pitfalls involved in such studies. Jayme and coworkers [147±150], CoÃte [53], Asunmaa and Steenberg [4,6], and, more recently, Nanko and coworkers [224,225] are among many workers who made notable contributions to the study of ®ber structure and ®ber bonding using the TEM. Ultrathin sectioning, employed for many TEM studies of ®ber and sheet cross sections, enabled Stone and Scallan [308] to present direct evidence of internal ®brillation of cell walls during beating (see Fig. 1). Internal ®brillation, a term coined by Campbell [39], is the splitting apart of the cell wall layers into their constituent lamellae, which, according to Gallay [98], are thus ``separated or resolved into a series of concentric sleeves, like layers on an onion.'' Stone and Scallan postulated that micro®brils were associated laterally into sheets or lamellae and, because the micro®brils had greater adhesion in the tangential direction than in the radial direction, ®ber wall swelling and internal rupture of lateral bonds caused the wall to delaminate into coaxial layers. But, because cleavage also occurs in the tangential plane, Scallan [286] extended the concept of concentric lamellae to a ``honeycomb'' structure. Water-swollen ®bers were dehydrated by solvent exchange. The ®bers were then embedded in methacrylate and mechanically sectioned. The methacrylate was then dissolved from the sections with chloroform. However, they recognized that their technique was responsible for exaggerating cell wall delamination and therefore precluded quantitative analysis. Tripp and Giuffria [336] had noticed earlier, using the same techniques, that the cross-sectional size of partially acetylated cotton ®bers swelled by 200±400%. They suggested that this phenomenon might ``be useful for studies where it is desirable and possible to break down the organization of the ®ber cell wall,'' and Dlugosz [71] took advantage of the expansion effects of methacrylate polymerization to reveal the cell wall structure of cotton ®bers. Jayme [147] had earlier shown similar cross sections of coaxially delaminated cell walls in spruce ®bers that had been deligni®ed with sodium chlorite after swelling with 5% NaOH.

162

Moss and Groom

Fig. 1 Cross section of beaten ®ber showing internal ®brillation. (From Ref. 308.) (Photo by G. M. A. Aberson.)

Gaining an insight into the structure of the cell wall is an important area of fundamental research. Pulping and re®ning processes produce changes in the molecular structure and chemical composition of cell wall components. For example, they cause redistribution of hemicelluloses, and knowledge of this is important to an understanding of ®ber bonding. A means of labeling hemicelluloses has long been sought. In 1949, Mark (see Ref. 295, p. 115) wanted a substance ``containing a heavy atom which reacts speci®cally with the hemicelluloses'' for use with electron microscopy in order to ®nd out something about their distribution in a sample. And in 1970, Page (see Ref. 53, p. 329) reiterated this need to develop ``methods for introducing heavy metal atoms into speci®c sites to provide electron contrast.'' He thought that the locations of lignin, hemicellulose, and cellulose could be found by this approach. Even today the interaction of the hemicellulose±lignin matrix with the cellulosic ®brils is not really known. However, some speci®c labeling techniques have been developed. Enzymes can be tagged with metal particles, and when these enzyme±metal complexes are incubated with the appropriate wood or ®ber samples the enzymes bind to the speci®c polysaccharide. Jackson et al. [139] give full details of the preparation techniques. Ruel and Joseleau [279] used enzyme±gold complexes to study the distribution of xylans in grasses and glucomannans in tracheids of spruce ®bers, and Mora et al. [214] employed the same technique to study the redeposition of xylans on birch kraft pulp ®bers. Immunolabeling techniques have also been used. Srebotnik and Messner [303,304] used gold-labeled rabbit antibodies conjugated with marker proteins of different sizes to examine the porosity of wood cell walls. Shijun [293] used a TEM to study the distribution of hemicelluloses in bagasse ®bers. He ®rst oxidized the aldehyde groups to carboxyl groups using NaClO2 and then stained the ®bers with KMnO4. Potassium permanganate is electron-opaque and is commonly used to provide contrast in TEM sections, but for lignin studies it is

Microscopy

163

used as a speci®c stain. The manganese ions react with the carboxyl groups and attach to them, so the presence of hemicellulose is denoted by dark areas. Energy-dispersive X-ray analysis (EDXA) can also be used in conjunction with the TEM to analyze various elements in a sample and determine their distribution. Saka and coworkers [283±285] undertook comparative studies of lignin distribution using TEM/EDXA and UV microscopy. Re¯ection Electron Microscope The re¯ection electron microscope (REM) was ®rst employed for examination of ®ber surfaces by Chapman and Menter [43], who used a commercial TEM modi®ed to operate in re¯ection mode. In the REM, specimens are illuminated by an electron beam at a glancing incidence. The advantages of REM lie in a greater depth of ®eld, giving images a three-dimensional appearance and allowing easy interpretation of the spatial relationship of structures. However, in early instruments beam damage caused polysaccharides to break down into gases, giving rise to bubble artifacts [3]. It is also necessary to use a higher beam intensity than with the TEM to make the image visible on the screen. The REM was used to examine pulp and paper for only a short period in the mid-1950s. Scanning Electron Microscope Goldstein et al. [103] provide a comprehensive review of scanning electron microscopy (SEM) and its applications. Vaughan [338] and DeNee and Abraham [64] provide further details of microanalytical techniques. Walbaum [342] gives a clear account of the SEM and its functions and discusses the factors that in¯uence image formation and contrast. Parham [247] outlines the principles of X-ray/SEM analysis and its application to paper. In the SEM the electron beam is focused into a very small spot on the specimen and traverses over each point of the specimen in turn, one scan line at a time, to form a raster. The electrons emitted from each point on the specimen are collected by a nearby photomultiplier tube and produce a signal on a cathode ray tube that is being scanned in synchrony with the specimen. The electron beam stimulates the emission of electrons from the target spot, and these vary in abundance with the topography of the specimen; crevices emit the fewest electrons and high points the most. The resolution of the SEM, generally 3±10 nm, is far better than that of the light microscope but inferior to that of the TEM. The great advantage of the SEM lies in its large depth of ®eldÐat least 300 times that of a light microscope at comparable magni®cation. The SEM is ideal for large bulky specimens and those of coarse structure. Small, whole specimens or thick pieces of material may be examined in the SEM, the size of the specimen being limited by the size of the sample holder, usually 1±2.5 cm in diameter, although larger samples can be accommodated. Dry specimens of wood and paper can be examined directly, but, as mentioned earlier, wet or moist specimens that are to be examined by conventional SEM have to be dehydrated. The favored method seems to be critical point drying rather than freeze drying because it is felt that critical point drying is more consistently reproducible; see Parham [249] and Sachs [281,282] for pulp ®bers, Boyde [28] for animal tissues, and Parsons et al. [251] for plant tissues. However, it is possible to examine a moist uncoated specimen, such as a fresh leaf, directly in the SEM at a low voltage. This is possible for only a short period because the specimen will dehydrate very rapidly in the vacuum chamber [74,251]. Whichever specimen preparation method is used, it is usually necessary to coat the specimen with a thin layer

164

Moss and Groom

(10 nm) of metal, usually gold or carbon. This provides a conductive surface that allows the electrons to drain off and thus prevents a buildup of electrostatic charge. The SEM also offers the possibility of elemental analysis using backscatter detection or X-ray analysis. The interaction of the electron beam with the specimen produces, in addition to secondary electrons (SE), backscattered electrons (BSE). As the incident beam enters the sample, secondary electrons are generated, and it is these that are normally detected to give the topographical information. Backscattered electrons are those that re-emerge from the sample and, as they exit, generate their own secondary electrons, which can be collected by a special detector. This is usually referred to as backscattered electron imaging (BEI or BSI). The amount of backscattering that occurs is dependent on the atomic number of the elements present. The higher the atomic number, the more electrons are backscattered. The emerging backscattered electrons carry information about the nature of the specimen over a range of depths, the depth depending on the exact nature of the specimen. Similarly, an X-ray detector can be used in conjunction with the SEM to identify elements, because each element generates characteristic X-rays, and elemental mapping can be used to show the location of an element over the surface of the specimen [51, pp. 87±89]. Scanning electron microscope energy-dispersive X-ray spectrometry (SEM/EDX or EDS) studies have become routine, and quantitative analysis can also be carried out using the appropriate standards and reference materials. The SEM is best suited for the study of surface structure, and its remarkable depth of ®eld means that the surface of rough or irregular samples can be brought into focus over an area of several millimeters, and dry samples of wood and paper can be examined directly. It has proved to be an invaluable aid to descriptive analysis of wood and non-wood ®bers, and several ®ne books of SEM micrographs have been published [38,51,52]. Details of the applications of the SEM to pulp and paper research can be found in Smith [301] and in Ilvessalo-PfaÈf¯i and Laamanen [135]. James et al. [140] describe some of its applications to the study of Australian newsprints with respect to product performance. Rezanowich [270] considered that the usefulness of SEM lay largely in its low magni®cation range, enabling studies of extensive surface areas. It is especially useful for studying large numbers of specimens that can be rapidly screened, with areas selected for more detailed examination. This application of SEM has yielded information about the in¯uence of various factors on ®ber±®ber bonding and sheet density, e.g., ®ber collapse and conformability, ®brillation, and pulp yield [32,129,273]. Klofta and Miller [172] describe a method for obtaining a ®brillation index from scanning electronmicrographs of freeze-dried ®bers. Teder [327,328] observed the effects of beating, drying, and heat treatment on the papermaking properties of spruce and birch pulps using the SEM and found differences between species (spruce and birch), ®ber type (springwood and summerwood), and pulping process (sulfate and sul®te). Pye et al. [258] studied the structural changes that occur in papermaking processes and characterized the appearance of ®bers at different stages. Samples of the web were taken from different positions along the paper machine, from wet end to reel-up, and freeze-dried for examination in the SEM. Washburn and Buchanan [345] outline the freezing procedure. Stratton and Colson [311] examined ®ber wall damage during rupture of individually bonded ®bers. The bonded area was ®rst measured using Page's polarized

Microscopy

165

light technique, and after elongation failure the surface areas were examined in the SEM for visual assessment of damage. Buchanan and Washburn [33,34] examined the fracture faces of samples of handsheets made of pure chemical pulp ®bers and groundwood pulp ®bers broken in tensile testing. In general, the number of ®bers broken, as opposed to the number being pulled out, was higher in the stronger, chemical pulp sheets. The fracture lines were more sharply de®ned in the stronger sheets and very irregular in the weaker sheets. Nordman and QvikstroÈm [230] made similar studies but compared handsheets made of pure earlywood and latewood ®bers. They showed that rupture zones were clearly distinguishable by broken ®bers in the stronger sheets made of earlywood ®bers, whereas rupture zones were discernible only as a loosening of the structure in the sheets made of latewood ®bers. They also examined the rupture zone in sections of wood subjected to tensile fracture and showed that ruptures proceeded in almost straight lines in earlywood sections but were very irregular in latewood sections. The installation of a tensile stage inside the specimen chamber of the SEM enables in situ monitoring and analysis of the straining behavior of the paper structure. Retulainen and Ebeling [265] employed this technique to examine the effect of sheet structure on the load±elongation behavior of ®ber±®ber bonds. KucÏera and Bariska [177] used the SEM to study the structural changes of spruce and aspen wood samples subjected to axial compression and produced excellent scanning electron micrographs showing different types of deformationÐmicroscopic and macroscopic compression failure lines, stress lines, and slip planes. Bergander and SalmeÂn [18] applied load to specimens of spruce sapwood to investigate the effects of compression on the development of cracks and delamination in cell walls. Thayer and Thomas [330] used scanning electron microscopy to examine and analyze the physical characteristics at each stage of the formation of the glue line in corrugated board. The SEM is the obvious tool for observing the micro- and macroroughness of surfaces, and stereoscopic images give an even greater perception of the surface topography and depth of ®eld. No special equipment is required to characterize the surface structure; the sample is simply tilted, typically at an angle of 30±458 to the beam. Settlemeyer [291] used this method to examine the topography of coated papers and found that conventional measurements of gloss did not agree with visual perception. Enomae et al. [83] used a novel analytical technique, employing four secondary electron detectors, for constructing pro®le curves in order to investigate the surface topography of sheets and make surface roughness measurements. Helle and Johnsen [123] used stereo images with BEI for studying the distribution of ink particles on paper and ®ber surfaces, the atomic number of the ink particles being suf®ciently different to distinguish them from ®bers. Gregersen et al. [106] used a combination of electron imaging and BEI to examine ink distribution on paper surfaces at the single ®ber level in order to study both variation in local ink coverage and ink penetration into inter®ber voids. The SEM has been used also to characterize ®ller and coating particles [27] as well as for examining their distribution within sheets. MacGregor [195] discusses the use of BSI to compare surface distribution of clay and calcium ®llers in sheets formed on different machines. He also used EDS to map the calcium, illustrating the differences in amount and distribution of calcium on the two sides of the sheets. Peterson and Williams [257] used SEM/BSI with image analysis to determine coating

166

Moss and Groom

thickness and base roughness from cross sections of resin-embedded sheets. Klungness et al. [173] investigated ®ber loading as a means of minimizing horni®cation by precipitating calcium carbonate within ®ber voids in order to prevent subsequent collapse of the voids during drying. They cut cross sections of pulp ®bers and handsheets with a razor blade and used SEM with X-ray microanalysis to observe the calcium carbonate deposits on the surfaces, and within the lumina, of the ®bers. Cross sections of sheets for SEM examination need to be cleanly cut, by using either a razor or a special guillotine or by ®rst embedding the sheet for sectioning in a microtome. Hellawell and Nelson [122] compared, and illustrated, various crosssectioning techniques they employed for examining coating structures: freeze-fracturing, free-hand razor cutting of paper immersed in a lubricant, and embedding in resin and microtoming. The disadvantage of the latter is that the sections may be distorted by knife chatter. Gibbon et al. [100] developed a new technique for obtaining cross sections of paper sheets for high magni®cation viewing, which overcomes this problem. Several precut samples of paper are inserted, between spacers, in a plastic mold (30 mm diameter); then resin is carefully added. After the resin has been cured, the block is ground and then polished on a machine to reveal the section plane. This results in a mount with artifact-free cross sections of paper. Gibbon used this technique in conjunction with SEM/EDX to study the distribution of inorganic ®ller particles in sheets. Williams and Drummond [356] used Gibbon's method of preparing sections of paper in resin blocks and then, after sectioning, removed a thin layer of the supporting resin according to the method of Maxwell [205]. This produces very high quality sections for studying sheet structure, and Williams et al. [357] include examples of these cross sections viewed ®rst in normal transmitted light and then in the SEM. Browne et al. [31] used the same technique to examine the way in which sheet structure responds to calendering. They examined cross sections of uncalendered and calendered samples of a thermomechanical pulp (TMP) newsprint and a bleached kraft pulp paper containing hardwood and softwood ®bers for evidence of deformation and fracture of ®bers resulting from calendering treatments. Lignin reacts with potassium permanganate and bromine, and treatment with these chemicals raises the atomic number of lignin, making it detectable in SEM/BEI and SEM/EDX studies. Saka et al. [285] developed a procedure for labeling lignin with bromine gas. The amount of reacted bromine was found to be directly proportional to lignin content and, using SEM/EDX, they made a qualitative and semiquantitative analysis of lignin concentration and distribution within ®ber cell walls. Eriksson et al. [84] and Westermark et al. [351] undertook similar studies in birch and spruce wood, but used a mercurization technique to label the lignin. Gregersen et al. [107] treated paper surfaces and paper cross sections with potassium permanganate or bromine gas to study the distribution of middle lamella on TMP ®bers, to identify the mechanical ®bers in blended pulps, and to compare kraft pulps of high and low yield. However, the method is not recommended for quantitative determination of lignin. A similar method for locating ®nes in mechanical pulp handsheets is described by de Silveira et al. [67]. The ®nes fraction of a mechanical pulp was halogenated with chlorine or bromine and then added to long ®bers for handsheet preparation. Handsheets samples were embedded and cross-sectioned. The ®nes were readily detectable with BEI, the lignin appearing much brighter than the carbohydrates.

Microscopy

167

Forseth and Helle [93] used the SEM to investigate the effects of moistening on calendered paper such as occur during pigment coating and offset printing. They prepared cross sections of sheets embedded in resin and then dissolved away the resin at the surface, using the method of Williams and Drummond. This enabled them to observe and quantify water-induced changes in cross sections of the same ®bers by imaging the same areas before and after wetting. They also exposed samples to bromine gas and used SEM/BEI to differentiate between the chemical and mechanical pulp components of the sheets [90]. Low Temperature Scanning Electron Microscope The low temperature scanning electron microscope (LTSEM) is a conventional SEM to which has been added a special cryogenic unit. Use of the LTSEM obviates the need for dehydration of wet specimens. Instead, the specimen is rapidly frozen and is examined in a frozen hydrated state, its frozen state being maintained at a temperature of 1858C on a special cryostage. It is also possible to use cold stages with the TEM and scanning transmission electron microscope (STEM), but these are con®ned to specialist ®elds and do not appear to have been applied to paper science research. The LTSEM is described in detail in Section V. Scanning Transmission Electron Microscope This uses the scanning beam in a transmission mode, thus allowing the examination of thicker samples than can be examined with conventional TEM. Environmental Scanning Electron Microscope The ESEM differs from the conventional SEM in that the vacuum column contains a differential pumping system, and a special detector allows examination of specimens in the presence of water vapor. Specimens can therefore be viewed under ambient conditions and no coating is required. In fact, wet, frozen, or dry samples may be examined with the ESEM. The ESEM was designed speci®cally for examining liquid and hydrated samples, thus enabling the effects of wetting or drying processes to be monitored. Danilatos [59] gives a concise description of the ®rst commercial ESEM and a bibliography [60] of its applications, and de Silveira et al. [66] discuss the advantages and drawbacks of the ESEM. So far the ESEM has been applied mainly to studies of the effects of water on the swelling behavior of ®bers [151] and on surface roughening of papers [91,92,108]. Surface pro®lometry of paper and board is described in detail in Chapter 11. Mott et al. [221] used the ESEM to examine microstraining of single wood pulp ®bers. A comparison of ®ber images obtained with ESEM and AFM is given in Section VI. Although both techniques revealed image details of the cell wall, the resolution of the AFM was superior to that of the ESEM. C.

Scanning Probe Microscopes

Atomic Force Microscope and Scanning Tunneling Microscope These are the most successful of a series of scanning probe microscopes that were developed for high resolution imaging. A review of scanning probe microscopes is given by Wickramasinghe [353]. They have several advantages over the TEM, including the possibility of atomic resolution and their ability to operate at ambient air pressure and even in liquid environments, allowing structures to be examined in their native

168

Moss and Groom

hydrated state. The AFM is now a well-established tool in three-dimensional topographical analysis of various materials at the molecular level. The AFM is described in detail in Section VI. IV.

THE CONFOCAL LASER SCANNING MICROSCOPE

The confocal laser scanning microscope (CLSM) is variously referred to in the literature as a laser scanning microscope (LSM), scanning optical microscope (SOM), confocal ¯uorescence scanning microscope (CFSM), confocal scanning optical microscope (CSOM), and confocal scanning microscope (CSM). References will also be found to a tandem scanning re¯ection light microscope (TSRLM or TSM). The latter are real-time, direct-viewing confocal microscopes. Illuminating and detecting apertures are scanned in tandem. This is achieved by using a rotating disk containing thousands of tiny holes (the apertures) arranged in a precise and symmetrical pattern and simultaneously illuminating the specimen with hundreds of scanning beams. All CLSMs currently on the market are re¯ection microscopes. Dixon and Cogswell [68] and Dixon et al. [69] reported on a confocal scanning microscope for transmission and re¯ection imaging, but confocal transmission microscopes are optically complex and not yet commercially available. More recently Dixon et al. [70] described a new confocal scanning beam laser macroscope that can image large areas of specimens (7.5 cm  7.5 cm). When combined with the confocal microscope this produces a hybrid imaging system, the microscope/macroscope, that uses the advantages of small area, high speed, high resolution microscopy. Again, this is not yet commercially available. The basic principles of confocal microscopy described below are based largely on knowledge and experience of a ®rst generation Leica CLSM. There are several different types of CLSMs on the market, most of which can be used for the applications described here. Their ability to perform certain tasks is usually just a matter of whether or not the appropriate software is installed, but in some cases appropriate hardware may not be available. Lichtman [187] gives a brief general review of confocal microscopes. There are a few reference books that provide detailed accounts of the different aspects of confocal microscopy [253,360,361]. The book edited by Pawley [253] includes chapters that describe and compare each of the main component parts of these microscopes, provides a bibliography of confocal microscopes [347], and, in Appendix 2, gives brief descriptions of the models currently on the market. Centonze and Pawley [42] provide an excellent tutorial on practical confocal microscopy for the novice. There is also a wealth of information in the literature on the application of confocal microscopy, but most of it pertains to cell biology. Of these, a review by Shotton [294] provides the beginner with a good overview of the principles. A.

Basic Principles of the CLSM

The CLSM consists of a conventional optical microscope to which are coupled a laser and associated optical system (mirrors, optic ®ber, pinholes), a scanning unit, and photodetectors. A fast central processor controls microscope functions, drives

Microscopy

169

the scanner, converts detected light to digital form, and controls the subsequent processing of data. A wavelength selector, beam splitters, and a series of bandpass ®lters enable the microscope to be used in re¯ection or ¯uorescence mode. The changeover from conventional use to the confocal mode is achieved simply by moving a lever. This operates a beam-diverting system that prevents entry of laser light into the eyepieces when it is being used in the confocal mode. In a conventional optical microscope the whole ®eld of view is uniformly and continuously bathed in light and can be viewed directly, but this is not possible with the confocal microscope. The optical system of a CLSM contains two pinholes that are optically separated by a beam splitter (see Fig. 2). The laser beam passing through the illumination pinhole is focused to a ®ne spot that illuminates only one point in the in-focus plane. Because the detection pinhole is confocal with the illuminated point, only the light that is re¯ected or emitted from this point can pass through it (placed in front of the detector). Any light emanating from above or below the in-focus plane is eliminated. Because the laser beam is focused to a single point it has to be moved across the sample. This is achieved by using a scanning mirror to de¯ect the beam across the specimen, point by point, in raster fashionÐthe same principle as in the scanning

Fig. 2

The principle of a confocal microscope.

170

Moss and Groom

electron microscope. This method of scanning (referred to as off-axis scanning) can be very fast, allowing many frames to be built up per second and thus enabling rapid changes within specimens to be observed. In some microscopes, however, it is the object that is mechanically scanned across a stationary point of light (referred to as on-axis scanning). This method produces undistorted images of very high quality. In this way an image is sequentially built up point by point. Data are recorded as a set of intensity values for every pixel in the area scanned and are stored in the computer for future processing and simultaneously displayed on the image monitor. Thus, the image produced by the beam is a thin slice of just those parts of the specimen that lie in the in-focus plane, i.e., an optical section. So, whereas in a conventional transmitted light image everything in the image ®eld of view can be seen but only part of the image will be in focus (Fig. 3a) in the confocal images only those structures that are in focus will be seen. Elimination of out-of-focus structures means that there are no blurred areas in the image (Figs. 3b±3d). By moving the microscope stage vertically through a set distance, one step at a time, a series of consecutive optical sections can be collected. In short, the CLSM can be considered a 3D sampling instrument for collecting information at different levels. Figure 4 shows a series of optical sections taken in the xy plane through a CTMP shive at incremental steps of 2:4 mm along the z axis. Nondestructive optical sectioning is the unique feature of the CLSM. It enables detection of structural details that are otherwise concealed by overlying material. The absence of various conformation-in¯uencing preparatory steps allows a more faithful rendering of specimen reality than the conventional optical or electron microscope. Optical sections can be collected not only in the xy plane but also in the xz plane, which gives a cross-sectional view of the sample (Fig. 5). Optical sectioning in the xz plane is achieved by moving the microscope stage very rapidly

Fig. 3 Pulp ®bers seen in conventional transmitted light (upper left) and the same ®bers seen in confocal mode at depths of 0, 16, and 31 mm.

Microscopy

171

Fig. 4 Nine of a series of 16 optical sections taken through a CTMP shive to a depth of 36 mm. Ray cells can be seen beneath the tracheids.

Fig. 5

Samples can be sectioned in the xy plane and in the xz plane.

172

Moss and Groom

up and down while a single line scan is being made by the laser beam. Figure 6 shows TMP ®bers seen in the normal xy plane and the cross-sectional view made through those ®bers in the xz plane. It is the xz plane (or z direction) sectioning capability that is the novel and potentially most useful imaging mode of the CLSM, and it is the only way that cross-sectional views can be obtained of wet pulp and paper samples in a fully hydrated state. The generation of cross-sectional images of ®bers, combined with image analysis, enables ®ber cross-sectional area, wall thickness, etc. to be rapidly and accurately determined. To obtain accurate measurements of transverse dimensions it is important that ®bers be oriented perpendicular to the scanning direction of the beam, and this can be quickly veri®ed by ®rst examining the ®bers in the normal xy plane. Interestingly, although the use of optical sectioning obviates the need to align ®bers in bundles for embedding in resin, some workers have used the method of Williams et al. [357] to measure transverse dimensions of ®bers. Figure 7a shows the cross sections of ®bers believed to be aligned in a parallel fashion. However, when the series of cross sections is combined, as shown in Fig. 7b, then it is clear that erroneous assumptions can easily be made regarding the correct alignment of ®bers. The microscope stages of most CLSMs have a computer-controlled traveling range of 100±300 mm in the z direction, but the depth to which a sample can be optically sectioned depends on its density and transparency. Terasaki and Dailey [329] found that with oil immersion objectives, ¯uorescence imaging seemed to be limited to a depth of 50±70 mm. Not all CLSMs on the market have the capability of sectioning in the xz plane. However, software is available that enables cross-sectional views to be obtained from a series of sections collected in the xy plane. This will be further referred to later. Pinhole Size The ``thickness'' of the optical sections collected in the xy plane depends on pinhole size and on the numerical aperture (N.A.), which governs how much light is let through the objective. The size of the illumination pinhole is ®xed, but that of the detection pinhole can be varied, up to a maximum diameter of about

Fig. 6 TMP ®bers seen (a) in the normal xy plane and (b) in the xz plane. The xz scan was made across the center of the image in the normal plane and goes through a bordered pit.

Microscopy

173

Fig. 7 (a) A single optical section of ®bers aligned in parallel fashion and embedded in resin. (b) The extended focus image, made of series of sections taken at incremental steps of 5 mm; reveals the orientation of the ®bers.

30 mm. The larger the pinhole, the more light can be detected, some of which will come from out-of-focus regions of the sample. On the other hand, too small a pinhole will reduce in-focus light as well as out-of-focus light and result in loss of information. So it is important to select a pinhole size appropriate to the objective being used and to the purpose of the investigation. A smaller pinhole increases resolution but also reduces the signal-to-noise ratio so the image is noisier, whereas a pinhole that is too large reduces the depth discrimination of the microscope. Choice of Objective Low magni®cation objectives are recommended for rapid screening of samples in the conventional transmitted light mode in order to selct areas for more detailed examination in the confocal mode. Low magni®cation objectives have a low N.A. and therefore a large depth of ®eld compared with objectives of high N.A. This makes them unsuitable for scanning in the xz plane. The size of the N.A. (the number inscribed on the objective after the magni®cation) is especially important for axial resolution, because the increase in axial resolution is inversely proportional to the square of the N.A. Therefore, the higher the N.A., the better the depth discrimination. Objectives of 40 magni®cation or greater should be used where accurate depth discrimination is necessary and for measuring height differences in the z direction of the volume sectioned. Recommended for xz plane sectioning are the 50/1.0 water immersion and the 40/1.3, 63/1.4, and 100/1.3 oil immersion objectives. New objectives are being developed speci®cally for confocal microscopy [161], including water immersion objectives of 40/1.2 and 60/1.2, which would be particularly useful for examining wet samples of pulp and paper. The working distance of the objective also requires consideration. A large working distance is needed for examination of thick specimens, and the Zeiss Plan-Apochromat 40/1.0 objective, with a working distance of 380 mm, is well suited for this purpose. However, the objectives required for good depth discrimination (e.g., 63/1.4 and 100/1.3) have a small working distance, which limits their

174

Moss and Groom

use to thin specimens or to a thin layer at the surface of a specimen. A hardware zoom function, which can be used during scanning, allows increase in the magni®cation of the image (by up to a factor of 8 with the Leica CLSM) without having to change the objective. Choice of objective also depends on the mounting medium being used; ideally, the refractive index of whatever lies between the objective and the glass cover slip should match that of the mounting medium. Therefore, a water immersion objective should be used to examine specimens mounted in water, an oil immersion objective for specimens in commercial mounting media of the same refractive index as oil, and a dry objective for specimens dried down onto microscope slides. If there is a mismatch in refractive indices of the media, then optical aberrations will distort the size and shape of images collected in the xz plane. However, when specimens are being examined under changing conditions, as in drying experiments, the ideal canot be attained. A compromise may be reached by using a multi-immersion objective. (This problem is discussed further in Section IV.F.) Step Size When a series of consecutive sections is being collected through a certain distance, it is necessary to select a step size, i.e., the distance the microscope stage is to be moved between successive planes. Choice of step size will depend on the N.A. and consequently on the depth of focus of the objective being used. Therefore the step size should be equal to, or less than, the depth of focus if all available information is to be acquired. This is especially important where height differences are being measured, and it is always better to oversample because undersampling may result in incomplete information. However, when selecting step size, the depth to which the specimen is to be sampled and the number of sections to be collected must also be borne in mind, because some older models of CLSMs may have limitations in data storage capability and computing power. B.

Imaging Modes

The optical system contains beam splitters and bandpass ®lters that allow the CLSM to be operated in different modes. The majority of CLSMs can be used in re¯ection or ¯uorescence mode, and most of them nowadays have two or more photodetectors. This enables simultaneous acquisition in different image modes. Re¯ection Mode All available light from the laser is used in the re¯ection mode. This mode is usually used for dry specimens, particularly for surfaces of paper samples where roughness measurements are to be made. Re¯ection mode is used where staining for ¯uorescence work (which requires wetting the sample) is impractical. Re¯ection mode is suitable for examining dry samples containing ®ller particles and can also be used for specimens that have been labeled with metal particles. Shotton [294] reports that clear images of individual gold particles of 40 nm diameter have been obtained with confocal re¯ection scanning microscopy using a 63/1.4 oil immersion objective. Fluorescence Mode Although there is a wide choice of lasers on the market, most of the earlier models of CLSMs came equipped with an argon ion laser. This has generally been replaced by a krypton/argon ion laser offering better choice of

Microscopy

175

excitation wavelengths. Although this mode usually requires the sample to be stained with a ¯uorescent dye, in some cases specimens may have enough auto¯uorescence to produce a reasonable image without the need for staining. This is particularly true of fresh plant material and lignin-rich ®bers such as those of mechanical pulps. As with the re¯ection mode, the ¯uorescence mode can be used for localization of ¯uorescently labeled subresolution particles. Although the particles themselves cannot be resolved, their location can be detected by the emitted ¯uorescence. Confocal ¯uorescence microscopy overcomes the problems associated with conventional epi¯uorescence microscopy where diffuse light originates from structures above and below the focal plane. The ¯uorescence mode is to be preferred for acquiring clear cross-sectional images of ®bers in the xz plane. However, in order that the structure of the sample is not affected by the wetting incurred during staining, experiments should be planned so that specimens are stained at the beginning of any beating or sheetmaking processes and are therefore ready for examination at any subsequent stage if required. Conventional Transmitted Light Mode It is also possible to collect non-confocal transmitted light images by using the laser if the specimen is suf®ciently thin or translucent. In this case laser light passing through the specimen and condenser is collected by an optic ®ber placed below the condenser and goes directly to a second detector, thus bypassing the detection pinhole. It is often very useful to be able to compare the area seen in the confocal image with the same area seen in the familiar transmitted light micrograph. Dual (or Multiple) Wavelength Imaging If there is more than one photodetector (and modern instruments may have up to four detectors), then it is possible, by the use of bandpass ®lters, to obtain images in different modes simultaneously from the same scanning spot. Images generated by the same scanning spot are in exact spatial register and can therefore be superimposed to give a complete picture. The ¯uorescence and transmission modes have proved to be a particularly useful combination. Similarly, two ¯uorescence images can be obtained simultaneously from a specimen containing structures that have been speci®cally stained with different ¯uorescent dyes. Use of appropriate ®lters separates the emission wavelengths, channeling each to a different detector. This particular technique of double, or even triple, staining is widely employed by cell biologists [120,174], though its application to the study of ®bers is very limited, for reasons that will be discussed.

C.

Specimen Preparation Techniques

Specimen preparation is the same as for normal light microscopy and therefore relatively simple. Specimens may require staining, and, of course, large samples may need to be cut down to a size that will ®t onto the glass microscope slide. Slide Preparation Wet ®bers, or drops of ®nes material in suspension, are mounted on a microscope slide in the normal way and covered with a glass cover slip. Very small amounts of material are being examined, so it is important that

176

Moss and Groom

samples be taken from a well-mixed pulp suspension to ensure that they are as representative as possible. Cover slips must be ®rmly secured to the slide so that the sample does not move when the microscope stage is moved during sectioning. Securing can be done simply with strips of adhesive tape, or the cover slip can be sealed in place with rubber solution or nail polish. Care should be taken to ensure that the pressure of the cover slip does not cause mechanical damage or deformation of the sample. Spacers can be used for this purpose, but they have to be of just the right size because for high magni®cation work the specimen needs to be as close to the cover slip as possible, because of limitations in working distance. Likewise, the microscope slide must be clamped ®rmly in position on the microscope stage. Optical sectioning in the xz plane involves rapid movement of the microscope stage in the z direction; therefore, failure to secure samples ®rmly enough will result in blurred or stretched images. This is particularly important when oil immersion objectives are being used because surface tension forces may be strong enough to lift the cover slip from the microscope slide or even lift the whole slide if it is not satisfactorily anchored to the stage. Generally, for investigations into the morphological structure of ®bers and ®nes material, the best images are obtained if the ®bers are stained with a ¯uorescent stain and examined in a wet state while ®brillar material is spread out. Certain types of paper and board, such as release paper, polyethylene-coated board, coated papers, and papers containing ®llers, may contain pigments and ®llers that have ¯uorescent properties and therefore do not require staining. In fact, some, such as polyethylenecoated board, may be impervious to staining, whereas for others, such as coated papers, staining is impractical because the application of an aqueous stain will disturb the structure of both the coating layer and the underlying base paper. If dry specimens are to be examined with an oil immersion objective, they should, ideally, be mounted either in immersion oil or in some commercial mounting medium of the same refractive index as the oil. (See Section IV.F.) Staining If staining is required it can usually be done quickly and easily. The stains are usually made up as aqueous solutions (1±2% w/v). Wet pulp and paper samples can be immersed directly in a few drops of stain and then rinsed in distilled water to remove excess stain; dry samples may be wetted with distilled water before staining. Suspensions of ®bers or ®nes material and starch solutions can be stained by carefully stirring stain into the suspension a drop at a time, because it is not easy to wash off excess stain without losing some of the material. Much biological light microscopy is carried out on specimens of low intrinsic contrast, and staining is needed to increase overall contrast without much concern for staining selectivity. But, more importantly, microscopy is a way of gaining information about small structures. It can tell us what they are, where they are, and how many there are or how much there is. A stain, therefore, is any visual label attached to the specimen that will help answer these questionsÐdyes, ¯uorochromes, metal ions, labeled antibodies, and enzyme substrates. Staining may be selective. Speci®c stains are those that have an af®nity for particular molecules or substances, thus enabling easy identi®cation and localization in a specimen of entities of a particular chemical nature. Staining is widely used for chemical analysis. Microscopical histochemistry (i.e., the localization and identi®cation of substances and enzyme activity within cells and tissues) was pioneered by the

Microscopy

177

French botanist Raspail as long ago as 1825, and histochemical staining techniques are routinely used by biologists to identify tissues and to make quantitative analyses. The standard test methods for ®ber analysis employ chemical analysis in conjunction with light microscopy. Several stains are used to identify the types of pulp present in a sample. Fibers are identi®ed by the color they assume, thus making it possible to differentiate between ®bers of hardwoods and softwoods, chemical and mechanical pulps, bleached and unbleached pulps. Unfortunately, these methods cannot be used with the CLSM, because laser light is monochromatic and cannot show the color differences. Instead, ¯uorescent stains must be used. Fluorochromes, the stains most commonly used in confocal laser scanning microscopy, are substances whose electrons are excited to a higher state by a particular wavelength of light. As the electrons return to their normal state they release energy in the form of light of a longer wavelength, and it is this emitted light that is detected. Some substances may be naturally ¯uorescent, creating primary ¯uorescence, but in most cases secondary ¯uorescence is also necessary. High image contrast is important for ¯uorescence microscopy; the structures of interest should be bright and the rest of the image dark. There are a large number of ¯uorochromes available for biological specimens (see Kasten [159]). Acridine orange appears to have been most widely used for both ®ber and paper samples and has proved to be a good general stain. Congo red, auramine, rhodamine, and safranin have also been used successfully. Some textile dyes (e.g., Neolan saurerot E-XB and Maxilon brillant¯avin 10 GFF, Ciba-Geigy) and optical brighteners (e.g., Resolinbrillantgelb 10 GN, Bayer) are also suitable, and further experimentation with such dyes is recommended. Fluorescently labeled dextrans can also be used. Dextrans are water-soluble polysaccharides of high molecular weight. They are commercially available in fractions of different molecular weights and have been used traditionally as inert colloids in perfusion studies because they are not absorbed signi®cantly by ®bers. Stone and Scallan [309] and Scallan [287] used dextran molecules in their solute exclusion experiments to determine the size of pores in cell walls. When labeled dextran is used to stain ®ber suspensions or sheets for ¯uorescence microscopy, a negative image is obtained, the inter®ber spaces giving a bright image and the ®bers remaining dark. Photobleaching can be a problem in ¯uorescence microscopy. Long excitation periods cause a ¯uorochrome to decompose, creating photobleaching, which results in a loss in intensity in the area being examined. The problem is particularly noticeable with acridine orange but can usually be overcome by using a very low concentration of the stain [145]. However, photobleaching can also be used to advantage, and ¯uorescence recovery after photobleaching (FRAP or FPR) has become a useful technique for studying the transport of ¯uorescently labeled substances in living cells. The laser light bleaches the ¯uorescent label in the small area being examined so that it no longer gives an intensity reading, but, because there is a constant through-¯ow of the labeled substance, continuous scanning of the same area shows when the intensity rises again as the photobleached material is replaced. Blonk et al. [24] describe methods for measuring the mobility of labeled particles in liquid systems and for calculating diffusion coef®cients. The choice of stain is governed by the excitation wavelengths of the laser and its speci®city for a particular specimen. Consequently, the use of CLSM for chemical analysis of pulp and paper samples imposes certain limitations. Most of the earlier

178

Moss and Groom

CLSMs were ®tted with argon ion lasers, which have excitation wavelengths of 488 and 514 nm. Nowadays krypton/argon ion lasers are more common. These have excitation wavelengths of 488, 568, and 647 nm and thus offer a little more scope. Helium/neon lasers are preferred by some researchers because they are lower powered and heating effects are therefore reduced [223]. Paper scientists are further limited in their choice of stain by the chemical composition of the material they are examining, mainly nonliving plant cells, and although there is a long history of botanical histochemistry, little quantitative work has been done with cell wall carbohydrates despite their importance in the cell wall. Ligning, cellulose, hemicelluloses, and pectins are the main constituents of the cell walls of wood. Lignins are fairly easy to deal with, because there are some stains, e.g., Schiff's reagent [152], that will preferentially stain lignin, and they also exhibit primary ¯uorescence in UV and blue light. It is the cellulose and hemicelluloses that present the dif®culties; there are many of them with considerable chemical similarity and, more important, they rarely occur singly. The cell wall contains large numbers of complex polysaccharides, and it is dif®cult to ®nd a stain speci®c to a particular polysaccharide. However, some speci®c labeling techniques have been developed for TEM studies (see Section III.B). Jackson et al. [139] describe methods for the preparation of enzyme±gold complexes enhanced with silver for observing the effects of cellulase attachment to secondary ®bers of a bleached softwood pulp with an optical microscope. Also, immunolabeling techniques are available for determining the presence and location of speci®c polysaccharides. Monoclonal antibodies are highly speci®c probes that can be directed against a range of substances. The application of this technique to the study of pulp ®bers is still in its early stages, and there are some problems. The antibodies are not always as speci®c as could be desired and may bind to other polysaccharides. Lackner et al. [180] employed the silver staining technique to enhance immunogold-labeled ligninases for optical microscopy. However, any labeling techniques using ¯uorochromes or metals can be used for confocal microscopy. A detailed account of the use of confocal re¯ection microscopy for imaging immunogold labels is given by Cogswell [48]. D.

Fluorescence Quanti®cation

Quantitative ¯uorescence analysis is widely used in cell biology. The amount of ¯uorescence emitted from speci®cally stained structures is related to the amount of the particular substance present. A densitometric program can display the distribution of ¯uorescence intensity in speci®cally labeled structures along a selected line, and the volume of ¯uorescent material can be determined. However, as far as is known, this has not yet been applied to pulp and paper research though it could be employed in conjunction with speci®c staining and immunolabeling techniques. E.

Image Acquisition and Processing

Reduction of signal noise is important both for image quality and for accuracy of subsequent measurements. Filtering can be achieved during image acquisition either by frame averaging or by scanning in a line accumulation mode, and the user can observe the improvement of the image quality during its acquisition. The noisier the

Microscopy

179

image, the greater the number of times it will have to be scanned. A ¯uorescently stained specimen should give a clean image, whereas an unstained or weakly auto¯uorescent sample will be noisy and may require a large number of scans. Raw confocal images are often less than perfect, but it is possible to enhance them by computational processing. Electronic contrast facilities enable observation of very ®ne structural detail in both weakly stained and unstained samples. This can be achieved with a variety of standard image analysis techniques. For example, linear ®lters can enhance the visual appearance of an image by smoothing or accentuating the edges of structures. There are also many logical and arithmetical image processing operations that can be used to increase contrast or enhance certain features. Laser light is monochromatic; therefore images are black and white, so pseudocolor is used. Because the human eye can perceive many more colors than gray levels, the use of pseudocolor may be considered a further technique for image enhancement. Once a series of images (or optical sections) has been collected it can be processed in different ways to make a three-dimensional representation of the image. A series of sections can even be viewed rapidly as a movie to give a more realistic idea of the depth of a sample as well as to show the relative positions of different structures. Although the principles of confocal imaging and methods of image collection are common to all makes of CLSMs, there will be differences in the methods of processing data, depending on the software used. Cox [56] gives details of data storage and of various systems for obtaining hard copies of images. Three-Dimensional Reconstruction Most CLSMs are supplied with basic image processing software and suf®cient power to reconstruct a 3D image. White [352] gives examples and details of some of the multidimensional imaging systems on the market. Each set of consecutive optical sections is in spatial register and can readily be superimposed to construct the three-dimensional structure. This can be done in a number of ways. Extended Focus Image An extended depth of focus image is produced by simply adding all the sections together and scaling down the intensity. Features hidden beneath overlying structures are revealed by this method. Various algorithms can be used that take into account the intensity values of pixels in the z direction through the image stack and thus emphasize different features in the image. Simulated Fluorescence Process This method of 3D image reconstruction simulates the ¯uorescence that would be produced by an incident light source angled obliquely. The intensity level is corrected for each section in the image stack according to its depth and to preset absorption and emission levels. This produces a shadowed effect (not unlike that produced by a scanning electron microscope) that emphasizes structural features on the surface. Topographical Image A topographical image is color-coded for height; the uppermost structures are the brightest, and the deepest are the darkest. Voids are black. This application takes advantage of the fact that as the object passes through focus, the image intensity shows a sharp maximum. The x; y; z coordinates of every piont in the sample are determined automatically by through-focusing, and the focal level where maximum intensity is recorded is determined for each pixel. It is from this information that the topographical image is derived, showing the relative positions of structures within the volume sectioned. Points of equal height on the topographi-

180

Moss and Groom

cal image can then be joined to produce a cartographic representation of the height structures (i.e., a contour map), or a three-dimensional surface plot can be produced to show the surface roughness pro®le. Stereo Images All the sections in the image stack are sequentially shifted by a constant number of pixels to the left or right and then combined pixel by pixel to generate a black-and-white stereo pair of images. A pair of images can be combined to give an anaglyph (two-color) stereo pair. Such images can be viewed directly in three dimensions by using special spectacles and give an excellent impression of depth and spatial structure. Figure 8 shows some of the different images, described above, that have been generated from the same data stack. Martinez-Nistal et al. [203] give some nice examples of 3D reconstructions of hardwood pulp ®bers. Rotation A stack of optical sections can be rotated about any, or all, of its axes so that the sample can be viewed from different angles. The rotation program calculates a series of projection images of the data stack calculated at different tilt angles, with rotation and projection being performed simultaneously. Tilt angles are ®rst selected and set. The best results will be obtained from a series of images having a small step size. Figure 9a shows CTMP ®bers constructed from a stack of 16 sections collected in the xz plane. The same image stack rotated through 208 about two of the axes (Fig. 9b) reveals a pit in a ®ber (top left-hand corner) not visible in Fig. 9a. It is also possible to cut an optical slice at any angle through the image stack. This enables cross-sectional (i.e., xz plane) views to be made through a stack of images collected in the xy plane if it is not possible to scan in the xz plane.

F.

Artifacts of CLSM Imaging

Artifacts must be avoided. Therefore, when a specimen requires some special preparation it is necessary to check the sample for any changes that might have been induced by the preparation technique. One way is to measure the specimen before and after treatment. Stelzer and Bacallao [307] suggest that, for ¯uorescence microscopy, an evaluation of the quality of the preparation procedure should be based on images recorded by conventional ¯uorescence microscopy; images recorded with a conventional ¯uorescence microscope using an objective of high N.A. (>1.25) will provide a standard against which images derived from a confocal ¯uorescence microscope may be tested. Diatoms have been traditionally used for testing the resolution of microscopes. Wilke [354] used the diatom Amphipleura pellucida recorded at a wavelength of 488 nm to show periodic structure resolved with good contrast using a Zeiss PlanApochromat 63/1.4 objective. Tsien and Bacskai [337] mixed small quantities of a commercial preparation of diatomaceous earth in solutions of ¯uorescent dyes. They used different ¯uorochromes in different solvents of varying concentrations to compare the effect of varying wavelength, signal strength, and refractive index. Aberrations are inherent in all optical microscopes. These may be optical (due to the design of various components), chromatic, or geometric (due to diffraction phenomena). Image ®delity can be affected by curvature-of-®eld effects. This produces distortion in 3D reconstruction; therefore, plan or ¯at®eld (F) objectives should be used. The thickness of the glass cover slip (normally 170 mm) is also very important, because it has a severe impact on xz resolution. Aberrations in

Microscopy

181

Fig. 8 The series of optical sections shown in Fig. 4 processed in different ways. (a) An extended focus image; (b) an extended focus image where weighting has been given to maximum intensity values; (c) a simulated ¯uorescence image; (d) a combination of a stereo image and simulated ¯uorescence process; (e) a topographical image; and (f) the surface pro®le constructed from it.

182

Moss and Groom

Fig. 9 (a) An extended focus image of a series of optical sections collected in the xz plane. (b)±(d). Views from different angles when the image stack is rotated 208 through the x and z axes, which reveal that the bright ®ber in the top left-hand corner of (a) has a bordered pit.

cross-sectional images collected in the xz plane are probably the main potential source of error in the analysis of confocal scanning microscopy images. Therefore images must be calibrated against a standard of known dimensions before any measurements are made. If the refractive index of the immersion medium for the objective does not match that of the mounting medium (see ``Choice of Objective,'' Section IV.A), then spherical aberrations will distort the size and shape of the xz plane image. A mismatch of media produces signi®cant distortion in the depth measurements, but lateral measurements are not affected. Such aberrations are described in more detail by Hell et al. [121] and Visser and Oud [340]. Consequently, careful calibration of the xz plane images is necessary. Fluorescently labeled latex spheres, or man-made ®bers, of known dimensions can be used, and a few can easily be included with pulp or sheet samples in order to check image shape in the z direction. Figure 10 illustrates the differences in apparent size and shape of cross-sectional images result-

Microscopy

183

Fig. 10 A cross section through a polyamide ®ber showing the spherical aberrations caused by mismatch of media.

ing from the use of different mounting media. Polyamide ®bers, which are hollow with a circular cross section of 20 mm diameter, have proved to be very suitable for calibration purposes. The effects of using a 40/0.7 dry objective to examine a polyamide ®ber mounted in air (top), water (middle), and oil (bottom) are shown on the left-hand side of Fig. 10, and the corresponding effects produced by using a 40/1.3 oil immersion objective to examine a ®ber are shown on the right-hand side. This shows clearly that the optimal conditions are obtained by using an oil immersion objective to examine a specimen mounted in a medium of the same refractive index. When in situ monitoring is being carried out under changing conditions, the refractive indices of the objective and mounting medium will not always match. This is a common problem with studies involving measurement of ®ber shape and crosssectional dimensions during drying. Distortion of cross-sectional images also occurs if an objective of low N.A. is used, and topographical images constructed from images collected with such an objective will give inaccurate height measurements. A step-height standard can be used for checking the height discrimination of objectives. Alternatively, a piece of paper of known thickness but translucent and thin enough for the laser to penetrate fully can be used. Lens cleaning tissue, and similar thin tissues, are suitable for this purpose. Experiments were undertaken to compare measurements obtained using a 40/0.7, a 16/0.45, and a 10/0.3 objective. A piece of lens cleaning tissue with a mean thickness of 39 mm was optically sectioned in both the xy and xz planes. Topographical images were constructed and height measurements made, and the

184

Moss and Groom

thickness of cross-sectional images was measured. The correct thickness was measured using the 40 objective, but mean thicknesses of 53 mm and 88 mm were recorded for the 16 and 10 objectives, respectively. Incomplete cross-sectional images are a problem with ¯uorescence microscopy. The signal weakens with depth, leading to a loss of intensity that is particularly noticeable in images scanned in the xz plane. Self-shadowing can also lead to loss of information in the image, which presents problems in image analysis because the lower half of a thick object is not clearly de®ned. Many drying experiments have been undertaken in which the starting point has been a wet-pressed sheet that is relatively dry. Under these circumstances a dry objective has been used, but, due to its low N.A., a complete cross section through a ®ber cannot always be obtained. In such cases the cross-sectional shape of the dried ®ber can be veri®ed by in®ltrating the sample with oil and examining the ®ber again using an oil immersion objective. This is easily done by pipetting oil along the edge of the glass cover slip and letting it soak in slowly so as not to disturb the structure of the sheet. Figure 11 shows the cross sections of the same TMP ®berÐwet pressed (top) and dried (center) imaged with a 40/0.7 dry objective, and after soaking in oil (bottom) imaged with a 40/ 1.3 oil immersion objective. Such images provide a great deal of information about changes that have occurred during drying but are clearly unacceptable for measurement purposes, and no method of image analysis could have determined the true shape of the lower side of this ®ber. However, in Fig. 12 the image of the dried ®ber immersed in oil (bottom) veri®es that the wet-pressed (top) and dried (center) ®ber cross sections are complete and can be measured accurately.

Fig. 11 A TMP ®ber cross section examined with a dry objective in the wet-pressed state (top) and after drying (middle). In®ltration of the sample with oil and use of an oil immersion objective (bottom) reveals that the cross sections are not suitable for measurement.

Microscopy

185

Fig. 12 A TMP ®ber cross section examined with a dry objective in the wet-pressed state (top) and after drying (middle). In®ltration of the sample with oil and use of an oil immersion objective (bottom) con®rms that the cross sections are acceptable for measurement.

G.

Applications of the CLSM to Pulp and Paper Research

Applications of the CLSM are many and varied, and almost certainly the full potential of the CLSM for studying pulp and paper has yet to be realized. Its unique feature is its nondestructive optical sectioning in the z direction, enabling the crosssectional appearance of beaten, wet ®bers to be examined readily and measurements made of their transverse dimensions. For many of the applications discussed below only cursory investigations have been made to evaluate the performance of the CLSM, but in other cases extensive studies have been undertaken. Raw Materials and Basic Fiber Properties There is a great deal of interest in identifying and characterizing the basic properties of ®bers from different sources, not only of wood but also of non-wood plants, which are being increasingly looked to as a papermaking resource. The CLSM can be used for investigating the structure and ligni®cation of various plants. Knebel and Schnepf [175] looked at fresh and ®xed stems of Clematis vitalba and Pinus sylvestris stained with Schiff's reagent to reveal differences in ligni®cation. Studies of lignin distribution across cell walls, previously undertaken with UV microscopy (see Section III.A), can be readily carried out using a CLSM without the need for embedding and sectioning techniques. Figure 13 shows lignin distribution across summerwood ®bers in a sample of Douglas ®r wood that had been stained with acridine orange. Mislankar et al. [212] used a CLSM to examine the distribution of orthoquinones in different regions of wood samples and in high yield pulps. The orthoquinones, believed to be responsible for photoyellowing in mechanical pulps,

186

Moss and Groom

Fig. 13 A cross-sectional image of Douglas ®r wood showing lignin distribution across the cell wall.

were labeled with o-phenylenediamine to produce ¯uorescent phenazine derivatives that could be clearly detected in cell corners, middle lamellae, and tori. The CLSM has been used to study the distribution of lignin in ®ber cell walls and to monitor changes with bleaching [218]. The embedding, grinding, and polishing technique described in Section III.B was used to obtain clear cross sections of handsheets made of unbleached and bleached kraft pulp ®bers. The intensity of ¯uorescence was assumed to correlate to lignin concentration, and this embedding technique was employed to eliminate the loss of intensity with depth that occurs when ®bers are optically sectioned in the xz plane. Lignin exhibits primary ¯uorescence at 488 nm; therefore the ®bers did not require staining. Commercial preparations of glucose, cellulose, xylose, and mannose, examined under the same microscopic operating conditions as for lignin, gave negligible ¯uorescence emission signals, thus verifying that the ¯uorescence was emanating from lignin components of the cell wall. Figure 14 shows the distribution of lignin in the ®ber walls of (a) an unbleached pulp (kappa number 25.4), (b) an oxygen-deligni®ed pulp (kappa number 16.9), and (c) a TCF-bleached pulp (kappa number 4.7). Cross-sectional images of a fully bleached pulp (not shown) were barely discernible. The CLSM lends itself well to the study of ®ber morphology and the characterization of ®ber types. Basic ®ber proprerties, such as length, width, and curl, are most readily determined by using existing methods. However, since the CLSM can be used to generate cross-sectional images of ®bers, it offers better possibilities for measuring width and coarseness (see Jang et al. [144]). Confocal microscopy has also been used to monitor the structural changes that occur when pulp ®bers are subjected to cyclical loading. Hamad and Provan [112] observed the propagation of

Microscopy

187

Fig. 14 Differences in lignin content of ®bers can be measured using the CLSM in ¯uorescence mode. (a) Unbleached pine kraft pulp, kappa 25.4; (b) oxygen deligni®ed pulp, kappa 16.9; (c) TCF bleached pulp, kappa 4.7.

cracks during straining of individual ®bers. Groom et al. [109] used a CLSM to determine the cross-sectional area of individual ®bers that had failed during tensile testing in the ESEM. The CLSM can be used to measure ®bril angle of wood pulp ®bers using the classical crossed polaroid technique. This overcomes the problems encountered in measuring the ®bril angle of whole ®bers using a conventional polarizing microscope (Section III.A), because with the optical sectioning capability of the CLSM only the wall layer of interest is focused. Verbelen and Stickens [339], using the CLSM in ¯uorescence mode, inserted a polarizing ®lter between the laser and the scanning unit of the microscope to determine the ®bril orientation in the walls of leaf cells cultured from Nicotiana tabacum. The cells were stained with Congo red. Jang [141] used both Congo red and acridine orange stains to measure ®bril angle in the softwood and hardwood ®bers of mechanical, chemical, and bleached pulps. Batchelor et al. [14] describe their method of measuring the ®bril angle of the S2 wall layer in wood ®bers using a CLSM in re¯ection mode.

188

Moss and Groom

Pulp Characterization and the Effects of Beating The CLSM is particularly useful for characterizing pulps and examining various structural changes brought about by pulping and re®ning processes. Wet samples can be observed in their natural state before ®brillar material has dried down onto ®ber surfaces. Xu et al. [365] used a CLSM to measure the wall thickness of mixed eucalypt kraft ®bers. Jang et al. [142] characterized the collapse behavior of mechanical ®bers produced under different re®ning conditions, and Sundin et al. [315] determined the collapse and coarseness of mechanical ®bers in laboratory handsheets. Similarly, the effects of recycling can be examined and quanti®ed. Jang et al. [143] used confocal microscopy to study the effects of laboratory recycling on the transverse dimensions of ®bers, evaluating the degree of collapse resulting from the recycling of mechanical and kraft pulp ®bers. Seth et al. [290] measured and compared the transverse dimensions of kraft pulp ®bers obtained from a variety of tree species. This information allows selection of ®ber sources for properties most appropriate for various end uses. Nanko et al. [223] undertook a carefully controlled experiment to measure shrinkage of individual ®bers using an SLM. The microscopy was carried out under controlled temperature and humidity, and a low power helium/neon laser was used to minimize heating effects. Fiber measurements were made at points where two ®bers crossed and also in free segments of ®bers. Fiber cross-sectional area measurements were made using the surface pro®le measurement system of the SLM. A real-time sample-weighing system was used to obtain a wet weight corresponding to each image collected, so that changes in ®ber width and cross section could be correlated with moisture content. The experiments undertaken by Page and Tydeman [243] to monitor controlled drying of ®bers using the porous plate apparatus (Section III.A) could perhaps be adapted for use with a CLSM to control the removal of water from a wet web during drying. This would allow easy observation of cross-sectional views of bond formation in ®bers. External Fibrillation External ®brillation is an important effect of beating, and the CLSM has been used to characterize external ®brillation of different types of ®bers and the effects of different laboratory beaters [370]. Fibers are best examined in a wet state when ®brillar material and peeling sheets of wall lamellae are still spread out. Fibers need to be ¯uorescently stained in order for the ®ne ®brillar material to be identi®ed, but even then this material can prove dif®cult to detect simply because much of it is too ®ne to take up enough stain to give a suf®ciently strong signal. However, it is possible to obtain good images, and image enhancement techniques can be used to bring up details of the ®ner structures. Ferro¯uids (Ferro¯uidics Corp., Nashua, NH) have been used to enhance ®brillation. Ferro¯uids are colloidal suspensions of very ®ne particles of magnetite (average particle size 10 nm) that when added to a suspension of ®brillated ®bers appear to bond to ®ne ®brillar and colloidal material. A sample of highly beaten kraft ®nes was stained with acridine orange, and a little of the ferro¯uids suspension was then stirred into the ®bers. A drop of this mixture was mounted on a microscope slide and examined wet. The sample was scanned simultaneously in the re¯ection and ¯uorescence modes. The re¯ection image (Fig. 15a) indicates that there is a great deal of ®brillar material to which the magnetite particles have bonded. The ¯uorescence image (Fig. 15b) shows the basic ®ber shape, although thick aggregations of particles have prevented detection of the ®ber surface in some areas.

Microscopy

189

Fig. 15 (a) A confocal re¯ection image of a ¯uorescently stained beaten kraft softwood ®ber shows nanoparticles of magnetite binding to ®brillar and colloidal material. (b) The ¯uorescence image reveals the basic ®ber shape.

190

Moss and Groom

Image analysis can be used to quantify ®brillation. Laamanen et al. [179] obtained a ®brillation index for ®bers dried down onto glass microscope slides, and the method described by Klofta and Miller [172] for obtaining a ®brillation index from scanning electronmicrographs of freeze-dried ®bers could well be applied to CLSM images of wet ®bers. Internal Fibrillation (or Cell Wall Delamination) Another important effect of beating, but one not as easy to study as external ®brillation and more dif®cult to quantify because of the scale involved, is internal ®brillation. Stone and Scallan's model (see Section III.B) describes internal ®brillation on a microscale in which the pores are only a few nanometers in diameter. It has been suggested that although these are important in effecting local plasticity in the cell wall, it is the larger visible splits in the cell wall that are important for increasing plastic ¯exibility in ®bers [78,236]. Internal ®brillation has been extensively studied, because it is believed by many workers to be the most important effect of beating [1,138,164,209,332]. Page [236] drew attention to the large cracks that are seen in a delaminated cell wall and pointed out that no technique had been developed to measure these cracks. Quanti®cation of their surface area would be relevant to determining the relative importance of pores and cracks to ®ber ¯exibility and ®ber±®ber bonding. Figure 16 shows what is achievable with the CLSM. Compare these cross sections with those obtained with TEM (Fig. 1) and LTSEM (Fig. 31). Measurement of pores in the cell wall is clearly beyond the scope of optical microscopy, but the larger cracks that occur during cell wall delamination could perhaps be quanti®ed using a program such as that developed by Kibblewhite and Bailey [166].

Fig. 16 Cross sections of fully hydrated bleached birch kraft ®bers beaten to 4000 rev in a PFI mill.

Microscopy

191

Investigations into wet ®ber ¯exibility such as those described earlier (Section III.A) could be employed using the CLSM, and, in addition to measuring ®ber ¯exibility, the cross-sectional appearance of the ®bers could also be monitored as they dried. Abitz and Luner [1] concluded, after examining the effects of re®ning on wet ®ber ¯exibility and its relationship to sheet properties, that wet ®ber ¯exibility might not depend solely on the amount of internal ®brillation. This could be veri®ed using the CLSM in conjunction with Steadman's method. Lammi and Heikkurinen [181] studied the effect of re®ning on the stiffness of latewood and springwood ®bers and the relationship between ®ber stiffness and cross-sectional shape. Fiber ¯exibility was measured using the method of Tam Doo and Kerekes [318]; then, for some of the pulp samples, a CLSM was used to obtain a series of 10 cross-sectional images of each ®ber after it had been de¯ected. From these Lammi and Heikkurinen determined cell wall thickness and calculated the moment of inertia. They concluded that ®ber stiffness cannot be predicted from the cross-sectional shape. Fines Material The importance of ®nes material is being increasingly recognized. The characteristics of a ®nes fraction vary with pulp type. Fines material also differs greatly according to whether it is from a mechanical or a chemical pulp, and further differences arise from subsequent re®ning processes. Image analysis can be used to characterize different types of ®nes on the basis of size and shape [193]. Fines vary in their chemical composition, and the distribution of lignin in mechanical pulp ®nes can be studied using dual-channel imaging in the CLSM. Mechanical ®nes are stained with acridine orange, and images are collected using the ¯uorescence mode and nonconfocal transmitted light. The transmitted light image shows all the structural components of the ®nes material, including ®ne ®brillar material, whereas the ¯uorescence image shows just the lignin-rich areas. The combined image shows the location of lignin on the ®nes particles, and the proportion of surface area that is lignin-rich can be quanti®ed. Luukko et al. [194], who used an optical microscope for measuring ®nes particles, assumed that light transmission attenuated exponentially in proportion to the thickness of the particles. They then tested this assumption using a CLSM to measure the cross-sectional thickness of particles and an optical microscope to take normal gray tone images of the same particles. The thickness values of the particles plotted against gray tone values showed a clear correlation. The effects of ®nes on sheet properties were investigated by Moss and Retulainen [219] and Retulainen et al. [266]. They found that additions of ®nes to the long ®bers of a TMP pulp produced handsheets of greater density and higher tensile strength properties than sheets made of just the long ®bers, and that the addition of kraft ®nes was even more effective. The CLSM was used to examine the cross-sectional appearance of potential bond sites in crossing ®bers, ®rst in wetpressed sheets and then in exactly the same areas after drying. This showed that ®nes were necessary to help wet-pressed ®bers maintain contact during drying and that better ®ber±®ber contact was achieved with kraft ®nes, shown in Fig. 17, than the bulkier TMP ®nes, as shown in Fig. 18. However, the bulky elements of TMP ®nes help to improve the optical properties by increasing light scattering. Distribution of Cell Wall Components Immunolabeling techniques, such as those described in Section IV.C, have been applied to investigations into the redeposition of xylans in kraft pulping using dual-channel scanning. Antibodies to xylanase were

192

Moss and Groom

Fig. 17 Secondary ®nes of a bleached pine kraft pulp beaten for 60 min in a Valley beater. (From Moss and Retulainen [219].)

Fig. 18

Primary ®nes of a spruce TMP. (From Moss and Retulainen [219].)

Microscopy

193

raised, labeled with acridine orange, and then conjugated with the xylanase. The enzyme±antibody complex was then added to birch kraft pulp ®bers and left to bind to xylans present in the kraft pulp. Images collected simultaneously in nonconfocal transmitted light and ¯uorescence mode were combined to show the location of the xylans. Fluorescently labeled cellulases have also been used in an attempt to elucidate the mechanism of enzymatic attack on ®bers. Sheet Structure The CLSM can be used to study network structure (pore volume, ®ber orientation, density, etc.) and to monitor changes that occur during sheet drying and the effects of humidity on network structure. Distribution of ®nes material and ®llers within a sheet and coating structures can also be examined. The CLSM can be used for examining surface roughness and is an excellent tool for height measurements and three-dimensional surface pro®ling. The effecs of calendering and wetting treatments can be monitored, and changes in sheet structure resulting from compression can be directly observed. Network Structure The nondestructive optical sectioning capability of the CLSM makes it eminently suitable for investigating network structure in both the xy and xz planes. However, there are limitations on the size of the area that can be sampled and the depth to which it can be sampled. The CLSM offers great potential as a means of assessing paper quality by direct measurement of ®ber orientation. SoÈremark et al. [302] used the CLSM to investigate the cause of ®ber orientation streaks affecting the quality of a kraft liner. Xu et al. [366] used confocal microscopy in conjunction with image analysis to evaluate ®ber orientation distribution in three dimensions. In-plane variation of ®ber distribution has important implications for paper properties such as printability and dimensional stability, and Xu et al. [367] employed the same techniques to determine ®ber distribution in the z direction. Naito et al. [222], in a study of the delamination resistance of paper, used the CLSM to determine the depth to which ®bers penetrate in the z direction of a sheet. Sheet density can be measured from images reconstructed from a series of sections, and sheet two-sidedness can also be investigated [220]. Distribution of surface pore area as a function of depth can be calculated from topographical images. Mangin and BeÂland [200] calculated surface pore areas of various newsprint samples and of a base stock (coated and uncoated) before and after calendering, to investigate structural changes induced by calendering. Lùvhaugen [190] also used the CLSM to make quantitative evaluations of pore volume of different unprinted newsprint samples, and Lùvhaugen et al. [191] related pore volumes of newsprint and uncoated liners measured by the CLSM to measurements made using the Parker Print-Surf and Bendtsen air leak instruments. Kunnas et al. [178] used the CLSM to investigate the effect of Condebelt drying on the structure of ®ber bonds. Under the high pressure Condebelt conditions, crossing ®bers appeared to weld together, with excess surface material (lignin and hemicellulose) ¯owing and ®lling the angles formed between the crossing ®bers. It was surmised that these bonds were better able to resist water than the thin, ®brillate bonds seen in the conventionally dried sheets. Nanko and Wu [227] studied the mechanisms of paper shrinkage during drying using a re¯ection-type SLM. They used silver grains as location markers

194

Moss and Groom

on ®ber surfaces and collected images of the sheet surface, before and after drying, to measure how much the ®bers had shrunk longitudinally. They also obtained surface pro®les along ®bers and examined the effects of restrained and free drying on the longitudinal shrinkage of ®bers. Optical sectioning in the xz plane allows easy monitoring of changes in drying sheets. Figure 19 shows an extended focus image and a cross section through a sample of handsheet at the just-formed stage (top) and after drying (bottom). Collapse of ®bers and their relative movement during drying are clearly indicated. Nanko and Tada [226] cut cross sections of sheets with a razor and used the SLM to observe deformation of ®bers due to changes in relative humidity. Samples were placed in a chamber into which the objective of the microscope could be inserted, and humidity-controlled air was supplied. They measured cross-sectional and longitudinal hygroexpansion of ®bers in the sheets. Mangin et al. [201] developed a special piece of apparatus to apply static compressive stress to the paper surface while imaging with the CLSM. Pressures of up to 5 MPa were applied during in situ studies of changes in surface compressibility. They applied this technique both to the effects of calendering and to the evaluation of surface roughness and compressibility in relation to printing processes. Ting et al. [335] also designed and manufactured their own compression rig in order to characterize network changes in paper under compression in the z direction. Direct visual observation of the cross-sectional shapes and locations of the ®bers enabled them to study the effect of loading rate and ®ber cell wall thickness on compressibility of the sheet.

Fig. 19 Extended focus image and a cross section of a wet handsheet made of spruce TMP (top) and the same area after drying (bottom).

Microscopy

195

Many earlier microscopists developed sophisticated techniques for investigating various properties of pulp ®bers and paper, many of which could be adapted for use with the CLSM. These include, for example, control of humidity in drying experiments [243,348] and techniques used for studying porosity. A study of relative porosity in paper, similar to that described by Lindsay [188], could be usefully undertaken using confocal microscopy. Lindsay observed and measured the movement of a dye injected into paper samples. One of the dyes he employed was blue dextran 2000, but ¯uorescently labeled dextran solution could be used to investigate the movement of the stain in both the xy and xz planes of the sheet. The rate of ¯ow of a liquid through a sheet could be measured using the FRAP technique described earlier (Section IV.C). High magni®cation examination of ®ber cell walls, seen in cross-sectional view, could be made in solute exclusion studies. Experiments could also be designed for use in conjunction with compression apparatus, such as that developed by Ting et al. [335], to study sheet permeability and the effects of pressing on the movement of water. Carlsson et al. [40], for instance, used dextran solutions of different molecular weights to demonstrate from which cavities in the ®ber matrix water was expressed during pressing. Distribution of Fines and Fillers in Sheets The simultaneous generation of images in ¯uorescence and nonconfocal transmitted light modes (see earlier) can be used to examine the distribution of ®nes material within a ®brous network providing that the ®ber and ®nes fractions are suf®ciently different chemically. One component needs to be from a mechanical pulp and the other from a chemical pulp, to enable differentiation of the components [220]. The same technique can be applied to investigations into the distribution of ®llers, using the re¯ection mode with the nonconfocal transmitted light mode, or the re¯ection mode with the ¯uorescence mode if the ®llers have ¯uorescent properties. Coated Papers Coated papers offer a challenging area of application. Much research has been done on the distribution of coating thickness that involved embedding sheets in resin and cutting thin sections (see Sections III.A and III.B). Crosssectional views of coated sheets can be obtained with the CLSM, but only to a limited depth due to the opaqueness of the coating layer. However, as shown in Fig. 20, the interface of the coating layer with the base paper can be clearly discerned. Measurements of variation in coating can be readily obtained from such images, and the void volume, critical to coating processes, can be calculated. Clearer cross-sectional images are obtained if the sample is stained with a ¯uorescent dye, but the application of an aqueous stain would not only affect the coating itself but also disrupt the structure of the base paper. This can be overcome if the sample is stained during sheetmaking. However, unstained samples can be examined quite successfully using the ¯uorescence mode, because coating layers usually include optical brighteners or other pigments that have ¯uorescent properties, and the base papers are often made of lignin-rich mechanical pulps that are inherently ¯uorescent. Polyethylene coatings, such as those used on drink cartons, and the adhesive layer of labels (see Fig. 21) can also be examined. Lange et al. [182] used several microscopical techniques, including conventional optical microscopy, CLSM, SEM, and AFM, to examine defects in solvent-free organic coatings that require rapid curing time. The AFM and CLSM both proved useful for characterizing coatings,

196

Moss and Groom

Fig. 20 Cross section through an LWC sheet showing the interface of the coating layer with the base paper.

Fig. 21

Cross section through the adhesive layer of a label.

Microscopy

197

but the CLSM was particularly useful in that the internal structure of transparent coatings could be viewed. Analysis of Surface Topography Surface roughness is an important property of paper in relation to printability, and the CLSM offers a non-contacting method of surface pro®lometry. This has certain advantages over conventional stylus instruments, which can cause damage to samples due to stylus pressure [83,341] (Chapter 11). This is even more true where there are raised or loose pieces of ®brillar or ®nes material, and the use of a laser overcomes these problems. The CLSM not only gives roughness values but also enables visual assessment of changes that occur at the ®ber level, because the same areas can be examined before and after treatment. The software for the Leica CLSM includes a program based on the international and DIN standards for the measurement of surface roughness parameters using stylus instruments, and such a program could be easily written for other makes of CLSMs that do not include it in the standard package. The Leica program calculates the arithmetic mean deviation from the pro®le, root-mean-square deviation from the pro®le, difference between maximum height and maximum depth of the pro®le, and skewness. Surface roughness measurements are made from topographical images constructed from a series of sections collected in the xy plane using the re¯ection mode. Samples should be dry, and a dry objective should be used. A 40 objective is recommended for such measurements. This shows a very small ®eld of view …125 mm  125 mm† with the Leica CLSM but is better able to discriminate height differences than objectives of smaller magni®cation, which have a greater depth of ®eld and consequently give falsely high values. Step size (see Section IV.A) should be kept as small as possible, and, in order to measure as accurately as possible, it is better to oversample than to undersample. There are no standards pertaining to image acquisition using a CLSM. This means there will be problems with reproducibility using different laser scanning microscopes. Experimentation is therefore necessary to determine the most suitable combination of microscope settings. Many variables (voltage, laser power, pinhole size, step size, objective, image format, the number of sections to be collected, and the number of scans needed to reduce noise) can affect the quality of topographical images. Optimal conditions may vary greatly for different paper surfaces, so these parameters should always be kept the same for any given paper so that meaningful comparisons can be made. Laser power is important and can affect the results of work being undertaken if there are ¯uctuations. Lùvhaugen [189] outlines some of these problems in relation to paper surface measurements. Because the area examined is very small, a very large number of areas would need to be sampled in order to acquire statistically meaningful data. However, if changes in roughness are being monitored and exactly the same areas are being re-examined after various treatments (as described in the following section), then a smaller sample size may suf®ce. Fjerdingen et al. [90] used a CLSM to investigate the effect of ®ber wall thickness on paper surface characteristics. The CLSM has been used by many workers to examine and measure changes in surface topography, some of which are detailed below. Monitoring the Effects of Calendering and Wetting Treatments Retulainen et al. [267] undertook a series of experiments to characterize the effects of calendering on the microstructure of the paper surface. They studied the effects of calendering

198

Moss and Groom

treatments applied to handsheets made of earlywood ®bers and to others made of latewood ®bers, with and without ®nes. Areas in an uncalendered paper sample were measured, then the same areas were measured again after calendering. The areas were randomly selected, but to facilitate location of areas for subsequent reexamination a few lines were lightly penciled onto the samples to serve as location markers. Roughness measurements were obtained and visual analysis was made of the topographical images and three-dimensional surface pro®les in order to see how calendering affected the surface structure. Figure 22 illustrates the surface pro®les of an uncalendered handsheet (a) and then exactly the same area after calendering treatments (b)±(d). After the ®rst calendering, ¯attening of ®bers and increases in ®ber width are observed. (Fig. 22b). Height differences are further diminished after the second calendering (Fig. 22c), but there is little change after the third calendering treatment (Fig. 22d), just some displacement and rearrangement of material in the lateral direction. Retulainen et al. [267] applied the same technique to study the effects of wetting printing papers, such as occurs during coating processes or in offset printing. A KRK printability tester was used for the dampening treatment. This tester simulates practical coating or printing processes and was used in these experiments to investigate the wetting behavior of calendered sheets and the effect of moisture on surface roughness. Roughness was measured, and images of the areas were examined for evidence of changes in ®ber dimensions, for ®ber swelling, ®ber rising, breaks at bonded areas, and redistribution of material. They found that in handsheets made of a furnish typically used for printing papers (a kraft pine whole pulp plus 60% spruce groundwood), the ®nal roughness and thickness of calendered sheets after wetting and redrying approached those of the uncalendered sheets. The higher areas on the surface were more compressed in calendering and more affected in subsequent rewetting. MacGregor [196] examined the roughening effect of water on the surfaces of untreated, water-treated, and printed papers and its relationship to reduction in gloss. In subsequent work [197] the CLSM was used to complement image analysis measurement of small-scale gloss variation in printed paper. Topographic data acquired with the CLSM were directly compared with gloss images of the same areas. Heat Conduction During Calendering A novel method was developed by Luong and Lindem [192] to determine the temperature distribution through the z direction of paper during calendering. Paper samples included monodispersed ¯uorescently stained latex spheres with an average diameter of 9:5 mm and a melting point of 1258C. Paper samples were calendered at various temperatures and speeds, and the calendered samples were then optically sectioned using a CLSM. The transient temperature distribution through the z direction of the sheet was determined by measuring the distances between the paper surface and the unplasticized spheres. Related Research Printing and Deinking BeÂland and Mangin [17] used the CLSM to study ink transfer mechanisms, comparing the same newsprint printed at high and low ink levels. They also examined the adhesion of toner to paper using re¯ection and ¯uorescence modes to record halftone dots before and after the samples were subjected to a peel test. Topographical images indicated whether halftone dots were situated on raised or depressed areas of the surface, whereas the ¯uorescence images showed the areas

Microscopy

199

of the dots. Images of the same areas taken after the peel test show clearly where toner had been detached from the surface. Christiani and Bristow [44], in a study of the drying mechanism of waterborne printing inks, used a CLSM to examine the diffusion of water within paper and observed a tendency for the water to follow the ®ber direction. Analysis of contaminants removed during recycling of waste paper is necessary for determining the ef®ciency of recycling equipment. The problem with characterizing toners and other three-dimensional particulate matter using conventional optical microscopy is that the size and shape of the particles cannot be fully ascertained from the two-dimensional binary images produced for image analysis. Borchardt et al. [25] found the CLSM to be a useful tool for studying the particle geometry of ink removed from laser printed and photocopied papers. It was thought that because the ink particles are three-dimensional, image analysis of particle length dimensions might not give a suf®ciently accurate measure of the actual mass of ink in particles. They used stereo images and simulated rotation to view the morphology of ink particles. Synthetic Polymers Furuta et al. [97] studied the structure of cellophane using both an SEM and a CLSM. They found the CLSM to be demonstrably more useful in the determination of three-dimensional structure. A major dif®culty with investigating the interface regions of heterogeneous materials is that the interface cannot be separated from the sample but must be studied in situ. Thomason and Knoester [333] used both polarized light microscopy and CSOM for in situ investigations of the interface region of the polymer composites. A series of optical sections taken through an aramid ®ber in polypropylene revealed structures not seen in polarized light. They also studied glass ®ber reinforced epoxy composites to see whether failure cracks in samples had propagated along the ®ber/matrix interface or through the matrix. An advantage of using CSOM was that samples required no further preparation other than cutting; also, the roughness of the cut surface disappeared at a depth of 5:3 mm and ®bers were clearly discernible. Previous methods of preparation had included polishing the cut surface, which often resulted in cracks getting ®lled in. Microbial Contamination of Pulp and Paper There is a great deal of interest in cellulose-related products, mainly because of the biodegradability of cellulose. Plastic ®lms used in food packaging have to be resistant to bacteria and fungi. Samples of ®lm made of discrete layers of starch and polyvinyl alcohol were incubated in soil for several days. They were then lightly stained with acridine orange and examined with the CLSM. Various species of bacteria and fungi were detected. Figure 23a shows the rami®cation of fungal hyphae at a depth of 35 mm from the surface, and scanning in the xz plane revealed that the fungi appeared to have a preference for speci®c layers, as shown in Fig. 23b. Microorganisms exist everywhere and are responsible for decomposition of both organic and inorganic substances. They cause decay in the wood used for production of pulp and are present in all systems in pulp and paper, producing damaging slime deposits. Work has been done to investigate the growth of bacteria on the metal surfaces of headboxes and also on the wires of paper machines and in the recirculating white water, which all provide ideal conditions for the growth of bacteria and fungi. Robertson and Taylor [275] describe their techniques for examining and evaluating bacterial bio®lms and bio®lms containing particulate calcium carbonate by using CLSM and ESEM. Bacterial ¯ocs from paper mill ef¯uent have

200

Moss and Groom

Fig. 22 The surface pro®le constructed from the topographical image (a) of an uncalendered handsheet made of summerwood ®bers of a pine kraft pulp and (b)±(d) of exactly the same area after repeated calendering treatments.

Microscopy

201

202

Moss and Groom

Fig. 23 (a) A single optical section taken at a depth of 35 mm below the surface of food packaging ®lm showing fungal hyphae and (b) a cross section showing that the fungal hyphae tend to be con®ned to discrete layers. The protuberance seen in (b) is a fungal spore that prevents the penetration of light and is the cause of one of the apparent voids seen in (a).

Microscopy

203

Fig. 24 Bacteria deep inside the base board of a polyethylene-coated milk carton. (Photo courtesy of I. Suominen.)

also been examined and analyzed, the size of the ¯ocs being an important factor in how quickly the bacteria break down the sludge. Suominen et al. [316] examined the growth and migration of bacteria in foodpackaging paperboards. Their study showed that microbially contaminated starches used as surface sizes and contaminated mineral pigment coatings posed a threat to hygiene, because the polythene layer separating the food from the paperboard displayed pores and indigenous damage. Base board used for milk cartons was also examined for bacterial contamination. The bacteria shown in Fig. 24 were located quite deeply in the base board and would not have been possible to detect directly using any other microscopical technique. Studies of deligni®cation are becoming of increasing environmental concern due to the rapid depletion of land®ll space. The degradation rate of laminated chip boards such as those used in modern furniture has been studied using compost heaps to simulate rubbish dumps. Fungi found in ligni®ed samples were isolated and cultured for identi®cation.

V.

THE LOW TEMPERATURE SCANNING ELECTRON MICROSCOPE

FernaÂndez-MoraÂn [88] reviewed early investigations into low temperature techniques for electron microscopy, but low temperature scanning electron microscopical (LTSEM) (also referred to as cryogenic or cryo-SEM) techniques were ®rst demonstrated for biological specimens by Echlin et al. [76]. A comprehensive account of

204

Moss and Groom

low temperature scanning electron microscopy is given by Beckett and Read [15], and de Silveira et al. [66] discuss its application to the analysis of pulp and paper. Until the advent of LTSEM, and later the ESEM, only fully dehydrated specimens could be examined satisfactorily in a scanning electron microscope because the electron beam is strongly scattered by air or water molecules. Unlike the conventional SEM, which entails rather elaborate and lengthy preparation for the examination of wet pulp and sheet samples, a low temperature stage makes it possible to examine fully hydrated material within a very short timeÐa specimen can be frozen, sublimed, coated, and examined in less than an hour. It also enables the examination of many previously ``impossible'' samples such as emulsions, oils, and inks. Material prepared in this way for examination with the LTSEM is usually referred to as frozen hydrated (FH). Low temperature scanning electron microscopy, which had been widely used in biological sciences for many years, was ®rst applied to the study of papermaking ®bers by Howard and Shef®eld [130] in 1987. Because the conversion of wood to paper is carried out almost entirely in water, the interactions of ®bers with water are of fundamental importance. Therefore, at that time, LTSEM seemed the most appropriate method to use for examination of wet samples. Howard also recognized the potential of developing a technique for obtaining cross sections of fully hydrated pulp and paper samples. An evaluation of the LTSEM as a means of examining the fully hydrated structures of ®bers and wet sheets was subsequently undertaken by Moss [216]. A.

The Cryosystem

The cryosystem of an LTSEM consists of a cryopreparation chamber (with cold stage, nitrogen dewar, fracturing device, sputtercoating head, airlock gate, and valve) and a cold stage (with nitrogen gas dewar and anticontaminator) interfaced with a conventional SEM. Figure 25 illustrates a typical system and shows the steps involved in cryopreparation. The use of an evacuated transfer system allows the frozen specimen to be placed directly onto the cold stage of the preparation chamber and then into the vacuum chamber of the SEM without coming into contact with the atmosphere or being returned to room temperature. The specimen is maintained, ideally, at 1858C by the circulation through the cold stage of nitrogen gas that has been cooled to 1908C with liquid nitrogen. It is essential that the temperature of the cold stage be maintained well below 1308C, because above this point sublimation begins to occur. Ice then has to be sublimed from the surface of frozen hydrated specimens, and this is achieved by heating the cold stage to a temperature of between 808C and 608C. An anticontaminator plate (shown at step 5 of Fig. 25) cooled to about 1908C prevents any recondensation of water molecules onto the specimen. The whole process can be monitored in situ and sublimation stopped when the relevant features of the specimen have been revealed. After sublimation of the surface ice, the specimen is withdrawn into the preparation chamber for sputtercoating and then returned to the cold stage for examination. Providing sublimation is not allowed to continue for too long there should be no morphological changes arising from shrinkage, and there should be little or no collapse because the bulk of the material is frozen, thus providing physical support.

Microscopy

Fig. 25

A typical cryosystem and preparation methods for LTSEM.

205

206

B.

Moss and Groom

Specimen Preparation

A small stub, similar to that used for conventional SEM work, is placed in a copper holder. The specimen, which may be a droplet of pulp suspension, a few wet ®bers, or a small piece of wet or couched sheet, is placed on the stub. A long rod (the transfer rod) is screwed into one end of the copper plate (see Fig. 25). This is used to plunge the specimen holder into the slushing chamber for freezing. It is then used for the subsequent transfer of the specimen to the cryopreparation chamber and to and from the cold stage. The size of the stub limits the size of the sample that can be examined, but this is not a real disadvantage because the smaller the sample, the faster the freezing. Special techniques are needed for obtaining cross sections of fully hydrated pulp and paper samples. Figure 26 illustrates the technique used for sectioning wet sheet samples. Here an SEM stub has been cut in half and the wet sheet sample is placed carefully between the two halves. The stub is inserted into the copper holder, and the sample is plunged into liquid nitrogen. The sample is then transferred to the cryopreparation chamber where the sheet is fractured. This is achieved by using a cooled scalpel that is inserted through the wall of the preparation chamber and manipulated from outside.

Fig. 26

Technique for obtaining cross sections of wet handsheet samples.

Microscopy

207

Figure 27 illustrates the technique developed for sectioning wet ®bers. Fibers are inserted into a hollow split rivet (2 mm diameter), the two halves of which are held together by a spring clip. Ideally, the ®bers are inserted into the rivet so that their long axes lie parallel to the long axis of the rivet. This method requires careful manipulation and considerable practice to achieve a uniform distribution of ®bers at an appropriate density; too many ®bers closely packed together make analysis dif®cult, and too few ®bers necessitate the processing of a large number of samples. The rivet is then inserted into a special stub that has a hollow center. This is placed in the copper holder, and the sample is plunged into liquid nitrogen. The transfer chamber is not immediately positioned over the top of the slushing chamber as it should be, and after a few seconds the frozen hydrated ®bers are fractured by giving the top half of the rivet a sharp blow with an appropriate tool (a spanner works well!)Ðall done while the ®bers are fully immersed in the liquid nitrogen. The transfer chamber is then ®tted over the top of the slushing chamber, and the rest of the procedure is as normal. Pretreatment of specimens is not necessary because they are physically ®xed by freezing rapidly in a liquid cryogen. Freezing specimens in liquid nitrogen slush at

Fig. 27

Technique for obtaining cross sections of wet pulp ®bers.

208

Moss and Groom

2108C was one of the earliest methods of freezing biological specimens and is one of the most commonly used today because it is cheap and safe to use. Other cryogens can be used that have a better freezing ef®ciency than liquid nitrogen, but they are not as readily available or are disadvantageous in some way or another. Likewise, other methods of freezing are available, but they generally require complex and expensive equipment [272]. C.

Comparative Evaluation of Preparation Techniques for LTSEM

Examination of wet pulp and paper samples in a conventional SEM necessitates dehydration of specimens either by air drying (AD), critical point drying (CPD), or freeze drying (FD). Many workers [e.g., 8,29,30] have reported on the deleterious effects of dehydration techniques. These include shrinkage, collapse of nonrigid cell walls, distortion, and even rupture of whole cells. Comparative studies should always be made of the different techniques being used, particularly if quantitative work is being undertaken. Ideally, objects of known dimensions should be used for calibration purposes. Pulp ®bers themselves are unsuitable because they are far too heterogeneous. Moss [216] used a single-cell alga, Euglena gracilis, to determine how much shrinkage and distortion of delicate structures was caused by air drying, critical point drying, and freeze drying compared with frozen hydrated cells. Light microscopy of living cells was undertaken as an experimental control. Euglena has been much used as an experimental organism, and its structure is well documented [36,102,111,278,364]. It was chosen for this evaluation because its relevant properties are in many respects similar to those of wet pulp ®bers. It has a protective membranous exoskeleton (the pellicle) that exhibits great ¯exibility and a distinctive, striated structural pattern. It also has a long, delicate ¯agellum. Comparison of the different methods of preparing Euglena for SEM examination indicated cryopreparation to be far superior to either FD or CPD. D.

Advantages and Disadvantages of Cryopreparation Techniques

All dehydration techniques used in preparing wet ®bers or sheets for examination using a conventional SEM cause shrinkage, collapse, and distortion of ®bers and networks to varying degrees. This is probably the reason electron microscopy has never been considered a satisfactory way of studying phenomena such as ®ber swelling. Critical point drying involves solvent exchange procedures, which may cause redistribution or loss of certain components. These problems do not occur in cryopreparation techniques, and samples can be examined in conditions more closely related to their natural state. Furthermore, cryopreparation is quicker and simpler. Cryopreparation obviates the need for chemical ®xation, and if specimens are rapidly frozen and maintained at a temperature below 1308C there should be no shrinkage or swelling. The shape and size of the specimen, the rate at which the specimen enters into the cryogen, and the depth to which it penetrates all have a pronounced effect on the cooling rate, so the choice of freezing technique ultimately depends on the nature of the specimen. The ideal method of cryo®xation should be such that the specimen retains good morphology, ultrastructure, and distribution of particulate components. However, cryopreparation does produce artifacts of its

Microscopy

209

own, namely the creation of eutectic membranes, which may be indistinguishable from genuine ®brillar material and sheets of ®ne lamellae peeling from the ®ber wall. Cryopreparation techniques appear to have yielded some excellent results. Cryo®xation physically ®xes and immobilizes dynamic processes and preserves the spatial distribution of particles, etc. within a sample. Many workers have found that this method gives a far superior state of specimen preservation compared with the standard dehydration techniques required for conventional SEM work [85,153,264]. A major disadvantage is that the specimens are very short-lived; unlike dried specimens, they cannot be stored in a desiccator and viewed again after an interval of weeks or months. E.

Artifacts of Cryopreparation

Artifacts Arising from Freezing Cryo®xation of specimens can cause problems, and the success of the process is largely dependent on the rate of cooling and the size and shape of the specimen. If freezing is not rapid enough there may be some redistribution of small cellular components or particles in suspension. The fastest rate of freezing currently available gives good ®xation to a depth of only about 10 mm. A slow rate of cooling produces large ice crystals that may disrupt cellular structure. It may also result in ice crystallization in the extracellular medium, and further slow crystal growth then causes water to be withdrawn across the cell wall. This may lead to shrinkage and deformation of the cell wall. Cryoprotectants, e.g., glycerol or DMSO, may be used to reduce ice crystal size for a given cooling rate, but their presence may pose further problems with artifacts, as described below. Cryo®xation of wet samples involves phase separation. As ice crystals form and grow, any solutes present in the sample become more concentrated. Consequently, freezing of solutions causes solutes to become concentrated into the liquid phase surrounding the growing ice crystals. This unfrozen solution (the eutectic phase) reaches a critical concentration (the eutectic concentration), at which point the solutes and remaining water freeze as a single component. Eutectic materials do not generally sublime, because their vapor pressure is relatively low. These eutectic regions remain and form a network surrounding the ice crystals. The process of cryo®xation and subsequent sublimation can produce a whole spectrum of artifacts [265], which may have repercussions on structural interpretation. Kistler and Kellenberger [171] demonstrated the formation of eutectic membranes during freeze drying of aqueous solutions, where the eutectic assumed the form of thin perforated sheets that, when completely dehydrated, appeared as a network of strands. Miller et al. [210] presented some very impressive electronmicrographs of lacy artifacts created by rapid freezing of buffer solutions commonly used in the preparation of biological material. Moss [216] and Moss et al. [217] demonstrated effects similar to those described above by freezing sugar solutions. Figure 28 shows a eutectic membrane formed by freezing a 0.03 mM solution of xylose. The membranous sheet seen in Fig. 28a splits apart as sublimation proceeds and forms the reticular structure shown in Fig. 28b. Other experiments showed that substances dissolved in the ®ltrates of beaten pulp suspensions also produced similar membranous and lamella-type structures that can mimic thin sheets of lamellae peeling off the walls of beaten pulp ®bers and produce ®ne strands that could easily be mistaken for external ®brillation [217].

210

Moss and Groom

Fig. 28 (a) Eutectic membrane formed during sublimation of a 0.03 mM xylose solution, which forms a reticular structure (b) as sublimation proceeds. (From Ref. 216.)

Microscopy

211

The disruption of cell wall layers during beating and the production of ®brillar and membranous material is a phenomenon that has been well documented in the literature (see, e.g., Giertz [101] and Iwasaki et al. [137]). Although no chemical analysis was made to determine the hemicellulosic content of the pulps used in the experiments, such membranous and ®brillar structures were attributed to the presence of soluble wood carbohydrates, and certainly the circumstantial evidence is strong. In a pulp suspension much of the ®brillar and membranous material will be of cellulosic origin and not an artifact of cryo®xation; the problem lies in distinguishing between fact and artifact. This therefore poses problems in the use of LTSEM for analysis of external ®brillation or classi®cation of ®nes material. Furthermore, it takes very little dissolved material to produce artifacts, as evidenced by studies of cotton ®bers, which have a very low hemicellulosic content. Likewise, the use of cryoprotectants can produce similar effects. These problems are also encountered in the examination of freeze-dried material using a conventional SEM. Artifacts Arising from Fracturing Techniques Fracturing a FH sheet involves planing the surface with a cooled scalpel blade. Use of knives in fracturing techniques may result in scuff marks across the fracture face and also ``smearing,'' due to localized surface melting, leading to loss of detail in parts of the specimen. Fracture methods that cleave the specimen in a single step, such as that employed with the split rivet, produce fracture faces that lack scuff marks but have a coarser surface topography. There is a greater tendency with such methods for fractures to occur along planes of weakness. Plastic deformation, a phenomenon that has been investigated by a number of workers, may also occur. Dunlop and Robards [73] experimented with polystyrene latex spheres in aqueous solutions of glycerol that they froze and fractured. Many of the spheres stretched by up to 50% before they fractured, whereas others were pulled out from the frozen matrix. However, Kellenberger et al. [160] reported that where plastic deformation had occurred the broken surfaces were usually heavily distorted with a rough appearance, and thus easily recognized. To quote Echlin [75], It has been said that in biology, image as related to truth is inseparable from artifact as related to the act of observing. It would appear that cryobiology is going to permit us to get nearer the truth than most of the methods we have been using in the past.

F.

Applications of the LTSEM to Pulp and Paper Research

Compared to other microscopes such as the CLSM, the LTSEM has been less used as a research tool for studying pulp and paper and may well be superseded by the ESEM, which can look at wet and moist samples directly. Fiber Structure Papermaking raw materials, both wood and non-wood, can be examined without the need for dehydration. Freeze-fracturing of ®bers provides a method by which comparisons can be made between the internal structure of the raw material, never-dried, once-dried, and recycled pulp ®bers.

212

Moss and Groom

External Fibrillation and Fines Material The LTSEM would seem to be the ideal instrument for detailed examination of ®brillation and ®nes material, because cryo®xation preserves the spatial arrangements of the ®bers and ®brillar material in a pulp suspension. However, the rate of development of external ®brillation cannot be accurately assessed owing to the fact that cryopreparation techniques can produce ®brillar material that may be mistaken for genuine ®brillation, as illustrated in Figs. 29 and 30. Cell Wall Structure and Internal Fibrillation It is in the study of cell wall structure and internal ®brillation that LTSEM would seem to offer the greatest potential. The relatively simple method by which cross-sectional views of frozen hydrated ®bers can be obtained for examination has enabled a study to be made of the internal changes in cell wall structure brought about by beating. It has also shown that differences exist between beating methods. In this application the presence of ®brils that may, in fact, be artifacts does not affect interpretation of cell wall delamination. Although the CLSM allows the ®bers to be cross-sectioned in their fully hydrated state without any pretreatment, the LTSEM allows observation of structures at a far higher resolution than can be achieved with optical microscopy. Examination of internal ®brillation of wet beaten ®bers has, in the past, meant that ®bers have had to be dehydrated, embedded, and mechanically sectionedÐwith all the attendant repercussions. Cryo®xation of specimens for use with the LTSEM overcomes these problems, as Figs. 31±33 show. The appearance of internal ®brilla-

Fig. 29 Fibrillar material seen in a suspension of bleached sul®te ®bers beaten for 45 min in a Lampen ball mill.

Microscopy

213

Fig. 30 Fibrillar material formed on nylon ®bers suspended in a solution of 2 mM arabinose solution.

tion in Fig. 31 clearly resembles Scallan's honeycomb model (cf. Fig. 1) except that the cell wall has separated into fewer lamellae. One cannot determine whether the ®brillar material spanning the cracks between wall layers is really part of the honeycomb pattern, but it is clear that the cell wall has separated into many distinct layers. It has been suggested that if the ®bers are not frozen rapidly enough then water within the cell wall could split the wall lamellae during the freezing process, but since the amount of delamination correlates so closely with the degree of beating there is no reason to suppose this has happened. Also, cell wall delamination was rarely observed in unbeaten ®bers even though they had been soaked and disintegrated according to the prescribed methods [319,331] (Fig. 32). However, with the subsequent advent of the CLSM it is now possible to verify whether or not cryopreparation techniques are inducing structural or dimensional changes in cross-sectional views of ®bers. Wet ®bers can be examined ®rst with the CLSM and then the same ®bers can be frozen and fractured and examined in the LTSEM. Quanti®cation of Cell Wall Delamination Experiments carried out by Moss [216] showed that the ratio of delaminated to ¯at ®bers in dry lap pulp samples correlated well with tensile index properties, and this ratio was proposed as a predictive index for the development of tensile strength. Unbeaten pulp ®bers consisted largely of collapsed ®bers in which the lumina were clearly visible but with the cell walls showing very little or no sign of delamination. Beaten ®bers showed a variety of cross-sectional shapes, from ¯at, through stages of swelling, to fully swollen, with

214

Moss and Groom

Fig. 31 (a) Cross section of internal ®brillation in a frozen hydrated bleached sul®te ®ber beaten for 120 min in a LampeÂn ball mill. (b) A close-up of the cell wall. (From Ref. 216.)

Microscopy

215

Fig. 32 Cross sections of frozen hydrated ®bers of an unbeaten bleached softwood sul®te pulp showing little evidence of cell wall swelling or delamination. (From Ref. 184.)

Fig. 33 Cross sections of frozen hydrated ®bers of an unbleached kraft pulp beaten for 90 min in a LampeÂn ball mill showing various degrees of wall swelling and delamination. (From Ref. 216.)

216

Moss and Groom

the cell walls showing various degrees of delamination (see Fig. 33). The number of ®bers showing swelling and cell wall delamination varied with beating time and method of beating. It must be emphasized that only internal ®brillation on the macroscale, i.e., what Page referred to as cracks, was being examined in this study. Sheet Structure Pye et al. [258] used a conventional SEM to study the structural changes that occur in paper during pressing and drying. In their experiments the wet sheets were critical point dried or freeze-dried, which would have induced shrinkage and changes in the geometry of the wet web as well as loss of ®nes through the solvent exchange process. LTSEM offers a better way of observing the changes that occur at the different stages in sheet consolidation from just-drained through couching to pressing and drying. However, although examination of frozen hydrated samples eliminates the effects of shrinkage and changes in microstructure incurred during dehydration, interpretation of ®ne detail may be confounded by the presence of artifacts. The technique of freeze-fracturing and examining cross sections of frozen hydrated sheets offers considerable potential for studying two-sidedness, both in terms of structure and in the distribution of ®llers such as titanium dioxide and calcium carbonate, particles of which are readily identi®ed in the LTSEM. It could also be used to investigate lumen loading. The LTSEM is probably the best method currently available for such studies, because preparative techniques used for conventional SEM can bring about a redistribution of the particles during dehydration. Coating Structure Sheehan and Whalen-Shaw [292] used high magni®cation cryo-SEM to examine microstructures in aqueous suspensions of coating components. They were able to observe associations between clay and latex pigments in coating suspensions containing kaolin, styrene-butadiene latex, and carboxymethyl cellulose in different combinations. They demonstrated how freeze-fracture techniques provide the visual link between wet and dry particles and how mechanisms of colloidal interactions can be deduced. Their paper also provides illustrations of the metal-coating and prepump chambers of the cryosystem they used. VI.

THE ATOMIC FORCE MICROSCOPE

Microscopes have been an integral part of material characterization since the early seventeenth century with the development of the ®rst optical microscope (see Section II). There have been many changes in optical microscopy, the most notable being altering of the light path via prisms, mirrors, and slits and the use of various light sources allowing ¯uorescent and optical sectioning. However, the basic operation of the optical microscope remains unchanged: an image is generated by passing light through, or re¯ecting it off, a sample. Although functionally quite different from optical microscopes, electron microscopes operate on a similar principle: an image is reconstructed from a scanning sequence acquired from a sample being illuminated with an electron beam rather than light. The resolution of the electron microscope is far greater than that of the optical microscope, primarily due to the short wavelength of electron radiation. The electromagnetic lenses used in electron microscopes are subject to the same type of diffraction limits as the lenses of optical microscopes.

Microscopy

217

Optical and electron-based microscopes have serious shortcomings with regard to the study of biological materials: optical microscopes are limited by their resolution, whereas electron-based microscopes generally require extensive material preparation, result in material degradation, or must be operated in nonambient conditions. These limitations can be overcome by the use of a new type of scanning probe technique referred to as atomic force microscopy, which allows for the visualization of topographic features from the macroscopic scale down to the atomic scale. The atomic force microscope (AFM) belongs to a family of microscopes known as scanning probe microscopes. These operate on a completely different principle than either the optical or electron microscopes: images are computer-generated by monitoring the x, y, and z positions of a probe scanning at or near the sample surface (Fig. 34). The microscope functions interactively to simultaneously monitor the position of the probe and maintain a desired tip±sample distance. The topogra®ner was the ®rst functional device to study surface imaging at near-atomic resolution [372]. This was the ®rst device that allowed surface imaging using the quantum-mechanical phenomenon of electron tunneling [198]. The topogra®ner was composed of a conductive tip attached to a piezoelectric scanner and was scanned over the surface of a conductive sample with a feedback-controlled motion. A similar apparatus was developed by Binnig and Rohrer [21]. They improved the feedback mechanism and electronics and were better able to monitor tip location such that resolution was improved to several angstroms. The scanning tunneling microscope was the ®rst instrument to generate images of surfaces with atomic resolution. An STM image is generated by monitoring the current of electrons ¯owing between a sample and a conducting tip that are separated by a distance  of no more than about 10 A . The STM, invented by Binnig and Rohrer and for which they were awarded the Nobel prize in physics in 1986, is the modern precursor of all scanning probe microscopes. The stunning images, resolution capability, and material characterization ®ndings prompted the rapid development of similar types of probe-based microscopical techniques. These microscopical techniques are generally lumped under the banner of scanning probe microscopy and include atomic force, interatomic force, electrochemical scanning tunneling, frictional force, near-®eld scanning optical, surface compliance, magnetic force, scanning thermal, and scanning capacitance. A prerequisite for this tunneling of electrons to occur across a bias voltage is that both the tip and sample must be conductors or semiconductors. This restriction limits the applicability of the STM in wood ®ber investigations due to the electrical insulating characteristics of lignocellulosic ®bers. The AFM is not subject to the same restrictions as the STM because tip displacement rather than current ¯ow is the primary sensor measurement. Therefore, this section deals with the applicability of atomic force microscopy to the study of lignocellulosic ®ber surfaces. A.

Principles and Operation of Atomic Force Microscopy

Principles of the AFM The AFM generates images by scanning the sample surface with a sharp tip mounted on the free end of a ¯exible cantilever (Fig. 35). The tip is lowered to the surface of the sample until suf®cient force is encountered to result in a de¯ection of the cantilever. De¯ection of the cantilever is monitored by re¯ection

Fig. 34

Schematic of an atomic force microscope. (Adapted from image courtesy of Digital Instruments, Veeco Metrology Group.)

218 Moss and Groom

Microscopy

Fig. 35

219

Side view of an intermittent contact cantilever and tip. (Courtesy of T. C. Pesacreta.)

of a laser signal from the back of the cantilever to a position-sensitive photodetector. Tip de¯ection data are necessary to carry out two primary functions: (1) to computer-generate a three-dimensional map of surface topography and (2) to establish the tip setpoint and subsequently maintain a constant tip±sample interaction. The photodetector signal, which thus monitors the de¯ection of the probe tip, is used for feedback to control a piezo-based scanner. For example, if the probe encounters a relatively high feature, the photodetector will monitor the upward probe displacement. This feedback signal will be relayed to the scanner, which will retract the tip enough to restore a predetermined tip±sample force setpoint. The same signal that controls the tip±sample setpoint is also used to generate the three-dimensional topographic representation of the sample surface. Hardware All AFM techniques depend on four components: probe, scanner, detection mechanism, and feedback electronics. A general description of each component is presented in the following sections. Probe The probe directly determines the type of image acquired and the ultimate lateral resolution possible. The probe is the least expensive AFM component and is therefore generally considered expendable after a certain number of scans have been obtained. A ``probe'' is actually a uni®ed assembly of two elements: a ``tip,'' which is a stylus with an extremely ®ne point that is in contact (or near contact) with the sample, and a ``cantilever,'' which is a ¯exible arm linking the tip to the scanner. The most common AFM tips are commercially fabricated from silicon or silicon nitride and are pyramidal, tetrahedral, or conical [198]. Tip selection is crucial, because the tip radius, along with the step size of the image, determines the lateral resolution of AFM images [132]. Conical tips are the sharpest and thus the  most fragile, with tip radii as small as 50 A . Pyramidal and tetrahedral tips have tip

220

Moss and Groom

radii of a few hundred angstroms but are very durable. The high aspect ratio of conical tips makes them ideal for imaging samples with deep troughs or sharp edges. Probe selection depends on the sample being imaged and the operating mode of the microscope. The majority of AFM imaging is done with either a contact or noncontact probe. Contact probes require intimate contact with the sample and therefore have cantilevers with low spring constants so as not to de¯ect the surface of the specimen. Contact probes are made of silicon nitride with a durable tip that has a large tip radius and low aspect ratio. Non-contact probes operate in forced vibration and must be rigid enough to allow for high frequency vibration. Non-contact probes are generally composed of an etched silicon cantilever and a tip with a small radius and high aspect ratio. Scanner The function of the scanner is to either move the tip over the sample or move the sample under the tip and is controlled by user-de®ned parameters and feedback regarding tip±sample interactions. Scanners are generally designed to allow translation in the three principal directions and are made of piezoelectric materials. Thus, scanner movement in the x, y, and z directions can be controlled by expansion and contraction in response to an applied voltage. Scanner movement is generally in a raster or back-and-forth pattern. The size of the scanner tube dictates the limits of lateral motion as well as vertical resolution [132]. Generally, lateral movement is restricted from a minimum of a few nanometers to 100 mm to a maximum of 100±200 mm. The range of vertical displacements are from 5±10 mm down to the subangstrom range. Detection Mechanism The choice of detection mechanism depends on the operating mode of the microscope. Traditional AFM is based on the vertical displacement of the tip and is monitored by the re¯ection of a laser beam from the back of the cantilever to a photodetector. The z-displacement photodetector data are then used to control the vertical response of the piezoscanner. Detection mechanisms that monitor such variables as probe rotation, lateral resistance, and disruption of vibrational integrity are also necessary to allow material characterization using various operating modes outside of traditional AFM. Examples of nontraditional AFM operating modes include force calibration, phase shift, and lateral force. Speci®cs of these detection mechanisms are beyond the scope of this section. Feedback Electronics The function of the feedback system is to maintain userde®ned tip±sample speci®cations by controlling the actions of the tip and the scanner. The maintenance of constant tip±sample interactions is necessary to collect accurate morphological representations of surface images. A constant interaction also ensures representative non-height data for techniques such as phase and force modulation. B.

Operating Modes

Atomic force microscopic images are generated by maintaining a constant de¯ection between the tip and the sample. De¯ection is held constant by monitoring the tip± sample force, which can be either attractive or repulsive. There are various AFM techniques that take advantage of these opposing forces. Contact AFM requires repulsive forces, non-contact AFM exists in the repulsive regime, and intermittent

Microscopy

221

contact (also known as ``tapping'') AFM is governed by the repulsive force but also resides in the attractive domain. These forces are summarized in Fig. 36. Contact Mode The contact mode of AFM is a nonvibrational technique that brings the tip of the probe to within a few angstroms of the sample surface. As the tip and sample are brought closer together, their atoms begin to attract each other; then their respective atoms encounter each other's Coulombic (repulsive) forces. This repulsive force prevents further intrusion by the tip and causes the cantilever to de¯ect due to the low spring constant of the cantilever. Surface topography is then determined by scanning this tip over the sample while monitoring tip de¯ection, maintaining a constant repulsive force. Non-Contact and Intermittent Contact Modes Non-contact AFM relies on monitoring the de¯ection of a vibrating cantilever as it is brought into close proximity of the sample surface. A rigid cantilever is vibrated at or near its resonance frequency, typically between 50 kHz and 1 MHz, and with a magnitude of 3±20 nm. This vibrating tip is lowered toward the sample surface. The tip encounters several boundaries during engagement before reaching the region dictated by molecular potential energy: ¯uid ®lm, electrostatic, and ¯uid surface. The ¯uid ®lm boundary

Fig. 36 Tip±sample interactions. (Adapted from image courtesy of Digital Instruments, Veeco Metrology Group.)

222

Moss and Groom

dampens vibration approximately 10 mm from the sample surface. However, this damping is temporary and disappears as the probe passes through the boundary. Electrostatic forces begin within 1 mm of the sample surface and can be either attractive or repulsive. Electrostatic forces in wood pulp ®bers tend to be rather high, due in part to their low conductivity. The ®nal pre-engagement boundary is the ¯uid surface boundary, which begins about 100 nm from the sample surface. This is an especially signi®cant layer in wood pulp ®ber investigations owing to the hydrophilic nature of the cellulosic material. This attractive force has the capability to pull the tip down onto the sample surface with a force strong enough to cause indentations. The probe is lowered until the tip begins to encounter van der Waals forces, typically 20±200 A from the sample surface. In a modi®ed version of non-contact atomic force microscopy referred to as intermittent contact microscopy, the oscillating tip is moved toward the surface until it begins to lightly touch or ``tap'' the surface. The tip is considered to contact the sample surface when the amplitude apex enters the repulsive regime of the molecular potential energy curve. In either non-contact mode, the force on the sample is automatically set and maintained by the feedback control mechanism. The vibrating tip is then scanned over the sample surface to determine topography. These modes are generally used for soft or elastic materials because the tip exerts very low forces on the sample surface. The intermittent nature of intermittent contact microscopy is most ideally suited for cellulosic ®bers. The oscillation amplitude is suf®cient to overcome any tip±sample interactions due to bound surface water. Also, sample damage is minimized by the vertical tip loading, there being no shearing forces to laterally displace the sample surface. Other Variants of AFM Many variants of conventional AFM techniques have evolved that either alter the detection mechanism or modify the probe. The variants are too numerous to describe in detail. However, those techniques most applicable to the examination of cellulosic ®bers are presented in this section. Lateral Force Microscopy Also known as frictional force microscopy, lateral force microscopy measures the twisting of the cantilever that arises from forces parallel to the plane of the sample surface [132]. Lateral force microscopy is useful for imaging variations in surface friction attributable to surface inhomogeneities. Further frictional studies can be conducted by coating the tip with various substrates. Force Modulation Microscopy This technique, also known as surface compliance microscopy, allows the user to ascertain the resistance of a sample to a load applied normal to the surface. The force applied to the sample surface is modulated with respect to z displacement. Traditionally, the AFM is operated in contact mode for force modulation microscopy, but recent advances in electronics allow for the intermittent contact mode to be employed. Changes in cantilever amplitude, which are a function of the elastic properties of the sample surface, are collected and converted to a compliance or elasticity image. Topographic data are collected simultaneously with force modulation data to maintain a constant tip±sample distance. The application of force modulation microscopy for wood pulp ®bers may prove useful in mapping various surface components.

Microscopy

223

Phase Detection Phase detection is done in intermittent contact mode AFM and monitors the difference between the cantilever driving signal and the cantilever output signal [132]. The phase lag, in part, indicates the surface adhesion or viscoelastic properties of a sample. In lignocellulosic ®bers, this technique can serve either to differentiate between bulk components such as lignin and cellulose or to identify surface contaminants. Phase detection, also known as phase imaging, can be conducted simultaneously with force modulation microscopy. Near-Field Scanning Optical Microscopy One of the newest and fastest growing variants of conventional AFM, near-®eld scanning optical microscopy (NSOM) uses an aluminum-coated optical ®ber as the scanning probe and produces visible light images. Laser light is guided through the ®ber, which acts as a subwavelength aperture to the sample, and is detected from either a transmitted or re¯ected signal. Scanning this laser light source over the sample allows for optical images with  resolution up to 150 A . Various types of illumination (e.g., polarization, ¯uorescence) can be used. The primary signi®cance of this technique lies in ultrahigh resolution optical imaging and single-molecule spectroscopy. A review of the technique and its applications is given by Harris et al. [119]. Electric Force Microscopy Electric force microscopy operates in a non-contact mode and measures the locally charged domains of the sample surface. A voltage is sent between the scanning tip and the sample, with cantilever de¯ection being proportional to the charge density of the sample surface. Although several offshoots of this technique exist, the primary purpose in lignocellulosic ®ber research would be to map static charges on the ®ber surface for ®ber±®ber or ®ber±substrate compatibility. Scanning Thermal Microscopy Thermal scanning microscopy is a non-contact mode AFM variant that utilizes a specialized cantilever composed of two different metals to ascertain the thermal conductivity of a sample surface. The difference in thermal conductivity responses of these metals causes the cantilever to de¯ect. Scanning the sample and monitoring this de¯ection results in a thermal conductivity map of the sample surface. Changes in vibrational amplitude are also collected simultaneously to map surface topography. Additional AFM Variants Several other techniques that may prove useful in the understanding of wood and other lignocellulosic ®bers include Chemical force microscopy, which uses chemically modi®ed tips to quantify adhesion and functional group distributions in molecular assemblies Magnetic force microscopy, which is similar in principle to electric force microscopy but uses a magnetically charged tip to image weak magnetic interactions

C.

Environmental Conditions

The AFM can image samples in one of four conditions: liquid, air, vacuum, and electrochemical. Electrochemical imaging is not relevant in pulp and paper research because it is used primarily for investigating metallic samples and biological materials in electrolytic solution. Imaging in a vacuum environment is predominantly for STMs but is also applicable for AFMs. The purpose of the vacuum is to minimize

224

Moss and Groom

electrostatic interference and debris contamination. The applicability of vacuum atomic force microscopy to wood pulp ®ber research is questionable due to the nonambient conditions and additional preparation time with little to no increase in image quality. Ambient air is the simplest, least expensive, and thus most popular scanning environment for the AFM. Sample preparation for this environment is minimal; samples are adhered to a magnetic AFM ``puck'' with a viscous adhesive or double-sided sticky tape. This type of environment is ideally suited for surface characterization of lignocellulosic ®bers, primarily because composites made from these materials are used in ambient conditions. Although AFM imaging of lignocellulosic ®bers is affected by various attractive and repulsive forces near the surface (see Fig. 36), it is still possible to get high resolution images of various ®ber types imaged in air. Imaging in liquids eliminates the ¯uid surface tension forces encounterd with imaging in air, resulting in images with improved detail. There are a few disadvantages to imaging in ¯uids. Lignocellulosic ®bers cannot be imaged in their dehydrated or near-dehydrated condition, with features possibly being obscured or distorted due to swelling of the cell wall. In addition, sample adhesion to the AFM puck must be adequate and able to withstand the liquid media. It should also be noted that liquid imaging requires more sample preparation time and is cumbersome due to the dif®culty of moving between samples within and among AFM pucks. However, the increased effort is rewarded in many cases by improving surface topographic details. D.

Applications of the AFM to Pulp and Paper Research

The applicability of the AFM to the pulp and paper industry lies in two general areas: qualitative and quantitative analysis. Qualitative analysis is necessary for observations of physical features and overall surface topography, often at extreme magni®cations. Collected data exist in the form of a three-dimensional array, thus allowing the development of algorithms that can then quantify individual features such as micro®brils and crystallites as well as more global properties such as surface roughness and spectral type analyses. It should be noted that most of the research done on cellulosic ®bers has been conducted in intermittent contact mode. The relatively rough surface of cellulosic ®bers can be problematic in contact mode, causing severe rotational stresses on the cantilever and tip as well as scratching and altering the surface of the specimen. The tapping nature of the intermittent contact mode results in negligible specimen damage and in images of high spatial resolution. Most of the AFM images shown in this section were obtained in intermittent contact mode. Qualitative Analysis There currently exist only a limited number of publications linking the art of atomic force microscopy and cellulosic ®bers. Excellent summaries of the applicability of the AFM to biological samples are given by Morris [215], with speci®c examples relating to wood and pulp ®ber research given by BeÂland [16]. A summary of the technique and the imaging of polysaccharides and organics in general can be found in Kirby et al. [170] and Radmacher et al. [262]. These publications stress the fact that the most common type of AFM analyses have been qualitative,

Microscopy

225

owing to the simplistic manner of image collection and analysis as well as the wealth of information generated. In essence, the qualitative analysis allows the investigator to ascertain the effect of various pulp generation schemes on the surface of cellulosic ®bers. Comparison with Various Scanning Microscopes Although the confocal laser scanning microscope (CLSM) is capable of characterizing surface details of cellulosic ®bers and networks, it is limited in its magni®cation and is therefore used primarily in the investigation of subsurface features as described in Section IV. The family of microscopes that most closely parallels specimen characterization by atomic force microscopy is that of electron microscopes. Some of the earliest transmission electron microscopical (TEM) work characterizing cellulose micro®brils was conducted by Asunmaa [4] on pine holocellulose and by Revol [268] on high resolution diffraction contrast images of transverse sections of the green algae Valonia ventricosa. Subsequent TEM studies by Revol [269] and Sugiyama et al. [312] con®rmed the earlier work. Hanley et al. [114] compared the images produced by AFM and TEM of Valonia ventricosa cellulose micro®brils. They found those microscopes to be complementary: specimen preparation for the TEM was more laborious but provided true detail of subsurface micro®bril structure, whereas the AFM provided information about surface features. Resolution for the AFM and TEM is similar,  around 2 A . Generally, TEM stains do not show cellulose well, if at all, except for negative stains of isolated cellulose strands. The SEM is more directly comparable than the TEM to the AFM because both the SEM and AFM use a raster-type motion to scan the surface of a specimen. In general, the surfaces to be studied in the SEM and AFM are similar. The primary differences in producing SEM and AFM images have been outlined by BeÂland [16] and are summarized as specimen preparation, ®eld of view, and resolution. Traditional SEMs require that specimens be conductive, thus requiring metal coating. However, this coating greatly restricts the level of detail attainable due to the size of the granular metal coating [110,355]. The 5±20 nm thick coating of gold typically used in SEM sample preparation often obscures the surface details of individual cellulosic ®bers, as can be seen in Fig. 37. A cotton ®ber (Gossypium hirsutum L.), with the waxy cuticle and pectin removed with an Updegraff reagent, was ®rst imaged with the AFM in intermittent contact mode (Fig. 37b). The ®ber was then sputtercoated with gold and imaged at a corresponding magni®cation with an SEM (Fig. 37a). It can be seen that the clarity of the cellulose micro®brils has been masked by the conductive coating. The problems associated with conductive coatings were minimized by Patnaik et al. [252] in evaluating thin ®lms of hydroxypropyl cellulose with an AFM and a low-voltage high resolution SEM (LVHRSEM) using a 2 nm layer of tungsten applied with a dual ion beam coater. However, the resolution and quantitative capabilities of the AFM were necessary to determine ®brillar trajectories. Figure 38 illustrates a direct comparison of an uncoated loblolly pine (Pinus taeda L.) macerated ®ber as imaged with an ESEM and an AFM. Figure 38a was imaged in an EnviroScan ESEM at a magni®cation of 170. The dashed region of Fig. 38a was rescanned at a higher magni®cation of 540, but it is still apparent that there is insuf®cient contrast to de®ne individual micro®bril features. This ®ber was removed from the ESEM chamber, and the boxed region of Fig. 38b was imaged with the AFM (Fig. 38c). The AFM image shows much greater detail than the

226

Moss and Groom

Fig. 37 A cotton ®ber with the cuticle removed after treatment with 6% NaOH and Updegraff reagent and imaged with (a) a scanning electron microscope and (b) an atomic force microscope (amplitude data, 2.5 mm scan). The ®ber was ®rst imaged with the AFM, sputtercoated with gold, then imaged in the SEM. (Courtesy of T. C. Pesacreta.)

original ESEM images, with individual micro®brils clearly evident. This increase in image detail was also observed in an investigation of cellulosic ®ber conducted by Suleman [314] in which the author compared uncoated cellulosic ®bers, primarily softwood pulp ®bers, with the AFM and the ESEM. Fiber Characterization The AFM is a surface probe, so the characterization of cellulosic ®bers is a direct function of the manner in which the ®bers were generated

Fig. 38 Varying levels of magni®cation of an uncoated macerated loblolly pine juvenile, latewood ®ber as imaged in (a) and (b) an environmental scanning electron microscope and (c) an atomic force microscope (amplitude data, 2.5 mm scan). [(a) and (b) courtesy of L. Mott and S. M. Shaler.]

Microscopy

227

and subsequently treated or modi®ed. Thus, the applicability of the AFM to cellulosic ®ber characterization becomes most meaningful in the area of pulping and its effect on surface properties. Although the potential for cellulosic ®ber characterization with the AFM is immense, little published information is available, primarily because the use of AFM is in its infancy in regard to biological sampling. Recent theses by Suleman [314] and Hanley [113] greatly expand our knowledge base of ®ber surfaces. Suleman evaluated the effect of sul®te versus sulfate processes, recycling, beating, and bleaching. In a complementary publication, Hanley and Gray [115] focused more on fundamental characterization, speci®cally with the description of cellulose micro®brils. The imaging of solid wood and cellulosic ®bers in the AFM involves a relatively simple procedure. The specimen of interest is adhered to an AFM puck by means of an adhesive or double-sticky tape and is then positioned under the AFM head. Figure 39 shows the height data of a typical AFM scan operating in intermittent contact mode of a macerated loblolly pine ®ber pit. The ®brillar structure of the pit chamber is clearly evident, as are remnants of the middle lamella and depositions from the maceration process. The applicability of the AFM to the pulp and paper industry will lie primarily in the ability of the AFM to ascertain surface features and properties of individual cellulosic ®bers as a result of the pulp generation method and subsequent treatments

Fig. 39 Intermittent contact mode 25 mm scan showing the primary cell wall, remnants of the middle lamella, and pit chamber of a single macerated loblolly pine ®ber.

228

Moss and Groom

and modi®cations. Nissan and Sternstein [229] state that paper strength is directly related to the frequency and strength of ®ber bonding. This reliance on ®ber-to-®ber bonding for the development of paper strength was reiterated by Page [235]. Some factors governing ®ber±®ber bonding are the surface topography and the primary constituents on the surface of individual ®bers. Figure 40 demonstrates the difference between ®bers that were generated mechanically and chemically. The ®ber in Fig. 40a was generated in a pressurized disk re®ner at a re®ning pressure of 40 psi and an ambient temperature well below 958C, the reported glass transition temperature of lignin [271]. The conditions resulted in intrawall damage as observed with the CLSM as well as a splotchy ligni®ed surface. The ®ber in Fig. 40b was chemically generated in a rather extreme kraft pulping regime, with an overall ®ber yield of 35%. The resulting pulp ®bers have had essentially all of the hemicellulose and lignin removed, thus revealing the random nature of the cellulose micro®brils of the primary cell wall. Figure 41a shows a typical mature loblolly pine pulp ®ber generated from a potassium/anthroquinone (KAQ) pulping regime to a pulp yield of 47% retain. The pulping was mild enough to not remove all of the lignin (pulp yield 47%), thus obscuring the micro®brillar network. Figure 41b shows a mechanically re®ned ®ber analogous to the conditions of Fig. 40a but under a re®ning pressure of only 10 psi. The longitudinal ®ber axes in Fig. 41 run horizontally and show the lamellar undertones of the S1 secondary cell wall layer. Steam-exploded ®bers and high yield kraft pulp ®ber exhibit similar surface details, as can be seen in Fig. 42, in which the longitudinal axis is oriented horizontally. Drying deformation in the form of wrinkles and creases can be seen as macro features. The micro®brils are present but not easily discernible in Fig. 42a, due most likely to a thin uniform coating of lignin. The high yield (kappa number of 100) commercial kraft pulp retained much of its original lignin, which formed a thick coating that masked the ®brillar network.

Fig. 40 Comparison of mechanically and chemically generated mature loblolly pine ®bers, showing 1 mm scans of (a) ®ber re®ned at ambient temperatures and at a re®ner head pressure of 40 psi and (b) low yield kraft ®ber (yield 35%).

Microscopy

229

Fig. 41 Scans (2.5 mm) of individual mature loblolly pine ®bers generated from (a) a potassium/anthroquinone pulping regime (yield 47%), and (b) a disk re®ner at ambient temperature and at a pressure of 10 psi.

The AFM is also important for distinguishing the manner in which various types of ®bers respond to pulping regimes. Figure 43 shows the response of juvenile and mature segments of a single loblolly pine to identical Franklin maceration schemes [246]. The surface features of the juvenile wood pulp ®bers (Fig. 43a) exhibited no signs of a primary cell wall, showing a ¯at micro®bril angle customary with the S1 secondary wall layer. In contrast, the mature wood pulp ®bers (Fig. 43b) possessed a random network of micro®brils, indicating that the primary wall was present. It is unclear whether the primary wall of the juvenile wood ®bers was present in the ®rst place or was so thin that it was removed during the pulping process.

Fig. 42 Scans (2.5 mm) of (a) a steam-exploded loblolly pine ®ber from growth ring 20, and (b) an unbleached linerboard pine ®ber generated by the kraft process.

230

Moss and Groom

Fig. 43 Intermittent contact mode 2.5 mm scans of loblolly pine ®bers subjected to identical maceration schedules of glacial acetic acid and hydrogen peroxide, showing (longitudinal ®ber axis oriented horizontally) (a) a juvenile and (b) a mature ®ber.

Although the AFM is used exclusively to evaluate surface features and topography, it is possible to image internal cell wall features by physical removal of outer layers. The S3 and S2 layers of mature loblolly pine ®bers are shown in Fig. 44, in which the longitudinal ®ber axis is oriented vertically. The S3 layer image was obtained by longitudinally splitting a ®ber with a knife, then mounting that bisected ®ber lumen side up (Fig. 44b). The S2 layer was imaged by initiating a cut with a knife and then peeling the ®ber back with forceps, thus removing the S1 layer and a portion of the S2 layer. In both cases, the micro®brils are clearly evident and are oriented accordingly.

Fig. 44 Intermittent contact mode 2.5 mm scans of macerated loblolly pine ®bers. Longitudinal ®ber axis runs from top to bottom of the images that show (a) the S3 layer as viewed from the lumen, and (b) the S2 layer, with the S1 layer physically removed.

Microscopy

231

The effects of chemical and physical treatments of the surfaces of various types of cellulosic ®bers can also be studied. Figure 45 shows the surface topography of macerated and re®ned loblolly pine ®bers. The macerated ®bers were chemically generated in a solution of acetic acid and hydrogen peroxide; speci®c details of the pulping regime are outlined by Groom et al. [109]. The re®ned ®bers were generated in a pressurized disk re®ner at 4 bar. The difference in the pulp generation method is evident from the surface topography. Individual micro®brils are evident on the surface of the macerated ®ber. Micro®brils are not visible on the surface of the re®ned ®ber surface because they are overshadowed by a coating that is more likely lignin in nature. It should also be noted that the AFM is being used to study native cellulose in situ, as demonstrated in the example of the green algae Valonia ventricosa (Fig. 46). Valonia ventricosa is commonly used in the study of cellulose because of the arrangement of its ®brils [126] and the high degree of crystallographic perfection within the ®brils [312]. In addition to qualitative assessment of cellulose micro®brils, current efforts are under way to quantitatively assess cellulose and its corresponding crystallites. This work is discussed in the next section. Level of Magni®cation The ability of the AFM to resolve features down to the angstrom level is one of the features that makes the AFM most useful for fundamental ®ber research, especially with respect to the constitution of cellulose micro®brils and their subsequent components. An illustration of the applicability to cellulosic ®ber research is shown in Fig. 47. A commercially processed cotton ®ber was adhered to an AFM puck and imaged at low, medium, and high magni®cations, with the ®ber longitudinal axis oriented horizontally. The 6 mm scan (Fig. 47a) reveals what appears to be a ®brillar network. Subsequent scans of 1:7 mm (Fig. 47b) and 530 nm (Fig. 47c) show that the undulating network was not ®brillar in nature but were wrinkles caused by processing and/or shrinkage, with the cellulose

Fig. 45 Atomic force height images (5 mm scans) of loblolly pine ®bers generated by (a) chemical maceration with hydrogen peroxide±acetic acid solution and (b) disk re®ned at a pressure of 4 bar. (Courtesy of R. Snell.)

232

Moss and Groom

Fig. 46 Intermittent contact mode 5 mm scan of the cell wall of Valonia ventricosa, clearly displaying the cellulose micro®brils. (Courtesy of S. Zauscher.)

micro®brils revealed at the higher magni®cations. A similar display of micro®brillar structure in a recycled wood ®ber is also demonstrated in Fig. 48. Imaging in air of cellulose micro®brils, as is the case with any hydrophilic specimen, becomes dif®cult at nanometer-scale resolution due to moisture present on the stylus and the specimen surface. These capillary layers of moisture coalesce as the stylus approaches the sample surface, resulting in attractive forces that distort or

Fig. 47 Intermittent contact mode images of processed cotton ®ber showing (a) a 6 mm scan with insert box magni®ed in (b) for 1.7 mm scan, and insert box there magni®ed in (c) for 530 nm scan. Longitudinal ®ber axis runs from left to right.

Microscopy

233

Fig. 48 Recycled (old newsprint) ®ber as imaged in air operated in intermittent contact mode at (a) a 15 mm scan, with the 1.6 mm box reimaged in (b). Micro®brils, as well as contaminants, are clearly de®ned.

destroy image quality [334]. The problem of undesirable force interactions was circumvented by Weisenhorn et al. [350] and Hansma et al. [118]. Hansma et al [118] submersed DNA samples in liquid to acquire extremely high resolution images. Weisenhorn et al. [350] showed that immersion in liquid reduced the tip-off forces by a factor of approximately 100. Although liquid submersion produces high resolution images, the technique has several disadvantages with regard to sample adhesion, contaminants, and cantilever±liquid interactions. Hanley et al. [114] were among the ®rst researchers to obtain extremely high resolution of Valonia cellulose. The researchers were able to resolve the glucose repeating distance on the cellulose surface by imaging in air. By contrast, these images were inferior in resolution compared with recent work by Baker [11] and Baker et al. [12,13]. These recent publications relied on high magni®cation aqueous (water and propanol) contact mode AFM to image native Valonia cellulose I microcrystals (Fig. 49). The AFM images were able to resolve both the intermolecular spacing of 0.5±0.6 nm between the cellulose chains (depending on the crystal face) and the glucose subunit interval of 0.52 nm along individual chains, shown in Figs. 50a and 50b, respectively. The high levels of magni®cation have also led to the physical identi®cation of the various phases of the cellobiose unit. The AFM image shown in Fig. 51 shows cellulose chains of Valonia, with con®rmation of the cellobiose unit achieved with Fourier transforms of the section data. Baker et al. [12] showed that the diagonal arrangement present in Fig. 51, due to the presence of a large O-6 hydroxymethyl group on the O-5±C-5 face of the glucose ring, is indicative of the Ia (triclinic) structure. The chains would have been in a more staggered arrangement if they were oriented in a I (monoclinic) structure [7,313]. Figure 51 shows evidence of the Ia (triclinic) structure over localized regions of the surface. Recent images have shown that this Ia phase exists over large areas of the surface and that the precise crystal face can be determined [13]. One of the most distinctive aspects of this work

234

Moss and Groom

Fig. 49 A typical contact mode topographic image of microcrystals of Valonia cellulose immersed in propanol. The long straight microcrystals have a cross section of approximately 20  20 nm, although some are much smaller but still easily distinguished in the AFM. (From Ref. 11.)

Fig. 50 (a) Contact mode de¯ection image showing molecules of cellulose aligned parallel to the double-headed white arrow. The spacing between the molecules is approximately 0.6 nm, and indicated by the M in the Fourier transform (insert). (b) An image showing the 0.5 nm glucose subunit interval along the chains. The Fourier transform shows both the glucose (G) and cellobiose (C) intervals of 0.52 and 1.04 nm, respectively. (From Ref. 11.)

Microscopy

235

Fig. 51 (a) Contact mode de¯ection image of the Valonia cellulose surface immersed in propanol. The cellulose chains run almost vertically in the image. The line section in (b) has been taken between the two arrows in (a) and shows that the spacing between the bright spots is 1.09 nmÐthe cellobiose interval. This measurement is shown in the Fourier transform in (c), taken along the line pro®le. The features along the molecules are arranged diagonally to the chain direction between adjacent molecules. This is characteristic of the Ia (triclinic) phase of cellulose I and is seen most clearly in the region near the two arrows. (From Ref. 11.)

was the achievement of the highest resolution AFM imaging of a biological specimen. High resolution AFM images of structural cellulosic ®bers are more limited than those for Valonia. A liquid cell ®lled with double-distilled water was used to image the micro®brils shown in Fig. 52 [256]. The individual micro®brils are of native cotton with the cuticle removed and are of suf®cient detail to show individual crystallites. Hanley and Gray [115] were the ®rst authors to image wood ®bers in a liquid cell (deionized water) showing the micro®brils of an unbleached beaten kraft pulp ®ber in an approximately 1:4  1:4 mm scan. A subsequent image of unbleached beaten kraft pulp by Hanley and Gray [116] is shown in Fig. 53. The scan size is approximately 250  250 nm and shows a marked increase in image quality. Vertical resolution of AFMs are in fractions of angstroms. Lateral resolution, regardless of operating mode or imaging medium, is based on two factors: step size and tip radius. The step size is controlled by the user, with smaller scan sizes resulting in greater lateral resolution. The radius of commercially produced tips can vary widely, but the tips generally possess a curvature of radius of 5±50 nm. Silicon (intermittent contact) tips are at the lower end of the curvature scale, whereas silicon

236

Moss and Groom

Fig. 52 Cellulose micro®brils of a cotton ®ber treated with 6% NaOH at 518C and subsequently rinsed in acetic acid and acetone. This high magni®cation image was produced by conducting the intermittent contact in a liquid cell ®lled with double-distilled water. (Courtesy of T. C. Pesacreta and B. Triplett.)

Fig. 53 AFM image of micro®brillar material produced by beating unbleached kraft ®bers dried down on a glass surface.The arrow, 50 nm long, points to an element with a measured height of only 2 nm. (From Ref. 116.)

Microscopy

237

nitride (contact) tips are at the higher end. The lateral resolution of an AFM with the  sharpest commercially produced tips is around 10±20 A . Resolution limits based on tip radius have been expanded with the advent of microfabricated tips, either through adhesively attached nanotubes [58] or by carbon deposition in an SEM [162]. An example of the resolution capabilities of a commercially produced tip is shown in Fig. 54, which shows oxygen atoms imaged in intermittent contact (air) mode on the surface of freshly cleaved mica. Surface Feature Quanti®cation Surface and pro®le studies of cellulosic ®bers require some degree of objectivity for valid statistical comparisons to be drawn, and this is accomplished through collection and analysis of quanti®able data. The AFM is ideally suited for this as the collected data are simply an x±y±z array of tip location. This array can then be analyzed quantitatively by using a series of algorithms that display speci®c morphometric details as well as overall surface topography. Morphology Routine application of the AFM in surface morphometry has been carried out primarily in the ®eld of polymer science to quantify features from banded polymer texture [89] to contact angles [163]. The application of morphometric algorithms to cellulosic ®bers has been less forthcoming, its usefulness generally restricted to characterization of nanostructure and three-dimensional features.

Fig. 54 AFM image (height data, intermittent contact mode) showing a 5 nm scan of an oxygen lattice on freshly cleaved mica. (Courtesy of T. C. Pesacreta.)

238

Moss and Groom

Nanostructure characterization is generally conducted with an algorithm referred to as a section analysis, which produces a two-dimensional pro®le along a user-de®ned digital slice. This algorithm is demonstrated in Fig. 55 for a loblolly pine ®ber previously described in Fig. 44b. Although this section analysis algorithm yields additional information regarding surface roughness, a section analysis generally depicts two-dimensional topography along with dimensions of speci®c features. In this case, the analysis shows that the micro®bril between the vertical arrows has a diameter of 117 nm. Another application of morphometric analysis to cellulosic ®bers is the threedimensional aspect of the ®bers under investigation. Figure 56 shows AFM height images of an unbeaten, unbleached kraft pulp ®ber taken during a hydration and dehydration series. The height images look identical in both the hdyrated and dehydrated states even though the cell wall dimensions have changed. This change in cell wall dimensions can be seen in the section analysis (Fig. 57), which then allows us to quantitatively assess the effect of moisture on such changes. Surface Roughness Of primary importance in the study of wood-®ber-based composites is the bond energy that is established between individual wood ®bers. Fiber± ®ber bond strength is generally estimated in the pulp and paper industry by several methods that evaluate the internal bond strength of paper and paperboards (Chapter 15, Vol. 1). These techniques are summarized by Koubaa and Koran [176]. The bulk of these tests measure the splitting resistance of paper or paperboard to a force applied normal to its plane [298,299]. These techniques determine bulk properties,

Fig. 55 Section analysis of the S2 layer of a macerated loblolly pine latewood ®ber. The primary wall, S1, and a portion of the S2 layer were physically removed with a microknife and forceps.The section analysis shows a cross-sectional view (upper left) of height data from a user-de®ned segment (lower left) oriented perpendicular to the micro®bril direction.

Microscopy

239

Fig. 56 AFM images of an unbeaten, unbleached kraft pulp ®ber. The same ®ber was imaged at ambient room conditions in air (a) after being allowed to dry down onto a glass cover slip. The ®ber was then imaged in deionized water (b) in a liquid cell. The ®ber was allowed to dry in ambient room conditions and imaged again (c). Each of the ®ber images shown is a montage of four AFM images. The images are shown in the height mode with the x, y scale indicated by the upper bar and the height information indicated by the lower bar. (From Ref. 116.)

Fig. 57 Section pro®les perpendicular to the same position on the ®ber axis of the ®ber shown in Fig. 56. Conditions for (a), (b), and (c) are the same as for Fig. 56. Note that the z scale is twice the width scale. (From Ref. 116.)

240

Moss and Groom

so they yield little information on individual ®ber±®ber bonding energies. An analogous situation is trying to measure individual ®ber mechanical properties by handsheet properties such as zero-span testing (Chapters 7 and 14, Vol. 1). Bond strengths within ®ber networks are a function of the ``effective external surface area'' [46]. Fiber surface area has been determined in a myriad of manners including pulp permeability to the ¯ow of water or air, stock freeness, and sorption of liquid nitrogen [46]. However, none of these techniques have been able to quantify the surface roughness of individual cellulose ®bers, which presents an opportunity for the AFM to provide data that will ®ll this void in the literature. Currently there exist only two studies on evaluation of the surface roughness of cellulosic ®bers using an AFM. Karlsson et al. [158] evaluated the surface roughness of cotton and wood ®ber surfaces and found that there was a strong relationship, with favorable interphases in cellulose ®ber±polymer composites. Pesacreta et al. [256] conducted a much more thorough study, focusing on the effect of cuticle removal of cotton ®bers as determined by the surface roughness. Cotton ®bers were evaluated in the AFM before and after removing the cuticle (Figs. 58 and 59). The surface roughness of the virgin and treated ®bers were characterized on the basis of the following analyses: root-mean-square (RMS), spectral density, and fractral. RMS is the square of the deviations of height data from the central plane, an RMS algorithm being analogous to the least squares method commonly employed in statistics. Spectral density is extracted from a Fourier transform analysis and generallly employs some type of graphic representation of feature distributions. The analysis reveals periodic surface features that might otherwise appear random. Pesacreta et al. [256] also employed the analysis of fractals, which relies on surface ``self-similarity'' over a wide range of scales of magni®cation [95]. Fractals are based on a numerical algorithm that evaluates the surface's geometrical complexity as determined by an iterative approach relating fractal cell size and frequency. A more detailed explanation of fractals can be found in Mandelbrot et al. [199].

Fig. 58 AFM intermittent contact mode 2.5 mm scans of (a) untreated cotton ®ber and (b) cotton ®ber with the cuticle removed by an Updegraff reagent. (From Ref. 256.)

Microscopy

241

Fig. 59 A fractal analysis of the cotton ®bers in Fig. 58, with (a) an untreated ®ber and (b) a ®ber with the cuticle removed. The primary number of importance is the fractal dimension, with the lower number expressing a ``rougher'' surface. (From Ref. 256.)

242

Moss and Groom

Pesacreta et al. [256] found that surface roughness increased as a result of removing the cuticle and that the RMS method was the least sensitive to surface roughness and fractals the most sensitive. It should be noted that this was not the ®rst time fractals were applied to wood ®ber research. A fractal analysis was conducted by Weise [349] to quantify ®ber wrinkling with the CLSM. Nanomechanical Properties Most AFMs now have the capability to monitor forces applied to the surface of specimens and, coupled with tip displacement data, are able to produce plots of force versus sample deformation. There are two general methods for ascertaining the surface moduli. One method applies displacements to the specimen surface and infers the force from the spring constant of the cantilever [127]. This is the most common method for force modulation, because no machine modi®cations are necessary and it allows the user to create a modulation map of the specimen surface. The operator does need the spring constant of the cantilever, either from the manufacturer or through calibration such as the technique described by Hutter and Bechhoefer [133]. A second force modulation method [146] probes the surface with known forces rather than displacements, allowing for absolute determination of contact compliance. This method eliminates erroneous readings due to secondary attractive or repulsive forces. The method indents the surface and is thus destructive, allowing only point location values for hardness and stiffness. The second force modulation method was used in the only extensive studies of nanomechanical properties of wood ®bers that have been carried out [362,363]. Wimmer and coworkers determined the stiffness and hardness values of the S2 layer and the corner middle lamella of red spruce (Picea rubens Sarg.). The studies showed that the modulus of elasticity of the S2 layer was approximately twice that of the corner middle lamella (Fig. 60). As mentioned previously, the nanoindentation method of force modulation allows point location values of hardness and stiffness to be determined but does not allow for surface mapping. The reduction in accuracy of the stiffness mapping method of force modulation is compensated for by the vast number of data points and the ability to compare mechanical properties of surface features. However, the applicability of the force modulation technique in wood and ®ber science is limited because of its inability to distinguish materials of similar stiffness. Figure 61 shows a surface stiffness map obtained by force modulation of a cross-sectional loblolly pine specimen of the latewood band in the 32nd growth ring. Distinctions in moduli are not realized owing to the relative similarity in stiffnesses across the cross-sectional area. Wimmer and Lucas [362] found that the S2 layer was only 2.9 times as stiff as the cell corner middle lamella. Although unable to adequately distinguish between the middle lamella and the primary/secondary walls, force modulation was able to accentuate ®brillar disruptions otherwise concealed and thus may be more useful for evaluating local disruptions. Material Interactions and Compatibility One of the newest applications of the AFM makes use of its ability to detect subtle changes in the vertical and lateral forces that act on the cantilever tip in response to changes in chemical and physical surface composition. The most prominent and, to the pulp and paper industry, most useful of these applications are described in detail in the following paragraphs.

Microscopy

243

Fig. 60 Nanoindentation of spruce, where In = indentation, S1 = S1 layer of the secondary cell wall, S2 = S2 layer of the secondary cell wall, P = primary cell wall, ML = middle lamella, and CC = cell corner middle lamella. (From Ref. 362.)

Fig. 61 Thirty-second growth ring of the latewood band of a loblolly pine specimen as shown by 17 mm scans. Diagonal lines are knife marks as a result of the hand sectioning. The cross-sectional view shows three thick-walled tracheids separated by a middle lamella, with the images displaying (a) height data and (b) force-modulation data.

244

Moss and Groom

Phase Imaging Phase imaging is conducted in intermittent contact mode and measures the phase lag of the cantilever oscillation relative to the oscillation of the drive piezo. Spatial variations in sample elastic and topographical properties cause phase shifts in the cantilever phase. These shifts are, in part, indicative of the local viscoelastic properties of the sample surface, with the out-of-phase component of the response amplitude proportional to the loss modulus [262]. The angular deviation from the excitation waveform is used to determine the phase shift, and the overall methodology is generally referred to as either phase contrast imaging or phase imaging. Phase imaging serves two general functions in the analysis of cellulosic ®bers. First, it allows the detection of different surface components based on the degree of phase shift. The primary application for this type of analysis seems most ideally suited for the identi®cation of local lignin and cellulosic components on ®ber surfaces. A study of this type was conducted by Pereira and da Silva [255] in which they were able to identify lignin and cellulose components on the surface of four different pulp types, due primarily to the inherent differences in viscosity between these two primary cell wall constituents. They were also able to further increase contrast by modifying the tips with various functional groups. The usefulness of this application is not limited to identi®cation of lignin and cellulose. Figure 62 shows the application of phase imaging for the distribution of coating components, the mineral pigment having a different hardness than the polymeric binder. Note that both edge detection of the boundaries and material hardness can be seen in the phase image. The second use of phase imaging lies in contrast enhancement. The images that result from phase imaging are higher in contrast than height images. This is due in part again to changes in the viscoelastic surface properties. However, phase imaging is also sensitive to surface topography, with even slight variations in topography

Fig. 62 AFM images (intermittent contact mode, 3 mm scans) of coated paper showing (a) conventional height data and (b) phase contrast imaging. (Courtesy of L. Mattson, L. Krummenacher, and M. C. BeÂland.)

Microscopy

245

altering the amount of surface area for attractive and repulsive tip±samples forces. This increase in contrast can be seen in Fig. 63, in which the phase image highlights the cellulosic micro®brils on the cell wall of a highly deligni®ed softwood. Lateral Force Microscopy Lateral force microscopy, also called frictional force microscopy, is a technique that identi®es and maps relative differences in frictional characteristics between the tip and sample surfaces. Lateral force microscopy measures the torsional ¯ex of the cantilever resulting from frictional forces during rastering of a surface. Cantilever de¯ection and twist are determined in most commercial AFMs with a quadrant photodetector from spatial changes in laser light intensity that is re¯ected off the tip of the force-sensing cantilever. Lateral force microscopy is useful for identifying surface compositional differences when the materials have differing frictional characteristics. It can thus be used to differentiate between lignin and cellulose. However, caution should be exercised in the interpretation of lateral force microscopic images of wood ®bers because of the possibility of errors being introduced as the result of extreme surface topography. A modi®cation of lateral force microscopy is chemical force microscopy, in which the tip is chemically modi®ed with a functional group and scanned over a sample to detect compatibility. Current research by Zauscher and Klingenberg at the University of Wisconsin, Madison, is looking at the frictional forces between cellulose surfaces in solutions of water-soluble polymers. To study cellulose±cellulose interactions these researchers adapted colloidal probe microscopy, a technique ®rst developed by Ducker et al. [72] for the measurement of normal forces between silica surfaces. In this technique a micromanipulator is used to adhere a cellulose bead with an approximate diameter of 20 mm to the end of a force-sensing cantilever with a UV-curing adhesive. Collidal probe microscopy thus allows the study of interactions on a model system that mimics a single ®ber contact point. Figure 64 shows the friction force±distance traces and the corresponding height±distance traces that

Fig. 63 Intermittent contact mode 5 mm image of a highly deligni®ed softwood cell wall showing (a) amplitude and (b) phase shift. (Courtesy of S. Zauscher.)

246

Moss and Groom

Fig. 64 Friction force and corresponding surface topography plotted as a function of sliding distance for two cellulose surfaces sliding past each other. (Courtesy of S. Zauscher.)

result when a cellulose-modi®ed tip is scanned across the surface of a regenerated cellulose ®lm. It is evident that the friction forces are correlated to surface topography. Frictional forces are dramatically reduced in the presence of water-soluble polymer solutions that are injected into the ¯uid cell of the AFM (Fig. 65). These results are directly applicable to papermaking processes in which the addition of water-soluble polymers is used to change pulp suspension rheology. It is clear that colloidal probe microscopy, applied to cellulosic systems, is a powerful tool; it may enhance through its access to nanonewton forces on nanometer length scales the understanding of friction between ®bers, of particle retention on ®bers, and of ®ber bonding.

Fig. 65 Friction force between two cellulose surfaces plotted as a function of polymer concentration. (Courtesy of S. Zauscher.)

Microscopy

VII.

247

CONCLUSIONS

For many of these different types of microscopy, but especially in the case of LTSEM and CLSM studies, it is biologists who have done the groundwork and so it is to their ®ndings that one has to turn initially for information. Hence the large number of references in this chapter to papers in biological journals. The choice of microscope is important, and the effects of preparative techniques have also to be borne in mind. Ultimately, the choice of microscopical technique depends on the material being examined and the resolution required. No single microscope is suitable for all purposes, and novel types of microscopes (i.e., CLSM, LTSEM, ESEM and AFM) do not replace the more traditional instruments, as evidenced by the fact that the established techniques continue to be used and new methods of microscopical examination are still being developed for them. Although the new microscopes open up all sorts of opportunities they should be regarded as being complementary to the older techniques. Samples need to be examined under conditions closely related to their natural state, and optical microscopy is ideal for observing fully hydrated pulp and paper samples. Accurate evaluation of ®ber morphology is not easy when ®bers have undergone the dehydration processes required for electron microscopy. However, the resolution of the optical microscope is inadequate to reveal detailed structural information. The CLSM offers little improvement in resolution over the conventional optical microscope. Its main advantage lies in its unique ability to visualize cross-sectional views of samples by nondestructive optical sectioning. The LTSEM, ESEM, and AFM, on the other hand, all offer possibilities of examining the wet structure of pulp and paper at high resolution without the need for dehydration and embedding techniques but can image only surfaces. All microscopical studies, by de®nition, enable examination of only very small areas of samples. They are also time-consuming. Thus speed of data collection is important because large amounts of data need to be collected for quantitative and statistical analysis. Microscopy has played a very important role in the past, and these new instruments, linked to image analysis systems, should ensure that it continues to be a necessary part of ®ber and paper research in the future. In conclusion, the words of Emerton and Watts [82] of more than 40 years ago can be reiterated: Progress in the ®eld of ®ber microscopy lies in the use of a prudent combination of the many techniques now available . . . rather than in the exclusive use of one technique.

ABBREVIATIONS AD AFM BEI BSE BSI CFSM CLSM

air drying atomic force microscope backscattered electron imaging backscattered electron backscattered electron imaging confocal ¯uorescence scanning microscope confocal laser scanning microscope

248

Moss and Groom

CPD CSM CSOM CTMP EDS EDX EDXA EM ESEM F FD FH FPR, FRAP LSM LTSEM LVHRSEM N.A. NSOM RBA REM SE SEM SLM SOM STEM STM TCF TEM TMP TSM, TSRLM UV

critical point drying confocal scanning microscope confocal scanning optical microscope chemithermomechanical pulp energy-dispersive X-ray spectrometry energy-dispersive X-ray energy-dispersive X-ray analysis electron microscope environmental scanning electron microscope ¯at®eld freeze-drying frozen hydrated ¯uorescence recovery after photobleaching laser scanning microscope low temperature scanning electron microscope low voltage high resolution SEM numerical aperture near-®eld scanning optical microscopy relative bonded area re¯ection electron microscope secondary electron scanning electron microscope scanning laser microscope scanning optical microscope scanning transmission electron microscope scanning tunnelling microscope totally chlorine-free transmission electron microscope thermomechanical pulp tandem scanning re¯ection light microscope ultraviolet

REFERENCES 1. 2. 3. 4. 5. 6.

Abitz, P., and Luner, P. (1989). The effect of re®ning on wet ®ber ¯exibility and its relationship to sheet properties. In: Fundamentals of Papermaking, Vol. 1. C.F. Baker and V. W. Punton, eds. Mechanical Engineering Publications, London, pp. 67±86. Alexander, S. D., Marton, R., and McGovern, S. D. (1968). Effect of beating and wet pressing on ®ber and sheet properteis. 1. Individual ®ber properties. Tappi 51(6):277± 283. Amboss, K., Emerton, H. W., and Walls, J. (1954). A new approach to the electron microscopy of ®bers. Proc.Tech. Sect. BPBMA 35(3):487±510. Asunmaa, S. K. (1966). The ®ber wall, micro®brils, and interfaces at electron microscopic resolution. Tappi 49(7):319±324. Asunmaa, S. K., and Marteny, W. W. (1963). Morphology of re®ned pulps of southern pine (longleaf pine) and black ash. Tappi 46(10):613±622. Asunmaa, S. K., and Steenberg, B. (1958). Beaten pulps and the ®ber-to-®ber bond in paper. Svensk Papperstidn. 61(18B):685±695.

Microscopy 7.

249

Atalla, R. H., and VanderHart, D. L. (1984). Native cellulose: A composite of two distinct crystalline forms. Science 223:283±285. 8. Attree, S. M., and Shef®eld, E. (1984). Scanning electron microscopy of protoplasts isolated from gametophytes of the fern Pteridium. I. Preparative methods. Micron Microsc. Acta 15(3):181±186. 9. Bailey, I. W. (1938). Cell wall structure of higher plants. Ind. Eng. Chem. 30(1): 411±417. 10. Bailey, I. W., and Kerr, T. (1935). The cambium and its derivative tissues. X. Structural, optical properties and chemical composition of the so-called middle lamella. J. Arnold Arboretum 16:273±300. 11. Baker, A. A. (1998). High resolution atomic force microscopy of polysaccharides. Ph.D. Thesis, University of Bristol, U.K. 12. Baker, A. A., Helbert, W., Sugiyama, J., and Miles, M. J. (1997). High resolution atomic force microscopy of native Valonia cellulose I microcrystals. J. Struct.Biol. 119(2):129±138. 13. Baker, A. A., Helbert, W., Sugiyama, J., and Miles, M. J. (1998). Surface structure of native cellulose microcrystals by AFM. Appl. Phys. AÐMat.Sci.Process. 66(PtISS):S559±S563. 14. Batchelor, W. J., Conn, A. B., and Parker, I. H. (1997). Measuring the ®bril angle of ®bers using confocal microscopy. Appita 50(5):377±380. 15. Beckett, A., and Read, N. D. (1986). Low-temperature scanning electron microscopy. In: Ultrastructure Techniques for Microorganisms. H. C. Aldrich and W. J. Todd, eds. Plenum Press, New York, pp. 45±86. 16. BeÂland, M.-C. (1996). CLSM and AFM applied in pulp and paper research. A literature review. IOF-Tech. Rep. TR 312, Inst. Opt. Res., Stockholm, Sweden, 18 p. 17. BeÂland, M.-C., and Mangin, P. J. (1995). Three-dimensional evaluation of paper surfaces using confocal microscopy. In: Surface Analysis of Paper. T. E. Conners and S. Banerjee, eds. CRC Press, Boca Baton, FL, pp. 1±40. 18. Bergander, A., and SalmeÂn, L. (1997). Lamellar cracks, do they appear during radial compression of wood? Nordic Pulp Paper Res. J. 12(4):216±219. 19. Bergh, N. O., and Thomin, W. H. (1976). Practical aspects of paper/liquid interaction during paper coating. In: The Fundamental Properties of Paper Related to Its Uses, Vol. 2. F. Bolam, ed. British Paper and Board Industry Fed., London, pp. 496±510. 20. Binnig, G., Quate, C. F., and Gerber, Ch. (1986). Atomic force microscope. Phys. Rev. Lett. 56(9):930±933. 21. Binnig, G., and Rohrer, H. (1982). Scanning tunneling microscopy. Helv. Phys. Acta 55(6):726±735. 22. Binnig, G., Rohrer, H., Gerber, Ch., and Weibel, E. (1982). Surface studies by scanning tunneling microscopy. Phys. Rev. Lett. 49(1):57±61. 23. Blanchette, R. A., Akhtar, M., and Attridge, M. C. (1992). Using Simons stain to evaluate ®ber characteristics of biomechanical pulps. Tappi J. 75(11):121±124. 24. Blonk, J. C. G., Don, A., Van Aalst, H., and Birmingham, J. J. (1993). Fluorescence photobleaching recovery in the confocal scanning light microscope. J. Microsc. 169(3):363±374. 25. Borchardt, J. K., Matalamaki, D. W., Lott, V. G., and York, G. A. (1994). Deinking of toner ink containing furnishes. Part 2. Deinking with mechanical cleaners. Prog. Paper Recycl. 4(1):44±56. 26. Boutelje, J., and Eriksson, I. (1984). Analysis of lignin in fragments from thermomechanical spruce pulp by ultraviolet microscopy. Holzforschung 38:249±252. 27. Bown, R. (1997). A review of the in¯uence of pigments on papermaking and coating. In: The Fundamentals of Papermaking Materials, Vol. 1. C. F. Baker, ed. Pira International, Leatherhead, U.K., pp. 83±137.

250

Moss and Groom

28. Boyde, A. (1974). Freezing, freeze-fracturing and freeze-drying in biological specimen preparation for the SEM. In: Scanning Electron Microscopy. O. Johari and I. Corvin, eds. IIT Res. Inst., Chicago, pp. 1043±1046. 29. Boyde, A. (1978). Pros and cons of critical point drying and freeze drying for SEM. In: Scanning Electron Microscopy/II. O. Johari and I. Corvin, eds. IIT Res. Inst., Chicago, pp. 303±314. 30. Boyde, A., and Franc, F. (1981). Freeze-drying and shrinkage of glutaraldehyde ®xed liver. J. Microsc. 122:75±86. 31. Browne, T. C., Crotogino, R. H., and Douglas, W. J. M. (1995). The effect of paper structure on behavior in a calender nip. J. Pulp Paper Sci. 21(10):J343±J347. 32. Buchanan, J. G., and Lindsay, R. A. (1962). Structure of paper as revealed by the scanning electron microscope. In: Formation and Structure of Paper. Vol. 1. F. Bolam, ed. British Paper and Board Makers' Assn., London, pp. 101±117. 33. Buchanan, J. G., and Washburn, O. V. (1962). The surface and tensile fractures of chemical ®ber handsheets as observed with the scanning electron microscope. Pulp Paper Mag. Can. 63(10):T485±T493. 34. Buchanan, J. G., and Washburn, O. V. (1964). The surface and tensile fractures of groundwood handsheets as observed with the scanning electron microscope Pulp Paper Mag. Can. 65(2):T52±T60. 35. Bucher, H. (1958). Discontinuities in the microscopic structure of wood ®bers. In: Fundamentals of Papermaking Fibers. F. Bolam, ed. British Paper and Board Makers' Assn., Kenley, England, pp. 7±34. 36. Buetow, D. E. (1968). The Biology of Euglena. Academic Press, New York. 37. Burgess, J., Linstead, P. J., and Harnden, J. M. (1977). The interpretation of scanning electron micrographs. Micron 8:181±191. 38. Butter®eld, B. G., and Meylan, B. A. (1980). Three-dimensional Structure of Wood. An Ultrastructural Approach. Chapman and Hall, London. 39. Campbell, W. B. (1933). The cellulose-water relationship in paper-making. Forest Service Bulletin 84, Department of the Interior, Canada. 40. Carlsson, G., LindstroÈm, T., and SoÈremark, C. (1978). Expression of water from cellulosic ®bers under compressive loading. In: Fiber-Water Interactions in Paper-Making, Vol. 1. Fundamental Research Committee, ed. British Paper and Board Industry Fed., London. pp. 389±402. 41. Catling, D., and Grayson, J. (1982). Identi®cation of Vegetable Fibers. Chapman and Hall, London. 42. Centonze, V., and Pawley, J. (1995). Tutorial on practical confocal microscopy and use of the confocal test specimen. In: Handbook of Biological Confocal Microscopy. J. B. Pawley, ed. Plenum Press, New York, pp. 549±569. 43. Chapman, J. A., and Menter J. W. (1954). A study of the shape, surface structure and frictional wear of ®bers by re¯ection electron microscopy. Proc. Roy. Soc. (Lond.), Ser. A 226:400±407. 44. Christiani, C., and Bristow, A. (1995). The drying mechanisms of water-borne printing inks. Proc. 47th TAGA Conf., pp. 231±227. 45. Clark, J. d'A. (1943). The nature of hydration and ®ber bonding. Proc. Tech. Sect. BPBMA 24:30±54. 46. Clark, J. d'A. (1978). Surface measurements. In: Pulp Technology and Treatment for Paper. Miller Freeman, San Francisco, pp. 533±540. 47. Clarke, B. (1972). Furnish blend optimization and evaluation. Papermaker 165(4):41. 48. Cogswell, C. J. (1995). Imaging immunogold labels with confocal microscopy: In: Handbook of Biological Confocal Microscopy. J. B. Pawley, ed. Plenum Press, New York, pp. 507±512. 49. Comer, J. J., Stetson, H. W., and Lyons, S. C. (1955). A new replica technique for making electron micrographs of surfaces of paper sheets. Tappi 38(10):620±624.

Microscopy 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72.

251

Coppick, S., and Fowler, W. F. (1939). The location of potential reducing substances in woody tissue. Paper Trade J. 109 (Sept. 14):TS135±TS140. Core, H. A., CoÃteÂ, W. A., and Day, A. C. (1979). Wood Structure and Identi®cation. Syracuse Wood Sci. Ser. 6. Syracuse Univ. Press, Syracuse, NY. CoÃteÂ, W. A. (1967). Wood UltrastructureÐAn Atlas of Electron Micrographs. Univ. Washington Press, Seattle. CoÃteÂ, W. A. (1970). The ultrastructural organization of wood and wood ®bers. In: Physics and Chemistry of Wood Pulp Fibers. D. H. Page, ed. STAP Ser. No. 8, TAPPI, Atlanta, GA, pp. 1±18. CoÃteÂ, W. A. (1980). Papermaking Fibers. A Photomicrographic Atlas. Syracuse Univ. Press, Syracuse. NY. CoÃteÂ, W. A., Koran, Z., and Day, A. C. (1964). Replica techniques for electron microscopy of wood and paper. Tappi 47(8):477±484. Cox, G. (1995). Mass storage and hard copy. In: Handbook of Biological Confocal Microscopy. J. B. Pawley, ed. Plenum Press, New York, pp. 535±548. Crosby, C. M., and Mark, R. E. (1974). Precise S2 angle determination in pulp ®bers. Svensk Papperstidn. 77:636±642. Dai, H., Hafner, J. H., Rinzler, A. G., Colbert, D. T., and Smalley, R. E. (1996). Nanotubes as nanoprobes in scanning probe microscopy. Nature 384:147±150. Danilatos, G. D. (1993). Introduction to the ESEM instrument. Microsc. Res. Techn. 25:354±361. Danilatos, G. D. (1993). Bibliography of environmental scanning electron microscopy. Microsc. Res. Techn. 25:529±534. Davies, G. W. (1966). Optical and electron microscopy of NSSC pulps. Appita 19(4):95± 110. Davies, G. W. (1966). Optical and electron microscopy of paper made from chemical pulps. Appita 20(1):21±30. Davies, G. W. (1968). A microscopic study of paper made from Pinus radiata. Tappi 51(10):454±461. DeNee, P. B., and Abraham, J. L. (1976). Backscattered electron imaging (application of atomic number contrast). In: Principles and Techniques of Biological Applications, Vol. 5. M. A. Hayat, ed. Van Nostrand Reinhold, New York, pp. 144±180. Dennis, D. T., and Colvin, J. R. (1964). Biosynthesis of cellulose. Pulp Paper Mag. Can. 65(9):T395±T399. de Silveira, G., Forsberg, P., and Conners, T. (1995). Scanning electron microscopy. A tool for analysis of wood pulp ®bers and paper. In: Surface Analysis of Paper. T. E. Conners and S. Banerjee, eds. CRC Press, Boca Raton, FL, pp. 41±71. de Silveira, G. Zhang, X., Berry, R., and Wood, J. R. (1996). Location of ®nes in mechanical pulp handsheets using scanning electron microscopy. J. Pulp Paper Sci. 22(9):J315±J320. Dixon, A. E., and Cogswell, C. (1995). Confocal microscopy with transmitted light. In: Handbook of Biological Confocal Microscopy. J. B. Pawley, ed. Plenum Press, New York, pp. 479±487. Dixon, A. E., Damaskinos, S., and Atkinson, M. R. (1991). A scanning confocal microscope for transmission and re¯ection imaging. Nature 351:551. Dixon, A. E., Damaskinos, S., Ribes, A., and Beesley, K. M. (1995). A new confocal scanning beam laser MACROscope using a telecentric, f-theta laser scan lens. J. Microsc. 178(3):261±266. Dlugosz, J. (1965). The ®ne structure of cotton ®ber as revealed by swelling during methacrylate embedding. Polymer 6(8):427±436. Ducker, W. A., Senden, T. J., and Pashley, R. M. (1992). Measurement of forces in liquids using a force microscope. Langmuir 8:1831±1836.

252

Moss and Groom

73. Dunlop, W. F., and Robards, A. W. (1972). Some artifacts of the freeze-etching technique. J. Ultrastruct. Res. 40:391±400. 74. Dwivedi, A. K., and Ahmad, K. J. (1985). Scanning electron microscopy of fresh, uncoated plant parts. Micron Microsc. Acta 16(1):55±57. 75. Echlin, P. (1971). The examination of biological material at low temperatures. In: Scanning Electron Microscopy/I O. Johari and I. Corvin. eds. IIT Res. Inst., Chicago, pp. 225±232. 76. Echlin, P., Paden, R., Dronzek, B., and Wayte, R. (1970). Preparation of labile biological material for examination in SEM. In: Scanning Electron Microscopy. O. Johari and I. Corvin, eds. IIT Res. Inst., Chicago, p. 307. 77. Elton, N. J., Hooper, J. J., and Gane, P. A. C. (1990). Optical microscopy of paper. Methodology and applications. Proc. 24th EUCEPA Conf., Stockholm, pp. 438±458. 78. Emerton, H. W. (1957). Fundamentals of the Beating Process. British Paper and Board Industry Assoc., Kenley, England. 79. Emerton, H. W. (1958). The outer secondary wall. 1. Its structure. In: Fundamentals of Papermaking Fibers. F. Bolam, ed. British Paper and Board Makers' Assoc., Kenley, England, pp. 35±54. 80. Emerton, H. W., Page, D. H., and Hale, W. H. (1962). The structure of papers as seen in their surfaces. In: Formation and Structure of Paper. Vol. 1. F. Bolam, ed. British Paper and Board Makers' Assoc., London, pp. 53±99. 81. Emerton, H. W., Page, D. H., and Watts, J. (1956). The uses of solid replicas for studying ®ber and paper surfaces. Proc.Tech. Sect. BPBMA 37(1):105±125. 82. Emerton, H. W., and Watts, J. (1953). Examination of spruce tracheids under polarized vertical illumination. Proc. Tech. Sect. BPBMA 34(2):269±287. 83. Enomae, T., Onabe, F., and Usuda, M. (1993). Application of new pro®lometry using topographic scanning electron microscope to paper surface topography. Tappi J. 76(1):85±90. 84. Eriksson, I., Lidbrandt, O., and Westermark, U. (1988). Lignin distribution in birch (Betula verrucosa) as determined by mercurization with SEM- and TEM-EDXA. Wood Sci. Technol. 22:251±257. 85. Eveling. D. W., and McCall, R. D. (1983). An evaluation of methods for preparing easily damaged cuticular surfaces of plants for SEM. J. Microsc. 129(2):113±122. 86. Fengel, D. (1970). The ultrastructural behavior of cell wall polysaccharides. In: Physics and Chemistry of Wood Pulp Fibers. D. H. Page, ed., STAP Ser. No. 8. TAPPI, Atlanta, GA, pp. 74±99. 87. Fergus, B. J., Procter, A. R., Scott, J. A. N., and Goring, D. A. I. (1969). The distribution of lignin in sprucewood as determined by ultraviolet microscopy. Wood Sci. Technol. 3(2):117±138. 88. FernaÂndez-MoraÂn, H. (1960). Low-temperature preparation techniques for electron microscopy of biological specimens based on rapid freezing with liquid helium II. Ann. NY Acad. Sci. 85:689±713. 89. Fischer, H., Miles, M. J., and Odell, J. A. (1994). Atomic force microscopy of the banded structure of lyotropic polymers. Macromol. Rapid Commun. 15(11):815±821. 90. Fjerdingen, H., Forseth, T. F., Gregersen, é.W., Helle, T., Johnsen, P. O., Kure, K.-A., and Reme, P. A. (1997). Some mechanical pulp ®ber characteristics, their process relationships and papermaking signi®cance. In: The Fundamentals of Papermaking Materials. Vol. 1. C. F. Baker, ed. Pira Int., Leatherhead, U.K. pp. 641±662. 91. Forsberg, P., and Lepoutre, P. (1992). A new insight into the ®ber-roughening phenomenon. Nordic Pulp Paper Res. J. 7:158±159. 92. Forsberg, P., and Lepoutre, P. (1994). ESEM examination of roughening of paper in high moisture environment. International Printing and Graphic Arts Conf., Atlanta, GA, pp. 229±336.

Microscopy 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114.

253

Forseth, T., and Helle, T. (1997). Effect of moistening on cross-sectional details of calendered paper containing mechanical pulp. J. Pulp Paper Sci. 23(3):J95±J100. Frey-Wyssling, A. (1964). Ultraviolet and ¯uorescence optics of ligni®ed cell walls. In: Formation of Wood in Forest Trees. M. H. Zimmermann, ed. Academic Press, New York, pp. 153±167. Friel, J., and Pande, C. (1993). A direct determination of fractal dimension of fracture surfaces using scanning electron microscopy and stereoscopy. Mater. Res. Soc. 8(1):100±104. Fukazawa, K., and Imagawa, H. (1983). Ultraviolet and ¯uorescence microscopic studies of lignin. 2nd Int. Symp. on Wood and Pulping Chemistry, Tsukuba, Japan, Vol. 1, pp. 20±23. Furuta, T., Morikawa, Y., Ito, K., and Ise, N. (1994). Cellophane structure as studied by confocal laser scanning microscope. Tappi J. 77(8):128±131. Gallay, W. (1949). Some factors in the strength of paper. Tappi 32(10):457±462. Gane, P. A. C., and Hooper, J. J. (1989). An evaluation of the interactions between coating color and base paper by coating pro®le analysis. In: Fundamentals of Papermaking. Vol. 2. C. F. Baker, ed. Mech. Eng. Pub., London, pp. 871±893. Gibbon, D. L., Simon, E. C., and Cornelius, R. C. (1989). New electron and light optical techniques for examining papermaking. Tappi J. 72(10):87±91. Giertz, H. W. (1977). Basic wood raw material properties and their signi®cance in mechanical pulping. Proc. Int. Mechanical Pulping Conf., Helsinki, Session 1:1±15. Godjics, M. (1953). The Genus Euglena. Univ. Wisconsin Press, Madison, Wisconsin. Goldstein, J. I., Newbury, D. E., Echlin, P., Joy, D. C., Fiori, C., and Lifshin, E. (1981). Scanning Electron Microscopy and X-Ray Microanalysis. Plenum Press, New York. Goring, D. A. I. (1970). Microscopic patterns of lignin removal during chemical pulping. In: Physics and Chemistry of Wood Pulp Fibers. D. H. Page, ed. STAP Ser. No. 8. TAPPI, Atlanta, GA, pp. 107±124. Graff, J. H., and Schlosser, M. A. (1942). Cross sectioning of paper. Paper Trade J. 114:TS87±TS91. Gregersen, é. W., Johnsen, P. O., and Helle, T. (1995). Small-scale topographical variations of newsprint surfaces and their effects on printing ink transfer distribution. J. Pulp Paper Sci. 21(10):J331±J336. Gregersen, é. W., Skinnarland, I., Johnsen, P. O., and Helle, T. (1995). Qualitative methods for the study of lignin distribution in wood and surface layers of unbleached pulp ®bers and paper. J. Pulp Paper Sci. 21(8):J285±J287. Gron, J., and Beghello, L. (1996). Cracking and roughening of coated paper surfaces. Paperi Puu (Paper Timber) 78(3):121±127. Groom, L. H., Shaler, S. M., and Mott, L. (1995). Characterizing micro- and macromechanical properties of single wood ®bers. Int. Paper Physics Conf., Niagara-on-theLake, ON, Canada, pp. 13±22. Gunning, A. P., Kirby, A. R., Morris, V. J., Wells, B., and Brooker, B. E. (1995). Imaging bacterial polysaccharides by AFM. Polym. Bull. 34:615±619. Guttman, H. N., and Styskal, R. C. (1971). Preparation of suspended cells for SEM examination of internal cellular structures. In: Scanning Electron Microscopy/I. O. Johari and I. Corvin, eds. IIT Res. Inst., Chicago, pp. 265±271. Hamad, W. Y., and Provan, J. W. (1995). Microstructural cumulative material degradation and fatigue-failure micromechanisms in wood-pulp ®bers. Cellulose 2:159±177. Hanley, S. J. (1996). Application of atomic force microscopy to cellulose, wood, kraft pulp ®bers, and paper. Ph.D. Thesis, McGill Univ., Montreal, Canada. Hanley, S. J., Giasson, J., Revol, J. F., and Gray, D. G. (1992). Atomic force microscopy of cellulose micro®brils: Comparison with transmission electron microscopy. Polymer 33(21):4639±4642.

254

Moss and Groom

115. Hanley, S. J., and Gray, D. G. (1994). Atomic force microscope images of black spruce wood sections and pulp ®bers. Holzforschung 48(1):29±34. 116. Hanley, S. J., and Gray, D. G. (1999). Imaging kraft pulp ®bers in air and water by atomic force microscopy. J. Pulp Paper Sci. 25(6):196±200. 117. Hanna, R. B. (1971). The interpretation of high resolution electron micrographs of the cellulose elementary ®bril. J. Polymer Sci., Part C 36:409±413. 118. Hansma, H. G., Vesenka, J., Siege®st, C., Kelderman, G., Morrett, H., Sinsheimer, R. L., Elings, V., Bustamante, C., and Hansma, P. K. (1992). Reproducible imaging and dissection of plasmid DNA under liquid with the atomic force microscope. Science 256(5060):1180±1184. 119. Harris, T. D., Grober, R. D., Trautman, J. K., and Betzig, E. (1994). Super-resolution imaging spectroscopy. Appl. Spectrosc. 48(1):14A±21A. 120. Haugland, R. P. (1989). Molecular Probes: Handbook of Fluorescent Probes and Research Chemicals. Molecular Probes Inc., Eugene, OR. 121. Hell, S., Reiner, G., Cremer, C., and Stelzer, E. H. K. (1993). Aberrations in confocal ¯uorescence microscopy induced by mismatches in refractive index. J. Microsc. 169(3):391±405. 122. Hellawell, J. M., and Nelson, R. (1971). Scanning electron microscopy studies of paper coating structures in cross-section. Tappi 54(10):1647±1654. 123. Helle, T., and Johnsen, P. O. (1994). Using stereoscopic and SEM BSI for studying ink distribution details on paper and ®ber surfaces. J. Pulp Paper Sci. 20(7):J189±J192. 124. Heyn, A. N. J. (1966). The microcrystalline structure of cellulose in cell walls of cotton, ramie and jute ®bers as revealed by negative staining of sections. J. Cell Biol. 29:181± 197. 125. Heyn, A. N. J. (1970). The elementary ®bril and supermolecular structure of cellulose in softwood ®ber. In: Physics and Chemistry of Wood Pulp Fibers. D. H. Page, ed. STAP Ser. No. 8, TAPPI, Atlanta, GA, pp. 27±43. 126. Hieta, K., Kuga, S., and Usuda, M. (1984). Electron staining of reducing ends evidences a parallel-chain structure in Valonia cellulose. Biopolymers 23:1807±1810. 127. Hoh, J. H., and Schoenenberger, C.-A. (1994). Surface morphology and mechanical properties of MDCK monolayers by atomic force microscopy. J. Cell Sci. 107:1105± 1114. 128. Hooke, R. (1665). Micrographia, Dover, New York. 129. Horn, R. A., and Simmonds, F. A. (1968). Microscopical and other ®ber characteristics of high-yield sodium bisul®te pulps from balsam ®r. Tappi 51(1):67A±73. 130. Howard, R., and Shef®eld, E. (1987). The wet structure of pulp and paper examined by cryo-SEM. Paper Technol. Ind. 27(2):425±427. 131. Howarth, P., Smith, F. H., and Fenton, F. (1975). The use of the interference microscope in a study of adhesive distribution in pigment coatings on paper. J. Phys., Ser. E (Sci. Instrum.) 8(12):1055±1058. 132. Howland, R., and Benatar, L. (1996). A practical guide to scanning probe microscopy. Parks Scienti®c Instruments, Sunnyvale, CA. 133. Hutter, J. L., and Bechhoefer, J. (1993). Calibration of atomic-force microscope tips. Rev. Sci. Instrum. 64(7):1868±1873. 134. Ilvessalo-PfaÈf¯i, M.-S. (1995). Fiber Atlas Identi®cation of Papermaking Fibers. Springer, New York. 135. Ilvessalo-PfaÈf¯i, M.-S., and Laamanen, J. (1970). Application of the scanning electron microscope in pulp and paper research. Paperi Puu (Paper Timber) 11:163±174. 136. Isenberg, I. H. (1967). Pulp and Paper Microscopy. The Institute of Paper Chemistry, Appleton, WI. 137. Iwasaki, T., Lindberg, B., and Meier, H. (1962). The effect of ultrasonic treatment on individual wood ®bers. Svensk Papperstidn. 65(20):795±816.

Microscopy 138. 139. 140.

141. 142. 143. 144. 145. 146. 147. 148. 149.

150. 151. 152. 153. 154. 155. 156. 157. 158.

255

Jackson, L. S., Heitmann, J. A., and Joyce, T. W. (1993). Enzymatic modi®cations of secondary ®ber. Tappi 76(3):147±154. Jackson, L. S., Heitmann, J. A., and Joyce, T. W. (1994). Visualization of enzymes binding to secondary ®ber silver-enhanced colloidal gold. Prog. Paper Recycling 3(2):32±41. James, C. J., Rees, A. J., and Plummer, M. K. (1996). Applications of the scanning electron microscope to the study of Australian newsprint mills papers: Product evolution/development. Proc. 50th Appita Annual General Conf., Auckland, Vol. 1, pp. 341± 345. Jang, H. F. (1998). Measurement of ®bril angle in wood ®bers with polarization confocal microscopy. J. Pulp Paper Sci. 24(7):224±230. Jang, H. F., Amiri, R., Seth, R. S., and Karnis, A. (1996). Fiber characterization using confocal microscopy. Collapse behavior of mechanical pulp ®bers. Tappi 79(4):203±210. Jang, H. F., Howard, R. C., and Seth, R. S. (1995). Fiber characterization using confocal microscopy. The effects of recycling. Tappi J. 78(12):131±137. Jang, H. F., Robertson, A. G., and Seth, R. S. (1992). Measuring ®ber coarseness and wall thickness distributions with confocal microscopy. 78th Annual Meeting, Technical Section, CPPA, Montreal, pp. B189±192. Jang, H. F., Robertson, A. G., and Seth, R. S. (1992). Transverse dimensions of wood pulp ®bers by confocal laser scanning microscopy and image analysis. J. Mater. Sci. 27:6391±6406. Jarvis, S. P., Oral, A., Weihs, T. P., and Pethica, J. B. (1993). A novel force microscope and point contact probe. Rev. Sci. Instrum. 64(12):3515±3520. Jayme, G. (1963). Production and characteristics of spruce sulfate pulps with biological properties. Tappi 46(7):415±420. Jayme, G., and Hunger, G. (1956). Horni®cation and cellulose ®ber structure seen in the electron microscope. Monatsh. Chem. 87(1):8±23 (in German). Jayme, G., and Hunger, G. (1958). The rearrangement of micro®brils in dried cellulose and the implication of this structure alteration on pulp properties. In: Fundamentals of Papermaking Fibers. F. Bolam, ed. British Paper and Board Makers' Assoc., Kenley, England, pp. 263±271. Jayme, G., and Hunger, G. (1962). Electron microscope 2- and 3-dimensional classi®cation of ®ber bonding. In: Formation and Structure of Paper. Vol. 1. F. Bolam, ed. British Paper and Board Makers' Assn., London, pp. 135±170. Jenkins, L. M., and Donald, A. M. (1997). Use of the environmental scanning microscope for the observation of the swelling behavior of cellulosic ®bers. Scanning 19:92± 97. Jensen, W. A. (1962). Botanical Histochemistry. Freeman, San Francisco. Jones, D., McHardy, W. J., and Tait, J. M. (1984). LTSEM of biological specimens. Trans. Br. Mycol. Soc. 82(1):164±170. Jordan, B. D., and O'Neill, M. A. (1994). The birefringence of softwood mechanical pulp ®nes. J. Pulp Paper Sci. 20(6):J172±J174. Jordan, B. D., and Page, D. H. (1981). Application of image analysis to pulp ®ber characterization. In: The Role of Fundamental Research in Paper Making, Vol. 2. J. Brander, ed. Mech. Eng. Pub., London, pp. 745±776. Kallmes, O. (1960). Distribution of the constituents across the wall of unbleached spruce sul®te ®bers. Tappi 43(2):143±153. Kallmes, O., and Eckbert, C. (1964). The structure of paper. VIII. The application of the relative bonded area concept to paper evaluation. Tappi 47(9):540±548. Karlsson, J. O., Andersson, N., Berntsson, P., Chihani, T., and Gatenholm, P. (1998). Swelling behavior of stimuli-responsive cellulose ®bers. Polymer 39(16):3589±3595.

256

Moss and Groom

159. Kasten, F. H. (1990). The origins of modern ¯uorescence microscopy and ¯uorescent probes. In: Cell Structure and Function by Microspectro¯uorometry. E. Kohen and J. G. Hirschberg, eds. Academic Press, New York, pp. 3±50. 160. Kellenberger, E., Carlemalm, E., and Villiger, W. (1985). Physics of the preparation and observation of specimens that involve cyroprocedures. In: The Science of Biological Specimen Preparation for Microscopy and Microanalysis. J.-P. Revel, T. Barnard, and G. H. Haggis, eds. SEM Inc., AMF O'Hare (Chicago), pp. 1±20. 161. Keller, H. E. (1995). Objective lenses for confocal microscopy. In: Handbook of Biological Confocal Microscopy. J. B. Pawley, ed. Plenum Press, New York, pp. 111± 125. 162. Keller, D. J., and Chou, C. C. (1992). Imaging steep, high structures by scanning force microscopy with electron-beam deposited tips. Surf. Sci. 268(1±3):333±339. 163. Kerle, T., Cohen, S. R., and Klein, J. (1997). Dihedral angle at solid/liquid-polymer interfaces determined by atomic force microscopy. Langmuir 13:6360±6362. 164. Kibblewhite, R. P. (1973). Structural modi®cations to pulp ®bers: De®nitions and roles in papermaking. Tappi 60(10):141±143. 165. Kibblewhite, R. P. (1989). Effects of pulp drying and re®ning on softwood ®ber wall structural organizations. In: Fundamentals of Papermaking, Vol. 1. C. F. Baker and V. W. Punton, eds. Mech. Eng. Pub., London, pp. 121±152. 166. Kibblewhite, R. P., and Bailey, D. G. (1988). Measurement of ®ber cross-section dimensions using image processing. Appita 41(4):297±303. 167. Kibblewhite, R. P., Bawden, A. D., and Hughes, C. (1991). Hardwood market kraft ®ber and pulp qualities. Appita 44(5):325±332. 168. Kibblewhite, R. P., and Brookes, D. (1977). Dimensions and collapse behavior of kraft and bisul®te ®bers in wet pulps and `in situ' in handsheets. Appita 31(2):111±118. 169. Kibblewhite, R. P., and Shelbourne, C. J. A. (1997). Genetic selection of trees with designer ®bers for different paper and pulp grades. In: The Fundamentals of Papermaking Materials, Vol. 1. C. F. Baker, ed. Pira International, Leatherhead, U.K., pp. 641±662. 170. Kirby, A. R., Gunning, A. P., and Morris, V. J. (1995). Atomic force microscopy in food research: A new technique comes of age. Trends Food Sci. Technol. 6: 359±365. 171. Kistler, J., and Kellenberger, E. (1977). Collapse phenomena in freeze-drying. J. Ultrastruct. Res. 59:70±75. 172. Klofta, J. L., and Miller, M. L. (1994). Effects of deinking on the recycle potential of papermaking ®bers. Pulp Paper Mag. Can. 95(8):41±44. 173. Klungness, J., Caul®eld, D., Sachs, I., Sykes, M., and Shilts, R. (1994). Fiber-loading: A progress report. Tappi Proc. 1994 Recycling Symp., Boston, MA, pp. 283±290. 174. Knebel, W. (1991). Dual-channel applications with the confocal laser scanning microscope CLSM of Leica Lasertechnik GmbH. Sci. Tech. Info. 10(2):46±52. 175. Knebel, W., and Schnepf, E. (1991). Confocal laser scanning microscopy of ¯uorescently stained wood cells: A new method for three-dimensional imaging of xylem elements. Trees 5(1):1±4. 176. Koubaa, A., and Koran, Z. (1995). Measure of the internal bond strength of paper/ board. Tappi J. 78(3):103±111. 177. KucÏera, L., and Bariska, M. (1982). On the fracture morphology in wood. Part 1: A SEM-study of deformation in wood of spruce and aspen under ultimate axial compression load. Wood Sci. Technol. 16:241±259. 178. Kunnas, L., Lehtinen, J., Paulapuro, H., and Kiviranta, A. (1993). The effect of Condebelt drying on the structure of ®ber bonds. Tappi 76(4):95±104. 179. Laamanen, J., Ora, M., and Eklund, H. (1986). In¯uence of post re®ning on the properties of TMP ®bers. International Paper Physics Conf., Adirondack, NY.

Microscopy 180.

181. 182. 183. 184. 185. 186. 187. 188. 189. 190. 191. 192. 193. 194. 195. 196. 197. 198. 199.

257

Lackner, R., Srebotnik, E., and Messner, K. (1992). Localization of extracellular ligninases of Phanerochaete chrysosporium. In: Ligno-Cellulosics: Science, Technology, Development and Use. J. F. Kennedy, G. O. Phillips, and P. A. Williams, eds. Ellis Horwood, London, pp. 41±44. Lammi, T., and Heikkurinen, A. (1997). Changes in ®ber wall structure during de®bration. In: The Fundamentals of Papermaking Materials, Vol. 1, C. F. Baker, ed. Pira International, Leatherhead, U.K., pp. 641±662. Lange, J., MaÊnson, J.-A. E., and Hult, A. (1994). Defects in solvent-free organic coatings studied by atomic force microscopy, scanning acoustic microscopy, and confocal laser microscopy. J. Coatings Technol. 66(838):19±26. Lange, P. W. (1958). The distribution of the chemical constituents throughout the cell wall. In: Fundamentals of Papermaking Fibers. F. Bolam, ed. British Paper and Board Makers' Assoc., Kenley, England, pp. 7±34. Leney, L. (1981). A technique for measuring ®bril angle using polarized light. Wood Fiber 13(1):13±16. Leopold, B., and McIntosh, D. C. (1961). Chemical composition and physical properties of wood ®bers. III. Tensile strength of individual ®bers from alkali extracted loblolly pine holocellulose. Tappi 44(3):235±240. Lepoutre, P., and de Silveira, G. (1991). Examination of cross-sections of blade- and roll-coated LWC paper. J. Pulp Paper Sci. 17(5):J184±J186. Lichtman, J. W. (1994). Confocal microscopy. Sci. Am. 271(2):30±35. Lindsay, J. (1993). Relative ¯ow porosity in paper. In: Products of Papermaking, Vol. 2. C. F. Baker, ed. Pira International, Leatherhead, U.K., pp. 935±968. Lùvhaugen, O. (1995). Possible effects of ¯uorescence bleaching/saturation on the de®nition of the paper surface in CLSM measurements. Zool. Stud. 34(Suppl. 1):230±231. Lùvhaugen, O. (1995). Average pore volume measurement using confocal laser scanning microscopy. Image Analysis for Pulp and Paper Research and Production. 3rd European Res. Symp., STFI, Stockholm. Lùvhaugen, O., Berli, E. L., and Bjerke, F. (1996). Pore volumes and their relations to standard surface parameters and print quality. Proc. Int. Printing and Graphic Arts Conf., pp. 111±114. Luong, C. H., and Lindem, P. E. (1995). A method to determine heat conduction in paper during calendering. Int. Paper Physics Conf., Niagara-on-the-Lake, ON, Canada, pp. P1±P3. Luukko, K. (1994). Characterization of chemical pulp ®nes. M.Sc. Thesis, Laboratory of Paper Technology, Helsinki University of Technology, Espoo, Finland (in Finnish). Luukko, K., Kemppainen-Kajola, P., and Paulapuro, H. (1997). Characterization of mechanical pulp ®nes by image analysis. Appita 50(5):387±392. MacGregor, M. A. (1992). The technical challenge of making better paper: Preparing for the 21st century. Proc. TAPPI Engineering Conf., Boston, MA, pp. 39±54. MacGregor, M. A. (1996). Measuring water-induced roughening of LWC paper and its effect on printed gloss. Tappi Proc. 1996 Process and Product Quality Conf., Cincinnati, OH, pp. 129±137. MacGregor, M. A., Johansson, P. R., and BeÂland, M.-C. (1994). Measurement of small-scale gloss variation in printed paper. Int. Printing and Graphic Arts Conf., Atlanta, GA, pp.33±43. Maganov, S. N. (1993). Surface characterization of materials at ambient conditions by scanning tunneling microscopy (STM) and atomic force microscopy (AFM). Appl. Spetrosc. Rev. 28(1&2):1±121. Mandelbrot, B. B., Passoja, D. E., and Pullay, A. J. (1984). Fractal character of fracture surfaces of metals. Nature 308(5961):721±722.

258

Moss and Groom

200. Mangin, P. J., and BeÂland, M.-C. (1993). Three-dimensional evaluation of paper surfaces using confocal microscopy: Applications to research and development. IARIGAI 22nd Int. Res. Conf., Munich, Germany. 201. Mangin, P. J., BeÂland, M.-C., and Cormier, L M. (1994). Paper surface compressibility and printing. Proc. 1994 Int. Printing and Graphic Arts Conf., Halifax, NS, pp. 19±31. 202. Mardon, J., Monahan, R. E., Carter, R. A., and Wilder, J. E. (1966). Dynamic consolidation of paper during calendering. In: Consolidation of the Paper Web, Vol. 1. F. Bolam, ed. British Paper and Board Makers' Assoc., London, pp. 576±634. 203. Martinez-Nistal, A., Alonso, M., GonzaÂles-Rio, F., Sampedro, A., and Astorga, R. (1998). 3D reconstruction of wood ®bers using confocal microscopy. Microsc. Anal. 63:19±20. 204. Marton, R., and Crosby, C. M. (1969). Distribution of phenolic resins in laminating papers. Tappi 52(4):681±688. 205. Maxwell, M. H. (1978). Two rapid and simple methods used for the removal of resins from 1:0 mm thick epoxy sections. J. Microsc. 112(2):253±255. 206. McCool, M. A., and Taylor, C. J. (1983). Image analysis techniques in recycled ®ber. Tappi 66(8):69±71. 207. McIntosh, D. C. (1967). The effect of re®ning on the structure of the ®ber wall. Tappi 50(10):482±488. 208. McIntosh, D. C., and Leopold, B. (1962). Bonding strength of individual ®bers. In: Formation and Structure of Paper, Vol. 1, F. Bolam, ed., British Paper and Board Makers' Assoc., London, pp. 171±193. 209. McKenzie, A. W., and Higgins, H. G. (1964). Conformability and surface texture of cellulose ®bers in relation to paper strength. Paper Technol. 5(2):155±158. 210. Miller, K. R., Prescott, C. S., Jacobs, T. L., and Lassignal, N. L. (1983). Artifacts associated with quick-freezing and freeze-drying. J. Ultrastruct. Res. 82:123±133. 211. Minsky, M. (1988). Memoir on inventing the confocal scanning microscope. Scanning 10(4):128±138. 212. Mislankar, A., Darabie, A., and Reeve, D. W. (1997). Deteciton of ortho-quinones in wood using confocal scanning laser microscopy. J. Pulp Paper Sci. 23(2):J73±J76. 213. Mohlin, U-B. (1975). Cellulose ®ber bonding: Part 5. Conformability of pulp ®bers. Svensk Papperstidn. 78(11):412±416. 214. Mora, F., Ruel, K., Comtat, J., and Joseleau, J.-P. (1986). Aspect of native and redeposited xylans at the surface of cellulose micro®brils. Holzforschung 40:85±91. 215. Morris, V. J. (1994). Biological applications of scanning probe microscopes. Prog. Biophys. Mol. Biol. 61:131±185. 216. Moss, P. A. (1990). A study of the frozen hydrated structure of pulp. Ph.D. Thesis, Department of Paper Science, University of Manchester Institute of Science and Technology, Manchester, U.K. 217. Moss, P. A., Kropholler, H. W., and Shef®eld, E. (1989). LTSEM: Great potential for pulp evaluation. Paper Technol. 30(9):12±14. 218. Moss, P. A., Nyblom, I., Sneck, A., and HyvaÈrinen, H.-K. (1999). The location and quanti®cation of lignin in kraft pulps using a confocal laser scanning microscope (CLSM) and image analysis. In: Microscopy as a Tool in Pulp and Paper Research and Development. STFI Proceedings, Stockholm, pp. 221±227. 219. Moss, P. A., and Retulainen, E. (1997). The effect of ®nes on ®ber bonding; crosssectional dimensions of TMP ®bers at potential bonding sites. J. Pulp Paper Sci. 23(8):J382±J388. 220. Moss, P. A., Retulainen, E., Paulapuro, H., and Aaltonen, P. (1993). Taking a new look at pulp and paper. Applications of confocal laser scanning microscopy (CLSM) to pulp and paper research. Paperi Puu (Paper Timber) 75(1/2):74±79.

Microscopy 221. 222. 223. 224. 225. 226. 227. 228. 229. 230. 231. 232. 233. 234. 235. 236. 237. 238. 239. 240. 241. 242. 243.

259

Mott, L., Shaler, S. M., Liang, B.-H., and Groom, L. H. (1995). The tensile testing of individual wood ®bers using environmental scanning electron microscopy and video image analysis. Tappi J. 78(5):43±148. Naito, T., Nishi, K., and Kawano, Y. (1995). Delamination resistance of paper. International Paper Physics Conf., Niagara-on-the-Lake, ON, Canada, pp. 125±130. Nanko, H., Ansano, S., and Ohsawa, J. (1991). Shrinking behavior of pulp ®bers during drying. Proc. Int. Paper Physics Conf., Kona, HI, Book 2, pp. 365±373. Nanko, H., and Ohsawa, J. (1989). Mechanisms of bond formation. In: Fundamentals of Papermaking, Vol. 1. C. F. Baker and V. W. Punton, eds. Mech. Eng. Pub., London, pp. 783±830. Nanko, H., Ohsawa, J., and Okagawa, A. (1989). How to see inter®ber bonding in paper sheets. J. Pulp Paper Sci. 15(1):J17±J22. Nanko, H., and Tada, Y. (1995). Mechanisms of hygroexpansion of paper. Int. Paper Physics Conf., Niagara-on-the-Lake, ON, Canada, pp.159±171. Nanko, H., and Wu, J. (1995). Mechanisms of paper shrinkage during drying. Int. Paper Physics. Conf., Niagara-on-the-Lake, ON, Canada, pp. 103±113. Nissan, A. H. (1951). Particle size analysis from sedimentation curves. Disc. Faraday Soc. 11:15±27. Nissan, A. H., and Sternstein, S. S. (1964). Cellulose-®ber bonding. Tappi 47(1):1±6. Nordman, L. S., and QvikstroÈm, B. (1970). Variability in the mechanical properties of ®bers within a growth period. In: Physics and Chemistry of Wood Pulp Fibers. D. H. Page, ed. STAP Ser. No. 8. TAPPI, Atlanta, GA, pp. 177±200. O'Brien, T. P., and McCully, M. E. (1981). The Study of Plant Structure: Principles and Selected Methods. Termacarphi Pty. Ltd., Melbourne. Page, D. H. (1960). Fiber-to-®ber bonds. Part 1. A method for their direct observation. Paper Technol. Ind. 1(5):519±530. Page, D. H. (1967). The collapse behavior of pulp ®bers. Tappi 50(9):449±455. Page, D. H. (1969). A method of determining the ®brillar angle of wood ®bers using polarized light. J. Microsc. 90(2):137±143. Page, D. H. (1969). A theory for the tensile strength of paper. Tappi 52(4):674±681. Page, D. H. (1989). The beating of chemical pulps: The action and the effects. In: Fundamentals of Papermaking. Vol. 1. C. F. Baker and V. W. Punton, eds. Mech. Eng. Pub., London, pp. 1±38. Page, D. H., and De GraÃce, J. H. (1967). The delamination of ®ber walls by beating and re®ning. Tappi 50(10):489±495. Page, D. H., El-Hosseiny, F. E., Winkler, K., and Lancaster, A. P. S. (1977). Elastic modulus of single wood pulp ®bers. Tappi 60(4):114±117. Page, D. H., and Emerton, H. W. (1959). The microscopic examination of ®bers, paper, board and wood. Part 1. Improved methods for the study of surfaces. Svensk Papperstidn. 62(9):318±332. Page, D. H., and Sargent, J. W. (1962). The ®ne structure of ®ber bonding. In: Formation and Structure of Paper, Vol. 1. F. Bolam, ed. British Paper and Board Makers' Assoc., London, pp. 193±200. Page, D. H., Sargent, J. W., and Nelson, R. (1966). Structure of paper in cross-section. In: Consolidation of the Paper Web, Vol. 1. F. Bolam, ed. British Paper and Board Makers' Assoc., London, pp. 313±352. Page, D. H., and Tydeman, P. A. (1960). Fiber-to-®ber bonds. Part 2. A preliminary study of their properties in paper sheets. Paper Technol. Ind. 1(5):519±530. Page, D. H., and Tydeman, P. A. (1966). Physical processes occurring during the drying phase. In: Consolidation of the Paper Web, Vol. 1. F. Bolam, ed. British Paper and Board Makers' Assoc., London, pp. 371±396.

260

Moss and Groom

244. Page, D. H., Tydeman, P. A., and Hunt, M. (1962). A study of ®ber-to-®ber bonding by direct observation. In: Formation and Structure of Paper, Vol. 1. F. Bolam, ed. British Paper and Board Makers' Assoc., London, pp. 171±193. 245. Page, D. H., Tydeman, P. A., and Hunt, M. (1962). The behavior of ®ber-to-®ber bonds in sheets under dynamic conditions. In: Formation and Structure of Paper, Vol. 1. F. Bolam, ed. British Paper and Board Makers' Assoc., London, pp. 249±263. 246. Panshin, A. J., and de Zeeuw, C. (1960). Textbook of Wood Technology. McGraw-Hill, New York. 247. Parham, R. A. (1973). X-ray/SEM analysis of materials: Principles and applications to paper. Paperi Puu (Paper Timber) 55(12):959±972. 248. Parham, R. A. (1974). Electron microscopy of pulp and paper. Wood Sci. 6(3):245±255. 249. Parham, R. A. (1975). Critical-point drying for ®ber microscopy. Tappi 58(3):138±140. 250. Parham, R. A., and Gray, R. L. (1982). The Practical Identi®cation of Wood Pulp Fibers. TAPPI, Atlanta, GA. 251. Parsons, E., Dole, B., Hall, D. J., and Thomas, W. D. E. (1974). A comparative survey of the techniques for preparing plant surfaces for the scanning electron microscope. J. Microsc. 101(1):59±75. 252. Patnaik, S. S., Bunning, T. J., Adams, W. W., Wang, J., and Labes, M. M. (1995). Atomic force microscopy and high-resolution scanning microscopy study of the banded surface morphology of hydroxypropyl cellulose thin ®lms. Macromolecules 28(1):393± 395. 253. Pawley, J. B., ed. (1995). Handbook of Biological Confocal Microscopy. Plenum Press, New York. 254. Pelton, R., Jordan, B., and Allen, L. (1985). Particle size distribution of ®nes in mechanical pulps and some aspects of their retention in papermaking. Tappi J. 68(2):91±94. 255. Pereira, D. E. D., and da Silva, E. C., Jr. (1995). Improvement of atomic force microscopy (AFM) as an analytical instrument for residual lignin characerization. Proc. 8th Int. Symp. on Wood and Pulping Chemistry, Helsinki, Finland, Vol.1, pp. 467±474. 256. Pesacreta, T. C., Carlson, L. C., and Triplett, B. (1997). Atomic force microscopy of cotton ®ber cell wall surfaces in air and water: Quantitative and qualitative aspects. Planta 202:435±442. 257. Peterson, R. A., and Williams, C. L. (1992). Determining paper-coating thickness with electron microscopy and image analysis. Tappi J. 75(10):122±126. 258. Pye, I. T., Washburn, O. V., and Buchanan, J. G. (1966). Structural changes in paper on pressing and drying. In: Consolidation of the Paper Web, Vol. 1. F. Bolam, ed. British Paper and Board Makers' Assoc., London, pp. 353±367. 259. Quackenbush, D. W. (1958). A photomicrographic technique for observing the contours of textured surfaces. Tappi 41(12):772±775. 260. Quackenbush, D. W. (1971). Faults in paper coatings and their relationship to base sheet structure. Tappi 54(1):47±52. 261. Quackenbush, D. W. (1988). Case studies in the use of the optical microscope in the analysis of coated paper. Tappi J. 71(5):70±75. 262. Radmacher, M., Tillmann, R. W., Fritz, M., and Gaub, H. E. (1992). From molecules to cells: Imaging soft samples with the atomic force microscope. Science 257:1900±1905. 263. RaÊnby, B. G. (1952). The ®ne structure of cellulose ®brils. In: Fundamentals of Papermaking Fibers. F. Bolam, ed. British Paper and Board Makers' Assoc., Kenley, England, pp. 7±34. 264. Read, N. D., and Beckett, A. (1983). Effects of hydration on the surface morphology of urediospores. J. Microsc. 132(2):179±184. 265. Retulainen, E., and Ebeling, K. (1985). Effect of paper on the load-elongation behavior of ®ber-to-®ber bonds. In: Papermaking Raw Materials, Vol. 1. V. Punton, ed. Mech. Eng. Pub., London, pp. 229±263.

Microscopy 266. 267. 268. 269. 270. 271. 272. 273. 274. 275. 276. 277. 278.

279. 280. 281. 282. 283. 284. 285. 286.

261

Retulainen, E., Moss, P., and Nieminen, K. (1993). Effects of ®nes on the properties of ®ber networks. In: Products of Papermaking, Vol. 2, C. F. Baker, ed. Pira Int., Leatherhead, U.K., pp. 727±769. Retulainen, E., Moss, P., and Nieminen, K. (1997). Effect of calendering and wetting on paper properties. J. Pulp Paper Sci. 23(1):J34±J39. Revol, J. F. (1982). On the cross-sectional shape of cellulose crystallites in Valonia ventricosa. Carbohydr. Polym. 2:123±134. Revol, J. F. (1985). Change of the D-spacing in cellulose crystals during lattice imaging. J. Mater. Sci. Lett. 4(11):1347±1349. Rezanowich, A. (1968). Some applications of the scanning electron microscope at the Pulp and Paper Research Institute of Canada. In: Scanning Electron Microscopy. O. Johari and I. Corvin, eds. IIT Res. Inst., Chicago, pp. 13±27. Rials, T. G., and Glasser, W. G. (1989). Multiphase materials with lignin. VI. Effect of cellulose derivative structure on blend morphology with lignin. Wood Fiber Sci. 21(1):80±90. Robards, A. W., and Sleytr, U. B. (1985). Low Temperature Methods in Biological Electron Microscopy. Elsevier, New York. Robertson, A. A., and Mason, S. G. (1962). The role of ®ber collapse in papermaking. In: Formation and Structure of Paper, Vol. 2. F. Bolam, ed. British Paper and Board Makers' Assoc., London, pp. 639±647. Robertson, L. R. (1993). The use of phase-contrast microscopy to assess and differentiate the microbial population of a paper mill. Tappi J. 76(3):83±87. Robertson, L. R., and Taylor, N. R. (1993). Bio®lms and dispersants: A less toxic approach to deposit control. Tappi Proc. 1993 Papermakers Conf., Atlanta, GA, pp. 631±638. Rollins, M. L., Cannizzaro, A. M., and Goynes, W. R. (1972). Electron microscopy of cellulose and cellulose derivatives. In: Instrumental Analysis of Cotton Cellulose and Modi®ed Cotton Cellulose. R. T. O'Connor, ed. Marcel Dekker, New York, pp. 171±271. Rollins, M. L., and deGruy, I. V. (1972). Light microscopy in the study of cellulose. In: Instrumental Analysis of Cotton Cellulose and Modi®ed Cotton Cellulose. R. T. O'Connor, ed, Marcel Dekker, New York, pp. 93±169. Rosowski, J. R., and Gilder, W. V.(1977). Comparative effects of metal coating by sputtering and by vacuum evaporation on delicate features of Euglenoid ¯agellates. In: Scanning Electron Microscopy/I. O. Johari and I. Corvin, eds. IIT Res.Inst., Chicago, pp. 471±480. Ruel, K., and Joseleau, J.-P. (1984). Use of enzyme-gold complexes for the ultrastructural localization of hemicelluloses in the plant cell wall. Histochemistry 81:573±580. Ruska, E. (1933). Formation of pictures of surfaces irradiated by electrons in the electron microscope. Z. Physik 83:492±497 (in German). Sachs, I. B. (1985). Preserving and recovering pulp ®brils subsequent to drying. Paper Tech. Ind. 20(1):38±41. Sachs, I. B. (1986). Retaining raised ®brils and micro®brils on ®ber surfaces. Tappi J. 69(11):124±127. Saka, S., and Thomas, R. J. (1982). Evaluation of the quantitative assay of lignin distribution by TEM-EDXA technique. Wood Sci. Technol. 16:1±18. Saka, S., Thomas, R. J., and Gratzl, J. S. (1978). Lignin distribution determination by energy dispersive analysis of X rays. Tappi 61(1):73±76. Saka, S., Whiting, P., Fukazawa, K., and Goring, D. A. I. (1982). Comparative studies on lignin distribution by UV microscopy and bromination combined with EDXA. Wood Sci. Technol. 16:269±277. Scallan, A. M. (1974). The structure of the cell wall of wood: A consequence of anisotropic inter-micro®brillar bonding? Wood Sci. 6(3):266±271.

262

Moss and Groom

287. Scallan, A. M. (1978). The accommodation of water within pulp ®bers. In: Fiber±Water Interactions in Paper-Making, Vol. 1. Fundamental Research Committee, ed. British Paper and Board Industry Fed., London, pp. 9±29. 288. Scott, J. A. N., Procter, A. R., Fergus, B. J., and Goring, D. A. I. (1969). The application of ultraviolet microscopy to the distribution of lignin in wood. Description and validity of the technique. Wood Sci. Technol. 3(1):73±92. 289. Sears, G. R., and Kregel, E. A. (1942). Application of the electron microscope to problems of the pulp, paper and paperboard industry. Paper Trade J. 114(19th March):TS139±TS145. 290. Seth, R. S., Jang, H. F., Chan, B. K., and Wu, C. B. (1997). Transverse dimensions of wood pulp ®bres and their implications for end use. In: The Fundamentals of Papermaking Materials, Vol. 1. C. F. Baker, ed. Pira Int., Leatherhead, U.K., pp. 473±503. 291. Settlemeyer, L. A. (1992). Topography and the corresponding effect on gloss as detailed with the scanning electron microscope. Proc. TAPPI Int. Printing and Graphic Arts Conf., Pittsburgh, PA, pp. 41±47. 292. Sheehan, J. G., and Whalen-Shaw, M. (1990). High-magni®cation cryogenic scanning electron microscopy (cryo-SEMs) of wet coating microstructures. Tappi J. 73(5):171±178. 293. Shijun, K. (1983). Ultrastructure of bagasse and distribution of hemicellulose across its cell wall. Cellulose Chem. Technol. 17:507±514. 294. Shotton, D. M. (1989). Confocal scanning optical microscopy and its applications for biological specimens. J. Cell Sci. 94:175±206. 295. Siegel, B. M. (1949). The structure of cellulose in the electron microscope range. Tappi 32(3):109±116. 296. Simons, F. L. (1950). A stain for use in the microscopy of beaten ®bers. Tappi 33(7):312±314. 297. Skowronski, J. (1990). Surface roughening of pre-calendered base sheets during coating. J. Pulp Paper Sci. 16(3):J102±J110. 298. Skowronski, J. (1991). Fiber-to-®ber bonds in paper. Part II: Measurement of the breaking energy of ®ber-to-®ber bonds. J. Pulp Paper Sci. 17(6):J217±J222. 299. Skowronski, J., and Bichard, W. (1987). Fiber-to-®ber bonds in paper. Part I. Measurement of bond strength and speci®c bond strength. J. Pulp Paper Sci. 13(5):J165±J169. 300. Sloane, C. M. (1993). A preliminary study of the effect of recycling on the bonding potential of unbleached radiata pine and Scandinavian spruce kraft pines. Proc. 47th Appita Annual General Conf., Rotorua, New Zealand, pp. 533±540. 301. Smith, K. C. A. (1959). Scanning electron microscopy in pulp and paper research. Pulp Paper Mag. Can. 60(12):T366±T371. 302. SoÈremark, C., Johansson, G., and Kiviranta, A. (1994). Characterization and elimination of ®ber orientation streaks. TAPPI Engineering Conf., San Francisco, Book 1, pp. 97±104. 303. Srebotnik, E., and Messner, K. (1991). Immunoelectron microscopical study of the porosity of brown-rot degraded pine wood. Holzforschung 45(2):95±101. 304. Srebotnik, E., and Messner, K. (1992). The use of immunoelectron microscopy to study the diffusion of proteins in wood and pulp. In: Ligno-Cellulosics: Science, Technology, Development and Use. J. F. Kennedy, G. O. Phillips, and P. A. Williams, eds. Ellis Horwood, London, pp. 775±778. 305. Steadman, R. K. (1982). The apparent density of paper. Ph.D. Thesis, Department of Paper Science, University of Manchester Institute of Science and Technology, Manchester, U.K. 306. Steadman, R., and Luner, P. (1985). The effect of wet ®ber ¯exibility on sheet apparent density. In: Papermaking Raw Materials, Vol. 1. V. Punton, ed. Mech. Eng. Pub., London, pp. 311±337.

Microscopy 307. 308. 309. 310. 311. 312. 313. 314. 315. 316. 317. 318. 319. 320. 321. 322. 323. 324. 325. 326. 327. 328. 329. 330. 331. 332.

263

Stelzer, E. H. K., and Bacallao, R. (1988). Confocal ¯uorescence microscopy of epithelial cells. SPIE 1028:167±168. Stone, J. E., and Scallan, A. M. (1965). A study of the cell wall structure by N2 adsorption. Pulp Paper Mag. Can. 66(8):T407±T414. Stone, J. E., and Scallan, A. M. (1967). The effect of component removal upon the porous structure of the cell wall of wood. II. Swelling in water and the ®ber saturation point. Tappi 50(10):496±501. Strachan, J. (1932). Further notes on the hydration of cellulose in papermaking. Proc. Tech. Sect. PMA. GB and Ireland 13:61±81. Stratton, R. A., and Colson, N. L. (1993). Fiber wall damage during bond failure. Nordic Pulp Paper Res. J. 2:245±249. Sugiyama, J., Harada, H., Fujiyoshi, Y., and Uyeda, N. (1984). High resolution observations of cellulose micro®brils. Mokuzai Gakkaishi 30(1):98±99. Sugiyama, J., Vuong, R., and Chanzy, H. (1991). Electron diffraction study on the two crystalline phases occurring in native cellulose from an algal cell wall. Macromolecules 24(14):4168±4175. Suleman, A. U. M. (1996). Atomic force microscope studies of cellulosic ®bers. M.Sc. Thesis, Dept. of Paper Science, Univ. Manchester Institute of Science and Technology, Manchester, U.K. Sundin, J., SundstroÈm, L., and Hartler, N. (1996). Determination of collapse and coarseness of ®bers in paper sheets from cross-sectional dimensions using confocal microscopy and image analysis. Nordic Pulp Paper Res. J. 1:56±57. Suominen, I., Suihko, M.-L., and Salkinoja-Salonen, M. (1997). Microscopic study of migration of microbes in food packaging paper and board. J. Ind. Microbiol. Biotechnol. 19:104±113. Szikla, Z., and Paulapuro, H. (1988). Changes in z-direction density distribution of paper in wet pressing. J. Pulp Paper Sci. 15(1):J11±J17. Tam Doo, P. A., and Kerekes,R. J. (1981). A method to measure wet ®ber ¯exibility. Tappi 64(3):113±116. TAPPI T 205 (1995). Forming handsheets for physical tests of pulp. TAPPI T 234 (1984). Coarseness of pulp ®bers. TAPPI T 246 (1989). Foreign particulate matter in pulp by transmitted light. TAPPI T 259 (1993). Species identi®cation of nonwood plant ®bers. TAPPI T 263 (1993). Identi®cation of wood and ®bers from conifers. TAPPI T 267 (1985). Compression wood identi®cation. TAPPI T 401 (1993). Fiber analysis of paper and paperboard. TAPPI T 421 (1991). Qualitative (including optical microscopic) analysis of mineral ®ller and mineral coating of paper. Teder, A. (1964). The properties of spruce pulps as related to paper structures observed with the scanning electron microscope. Svensk Papperstidn. 67(17):670±685. Teder, A. (1964). The effect of drying, heat-treatment and beating on the papermaking properties of birch pulps. Svensk Papperstidn. 66(22):911±923. Terasaki, M., and Dailey, M. E. (1995). Confocal microscopy of living cells. In: Handbook of Biological Confocal Microscopy. J. B. Pawley, ed. Plenum Press, New York, pp. 327±344. Thayer, W. S., and Thomas, C. E. (1971). Analysis of the glue lines in corrugated board. Tappi 54(11):1853±1858. The Second Report of the Pulp Evaluation Committee to the Technical Section of the Paper Makers' Association, 1936, London. Thode, E. F., and Ingmanson, W. L. (1959). Factors contributing to the strength of a sheet of paper. I. External speci®c surface and swollen speci®c volume. Tappi 42(1):74± 83.

264

Moss and Groom

333. Thomason, J. L., and Knoester, A. (1990). Application of confocal scanning optical microscopy to the study of ®ber-reinforced polymer composites. J. Mater. Sci. Lett. 9:258±262. 334. Thundat, T., Allison, D. P., Warmack, R. J., and Ferrel, T. L. (1992). Imaging isolated strands of DNA-molecules by atomic force microscopy. Ultramicroscopy 42(B):1101± 1106. 335. Ting, T. H. D., Chiu, W. K., and Johnston, R. E. (1997). Network changes in paper under compression in the z-direction: The effect of loading rate and ®ber wall thickness. Appita 50(3):223±229. 336. Tripp, V. W., and Giuffria, R. (1954). Concentric cellulose layers in cotton ®ber cell wall. Textile Res. J. 24:757±758. 337. Tsien, R. Y., and Bacskai, B. J. (1995). Video-rate confocal microscopy. In: Handbook of Biological Confocal Microscopy. J. B. Pawley, ed. Plenum Press, New York, pp. 459± 477. 338. Vaughan, D., ed. (1983). Energy-Dispersive Xray Microanalysis: An Introduction. Kevex Corp., Foster City, CA. 339. Verbelen, J.-P., and Stickens, D. (1995). In vivo determination of ®bril orientation in plant cell walls with polarization CLSM. J. Microsc. 177(1):1±6. 340. Visser, T. D., and Oud, J. L. (1994). Volume measurements in three-dimensional microscopy. Scanning 16(4):198±200. 341. WaÊgberg, P., and Johannson, P.-AÊ. (1993). Surface pro®lometry: A comparison between optical and mechanical sensing on printing papers. Tappi J. 76(12): 115±121. 342. Walbaum, H. H. (1972). Pigment dispersion property evaluation by scanning electron microscopy. Tappi 55(7):1108±1114. 343. Wardrop, A. B. (1962). The path of penetration of pulping media into wood. In: Formation and Structure of Paper, Vol. 2. F. Bolam, ed. British Paper and Board Makers' Assoc., London, pp. 621±637. 344. Wardrop, A. B., and Preston, R. D. (1951). The submicroscopic organization of the cell wall in conifer tracheids and wood ®bers. J. Exp. Bot. 2:20±30. 345. Washburn, O. V., and Buchanan, J. G. (1964). Changes in wet structure on pressing and drying. Pulp Paper Mag. Can. 65(9):T400±T408. 346. Watts, J., and Emerton, H. W. (1954). The interpretation of electron micrographs. Proc. Tech. Sect. BPBMA 35(2):303±318. 347. Webb, R. H. (1995). Bibliography of confocal microscopes. In: Handbook of Biological Confocal Microscopy. J. B. Pawley, ed. Plenum Press, New York, pp. 571±577. 348. Weidner, J. P. (1939). The in¯uence of humidity on changes in diameter and length of sul®te ®bers. Paper Trade J. 108:T1±T10. 349. Weise, U. W. H. (1995). Monitoring of ®ber wrinkling in drying: An approach by fractal analysis. Proc. Third Eur. Res. Symp., Image Analysis in Pulp and Paper Research and Production, Stockholm, Sweden. 350. Weisenhorn, A., Schmitt, F., Knoll, W., and Hansma, P. (1992). Streptavidin binding observed with an atomic force microscope. Ultramicroscopy 42(B):1125±1132. 351. Westermark, U., Lidbrandt, O., and Eriksson, I. (1988). Lignin distribution in spruce (Picea abies) determined by mercurization with SEM-EDXA technique. Wood Sci. Technol. 22:243±250. 352. White, N. S. (1995). Visualization systems for multidimensional CLSM images. In: Handbook of Biological Confocal Microscopy. J. B. Pawley, ed. Plenum Press, New York, pp. 211±254. 353. Wickramasinghe, H. K. (1989). Scanned-probe microscopes. Sci. Am. 261(4):74±81. 354. Wilke, V. (1985). Optical scanning microscopy: The laser scan microscope. Scanning 7(2):88±96.

Microscopy 355.

356. 357. 358. 359. 360. 361. 362. 363. 364. 365. 366. 367. 368.

369. 370. 371. 372.

265

Wilkins, M. J., Davies, M. C., Jackson, D. E., Mitchell, J. R., Roberts, C. J., Stokke, B. T., and Tendler, S. J. B. (1993). Comparison of scanning tunnelling microscopy and transmission electron microscopy image data of a microbial polysaccharide. Ultramicroscopy 48:197±201. Williams, G. J., and Drummond, J. G. (1994). Preparation of large sections for the microscopical study of paper structure. Proc. TAPPI Papermakers Conf., pp. 517±523. Williams, G. J., Drummond, J. G., and Cisneros, H. A. (1994). A microscopical approach for examining ®ber and paper structures. J. Pulp Paper Sci. 20(4):J110±J113. Williams, R. C., and Wykoff, R. W. G. (1944). The thickness of electron microscopic objects. J. Appl. Phys. 15:712±716. Williams, R. C., and Wykoff, R. W. G. (1946). Applications of metallic shadow-casting to microscopy. J. Appl. Phys. 17:23±33. Wilson, T., ed. (1990). Confocal Microscopy. Academic Press, New York. Wilson, T., and Sheppard, C. (1984). Theory and Practice of Scanning Optical Microscopy. Academic Press, New York. Wimmer, R., and Lucas, B. N. (1997). Comparing mechanical properties of secondary wall and cell corner middle lamella in spruce wood. IAWA J. 18(1):77±88. Wimmer, R., Lucas, B. N., Tsui, T. Y., and Oliver, W. C. (1997). Longitudinal hardness and Young's modulus of spruce tracheid secondary walls using nanoindentation technique. Wood Sci. Technol. 31(2):131±141. Wolken, J. J. (1961). Euglena. Rutgers, The State University, New Brunswick, NJ. Xu, L., Filonenko, Y., Li, M., and Parker, I. (1997). Measurement of wall thickness of fully collapsed ®bers by confocal microscopy and image analysis. Proc. 51st Appita Annual General Conf. Xu, L., Osborne, C., and Parker, I. (1995). Determination of ®ber 3D orientation in paper using confocal microscopy and image analysis. Dicta-95, Brisbane, pp. 515±520. Xu, L., Parker, I., and Osborne, C. (1996). Technique for determining the ®ber distribution in the z direction using confocal microscopy and image analysis. Proc. 50th Appita Annual General Conf., Auckland, Vol. 2, pp. 603±607. Yang, C.-F., Eusufzai, A. R. K., Sankar, R., Mark, R. E., and Perkins, R. W. (1978). Measurements of geometrical parameters of ®ber networks. Part 1. Bonded surfaces, aspect ratios, ®ber moments of inertia, bonding state probabilities. Svensk Papperstidn. 81(13):426±433. Ye, C., and SundstroÈm, O. (1997). Determination of S2-®bril angle and ®ber-wall thickness by microscopic transmission ellipsometry. Tappi 80(6):181±190. Ylikoski, J. (1992). Characterization of ®ber surface properties and ®brillation by confocal laser scanning microscopy. M.Sc. Thesis, Helsinki Univ. Technology, Espoo, Finland (in Finnish). Young, J. Z., and Roberts, F. (1951). A ¯ying-spot microscope. Nature 167:231. Young, R., Ward, J., and Scire, F. (1971). Observations of metal-vacuum-metal tunneling, ®eld emission, and the transition region. Phys. Rev. Lett. 27(14):922±924.

6 POROSITY AND GAS PERMEABILITY TATSUO YAMAUCHI and KOJI MURAKAMI Kyoto University Kyoto, Japan

I. Internal Pore Structure A. Description and De®nition B. Measurement of Porosity C. Measurement of Speci®c Surface Area D. Measurement of Pore Size Distribution II. Gas Permeability A. Fundamental Description B. Measurement of Gas Permeability C. Measurement of Water Vapor Permeability References

I.

267 267 269 270 272 282 282 288 294 298

INTERNAL PORE STRUCTURE

A.

Description and De®nition

Pore space forms a considerable amount of the volume contained within a piece of paper or paperboard. The pores range in size from inter®ber gaps to molecular interstices. Most of the pore spaces connect with each other to form complicated three-dimensional channels through the sheet. Because paper is anisotropic due to preferential ®ber orientation and is deformable in its bulk dimensions under an external load or upon the sorption of moisture, the structure of the pore system is also anisotropic and deformable. The pore spaces in paper can be classi®ed according to their origin: 

Professor Emeritus, Kyoto University. 267

268

Yamauchi and Murakami

. The pore spaces that originally occurred in papermaking ®bers or converting materials such as ®llers and coating materials .. The pore spaces made by the papermaking or converting process, e.g., inter®ber gaps, toroidal voids around ®ber±®ber bonds or ®ber±®ller contacts, and pores in coatings The pore spaces in porous materials can also be classi®ed by considering whether they are interconnected with each other [22,61]. That is, interconnected pores are accessible from the outside through one or both ends. If accessible from only one end, the pore is referred to as a dead-end pore. Noninterconnected pores are not accessible at all. The interconnected part of a pore system is called the effective pore space of a porous medium. It is possible to quantify some aspects of the porous structure of paper, employing space-related terms such as porosity, speci®c surface area, pore size, or pore size distribution, and others. Using these terms, many attempts have been made to elucidate the porous properties of paper and paperboard [8,16,50,74]. Porosity is generally recognized as the ratio of pore volume to total volume, although it was originally a generic term denoting the quality or state of pore structure. Sometimes the term is used for a speci®ed property such as pore volume or air permeability. Because the term is often used for air permeability of paper, ``porosity'' in this chapter means the pore volume ratio to avoid confusion. Speci®c internal surface area is de®ned as the internal surface area of pores either per unit mass or per unit bulk volume of a specimen. In the case of paper, the former is preferable because for paper mass is a more accurate quantity than its bulk volume. Internal surface area plays an important role in the collection ef®ciency of ®lter paper. Pore size or pore size distribution would be a very desirable quantity for characterizing pore structure if it could be ascertained with precision. Unfortunately, there is no acceptable experimental method for obtaining a geometrically de®ned quantity describing the size of an irregularly shaped pore. In spite of this uncertainty, the term pore radius or pore diameter as an expression of pore size is widely used in paper science, with the tacit understanding of the difference from the actual features of pores. Pore size or pore size distribution expressed in terms of pore radius can be de®ned only by both the experimental technique and the analytical method used to determine this quantity. Therefore, one must pay careful attention when comparing data resulting from different test methods. Another attempt to describe the porous structure is an approach from statistical geometry developed by Corte and Kallmes [17] and Corte and Lloyd [18]. Although their theory, which assumes for paper a multiplanar model consisting of a pile of ultrathin two-dimensional sheets, intends to describe the effective pore size distribution and the maximum pore size in terms of ®ber and sheet dimensions, it requires further development to interpret the porous properties of paper suf®ciently. Much information on the pore structure of paper can be obtained intuitively by using electron microscopy and other microscopic techniques. Here, scanning electron and video microscopes are very helpful tools for obtaining real three-dimensional images of the pore structure with their easy sample preparation procedures (see Chapter 5 of Volume 2).

Porosity and Gas Permeability

B.

269

Measurement of Porosity

General Methods Porosity, as de®ned earlier, is the ratio of pore volume to total volume of paper, that is Porosity ˆ

pore volume of paper total volume of paper

ˆ1

bulk density of paper solid density of paper

…1†

Therefore, the methods for determination involve either volume or density measurement. If the bulk density is measured and the solid density of a specimen is known or calculated from the density and the fractional ratio of each material making up the specimen, the porosity can be calculated. Methods for Bulk Density Measurement Bulk density of paper can be calculated from measurements of sheet grammage and thickness. The values obtained, however, vary slightly depending upon the thickness-measuring method under the conditions used due to the surface roughness and compressibility of paper [88]. Caliper Method The thickness can be conventionally obtained with a standard paper caliper. In the TAPPI standard method a rubber platen caliper is preferable to minimize the effects of surface roughness [85]. Mercury Method Another method to measure the bulk density of paper involves using mercury as a nonwetting liquid for paper substances. Several methods, such as the mercury capacitance [9], mercury pycnometry [63,88], and mercury buoyancy techniques [81,88], are available for this purpose. The values of bulk density obtained from these methods may be higher than that obtained from the usual caliper method, because the caliper method yields a greater thickness (and therefore less volume). Methods for Solid Density Measurement The density of the solid fraction of a porous medium can be measured accurately by the displacement method. The basic principle of this method is direct measurement of the volume of gas or liquid that is contained in the pore spaces. In the case of paper and paperboard, it is necessary to take care in the selection of the displacement ¯uid. The following requirements need to be met: . The size of the ¯uid molecule must be suf®ciently small compared with that of the pore spaces. .. The ¯uid must be inert to the sample, that is, chemical and physical effects must be absent. ... The liquid surface tension must be suf®ciently small. Liquid Displacement Method Although the most desirable ¯uid is helium gas, it has been reported that comparable values can be obtained with the use of nonpolar liquids such as heptane, benzene, and toluene [32]. When the measurement is made, the pore spaces of the specimen must be ®lled completely with the displacement liquid. For this purpose, it is advisable to use a pycnometer with a device for degassing and ®lling the liquid under vacuum [82,84]. Mercury porosimetry is appro-

270

Yamauchi and Murakami

priate for this purpose. The density is determined as the ratio of ®ber mass to ®ber material volume. The volume value is determined by measuring the difference in mass between a pycnometer ®lled with mercury and one ®lled with mercury and paper intruded with mercury, while the mass is measured as oven-dry mass [91]. Finally, the solid density of a specimen s can be calculated by s ˆ where

mS mL mLS L

mL = = = =

mS L …mLS mS †

…2†

oven-dry mass of specimen mass of displacement liquid required to ®ll pycnometer mass of specimen ‡ liquid required to ®ll pycnometer density of liquid at temperature used

Gas Displacement Method The gas displacement method involves measuring the change in either pressure or volume to maintain a given pressure at a constant temperature when a known amount of gas is introduced into a sample chamber that contains a known mass of the specimen. The solid volume of the specimen can be calculated from the ideal gas law equation. For example, if an evacuated sample chamber is connected to a gas reservoir containing a known amount of gas, the solid volume of a specimen, VS , can be calculated as VS ˆ V1 ‡ V2 where

p1 V1 p12

…3†

V1 = volume in gas reservoir p1 = pressure in gas reservoir V2 = volume of sample chamber p12 = pressure when gas reservoir is connected to sample chamber

A variety of apparatuses using helium gas have been designed [62,84]. A relatively large amount of specimen material is required to obtain accurate results. However, an apparatus designed by Schumb and Rittner [62] allows accurate measurement on small specimens, for example, 0:5 cm3 of specimen having about 50% porosity, with an accuracy of 0:1%. An apparatus for measurement of adsorption phenomena can also be used for this purpose. With this procedure, it is essential to completely remove any moisture and air that might be adsorbed on the specimen and to allow enough time for the helium molecules to penetrate throughout the micropores of the specimen. C.

Measurement of Speci®c Surface Area

General Methods Gas Adsorption Method If the area occupied by a molecule adsorbed on a solid surface m is known, the speci®c surface area can be calculated from the following equation by measuring the amount of gas vm required to cover the entire internal surface of a specimen as a monomolecular layer. That is,

Porosity and Gas Permeability

271

vm NA m mS

…4†

Sˆ where

S = speci®c surface area NA = Avogadro's number mS = mass of specimen

Methods to determine the amount of gas adsorbed can be classi®ed into two groups. The ®rst is a volumetric method that accurately measures the pressure and volume of the gas before and after adsorption at a constant temperature (in practice this is achieved either at a given constant volume or at a given constant pressure). The second is a method to directly measure the mass of gas adsorbed using a special balance such as an electric balance. A wide variety of apparatus modi®cations have been devised, and various types of commercial instruments are now available for the measurement; readers are referred to the books of Young and Crowell [96] and Gregg and Sing [28] for further details. In practice, these methods require rather elaborate high vacuum apparatus in which very precise pressure±volume measurements must be made. As improvements, Nelsen and Eggertsen [46] and, later, Stone and Nickerson [66] developed dynamic nitrogen adsorption methods using conventional thermal conductivity cells with successful application to paper [53]. The amount of gas adsorbed as a monolayer can be calculated from the dependence of the amount of adsorbed gas upon the pressure at a constant temperature, that is, the adsorption isotherm. Although there are many theoretical controversies over the calculation methods employed for this purpose, the Brunauer± Emmett±Teller (BET) method [28] is generally used in practice in preference to others, such as the B-point method [28], the relative method of Harkins and Jura [29], and the Fu±Bartell method [24]. A widely accepted method to measure the internal surface area of paper is a combination of the determination of the nitrogen adsorption isotherm at liquid nitrogen temperature with calculation by the BET equation. Haselton [30] studied several gases as adsorbates and considered the analytical methods for determining the surface area. He concluded that the aforementioned combination could be recommended for surface area studies on pulp and paper. Solution Adsorption Method The internal surface area can also be measured, using the same principle as the gas adsorption method, by adsorption of a solute from its solution in a nonswelling solvent. Stamm and Millett [65] applied this method to evaluate the internal surface area of wood and ®lter paper using stearic acid solution in benzene. Optical Method It has been well recognized that the speci®c scattering coef®cient calculated by the Kubelka±Munk equation from re¯ectance measurements of pulp sheets is proportional to the BET surface area [31]. This method has been widely used in relative comparisons of the internal surface area of various pulp sheets. Readers may consult the chapter on optical and appearance properties (Chapter 4 of Volume 2) for greater detail concerning the interpretation of re¯ectance measurements. Other Methods Attempts have been made to determine the hydrodynamic surface area of pulp mats from ¯uid permeability measurements by application of the

272

Yamauchi and Murakami

Kozeny±Carman equation or its modi®cations [7,9]. This method has some problems that need to be overcome in order to be applicable to paper and paperboard as described in Section II.A. For wet pulp ®bers or pulp sheets, the silvering-catalytic method of Clark [10] is available for determining surface area. D.

Measurement of Pore Size Distribution

General Methods Although the pore size of paper and the pore size distribution are generally expressed in terms of pore radius or pore diameter, one should keep in mind that the pore radius speci®es only an experimentally determined value and not a geometrically de®ned value, as discussed in Section I.A. Capillary Pressure Method According to basic capillary theory, the pressure difference across a curved interface between two ¯uid phases,
p, is given by the Laplace equation,   1 1
p ˆ …5† ‡ r1 r2 where

= interfacial surface tension r1 ; r2 = principal radii of curvature of an interfacial meniscus

If the harmonic average of the two radii of curvature is r, Eq. (5) becomes, for an imperfect wetting ¯uid,
p ˆ where

2… cos † r

…6†

r = capillary radius  = contact angle between a ¯uid and a solid body

There are some useful methods, including the mercury intrusion porosimetry method, that are designed to determine the pore size distribution by using this equation. As another method, the capillary suction method was applied by White and Marceau [83] to determine the pore size distribution of wetted blotter sheets. Although one cannot expect it to yield accurate results, this method is applicable to dry sheets if a nonswelling liquid is used as a saturant. The capillary rise method [74], which is based on an analogous principle, is described in Chapter 7 of Volume 2. Gas Drive Method The gas drive method was ®rst introduced to paper by Corte [16] as the dioxane method and has since been applied to a variety of wood ®ber sheets [7], coatings [27], and, recently, to air ®lters using a commercially available instrument [70]. The determination is conducted by covering a test specimen with a wetting liquid that is allowed to saturate the pore structure and subsequently applying gas pressure below the specimen to eject the liquid from the pore structure. Any path that passes through the sheet will be opened for gas ¯ow when the applied pressure exceeds the capillary pressure. The relationship between the ¯ow rate of gas as an immiscible ¯uid and applied pressure is recorded. Assuming that the paths in the specimen consists of a group of cylindrical capillaries that have equivalent capil-

Porosity and Gas Permeability

273

lary pressures and the same length (and if the ¯ow rate of gas through opened paths can be described by the Hagen±Poiseuille equation for a cylindrical capillary), the pore size number distribution can be obtained from the ¯ow rate±capillary pressure data. It has been pointed out that the pore size distribution estimated by this method contains several uncertainties introduced by oversimplifying the model and the related assumptions [27]. In spite of this, it is also true that this method can provide valuable information on the pore paths through paper, especially ®lter papers [3]. Quantitative Stereology Method During the past decade remarkable developments were achieved in the theory and equipment for quantitative image analysis for quantitative stereology [20,77]. This method, using optical means to analyze sections of the sample, has some advantages over other methods. However, a critical problem of this method is the need to prepare cross sections without a structural change. Generally a change of internal structure or the creation of artifacts in the thin sections of paper during their preparation is unavoidable even in a sample embedded in epoxy resin. In recent years confocal laser scanning microscopy (Chapter 5 of Volume 2) and focused ion beam sectioning have been applied to the analysis of paper cross sections without the structural changes or artifacts inherent to mechanical sectioning [76]. If either of these new techniques is available, the quantitative stereology method will provide a well-de®ned pore size distribution. Comparisons of the pore volume distribution obtained with this method with that from mercury porosimetry were made by Dullien and Dhawan [23] for a sandstone structure and by Climpson and Taylor for coating layers [13]. They found, as expected, that the distribution calculated from mercury penetration analysis has a much lower mean value and is narrower than that calculated from quantitative stereology, as shown in Fig. 1. Probably a combination of this method with mercury porosimetry using penetration and retraction curves gives the best results for determination of pore size distributions at the present time [22,23]. Gas or Vapor Sorption Method Below the critical temperature of a gas, gas molecules are adsorbed on a solid surface to form one or more layers, the number of layers increasing with gas pressure. If micropores are present, capillary condensation

Fig. 1 Comparison of mercury porosimetry (a) with quantitative microscopy (b) for clay structure. (From Ref. 13.)

274

Yamauchi and Murakami

can also occur simultaneously. Using these relations the pore size distribution can be determined. Nitrogen gas [67,73] and benzene vapor [44] have been used as adsorbates for pulp ®bers and papers. Because the measurable range is limited to a few decades of nanometers in radius, this method does not detect the bulk pore spaces of paper and paperboard. Other Methods The X-ray small-angle scattering method, based on measurement of the X-ray scattering pattern corresponding to the electron density distribution in the solid part of a sample, is useful for detecting micropores less than 100 nm in diameter irrespective of their interconnection. Hermans et al. [33] applied this method in a study of micropores in cellulosic ®bers. Banacki and Bowers [4] advanced a unique method for determining pore size distributions within ®lter papers. It involves passing a nonswelling liquid, in which a large number of spherical beads of different sizes are suspended, through a specimen. The beads passed through the specimen are analyzed in size and number. The pore radius obtained by this method implies that of a circle inscribed in the narrowest part of the pore through which the beads have passed. Mercury Intrusion Porosimetry This method involves the intrusion of mercury as a nonwetting ¯uid into an evacuated porous medium. The volume of mercury intruded or retracted and the pressure applied are measured simultaneously. Based on the idea of Washburn [80], the mercury porosimeter was ®rst developed by Ritter and Drake in 1945 [21,54]. Since then this method has been widely applied to a very large number of different materials from incompressible inorganic materials to compressible organic materials such as cigarette ®lters [15], ®bers [49], and papers [44]. There is an extensive literature on the use of this technique, and it has become one of the most popular standardized methods of determining the pore size distribution [22,23]. Mercury intrusion porosimetry is applicable to a wide range of pore sizes, including the whole spectrum of pore sizes of paper. Relative Macro- or Micropore Structure of Paper In analyzing the pore structure of paper by using mercury porosimetry, it is useful to categorize all pores into relative macro- and micropores [95]. The conceptual understanding of relative macro- and micropore structures within paper webs is as follows. ``Relative macropores'' refers to the larger voids in paper webs made up mainly of inter®ber voids and voids between other furnish components with equivalent diameters greater than 1:5 m, which corresponds to mercury intrusion into paper webs at pressures below 0.98 MPa. Relative micropores, on the other hand, have diameters of 1:5 m or less and consist of pores or voids within ®ber walls and pit apertures and also the ®ne voids associated with inter®ber bonds and/or bonds between ®ber and ®brillar or particulate ®nes material, between ®bers and ®ller particles, and between adjacent ®ller and other furnish components. Usually over 90% of the pore volume of paper is categorized into the macropore component [91]. Principle and Limitations Mercury intrusion porosimetry involves enveloping an evacuated sample with mercury and applying pressure to the meniscus. The mercury penetrates into the pores that have capillary pressures equivalent to the applied pressure. If each change in the meniscus level of the mercury is measured versus stepwise rising or descending pressures, a cumulative volume±capillary pressure

Porosity and Gas Permeability

275

curve with penetration and retraction branches can be obtained. As show in Eq. (6), the capillary pressure is inversely proportional to the capillary radius. Using this relation, the penetration branch of the cumulative volume±capillary pressure curve is usually converted to a cumulative pore size distribution curve. Since mercury penetrates through the most accessible pore path, regardless of its direction and tortuosity, this method is effective only for interconnected pores. In addition, no information is obtained about the anistropy of a porous structure. Furthermore, this method will not detect the dimension of those pores that are accessible only through necks narrower than the pore itself; those pores are assigned the diameter of the narrow neck. In applying this method to paper and paperboard, one item of concern is whether or not compression of a specimen and collapse of the voids take place during measurement under high pressure. However, the mechanism of compression in the mercury intrusion method would be quite different from that in compression testing of paper. The intruded mercury will not act to collapse the lumen. It would rather tend to support the pore structure. Although there is no experimental solution to this problem, data suggesting the occurrence of signi®cant compression have not been reported, at least not in the macropore region. Finally, the contact angle that mercury makes with a solid surface plays an important role in determining the capillary pressure, as seen in Eq. (6), but the true value of the contact angle is a rather elusive quantity. In practice, assuming that the contact angle is constant and independent of the applied pressure, one particular value between 130 and 142 is generally used. For example, Ritter and Drake [54] and others [44,49] assumed the value of 140 for the contact angle of mercury as an average for most organic materials [1]. On the other hand, Rootare and coworkers [56,57] recommended using the value of 130 because of the good agreement between mercury porosimetry and BET surface areas for inorganic powders that have low energy surfaces. These values have also been used for paper materials [27]. Mercury penetration and retraction hysteresis may be partly caused by the contact angle hysteresis between advancing and receding angles. However, the contact angle measurement for paper is not well de®ned, and therefore the pore radius size may not be exact. It would be advisable to express the result by simultaneous use of capillary pressure and equivalent pore radius. Apparatus The mercury porosimeter is essentially a combination of pycnometer, pressure-applying and vacuum systems. A typical commercial instrument consists of a penetrometer cell with an accessory to measure the volume change in mercury, mercury and pressure-transmitting liquid reservoirs, a pressure vessel and pressure generator, a vacuum pump, and a series of pressure gauges. Because mercury penetration and retraction hysteresis of paper is often observed at a fairly low pressure region, as shown in Fig. 2, it is desirable that the instrument be able to measure the retraction curve as the pressure is lowered even below atmosphere pressure [94]. Procedure 1. Specimen setting. A known mass of dry specimen material is placed carefully into the penetrometer cell. It is essential to avoid the formation of pore spaces that might result from overlapping the specimen surfaces, which might lead to a false pore size distribution [47,91].

276

Yamauchi and Murakami

Fig. 2 Cumulative mercury volume percent±pressure curves of four-cycle experiments for handsheets of unbeaten pulp. (From Ref. 92.)

2. Evacuation. The cell with the specimen is evacuated completely to remove gases and vapors adsorbed in the specimen. 3. Intrusion of mercury. The space within the cell is ®lled with mercury below atmosphere pressure, and the relationship between the applied pressure up to atmosphere pressure and the volume change in mercury is measured. After the cell with the specimen is moved to the pressure vessel ®lled with the pressure-transmitting liquid under atmosphere pressure, the relationship between the applied pressure and the mercury volume change is measured during stepwise pressurizing or depressurizing of the liquid. Note that in very high pressure regions, the value of the volume intruded into voids should be corrected for the compressibility of mercury itself. Mercury Penetration and Retraction Hysteresis Relating to Pore Structure The pore size distributions are usually calculated from mercury penetration curves on the basis of the ``bundle of capillary tubes'' model. However, real pore structures of paper and coatings are far from this model. Alternatively, the ink bottle (bottleneck) pore model or the interconnected pore (packed spheres) model will give a much better approximation to the actual pore spaces in papers and coatings. An analysis of the mercury penetration and retraction hysteresis curve is required to investigate the details of pore structure related to these models [5,38,71]. For example, no mercury retraction on depressurizing suggests ink-bottle type pore structure [49]. On the other hand, the characteristic behavior for the packed spheres model consists of abrupt mercury penetration, mercury ®lling of the toroidal void volume around contacts between packed spheres following the abrupt (breakthrough) penetration, then an abrupt mercury retraction [5,38,43]. Kruyer [38] argued that when the pressure is decreased after all pores are ®lled with mercury, toroidal voids will be emptied ®rst and, ®nally, at the point where isolated mercury surfaces interfere with each other, the entire pore space will be emptied of mercury. This point is determined by

Porosity and Gas Permeability

277

the radius of the spheres inscribed in the interstices of the packings. The size at breakthrough (abrupt) penetration, on the other hand, is approximately equal to the radius of the circle inscribed in the opening of the packings. The radius of this circle is invariable smaller than the radius of the sphere controlling the abrupt retraction of mercury from the interstices, resulting in hysteresis. Massive retraction of mercury characterizes the hysteresis for the model. The speci®c pressures at which abrupt mercury penetration and retraction occur, breakthrough pressure and withdrawal pressure, are the characteristic parameters in the packed spheres model [5,58]. The observed mercury penetration and retraction curves of papers are similar to those of packed spheres. Therefore, the mercury penetration and retraction behavior of paper can be described as follows in order of pressurizing and depressurizing: A slight mercury penetration before abrupt penetration corresponds to the mercury ®lling in large surface pores (and/or compressive deformation); abrupt penetration, penetration into interconnected large spaces between ®bers through openings; gradual penetration after the breakthrough, mercury ®lling around ®ber-®ber bonds (on pressurizing); mercury retraction at higher pressure, the retraction from around ®ber-®ber bonds; and abrupt retraction, mercury retraction from large spaces (on depressurizing). Actually, the sizes of openings and interstices between ®bers in paper or coatings are distributed, so apparent breakthrough and withdrawal pressures are determined as the pressures at which mercury penetrates and retracts at the maximum rate [91,92]. Scanning curves within the main hysteresis loop can be obtained by reversing the direction of pressure change at some intermediate points along the penetration curve to obtain further information about the relationship between mercury porosimetry and the pore structure [5]. An example for handsheets of unbeaten kraft pulp is shown in Fig. 2. The penetration branch rises steeply in the range of relatively low pressure where mercury is intruded into relatively large interconnected pore channels in the ®ber network structure. The retraction branch shifts toward the lower pressure side compared with the penetration branch. Alternate penetration and retraction of mercury will produce a hysteresis loop. The pore model expressed by a closed hysteresis loop is the interconnected pores (packed spheres) model for mercury porosimetry. The remaining pores, in which mercury is retained after retraction, may be interpreted as being related to drops of mercury within a sample that become isolated and disconnected from the bulk continuum mercury during the retraction process. The ratio of mercury retraction volume to total mercury penetration volume at the beginning point of the experiment, termed the mercury retraction ratio, decreases with an increase in beating degree [92]. The decrease in the ratio as beating increases indicates that the pore structure shifts from the packed spheres model type to the ink bottle model type, that is, a decrease in the degree of interconnection [92,95]. For the packed spheres model, there is a theoretical relationship between breakthrough pressure, porosity, and sphere diameter [42]. In spite of beating degree and wet pressure, the relation for handsheets from softwood kraft pulp agrees with that for packed spheres of 14 m diameter and the assumed extension as shown in Fig. 3. Thus, conceptually such trends could indicate that sheet consolidation brought about by pulp re®ning and/or wet pressing corresponds to the packing behavior of packed spheres of 14 m diameter and that individual spheres could be considered equivalent to the basic structural elements such as ®ber segments

278

Yamauchi and Murakami

Fig. 3 Relationship between porosity and apparent breakthrough pressure for kraft pulp handsheets. Solid line shows the theoretical curve for packing spheres of 14 m diameter. (From Ref. 92.)

within a paper structure [92]. The theoretical relationship and the assumed extension is very useful for a discussion of the relationship between pore structure and substantial ®brous structure [91]. Clay size and pore size of coating layers have also been successfully estimated by using the relationship between breakthrough pressure, porosity, and sphere diameter [89]. Calculation for Differential Pore Size Distribution As a pore size distribution the differential distribution has been used more than the cumulative distribution. If dV is the pore volume in the range of pore radii between r and r ‡ dr, we may write dV ˆ D…r†dr

…7†

where D…r† is a volume distribution function of pore radii. If the mercury contact angle and its surface tension are assumed to be constant, it follows from Eqs. (6) and (7), that D…r† ˆ

p2 dV …2 cos † dp

or

D…r† ˆ

p dV 2… cos † d log p

…8†

where the right-hand terms of both equations consist of known or measurable quantities. That is, if the cumulative volume of mercury intruded is plotted as a function of the applied pressure, the slope of this curve is equal to the value of dV=dp. Thus, a derivative function D…r† is easily obtained by graphic differentiation, and a plot of the value of D…r† as a function of r gives the differential pore size distribution curve. A computer program for the pore volume and pore area distribution calculations from mercury porosimetric data was reported by Rootare and Spencer [58]. The pore size distribution has been exclusively given from mercury penetration curves

Porosity and Gas Permeability

279

obtained experimentally. However, the distribution obtained from the penetration curve is a distribution of the neck pores rather than one of pores (interstices). Alternatively, the distribution obtained from the retraction curve might be close to that of pores according to the packed spheres model unless a considerable amount of mercury remains in the sample tested. Expression of Result The result obtained is usually expressed as a volume distribution curve of either pore radius or capillary pressure in a cumulative or differential form. If the speci®c volume displaced by the specimen at a given applied pressure can be calculated, it is possible to plot the speci®c displacement volume as a function of capillary pressure or pore radius [68]. This expression, which is called a Stone± Scallan plot, may be very useful in investigating micropore structure such as the surroundings of ®ber±®ber contacts [91,95]. Taking into consideration the uncertainty of the pore size distribution, four parameters obtained from mercury porosimetry can be recommended to characterize the pore structure of paper: porosity, (apparent) breakthrough pressure corresponding to the mean diameter of openings, (apparent) withdrawal pressure corresponding to the mean diameter of pores, and the mercury retraction ratio showing the degree of interconnection [91,95]. An example showing the effects of wet pressure and degree of beating on the apparent breakthrough pressure of kraft pulp handsheets at the midstage of sheet consolidation (after wet pressing) is shown in Fig. 4. Surface and Internal Pore Structures Although ``pore structure'' generally refers to the internal pore structure, the pore structure of paper consists of surface and internal pore structures. It is dif®cult to measure and analyze the surface and internal pore structure separately. However, analysis of low grammage sheets, which mainly consist of surface layers, may provide information on the surface pore structure. An estimated grammage of the surface layer for one side of paper is within a range of 10±20 g=m2 depending on ®ber length and type of pulp [88]. Since surface pore volume is constant and internal poor volume increase with an increase in grammage, the ratio of the internal pore part to surface pore part increases with increasing grammage. For example, a paper of 30 g=m2 is thought to be mainly composed of two-surface pore structure from the wire and felt sides and has little internal pore structure. Mercury penetration and retraction curves for a series of handsheets varying in grammage from 30 to 120 g=m2 are shown in Fig. 5. A signi®cant difference in the penetration curves is found for sheets between 30 and 60 g=m2 , that is, the mercury penetration volume ratio at the lower pressures before the abrupt penetration decreases remarkably with an increase in the grammage from 30 to 60 g=m2 and slightly decreases from 60 to 120 g=m2 sheets. Because almost all of the pore volume of 30 g=m2 sheets is derived from surface pores, abrupt mercury penetration through openings does not occur here. On the other hand, the penetration volume at higher pressure after abrupt penetration is almost the same for these various grammage sheets, and, further, the mercury retraction curves representing mainly the interstices in the ®ber network show little effect of grammage. Taking into consideration that the toroidal pore structure and the interstices between ®bers are derived from the fundamental small-scale and large-scale ®ber network structure, respectively, these results suggest that surface and internal pore structures are fundamentally the same, that is, the toroidal pore structure around the ®ber±®ber bonds and the pore structure between ®bers are common basic structures for surface and internal pores.

280

Yamauchi and Murakami

Fig. 4 Effect of beating and wet pressing on apparent breakthrough pressure of wet handsheets after wet pressing.

Other Information Concerning Substantial Structure A change in the pore structure of paper due to differences in papermaking and converting followed by mercury penetration or retraction curves provides information on the substantial structure of papers [94,95]. For example, the spatial distribution of rubber in latex-impregnated paper sheets was investigated by comparing cumulative pore volume curves before and after latex impregnation [94]. The theoretical relationship between porosity, breakthrough pressure, and diameter of packed spheres is very useful for investigating the substantial ®ber network structure related to the component ®bers and sheetmaking parameters as described earlier. Methods for Wet Pore Structure It is often necessary to investigate the pore structure of paper in wet condition as well as in dry condition. For relative macropores in ®ber network structures, the above-mentioned methods including mercury porosimetry have been successfully employed after water in the wet webs was removed by freeze drying or other drying methods [44,95]. For relative micropores such as those in wet ®ber walls, some speci®c methods such as the solute extrusion method have been employed. An application of the solute extrusion method to pulp

Porosity and Gas Permeability

281

Fig. 5 Effect of sheet grammage on mercury penetration and retraction curves of kraft pulp handsheets. (From Ref. 90.)

®bers is based on the work of Stone et al. [69]. The fundamental idea is that a probe polymer will penetrate into pores of equivalent size, resulting in a change in the polymer concentration. By using a series of polymers of different sizes such as dextrans, the pore size distribution ranging from 1 to 150 nm is displayed as a relationship between inaccessible water and polymer size. However, the assumption of penetration may not be fully applicable to the ®ber wall, thus making the distribution obtained questionable. This method is now employed as a standard method to determine ®ber saturation point rather than as a method of measuring pore size distribution [41]. Recently, alternative methods for the wet pore size distribution have been developed and applied to paper and coatings using differential scanning calorimetry (DSC) or nuclear magnetic resonance (NMR) spectroscopy. DSC Method The basic principle of the DSC method is the relationship between the size of pores and the degree of ice melting temperature depression. The DSC endotherm obtained can be converted into pore size distributions ranging up to 100 nm. The application of the method to silica gels or porous glasses is quite good [40], but for wet paper webs the situation is complicated by the gel effect of polysaccharides, because the melting point depression is also affected by the lower activity and increased entropy of the water in the hydrogel. NMR Method Various researchers have tried spin-lattice relaxation measurements of water in wet pore structures by using nuclear magnetic resonance [19,25]. The basic principle of the NMR method is that liquid in close proximity to the pore wall will undergo spin-lattice relaxation at a much faster rate than liquid far removed from the pore surface. The ratio of surface-affected liquid to bulk liquid will be

282

Yamauchi and Murakami

different in pores of different sizes. The pore size distribution derived from a distribution of spin-lattice relaxation time has been ®tted with that from mercury porosimetry and is used mainly to investigate relative micropore structure up to 100 nm [25]. Probably, the above-mentioned gel effect should also be considered in this method.

II.

GAS PERMEABILITY

A.

Fundamental Description

The gas permeability of paper and other sheet materials is the ability to allow the ¯ow of gas or vapor through the structure of a sheet under a pressure gradient. It is expressed by a permeability coef®cient or by various equivalent quantities in arbitrary units. For air, the terms air permeability and air resistance are used. These are important properties from the fundamental structural standpoint and also as practical descriptive measurements for such papers as ®lter papers and packaging papers. The gas ¯ow through a pore path of porous paper is treated as a laminar or viscous ¯ow, whereas gas permeation through nonporous papers, such as barriercoated papers and ®lm-laminated papers, is treated as a diffusion-type permeation. Diffusion-type permeation is also observed in water vapor permeation of paper. But the behavior of water vapor passing through paper is quite complicated, as described later, because of the great af®nity for water and water vapor shown by the ®ber polysaccharides. Flow-Type Permeation (Darcy's Law) A quantitative description of ¯ow through porous media begins with an experimental rule found by Darcy. This rule, describing that the rate of ¯ow passing through a powder bed is proportional to the pressure gradient of the ¯ow, has been extended to a variety of porous media. There are various equations to express this rule and its modi®cations, known collectively as Darcy's law [61]. We consider the ¯ow to be of a viscous streamline or laminar type. If a ¯uid ¯ows across a cross-sectional area A at a volume rate Q, Darcy's law is described by the equation Q ˆ A

K

@p
p ˆK @x L

…9†

where the ¯ow is assumed to be parallel to the X axis and
p is the total pressure drop across a sample of length L. The proportionality constant K de®ned by Eq. (9) depends on both the properties of the porous medium and the nature of the ¯uid. It is called the volume permeability coef®cient or permeability constant [61], and its dimensions are L3 TM 1 Another permeability coef®cient, one that depends only on the structure of the porous medium, can be de®ned by introducing a relationship stating that the rate of ¯ow is inversely proportional to the viscosity of the ¯uid. That is, for incompressible ¯uids (a liquid, for example), Q ˆ A

Kv

@p
p ˆ Kv @x L

…10†

Porosity and Gas Permeability

283

where  is the viscosity of the ¯uid. The proportionality constant Kv is called the permeability coef®cient for viscous ¯ow or speci®c permeability [61], and its dimensions are L2 . Although mixed units, ``darcy'' or ``cm/s for water,'' are widely employed in petroleum technology and groundwater hydrology, consistent SI units, m2 , are conventionally used for a wide variety of porous sheet materials. In the case of gases, the volume rate of ¯ow changes continuously as the pressure varies. However, the mass rate of ¯owÐthe product of the volume rate and the density at a given pointÐmust remain constant, and, further, for isothermal ¯ow the product of the volume rate and the pressure also remains constant. Equation (10) becomes p1 Q1 ˆ pQ ˆ or

Kv

  Q1 Kv pav
p ˆ  p1 L A

where

Q1

pav

p @p A  @x

…11a†

…11b†

= apparent volume ¯ow rate at either surface of a specimen at which the pressure is p1 =mean pressure of system

If
p is very small compared with p1 , the ratio pav =p1 is near unity. In such cases the compressibility of the gas can be ignored. For example, if the volume change in compression can be held to within less than 1%, in the case of air at STP (273:15 K ˆ 0 C; 1:01325  105 Pa ˆ 760 mmHg), Eq. (10) can be applied instead of Eq. (11a) or (11b) for pressure drops in the range of less than 2 kPa (approximately 20 cmH2 O). Limitations of Darcy's Law Darcy's law is valid only for viscous streamline or laminar ¯ow. Various effects that cause deviation from Darcy's law have been reported. Two effects that may occur at a high ¯ow rate and under reduced pressure require special caution when gas permeability experiments are performed. These are (1) turbulence and (2) molecular effects. Turbulent Flow If the rate of ¯ow through a pore path becomes large, an inertial force acts on the ¯ow in addition to the viscous force and the manner of ¯ow turns from laminar to turbulent. The Reynolds number at which turbulence starts is called the critical Reynolds number. The value of the critical Reynolds number varies, dependent on the shape and structure of the pore paths. A value of around 2000 for a straight, circular tube and values ranging from 0.1 to 75 for porous rocks and ceramics have been reported [61]. Molecular Flow If the pressure of a gas ¯owing in a capillary is reduced until its mean free path is an appreciable fraction of the capillary radius, the gas molecules 

Reynolds number is a dimensionless quantity and is de®ned by the equation Re ˆ ruav =, where r is the radius of the tube, uav is the mean velocity of ¯ow, and  is the density of the ¯uid.

284

Yamauchi and Murakami

slip at the capillary walls to cause deviation from Darcy's law. If the pressure is further decreased until the mean free path is much greater than the capillary radius, viscosity plays no part in the ¯ow, and such ¯ow is called the free molecular ¯ow or Knudsen ¯ow. The mean free path of a gas is inversely proportional to the molecular density. For example, the mean free path of air is about 60 nm under a pressure of 101 kPa (1 atm) at 273 K (0 C) and about 4:5 m under 1.3 kPa (1 cmHg) at 273 K. Because the pore spaces between ®bers in a paper sheet may range in size from 0:1 m to several tens of micrometers, special attention to the effect of molecular ¯ow should be paid in evaluating results of gas permeability experiments at pressures lower than atmosphere pressure. Relationship Between Permeability and Pore Structure of Paper The ideal goal of pore structure analysis is to develop a mathematical model that can quantitatively account for all the properties related to pore structure such as air permeability [22,52]. Before considering a relationship between the permeability and pore structure of paper, the effect of sheet grammage on permeability should be clari®ed. The permeability coef®cient of paper sheet decreases with an increase in its grammage and levels off at a grammage higher than about 60 g=m2 [93]. The change of permeability with grammage corresponds to completion of the internal pore structure as suggested by the change in the mercury penetration curve with grammage (see Fig.5). The resistance to ¯ow through a porous medium must be controlled by the neck pores such as openings of the packed spheres type of pores because they are paths of ¯ow [22]. Further, ¯uid cannot pass through dead-end pores; it passes only through interconnected pores. Thus, two parameters, porosity and the degree of pore interconnection, which is experimentally determined as the mercury retraction ratio, must be considered in discussing the relationship between permeability and pore structure. A logarithmic relationship between the modi®ed permeability coef®cient (permeability coef®cient divided by the two parameters) and the apparent breakthrough pressure, representing the size of openings, is examined. A linear relationship between these is found for a series of handsheets of softwood kraft pulp irrespective of wet pressure and beating degree as shown in Fig. 6. The difference in ®ber shape, rather than ®ber length, causes a shift of the linear relationship [45]. The behavior of ¯ow through a group of cylindrical capillaries is also analyzed by the use of the Hagen±Poiseuille equation. Thus, the volume rate q through a capillary is qˆ

r4
p 8L

…12†

where r is the capillary radius and L is the length of the capillary. The velocity of the ¯ow is q r2
p ˆ 8L r2

…13†

p The mean free path of a gas molecule can be described by the equation  ˆ …2 2Nd 2 † 1 , where d is the diameter of the gas molecule, N is the gas density, and  is the mean free path. 

Porosity and Gas Permeability

285

Fig. 6 Relation between modi®ed permeability coef®cient and apparent breakthrough pressure for kraft pulp handsheets. (*) Whole pulp; (*) long ®ber fraction; (~) short ®ber fraction. (From Ref. 93.)

Assuming that a porous medium consists of a group of cylindrical capillaries with uniform radii, where both ends of all capillaries are open to the direction of ¯ow, Eq. (13) becomes ! Q " r2
p ˆ …14† A 8L where " is the porosity of the medium and ˆ tortuosity ratio ˆ

effective length of capillary thickness of the specimen

Combining Eq. (14) with Eqs. (9) and (10) yields ! " r2 Kv ˆ K ˆ 8

…15†

This equation shows that the coef®cient Kv is proportional to both porosity and the second power of the capillary radius and that it is inversely proportional to the tortuosity ratio. Unfortunately, there is no method available to determine the value of the tortuosity ratio directly in the case of paper. Many attempts to apply this theory for pulp ®ber mats and papers have been made by Brown [9], Bliesner [7], Labrecque [39], and Peterson [48]. However, Scheidegger [60] and later Higgins and De Yong [34] pointed out that such an application is dubious because of the unrealistic assumption of a cylindrical capillary model and an unreasonably large value for the tortuosity ratio for paper.

286

Yamauchi and Murakami

Diffusion-Type Permeation (Fick's Law) A quantitative description of diffusion-type permeation is based on Fick's law. Consider a system in which two mixed gas phases having different compositions are in contact with each other at a constant pressure and temperature. There will be a transfer of gas molecules from the high concentration side to the low concentration side of the interface. If gas molecules are transferred a distance x along the X axis perpendicular to the plane, the number of gas molecules that pass through the plane is proportional to a partial concentration gradient, n ˆ A

D

@c @x

…16†

where n is the number of gas molecules permeated in unit time across an area A of the plane and c is the concentration of gas. The proportionality constant D is called the diffusion coef®cient, and its dimensions are L2 T 1 . The negative sign of the right-hand side of Eq. (16) indicates that the molecules transfer to the low concentration side. If n1 is the number of gas molecules ¯owing through the plane at position x, and n2 is the number ¯owing through the plane at position x ‡ dx, the concentration gradient across the thickness dx is   @c n n2 @c @ @c dx ˆ 1 c ‡ dx …17† ˆ D ‡D @t @x @x @x A Therefore, @c @2 c ˆD 2 @t @x and if D is variable dependent on C it can be expressed by   @c @ @c ˆ D…c† @t @x @x

…17a†

…17b†

Equations (16) and (17) are called, respectively, the ®rst and second laws of Fick. Diffusion-Type Permeation Through Nonporous Paper For a simple description, we consider a polymer ®lm that has no cracks or pinholes. When there is a difference in total or partial pressure between the two sides, gas molecules will transfer across the ®lm. The permeation of a gas or vapor through a polymer ®lm typically includes the following phenomena: 1. The gas or vapor dissolves in the ®lm at one surface. 2. The dissolved molecules diffuse through the ®lm under a concentration gradient. 3. The diffused molecules evaporate from the surface on the low concentration side. If we assume that the diffusion coef®cient is independent of concentration, we ®nd for the steady-state condition that n ˆ A

D

@c
c
p=RT ˆD ˆD @x L L

…18†

Porosity and Gas Permeability

287

where R is the gas constant. If Henry's law can be expected to hold at both surfaces of the ®lm, the concentration of gas or vapor at the ®lm surface will be proportional to its partial pressure. Therefore, since c is equal to n=Q, Eq. (18) becomes Q
p ˆ DSsol A L

…19†

The proportionality constant Ssol is called the solubility coef®cient, and its dimensions are LT2 M 1 . If Eqs. (9) and (19) are combined, the relation …20†

K ˆ DSsol

is obtained. Coef®cients D and S are both temperature-dependent and obey the Arrhenius equation [55]: D ˆ D0 exp… Ed =RT †

…21a†

Ssol ˆ Ssol;0 exp…
H=RT†

…21b†

where Ed ˆ activation energy for the diffusion process,
H ˆ heat of solution, and D0 and Ssol;0 are the diffusion and solubility constants irrespective of temperature, respectively. Now we consider a laminate consisting of n layers of ®lm or other sheet materials. The number of gas molecules that pass through each layer is uniform under a steady-state condition [6]. If
pi is the pressure difference across the ith layer, which has a permeability coef®cient Ki and thickness Li , this relationship can be expressed on a volume basis by Q
p1
p2
pn ˆ K1 ˆ K2 ˆ


Kn A L1 L2 Ln

…22†

The permeability coef®cient for the laminate overall is de®ned by Q
p ˆK A L

…23†

where
p ˆ

iˆn X iˆ1

pi

and

Lˆ

iˆn X iˆ1

From Eqs. (22) and (23), one obtains   iˆn 1 X Li 1 ˆ K L Ki iˆ1

Li

…24†

Water Vapor Permeation Through Paper Water vapor permeability is a very important property in terms of practical uses for such materials as packaging papers and building papers. The mechanism of water vapor permeation through paper is much more complicated than that of dry air. The reason for this is the interaction between paper substances and water vapor. Vollmer [78] determined the permeability of dense paper for air and water vapor and found that the permeability for water vapor is much greater than that expected on the basis of air permeability measure-

288

Yamauchi and Murakami

ments. The amount of water vapor transferred to the zero relative humidity side of a sheet increases exponentially with increasing relative humidity at the other side [59,64]. Part of the water vapor transport through paper is the result of ¯ow through relatively large pore paths. The remainder is of the diffusional transport type, caused by the interaction between water vapor and cellulosic ®bers. Water molecules adsorbed onto the surface of a cellulosic ®ber are often subsequently adsorbed or dissolved into the matrix of polysaccharides in the ®ber. They thus penetrate the ®ber wall progressively. In a manner similar to that described for the case of gas permeation through a polymer ®lm, the water vapor dissolves in the polysaccharide matrix at or near the cellulosic ®ber surface on one side of the sheet, diffuses through the matrix under a concentration gradient, and evaporates from the other surface at the lower concentration. This process is called solid-solution diffusion. If the relative humidity becomes very high, some of the microvoids in a sheet of paper will be ®lled by capillary condensation water. It is suggested that the transport of such condensed water results from a gradient of the capillary pressure [2]. It is true that permeation of water vapor through paper or paperboard occurs by the additive effects of the above-mentioned mechanisms, but there is an additional effect that results from the swellability of the ®bers by water. Although a number of theoretical investigations have been made [2,59,64], a complete explanation of water vapor permeation through paper is not well formulated because of the complexities of the structure of paper and the interaction between water and the structural components of paper.

B.

Measurement of Gas Permeability

Measurement of Flow-Type Permeation The permeability of a sample of paper or paperboard to the ¯ow of an inert gas can be measured by an apparatus that provides the pressure levels to achieve ¯ow rates that would be expected in actual use. Principle and Calculation The principle of the measurement is quite simple. It can be achieved by simultaneous measurements of the pressure drop and the ¯ow rate when a gas ¯ows through a specimen at a constant temperature. The permeability, measured as a permeability coef®cient or other arbitrary unit, can be calculated in either of two ways: (1) from Eqs. (9) and (10) if the compressibility of the gas can be ignored, that is, p1 =pav  1, or (2) from Eq. (11) if compressibility cannot be ignored. The thickness value, needed to calculate the permeability coef®cient, should be measured by using rubber platen caliper or equivalent methods to minimize the surface roughness effects [85,93]. Apparatus The apparatus consists of three partsÐthe gas supply system or air pump, the permeability cell, and the ¯ow rate measurement system. The permeability cell should be equipped with manometers to measure the pressure drop across the specimen and the absolute pressure in the cell if possible. In the case of air permeability, it is convenient to use the atmospheric pressure as a constant-pressure gas source and an air pump to be connected to the return line of the ¯ow rate measure-

Porosity and Gas Permeability

289

ment system. A capillary ¯owmeter [11] or a rotameter [93] is used to measure the rate of ¯ow. Procedure 1. The specimen is brought to equilibrium at the desired experimental temperature and humidity. 2. The specimen is clamped onto the permeability cell, and then a gas stream is introduced into the apparatus under the prescribed pressure. 3. The following quantities are measured: the ¯ow rate and the pressure drop across the specimen, the absolute pressure in the cell, and the temperature of the ¯owing gas. Measurement of Diffusion-Type Permeation Principle and Calculation If the volume on the low pressure side of the specimen in the apparatus is maintained constant during the measurement, the amount of gas that passes through the specimen can be measured as a pressure change. There are two types of methods: . Low vacuum methods such as that of ASTM D 1434, in which the volume in the low pressure side in the apparatus is maintained as small as possible .. High vacuum methods, in which very small pressure changes in a high vacuum are measured The latter may be preferable for high accuracy. One of the high vacuum methods, such as that used by Rogers et al. [55], is described below. The calculation is carried out as follows. From Eqs. (19) and (20), the volume permeability coef®cient K is expressed as Kˆ

…p1

QL p2 †A

…25†

where p1 is the pressure at the high pressure side and p2 is the pressure at the low pressure side. In Eq. (25), Q stands for the volume of gas at STP (273:15 K ˆ 0 C; 1:01325  105 Pa ˆ 760 mmHg) transmitted in unit time in a steady state. From a plot of p2 as a function of time,
p2 =
t in the steady state can be obtained. If V2 is the total volume on the low pressure side, K can be calculated as  
p2 273V2 L …26† Kˆ
t 1:013  105 Tp1 A where p2 is very small in comparison with p1 . The units of K are, of course, m4 =…Ns† ˆ m3 s= kg, but other units are often used for barrier materials, namely, cubic centimeters (STP) per second per square centimeter of cross section per millimeter of thickness per centimeter of mercury pressure drop across the specimen [1 cm3 …STP† mm=…s cmHg cm2 † ˆ 7:5  10 9 m4 =…N s† ˆ 7:5  103 mm4 =…N s†Š: If the diffusion coef®cient is independent of the concentration, the diffusion coef®cient can be calculated by using the equation Dˆ

L2 6

…27†

290

Yamauchi and Murakami

where  is the time lag. The time lag is equal to the intercept on the time axis when the linear part of the above-mentioned pressure±time plot is extrapolated to p2 ˆ 0. Apparatus A permeability apparatus consists of a gas reservoir, gas puri®cation traps, a permeability cell, a solvent trap, and a McLeod gauge; it is connected to a high vacuum system and a gas supply. Because several days are required for a measurement, air tightness of the apparatus is particularly important. The permeability cell consists of two halves; these are assembled with a guard ring that allows a mercury seal to form between the outer wall of the upper half and the inner wall of the lower half. The pressure at the high pressure side of the cell is measured by a manometer. Procedure 1. The specimen is clamped in the cell; then evacuation of the apparatus is performed several times. 2. A prescribed pressure is applied on the high pressure side of the cell. 3. At this moment, the measurement is started. The following quantities should be recorded as a function of time: the pressure at the high pressure side of the cell by the manometer, the pressure at the low pressure side by the McLeod gauge, and the cell temperature. The total volume on the low pressure side of the cell should also have been measured previously. Methods for Routine Control of Air Permeability A number of instruments for air permeability measurements have been designed for the quality control of paper and paperboard. Some of them are listed in Table 1. As expected from Darcy's law, Eq. (9), the volume of air permeated is proportional to the pressure difference across the sheet, the effective area of the sheet, and the time and is inversely proportional to the sheet thickness. Among these variables, the thickness is a speci®c property of the paper, and the pressure difference and the effective area are quantities determined by the instrument used. Therefore, if one of the remaining variables is measured by the volume of permeating air in a given period of time (rate of ¯ow) or the permeation time for a given volume of air, a relative value that speci®es the air permeability of a sample can be obtained. In the majority of these instruments atmosphere pressure is used as the constant-pressure source on one side of the specimen, and then positive or negative pressure is applied on the other side. The results from these instruments are expressed in terms of the arbitrary units speci®ed for each instrument or method. These values are often interchangeable theoretically [79] and experimentally [12]; some examples are listed in Table 2. Permeation Time Method This method involves measuring the time in seconds required for a constant volume of air to pass across a specimen of constant area under a speci®ed pressure gradient. One of the most widely used instruments is the Gurley densometer, which has been adopted for use in the standard methods in many countries. The result obtained by this tester is often called the air resistance or Gurley seconds. This tester can be applied in cases where the papers have permeation times ranging from 2 to 1800 s for 100 cm3 of air. For more impervious papers, an instrument in which the rate of ¯ow is measured directly is more desirable, because Gurley densometer measurements

cm3 /60 s cm3 /15 s cm3 /60 s Shef®eld units (by calibrated rotameter) Emanueli units (pressure difference in capillary ¯owmeter) mmH2 O (pressure difference in capilary ¯owmeter) mmHg (pressure difference in capillary ¯owmeter)

0.981 (100 mmH2 O) 5.35 (546 mmH2 O) 1.47 (150 mmH2 O) 10.3 or 68.9 (1.5 or 10 psi) Compressed air (depends on the sample) < 4:90 (500 mmH2 O) (depends on the sample Suction by vacuum pump (depends on the sample)

Emanueli porosity tester

Oken type denso-asperometer (backpressure-type air micrometer) Smooster (vacuum air micrometer)

Japan TAPPI No. 5 [35]

Japan TAPPI No. 5

TAPPI T-547

DIN DVM 3413, [79] JIS C2111

s/100 cm3 s/100 cm3 s/25 cm3

Gurley-Hill S-P-S tester Bekk tester Williams tester Volumetric methods Schopper densometer Emiel Greiner porosimeter Flowmeter methods Bendtsen porosity tester Shef®eld porosimeter

1.22 (124 mmH2 O) 64.3±37.6 (482±282 mmHg) Ð

Standard method and reference TAPPI T-460, CPPA-TS D14, SCAN-P19, JIS P8117, etc.

1.22 (124 mmH2 O†

Units s/100 cm3

Permeation time methods Gurley densometer

Pressure difference

Air Permeability Testers for Routine Control

Instrument

Table 1

Porosity and Gas Permeability 291

Gurley densometer vs. Oken type densoasperometer

Gurley densometer vs. Emanueli porosity tester

Gurley-Hill S-P-S tester vs. Shef®eld porosimeter

Gurley densometer vs. Emiel Greiner porosimester

log aG ˆ 0:965 log bO ‡ 0:0818

aG ˆ 523:023 P1 =P2

or

43:478aG L

1:134 log bSh

log aGH ˆ 3:70975 bE ˆ

1:16 log bSh

log aGH ˆ 4:3393

979 ‡ 0:675be be 1033

83046:6 ; 0:803aG ‡ 513:34

bEG ˆ

be ˆ

24096:38 bs

Gurley densometer vs. Schopper densometer aG ˆ

Conversion equation

Instrument

Table 2 Conversion Equations Between Air Permeability Testers

bO ˆ Oken reading in s/100 cm3

P2 ˆ pressure drop across the specimen

P1 ˆ pressure drop across a standard capillary

L ˆ thickness of the specimen in cm

bE ˆ Emanueli reading in the gauge

bSh ˆ Sheffield reading in the gauge

aGH ˆ Gurley-Hill reading in s/100 cm3

bEG ˆ corrected Emiel Greiner value in cm =300 s

3

be ˆ uncorrected Emiel Greiner reading in cm3 =300 s

bS ˆ Schopper reading in cm =300 s

3

aG ˆ Gurley reading in s/100 cm3

Symbol

[79]

[79]

[12] TAPPI T-547

[79]

[79]

Reference

292 Yamauchi and Murakami

Porosity and Gas Permeability

293

tend to be time-consuming. Another limitation of this tester is that it cannot be used for papers that have rough surfaces such as creped and corrugated papers. The Gurley-Hill S-P-S tester is quite similar to the Gurley densometer; the two instruments differ only in the manner of clamping the specimen. In the Gurley-Hill S-P-S tester, the specimen is held with a constant clamping pressure. Results from the two testers usually correlate almost perfectly, but for papers with poor surface smoothness one observes some differences caused by air leakage at the clamps. The Williams tester is also a constant-pressure type of tester. This tester was designed originally to evaluate the surface smoothness of paper by measuring the air leakage from the gap between adjacent surfaces formed when a specimen is folded in two. The value measured must be corrected for the volume of air that passes through the internal pore paths of the specimen. The measurement for the correction is accomplished by inserting a soft rubber plate with a circular hole in it between the two surfaces of the specimen. The value obtained by this procedure provides a measure of the permeability in the direction perpendicular to the paper thickness. The Bekk tester is a differential pressure gradient type of tester. A relatively high pressure gradient is used in this tester. Volumetric Method In the volumetric method, the volume of air passed in a constant interval of time in seconds across a constant area of the specimen under a speci®ed pressure gradient is measured. The instruments for this purpose may be classi®ed into two types: . ..

Constant-pressure gradient type tester such as the Schopper air permeability meter. Differential pressure gradient type tester such as the Emiel Greiner porosity meter. This tester may be used to measure relatively dense papers such as condenser papers.

Flowmeter Method The ¯ow rate of air passing across a constant specimen area under a speci®ed pressure gradient is determined. Instruments designed for the ¯owmeter method provide rapid measurements and can be used for a wide variety of papers having various air permeabilities. The Bendtsen porosity tester and the Shef®eld porosimeter are quite similar. In both instruments, a group of rotameters (two to four) covering an appropriate range of ¯ow rates is inserted on the line from a constant-pressure source of compressed air to a sample holder. The rotameters are selected according to the air permeability of the sample. Another type of instrument is made by applying the same principle to a capillary ¯owmeter. In this type of tester, a standard capillary is inserted into the air line on one side of the cell, and then the pressure drop across the calibrated standard capillary is measured. Two types of instruments are available. One type uses compressed air as a pressure source; examples are the Emanueli porosity tester and the Oken-type denso-asperometer. The other type, of which the Smooster [35] is an example, uses a vacuum pump as a suction source. In both types of testers, the pressure difference across a specimen depends on the permeability of the specimen and therefore will become larger in the case of denser paper. In SCAN-P26 a method is given to evaluate the air permeability of dense paper by measuring the pressure drop and the rate of ¯ow across a constant area of specimen under conditions of constant compressed air pressure. The pressure drop

294

Yamauchi and Murakami

across the specimen is measured by a mercury manometer, and the rate of ¯ow is measured by one of three rotameters in series or by a capillary ¯owmeter. The air permeability in this method, P, is de®ned as u …28† P ˆ av A
p where

uav = mean ¯ow rate A = test area
p = pressure difference across the specimen

The air permeability can be calculated as a P ˆ 1500 b

…29†

where a is the ¯ow rate in cubic centimeters per second, b is the difference between the mercury levels of the manometer in millimeters, and the units of P are mm3 =N s†: For highly porous sheet materials (such as porous papers, fabrics, and unbeaten pulp handsheets), another air permeability is de®ned in TAPPI T 251. The method involves measuring the rate of ¯ow through the specimen under a speci®ed pressure drop, that is, 124.4 Pa (12:7 mmH2 O; 0:50 in:H2 O). The air permeability is usually expressed as cubic centimeters per second per square centimeter or cubic feet per minute per square foot of area at 12.7 mm (0.50 in). water pressure. C.

Measurement of Water Vapor Permeability

The water vapor permeability varies not only with the partial pressure difference across a specimen but also with the concentration of water vapor (i.e., the relative humidity) and with temperature. In addition to these factors, some papers and heavy paperboards require very long periods to establish a steady state; for example, a barrier paper coated on both sides with a hydrophobic material usually requires several weeks or longer [36]. The test conditions are normally speci®ed. In TAPPI standards the following conditions are designated for measurement of the water vapor transmission rate: . At 293 K (23 C; 73 F) with an atmosphere of 50% RH on one side and a desiccant on the other (TAPPI T-448) .. At 310.8 K (37:8 C; 100 F) with an atmosphere of 90% RH on one side and a desiccant on the other (TAPPI T-464) For homogeneous sheet materials, water vapor permeability is de®ned by Eq. (25); results are customarily expressed in units of gram-centimeters per square centimeter per second per centimeter of mercury ‰g cm=…cm2 s cmHg†Š or gram-centimeters per square meter per 24 h per millimeter of mercury ‰g cm=…m2 24 h mmHg†Š. On the other hand, paper is a nonhomogeneous material, and its water 

In Eq. (28), uav ˆ a cm3 =s ˆ a  10 10 3 † Pa.

3

m3 =s; A ˆ 50 cm2 ˆ 50  1004 m2 ; and
p ˆ b= …7:50

Porosity and Gas Permeability

295

vapor permeability varies markedly, depending on the test conditions. Therefore, the water vapor permeability of paper and paperboard is often expressed as the mass rate of water vapor transfer through a unit area of a specimen under the test conditions. This value (under a steady state) is called the water vapor transmission or the water vapor transmission rate and is expressed in grams per square meter per day. Methods to measure the water vapor permeability may be classi®ed into three types . Pressure gradient method .. Gravimetric method ... Hygrometric method The ®rst method was described earlier. In this case water vapor is used instead of an inert gas. Gravimetric Method Principle and Calculation The gravimetric method involves the permeation of water vapor across a constant area of a specimen under a speci®ed humidity gradient and the measurement of the amount of water vapor permeated under a steady state as the weight change in either a moisture absorbent or the moisture supply. Although many devices and modi®cations for practical measurement have been proposed [97,98], a simple method called the cup test, as described in TAPPI T 448, SCANP22, and elsewhere, will be covered here. This method requires several days for an experiment, but it enables simultaneous measurement for a number of samples. The water vapor transmission rate X can be calculated as X ˆ 240 where

a b

…30†

a = increase in mass per unit time, mg/h b = exposed area, cm2

Apparatus It is necessary for the gravimetric method to provide a light, shallow, nonpermeable test dish, such as a Vapometer cup, and a cabinet or room maintained at prescribed temperature and relative humidity. The humidity gradient is determined by a combination of water vapor pressure in the cabinet and the water vapor pressure of a material that either supplies moisture or absorbs it, placed in the test dish. Hence, the humidity gradient can be varied to a wide extent by putting a desiccant, a saturated salt solution, or distilled water into the test dish. Procedure 1. The specimen is placed on the top of the test dish containing the moistureabsorbing or -supplying material and is sealed by wax at the edge around the cup, leaving at test area with a diameter equal to that of the top of the cup. 2. The dish, assembled with contents and specimen, is placed in a cabinet maintained at prescribed conditions. After a steady state is achieved, the weight change of the assembled dish is recorded as a function of time. Note that the hygroscopicity of the test dish itself should be checked.

296

Yamauchi and Murakami

Hygrometric Methods

The following methods are considered here:

. Closed-cell method .. Sweep gas method ... Comparison method Each of the ®rst two methods involves measurement of the change in the concentration of water vapor at one side of a cell and the calculation of the average water vapor transfer rate from the concentration change as a function of the permeation time. These methods are often preferred for papers that have low water vapor permeability, such as water vapor barrier materials, and they provide rapid measurements. Although the theoretical relationships to convert the values from these dynamic methods into those from the standard cup method under a given steady state have not been established yet, calibration curves can be made by testing a series of specimens by both dynamic and gravimetric procedures. The third method is a steady-state process and requires the establishment of equilibrium. A value comparable to that from the standard cup method is obtained. In the case of composite materials, the measurement by each of the three methods must be made with the relatively impermeable surface of the specimen facing the low humidity side of the cell [26,87]. Closed-Cell Method The tester consists of two half-cells that are bolted together with a specimen between them. The down side of the cell is maintained at a prescribed high humidity by means of distilled water or a saturated salt solution. The specimen is conditioned by passing a stream of dry air through the upper side of the cell for a prescribed period, then the cell is closed. The humidity change in the upper side of the cell is measured by means of a hygrometer using an electrical resistance element [51] or an infrared instrument [36]. The average water vapor transfer rate X 0 can be calculated as X 0 ˆ 2:4 where


RHsat a b

…31†


RH

sat a b

= relative humidity change per unit time at the low humidity side of the cell, %/h = saturated water vapor density at temperature used = volume of low humidity side, cm3 =exposed area, cm2

Since the permeated water vapor accumulates in the low humidity side of the cell in this method, one might not expect to obtain an accurate result. However, in the case of a sample having relatively low water vapor permeability, the moisture content of the atmosphere in the low humidity side of the cell would be maintained at less than 0.5% RH during the measurement period [36]. Sweep Gas Method The specimen divides the test cell into two parts, the part on one side of the specimen being at low humidity and the other at high humidity. The method involves passing a dry gas stream at a slow rate through the low humidity side of the cell. The relative humidity difference between the ef¯uent gas and the in¯uent gas is measured by means of an electrical hygrometer [86] or an electrolytic cell [14,37,75]. The mean water vapor transfer rate can be calculated as

Porosity and Gas Permeability

297

X 0 ˆ a
RH sat b

…32†

where

a = a constant determined by the apparatus
RH = relative humidity difference between ef¯uent and in¯uent gases sat = saturated water vapor density at temperature used b = total volume of gas passed through low humidity side in unit time

Another measurement can be achieved by passing two gases at equal pressure simultaneously through the two sides of the cell at the same ¯ow rate. If two streams of nitrogen gas containing different concentrations of water vapor are used, the water vapor concentration of the ef¯uent gases can be measured by a calibrated thermal conductivity cell. An accurate measurement of water vapor permeation can be made by recording the ¯ow rate and pressure of the in¯uent gases, the water vapor concentration of the in¯uent and the ef¯uent gases, and the temperature [2]. This method is also generally applicable for gases and vapors other than water vapor. Note: In experimental work, it is particularly important that there should be (1) prevention of adsorption or condensation of water vapor on the cell walls, (2) application of exactly the same pressure on both sides of the cell, and (3) assurance that the gas composition in the cell is homogeneous. Comparison Method As in the previous case, the test cell is divided by the specimen into two halves. The lower half of the cell is maintained at a constant high humidity. The upper part of the cell is further divided into two chambers by a partition with a standardized humidity resistor (such as glass capillary tubing, ¯ared and covered at each end with thin permeable sheeting) that possesses a known resistance to water vapor transfer. The chamber on one side of the specimen is equipped with an electrical hygrometer; the chamber on the other side is the dry side and contains a desiccant. At equilibrium, the rate of water vapor transfer through the specimen is equal to that of water vapor passing through a standardized resistor. If the rate of water vapor transfer through the specimen is directly proportional to the relative humidity difference across the specimen, the water vapor transfer rate can be calculated as X 0 ˆ Xs 0 where


RH2
RH1

…33†

Xs 0 = water vapor transfer rate of standardized resistor
RH1 = relative humidity difference across specimen
RH2 = relative humidity difference across standardized resistor

Because paper and paperboard have a water vapor permeability that depends on relative humidity itself at one side of the material, the standardized resistor should be selected from a set of resistors covering an appropriate range to give a prescribed relative humidity in the middle chamber between the specimen and the resistor. A plastic ®lm having a known humidity±water vapor permeability relation is also used as a resistor [72].

298

Yamauchi and Murakami

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22.

Adam, N. K. (1941). Physics and Chemistry of Surfaces. Oxford Univ. Press, New York, p. 185. Ahien, A. T. (1970). Diffusion of sorbed water vapor through paper and cellulose ®lm. Tappi 53(7):1320±1326. ASTM Standard Test Method F316-86. Pore size characteristics of membrane ®lters by bubble point and mean ¯ow pore test. Banacki, W. J., Jr., and Bowers, R. L. (1962). Pore size testing of ®lter papers. Tappi 45(10):805±807. Bell, W. K. (1972). Mercury penetration and retraction hysteresis in closely packed spheres: A non-independent domain system. Ph.D. Thesis, Univ. Technology, Delft. Bhargava, R., Rogers, C. E., Stannett, V., and Szwarc, M. (1957). Studies in the gas and vapor permeability of plastic ®lms and coated papers, 4. Effect of a paper substrate. Tappi 40(7):564±567. Bliesner, W. C. (1964). A study of the porous structure of ®brous sheets using permeability techniques. Tappi 47(7):392±400. Brecht, W. (1962). Effect of structure on major aspects of paper behaviour with ¯uids. In: The Formation and Structure of Paper, Vol. 1. F. Bolam, ed. British Paper and Board Makers Assn. London, pp. 427±460. Brown, J. C. (1950). Determination of the exposed speci®c surface of pulp ®bers from air permeability measurements. Tappi 33(3):130±137. Browning, B. L. (1950). The speci®c surface of pulps by the silvering method. Tappi 33(8):410±412. Bublitz, W. J., Dappen, J. W., and Van den Akker, J. A. (1948). An investigation of the relationship between the permeabilities of paper to oil and air. Tech. Assn. Papers 31:305±315. Chesley, K. G., Jones, E. D., and Truax, R. L. (1959). Correlation of the Shef®eld porosimeter and the Gurley densometer. Tappi 42(4):299±300. Climpson, N. A., and Taylor, J. H. (1976). Pore size distributions and optical scattering coef®cients of clay structures. Tappi 59(7):89±92. Cole, L. G., Czuha, M., and Mosley, R. W. (1959). Continuous coulometric determination of parts per million of moisture in organic liquids. Anal. Chem. 31(12):2048±2050. Corte, H. (1955). Contribution to porosity analysis as exempli®ed by a study of cigarette ®lters. Das Paper 9(7):290±295 (in German). Corte, H. (1958). The porous structure of paper: Its measurement, its importance and its modi®cation by beating. In: Fundamentals of Papermaking Fibers. F. Bolam, ed. British Paper and Board Makers Assn., Kenley, England, pp. 301±331. Corte, H., and Kallmes, O. J. (1962). The interpretation of paper properties in terms of structure. In: The Formation and Structure of Paper, Vol. 1. F. Bolam, ed. British Paper and Board Makers Assn., London, pp. 351±368. Corte, H. K., and Lloyd, H. (1966). Fluid ¯ow through paper and sheet structure. In: Consolidation of the Paper Web, Vol. 2. F. Bolam, ed. British Paper and Board Makers Assn., London, pp. 981±1009. Davis, P. J., and Smith, D. M. (1989). Using NMR spectroscopy to analyze the pore structure of coatings. Tappi J. 85(5):85±89. Dehoff, R. T., and Rhines, F. N., eds. (1968). Quantitative Microscopy. McGraw-Hill, New York. Drake, L. C., and Ritter, H. L. (1945). Macropore-size distributions in some typical porous substances. Ind. Eng. Chem., Anal. Ed. 17(12):787±791. Dullien, F. A. L. (1979). Porous Media, Fluid Transport and Pore Structure. Academic Press, New York.

Porosity and Gas Permeability

299

23. Dullien, F. A. L., and Dhawan, G. K. (1974). Characterization of pore structure by a combination of quantitative photomicrography and mercury porosimetry. J. Coll. Interface Sci. 47(2):337±349. 24. Fu, Y., and Bartell, F. E. (1951). Surface area of porous adsorbants. J. Phys. Coll. Chem. 55(5):662±675. 25. Furo, I., and Daicic, J. (1999). NMR cryoporometry: A novel method for the investigation of the pore structure of paper and paper coatings. Nordic Pulp Paper Res. J. 14(3):221±225. 26. Gaibraith, A. D., and Kitchen, E. A. (1962). Permeability theory vs. practice in packaging liquid products. Tappi 45(2):173±176. 27. Garey, C. L., Leekley, R. M., Hultman, J. D., and Nagel, S. C. (1973). Determination of pore size distribution of pigment coatings. Tappi 56(11): 134±138 28. Gregg, S. J., and Sing, K. S. W. (1967). Adsorption, Surface Area and Porosity. Academic, London. 29. Harkins, W. D., and Jura, G. (1944). Surfaces of solids, 13. A vapor adsorption method for the determination of the area of a solid without the assumption of a molecular area, and the area occupied by nitrogen and other molecules on the surface of a solid. J. Am. Chem. Soc. 66(8):1366±1373. 30. Haselton, W. R. (1954). Gas adsorption by wood, pulp, and papers, 1. The low-temperature adsorption of nitrogen, butane, and carbon dioxide by sprucewood and its components. Tappi 37(9):404±412. 31. Haselton, W. R. (1955). Gas adsorption by wood, pulp, and papers, 2. The application of gas adsorption techniques to the study of the area and structure of pulps and the unbonded and bonded area of paper. Tappi 38(12):716±723. 32. Hermans, P. H. (1949). Physics and Chemistry of Cellulose Fibers. Elsevier, New York. 33. Hermans, P. H., Heikens, D., and Weidinger, A. (1959). A quantitative investigation on the X-ray small angle scattering of cellulose ®bers, 2. The scattering power of various cellulose ®bers. J. Polym. Sci. 35(128):145±165. 34. Higgins, H. G., and De Yong, J. (1966). Visco-elasticity and consolidation of the ®ber network during free water drainage. In: Consolidation of the Paper Web, Vol. 1. F. Bolam, ed. British Paper and Board Makers Assn., London, pp. 242±267. 35. Horii, T., and Chino, T. (1960). Trial manufacture of the smoothness tester applied vacuum air micrometer, 2. Jpn. Tappi 14(110):312±321, 327 (in Japanese). 36. Husband, R. M., and Petter, P. J. (1966). An infrared instrument for the rapid measurement of water vapor permeation through barrier webs. Tappi 49(12):565±572. 37. Keidel, F. A. (1959). Determination of water by direct amperometric measurement. Anal. Chem. 31(12):2043±2048. 38. Kruyer, S. (1958). The penetration of mercury and capillary condensation in packed spheres. Trans. Faraday Soc. 54:1758±1767. 39. Labrecque, R. P. (1968). The effects of ®ber cross-sectional shape on the resistance to the ¯ow of ¯uids through ®ber mats. Tappi 51(1):8±15. 40. Maloney, T. C. (1999). Thermoporosimetry by isothermal step melting. Proc. PreSymposium of the 10th ISWPC, Seoul, Korea, pp. 245±253. 41. Maloney, T. C., Li, T. Q., Weise, U., and Paulapuro, H. (1997). Intra- and inter®ber pore closure in wet pressing. Appita J. 50(4):301±306. 42. Mayer, R. P., and Stowe, R. A. (1965). Mercury porosimetryÐbreakthrough pressure for penetration between packed spheres J. Colloid Sci. 20:893±911. 43. Mayer, R. P., and Stowe, R. A. (1966). Mercury porosimetry: Filling of toroidal void volume following breakthrough between packed spheres J. Phys. Chem. 70(12):3867± 3873. 44. McKnight, T. S., Marchessault, R. H., and Mason, S. G. (1958). The distribution of pore sizes in wood-pulp ®bers and paper. Pulp Paper Mag. Can. 59(2):81±88.

300

Yamauchi and Murakami

45. Murakami, K., and Yamauchi, T. (1981). Porous structure of pulp sheets evaluated by mercury intrusion method. Preprints of 1981 Papermaking Conference, Tokyo, pp. 97±103. 46. Nelsen, F. M., and Eggertsen, F. T. (1958). Determination of surface area, adsorption measurements by a continuous ¯ow method. Anal. Chem. 30(8):1387±1390. 47. Onogi, Y., Yamauchi, T., Murakami, K., and Imamura, R. (1974). The porous structure of paper: An approach from mercury porosimetry. Jpn. Tappi 28(3):99±107 (in Japanese). 48. Peterson, R. M. (1970). Two-dimensional ¯ow of incompressible ¯uids through deformable porous media. Tappi 53(1):71±77. 49. Quynn, R. C. (1963). Internal volume in ®bers. Textile Res. J. 33(1);21±34. 50. Rance, H. F. (1958). The porous structure of paper. In: The Structure and Properties of Porous Materials. D. H. Everett and F. S. Stone, eds. Butterworths, London, pp. 302± 321. 51. Ranger, H. O., and Gluckman, M. J. (1964). A faster WVT method. Modern Packaging 37(11):153±156, 202. 52. Rasi, M., Koponen, A., Aaltosalmi, U., Timonen, J., Kataja, M., and Niskanen, K. (1999). Permeability of paper: Experiments and numerical simulations. Proc. 1999 Int. Paper Physics Conf., San Diego, pp. 297±306. 53. Rennel, J. (1969). Opacity in relation to strength properties of pulps, 1. Method for producing unbonded ®bers and determining their light-scattering coef®cient and surface area. Svensk Papperstidn. 72(1):1±8. 54. Ritter, H. L., and Drake, L. C. (1945). Pore-size distribution in porous materials, pressure porosimeter and determination of complete macropore-size distributions. Ind. Eng. Chem., Anal. Ed. 17(12):783±786. 55. Rogers, C. E., Meyer, J. A., Stannett, V., and Szwarc, M. (1956). Studies in the gas vapor permeability of plastic ®lms and coated papers, 1. Determination of the permeability constant. Tappi 39(11):737±741. 56. Rootare, H. M., and Nyce, A. C. (1971). The use of porosimetry in the measurement of pore size distribution in porous materials. Int. J. Powder Met. 7:3±11. 57. Rootare, H. M., and Prenzlow, C. F. (1967). Surface area from mercury porosimeter measurements. J. Phys. Chem. 71(8):2733±2736. 58. Rootare, H. M., and Spencer, J. (1971). A computer program for pore volume and pore area distribution calculations from mercury porosimeter data on particulate or porous materials. Powder Technol. 6:17±23. 59. Rounsley, R. R. (1964). Vapor transport through paper. Tappi 47(2):95±98. 60. Scheidegger, A. E. (1962). On the validity of the Kozeny equation. In: The Formation and Structure of Paper, Vol. 2. F. Bolam, ed. British Paper and Board Makers Assn., London, pp. 829±831. 61. Scheidegger, A. E. (1974). The Physics of Flow Through Porous Media. 3rd ed. Univ. Toronto Press. 62. Schumb, W. C., and Rittner, E. S. (1943). A helium densitometer for use with powdered materials. J. Am. Chem. Soc. 65(9):1692±1695. 63. Setterholm, V. C. (1974). A new concept in paper thickness measurement. Tappi 57(3):164. 64. Stamm, A. J. (1956). Diffusion of water into uncoated cellophane. 2. From steady state diffusion measurements. J. Phys. Chem. 60(1):83±86. 65. Stamm, A. J., and Millett, M. A. (1941). The internal surface of cellulosic materials. J. Phys. Chem. 45(1):43±54. 66. Stone, J. E., and Nickerson, L. F. (1963). A dynamic nitrogen adsorption method for surface area measurements of paper. Pulp Paper Mag. Can. 64(3):T155±T161. 67. Stone, J. E., and Scallan, A. M. (1965). A study of cell wall structure by nitrogen adsorption. Pulp Paper Mag. Can. 66(8):T407±T414.

Porosity and Gas Permeability

301

68. Stone, J. E., Scallan, A. M., and Aberson, M. A. (1966). The wall density of native cellulose ®bers. Pulp Paper Mag. Can. 67(5):T263±T268. 69. Stone, J. E., Scallan, A. M., and Abrahamson, B. (1968). In¯uence of beating on cell wall swelling and internal ®brillation. Svensk Papperstidn. 71(10):687±694. 70. Soyama, T. (1997). A requirement on air ®lter media for semiconductor industry in Japan and high performance air ®lter media. Proc. Tappi 1997 Nonwovens Conference, pp. 151±155. 71. Svata, M. (1971). Determination of pore size and shape distribution from porosimetric hysteresis curves. Powder Technol. 5:345±349. 72. Takahashi, F., and Naito, Y. (1965). An electric hygrometer method of test for water vapor transmission rate of coated paper. Japan Tappi 19(5):245±250 (in Japanese). 73. Thode, E. F., Swanson, J. W., and Becher, J. J. (1958). Nitrogen adsorption on solventexchanged wood cellulose ®bers, indications of ``total'' surface area and pore size distribution. J. Phys. Chem. 62:1036±1039. 74. Tollenaar, D. (1967). Capillarity and wetting in paper structures: Properties of porous systems. In: Surface and Coatings Related to Paper and Wood. R. H. Marchessault and C. Skaar, eds. Syracuse Univ. Press, pp. 195±219. 75. Toren, P. E. (1965). Application of electrolytic moisture meter to measurement of water vapor transmission through plastic ®lms. Anal. Chem. 37(7):922±923. 76. Uchimura, H., and Kimura, M. (1998). A sample preparation for paper cross-sections using a focused ion beam. Sen-i Gakkaishi 54(7):360±366 (in Japanese). 77. Underwood, E. E. (1970). Quantitative Stereology. Addison-Wesley, Boston. 78. Vollmer, W. (1954). Transport of gas and vapor in paper. Chem. Ing. Tech. 26(2):90±94 (in German). 79. Warashina, Y., Takeshita, H., and Take, Y. (1962). Studies on the measurement method of porosity of paper. Japan Tappi 16(137):589±600 (in Japanese). 80. Washburn, E. W. (1921). Note on a method of determining the distribution of pore sizes in a porous material. Proc. Natl. Acad. Sci. USA 7:115±116. 81. Wasser, R. B. (1974). A mercury buoyancy technique for determining apparent density and thickness of paper. Tappi 57(3):166. 82. Weatherwax, R. C., and Tarkow, H. (1968). Cell wall density of dry wood. Forest Products J. 18(2):83±85. 83. White, R. E., and Marceau, W. E. (1962). The capillary behavior of paper. Tappi 45(4):279±284. 84. Wilfong, J. G. (1966). Speci®c gravity of wood substance. Forest Products J. 16(1): 55±61. 85. Wink, W. A., and Baum, G. A. (1983). Rubber platen caliper gauge: New concept in measuring paper thickness. Tappi J. 68(9):131±133. 86. Wink, W. A., and Dearth, L. R. (1949). Measurement of water-vapor permeability at low temperature: The hygrometric sweep gas method. Tappi 32(3):232±238. 87. Winrich, K. M. (1971). Dynamic WVTR testing, an experience report on the St. RegisHoneywell WVTR tester. Tappi 54(8):1302±1304. 88. Yamauchi, T. (1987). Measurement of paper thickness and density. Appita J. 40(5):359± 369. 89. Yamauchi, T. (1994). Structural characterization of the coating layer using mercury porosimetry. Sen-i Gakkaishi 50(12):615±617. 90. Yamauchi, T. (1999). Pore structure of paper materials. In: Porous Materials: Characterization, Production and Application. Y. Takeuchi, ed. Fuji Technosystem, Tokyo, pp. 223±234 (in Japanese). 91. Yamauchi, T., and Kibblewhite, R. P. (1988). Pore structure of paper webs from Radiata pine thermomechanical pulps. Appita J. 41(1):37±42.

302

Yamauchi and Murakami

92. Yamauchi, T., Murakami, K., and Imamura, R. (1975). The porous structure of paper, analytical research on the behavior of mercury penetration and retraction. Jpn. Tappi 29(9):492±497 (in Japanese). 93. Yamauchi, T., Murakami, K., and Imamura, R. (1976). The air permeability of paper related to the porous structure. Jpn. Tappi 30(5):273±280 (in Japanese). 94. Yamauchi, T., Murakami, K., and Imamura, R. (1978). Studies on latex impregnated paper, 1. The porous structure and rubber distribution of latex impregnated handsheets. Jpn. Tappi 32(9):434±540 (in Japanese). 95. Yamauchi, T., Murakami, K., and Imamura, R. (1979). Development of the pore structure of paper web during consolidation. Mokuzai Gakkaishi 25(6):414±421. 96. Young, D. M., and Crowell, A. D. (1962). Physical Adsorption of Gases. Butterworths, London. 97. Ziegler, R. D. (1957). A procedure for measuring the water vapor permeability of insulation board. Tappi 40(11):881±884. 98. Ziegler, R. D. (1960). The effect of temperature and moisture content on the moisture permeability of building boards. Tappi 43(11):913±918.

7 WETTING AND THE PENETRATION OF LIQUIDS INTO PAPER M. BRUCE LYNE International Paper Tuxedo, New York

I. Introduction II. Theory of Dynamic Wetting, Spreading, and Capillary Penetration

303 305

III. Test Methods A. Wetting B. Penetration of Liquids C. Total Uptake of Aqueous Liquids D. Pressure Penetration

314 314 321 325 326

IV. Concluding Remarks

329

References

I.

330

INTRODUCTION

Wetting is a truly surface phenomenon; therefore, it is unsatisfactory to view paper as simply being composed of cellulose, hemicelluloses, and lignin, because the chemical composition of surface layers down to monomolecular thicknesses determines the wetting characteristics. Low surface energy resin and fatty acids present in wood and surface-active agents added during pulping and papermaking can have a profound effect on wetting and sorption, particularly with respect to interactions with polar liquids such as water and aqueous solutions, emulsions, and dispersions [9,45,46].

303

304

Lyne

Mechanically prepared wood pulps, such as those used in newsprint and magazine grades, have a relatively heterogeneous surface chemistry. In mechanical pulping, ®bers are liberated from wood by physical degradation of middle lamellae between the ®bers. In order to increase opacity and promote inter®ber bonding in paper, the surface area of the ®bers is further ``developed'' by mechanically peeling the outer layers of the ®ber wall. Lignin is more concentrated in the outer layers of the ®ber, and because lignin is more hydrophobic than either cellulose [23,25] or hemicelluloses, surface chemistry varies locally with the degree of surface development and ®nes concentration. Resins, which are originally contained in ray cells in softwoods, are liberated during mechanical pulping and spread readily when heated to cover the surfaces of ®bers and reduce surface free energy, a phenomenon referred to as self-sizing. Surface morphology also plays a role in wetting. Drops of nonwetting liquids tend to exhibit higher contact angles on rough surfaces and tend to extend more readily along grooves or ®bers than across them [37]. Conversely, liquids that wet a surface will exhibit a lower contact angle on a rougher surface of the same material. This can be explained by the tendency of all systems to move in the direction of reducing surface free energy. Thus, when a liquid wets a solid surface it has the lower surface free energy and will move to cover all of the rough surface, whereas a nonwetting liquid will move to expose the rough surface, which has a lower surface free energy. Penetration of liquids is also interrupted by dislocations in local surfaces such as ridges on ®bers or the sharp edges of mineral ®llers such as kaolin and calcium carbonate in paper [22]. Therefore, even paper made from chemically prepared wood pulps having relatively homogeneous ®ber surface chemistries will exhibit local nonuniformities and anisotropy in wetting behavior due to variation and orientation in ®ber and network morphology. Sizing agents such as alkyl ketene dimer (AKD) do not change the overall surface free energy of paper because they cover less than 20% of the ®ber surface. Local chemical discontinuities caused by AKD deposition disrupt the spread of aqueous liquids and their penetration of inter®ber voids but do not prevent the ultimate penetration of the ®ber wall and resultant ®ber swelling. In fact, morphological and surface chemistry discontinuities can have a much greater in¯uence on the rate of penetration of liquids into paper than does pore size distribution [40]. The penetration of aqueous liquids into paper is further complicated by absorption into ®ber walls and consequent increases in ®ber wall thickness. It is thought that swelling occurs as aqueous liquids break and replace interchain hydrogen bonds in cellulose. Swelling appears to be proportional to the amount of liquid absorbed [8,19,48], and swelling of the ®ber walls generally tends to close voids in the ®ber surfaces while enlarging inter®ber voids in the ®ber network. Thus the rate of capillary imbibition generally increases with time for the absorption of aqueous liquids [26]. Techniques for measuring wetting and the penetration of aqueous liquids into paper have traditionally not separated the surface chemistry, morphological, and swelling aspectsÐnor have they attempted to replicate the highly dynamic conditions that pertain in most industrial processes where wetting and penetration are important. Although this chapter does deal with the more traditional techniques, such as the Cobb test [44], contact angle measurement [42], and drop penetration tests [43], it concentrates on more modern and informative techniques.

Wetting and Penetration of Liquids into Paper

II.

305

THEORY OF DYNAMIC WETTING, SPREADING, AND CAPILLARY PENETRATION

Liquids will wet and spread over the surface of a solid if the total surface free energy is thereby reduced: S ˆ S

LS

L

…1†

That is, spreading will occur when S is positive and the surface free energy of the solid S exceeds that of the liquid L plus the interfacial tension between the liquid and the solid LS . Spreading can also be described as occurring when the work of adhesion between the liquid and solid WA…LS† exceeds the work of cohesion in the liquid WC…L† : WA…LS† ˆ S ‡ L

LS

…2†

WC…L† ˆ 2 L S ˆ WA…LS†

…3† WC…L† ˆ S

LS

L

…4†

Because spreading occurs only when S is positive, any process that suf®ciently reduces the surface free energy of the solid S will retard or prevent wetting and spreading. As discussed above, wetting can be viewed as an interfacial adhesion phenomenon between a liquid and a solid surface. There are two basic types of intermolecular interactions that contribute to wetting and adhesion: dispersive and acid±base. Dispersive forces are weak attractions between ¯uctuating dipoles that operate at relatively long range (in excess of 100 AÊ). These forces, which are referred to as London±van der Waals or Lifshitz±van der Waals forces, are apolar interactions during which the electromagnetic ®eld created by random ¯uctuations in the dipole moment of a molecule induce a corresponding ¯uctuation in a neighboring molecule and attraction across the liquid/solid phase boundary. In contrast, acid±base interactions involve ®xed overlap of electron densities, or electron donor±acceptor exchange between two molecules on opposite sides of the phase boundary. Acid±base interactions are much stronger than dispersive forces, but they act over shorter ranges (as little as 5 AÊ) and thus require intimate conformation of the liquid to the solid surface. Wetting is generally studied under equilibrium conditions and reported as a static interfacial contact angle. However, liquids that are forced into contact with solids under pressure will wet and adhere even though nonwetting occurs under equilibrium conditions. This is important in industrial processes such as printing, where critical liquid transfer processes occur in nips under pressure. As shown in Fig. 1, the transfer curves for mineral-oil-based ink to polytetra¯uoroethylene (Te¯on), polypropylene, and polyester are the same although the surface free energies are 18, 35, and 42 mN/m, respectively. In fact, mineral oil forms a high contact angle with Te¯on under equilibrium conditions, and the ink beads up on the Te¯on after the printing nip. In general, when pressure is applied to force a liquid into intimate contact with a solid surface, dispersive interactions are enhanced and wetting may occur when it is not energetically favorable under equilibrium conditions.

306

Lyne

Fig. 1 Fraction of oil-based ink transferred from printing plate to polymer sheets in a GFL rotary printing press. (From Ref. 28.)

Substituting the Young±Dupre contact angle equation in Eq. (2) and assuming negligible equilibrium spreading pressure, the work of adhesion can also be expressed as WA ˆ LV …1 ‡ cos †

…5†

Recently, it has become common to partition the work of adhesion into dispersive and acid±base interactions as proposed by van Oss [49]:  1=2  1=2  1=2  d d a b b a …6† W A ˆ 2 S L ‡ S L ‡ S L in which the superscripts d, a, and b denote contributions to surface free energy from dispersive forces, acidity (electron acceptor), and basicity (electron donor), respectively. Thus, knowing the dispersive, acid, and base components of the surface free energy of the solid and liquid it is possible to calculate the dispersive and acid±base components of their adhesive interaction. If probe liquids having known dispersive, acid, and base components are applied to a solid and the contact angles are determined, it is possible to calculate the dispersive, acid, and base characteristics of the solid. When cellulosic or lignin-containing ®lms absorb moisture, their surfaces progressively approach the more polar and higher surface free energy of a surface water layer [23,25]. Conversely, when a surfactant is added to water it becomes less polar and exhibits a lower surface energy. Both tend to facilitate wetting and increase spreading. The effects of dispersive and acid±base interactions on wetting can be illustrated by comparing the way in which water and cyclohexane wet polyethylene and glass. Because polyethylene has no acid or base component, interactions with water and cyclohexane are purely dispersive. Because the dispersive component for

Wetting and Penetration of Liquids into Paper

307

cyclohexane is greater than that for water (25.5 and 21.8 mN/m, respectively), cyclohexane wets polyethylene and water does not. Conversely, glass, like water, has pronounced acid±base characteristics in addition to a dispersive component. It follows that the overall work of adhesion between water and glass is greater than the work of adhesion between a purely dispersive liquid such as cyclohexane and glass. Thus, water wets glass more aggressively. This can be further appreciated in Table 1, where the static contact angles between water and glass and between water and various polymers can be compared with the dispersive and acid±base components of the work of adhesion between water and these surfaces. As a general rule of thumb, the total work of adhesion between a solid and a liquid must exceed 100 mN/m for wetting to occur under equilibrium conditions. However, the effect on absorption and capillary imbibition of increasing the moisture content of cellulose and adding surfactants to water is not so simple. Consider a liquid in a capillary that wets the surface of the capillary walls. The liquid spreads parallel to the surface of the capillary walls, but the surface of the bulk liquid lies perpendicular to the capillary walls. Therefore, the surface of the liquid must have a curvature. This curvature, or meniscus, is a re¯ection of the capillary pressure exerted by the liquid spreading of the pore walls (see Fig. 2). The Laplace equation describes this capillary pressure P in terms of the contact angle  between the liquid and the capillary walls, the surface free energy of the liquid L , and the radius of the capillary r: P ˆ 2 L

cos  r

…7†

This capillary pressure results in a net ¯ow of liquid along the capillary. In the case of a horizontal capillary, once the ¯ow has been initiated the rate of penetration dh=dt can be described by the Lucas±Washburn equation dh r2 p r L cos  ˆ ˆ dt 8h 4h

…8†

Table 1 Hydrophilicity and Acid±Base Interactionsa

Substrate Glass Cellophane 15% RH 50% RH 95% RH PMMA Polyethylene Polypropylene Te¯on a

YSd (mN/m)

Contact angle with water (deg)

WA (mJ/m2 )

WAd (mJ/m2 )

WAab (mJ/m2 )

76.0

0

145.6

81.4

64.2

50.2 48.6 41.6 36.1 33.1 28.4 19.0

68 28 5 74 105 108 116

100.1 137.1 145.3 92.9 54.0 50.3 40.9

66.2 65.1 60.2 56.1 53.7 49.8 40.7

33.9 72.0 85.1 36.8 0.7 0.5 0.2

For water: YLV ˆ YLd …21:8† ‡ YLab …51:0† ˆ 72:8 mN=m.

308

Fig. 2

Lyne

Parallel-side capillary showing contact angle  with rising liquid.

where h = length of ®lled portion of capillary  = viscosity of the liquid If a surfactant is added to an aqueous liquid, both L and  will decrease. The penetration rate dh=dt will be altered according to whether L cos  is increased or decreased. For example, if the addition of a surfactant reduces the contact angle of water to zero and reduces the surface free energy from 72 to 30 dyn/cm, the penetration rate will increase only if 30 cos 0 > 72 cos w or the contact angle for water w is greater than 65 . The contact angle with the pore walls is a function not only of the dispersive and acid±base chemistry of the liquid and solid surfaces but also of pore wall geometry. As shown in Fig. 3 [22], the angle of convergence of the pore walls should be

Fig. 3 Capillary penetration into a conical pore having a convergence angle ! where the contact angle with the capillary walls equals 90 . (From Ref. 22.)

Wetting and Penetration of Liquids into Paper

309

subtracted from the contact angle of the liquid with a ¯at surface of the same material when calculating capillary suction pressure. Thus, a liquid that exhibits a contact angle greater than 90 with a ¯at ®lm of cellulose can still penetrate the surface pores of a cellulose ®ber network, because these pores are generally convergent and can be modeled as truncated cones [12]. This explains why water, which exhibits an initial contact angle with sized paper that exceeds 90 , penetrates the surface pores in a Bristow test [29] without any measurable delay. Marmur [31] also pointed out that in realistic sorption situations the ¯uid reservoir is often limited and has a curvature of its own. If a discontinuous ®lm of liquid is applied to paper, as in a printing situation, the curvature of the liquid reservoir above the pore also exerts a pressure that speeds capillary penetration. Danino and Marmur [13] studied the penetration of nonpolar liquids radially into paper from both ®nite drops and unlimited liquid reservoirs. They concluded that penetration is much slower in the former case because of the redistribution of liquid from larger to smaller pores in the paper once the drop has been absorbed, whereas the liquid continues to rapidly penetrate out from the in®nite reservoir through the larger pores in paper. With irregularly shaped pore walls, capillary penetration can be observed to accelerate in converging portions of the pore and decelerate in diverging portions. As shown in Fig. 4, this is a re¯ection of the change in capillary pressure as calculated by inserting the local contact angle between the liquid front and the pore walls in the Laplace equation. A further complexity in capillary penetration can also be observed at sharp dislocations in the pore wall such as might be encountered at the edge of the mineral ®ller particle. The liquid can be observed to stick at this point and hinge around the dislocation. As pictured in Fig. 5, this behavior is governed by the Gibbs inequality [18] and represents a suspension of the Young±Dupre contact angle equa-

Fig. 4 Acceleration of liquid into a converging pore, jump in velocity at a dislocation of less than 180 , and stick at sharp dislocations greater than 180 . (From Ref. 22.)

310

Fig. 5

Lyne

Hinging behavior of a liquid at a sharp discontinuity. (From Ref. 22.)

tion. If and when the liquid reaches the second surface, capillary suction and penetration resume. In mineral coatings and ®lled papers the frequency of these dislocations and divergent pore geometries are more important than pore size in determining penetration rate. This is illustrated in Table 2 for ®ve pigments coated on millipore ®lters with a starch binder. Note that the penetration rate for oil in the Bristow test is much slower for the very porous ground calcium carbonate than for the large-plate mica. Presumably, this is due to the frequency of diverging pore sections in the gravel-like carbonate and the frequency of convergent pores in the mica. Because the samples were uncalendered, disorder in plate alignment results in the pores between the plates having a wide end and a narrow end. Oil penetrates from the wide end to the narrow end of these pores. As legislation forces the removal of volatile organic liquids from inks, fountain solution, and varnishes, new problems arise from unwanted sorption and spreading behavior. For example, as shown in Fig. 6, isopropanol was very effective in speeding wetting and absorption of fountain solution into paper. Although surfactants can also reduce the surface tension of fountain solution, they can take more than a second to migrate to the surface of a newly split ®lm of fountain solution such as would be found on the outgoing side of a printing nip. As shown in Fig. 7, the equilibrium surface tension of a surfactant solution and a solution of isopropanol in water can be the same, but the dynamic surface tension of the surfactant solution can remain the same as that of water for such a long time that little effect is seen on the speed of penetration in the Bristow test. Papers that exhibit a slow rate of absorption for fountain solution may be problematic for the transfer of oil-based ink in subsequent printing nips.

Shape Aspect ratio Mean E.S.D. (m) Bristow ABS rate [mL=…m2 s1=2 †Š Hg intrusion pore diameter (m)

Table 2

Platey 10 1 6 0.4

Fine kaolin Platey 15 2 16 0.6

Std. kaolin Platey 25 10 190 0.6

Mica

Mineral

Amorphous 1 1 170 0.6

Precipitated silicoaluminate

Blocky 1 2 30 2.5

Ground calcium carbonate

Wetting and Penetration of Liquids into Paper 311

312

Lyne

Fig. 6 Sorption curves for water and for water with increasing amounts of isopropyl alcohol added. Surface tension shown on each curve in milliNewtons per meter. (From Ref. 1.)

Fig. 7 Absorption curves for water, isopropyl alcohol in water, and a surfactant solution having the same equilibrium surface tension. The dynamic surface age of 3 ms is also shown. (From Ref. 1.)

Wetting and Penetration of Liquids into Paper

313

Further, when alcohol evaporates from a water-based ink or fountain solution, the remaining liquid increases in surface tension. Conversely, when water evaporates from an ink or fountain solution containing a surfactant, the surface tension of the remaining liquid is decreased as the surfactant becomes more concentrated. If an uneven layer of ink or fountain solution is transferred to a paper or board, the thin areas will become higher in surface tension for the case of evaporating alcohol, but the thin areas will decrease in surface tension with the evaporation of water from a surfactant solution. Thus, the thick areas move toward thin areas for the alcohol solution, but thin areas try to spread out in the case of surfactant-based systems. The net result can be a worsening of the nonuniformity in thickness of a surfactant solution. This is particularly grievous for water-based inks because perceived mottle will increase with migration from the thin areas. Flow of this sort caused by gradients in surface tension is a manifestation of the Marangoni effect. This is distinct from diffusion caused by concentration gradients and is simply an example of migration that leads to minimization of the total surface free energy of the system. Liquids penetrate dense and sized papers more by diffusion than by capillary imbibition [19,23]. For the diffusion of aqueous liquids, swelling of ®bers as the result of the absorption of moisture will cause expansion of the network. For porous papers the penetration rate will increase because the inter®ber capillary radii will increase with the attendant expansion of the ®ber network. For dense papers, diffusion of aqueous liquids can be monitored as a local increase in the thickness of the paper. Fick's second law describes the change in concentration of a diffusing liquid in time t at any point x along the direction of diffusion as a function of a diffusion coef®cient D: @c @2 c ˆD 2 @t @x

…9†

Hoyland [19] proposed that the following solution of Fick's law should describe the penetration of aqueous liquids into paper by diffusion:
Z m ˆ t
Zmax m1 where
Z=
Zmax m0 ; mt ; m1 t L

  m0 2 Dt ˆ m0 L 

…10†

= change in thickness of paper as aqueous liquid diffuses into ®ber walls relative to maximum thickness change at ®ber saturation = moisture absorbed at times 0, t, and ®ber saturation, respectively = time elapsed from introduction of liquid = initial thickness of paper

Thus, the diffusion coef®cient D can be calculated by measuring the thickness change as a function of time after introducing the aqueous liquid to the paper surface. Also, the change in the diffusion coef®cient with depth of diffusion into the paper web can be studied with Hoyland's apparatus [19,20].

314

III.

Lyne

TEST METHODS

From the foregoing, it is apparent that wetting and sorption of liquids depend on the initial moisture content of paper. Temperature also affects chemical reactivity, viscosity, contact angle [4], and paper rheology. Therefore it is important that all tests be made on conditioned paper samples in an atmosphere at 50% relative humidity at 23 C (ISO standard atmosphere). When testing with aqueous liquids, minor impurities in the water used can have considerable in¯uence on sorption. It should not be expected that tests made with tap water can be replicated, because ionic concentration and pH can vary considerably. Where process conditions require tests with tap water, the amount and character of dissolved ionic material and the pH should be recorded or, if possible, controlled. Bristow [6] showed that for kraft liner, the absorption rate is strongly in¯uenced by the pH of the aqueous liquid and by the pH of the paper. More rigorous testing may also require deionized or distilled water; in some cases, the water must be deaerated to ensure reproducibility. Finally, test apparatus should be made of nonreactive materials such as stainless steel, glass, or gold-plated brass. Every effort should be made to keep all test surfaces free from contamination, particularly by liquids with low surface free energy such as oils and silicones. A.

Wetting

Contact Angle Measurement Contact angle measurements between sessile drops and the surface of paper have long been used as an indication of wettability. As shown in Fig. 8, the contact angle  for a drop of aqueous liquid increases as the paper surface becomes more hydrophobic and, conversely, decreases as the surface becomes more wettable. Methodology for measuring static contact angles is described in Ref. 42. Essentially, the method consists of depositing a drop of liquid from a small hypodermic need (1 mL) onto the paper surface and measuring the static contact angle after 5 s with the aid of a microscope or projection apparatus. Because the contact angle generally decreases as wetting and absorption proceed, a system capable of recording and estimating the contact angle at 10 ms intervals such as the STFI dynamic absorption tester can be useful for determining contact angles on industrial relevant time scales (Fig. 9). The dependence of dynamic contact angle on drop volume can also tested with the instrument. Finally, judgement of contact angles between rough, heterogeneous surfaces such as paper and a sessile drop is subjective and tedious. Having the experimental factors controlled by an instrument and having criteria for determining the contact angle ®xed by the image analysis program greatly improve reproducibility of contact angle measurement [43]. If wetting force measurements are performed on single ®bers using a capillograph (Fig. 10) [2,36], then advancing and receding contact angles, a and r , can be calculated from the wetting force with the liquid advancing F…a†, receding F…r†, and leaving the tip of the ®ber. The last gives the maximum force F…m† at near zero contact angle (neglecting any buoyancy forces): F…m† ˆ L P

…11†

Wetting and Penetration of Liquids into Paper

Fig. 8

Contact angle between a sessile drop and a ¯at surface. (From Ref. 42.)

Fig. 9

STFI dynamic absorption tester. (From Ref. 43.)

315

316

Lyne

Fig. 10

Schematic of capillograph.

where P is the ®ber perimeter. Also, F…a† ˆ F…m† cos a

…12†

F…r† ˆ F…m† cos r

…13†

Miller et al. [33] introduced an elegant method of determining pore size distribution that can also be used to determine advancing and receding contact angles. As shown in Fig. 11, the paper sample is suspended on a sintered membrane and is fully saturated with the test liquid before the test begins. Air pressure is then applied in increments to the sealed chamber, driving the liquid from progressively smaller pores in the paper. By measuring the amount of liquid displaced with each pressure step it is possible to build up a pore size distribution diagram using the Washburn equation. Further, it is also possible to release the pressure incrementally and let the liquid back into the sample in a stepwise fashion. Advancing contact angles can be measured during the intrusion of a liquid, and receding contact angles measured during extrusion. Hysteresis for the intrusion and extrusion of the liquid can be used to deduce the distribution of bottleneck and transverse pores that are frequent in paper. Also, by making evaluations with water and then with a nonswelling liquid such as

Fig. 11 Porosimeter in which a saturated paper sample is supported on a porous membrane and applied air pressure causes extrusion of liquid measured by the balance at the right. (From Ref. 33.)

Wetting and Penetration of Liquids into Paper

317

hexadecane, the effect of swelling on the pore size distribution can be determined (see Fig. 12). This appears to be a valuable approach for analyzing the dynamic contact angle and the pore structure of paper down to a pore radius of about 1 m. At the lower end of the sensitivity range, air diffusion through the support membrane becomes problematic, but research is under way to extend the useful range of the instrument. Dynamic Wetting Instruments Useful information about wetting and absorption can be gained without reference to contact angles or effective capillary radii. Hawkes and Bedford [17] and later Bristow [5] designed instruments to study dynamic penetration and wetting of paper. Figure 13 shows Bristow's instrument. A paper strip is mounted on a rotating wheel and is drawn past a miniature headbox that is deadweighted so that it rests on the paper strip with a pressure of 0.1 MPa. The headbox is ®lled with a known amount of liquid that has time available for absorption determined by the peripheral speed V and the width of the slice opening (l ˆ 1 mm). Because this speed can be varied, the relationship between the amount of liquid transferred and the time available for absorption can be established as in Fig. 14. The amount of liquid that transfers to the paper in time t can be described by Kr ‡ Ka t1=2

…14†

where Kr = roughness index, or amount of liquid that theoretically would ®ll the surface of the paper at time zero Ka = absorption coef®cient or slope of the curve

Fig. 12 Distribution of pore throats in nonwoven sample showing the effect of swelling of ®bers on pore sizes. (From Ref. 33.)

318

Lyne

Fig. 13 Bristow's instrument for dynamic wetting for wetting and absorption studies. (From Ref. 7.)

It is apparent from the units for liquid transferred that Kr is effectively the mean depth to which the ¯uid may penetrate before capillary forces are initiated (i.e., a mean topographic depth from the plane of the headbox opening). It is also apparent in Fig. 14 that the curves for water show a measurable wetting time, whereas those for oil do not. This wetting delay for polar liquids has been observed by several researchers [5,19,26,50], but the mechanism has never been clearly elucidated. Bristow [5] concluded that the wetting time was a two-stage process involving the time for water to run into the surface topography and the chemical wetting time of the ®ber surfaces. Wetting times for water may vary from about 10 ms for chemically prepared ®ber networks of celluloseÐfor example, ®lter paper [19]Ðto about 50 ms for newsprint made from fresh mechanical pulp [26], to several seconds for sized papers (hydrophobic agent added to the network [5,7]). In the case of self-sized papers made from mechanical pulp, wetting times of about 1 s have been observed [26].

Fig. 14

Amount of liquid transferred to paper on a square-root time axis. (From Ref. 5.)

Wetting and Penetration of Liquids into Paper

319

Extraction with acetone is necessary to remove physically absorbed resin and fatty acids [45,46]. There is evidence to suggest that chemisorption may occur when paper is treated with very small amounts of stearic acid (0.003%) or natural resin and fatty acids and heated to above 100 C [1]. Sodium methoxide extraction has been found to eliminate hydrophobicity in these cases, whereas acetone extraction was found to be ineffective [46]. As shown in Fig. 15, the onset of water penetration into the ®ber wall and associated network swelling can be seen as a deviation upward in the rate of penetration (slope) in a Bristow absorption curve. If water penetrates only into inter®ber pores, then no swelling or network expansion takes place on the time scale of the test (nominally, 2 s). However, if absorption into the ®ber walls occurs on this time scale, the resulting network expansion will cause an increase in the size of the inter®ber pores and an associated increase in the rate of penetration. Penetration of water into the ®ber wall can be delayed by the addition or application of hydrophobic sizing agents. However, water-based inks, varnishes, coatings, and offset fountain solution must be sorbed to some degree in order to facilitate transfer, setting, and/or drying. Thus, it is generally not advisable to size so hard that water cannot penetrate inter®ber pores in the ®ber network. A more practical solution to the hydroexpansion problems is to limit swelling of the ®bers by suitable chemical treatment. Salminen [39] increased the versatility of the Bristow test by crafting a headbox that could be pressurized (see Fig. 16). The surface tension of an aqueous test liquid was varied by controlling the addition of isopropanol. Note the differences in penetration rate for the test liquids of varying surface tension when tested at atmospheric pressure (Fig. 16b, left), whereas the absorption curves for the same liquids show virtually no difference in penetration behavior when 0.5 atm pressure is applied to the test liquid (Fig. 16b, right). This is further evidence that wetting of the surface of paper can be forced by applied pressure and that surface tension and contact angle

Fig. 15 Effect of ®ber swelling on the rate of penetration of water into paper in a Bristow apparatus.

320

Lyne

Fig. 16 (a) Salminen's apparatus for pressure penetration of liquids into paper. (b) Comparison of atmospheric sorption (PE ˆ 0) and penetration under 0.5 atm external pressure of water and aqueous isopropanol solutions into LWC base paper. (From Ref. 39.)

do not come into play for pressure penetration. This begs the question as to how low the applied pressure needs to be to force wetting and negate surface tension effects. Doubtless, forced wetting must occur in printing and converting operations where even low pressure processes such as ¯exography exceed 0.5 atm of applied pressure. Repellency Tests Water repellency is an undesirable property for paper and board that is to be printed by offset lithography. It is necessary that water (or fountain solution) in nonimage area of lithographic printing be absorbed rapidly. In multicolor printing, ink is often printed in areas that received water in a previous printing unit. The oil-based ink will not wet the surface of the paper unless this water ®lm has been absorbed. The IGT [14] and PruÈfbau [34] laboratory print testers have wetting accessories that allow a metered ®lm of water to be applied to the surface of a test paper strip prior to printing with oil-based ink. As shown in Fig. 17, the interval between the application of the water and the printing can be varied in order to assess the time required for absorption of the water ®lm. However, the water ®lm thickness applied in offset printing is generally of the

Wetting and Penetration of Liquids into Paper

321

Fig. 17 Ink repellence as a function of time between the application of water and printing ink. (From Ref. 28.)

order of 1 m, which is dif®cult to replicate reliably in these laboratory printing units.

B.

Penetration of Liquids

As described above, the Bristow type of instrument can be used to directly measure the rate of penetration (absorption coef®cient) for oil or aqueous liquids [5]. However, most methods of measuring penetration rates for aqueous liquids rely on detection of changes in paper properties as a function of liquid penetration. These include measuring changes in electrical conductivity [3,19,20,41], ultrasonic wave propagation [10,47], infrared absorption [21], and surface polarity. The last is most easily measured by the IGT [14] and PruÈfau [34] tests mentioned earlier, where the time required for absorption of a water ®lm is detected by overprinting with an oil-based ink at varied time intervals. The simplest penetration test is doubtless the measurement of the time required for a drop of test liquid to be absorbed after it is allowed to fall onto the test paper. As speci®ed in TAPPI T 492, the drop volume is controlled by using a burette and the endpoint is judged by the sudden loss in gloss of the drop upon absorption. In a variation of this test, the burette can be brought into contact with the paper and opened. the rate of penetration can then be calculated by noting the level of the liquid in the burette at various time intervals. However, the column of liquid in the burette causes a hydraulic head that decreases with the falling liquid level. This problem can be circumvented by introducing the liquid to the underside of the

322

Lyne

paper sample from an ori®ce level with the surface of a large reservoir of the liquid. In this case the absorption of the liquid must be monitored gravimetrically. The simplest method of measuring the rate at which an aqueous liquid penetrates through paper is to ¯oat the paper sample on the liquid and measure the time required for an indicator dye to be wetted on the upper surface of the sample. A mixture of powdered sugar, soluble starch, and methyl violet dye is sprinkled on the back side of the paper sample. The edges of the paper sample are sealed in molten wax, and a watch glass is placed over the powder mixture to prevent vapor and penetration in the plane of the paper web from interfering in the test. Alternatively, a ¯otation device with a desiccator can be used. However, the powder containing the dye can be triggered by the penetration of water vapor through the sheet of paper ahead of the penetration of the liquid. Also, liquid penetration may not be uniform, making subjective judgment of the penetration time dif®cult. These shortcomings can be overcome by using a ¯uorescent dye that is relatively insensitive to water vapor or by using a dyed water solution and measuring the mean drop in re¯ectance form the reverse side of the test sheet. The Hercules size test employs this optical measuring technique. The dye solution may also contain formic acid to speed the penetration of hard-sized papers. However, this practice should be avoided, because the acid changes the chemisty of the absorption process. Conductivity Change Method Tap water contains suf®cient ionic material to be electrically conductive. Therefore, conductivity between electrodes placed on either side of a paper sheet can be used as an indicator of penetration of water from one side of a sheet of paper to the other [3,19,20,41]. Low voltage DC current (e.g., 9 V) can be used in conjunction with a standard conductance meter. To avoid polarizing the liquid, a low frequency (e.g., 1000 Hz) AC current can also be used. The upper electrode can be used to introduce the aqueous liquid, and electrodes can be positioned at various locations under the paper sheet to test penetration times not only through the thickness of the sheet but in any set direction in the plane of the paper sheet (Fig. 18). One shortcoming of the conductance method for detection of penetration through the thickness of paper is that pinholes in the paper may cause conduction paths to be created ahead of bulk penetration. Multiple Internal Re¯ection Spectroscopy In certain frequency bands, infrared light is strongly absorbed by water. In fact, the concentration of water aqueous liquids such as adhesive solutions can be assessed by the degree to which they absorb infrared light. Multiple internal re¯ection is a means of measuring this absorption over a broad sample area and to a depth not exceeding that of a ®lm of adhesive [21]. As shown in Fig. 19, if light is directed across an interface between two media (e.g., glass and an adhesive ®lm) above the critical angle for total internal re¯ection, the light will be re¯ected. However, there will be a certain depth of penetration d into the second medium described by dˆ where

 2…sin  n221 †1=2 2

…15†

Wetting and Penetration of Liquids into Paper

323

Fig. 18 Con®guration of electrodes for conductivity measurement relating to penetration in the thickness direction and in the plane of the paper sheet.

  n21

= wavelength in infrared (3400 cm 1 , or about 3 m) = angle of incidence = refractive index from medium 1 to medium 2

Using a con®guration such as the one in Fig. 20, it is possible to cause multiple internal re¯ections and therefore multiple penetrations of the light into the second medium. If the second medium is an infrared-absorbing liquid, the light will be attenuated each time it passes through the liquid. In practice, the internal re¯ection plate is made of germanium or sapphire with edges beveled to allow incidence and detection of the appropriate angles. The instrument compares the ratio of the light

Fig. 19

Total internal re¯ection in a glass prism. (From Ref. 21.)

324

Fig. 20

Lyne

Multiple internal re¯ections in a plate of germanium or sapphire. (From Ref. 21.)

intensities exiting the internal re¯ection plate when it is in contact with the liquid ®lm and when the surface of the plate is dry (no attenuation). Calibration of this ratio of light intensities to moisture content is made by using liquid ®lms of known moisture content. The instrument can then be used to monitor the penetration of aqueous liquids into paper and paperboard (Fig. 21). One shortcoming of this technique is that during the penetration of the aqueous portion of the adhesive into the paperboard, a concentration gradient is established across the adhesive layer. Therefore, there is a delay in detecting the penetration that depends on the rate of diffusion of water across the adhesive ®lm. Ultrasonic Propagation Technique If transducers are attached to two ends of the test strip of paper, the rate of propagation of ultrasonic wave trains between the transducers can be measured [10,47]. As water is absorbed into the ®bers in the paper strip, the speed of propagation decreases. Thus, the rate of propagation can be calibrated against moisture content in the paper strip. Alternatively, if the aqueous liquid is introduced to one side of the paper strip, the mean rate of penetration of the liquid across the paper can be monitored.

Fig. 21 Solids content in adhesive ®lm as a function of time after application to various types of paper. (From Ref. 21.)

Wetting and Penetration of Liquids into Paper

325

Pan et al. [38] correlated the attenuation of an ultrasonic beam propagated perpendicular to the surface of the paper sheet with the penetration of liquid through the paper. The paper sheet was immersed in the liquid, and the attenuation of the ultrasonic transmission through the sheet was monitored. The increasing interface between the liquid and the interior surfaces of the sheet of paper as the liquid penetrated apparently causes greater dispersion of the ultrasonic beam and attenuation in transmission through the sheet. Although this technique permits relative pro®ling of dynamic penetration perpendicular to the surface, the exact relationship between increase in ultrasonic attenuation and liquid penetration is not de®ned and may depend on morphological aspects of the paper. A commercial version of the instrument is available from Emco that features various modular accessories that permit penetration monitoring under pressures up to 10 bar [15]. C.

Total Uptake of Aqueous Liquids

Immersion Test An immersion test consists of submerging a sample of paper or paperboard under a ®xed height of aqueous liquid (e.g., 7.6 cm in TAPPI T 491), holding it down with a 12 mesh wire screen for a set time interval (e.g., 10 min), then blotting both sides of the sample with a blotter having a standardized rate of absorption. The weight increase of the sample as the result of immersion is taken to be a measure of the amount of liquid absorbed. Because the blotting action of the standardized blotter is crucial to the ®nal weight increase of the sample, it must be assumed that the blotter acts to remove all free liquid, leaving only the absorbed liquid. The absorptiveness of the blotter is checked by cutting strips 15 mm  200 mm in the machine and cross-machine directions, respectively, and immersing one end of the strips to a depth of 10 mm in deionized water. The height of rise after 10 min should be between 50 and 100 mm in both directions. However, this method does not guarantee that the water uptake in a blotting action (in the thickness direction rather than in the plane of the blotter) will be constant, nor does it guarantee that all free water will be removed from the sample. According to Eq. (7), capillaries in the paper sample smaller than those in the blotter would be expected to retain their free water by virtue of higher capillary pressures. Van den Akker proposed a water-holding capacity test that obviates the need for blotting paper. A 3 in:  3 in: sample cut and weighed to the nearest milligram. The sample is immersed in water for 60 s, then removed and placed on a gridlike water extraction device. The device exerts suction on the sample equivalent to a 5 mm column of water for 15 s, after which the sample is reweighed and the water uptake calculated. Finally, air bubbles trapped on the under surface of the sample may also restrict access to the aqueous ¯uid. A light brushing with the liquid prior to immersion of the sample may remove most of these bubbles, but their presence does make the test harder to standardize. One-Side Immersion: Cobb Test As shown in Fig. 22, the Cobb test [11] amounts to sealing one side of the paper sample to a rubber mat by means of a cylindrical ring clamped over the upper surface of the sample. The cylindrical ring has an inside area of 100 cm2 (11.28 cm inside diameter) and a height of 2.5 cm. The

326

Lyne

Fig. 22

Schematic of apparatus for the Cobb test. (From Ref. 44.)

aqueous liquid is poured into the right to a depth of 1 cm (100 cm3 ), and 10 s before the end of a speci®ed time interval (120 s in Ref. 44) the liquid is poured off. The sample is removed and blotted with a standardized blotting paper (see above). Exactly at the end of the present time interval, a second blotter is put on top of the ®rst and a 10 kg roller is passed over the three sheets to extract the free water. The weight increase in the paper sample is taken to be the amount of liquid absorbed in the present time interval. Slightly different test geometries and procedures are followed in CPPA and SCAN test methods. The criticisms mentioned in relation to the immersion test apply to the Cobb test as well, and Bristow [7] further criticized the Cobb test. D.

Pressure Penetration

Darcy's law is invariably the starting point for any discussion of forced penetration of porous media: vˆK where v K 
P=L

= = = =


P L

…16†

volumetric ¯ow rate into the medium per unit area permeability viscosity of the ¯uid pressure drop per unit of penetration distance

The assumptions inherent in this model of pressure ¯ow limit its applicability. Although the pressure penetration of low density pulp pads such as those used in diapers can be modeled using Darcy's law, more anisotropic materials such as printing papers cannot be characterized by a unidimensional permeability factor. Lindsey [24] examined the anisotropic permeability of paper using the following vector form of Darcy's law: vˆ

K

rP 

…17†

where v is now a velocity vector, rP is the pressure gradient, and K is a second-order tensor of the form

Wetting and Penetration of Liquids into Paper

KX K ˆ 0 0

0 KY 0

327

KZ 0 0

where KX , KY , and KZ are the permeability components in the x, y, and z directions (CD, MD, and transverse directions, respectively). The direction of ¯ow need not be the same direction as the pressure gradient in an anisotropic medium. Therefore, in order to calculate KX , KY , and KZ it is necessary to measure velocity and pressure drop in the same direction and then calculate an apparent permeability in that direction. By repeating the process in different directions, it is then possible to calculate KX , KY , and KZ via direction cosines. Because this is not practical experimentally, several investigators have opted to constrain the ¯ow. Nguyen and Durso [35] studied the ¯ow in one direction by suspending the paper vertically and measuring the steady-state ¯ow down the saturated paper strip as a form of siphon (see Fig. 23).
P in the vertical direction is then simply the hydrostatic head, and the permeability at saturation, K0 can be calculated from the ¯ow rate of liquid that is required to maintain a constant level in the reservoir above the sample. They also studied capillary ¯ow in the vertical direction by dipping one end of the sample in a liquid reservoir, then cutting the sample into horizontal strips at different times after immersion. By weighing these strips they were able to assess the rate at which the permeability coef®cient changed with the degree of saturation in the sample. As shown in Fig. 24, their results con®rmed the cubic relationship proposed by Gillespie [16] in his classic studies of imbibition in ®lter papers. Thus, it should be remembered that in most processes involving pulsed impression of liquids into paper, such as in printing or coating, there is a gradient in saturation that will in¯uence local permeability. The rate at which the pore system in paper saturates locally under pressure penetration depends principally on the distribution of pore size and shape.

Fig. 23

Fiber web siphon for calculating permeability at saturation. (From Ref. 35.)

328

Lyne

Fig. 24 Change in the permeability coef®cient K with the degree of saturation s. (From Ref. 35.)

Lindsey [24] and Miller et al. [32] constrained the ¯ow to the plane of the ®ber web by compressing the sample between the parallel plates. Whereas Lindsey measured the amount of liquid pushed into the compressed ®ber web under pressure, Miller weighed the amount drawn into the sample by an applied vacuum. Both were able to study the effect of varying sample compression on porosity and permeability Classically, investigators have used the Kozeny±Carmen equation to relate permeability to porosity: Kˆ

3 de2 36…1 †2 

…18†

The Kozeny constant  is set according to the type of pore structure, and  is the void fraction in the ®ber web calculated as 1 m=V, where m is the mass of the sheet, V is the compressed volume between the parallel plates, and  is the density of solid cellulose. The model assumes a uniform bed of packed particles having an effective particle diameter of de . Although the model works reasonably well for relatively simple structures, such as woven polypriopylene fabrics, it does not accommodate the complex, irregular, and microscopically rough pore systems associated with machine-made paper. Therefore, Lindsey [24] ®t the permeability to the void fraction of paper samples using an exponential model, thereby avoiding the dif®culty in setting the Kozeny constant and in establishing the actual physical volume of the sample. In either case, compression of the samples leads to a highly nonlinear reduction in permeability. Lindsey [24] was able to show that the lateral permeability (KX or KY ) or machine-made papers exceeded the transverse permeability (KZ ) by approximately a factor of 2 and that the permeability in the machine direction was invariably greater than in the cross-machine direction (KY > KX ). Thus, liquids that are impressed into paper will have a natural tendency to spread in the plane of the web, forming an elliptical pattern oriented in the machine direction.

Wetting and Penetration of Liquids into Paper

329

It should be remembered that these models also assume a uniform z-direction structure that is becoming less likely with modern forming, pressing, drying, surface sizing and pigmentization, and calendering methods. Structural variation in the xy plane my result in uneven pressure penetration, which can be a major cause of mottling and related phenomena. Thus, it would be useful to extend these permeability studies to models that more closely resemble real paper structure. Many of the liquids impressed into paper during converting, printing, and enduse operations are non-Newtonian. Miller et al. [32] demonstrated that the penetration of an aqueous glycerol solution into a woven fabric followed a linear relationship between applied pressure and the square of the volume of liquid entering the sample per unit time over a broad range of applied pressures. However, an aqueous solution of polyacrylamide showed only a fraction of the penetration rate at higher pressures that would be expected by a linear extrapolation of its penetration behavior at low pressures. The onset of this non-Newtonian behavior could be determined as a function of the applied pressure in this manner. As shown by Lyne and Aspler [28], resistance to pressure penetration is desirable in printing because it leads to higher print density for the same amount of applied ink. Similarly, resistance to penetration during blade coating is desirable. Long-chain polymers are commonly added to printing inks and coating colors in order to provide greater penetration resistance. However, it is not clear which rheological property or properties predominate in this effect. Shear thickening, elasticity, and resistance to extensional ¯ow may all play a role. Lyne [27] pointed out that the surface pores in paper can be modeled as truncated cones, and thus impression of liquids into the surface of paper involves extensional deformation. To the extent that polymeric additives can increase extensional viscosity on a time scale that is relevant to printing and coating, it may be possible to signi®cantly improve hold-out on porous stocks. Finally, paper and paperboard are often made suf®ciently hydrophobic by hard sizing that liquids will not penetrate under a low hydrostatic pressure. For example, the exposed edges of a seamed milk carton must resist penetration of the milk, or the stiffness of the board will decrease and the container will bulge. Miller et al. [32] used the parallel-plate compression cell pictured in Fig. 25 to determine the critical pressure at which edge wicking of water begins. IV.

CONCLUDING REMARKS

Modern industrial processes, such as lithographic printing, printing with waterbased inks, gluing operations in paper and paperboard converting, and waterbased coating of paper, require wetting and penetration by aqueous liquids at rates measured in tens of milliseconds. Contact angle measurements on sessile drops, the Cobb test, and similar industry standard tests operate on time scales that are orders of magnitude too long to be useful in these applications. Newer tests have been reviewed that have the necessary time resolution. For example, the Bristow apparatus is designed to measure absorption and wetting delays down to 5 ms. The conductivity change method and the ultrasonic propagation technique can measure penetration times to a similar resolution. These techniques should be considered by the paper industry, because it must cope with

330

Lyne

Fig. 25

Parallel-plate permeability cell. (From Ref. 32.)

ever-increasing process speeds and product applications requiring rapid wetting and absorption by aqueous liquids. The methods reviewed in this chapter were selected for discussion because they are of potential use in industrial laboratories. There are, of course, more sophisticated approaches to analyzing wetting and absorption. For example, electron spectroscopy for chemical analysis (ESCA) can be used to probe the chemical composition of the immediate surface (< 5 nm) of papermaking ®bers. Fiber capillography can give a detailed physical characterization of the surface of single ®bers. Inverse gas chromatography (GC) is also used to measure certain thermodynamic properties of cellulose, cellulose derivatives, and paper. Among the properties are the heats of wetting and absorption. However, techniques such as ESCA, capillography, and inverse GC are used primarily for fundamental research into the surface chemistry of materials. They are usually applied under strictly controlled conditions, and experiments must be carried out with care. In contrast, the methods reviewed in this chapter are meant to be applicable to the problems encountered in the high-speed processes used in commercial production. As the process problems become clearly de®ned, the more fundamental techniques of surface chemistry may be used to ®nd the origin of the problems and to suggest approaches for their solution.

REFERENCES 1. 2.

Aspler, J. S., Davis, S., and Lyne, M. B. (1987). The surface chemistry of paper in relation to dynamic wetting and sorption of water and lithographic fountain solution, J. Pulp Paper Sci. 13(2):J55±J60. Bendure, R. L. (1973). Dynamic adhesion tension measurement. J. Colloid Interface Sci. 42:137±144.

Wetting and Penetration of Liquids into Paper 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26.

331

Bohmer, E., and Lute, J. (1966). Adhesive migration and water retention with reference to blade coating. Svensk Papperstidn. 69(18):610±618. Borgin, K. (1959). The properties and nature of the surface of cellulose. 1. Cellulose in contact with water: Experimental results and their interpretation. Norsk Skogind. 13(11):429±442. Bristow, J. A. (1967). Liquid absorption into paper during short time intervals. Svensk Papperstidn. 70(19):623±629. Bristow, J. A. (1968). The absorption of alkaline solutions by paper. Paper Timber (Finland) 11:639±646. Bristow, J. A. (1968). The absorption of water by sized papers. Svensk Papperstidn. 71(2):33±39. Bristow, J. A. (1971). The swelling of paper during short time intervals. Svensk Papperstidn. 74(20):645±652. Carlsson, G. (1996). Composition of wood pulp ®bers: Relevance to wettability, sorption and adhesion. Doctoral Thesis. The Royal Institute of Technology, Stockholm. Chatterjee, P. K. (1971). The sonic velocity response during the absorption of water in paper. Svensk Papperstidn. 74(17):503±508. Cobb, R. M., and Lowe, D. V. (1934). A sizing test and sizing theory. Tech Assoc. Papers 17:213±216. Corte, H., and Kallmes, O. J. (1962). The interpretation of paper properties in terms of structure. In: The Formation and Structure of Paper. F. Bolam, ed. Tech. Div., Br. Paper Board Ind. Fed., London, pp. 351±368. Danino, D., and Marmur, A. (1994). Radial capillary penetration into paper: Limited and unlimited liquid reservoirs, J. Colloid Interfacial Sci. 166:245±250. IGT. (1969). Determination of the wet pick and wet repellence by means of the IGT damping unit. IGT Info. Lea¯et No. W32. IGT, Amsterdam. www.emco.Leipzig.de Gillespie, T. J. (1959). The capillary rise of a liquid in a vertical strip of ®lter paper. J. Colloid Sci. 14:123±130. Hawkes, C. V., and Bedford, T. (1963). The Absorption Characteristics of Paper I. PATRA Lab Report No. 51. Gibbs, J. W. (1961). The Scienti®c Papers of J. W. Gibbs, Vol. 1. Dover, New York, p. 288. Hoyland, R. W. (1978). Swelling during the penetration of aqueous liquids into paper. In: Fiber±Water Interactions in Paper-Making. F. Bolam, ed. Br Paper Board Ind. Fed., London, pp. 557±577. Hoyland, R. W., Howarth, P., and Field, R. (1976). Fundamental parameters relating to performance of paper as a base of aqueous coating. In: The Fundamental Properties of Paper Related to Its Uses. F. Bolam, ed. Br Paper Board Ind. Fed., London, pp. 464±510. Huynh, H. K., Lancaster, P. E., Lepoutre, P., and Robertson, A. A. (1978). The setting of aqueous adhesives on paper. Tappi 61(12):63±65. Kent, H. J., and Lyne, M. B. (1989). In¯uence of paper morphology on short term wetting and sorption phenomena. In: Fundamentals of Papermaking. C. F. Baker, ed. Mech. Eng. Publ., pp. 895±920. Lee, S. B., and Luner, P. (1972). The wetting and interfacial properties of lignin. Tappi 55(1):1116±121. Lindsey, J. D. (1988). The Anisotropic Permeability of Paper: Theory, Measurements, and analytical tools. IPC Tech. Paper Ser. No. 298. Luner, P., and Sandell, M. (1969). The wetting of cellulose and wood hemicelluloses. J. Polym. Sci. C28:115±142. Lyne, M. B. (1978). The effect of moisture and moisture gradients on the calendering of paper. In: Fiber±Water Interactions in Papermaking. F. Bolam, ed. Br. Paper Board Ind. Fed., London, pp. 641±665.

332

Lyne

27. Lyne, M. B. (1990). The importance of extensional rheology in the impression of ink into paper. Adv. Printing Sci. Technol. 20:236±248. 28. Lyne, M. B., and Aspler, J. S. (1981). A review of ink±paper interactions in printing. ACS Symp. Ser. Colloids Surf. Reprog. Technol. 200:385±420. 29. Lyne, M. B., and Aspler, J. S. (1982). Wetting and the absorption of water by paper under dynamic conditions. Tappi 65(12):98±101. 30. Lyne, M. B., and Huang, Y. C. (1993). Measuring acid±base and dispersive interactions with paper surfaces under dynamic conditions. Nordic Pulp Paper 8(1):120±122. 31. Marmur, A. (1992). Adv. Colloids Interface Sci. 39:13±33. 32. Miller, B., Friedman, H. L., and Amundson, R. (1991). In-plane ¯ow of liquids into ®brous networks. Proc. Tappi Nonwovens Conference, pp. 19±21. 33. Miller, B., Tyomkin, I., Wesson, S. P., and Mahale, A. D. (1992). Interactions of Fluids with Fibrous Materials. Textile Res. Inst. Rep. 49. Textile Res. Inst., Princeton. 34. Niesser, G. (1972). The interaction of printing ink and paper, 2. Druck 10:687 (in German). 35. Nguyen, H. V., and Durso, D. F. (1983). Absorption of water by ®ber webs: An illustration of diffusion transport. Tappi 66(12):76±79. 36. Okagawa, A., and Mason, S. G. (1978). Capillarography: A new surface probe. In: Fiber±Water Interactions in paper-Making. F. Bolam, ed. Br. Paper Board Ind. Fed., London, pp. 581±586. 37. Oliver, J. F., and Mason, S. G. (1976). Scanning electron microscope studies of spreading of liquids on paper. In: The Fundamental Properties of Paper Related to Its Uses. F. Bolam, ed. Br. Paper Board Ind. Fed., London, p. 209. 38. Pan, Y.-L., Kuga, S., and Usuda, M. (1988). An ultrasonic technique to study wetting and liquid penetration of paper. Tappi 71(5):119±123. 39. Salminen, P. J. (1988). Water transport into paper: The effect of some liquid and paper variables. Tappi 71(9):195±200. 40. Senden, T. J., Knackstedt, M. A., and Lyne, M. B. (2001). Droplet penetration into porous networks: Role of pore morphology. Nordic Pulp Paper. In press. 41. Stinch®eld, J. C., Cliff, R. A., and Thomas, J. J. (1958). The water retention test in evaluating coating color. Tappi 41(2):77±79. 42. TAPPI (1994). TAPPI T 458. Surface Wettability of Paper (Angle of Contact Method). 43. TAPPI (1997). TAPPI T 558. Surface Wettability and Absorbency of Sheeted Materials Using an Automated Contact Angle Tester. 44. TAPPI (1998). TAPPI T 441. Water Absorptiveness of Sized (Non-Bibulous) Paper and Paperboard (Cobb Test). 45. Takeyama, S., and Gray, D. G. (1979). Surface analysis of some sulphite pulps by ESCA. Trans. Tech. Sect. Can. Pulp Paper Assoc. 6(3):61±64. 46. Takeyama, S., and Gray, D. G. (1982). An ESCA study of the chemisorption of stearic acid on cellulose. Cellulose Chem. Technol. 16(2):133±142. 47. Taylor, D, L., and Dill, D. R. (1967). Water retention and coating colors: A study of the sonic velocity method and the effect of color composition on water retention. Tappi 50(11):536±541. 48. Tremaine, P. R., Mohlin, U., and Gray, D. G. (1977). The adsorption of n-decane on the surface of water-swollen cellulose ®bers. J. Colloid Interface Sci. 60(3):548±554. 49. van Oss, C. V. (1991). Dispersion Sci. Technol. 12(2):201±219. 50. Windle, W., Beazley, K. M., and Climpson, M. (1970). Liquid migration from coating colors. Tappi 53(12):2232±2242.

8 ELECTRICAL PROPERTIES: I. THEORY GARY A. BAUM The Institute of Paper Chemistry Appleton, Wisconsin

I. Introduction II. De®nitions III. Fundamental Considerations A. Direct Current Conductivity B. Dielectric Constant C. Dielectric Anisotropy

I.

333 334 337 338 346 351

Symbols

355

References

357

INTRODUCTION

The electrical properties of paper and board are important from a number of points of view. As a dielectric in capacitors or as an insulator for wires, coil windings, or cables, paper provides good dielectric properties together with desirable mechanical properties at a relatively low cost. In combination with other materialsÐfor example, saturating oils or resinsÐit is used in transformers, electronic circuit boards, and numerous other applications. The electrical properties are also important in many reprographic processes. For example, in electrographic and dielectric printing methods, the paper must meet certain minimum electrical conduction requirements. Usually it must be rendered 

Current af®liation: Institute of Paper Science and Technology, Atlanta, Georgia. 333

334

Baum

conductive by the application of special materials prior to being coated with the photoconductive or dielectric ®lm. In electrophotographic printing, plain paper properties must meet speci®c requirements as described in Chapter 9. The importance of the electrical characteristics is obvious in the applications mentioned above. Perhaps not so obvious is the need to understand these properties of paper (or ®bers) with respect to other processes. For example, electrostatically assisted rotogravure, radio frequency or microwave dryers, and radio frequency or microwave moisture gauges all involve the interaction of electromagnetic ®elds with paper and water systems. A fundamental understanding of this interaction is essential to the proper use of these devices. The same understanding is needed with respect to ®bers in certain air-lay processes or solid±solid separation processes. Although much research has been carried out over the years with respect to the electrical properties of paper and board, there is still an incomplete understanding of such basic phenomena as the nature of the electrical conduction process itself, the role played by water and the other polar groups in the dielectric response of cellulose, time and temperature effects on the conduction process, and the effects of additives or sheet process variables. A greater awareness is needed of the fact that the electrical properties of paper are anisotropic. Just as paper must be considered an orthotropic medium in terms of its elastic properties, it should also be considered an orthotropic medium in terms of its dielectric properties. Although this may not be very important in terms of the more conventional electrical uses of paper as an insulator, it has much signi®cance in terms of electromagnetic dryers, moisture gauges, and other devices.

II.

DEFINITIONS

The electrical properties of interest with respect to paper and board include volume and surface conductivities (or resistivities), dielectric constants and dielectric loss, permittivities, and dielectric breakdown strength. When paper is used as an insulator, the requirements would be for low conductivity, low dielectric loss, and high dielectric breakdown strength. In other applications, such as certain reprographic processes, the paper must have relatively high volume or surface conductivities. The volume conductivity v is the ratio of the current density J to the applied electric ®eld strength E or J ˆ v E

…1† 2

Both E (newtons per coulomb; NE) and J (amperes per square meter; A=m ) are vector quantities. In a homogeneous isotropic material, the conductivity, (ohm cm† 1 , will be a scalar. In a homogeneous anisotropic material, where the properties vary in a different manner along different directions at a point, the conductivity v forms a symmetrical second-rank tensor. In terms of mechanical properties, paper and board behave as (three-dimensional) orthotropic materials in which there are three mutually perpendicular symmetry planes. The elastic properties are different in the three directions. Similar behavior is expected for the electrical properties. Three independent conductivities are anticipated, one along each of the three principal directions in the paper (the machine direction, cross-machine direction, and thickness or Z direction). Although no measurements appear to have been reported

Electrical Properties I

335

in the literature for conductivities of paper, Lin [36,37] did study this anisotropy in wood. In most uses of paper where its electrical properties are important, the direction of concern is the thickness (Z) direction, and values reported for the volume conductivity of paper are measured in this direction. In many applications, the in-plane conductivities may be of little importance. As discussed later, a similar anisotropy is expected (and observed) for the dielectric constant. The volume resistivity v in a given direction in a material is the reciprocal of the volume conductivity in the same direction. Both v and v are material properties independent of the specimen and electrode geometries. The volume resistance R is de®ned between two electrodes contacting the specimen as the ratio of the applied voltage to the volume current. It is a function of both the resistivity and the geometry. The volume resistivity and R are related by R ˆ v L=A (ohm), where L is the distance between electrodes and A is the cross-sectional area of the specimen between electrodes. It is often useful to speak in terms of a surface resistivity, especially in those situations where the surface of the material has been altered in some manner. The surface resistivity s is taken as the (surface) resistance measured between the opposite sides of a square on the surface. It is independent of the size of the square and is expressed in ohms. The surface conductivity s is the reciprocal of s . The surface conductivity is frequently used to characterize reprographic papers where the surface has been rendered conductive by the application of special resins (Chapter 9). The insulation resistance of a material is de®ned as the ratio of the applied voltage to the resultant current, and it involves both the volume and surface resistances. Actual measurements of any of the above parameters may be made using either direct current (DC) or alternating current (AC) equipment. To avoid confusion, the quantity involved is usually prefaced with the letters AC or DCÐfor example, AC conductivity or DC conductivity. In electromagnetic theory, the proportionality constant between the electric displacement vector D (C=m2 ) and the electric ®eld strength E in a material is called the permittivity  (F/m) or D ˆ E

…2†

Just as in the case of the conductivity,  is a scalar if the material is homogeneous and isotropic. If the material is anisotropic,  is a symmetrical second-rank tensor. For paper, assuming orthotropic symmetry, this permittivity tensor has three independent nonzero components. Each component relates the electrical displacement to ®eld strength in one of the three principal directions. The permittivity itself is usually represented as a complex number. The signi®cance of this is perhaps most easily understood by considering a parallel-plate capacitor. If a capacitor with a vacuum dielectric is connected to a sinusoidal voltage source, V ˆ V0 exp…i!t†, it will store a charge (coulombs), Q ˆ C0 V. In the foregoing expression for V, V0 (volts) is the voltage at time t equal to zero, ! is the angular frequency (ˆ 2f ), and i ˆ … 1†1=2 . The quantity C0 (farads) is the geometric (or vacuum) capacitance of the capacitor and, neglecting any fringing ®elds, is equal to 0 A=d, where 0 is the permittivity of free space, A the surface area of one capacitor

336

Baum

plate, and d the distance between plates. The charging current IC (amperes) is given by IC ˆ i!C0 V. If the vacuum capacitor is now ®lled with some material having a permittivity , two things happen. The capacitance increases to C ˆ  0 A=d, where  0 is the real part of the permittivity of the substance, and at the same time energy losses occur because the material is not a perfect insulator. The substance-®lled capacitor must be characterized by both a capacitance parameter and a loss parameter. Thus, in addition to the charging current, we now must include a loss current IL . so the total current is IT ˆ IC ‡ IL . The loss current might arise because of ion migration or because dipoles induced in the material as a result of the impressed voltage are hindered in their attempt to follow or align themselves with the applied ®eld. Such loss mechanisms may be represented as either a parallel or a series combination of a resistance and a capacitance. In the parallel circuit of Fig. 1A, R represents the loss component and C the pure capacitive component. The loss current is IL ˆ V=R, so that the total current is IT ˆ i!CV ‡ V=R. A vector diagram for this circuit is shown in Fig. 1B. In the ®gure,  is the phase angle and  is the dielectric loss angle. The dissipation factor D is de®ned as the cotangent of the phase angle or D ˆ cot  ˆ tan  ˆ

1 !RC

…3†

This quantity is also referred to as the loss tangent.

Fig. 1 (A) Schematic circuit depicting charging and loss currents. (B) Vector representation of Fig. 1A.

Electrical Properties I

337

It is customary to describe the charging current and the loss current by the introduction of a complex permittivity,  ˆ 0

i 00

…4†

or a complex dielectric constant (dimensionless), k ˆ =0 ˆ k 0

ik 00

…5†

where  00 and k 00 represent the loss parts of the permittivity and dielectric constant, respectively. In terms of the complex dielectric constant, the total current may now be written IT i!kC0 V, and the loss tangent becomes tan  ˆ  00 = 0 ˆ k 00 =k 0

…6†

In the case of paper, the dielectric parameters reported are usually the real and imaginary parts of the dielectric constant, k 0 and k 00 , although the latter is usually called the loss factor. It is also common, however, to see reference to the dissipation factor D or loss tangent tan . The quantity R ˆ 1=…!C tan ) is referred to as the AC resistance. In addition to the above parameters, there is another that is important in describing the electrical properties of paper. This is the dielectric strength. The dielectric strength is a measure of the ability of the dielectric to withstand high voltages and maintain its high resistance to current ¯ow. It is usually de®ned as the maximum electric ®eld strength that can be applied without causing an irreversible failure of the material. If the dielectric strength is exceeded, the material may break down, meaning that the material has lost its high resistance to current ¯ow. That is, it becomes conductive.

III.

FUNDAMENTAL CONSIDERATIONS

In a literature review on electrical papers in 1972, Raman and Walker [54] concluded that even though a considerable amount of research had been carried out, there was a distinct lack of detailed understanding of many aspects concerning the electrical characteristics of papers. Although some progress has been made in recent years, the general conclusion still appears to be valid. With respect to its electrical properties, paper has some interesting and somewhat unusual characteristics. This section is intended to acquaint the reader with those characteristics and attempts to provide an overview of our current understanding of electrical phenomena in paper and board. A word of caution is in order. Much of the work reported in the literature refers to ``cellulose.'' This term may encompass a number of substances and cellulose-based materials with varying amounts of lignin and hemicelluloses; therefore, it is sometimes dif®cult to directly compare experimental results reported in the literature. In this summary we attempt to include as complete a description of the specimen as possible.

338

A.

Baum

Direct Current Conductivity

Introduction If a DC potential is suddenly applied to paper, the initial current is a function of time, eventually decaying to some steady-state (time-independent) value. Such behavior is not unique to paper and can be attributed to electrolytic polarization and high contact resistances, among other things. This time dependence is discussed later. It does give rise to differences between measured AC and DC conductivities in paper. Friedrich and Chiu [19], studying surface conductivities, observed that at long times (several minutes) the DC conductivity approached the AC value (from the high side). The following discussion refers to the steady-state DC conductivity. The very early literature dealing with the electrical properties of paper and related materials has been reviewed elsewhere [5,28,49,54] and will not be presented in detail here, because our objective is to provide an overview of current understanding concerning these properties. A series of papers published in 1947 and 1948 by O'Sullivan [49±53] considered the conduction of electricity in cellulosic materials from a fundamental viewpoint. He demonstrated that conduction in cotton or regenerated cellulose was dependent on an association of moisture and naturally occurring electrolytes in the material. If the electrolytes were removed, the conductivity at a given relative humidity decreased. For a ®xed salt content, the conductivity was an exponential function of the conditioning relative humidity. In addition, at salt contents greater than about 1% the conductivity was largely governed by the moisture content, being insensitive to the amount or type of salt present. At low salt concentrations (< 1%), the conductivity was affected by the amount present (at a ®xed relative humidity). O'Sullivan showed that ions could migrate between electrodes and concluded that the conduction process in cellulose was ionic rather than electronic. He strengthened that argument by further demonstrating that the moisture dependence observed for conduction in the cellulosic sheet was similar to the dependence found for the mobility of ions as a function of moisture content. Similar observations were made by Hearle [23±27] in textile materials. Phenomenologically, the conductivity of any material may be expressed in terms of the density of charge carriers n, the electronic charge on the carrier e, and the mobility of the carrier in the solid , as v ˆ ne

…7†

The mobility is de®ned as the average drift velocity of the carrier per unit electric ®eld strength. If more than one type of ion is present, X v ˆ n i e i i …8† i

where the subscript i refers to a particular ion species. In paper, the alkali metal ions are often considered to be the predominant current carriers, although anions may also play an important role [56]. The metal ions are assumed to be bound at sites along the molecule but capable of being dissociated. In this case, ni would be expected to be proportional to the salt content, according to a Boltzmann expression:   Ui ni ˆ Ni exp …9† 2kB T

Electrical Properties I

where Ni Ui kB T

ˆ ˆ ˆ ˆ

339

concentration of ionizable sites of species i energy required to dissociate the ion Boltzmann's constant absolute temperature

The effect of changing moisture on ni (or i ) is not clear at this point, but O'Sullivan's observation that the conductivity and mobility (measured separately) varied in the same manner with increasing moisture implies that the mobility is the dominant factor controlling the conductivity in the case of high salt contents. At low salt contents, the concentration of charge carriers is limited, and this factor must outweigh the strong effect of moisture on the mobility. This picture is probably too oversimpli®ed, however, in that Ui in Eq. (9) would be expected to be a function of the moisture content because the energy required for dissociation will depend upon the local conditions at the ionizable site, in particular on the dielectric constant at the site. The dielectric constant will change as moisture is added, because the dielectric constant of water is much greater than that of dry paper or other cellulosic materials. The increasing local dielectric constant at the ionizable site as moisture is added causes a decrease in the energy required for dissociation, with a subsequent increase in ni (assuming that all sites are not already ionized). The same type of argument may apply to the observed increase in mobility with increasing moisture, but there does not appear to be a straightforward method of separating the contributions due to mobility and carrier concentration. The effect of changing moisture content on the conductivity is very dramatic. Figure 2, for example, shows that the conductivity of a natural cellulosic (cotton) varies over 12 orders of magnitude as the moisture content is varied from bone dry to saturation. Murphy [42] studied this phenomenon in some detail. He found that the conductivity could be expressed in terms of the moisture content M (expressed in percent of the dry weight) as v ˆ sm …M=Ms †m

…10†

where Ms ˆ moisture content at saturation sm ˆ conductivity at the saturation water content For his cotton samples, Murphy found m to be 9.3. To explain this empirical result, he proposed that there were metal ion±generating sites that could be ionized according to Eq. (9). These sites were assumed to be periodically distributed throughout the material and connected by a continuous chain of water adsorption sites. Murphy argued that a contribution to the conductivity occurred when there was a complete chain of water molecules adsorbed between the ion-generating sites. The probability that only one water adsorption site is occupied is M=Ms . The probability that m of these would be occupied simultaneously is …M=Ms †m . Hence, in cotton, where experimentally m ˆ 9:3, Murphy proposed that there were nine water adsorption sites between each neighboring pair of ion-generating sites. Barker and Thomas [2], studying alkali halide±doped cellulose acetate, suggested some modi®cations to Murphy's model that better explain their results and

340

Baum

Fig. 2 Conductivity of a natural cellulosic (cotton) as a function of relative water content, M=Ms . Filled circles refer to an increasing sequence of relative humidities, open ones to a decreasing sequence. Solid lines are the empirical curves corresponding to the power 9.3 in Eq. (10). Broken lines correspond to the power 9 given by Murphy's model when m is restricted to integer values. (From Ref. 42.)

provide a different viewpoint as to the role of the water. They assumed that the sample could be divided into small cells, each containing a site that can generate an ion if the appropriate number of water molecules enter the cell. Thus, in this case, m is the number of water molecules necessary for ion generation. If the distribution of water molecules among the cells is individually and collectively random, Poisson's distribution applies, leading to the expression  v ˆ Asm

M Ms

m



 exp ms 1

M Ms

 …11†

where A ˆ constant ms ˆ number of water molecules in the cell at saturation In this model, m is perhaps related to the hydration numbers of the ions. Although the mechanism by which moisture in¯uences the conductivity may not be certain, the notion that discrete water molecules can affect the electrical properties is consistent with the phenomenological argument presented earlier.

Electrical Properties I

341

Ionic Conductivity and Other Models A useful semiquantitative model of the conduction process in paper is similar to one described by Seitz [59] for ionic conduction in alkali halides, which has also been applied to polymers [1,33]. This model assumes that ions are generated according to Eq. (9) and that, additionally, the ions experience potential barriers due to the other constituents of the material. [The ions can recombine with a site of opposite sign, of course. Equation (9) describes the net concentration of carriers at equilibrium.] Some of the ions will have suf®cient energy to overcome these barriers and thus migrate through the material. In the absence of an electric ®eld, there will be as many ions moving to the right as to the left. Ion migration is assumed to be equally probable in any direction. When an electric ®eld is present, however, the pattern of migration of the ions is altered such that more (positive) charge carriers move parallel to the ®eld than antiparallel. The net movement of ions constitutes a current that is detectable in the external circuit. The drift velocity of the ions parallel to the ®eld (the mobility) will depend on the local environment. For this ionic conductivity model, it is shown that the behavior is non-Ohmic [does not obey Ohm's law, Eq. (1)]. In fact, J ˆ A sinh…BE† where

 A ˆ GebNi exp

…12† Ui =2 ‡ Ub kB T



B ˆ eb=2kB T where G ˆ constant involving, among other things, the vibrational frequency of the ion e ˆ magnitude of the ion charge b ˆ ``jump distance,'' the distance traveled by the ion to overcome the potential barrier of height Ub The quantity Ub is sometimes called the activation energy for mobility. The other parameters have been de®ned previously. Theory and experiment are compared in Fig. 3, which gives J versus E for several papers and a regenerated cellulose ®lm. Ohm's law is also shown on the semilogarithmic plot for comparison. The agreement between theory and experiment is good, and values can be found for the parameters A and B. From B a value can be estimated for the jump distance b. For the cellulosic ®lm for which data are shown in Fig. 3, if one assumes a monovalent ion, this distance turns out to be around 20 nm (200 AÊ), a surprisingly large value compared to, say, the length of a cellobiose unit (1.03 nm). If a divalent ion were assumed, b would be halved. It is not clear whether values of this magnitude are reasonable or whether the ionic conductivity model is inadequate when it comes to describing a heterogeneous material such as paper. The samples of Fig. 3 suffered dielectric breakdown after the maximum value given for E for each material. That is, the dielectric strength was about 12  106 nT=C. These samples were nondielectric papers and were not intended for

342

Baum

Fig. 3 Steady-state current density versus electric ®eld strength for several nondielectric papers and a cellulosic ®lm at low moisture contents (M.C.). The solid lines correspond to the model given by Eq. (12).

use as insulating materials. Nonimpregnated dielectric papers should be able to withstand much higher ®eld strengths than those in Fig. 3. At very high values of ®eld strength, the hyperbolic sine function will be approximated by an exponential term. That is, J should increase exponentially with increasing E. Such behavior is observed [22,44]. Murphy [44] studied conduction in capacitor tissue as a function of E at ®eld strengths up to 5:0  107 nT=C (500 kV/cm). He found that the conductivity can be expressed as a sum of two exponential terms. The slope of the curve of log resistivity versus E decreases around 200 kV/cm. Murphy concluded that although the paper is an ionic conductor, the behavior is suggestive of a local electronic conduction in the vicinity of transient thermally generated defects.

Electrical Properties I

343

Equation (12) correctly predicts the temperature dependence of the conductivity over a broad temperature range. If E is held constant, the hyperbolic sine term is relatively insensitive to changing T, compared to the exponential term. Thus for constant E and changing T,   J U v ˆ ˆ 0 exp …13† E kB T where the sinh term and constants are contained in the parameter 0 and U ˆ Ui =2 ‡ Ub . A typical plot of log resistivity (ˆ 1=v ) versus T 1 is shown in Fig. 4 for a regenerated cellulose ®lm. The energy U can be determined from the slope of this line, as 0.92 eV (21.2 kcal/mol). Murphy [43] studied the conductivity of dry capacitor tissue as a function of temperature in the range 25±175 C. He found that from room temperature to about 60 C the energy U was 10.6 kcal/mol, which he attributed solely to an activation energy for mobility, Ub , for impurity ion conduc-

Fig. 4

Log resistivity versus inverse temperature for a plasticized cellulose ®lm.

344

Baum

tion. At higher temperatures, however, the slope of the curve ln v versus T 1 increased, yielding an energy of 30.7 kcal/mol. Murphy argued that this value could be broken into a dissociation energy Ui of 40.2 kcal/mol and a mobility activation energy of 10.6 kcal/mol. That is, Ui =2 ‡ Ub ˆ 30:7 kcal/mol. According to Murphy, the 40.2 kcal/mol agrees favorably with the activation energy for thermal decomposition of regenerated cellulose of 39.5 kcal/mol, and 10.6 kcal/mol is approximately equal to the activation energy for conduction in ice. The ionic conductivity model can predict the effects of moisture on the conductivity only qualitatively. As suggested earlier, the presence of water alters the local dielectric constant, which in turn would affect U (or Ui or Ub ). The value of U should decrease with increasing moisture. Data testing this effect in paper appear to be scarce. Lin [36] has discussed the effect in the case of wood. Other models have been proposed for conduction in paper. Hanneson et al. [22], studying capacitor tissue paper, suggested a model in which ionic space charges accumulate near one or both electrodes and thereby limit a predominantly electronic steady-state current. The contribution of the ionic current is presumed to be transient. Although the idea that electrons carry the current in paper (or cellulosic materials) is contrary to all previous thinking, the model is successful in describing some experimental observations, and it should be given some attention. For example, in the case of the dependence between current density and ®eld strength, the space-charge-limited electronic model predicts that J should vary exponentially with the square root of E. If the data in Fig. 3 are replotted as log J versus E 1=2 they do fall in a straight line (except at the lowest values of E). The model supposes that a carrier exists at the (electrode) metal/paper interface. The height of the barrier, in part, depends on the work function of the metal. If different electrode metals were to be used, differences would be expected in the measured currents. Such behavior, however, is not observed [44]. Time Effects If a voltage step function is applied to paper, several distinct time regions are observed in which the behavior of the current with time differs signi®cantly. At very short times, on the order of microseconds, paper behaves as a typical dielectric material, with the current decaying exponentially with time according to h V t i exp Iˆ …14† R RC where R ˆ resistance (ohms) of the paper sample C ˆ capacitance (farads) of the paper sample V ˆ applied voltage For typical papers, the product RC, the relaxation time, is on the order of a few microseconds [4]. At longer times, greater than about 100 ms, the current tends to approach a ``nearly'' steady-state value. The current may continue to slowly decrease for minutes or hours before a true steady-state value is reached, if at all. In the range from about 10 s to 100 ms, however, the current decays according to I ˆ at

p

…15†

where a and p are constants. Such behavior is observed in many insulating materials.

Electrical Properties I

345

Figure 5 shows a plot of log current versus log time, from 1  sec to 1 min, for a coating base stock paper (solid line). If the steady-state current at 1 min is subtracted from the curve, the dashed line is obtained, showing the exponential behavior at short times and the t p behavior at the intermediate times. For the data shown, the value of p is 0.67. Although studies of this phenomenon in paper are limited, this value appears to be typical for paper and organic polymers [35,71,72]. Several explanations have been offered for the time dependence observed in polymers [70,71]. These involve electrons as charge carriers with trapping sites distributed over a range of energy levels. At this time it is clear that neither the ionic model presented here nor any of the various electronic models proposed for other polymeric materials [30,55] are completely in accord with the available experimental evidence for paper. Perhaps in the future a de®nitive theory will be found, but in the interim much additional experimental work needs to be done.

Fig. 5 Time dependence of the current when a constant voltage of 175 V is suddenly applied to a paper sample.

346

Baum

Wood Pulp Fibers A number of theories have been put forth attempting to relate the mechanical properties of paper to the ®ber properties and the ®ber orientations in the web. Similar models have not been put forth in regard to the electrical properties of paper and their relationships to the individual ®ber characteristics and distribution in the sheet. We will ®nd in the next section that the anisotropy observed in the dielectric properties of machine-made papers can be related to geometric effects and probably anisotropy in the ®bers themselves. Norimoto el al. [46] related the anisotropy of the dielectric constant in the cell wall to the anisotropy of the dielectric constant in coniferous wood. Very few data on individual wood pulp ®bers are available. Smith [64,65] studied parallel groups of ®bers and attempted to relate sheet conductivities to ®ber conductivities. Lowe and Baum [38] made conductivity measurements on individual loblolly pine ®bers. They found that the conductivity of both earlywood and latewood ®bers varied exponentially with relative humidity. The in-plane conductivity of small handsheets made from the same ®bers showed a similar dependence. Over the relative humidity range examined (11±84%) the ®ber conductivity was 50±100 times greater than the paper conductivity. B.

Dielectric Constant

Introduction Historically, the dielectric constants of paper have received more attention than the conductivity, perhaps because the former are more closely related to the actual end-use requirements for most electrical papers. As expected, the ion and water contents still play important roles. Delevanti and Hansen [12] provided a review of the early literature concerning the dielectric properties of chemical pulps, and Raman and Walker [54] more recently reviewed the literature on electrical papers. In order to understand the dielectric properties of paper, it is worthwhile to review the factors that in¯uence the dielectric constants of materials in general. The dielectric constant or permittivity of any material is a measure of the polarizability of its constituents. The polarizability is an atomic property, whereas the dielectric constant also depends on the concentration and spatial distribution of the constituents. The polarizability can usually be separated into several parts. In the presence of an electric ®eld, an electronic contribution arises from a displacement of the electron shell relative to a nucleus and an ionic contribution from the displacement of a charged ion with respect to other ions. In materials that possess molecular groups having permanent dipole moments, such as water or the hydroxyl and carboxyl groups in lignin or cellulose, the molecular dipoles will also make a contribution. At low frequencies all of these parts contribute to the dielectric constant, as will any free ions (space charges) in the material. The space charges and permanent dipoles attempt to oscillate back and forth in phase with the electric ®eld vector. Their ability to do this depends upon the inertia of the group, the nature of the local environment (steric hindrances), temperature, and frequency. Any interactions with the surrounding atomic structure can diminish the contribution of the group to the overall polarizability of the material, and any energy dissipated in the process will be re¯ected in the magnitude of the loss factor. As the frequency increases, the space charges and permanent dipoles ®nd it increasingly dif®cult to keep in phase with the rapidly changing electric ®eld vector, and at some frequency they ``relax out.'' That is, they no longer make a contribution

Electrical Properties I

347

to the polarizability or the real part of the dielectric constant. The relaxation is accompanied by a large increase in the imaginary part of the dielectric constant (loss factor). At still higher frequencies the group does not contribute to either k 0 or k 00 [refer to Eq. (6)]. Space charges are usually the ®rst to relax out, followed by the permanent dipole groups (at frequencies in the radio frequency or microwave regions). Ionic and electronic polarizabilities still make contributions at infrared frequencies and above. Cellulosic materials contain space charges and permanent polar groups, and both are important at low frequencies. If water is present, however, even in relatively small amounts, it is apt to make the major contribution owing to its polar nature. Machine-made papers are anisotropic in their mechanical properties because of ®ber alignment with the wire during formation and because of restraints imposed on the web during drying. This anisotropy at the ®ber level governs the distribution of polar groups in the material. Because the dielectric properties are sensitive to the manner in which the polar groups are arranged in space, the anisotropy of the paper would be expected to be manifest in the dielectric properties. This has been observed and is discussed in detail later. It is important to understand that the dielectric constant has different values along each of the three principal directions: the machine direction, the cross-machine direction, and the thickness or Z direction, in accordance with Eq. (2). In the following discussion the dielectric constant referred to is the one measured in the thickness direction. Low Moisture and Free Ion Concentrations At low moisture contents and low free ion concentrations, the polar groups of cellulose should be a major contributor to the dielectric constant at low frequencies. For dry, low ash papers, dielectric constants around 1.3±1.8 have been reported [13,21,58,63,66,67,69]. Table 1 gives some literature values for dry paper and regenerated cellulose samples. The concentration of polar groups has an important bearing on the overall dielectric constant. Because the concentration depends on the density of the material, the dielectric constant would be expected to vary with density. Any process whose effect is to increase paper density, such as re®ning or wet pressing, should also increase the dielectric constant. Seidman and Mason [58] estimated that for their ®lter paper samples with a dielectric constant of 1.35 (10 kHz), the dielectric constant of the cellulosic ®ber itself was 5.5. They also reported values for transverse and axial dielectric constants of the cellulose crystallite as 5.27 and 7.19, respectively, at 300 kHz. Delevanti and Hansen [12] and Calkins [8] reported that the relationship between the density d and the (real part of the) dielectric constant of paper satis®ed a Clausius±Mosotti type relationship: k0 1 ˆ Kd k0 ‡ 2

…16†

where K is a constant for a particular ®ber and involves the polarizability of the atoms and their arrangement in space. The ability of a polar group to respond to an electric ®eld also depends upon the temperature and the local environment of the group in question. In crystalline regions, for example, the groups will not be as labile as similar groups in a noncrystalline region. The differences in steric hindrance between noncrystalline and

348 Table 1 Dielectric constanta k0 1.33 1.32 1.31 1.4 1.4 1.4 1.8 2.67 2.60 1.67 1.87 2.86 2.64 2.50 2.42 1.6 1.8 a b c

Baum Dielectric Constant (k 0 ) for Dry Paper and Other Cellulosic Materials

Frequency 100 kHz 550 kHz 1 MHz 0.1±200 kHz 27 MHz 27 MHz 27 MHz 1 kHz 1 kHz 10 kHz 10 kHz 200 kHz 2 MHz 5 MHz 10 MHz 9.6 GHz 9.6 GHz

Moisture content b (%)

Density (g/cm3 )

Temperature ( C)

Ref.

Low ash ®lter paper

0

0.496

25

58

Printing paper Newsprint Linerboard Tracing paper Viscose rayon Alpha wood pulp Cotton linters Bleached sul®te Cotton cellulose

2.3 0c 0c 0c 0 0 0 0 0

Ð 0.49 0.70 1.03 0.627 0.721 0.404 0.485

Ð

66 13

105 105

69

Newsprint Linerboard

0c 0c

0.375 0.653

Material

67

25

63

25 25

21

Dielectric constant not corrected for density. See text. Most measurements carried out in vacuum. Obtained by extrapolating plots of dielectric constant versus moisture to 0% moisture.

crystalline regions have been found to result in measurable differences in the dielectric constant. Calkins [8] and Verseput [69] ®rst studied the relationship between dielectric constant and the accessibility of cellulose, with additional work undertaken by Kane [29]. Venkateswaran and Van den Akker [68] examined the dielectric properties and crystallinity of ramie, cotton linters, bleached sul®te pulp, and cellophane when the materials were treated with ethylamine in water. They found a linear relationship between the crystallinity and dielectric constant, the latter decreasing as the crystallinity increased. The effect of crystallinity and other paper variables is shown in more detail in Chapter 9. Frequency Effects As noted earlier, increasing frequencies will tend to decrease k 0 and increase k 00 . The decrease in k 0 can be seen in some of the data in Table 1. When relaxation occurs, the rising loss factor will reach a maximum and then decrease to some new low value, and the dielectric constant k 0 will also decrease to a new value. The polar group in question no longer makes a contribution to the complex dielectric constant. If more than one type of polar group is present, and they relax out at suf®ciently different frequencies, separate peaks or bands may be observed in the loss part of the dielectric constant when it is plotted against frequency. The relaxation frequency increases with temperature. Because of the strong dependence upon the conditions of the local environment, such bands are apt to be broad and overlap one another, rather than sharp distinct bands, a re¯ection of the

Electrical Properties I

349

fact that the local conditions are likely to vary on an atomic scale from point to point. No well-de®ned bands are observable for paper near room temperatures. Renne [56] observed a loss maximum at 55 C and 1 kHz for condenser paper and attempted to relate it to the cation concentration. Mikhailov et al. [41] reported a similar loss maximum at 67 C and 1 kHz and attributed it to the movement of the primary hydroxyl groups. Seidman and Mason [58] investigated the dielectric relaxation in dry cellulosic material and cellulose-containing sorbed vapors over the temperature range from 60 C to 30 C and a frequency range of 10±1000 kHz. For dry material, they found that the dielectric constant k 0 increased over this temperature range, whereas the loss component k 00 went through a maximum. The temperature of the maximum increased as the measuring frequency increased. Klason and KubaÂt [32] investigated loss factors in cellulosic materials for both the mechanical and dielectric cases. The samples used were spruce sul®te pulp (containing cellulose I), regenerated cellulose (containing cellulose II), and a cellulose III, prepared by immersing the sul®te pulp in liquid ammonia for 24 h. They used a torsion pendulum for the mechanical measurements and a capacitance bridge for the dielectric measurements. For the latter, on dry samples, after plotting loss factor versus temperature, Klason and KubaÂt found a loss maximum at 60 C and 1 kHz. The peak temperature increased with frequency and drying intensity, reaching about 20 at 10 kHz for the most intensely dried samples. The dielectric loss peak had a counterpart in the mechanical case if a measuring frequency of 1 Hz was assumed, with a maximum at 73 C. The activation energy in either case was found to be 50 kJ/mol. Shinouda and Hanna [63] measured the dielectric constants of dry cotton and cellulose derivatives at frequencies from 0.1 to 12 MHz and temperatures from 0 C to 70 C. They found that for the cellulose derivatives at a given frequency, k 0 fell between the values for the cotton (lowest) and regenerated cellulose (highest). The results for the cotton are shown in Fig. 6. The values of k 0 steadily increased with temperature and decreased with frequency, as expected. Results for the cellulose derivatives behaved in a similar way. Values for the dielectric loss were not reported. The dielectric results, together with infrared and speci®c volume measurements, were explained in terms of the nature of the side groups and the degree of hydrogen bonding. It was argued that in cotton ®bers the OH groups are relatively strongly bonded, whereas in regenerated cellulose they are more weakly bonded, resulting in a higher dielectric constant. In general, for dry, low ash paper at ordinary temperatures and low frequencies (10 Hz to 10 kHz), little dispersion is exhibited. In some instances where effects have been noted, it is likely that water or some other impurity was present, or excessive ions in the system led to the loss [12]. Maxima in the dielectric loss factor have been reported when a DC bias voltage was applied simultaneously with the AC measuring ®eld [48]. These have been attributed to impurities. Salt Effects The effects of salts on the dielectric properties were alluded to earlier. Delevanti and Hansen [12] demonstrated that metallic ions were responsible for much of the loss factor in paper at low frequencies. Driscoll [13] measured the loss factor of boxboard samples that had been soaked in various salt solutions. Very high loss factors were reported for samples soaked in tap water or alum in

350

Baum

Fig. 6 Variation of the dielectric constant k 0 of cotton with temperature and frequency. (From Ref. 62.)

tap water in comparison with deionized water. The effect decreased signi®cantly with increasing frequency, suggesting that no salt effect would be observed above several hundred megahertz. Chu and Wyslouzil [11] examined the addition of salt and alum to paper at three microwave frequencies (10, 21, and 33 GHz). They reported a small effect due to the impurities at the lowest frequency but none at the highest frequency. Such a result is to be expected, because any contribution to the dielectric loss of paper due to metallic ions should relax out well before the microwave range. Effects of Water The effects of moisture in the low frequency region have been studied extensively. References 54 and 28 review the literature for paper and wood and hardboard. At frequencies below the microwave region, the dielectric constant of water is around 80, a value signi®cantly higher than the other normal constituents of paper. Thus if water is present in even relatively small quantities, it can signi®cantly in¯uence the dielectric properties of paper. Tsuge and Wada [66] found that above a critical water content a dielectric dispersion, the intensity of which increased linearly with increasing water content, was observed in their printing paper and cellophane samples. They argued that the ®rst water molecules sorbed in polysaccharide are associated with unbound hydroxyl groups, that is, groups that are not involved in inter- or intramolecular hydrogen bonds. At some critical moisture content, however, the sorbed water begins to break

Electrical Properties I

351

up the inter- and intramolecular bonds (in noncrystalline regions) and in so doing allows more and more polar groups to contribute to the polarizability. For paper and cellophane, these critical moisture levels were estimated to be about 3% and 6% water, respectively. It was proposed that below the critical value the water had essentially no effect and the polar groups were unable to contribute signi®cantly because of steric hindrances due to hydrogen bonding. At moisture contents above the critical value, the polar groups are more labile and dielectric dispersion was observed. The higher critical value for cellophane, compared to paper, was assumed to be a result of the greater crystallinity for the cellophane and the presence of other organic materials in the case of the paper. As more and more water is added to paper, a point may be reached where each additional water molecule added behaves as if it were free. It is quite probable that the bound and free water potions behave dielectrically quite differently. In this case, it would be necessary to de®ne different dielectric constants for the bound and free water. Pure water undergoes relaxation at about 22 GHz. When it is sorbed in paper, it is likely that this relaxation would occur at much lower frequencies. That is, the bound and free portions of water would also be expected to behave differently in terms of their frequency dependence. A number of researchers have measured the dielectric constants of paper at higher frequencies [7,9±11,13,14,17,21,34,39,60±62]. The contributions of bound water and free water to the dielectric properties at microwave frequencies have been discussed by Busker [7], Dusoiu [14], and M'Baye and Pellissier [39]. M'Baye and Pellissier estimated the free and bound water concentrations in paper from microwave measurements of dielectric constants using a form of the Kirkwood equation [31]. These authors applied the equation to relate the dielectric constants of paper to the dielectric constants and volume fractions of its constituents. Assuming values for the dielectric constants of bound and free water, the volume fractions of these two constituents were estimated. The results were dependent upon the relative orientation of the electric ®eld vector with the normal vector to the paper sheet. The inhomogeneity of the mixture was ignored in these calculations but properly should be included [21]. At present, however, the appropriate way to do this is not obvious. C.

Dielectric Anisotropy

Introduction Measurements of dielectric constants in the radio frequency and microwave ranges reveal that paper and board are anisotropic with respect to these parameters. Complex dielectric constants must be de®ned for each of the three principal directions: the machine direction, cross-machine direction, and Z or thickness direction A number of researchers have studied this anisotropy [13,15,16,18,21,60,62]. Figure 7 shows the real and imaginary parts of the complex dielectric constant of a machine-made newsprint, measured at 9.6 GHz and a constant moisture content of 9%, as a function of the angular displacement between the electric ®eld and the machine direction of the paper. The highest values are obtained when the ®eld is parallel to the machine direction. This in-plane anisotropy is believed to arise from the ®ber orientation in the machine direction resulting from formation of the sheet on the wire. Fainberg et al. [18] used the dielectric anisotropy to estimate the degree of orientation of molecular chains and aggregates.

352

Baum

Fig. 7 In-plane dielectric constant for machine-made newsprint as a function of the orientation of the electric ®eld with respect to the machine direction. (From Ref. 21.)

Driscoll [13] examined the three-dimensional dielectric anisotropy in machinemade papers at radio frequencies (13.56±100 MHz). Figures 8 and 9 show k 0 and k 00 , respectively, for a boxboard sample as a function of moisture content. In each ®gure, the three curves correspond to the electric ®eld vector aligned with each of the three principal directions in the paper. As noted above, the highest values are obtained when the ®eld is parallel to the ®bers. On the other hand, the lowest values of the dielectric constants are obtained when E is perpendicular to the sheet and thus perpendicular to the majority of ®bers. The data indicate that the sorbed water apparently re¯ects the anisotropy of the ®brous structure, although the anisotropy ratio (the ratio of two dielectric constants at a given moisture) does increase with increasing moisture. Clearly more study of this phenomenon is needed. Thermally Stimulated Depolarization Currents The ionic thermocurrent or thermal depolarization current technique of Bucci et al. [6] (sometimes also referred to as thermally stimulated currents, although this is a different technique altogether) is a method of investigating the dielectric properties of solids. In particular, it provides a means of studying the response of polar groups in a material to an applied electric ®eld. Since the response will also be sensitive to the local environment, the effect of any thermal, mechanical, or electrical treatments that alter the number, character, or environment of the polar groups can be examined. The method involves the application of an electric ®eld to a specimen at a constant temperature (Fig. 10). This causes a polarization of the specimen or, if it has permanent dipole groups, the partial alignment of these groups with the applied

Fig. 8 The real part of the dielectric constant k 0 in the three principal directions of a boxboard sample as a function of moisture content. (From Ref. 13.)

Fig. 9 The imaginary part of the dielectric constant k 00 in the three principal directions of a boxboard sample as a function of moisture content. (From Ref. 13.) 353

354

Baum

Fig. 10 The thermal depolarization technique. (A) At the polarization temperature Tp with no external electric ®eld, the dipoles in the material are randomly oriented. (B) When the polarization electric ®eld ED is applied the dipoles will attempt to align with Ep . This results in a polarization current that decays with time to some steady-state value. (C) With Ep applied the temperature of the material is lowered to temperature T0 such that the dipoles are frozen in their aligned con®guration. In this state, if the ®eld is removed and the sample electrodes are shorted, no current will ¯ow. (D) If the sample is heated, eventually the dipoles can relax back to a random orientation. This causes a current in the external circuit, the depolarization bond. A linear heating rate is not essential but makes analysis of the data simpler.

®eld. The temperature of the material is then lowered while the electric ®eld is maintained. The temperature is reduced to a value suf®ciently low that the polarization due to the permanent dipoles is ``frozen in.'' That is, when the ®eld is removed at the low temperature and the sample electrodes are shorted, the induced (ionic and electronic) polarizations will relax out, whereas the polarization caused by the permanent groups cannot. This is because the time constant for the dipoles to return to a random orientation is very long at the low temperature. If the sample electrodes are connected to an ammeter and the specimen is slowly warmed up, currents will be measured by the meter over certain temperature ranges. Typically, current peaks or bands are observed, the current increasing from zero to a maximum at some temperature and then decaying back to zero. The shape and area of this current band depends upon the magnitude of the dipole moment, the polar group involved, the concentration of groups present, the local environment of the polar group, the temperature at which the specimen was polarized, and the electric ®eld strength. The temperature of the current maximum depends upon the dipole relaxation time, among other things, which in turn is sensitive to the nature of the local environment. For example, a polar group such as the hydroxyl group in cellulose would be expected to behave quite differently in crystalline and noncrystalline regions or in the presence of water or some other plasticizer. A number of theoretical treatments of the phenomenon have been published [6,20,45]. McKeever and Hughes [40]

Electrical Properties I

355

analyzed and described the reverse procedure, which involves cooling the specimen in the absence of an electric ®eld and then warming the material in the presence of the ®eld. The thermal depolarization current technique is especially suitable for materials such as cellulose that have polar side groups. A number of researchers have applied the method to cellulosic ®bers and their derivatives. Baum [3] studied dry regenerated cellulose ®lms in the temperature range 100 C to ‡50 C and reported depolarization bands at 58, 2, and 25 C. the most intense band was that at 2 C and was shown to be caused by the presence of a glycerol plasticizer in the celllulosic ®lm. This band completely disappeared when the glycerol was extracted. The low temperature band was not in¯uenced by the presence of the plasticizer and was thought to arise from the rotation of the primary hydroxyl group on the glucose unit. The band with the maximum at 25 C was believed to be related to the transition that is reported to occur in dry cellulose at this temperature. According to Wahba [70], extensive hydrogen bond breakage occurs above the transition temperature. If this is true, the thermal depolarization method should be ideally suited for investigation of such an event. Since the method is sensitive to the density of polar groups and the ease with which they can respond to the external ®eld, the presence of the electric ®eld, which causes changes in the orientation of the dipoles, would prohibit re-formation of the hydrogen bond, especially as the specimen is cooled. Sawatari et al. [57] studied cellulosic powder and some cellulose derivatives in the temperature range from 175 C to ‡120 C using various techniques for drying the specimens. They observed a broad triplet peak at 140 C, which they attributed to the primary hydroxyl group in the amorphous region. The magnitude of a band at 70 C decreased as the drying conditions became more severe and was thus attributed to sorbed moisture. A third peak at 70 C was believed to be related to the motion of the main chain, although in this case the nature of the polar group giving rise to the effect is not clear. Okabe and coworkers [47] used the thermal depolarization method to study interfacial polarization mechanisms in oil-impregnated kraft paper systems. It seems reasonable to assume that in the future this technique will ®nd wider applications with respect to paper systems. SYMBOLS a A A b B C C0 d D D E

constant, Eq. (15) area (m2 ) constant, Eq. (12) jump distance constant, Eq. (12) capacitance (farads) geometric capacitance (F) with vacuum dielectric caliper electric displacement vector (C/m2 ) dissipation factor (unitless) electric ®eld strength (N/C)

356

e ei f G i i IC IL IT J k k0 k 00 kB K L m ms M Ms n ni Ni p Q R t T U Ub Ui w V V0   0  00 0  i  d v s v s sm

Baum

electronic charge (C) electronic charge (C) of ion of species i frequency (Hz) constant, Eq. (12) dummy index … 1†1=2 charging current (A) loss current (A) total current (A) steady-state current density (A/m2 ) complex dielectric constant real part of complex dielectric constant imaginary part of complex dielectric constant Boltzmann's constant constant, Eq. (16) electrode separation distance (m) constant, Eq. (10) constant, Eq. (11) moisture content moisture content at saturation density of charge carriers (cm 3 ) density of charge carriers of ion species i (cm 3 ) concentration of generating sites of ion species i (cm 3 ) constant, Eq. (15) electric charge (C) electric resistance (ohms) time (s) temperature (K) activation energy at constant ®eld strength (eV) activation energy for mobility (eV) ionization energy to liberate ion species i (eV) angular frequency (s 1 ) voltage (V) voltage at time zero (V) dielectric loss angle complex permittivity (F/m) real part of complex permittivity (F/m) imaginary part of complex permittivity (F/m) permittivity of free space (F/m) carrier mobility [cm2 =…V s†Š carrier mobility of ion species i [cm2 =…V s†Š phase angle density (g=cm3 ) volume resistivity (ohm cm) surface resistivity (ohms) volume conductivity (mho/cm) surface conductivity (mho) conductivity at moisture saturation (mho/cm)

Electrical Properties I

357

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.

Amborski, L. E. (1962). Structural dependence of the electrical conductivity of polyethylene terephthalate. J. Polym. Sci. 62:331±346. Barker, R. E., and Thomas, C. R. (1964). Effects of moisture and high electric ®elds on conductivity in alkali-halide-doped cellulose acetate. J. Appl. Phys. 35(11):3203±3215. Baum, G. A. (1973). Thermal depolarization currents of regenerated cellulose ®lms. J. Appl. Polym. Sci. 17:2855±2866. Brodie, I., Dahlquist, J. A., and Sher, A. (1968). Measurement of charge transfer in electrographic processes. J. Appl. Phys. 39(5):1618±1624. Brown, J. H., Davidson, R. W., and Skaar, C. (1963). Mechanism of electrical conduction in wood. Forest Prod. J. 13(10):455±459. Bucci, C., Fieschi, R., and Guidi, G. (1966). Ionic thermocurrents in dielectrics. Phys. Rev. 148(2):816±823. Busker, L. H. (1968). Measurement of water content above 30% by microwave absorption methods. Tappi 51(8):348±353. Calkins, C. R. (1950). Studies of dielectric properties of chemical pulps. 3. Dielectric properties of cellulose. Tappi 33(6):278±285. Chene, M., Coumes, A., and Lafaye, F. (1965). Cellulose permittivity in the 9000 megahertz wavelength band. C. R. Acad. Sci. (Paris) 260:3632±3635 (in French). Chene, M., Revlo, N., Pellissier, J. P., and Mesnard, G. (1967). Interferometric measurement in band X of the dielectric constant of paper and its application to moisture content determination. Cellulose Chem. Tech. 1:597±600 (in French). Chu, F. Y., and Wyslouzil, W. (1977). Frequency dependence of microwave moisture measurements of paper. Tappi 60(10):144±145. Delevanti, C., and Hansen, P. B. (1945). Studies of dielectric properties of chemical pulps. 1. Methods and effects of pulp purity. Paper Trade J. 121(26):25±33. Driscoll, J. L. (1976). The dielectric properties of paper and board and moisture pro®le correction at radio frequency. Paper Tech. Ind. (April):T42±T46. Dusoiu, N. (1975). Decrease of high-frequency dielectric hysteresis by water desorption at different temperatures. C. R. Acad. Sci. B280:(24):777±779. Dusoiu, N. (1976). Dielectric anisotropy of stacks of paperboard and paper in the microwave range. Rev. Roum. Chim. 21(4):563±570. Dusoiu, N., Balanescu, G., and Liviu, T. (1976). Dependence of dielectric permittivity of paperboard and paper on density. Celul. Hirtie 25(1):34±37. Dusoiu, N., Balanescu, G., and Liviu, T. (1976). Method of measuring the complex permittivity of paperboard and paper in the microwave range. Celul. Hirtie 25(2):57±63. Fainberg, E. Z., Eifer, I. Z., and Mikhailov, N. V. (1966). Electric anisotropy of regenerated cellulose ®bers. Khim. Volokna 4:38±41. [ABIPC 37:541 (1967).] Friedrich, R. E., and Chiu, T. T. (1970). Comparison of AC and DC methods of measuring conductivities in electrophotographic papers. Tappi 53(2):382±384. Gross, B. (1975). On the analysis of thermally stimulated depolarization effects. J. Phys. D: Appl. Phys. 8:L12±L128. Habeger, C. C., and Baum, G. A. (1983). The microwave dielectric constant of waterpaper mixtures: The role of sheet structure and composition. J. Appl. Polym. Sci. 28:969± 981. Hanneson, J. E., Raman, R., and Hart, J. (1971). Electrical conductivity in tissue paper. Tappi 54(6):955±958. Hearle, J. W. S. (1952). The electrical resistance of textile materials: A review of the literature. J. Textile Inst. 43:P194±P223. Hearle, J. W. S. (1953). The electrical resistance of textile materials: The in¯uence of moisture content. J. Textile Inst. 44:T117±T143.

358

Baum

25. Hearle, J. W. S. (1953). The electrical resistance of textile materials: The effect of temperature. J. Textile Inst. 44:T144±T154. 26. Hearle, J. W. S. (1953). The electrical resistance of textile materials: Miscellaneous effects. J. Textile Inst. 44:T155±T176. 27. Hearle, J. W. S. (1953). The electrical resistance of textile materials: Theory. J. Textile Inst. 44:T177±T198. 28. James, W. L. (1975). Dielectric properties of wood and hardboard: Variation with temperature, frequency, moisture content, and grain orientation. USDA Forest Service Res. Paper FPL 245, Madison, WI. 29. Kane, D. E. (1955). The relationship between the dielectric constant and water-vapor accessibility of cellulose. J. Polym. Sci. 18:405±410. 30. Karasz, F. E., ed. (1972). Dielectric Properties of Polymers. Plenum, New York. 31. Kirkwood, J. G. (1939). The dielectric polarization of polar liquids. J. Chem. Phys. 7:911±919. 32. Klason, C., and KubaÂt, J. (1976). Thermal transitions in cellulose. Svensk Papperstidn. 79:494±500. 33. Kosaki, M., Sugiyama, K., and Ieda, M. (1971). Ionic jump distance and glass transition of polyvinyl chloride. J. Appl. Phys. 42(9):3388±3392. 34. Kumar, A., and Smith D. G. (1976). The measurement of the complex permittivity of paper at microwave frequencies. Tappi 59(1):149±151. 35. Lengyel, G. (1966). Schottky emission and conduction in some organic insulating materials. J. Appl. Phys. 37(2):807±810. 36. Lin, R. T. (1965). A study on the electrical conduction in wood. Forest Products J. 15(11):506±514. 37. Lin, R. T. (1973). Wood as an orthotropic material. Wood Fiber 5(3):226±236. 38. Lowe, G. R., and Baum, G. A. (1979). Electrical conductivity of single wood pulp ®bers. Tappi 62(6):87±89. 39. M'Baye, K., and Pellissier, J. P. (1975). Determining the nature of water sorbed on cellulosic ®bers through microwave measurements of dielectric constants. Rev. ATIP 29(2):51±54 (in French). 40. McKeever, S. W. S., and Hughes, D. M. (1975). Thermally stimulated currents in dielectrics. J. Phys. D: Appl. Phys. 8:1520±1529. 41. Mikhailov, G. P., Artyukov, A. I., and Borisova, T. I. (1967). Characteristics of relaxation of cellulose hydroxyl groups at low temperatures. Vysokomol. Soed. 9(B)(2):138±141 (in Russian). 42. Murphy, E. J. (1960). The dependence of the conductivity of cellulose, silk and wool on their water content. J. Phys. Chem. Solids 16:115±122. 43. Murphy, E. J. (1960). The temperature dependence of the conductivity of dry cellulose. J. Phys. Chem. Solids 15:66±71. 44. Murphy, E. J. (1974). High ®eld conduction in native cellulose and its structural implications. J. Coll. Interface Sci. 49(3):442±452. 45. Nedetzka, T., Reichle, M., Mayer, A., and Vogel, H. (1970). Thermally stimulated depolarization: A method for measuring the dielectric properties of solid substances. J. Phys. Chem. 74(13):2652±2666. 46. Norimoto, M., Hayashi, S., and Yamada, T. (1978). Anisotropy of dielectric constant in coniferous wood. Holzforschung 32(5):167±172. 47. Okabe, Y., Yamashita, H., and Amano, H. (1978). Thermally stimulated current in oil impregnated kraft paper systems. Proceedings of the 11th Symposium on Electrical Insulating Materials in Japan, September, pp. 113±116. 48. Olach, O., and Calderwood, J. H. (1977). The effect of simultaneous application of AC and DC voltages on dielectric loss. J. Phys. D: Appl. Phys. 10:L257±L259.

Electrical Properties I

359

49. O'Sullivan, J. B. (1947). The conduction of electricity through cellulose. 1. The conductance of cellulose sheet impregnated with salts. J. Textile Inst. 38:T271±T284. 50. O'Sullivan, J. B. (1947). The conduction of electricity through cellulose. 2. The chemical effects of the current. J. Textile Inst. 38:T285±T290. 51. O'Sullivan, J. B. (1947). The conduction of electricity through cellulose. 3. The mobility of hydrogen and hydroxyl ions in cellulose sheet. J. Textile Inst. 38:T291±T297. 52. O'Sullivan, J. B. (1947). The conduction of electricity through cellulose. 4. The mobility of various ions in cellulose sheet. J. Textile Inst. 38:T298±T306. 53. O'Sullivan, J. B. (1948). The conduction of electricity through cellulose. 5. The effect of temperature. J. Textile Inst. 39:T368±T384. 54. Raman, R., and Walker, S. (1972). Electrical papers: A literature review. Paper Technol. (April):126±132. 55. Rembaum, A., and Landel, R. F., eds. (1967). Electrical conduction properties of polymers. J. Polym. Sci. C. Polym. Symp. 56. Renne, V. T. (1968). In¯uence of inorganic impurities on the dielectric properties of electric insulating paper. Zellstoff Papier 17:343±345. 57. Sawatari, A., Kurihara, H., and Takashima, T. (1978). Thermally stimulated current of cellulose powder. J. Jpn. Wood Res. Soc. (Mokuzai Gakkaishi) 24(4):224±229 (in Japanese). 58. Seidman, R., and Mason, S. G. (1954). Dielectric relaxation in cellulose containing sorbed vapors. Can. J. Chem. 32:744±762. 59. Seitz, F. (1940). The Modern Theory of Solids. McGraw-Hill, New York. 60. Servant, R., and Cazayus-Claverie, J. (1957). Dielectric anisotropy of paper at 3400 megahertz. The in¯uence of moisture. Comptes Rendus 245:509±511 (in French). 61. Servant, R., and Gougeon, J. (1960). Birefringence and rectilinear dichroism of paper at 3000 MHz. Comptes Rendus 242:2318±2320 (in French). 62. Servant, R., and Weever, J. W. (1960). Dielectric anisotropy of boards at 3000 megahertz. J. Phys. Rad. 21:95S±96S. 63. Shinouda, H. G., and Hanna, A. A. (1977). Dielectric and infrared study of some cellulose derivatives. J. Appl. Polym. Sci. 21:1479±1488. 64. Smith, W. E. (1965). Determination of the relative bonded area of handsheets by directcurrent electrical conductivity. Tappi 59(1):476±480. 65. Smith, W. E. (1968). An investigation of a method for measuring inter®ber bonding in pulp handsheets based on sheet and ®ber DC electrical conductivities. Ph.D. Dissertation, North Carolina State University, Raleigh. 66. Tsuge, K., and Wada, Y. (1962). Effect of sorbed water on dielectric dispersion of cellulose at low frequencies. J. Phys. Soc. (Jpn.) 17(1):156±164. 67. Venkateswaran, A. (1965). Formulas for the dielectric constant and dissipation factor of mixtures and their application to the cellulose system. J. Appl. Polym. Sci. 9:1127±1138. 68. Venkateswaran, A., and Van den Akker, J. A. (1965). Effect of ethylamine treatment on the dielectric properties and crystallinity of cellulose. J. Appl. Polym. Sci. 9:1149±1166. 69. Verseput, H. W. (1951). Studies of dielectric properties of chemical pulps. 4. The relationship between the dielectric constant and crystallinity of cellulose. Tappi 34(12):572± 576. 70. Wahba, M. (1968). the effect of drying and of temperature on the infrared spectra of regenerated cellulose ®lms, in relation to the second order transition of cellulose around 25 C. Arkiv Kemi 29(32):395±413. 71. Wilcox, P. (1972). A dielectric loss model based on interfacial electron tunneling. Can. J. Phys. 50:912±924. 72. Wintle, H. J. (1973). Absorption current, dielectric constant, and dielectric loss by the tunneling mechanism. J. Appl. Phys. 44(6):2514±2519.

9 ELECTRICAL PROPERTIES: II. APPLICATIONS AND MEASUREMENT METHODS SAMI SIMULA* The Finnish Pulp and Paper Research Institute (KCL) Espoo, Finland

I. Applications A. Electrophotography B. Electrical Grade Papers C. Static Electricity II. Measurement Methods A. Resistivity Measurements B. Charging and Charge Decay Measurements C. Dielectric Testing References

I.

361 361 365 372 375 375 380 381 385

APPLICATIONS

A.

Electrophotography

Electrophotography is the technology used in virtually all commercially available copiers. Laser printers and some digital printing machines are also based on this technology. The speed of electrophotographic printers is continuously increasing, and color printing is becoming more popular. High-speed color electrophotography places stringent requirements on the printability and runnability of paper. There are six steps in electrophotography, as shown in Fig. 1. In a typical print engine they are (1) charging of the photoconductor, (2) creation of an electrostatic latent image by discharging non-image areas with light, (3) development of the latent 

Current af®liation: Okmetic Oyj, Vantaa, Finland. 361

362

Simula

Fig. 1 The six steps of electrophotography: (1) Charge; (2) exposure; (3) development; (4) transfer; (5) fusing; and (6) cleaning. Paper path is indicated by the line with arrows.

image by placing pigmented charged powder (toner) on the charged areas, (4) transfer of toner image onto paper by an electric ®eld, (5) ®xing of the image on paper by heat and pressure, and (6) cleaning the toner residue from the photoconductor and discharge of the residual charge. It is the fourth step, toner transfer, that makes the electrical properties of paper important. In the transfer step the paper is charged by a high voltage corona (or a bias roll) so that the surface of the paper attains a charge opposite that of the toner. A strong electric ®eld is formed between the photoconductor and the paper, and toner particles are transferred to the paper surface. The amount of toner that is transferred depends on the electrical properties of the paper. It is therefore important for image quality that these properties be in agreement with the requirements set by the transfer system. In the following section we will discuss the effect of electrical properties on image quality and runnability. The reader interested in electrophotography in general is referred to books written on the subject [35,46]. Requirements on the Electrical Properties of Paper The electrical properties of paper used in electrophotographic copiers and laser printers affect both image quality and runnability. The requirements on the electrical properties of paper become more stringent with increases in printing speed and with the use of colors. Speed adds dynamics to the electrostatic transfer system and increases vulnerability to static electricity. The transfer of four toner layers instead of one complicates the prediction of the electrical behavior of paper in the transfer subsystem. Hence, knowledge of the meaning of the electrical properties of paper is becoming more valuable. A schematic view of the toner transfer process is shown in Fig. 2. Toner is transferred from the photoconductor drum to paper by an electric ®eld. The transfer ®eld is generated by charging the paper with the help of a corona wire (or a bias roll). The paper is charged more than the photoconductor so that the toner particles attach to the surface of the paper. Contact between paper and toner is necessary at the transfer step. If there is an air gap between toner and paper, an image quality deteriorating discharge could occur due to air breakdown.

Electrical Properties II

363

Fig. 2 A schematic view of the toner transfer process. A high voltage corona charges the paper, generating a transfer ®eld between the paper and photoconductor, and toner is transferred to the paper surface.

There are usually three parameters used to characterize the electrical properties of paper used in electrophotography. They are surface resistivity s , volume resistivity v , and dielectric constant k (relative permittivity) (Chapter 8 of this volume). Surface resistivity is usually deemed the most important one because it is responsible for charge relaxation in the transfer nip caused by lateral conduction effects [15,16]. Sometimes a parameter called the charge relaxation time constant is used. It is the product of surface resistivity and dielectric constant and has been reported to give a better correlation with transfer ef®ciency than surface resistivity alone [8]. In blackand-white electrophotography, knowledge of surface resistivity is adequate for controlling the electrical properties of paper. As a rule of thumb it can be said that the higher the transfer ®eld, the better the transfer ef®ciency. The magnitude of the transfer ®eld depends on the resistivity and dielectric constant of the paper. The sensitivity of the transfer ®eld to the resistivity and dielectric thickness of paper (paper thickness divided by permittivity) can be divided into three regions according to the resistivity value as follows [16]. At high resistivity the charge decay time is much longer than the time spent in the electric ®eld of the transfer nip. In this resistivity region the system is insensitive to variations in resistivity because charge decay is slow relative to the velocity of paper. However, the dielectric constant has a small effect on transfer ef®ciency. At intermediate resistivities, when the charge decay time is comparable to the charging time in the nip, the transfer ®eld is very sensitive to variations in paper resistivity but insensitive to the dielectric thickness. Finally, at low resistivities the charge is very quickly relaxed and the transfer ®eld is insensitive to both resistivity and dielectric thickness because it is almost zero. The effect of surface resistivity on print density can be seen in Fig. 3 [19]. It suggests that paper resistivity should be in the high resistivity region. However, one should keep in mind that the value of the high resistivity threshold depends on both the machine speed and the transfer ®eld width. The time spent in the transfer ®eld is the width of the nip divided by the velocity of the paper. A simple physical model for electrostatic transfer was given by Yang and Hartmann [47]. According to their model, transfer ef®ciency can be maximized by minimizing the dielectric thickness of paper. Hence, transfer ef®ciency should increase with increasing dielectric constant and decreasing thickness. However, contrary to the latter statement of the model it has been found that transfer ef®ciency increases slightly with increasing paper thickness [35]. The reason is the porous structure of paper. Ions created by the transfer corotron polarize the paper more

364

Simula

Fig. 3 Print density versus surface resistivity. If the resistivity is lower than a certain threshold value, a dramatic decrease in print density as a function of the surface resistivity is observed. (From Ref. 19.)

effectively if they cannot ¯ow through the paper. This has been experimentally veri®ed. Transfer ef®ciency was seen to increase with decreasing porosity [35]. Some of the smaller toner particles cannot be transferred at all because of the strong, short-range van der Waals type of adhesion forces between the particles and the photoconductor. Hence, the maximum transfer ef®ciency is limited to approximately 85±90%. In addition to image quality the electrical properties of paper have a considerable effect on runnability. As far as paper handling is concerned, resistivity should be as low as possible to facilitate static charge dissipation. Too high a resistivity will cause problems with triboelectric charging of paper in the printer. Because most machines are sheet-fed, static charge can lead to misfeeds, double feeds, jams, stacking problems in the bin, sparking, or even interference with the printer control signals. The problem is most obvious in high speed machines and for transparency ®lms whose resistivities are a couple of orders of magnitude higher than that of paper. Consequently, modern electrophotographic printers, copiers, and printing machines employ alternating current (AC) corotrons for static charge removal. The wide range of current electrophotographic designs makes it impossible to give exact values for the desired electrical properties of paper. To get the best possible paper performance in electrophotography, one should optimize the electrical properties according to machine speed and design (paper path, charging mechanism, etc.). Fortunately, electrical properties are crucial to paper performance only if

Electrical Properties II

365

they are not within certain limits. It is therefore possible to give some general outlines on what kind of electrical properties are desirable in electrophotography. First, for optimal printability the surface resistivity of paper should be higher than the lower resistivity limit of effective toner transfer. Second, to facilitate runnability the charge dissipation time (strongly affected by resistivity) should be kept as low as possible without affecting printability. Third, printability and runnability of paper should be maintained even if the environmental conditions (humidity and temperature) are changing. Minor changes in the transfer ef®ciency have little effect on print quality because it is usually the toner fusing process that determines the quality of the ®nal print.

B.

Electrical Grade Papers

Electrical grade papers are used as an insulating material in capacitors, power cables, power transformers, and some other applications. The main requirements of insulating materials in these applications are low dielectric loss and high dielectric breakdown strength. These requirements can be met by using very highly beaten, specially puri®ed kraft pulp in the making of insulating paper and by proper impregnation with oil [18,22,26,48]. Paper is an inexpensive and reliable insulating material. The advantage in using paper as the insulator can be summarized as follows: Low cost Excellent mechanical properties Uniform thickness attainable over a large area Chemical stability and long life Excellent electrical properties in dry condition High dielectric breakdown strength achievable by oil impregnation The good passage for liquids, allowing good impregnation, provided by the roughness of paper on the microscopic scale Excellent self-healing properties The major disadvantages of paper are its high dielectric loss, especially at high temperatures and high frequencies, and relatively large thickness that increases capacitor volume. Paper has been replaced to a great extent by insulators made of plastics such as polyethylene terephthalate, polyethylene, polypropylene, and polyester. However, there are applications such as high voltage, direct current (HVDC) power cables in which paper is still a good choice as the dielectric. Paper is sometimes used as the dielectric in capacitors that are used in high voltage, high power applications. To ®ll the voids, paper is impregnated with oil or epoxy. Due to their long history, paper capacitors are inexpensive and reliable. On the other hand, they suffer from limited operating temperature range ( 40 C to 70 C), large losses, and large size. The large size of paper capacitors follows from the fact that the thinning of paper has not been as successful as that of plastic ®lms. Figure 4 shows the composition of two different types of paper capacitors. Dielectric and metal sheets are chosen according to the application, placed on top of each other as indicated in Fig. 4, and wound into a cylinder. The component is completed by making contacts to the electrodes and enclosing the capacitor in a protective casing.

366

Simula

Fig. 4 The composition of (a) a metallized oil-impregnated power capacitor and (b) a metallized epoxy-impregnated electromagnetic interference (EMI) suppression capacitor. (MP ˆ metallized paper; PP ˆ polypropylene.)

The desired electrical properties of capacitor paper include high dielectric constant, low dielectric losses, and high dielectric breakdown strength. These requirements can be ful®lled by having good chemical and mechanical purity, no voids, and high density. For this purpose, specially puri®ed, very highly beaten kraft pulp is used. Some chemicals may be added to the pulp to increase the dielectric constant. Oil impregnation is used to improve the dielectric strength. An important feature of metallized ®lm paper capacitors is their ability to selfheal [22]. When a local dielectric breakdown occurs, a conductive channel is formed between the electrodes. Current rises very rapidly to a high value up to the point where the dielectric and the metal near the channel vaporize. The effect is such that the dielectric breakdown insulates itself, i.e., self-heals. The self-healing ability of a capacitor depends on several factors including electrode thickness and chemical composition of the dielectric. Paper has proven to be the best quali®ed dielectric to self-heal. Dielectric Loss Because dielectric losses are the major problem for electrical grade papers we shall deal with them ®rst. Dielectric loss in terms of power consumption P (W) is de®ned by the equation P ˆ IL V ˆ !CV 2 tan  where IL ˆ loss current V ˆ applied voltage

…1†

Electrical Properties II

367

! ˆ 2f ˆ angular frequency C ˆ capacitance tan  ˆ dielectric loss tangent (dissipation factor) as de®ned in Chapter 8 of this volume. To reduce dielectric loss we must lower tan . The dielectric loss tangent is the ratio of the complex and real parts of the permittivity of the dielectric, tan  ˆ  0 = 00 where

 0 ˆ the real part of (complex) permittivity ( ˆ  0  00 ˆ the complex part of permittivity

…2† i 00 , where i ˆ

p 1)

as also shown in Chapter 8 of this volume. Both  0 and  00 are frequency-dependent. In paper the complex part of permittivity can be attributed to ionic conduction and polarization losses. Polarization losses are due to rotation and oscillation of polar groups (dipoles) in the constituents of the ®bers. Conduction losses are dominant at high temperatures, whereas polarization is the cause of dielectrics loss at low temperatures. In the following we shall brie¯y review the effect of lignin content, hemicellulose content, carboxyl group content, crystalline fraction of cellulose, and metallic ion content on the dielectric loss tangent. Effect of Lignin Content The effect of lignin is to increase dielectric losses because lignin is essentially polar. However, some studies show that excessive removal of lignin (to below 2%) can cause an increase in the loss tangent as shown in Fig. 5 [40]. Effect of Hemicellulose Content The effect of hemicellulose content depends on the temperature [41]. At 80 C the loss tangent is independent of hemicellulose con-

Fig. 5

Relations between lignin content and dielectric loss tangent of paper. (From Ref. 40.)

368

Simula

tent, whereas at higher temperatures an increase in the hemicellulose content will result in a lower loss tangent (Fig. 6). However, the effect is reversed below 80 C, and the loss tangent will increase slightly with increasing hemicellulose content. Since the temperature range of interest is below 80 C, hemicellulose content can be low. Effect of Carboxyl Groups Carboxyl groups do not exist in pure cellulose but are present in large quantities in wood pulps. They are powerful ion absorption centers and therefore have a strong in¯uence on the loss tangent of paper. The in¯uence of carboxyl groups is a bad one [38], and it is at its greatest at high temperatures and low frequencies [7]. An interesting theory related to carboxyl groups and the loss tangent has been suggested [39,42]. When the number of carboxyl groups is reduced, metallic ions are liberated, causing an increase in ionic conduction and in the loss tangent. The increase in tan  with decreasing lignin or hemicellulose content at high temperatures could then be attributed to metallic ions liberated from the carboxyl groups contained in lignin and hemicellulose. In¯uence of the Crystalline Fraction The hydroxyl groups of the cellulose chain are believed to have large associated dipole moments when not part of a crystalline region. Hence, a decrease in tan  with increasing crystalline fraction has been observed at room temperature (mineral oil±impregnated paper at 208C in Fig. 7) [39]. However, it is not possible to change the crystalline fraction independently from other properties, i.e., the amount of hemicellulose and lignin decrease with increasing crystalline fraction. It has also been seen that above 100 tan  increased when the

Fig. 6 Relation between pentosan content of prehydrolyzed sulfate pulp and dielectric loss tangent of unimpregnated paper. (From Ref. 41.)

Electrical Properties II

369

Fig. 7 Relation between the amount of crystalline area and dielectric loss tangent of paper. (From Ref. 39.)

crystalline fraction was increased (unimpregnated paper in Fig. 7) [39]. However, it is not possible to change the crystalline fraction independently from other properties, i.e., the amount of hemicellulose and lignin decrease with increasing crystalline fraction. It has also been seen that above 100 C tan  increased when the crystalline fraction was increased (unimpregnated paper in Fig. 7). This could be due to the effect of carboxyl groups as described in the previous paragraph. Effect of Ions At high temperatures and at commercial frequencies, the dielectric loss tangent is dominated by conduction losses due to metallic ions (Chapter 8 of this volume). Even though the number of divalent ions in paper is much greater than the number of univalent ions, the in¯uence of the latter is much greater. The most important ions are the cations Na‡ , Li‡ , and K‡ . Cations are more important for electrical conduction than anions, but in the absence of an effective cation the contribution from anions such as Cl can become signi®cant. Ionic conduction is enhanced by the presence of even minute amounts of moisture in paper. Consequently, ef®cient water removal is very important for electrical grade papers. Dielectric Constant Instead of permittivity it is often much more convenient to use the concept of relative permittivity or dielectric constant. The dielectric constant k of a material is the ratio of the permittivity of the material  and the permittivity of vacuum 0 . Dielectric constant is a dimensionless quantity and a very convenient one because it is always greater than 1 and is below 10 for most polymers. The dielectric constant of paper is typically between 2 and 6 depending on the density, moisture content, and composition of the paper. For dry, low ash papers, somewhat lower dielectric constants have been reported (1.3±1.8 in Table 1 of Chapter 8 of this volume). Studies on dielectric properties have mainly been concentrated on the dielectric loss because it is usually the crucial parameter in applications. However, some

370

Simula

remarks on how the dielectric properties of paper are affected by material and other variables are appropriate in this context also. As shown in Chapter 8 of this volume, Delevanti and Hansen [13] and Calkins [7] found that the dielectric constant of chemical pulps roughly follows the Clausius± Mossotti relation [33], k 1 N  ˆ 0  k ‡ 2 30 M d

…3†

where N0  ˆ polarizability per mole M ˆ molecular weight d ˆ density Moisture content has a substantial effect on the dielectric loss, but its effect on the dielectric constant is less pronounced. The dielectric constant of free water is as much as 80, but the effective dielectric constant of bound water is much less because the dipoles of water molecules cannot align themselves freely with the electric ®eld. At high moisture contents water is expected to have a strong in¯uence on the dielectric constant because there will be several layers of water molecules on the surface of ®bers and they become relatively free. On the other hand, moist paper is a relatively good conductor of electricity, so it is no longer a ``pure'' dielectric. Because the freedom of movement of the polar groups in cellulose is greater in the noncrystalline region than in the crystalline region, the dielectric constant of cellulose decreases with increasing crystalline fraction. As for most polymer materials, the dielectric constant of paper increases with increasing temperature and decreasing frequency [30]. Theoretical aspects of the dielectric constant and its variation in paper are discussed in more detail in Chapter 8 of this volume. Dielectric Breakdown Strength The dielectric breakdown strength, expressed in units of kilovolts per millimeter (kV/mm) or megavolts per meter (MV/m), is the value of the breakdown voltage divided by the thickness of the material. If the electric ®eld is uniform throughout the sample, it is also the magnitude of the electric ®eld that causes the breakdown of the dielectric. It is a parameter of great importance in high voltage, high power applications. Dielectric breakdown phenomena in polymers can be roughly divided into four categories [37]: 1. 2. 3.

Intrinsic breakdown is electronic in nature and depends on the presence of electrons that can migrate through the polymer. Thermal breakdown is related to changes in electrical conductivity with temperature. If at some point heat is generated faster than it is dissipated, a catastrophic local increase in temperature and current can occur. Discharge breakdown is related to the presence of voids in the material. Gaseous breakdown occurs at a lower potential than that at which the solid will break down, and spark discharges can occur. This will lead to the oxidation and carbonization of the walls of the void. Gaseous products of

Electrical Properties II

4.

371

oxidation can be chemically active or build up the pressure in the void, leading to further damage. Electromechanical breakdown is due to mechanical deformation caused by electrical stress.

In addition to the intrinsic breakdown mechanisms listed above, there are other detrimental effects that lower the dielectric breakdown strength. Electrochemical effects, corona degradation, or mechanical stress can create conducting paths, degrade mechanical properties, or decrease the intrinsic dielectric breakdown strength. For insulating material used in capacitors or other high voltage applications, the third mechanism, discharge breakdown, is the most serious problem. If all the gas pockets within the dielectric cannot be removed by impregnation or some other method, they become the prime cause of dielectric breakdown and a source of electrical noise within the insulation. Partial discharges leading to dielectric breakdown can result even at much lower applied voltages than the inception voltage for gaseous breakdown, because the electric ®eld within the void is larger than within the dielectric itself [37]. Because paper is by nature a porous material it has to be impregnated with insulating oil or varnish if it is to be used in high voltage applications. An impregnant that has a higher dielectric strength than air or vacuum is used to ®ll the voids, thus rendering the paper more resistant to dielectric breakdown. Paper properties that have the most signi®cant effect on dielectric breakdown strength are thickness, density, air impermeability, and the variation (uniformity) of these parameters. The effects of these parameters on dielectric breakdown strength are summarized in Table 1. Conducting inclusions have a detrimental effect on dielectric breakdown strength because they provide low resistance areas that provide possible conducting paths at high voltages [27]. Beating of the ®bers not only increases the density but also ®lls many of the voids with ®nes, thereby increasing the barrier effect of paper and it dielectric strength. Various factors that affect the dielectric properties of paper are listed in Table 2. Stages in the manufacturing process of paper that affect the dielectric properties of paper are listed in Table 3 with the corresponding properties that are changed. Table 1 Summary of the Effect of Sheet Parameters on Dielectric Breakdown Strength Sheet property Air impermeability Density Thickness Uniformity

Effect on dielectrics breakdown strength AC short-term, DC, and impulse breakdown strengths will increase with increasing air impermeability. AC long-term breakdown strength is not affected. AC-short term, DC, and impulse breakdown strengths should increase with density (some contradictory reports exist). AC short-term and impulse breakdown strengths should increase with decreasing thickness, but there is evidence that thickness has no effect on the breakdown strength. Impulse breakdown strength increases with improving uniformity.

372

Simula

Table 2 Factors that In¯uence the Dielectric Properties of Paper: In¯uence of Wood Material, Pulp, and Papermaking Conditions Dielectric property Dielectric constant, k

Paper property 0

Dielectric loss tangent, tan 

Dielectric breakdown strength

C.

Apparent density of paper Sheet structure (®ber orientation) Crystalline cellulose Fundamental elements in pulp (lignin, hemicelluloses, etc.) Ionic conduction losses Inclusion of other materials: organic ions; inorganic ions Adsorbed ions: carboxyl group etc. in material; morphological structure of ®ber Polarization losses Rotation and oscillation of polar material Fine structure of cellulose Fundamental elements in pulp (lignin, hemicellulose, etc.) Long-term AC breakdown strength Apparent density of paper Barrier effect of paper Inclusion of other materials Dielectric loss Short-term breakdown strength (AC, pulse) Barrier effect of paper; mesh structure of paper Bonding structure of ®bers Morphological structure of ®bers Macroscopic structure of paper Inhomogeneity of paper

Static Electricity

The source of static electricity is an imbalance in the number of electrons and protons in the material. This imbalance causes the material to have either a positive net charge (de®ciency of electrons) or a negative net charge (surplus of electrons). When two different materials are brought into contact, the difference in their chemical potentials causes some of the electrons to move from one material to the other. When the materials are separated again the charge imbalance remains. In insulators this imbalance of charge remains for a suf®ciently long time to be noticed. Static charge can also be generated by rubbing two pieces of the same material together. Even if there is no difference in the chemical potentials, charge and parts of the material are interchanged, resulting in a similar charge imbalance. Charge caused by friction between materials is called triboelectricity. There are three main characteristics to consider: the polarity of the charge, the magnitude of the charge, and the charge decay rate. The polarity of the material, i.e., whether a material will give away or receive electrons, is determined by the so-called triboelectric series [1]. The amount of charge transferred depends on either the

Electrical Properties II

373

Table 3 Various Factors of the Manufacturing Process and Paper Properties Affected by Them Process variable Wood material

Cooking

Washing Re®ning (beating)

Papermaking process

Calendering and supercalendering

Paper property affected Structure of crystalline cellulose Fundamental elements in pulp (lignin, hemicellulose, etc.) Organic resins Carboxyl groups, etc. Morphological structure of ®bers Degree of polymerization Structure of crystalline cellulose Fundamental elements in pulp (lignin, hemicellulose, etc.) Organic resins Carboxyl groups, etc. Morphological structure of ®bers Bonding structure between ®bers Degree of polymerization Inclusion of other materials Inorganic ions (N‡ , K‡ , Ca2‡ , etc.) Organic resins Apparent density Con®guration of ®bers in paper Morphological structure of ®bers Barrier effects of paper Mesh structure of paper (air impermeability) Bonding structure between ®bers Macroscopic structure of paper Inhomogeneity of paper Apparent density Con®guration of ®bers in paper Barrier effects of paper Mesh structure of paper (air impermeability) Bonding structure between ®bers Macroscopic structure of paper Inhomogeneity of paper Apparent density Barrier effects of paper Mesh structure of paper (air impermeability)

difference between the chemical potentials or the number of available surface states for the electrons as well as on the number of electrons available for charge transfer [36]. The charge decay time depends mainly on the resistivity of the material in question. Static electricity can cause a lot of problems in manufacturing, converting, and printing operations if it cannot be properly prevented or eliminated. Some examples of situations where static electricity is generated are illustrated in Fig. 8. To avoid problems with static electricity one can either try to prevent the formation of static charge or try to control it. The formation of static charge depends mainly on the following factors:

374

Simula

Fig. 8

Examples of situations in paper handling where static electricity is generated.

The difference between the chemical potentials of the materials The separation velocity of the materials Contact pressure Roughness of the contacting surfaces Friction coef®cients (rubbing) Environmental conditions (temperature, humidity of air) In practice it is impossible to completely avoid the formation of static charge. Consequently, it is easier to control the formed static charge. There are three basic ways to control static electricity [2]: 1. 2. 3.

Grounding of the charged material Use of antistatic agents in coatings, additives, or sprays Ionization of the air around the charged material

In the case of paper, grounding is not very practical. It is dif®cult to bring all parts of paper into intimate contact with the ground. This would be necessary because of the high resistivity of paper. On the other hand, when paper is separated from the grounding object, charge is again created (see Fig. 8). Hence, grounding of paper is rendered ineffective by its high resistivity unless the resistivity is lowered by keeping the relative humidity high. In fact, some printing presses employ steamers whose function is to reduce static electricity problems by enhancing the dissipation of static charge. The function of antistatic coatings and sprays is to increase the surface conductivity in order to let the charge leak into the ground. Their use is somewhat limited by cost, the need for repeated application, and the possible effects on the ®nal product. Conductive additives such as salts or polyelectrolytes are added to the surface size of copy papers to increase their conductivity. This is necessary because they need to have low moisture contents at relatively low resistivities. The most practical way to control static is air ionization, and it is widely employed in paper handling. There are three types of static neutralizers based on

Electrical Properties II

375

air ionization: induction devices, high voltage ionizers, and alpha emitters (nuclear) [10]. A summary of the methods and their advantages and disadvantages are given in Table 4.

II.

MEASUREMENT METHODS

A.

Resistivity Measurements

There are usually two kinds of resistivity values given for paper: volume (or bulk) resistivity and surface resistivity. Due to the anisotropy of paper, resistivity has a directional dependence. Usually the term volume resistivity is used to denote resistivity measured in the Z direction, i.e., the thickness direction of paper. Volume resistivity (Ohm cm) is a material parameter and does not depend on paper thickness. The surface resistivity of paper is measured along the surface and has the units of Ohm/&. The square in the denominator is used only to separate surface resistivity values from resistances. It is equally correct to use ohms for surface resistivity. One should bear in mind that electrical conduction is not entirely limited to the surface layer of the paper unless its resistivity is signi®cantly lower than that of the rest of the paper. Standard Method The most widely adopted method for the measurement of volume and surface resistivities in the paper industry is the method based on ASTM Standard D 257 [5]. This standard is relatively loose in the sense that it allows a considerable amount of latitude in the selection of electrode materials and con®gurations as well as other measurement parameters. In the following the most common measurement setup is discussed. The usual electrode con®guration has two circular electrodes and a ring electrode around one of the circular electrodes (Fig. 9). Sometimes conductive rubber is used in one or two of the electrodes to improve the contact between the specimen and the electrode. When volume resistivity is measured the ring electrode acts as the guard electrode, and when surface resistivity is measured the upper circular electrode is the guard electrode. The purpose of the guard electrode is to minimize errors due to surface conduction while volume resistivity measurements are being made and those due to volume conduction during measurements of surface resistivity. The resistance of the specimen is obtained by measuring the current ¯ow through the specimen with a sensitive electrometer or picoammeter. The values of the volume and surface resistivities can then be calculated from the measured resistance, because the geometry of the current path is known [5]. Due to the very small current, proper shielding must be provided to minimize electromagnetic interference from the environment. Because we are dealing with a material that is essentially an insulator, several precautions have to be taken. The humidity of ambient air should be controlled within 0:5% RH because the resistivity of paper can change by as much as one order of magnitude when the moisture content of the paper changes by 1% [11]. Paper samples have to be conditioned for a suf®ciently long time that the equilibrium moisture content can be reached. The measurement temperature has to be controlled accurately also. The reason is that resistivity has an exponential temperature depen-

High voltage applied to a row of needle point ionizers ionizes air Alpha particle (210 Po) radiation ionizes air molecules.

High voltage ionizers

Alpha emitters

Source: Refs. 2 and 10.

Conductive bristles that induce an electric ®eld between the tips and the charged material

Principle

Easy installation Does not require an external power source Maintenance not required

Inexpensive Long life Does not require an external power source Moderate cost High ef®ciency Long life

Advantages

Types of Static Control Devices and Their Advantages and Disadvantages

Induction devices

Type

Table 4

Power source needed Regular maintenance required EMI (electromagnetic interference) Ozone formation Limited capacity (high charges, fastmoving webs) Needs annual replacement due to radioactive decay

Works only for high potentials and reduces them to 5 kV Requires maintenance

Disadvantages

376 Simula

Electrical Properties II

377

Fig. 9 Electrode arrangements for (a) volume resistivity and (b) surface resistivity measurement.

dence in paper [32]. The equilibrium moisture content of paper also depends on temperature; though not much at the usual measurement temperatures. The recommended applied voltage in the standard is 500  5 V unless otherwise speci®ed. The speci®cation of voltage is necessary because the volume resistivity value may depend rather strongly on the applied voltage as shown in Fig. 10. However, 500 V is rarely used in paper testing because it easily leads to current over¯ow, especially in the volume resistivity measurement (see Fig. 10). Another important parameter speci®ed in the standard is the electri®cation time, which has been chosen to be 60 s. This time is rather suitable for paper samples because the time dependence of the resistivity value is usually very small after 1 min. Shorter (or longer) electri®cation times can also be used if the time is quoted. Using speci®c testing equipment (the Keithley 6105 resistivity adapter), ASTM D 4949 recommends 90 V and 60 s for surface resistivity measurements of plain papers [6]. A recommendation of ASTM D 257 [5] is that with the electrode selection speci®ed a pressure 140±700 kPa should be applied to minimize the effect of contact resistance. If the electrode diameter is 5 cm (a realistic value), then the load corresponding to the recommendation would be between 275 and 1400 N. This means that the weight of the top electrode should be 27±135 kg. This recommendation is apparently not met in commercial devices. With typical loading of commercial measurement cells, less than 5% of the paper surface covered by the electrode area is in actual contact [20]. On the other hand, such high pressures would cause permanent deformation in paper, which may change the ®ber-to-®ber conductivity from that of an undeformed specimen.

378

Simula

Fig. 10 Surface resistivity (*) and volume resistivity (*) of copy paper as a function of applied voltage. The dotted lines are single exponential decay curves ®tted to the resistivity data. The charging time was 30 s for an HP4339A high resistance meter with an HP16008B resistivity cell.

From the purely physical point of view it can be argued that the whole concept of paper resistivity is questionable, because resistivity (or conductivity) is de®ned by Ohm's law, which states that the current density is directly proportional to the applied electric ®eld and that the proportionality constant is the conductivity of the material. In the case of paper (and other insulating materials), current density does not increase linearly as a function of electric ®eld strength. This means that the very de®nition of conductivity is not valid in the strictest sense of the de®nition. Even worse is the effect of contact resistance, which can lead to great errors in the measurement. Even though the reproducibility of the experiments might be good, the accuracy suffers a great deal from contact resistance. Like resistivity, contact resistance also depends on the moisture content of paper. However, the dependence is unknown and might be different from that of resistivity, making it dif®cult to say which moisture effects are due to paper properties and which are due to contact resistance effects. Nevertheless, by following the de®nitions and recommendations of the standard one can get results that can be reasonably well justi®ed as electrical parameters of the material. The advantages and disadvantages of the standard method are summarized in Table 5. For the reasons mentioned above and because the recommendations of the standard are not rigorously followed, it is rather dif®cult to compare resistivity values found in the literature. In spite of the shortcomings, this is still the most widely employed method in the paper industry [29]. In the next section, some suggested improvements to the standard method are discussed.

Electrical Properties II

379

Table 5 Advantages and Disadvantages of Using the Standard DC Resistivity Measurement Method Advantages Integrated devices commercially available Ease of operation ``Easy'' data No directional dependence in the surface resistivity measurement (circular electrodes)

Disadvantages Contact resistance Time dependence Voltage dependence Pressure dependence (if pressure too low) Electrode material affects results (charge injection, electrochemical effects) Surface roughness of paper in¯uences results (contact area unknown) Physical concepts not clear (apparent resistivity measured)

Improvements to the Standard Method To improve the accuracy of resistivity measurements, several improvements to the standard method have been suggested. Because the most serious problem in the standard resistivity measurement is the contact resistance, practically all practitioners have aimed at the removal or reduction of the contact resistance. Several authors have suggested improvements to the measurement of surface resistivity [11,12,17,20,21,25,29,34]. Some authors talk about lateral resistivity, which emphasizes the fact that current conduction is not completely limited to the surface of the paper. Perhaps the best account is given by Griesmer [20], who studied different electrode con®gurations, electrode materials, and charging methods (AC, DC, pulse). He also gave recommendations based on his results and presented a new electrode arrangement to be used in surface conductivity measurements. The recommendations of the ASTM standard were also reviewed. The measures that are needed to reduce contact resistance in the surface resistivity measurement to a negligible value according to Griesmer [20] are At least 500 V of applied voltage Adequate pressure Electrode gap of at least 2 cm or more (Fig. 11) Unfortunately, the removal of contact resistance from the volume resistivity measurement is not that simple. The geometry does not allow the distance between the electrodes to be varied, and the thinness of paper is such that the contact resistance in series with the paper resistance is signi®cant. The uncertainty in the contact area of the electrode/paper interface is also a more severe problem than the uncertainty in the surface resistivity measurement. One could use an AC measurement to overcome the contact resistance, but the interpretation of AC data is rather dif®cult due to the capacitive elements involved [20]. A pressure cell was constructed for in situ measurement of volume resistivity [24], but it does not seem to be a practical solution for commercial purposes.

380

Simula

Fig. 11 Lateral resistivity as a function of electrode distance L measured with mercury electrodes. (From Ref. 11.)

Completely new measurement systems have been suggested [12,17,20,25,34]. These include four-probe contacts, piercing electrodes, measurement of transient currents, AC measurements, measurement of the voltage drop along the length of a paper strip, and others. However, none of the methods is as practical a testing method as the standard method, especially if a commercial device is used. Traditional methods are often preferred in the paper industry, especially if the results are reproducible and can be used in paper testing for paper speci®cation. Consequently, the standard method is likely to be the most widely used in the paper industry for a long time. Recently, new methods have been developed for additional characterization of the electrical properties of paper with charging and charge dissipation measurements described in the next section. B.

Charging and Charge Decay Measurements

Some of the most interesting electrical properties are the charging capacity and the charge decay rate of paper. The principle of the measurement apparatus is shown in Fig. 12. Charge is deposited on the surface of the specimen by using a set of high voltage corona wires. The great potential difference between the paper and the wires ionizes the surrounding air, and, depending on the polarity of the corona, either electrons or positive ions are deposited on the paper surface. Paper is polarized due to both the charges deposited on the surface and the potential difference between the paper and the corona wires. After a predetermined charging time the corona wires are rapidly moved aside and the measurement is begun. The potential due to the charge surplus on the surface of the paper is measured with a high speed electrostatic ®eld meter so that it can be measured as a function of time [9]. The measurement parameters of interest are the maximum potential, Vmax , and the decay time to half of the maximum voltage, t1=2 . The maximum surface potential Vmax is related to the ability of a paper to be charged with the help of a high voltage corona and depends on the resistivity and permittivity of the sample. If the resistivity of paper is below a certain limit, the measured Vmax will depend strongly on the resistivity. This is due to the fact that it

Electrical Properties II

381

Fig. 12 The principle of a charge decay measurement. The specimen is charged by a high voltage corona for a predetermined time. When the charging ends, the corona device is rapidly moved aside and a fast electric ®eld meter measures the potential caused by the charges at surface of the specimen.

takes some time ( 20 ms) to move the corona wires and start the measurement. If the resistivity is low, some of the charge decays before the actual measurement begins. Above the high resistivity limit, Vmax is nearly constant as a function of resistivity. According to theory, the measured potential V should be a single exponential decay curve if there is only one type of charge carrier present. In this case the measured curve should follow the equation [23] V ˆ V0 e

t=

…4†

where the time constant is  ˆ v k. This implies that it should be possible to determine the effective dielectric constant of a material by measuring its resistivity and charge decay rate. Unfortunately, the measured curve is not a single exponential decay but contains perhaps several decay constants that are dif®cult to separate from each other. The determination of the ``true'' resistivity is also very dif®cult, as discussed earlier. Charge decay times decrease exponentially with the increase of moisture content, which is in agreement with the assumption of resistivity dependence. Charge decay curves for different papers and a polyethylene terephthalate (PET) ®lm are shown in Fig. 13. The electrostatic charging method discussed above has gained increasing interest recently due to the fact that it charges the paper in much the same way as in electrophotography, i.e., with a corona device. It is also suitable for the study of problems caused by static electricity because it gives an idea of the charge dissipation characteristics of a material. Unfortunately, the tendency of a material to get charged in contact with another material cannot be examined in this manner. For this purpose, charging should be accomplished by rubbing or passing the paper through a nip. However, these methods are less satisfactory due to poor reproducibility in the amount of induced charge. C.

Dielectric Testing

Measurement of Dielectric Constant and Dielectric Loss The technique used in the measurement of dielectric properties of paper depends on the frequency range of interest. In the range from 1 Hz to 1 MHz, bridge techniques are preferred. The Schering bridge and transformer bridge are variations of the Wheatstone bridge used for the measurement of DC resistances. From 1 to 500 MHz, resonant circuits are more appropriate than bridges, and at even higher frequencies one must resort to transmission line techniques. Below 1 Hz, time domain techniques are used [28,44,

382

Simula

Fig. 13 Charge decay curves at 40% RH and 23 C for copy paper, supercalendered (SC) paper, and polyethylene terephthalate (PET) ®lm.

45]. Some reviews are available for readers interested in the details of these techniques [44]. The classical dielectric cell used for the measurement of isotropic solids is a uniform ®eld, parallel-plate capacitor ®lled with the material under test [14]. However, such simple requirements are surprisingly hard to meet even if the fringing ®elds at the edges of the electrodes are neglected. In the case of a thin, microscopically rough sample like paper there is usually bound to be some air trapped in the test cell, which will cause an error in the measurement [44]. If the plates are not completely parallel the effect is even worse. Even if one is using commercial equipment one should be absolutely sure that the electrodes are parallel. The simplest way to measure the dielectric constant of paper is to place the specimen under test between the electrodes of a dielectric cell such as the one described above and measure the parallel resistance R and parallel capacitance C of the cell as shown in Fig. 1, Chapter 8 of this volume. Then the sample is removed and the capacitance of the empty cell, C0 , is measured. The dielectric constant is simply [14] k ˆ C=C0

…5†

and the loss tangent is tan  ˆ 1=2fCR

…6†

where f is the measurement frequency in hertz (Chapter 8 of this volume, Section II). However, one should be very careful not to leave an air gap between the specimen and the electrode, for that could cause a great error (easily as much as 40%) in the result [44]. There are commercial material analyzers available with test ®xtures for dielectric testing for a variety of frequency ranges. They will directly calculate permittivity

Electrical Properties II

383

parameters such as dielectric constant, real and complex parts of dielectric constants, and the loss tangent. As with all electrical measurements concerning paper, special attention should be given to sample conditioning and humidity and temperature control during the measurement. Useful guidelines and measurement terminology for the measurement of dielectric constant and dielectric loss can be found in ASTM Standard D 150 [4]. Measurement of Dielectric Breakdown Strength The dielectric breakdown strength of insulators is usually measured for AC voltage, DC voltage, or impulse voltage. One also has to select how the voltage will be applied, i.e., whether one will make a short-term, step-up, or long-term test. The voltage and how it is applied should be chosen according to the application. In a short-term test the voltage is increased at some rate to a high value and kept there for a predetermined time. The voltage is chosen in such a way that a weak specimen will fail within 10 s or so. In the step-up test the voltage is rapidly increased from zero to about half the breakdown voltage measured with the short-term test. The voltage is then increased stepwise and held constant at each voltage level for some predetermined time, e.g., 1, 10, or 30 min. This will allow charges to redistribute before the voltage is increased again. Long-term tests are, in a way, endurance tests. A voltage is applied to the specimen until it breaks down. The time needed for breakdown is measured at several voltages to obtain a time to breakdown versus voltage curve called the V±t curve (Fig. 14) [31]. V±t characteristics are usually measured for AC voltage.

Fig. 14

Voltage±time characteristics of oil-impregnated pressboard. (From Ref. 31.)

384

Simula

Fig. 15 In¯uence of electrode area on dielectric breakdown strength of insulating paper. (From Ref. 43.)

After the choice of the test voltage and its application method, one has to choose the electrode to be used. The probability of an imperfection lying under the electrode during the test increases with electrode size. Hence, the dielectric breakdown strength decreases with increasing electrode size. However, this probability will increase only up to some point, after which the dielectric breakdown strength stays constant. In some tests it was found that the dielectric breakdown strength did not decrease above a 6 cm2 electrode area (Fig. 15) [43].

Fig. 16 Comparison of AC breakdown strength of mineral oil±impregnated paper between (a) an electrode speci®ed by the Japanese Industrial Standard JIS C-2111 and (b) one having an epoxy-reinforced edge.

Electrical Properties II

385

Uniformity of the electric ®eld applied to the specimen is also important. The size and shape of the electrodes used in the test can have a signi®cant effect on the results. Completely ¯at electrodes cannot be used, because, due to the fringing ®elds, the electric ®eld at the edges of the electrode is higher than the applied voltage implies. Thus, breakdown voltages lower than the real value are obtained. For this reason the edges of the electrodes are rounded. The use of epoxy can further reduce the edge effects as shown in Fig. 16. The standard ASTM D 149 [3] gives some recommendations for testing the dielectric strength of insulating materials. It should be noted that the environmental conditions, especially the relative humidity of air, have a marked effect on the measured dielectric breakdown strength.

ACKNOWLEDGMENT Part of this chapter's content is based on the ®rst edition chapter, ``Electrical Properties: II. Practical Considerations and Methods of Measurement of Electrical Properties,'' written by S. Matsuda.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.

Aaltio, H. J. (1960). Static electricity of paper. Paperi ja Puu 4a:167±184. Anon. (1988). Static causes problems in converting operations. Paper. Film Foil Converter 62(9):74±78. ASTM D149. (1995). Dielectric breakdown voltage and dielectric strength of solid electrical insulating materials at commercial power frequencies. ASTM D150. (1995). AC loss characteristics and permittivity (dielectric constant) of solid electrical insulation. ASTM D257. (1993). DC resistance or conductance of insulating materials. ASTM D4949. (1994). Standard test method for determination of D-C resistivity of writing paper (Keithley method). Calkins, C. R. (1950). Studies of dielectric properties of chemical pulps. 3. Dielectric properties of cellulose. Tappi 33:278±285. Chen, A. C., Borch, J., Garcia, C. E., and Linn, B. J. (1985). Effect of variations in paper parameters on transfer ef®ciency in the electrophotographic process. J. Imaging Sci. 29(1):11±17. Chubb, J. N. (1988). Measurement of static charge dissipation. Electrostatic charge migration. Institute of Physics Short Meeting. March 1988, London, pp. 73±81. Clarke, M. (1987). Static electricity: A wider choice of control. Converter 3: 16±17. Cooprider, T. E. (1968). Resistivity testing methods for conductive base paper. Tappi 51(11):520±527. Cronch, R. D. (1979). Novel method of measuring the base resistivity of electrographic papers. TAPPI Printing and Reprography Conference. Technical Association of the Pulp and Paper Industry, Atlanta, GA, pp. 89±94. Delevanti, C., and Hansen, P. B. (1945). Studies of dielectric properties of chemical pulps. 1. Method and effects of chemical purity. Paper Trade J. 121(26):25±33. Driscoll, J. L. (1976). The dielectric properties of paper and board and moisture pro®le correction at radio frequency. Paper Tech. Ind. 17(2):71±75.

386

Simula

15. Fletcher, G. M. (1993). Lateral conduction in electrostatic systems. Proc. IS&T 9th International Congress on Advances in Non-Impact Printing Technologies, Vol. 1, pp. 157±166. 16. Fletcher, G. M. (1995). Techniques for estimating resistivity ranges of interest in xerographic systems. Proc. IS&T 11th International Congress on Advances in Non-Impact Printing Technologies, pp. 234±237. 17. Friedrich, R. E., and Chiu, T. T. (1970). Comparison of AC and DC methods of measuring conductivities in electrophotographic papers. Tappi 53(2):282±284. 18. Gould, F. R. (1976). Specialty Papers. Noyes Data Corporation, Park Ridge, NJ, pp. 109±112. 19. Green, C. J. (1981). Functional paper properties in xerography. Tappi 64(5):79±81. 20. Griesmer, J. J. (1976). Accurate and meaningful conductivity measurements on electrofax base paper. TAPPI Conference PapersÐReprography, pp. 71±83. 21. Griesmer, J. J. (1978). Electrofax base paper. Accurate and meaningful resistivity measurements. Tappi 61(3):97±99. 22. Jeroense, M. J. P., and Kreuger, F. H. (1995). Electrical conduction in HVDC massimpregnated paper cable. IEEE Trans. Dielectrics Elec. Insul. 2(5):718±723. 23. Jonassen, N., Hansson, I., and Nielsen, A. R. (1979). On the correlation between decay of charge and resistance parameters of sheet materials. Inst. Phys. Conf. Ser. No. 48, Institute of Physics, London, pp. 215±224. 24. Josefowicz, J. Y., Anczurowski, E., Jones, A. Y., and Deslandes, Y. (1981). Bulk conductivity measurement of paper by a new in situ pressure cell. Rev. Sci. Instrum. 52(6):926±932. 25. Josefowicz, J. Y., and Deslandes, Y. (1982). Electrical conductivity of paper: Measurement methods and charge transport mechanisms. In: Colloids and Surfaces in Reprographic Technology. M. Hair and M. D. Croucher, eds. Am. Chem. Soc., Washington DC, pp. 493±530. 26. Kamata, Y., Ohe, E., Endoh, K., Furukawa, S., Tsukioka, H., Maejima, M., Fujita, H., Nozaki, M., Ishizuka, F., and Hyohdoh, K. (1991). Development of low-permittivity pressboard and its evaluation for insulation of oil-immersed EHV power transformers. IEEE Trans. Elec. Insul. 26(4):819±825. 27. Kelck, E., Hinde, R. M., and Wilson, I. O. (1971). Detection and identi®cation of conducting paths in capacitor tissue. IEEE Trans. Elec. Insul. 6(1):32±39. 28. Kumar, A., and Smith, D. G. (1976). The measurement of the complex permittivity of paper at microwave frequencies. Tappi 59(1):149±151. 29. Lim, P. Y. W. (1995). Resistivity of non-impact printing paper. Proc. IS&T 11th International Congress on Advances in Non-Impact Printing Technologies, pp. 401±403. 30. Link, G. L. (1972). Dielectric properties of polymers, I. Polym. Sci., pp 1281±1295. 31. Montsinger, V. M. (1925). Elec. World (Oct.). Summarized in Handbook of Electric Discharge (1963). Denki Gakkai, Tokyo, p. 277 (in Japanese). 32. Murphy, E. J. (1960). The temperature dependence of the conductivity of dry cellulose. J. Phys. Chem. Solids 15:66±71. 33. Parker, T. G. (1972). Dielectric properties of polymers, II. Polym. Sci., pp. 1297±1327. 34. Sapieha, S., and Lepoutre, P. (1982). On the measurement of conductivity in paper. Tappi 65(6):99±101. 35 Schein, L. B. (1993). Electrophotography and Development Physics. Rev. 2nd ed. Laplacian Press, Morgan Hill, CA. 36. Schein, L. B., LaHa, M., and Novotny, D. (1992). Theory of insulator charging. Phys. Lett. A167(1):79±83. 37. Seanor, D. A. (1972). Electrical properties of polymers. Polym. Sci., pp. 1187±1280.

Electrical Properties II

387

38. Shimoyada, T., and Satoh, H. (1958). Thermal deterioration of insulating paper for power cable (interrelationships between dielectric properties and drying conditions). Hitachi Rev. 40:1235±1242 (in Japanese). 39. Take, Y. (1960). Investigation of dielectric losses in insulating paper. Ph.D. Thesis, Tokyo Institute of Technology (in Japanese). 40. Take, Y., Suzuki, Y., and Itohara, F. (1964). The effects of lignin contents on the dielectric properties of insulating paper. J. Inst. Elec. Eng. (Jpn.) 84±4(907):601±608 (in Japanese). 41. Take, Y., Suzuki, Y., and Itohara, F. (1964). Effects of hemicellulose on the dielectric properties of insulating paper. J. Inst. Elec. Eng. (Jpn.) 84±4(907):609±614 (in Japanese). 42. Take, Y., Suzuki, Y., Itohara, F., and Matsunaga, R. (1964). In¯uence of carboxyl group on dielectric properties of insulating paper. Conference Papers of Electrical and Electronic Engineers in Japan No. 495 (in Japanese). 43. Toriyama, Y., and Inada, K. (1957). High Voltage Engineering. Corona Co., Tokyo, p. 92 (in Japanese). 44. Van Roggen, A. (1990). An overview of dielectric measurements. IEEE Trans. Elect. Insul. 25(1):95±106. 45. Von Hippel, A. R. (1954). Dielectrics and Waves. MIT Press, Cambridge, MA. 46. Williams, E. M. (1984). The Physics and Technology of Xerographic Processes. Wiley, New York. 47. Yang, C. C., and Hartmann, G. C. (1976). Electrostatic separation of a charged-particle layer between electrodes. IEEE Trans. Elec. Dev. 23:308. 48. Yoshida, Y., and Nishimatsu, M. (1986). Power capacitors. IEEE Trans. Elec. Insul. 21(6):963±973.

10 THERMAL PROPERTIES EINAR BéHMER Cogito Consulting A.S. Oslo, Norway

I. Introduction II. Thermal Conduction A. Speci®c Heat B. Thermal Conductivity

390 390 391 393

III. Thermal Expansion

400

IV. Combustion A. Flammability B. Ignition Temperature C. Flash Point D. Oxygen Index E. Heat of Combustion

404 405 406 406 407 407

V. Thermal Decomposition of Cellulosic Materials A. Low Temperature Reactions B. High Temperature Reactions C. Signi®cance D. Theories of Fire Protection E. Analytical Methods References

408 409 414 417 419 421 425

389

390

I.

Bùhmer

INTRODUCTION

By ``thermal properties'' of a material we mean in general its properties at high temperatures. Actually, the term should also encompass the behavior at very low temperatures, for instance how sack paper will perform in subzero weather. In practice, however, we are more concerned with high temperature problems such as the ®re hazards of wallpaper, the quality of release paper for baking purposes, and the plates used for pizzas in the microwave oven [11]. Some processing problems may also justify an interest in thermal problems such as the heat transfer during drying and hot calendering [6,25±27,41]. The technology used in non-impact printing processes may also rely on heat transfer for imaging, and the problems in this connection have been dealt with in the literature [32,43]. The growing importance of this technology is expected to increase the interest in this particular ®eld. Considering the general interest in the thermal properties of other polymers, surprisingly little has been written about the properties of paper and paperboard. One reason may be that it is believed that these properties might be estimated if the thermal properties of cellulose, hemicellulose, and lignin are known. These are the basic ingredients of wood pulp, which is by far the most important component of paper and paperboard. Another complication is that the characteristics of wood ®ber are greatly affected by the moisture content. There may therefore be a widespread belief that the properties of paper and board under moisture-free conditions will not be important from a practical point of view. Also, this is a very fertile ®eld for the combination of different ®bers and chemicals, and the number of patents and papers is great. The thermal properties of these composite products may be markedly different from those of pure paper and board, and greater emphasis on the interaction of chemicals, polymers, and ®bers in a high temperature environment should certainly be expected. Although research efforts have been extensive in some areas such as ®re retardants, there have in general been relatively few reports on thermal properties. In this chapter thermal properties include thermal conduction, thermal expansion, combustion, and thermal decomposition. II.

THERMAL CONDUCTION

Thermal conduction will be in¯uenced primarily by the speci®c heat and thermal conductivity of the material. Speci®c heat is strongly related to the physical structure of polymers, and the importance of its measurement is therefore well recognized in polymer science. Such measurements are seldom applied to paper, and there is not much useful information available regarding values of speci®c heat. Although the heat conductivity of the web is important in paper drying, most of the energy consumed is spent on evaporating water, and other parameters may also be more critical in the heat transfer process. The lack of interest in speci®c heat or conductivity may therefore be explained and partly excused. With new applications of paper and board, interest in the different physical and mechanical properties of ®ber products has increased. Measurements of speci®c

Thermal Properties

391

heat and conductivity must still be regarded as a specialty, however, and the importance of these properties is not generally acknowledged. A.

Speci®c Heat

De®nitions Speci®c heat is de®ned as the energy required per unit mass of material to raise the temperature of the mass by 1 C. In SI units the speci®c heat is therefore expressed in units of energy per unit weight per degree Celsius, or kJ/ (kg K). [1 kJ=…kg K† ˆ 0:239 cal=…g C).] In some cases, such as in ASTM D 2766 [3], speci®c heat is de®ned as a unitless number. It is measured and classi®ed in two ways, at constant pressure, Cp , or at constant volume, Cv . In the case of a solid such as paper or board only Cp is measured. In dealing with heat conduction, thermal capacity is normally used along with speci®c heat. It is expressed as the ratio Q=dT, where a temperature rise dT is observed when Q calories are added to the material. It can be calculated by multiplying speci®c heat by the mass of the material. Measurement For the measurement of speci®c heat, several methods are available, for example, the adiabatic method, the thermodifferential method (heat conductive method), and the mixing method. The adiabatic method is recommended for the measurement of a cellulosic material such as paper [15]. The improved measuring device invented by Gùtze and Winkler is shown in Fig. 1 [16]. This device consists of a calorimeter, a switchboard, a regulator with attachment, a galvanometer, a Weston standard cell, and a 6 V accumulator. The main part of the measurements is the calorimeter, in the center of which is a 50 cm3 copper container with a platinum resistance thermometer and a wound ®lament for the continuous heating of a sample placed in the container. To prevent heat exchange with the surroundings, the container is encircled with a shield, which is automatically adjusted by means of a heater to avoid any temperature difference with the container. The shield is double-walled, and the outer shield is also ®tted with a heater for automatic temperature control to prevent any temperature difference. Furthermore, the shield is put in an aluminum pot, which in turn is put in a Dewar's vessel containing a freezing mixture (alcohol±dry ice) for cooling. The shield is kept at a temperature slightly lower than that of the sample with the heater. The temperatures are detected with copper-constantan thermocouples placed between the container and the shield and also between the shields. The upper part of the Dewar's vessel and the pot are ®tted with a polyvinyl chloride (PVC) cap. To prevent loss of insulation due to water condensate, silica gel or another desiccant is placed between the pot and the shield. First the heat capacity of the container is measured without a sample, and then a fully dried sample whose mass is already known is put in the container. The calorimeter is assembled and put in the Dewar's vessel, which contain alcohol, and the vessel is capped with the PVC cap. Dry ice is put through a hole in this cap. The galvanometer is adjusted so that the reading is zero in the absence of current ¯ow (after the electrical connections are made). The temperature is raised by applying constant voltage and current to the heater of the container. Next, an adjustment is made by applying currents to the heaters of the two shields so that there will be no heat ¯ow between the container and the shield. The operation is

392

Bùhmer

Fig. 1 Main measuring part of GoÈtze±Winkler calorimeter. 1, Copper container; 2, platinum resistance thermometer; 3,4, double-walled shield; 5,6, alumite pot; 7, PVC cap; 8, Dewar's vessel; 9, freezing mixture; 10, desiccant; Th, copper-constantan thermocouple; RH , heater. (From Ref. 16.)

repeated so as to raise the temperatures of the sample and the container gradually in order to obtain the rate of temperature rise for the sample under the particular heating schedule employed. From this value, the speci®c heat Cp can be calculated according to the equation mCp ‡ Ce ˆ E=
t

…1†

where Cp ˆ speci®c heat of the sample [kJ/(kg K)] Ce ˆ heat capacity of the container (kJ/K) E ˆ calories (product of voltage and current) (kJ) applied to the container heater m ˆ mass of the sample (kg)
t ˆ temperature difference (K) Typical Results Although few results have been published regarding the speci®c heat of paper and paperboard [60], a number of Cp data have appeared in the

Thermal Properties

393

literature for wood pulp and various types of ®bers. These results have been converted to kJ/(kg K) and are shown in Table 1. In general, the speci®c heat of paper seems to be quite close to that of natural wood pulps or various vegetable ®bers. In cases where the paper contains other materials, however, some deviation must be expected. Magazine paper, for instance, will contain up to 30% inorganic pigments, and this will certainly have some effect on the speci®c heat. If the paper has a high moisture content, the speci®c heat will appear to be greater because of the in¯uence of the water. This is particularly important if the speci®c heat measurements are made on paper machine wet webs. Schneider and Tùppel [44] surveyed the different thermoanalytical tools available and described the use of a differential scanning calorimeter. Olsson and Back [33] measured the speci®c heat of paper and hardboard with the differential scanning method and used two other methods for comparison. In the ®rst method, a paper web was treated over air ¯oat decks with impinging hot air of up to 425 C. The speci®c heat was evaluated from the times required to reach web temperatures up to 260 C. In the other method a commercial heating chamber for hardboard was used. Hot air was blown through the chamber, and the air velocity and the temperature difference between ingoing and outgoing air were measured. After some compensations for the heat of reaction, the speci®c heats shown in Table 2 were observed. The impingement method gave a mean speci®c heat of 3.5 kJ/(kg K) between 180 and 250 C, but the speci®c heat obtained with the differential scanning calorimeter was signi®cantly lower and closer to the values given in Table 1. The main reason for the higher values of the impingement method according to Olsson and Back is the overall exothermic reaction taking place. The heat release connected with this reaction can be neglected only with extremely rapid heating. It should also be remembered that the moisture content must be taken into account at temperatures below 100 C. B.

Thermal Conductivity

Reported data on thermal conductivity are abundant compared with those for other thermal properties. This probably re¯ects the fact that in practice conductivity is more important than other properties. Because paper is a porous and quite hygroscopic substance, its thermal conductivity is in¯uenced by its apparent density and Table 1 Measured Speci®c Heats for Paper and Cellulosic Materials Material

Speci®c heat [kJ/(kg K)]

Paper Sul®te pulp Soda pulp Cotton Flax Hemp Viscose ®ber

1.17±1.34 1.34 1.35 1.21±1.36 1.34±1.35 1.33±1.35 1.36

Reference 60 53 53 8, 16 8, 16 8, 16 14

394

Bùhmer

Table 2

Repeated Measurements of Speci®c Heat of Hardboard at 160±170 C Heat consumption (kJ)

Temperature change ( C) 11.3 11.5 11.3 9.8

Total

After correctiona

438 470 457 545

390 422 409 503

Speci®c heat calculated on board kJ/kg

kJ/(kg K)

27.8 30.1 29.2 36.0

2.47 2.63 2.59 3.68

a For supporting steel beams, etc. Source: Ref. 33.

moisture content. These effects were studied by Kartovaara et al. [25] by means of a thermoacoustic method, and both this method and their results will be reported later in this chapter. De®nition Thermal conductivity is a measure of the quantity of heat that passes through a unit area per unit time as expressed in the equation dQ ˆ d

kB

@t @L

…2†

where kˆ Qˆ dQ=d ˆ tˆ Bˆ Lˆ @t=@L ˆ

thermal conductivity heat quantity time rate of heat ¯ow temperature sample area perpendicular to the heat ¯ow distance temperature gradient in the direction of the heat ¯ow

Measurement If the ®ber product is absolutely dry or close to it, its thermal conductivity can be measured in a steady-state condition. However, if the sample contains a signi®cant amount of water, steady-state measurements are dif®cult to execute. It will then be necessary to carry out measurements in an unsteady state in order to take into account the water release caused by the heating. Steady State ASTM Method This method, now described in TAPPI T 1000 [55], yields the value for k that is used in Eq. (2). See Fig. 2. In the center of the hot plate device there is a heater, and the samples are placed on both sides. The thermal conductivity can be obtained from the quantities of heat generated by the heater and the temperatures attained by the samples on both sides. The heater is controlled electrically and is ®tted with a guard to prevent heat loss. The sizes of the samples and the guard are stipulated. The sample for measurement should be homogeneous with a ¯at surface.

Thermal Properties

395

Fig. 2 General features of the metal surface hot plate in TAPPI 1000 method. A, Central heater; B, central surface plates; C, guard heater; D, guard surface plates; E, cooling units; Es , cooling unit surface plates; F, differential thermocouples; G, heating unit surface thermocouples; H, cooling unit surface thermocouples; I, test samples. (From Ref. 55.)

At the outset, the temperature difference between the hot plate and the cold plate is set greater than 12 C, but the temperature on the cold side should not be below the ambient dew point for obvious reasons. The thickness of the sample is determined by the distance between the hot and cold plates. The temperature measurement is conducted by reading the electromotive force up to 0:5 V on the potentiometer in a steady state. ``Steady state'' means here that the variation of the temperature difference between the hot plate and the cold plate should be less than 0.5% of the measured temperature difference and that the temperature drop through the two specimens should not differ by more than 1%. Thermal conductivity is calculated with the equation kˆ

qL A…t1

t2 †

where k ˆ thermal conductivity (W/mK) q ˆ time rate of heat ¯ow (W)

…3†

396

Bùhmer

Lˆ Aˆ t1 ˆ t2 ˆ

thickness of sample (m) two-sided area of heating plates (m2 ) temperature of hot plate (K) temperature of cold plate (K)

Terasaki's Method This is a method that measures the thermal conductivity of paper at constant temperature and humidity [58]. It is illustrated in Fig. 3. The sample is wound three times around a copper tube containing an electric heater. Air regulated with a thermostat is in contact with the sample on the outer surface. The temperature difference between the copper tube surface and the air is adjusted to 2±3 C. A copper-constantan thermocouple is used for the actual temperature measurement. In addition to the method described here, Terasaki and coworkers also devised two other methods: a method to measure heat transfer coef®cients by passing hot and cold currents inside and outside a copper tube [57] and a method to measure effective thermal conductivity of the paper with a plate heater [59]. Terada's Method Terada et al. [56] devised a special method to measure the thermal conductivity of electrical insulating papers (Fig. 4). A copper-stranded cable is used as a high temperature source in this case, and electrical insulating paper is wound around it to a thickness up to 23 mm. The thermal conductivity of paper for this particular purpose is generally measured in vacuum after completely drying the samples and also in dry nitrogen gas at three pressure levels (100, 380, and 760 mmHg). Unsteady State Whereas in the steady-state methods the heat is furnished continuously to one side of the sample and removed at the other, the heat supply is in general variable in the unsteady-state method. The time and temperature gradients and rates of heat supply and removal are measured and used in the appropriate forms of the general heat transfer equation.

Fig. 3 Terasaki's device. 1,2, Heating medium; 3, panel heater; 4, foamed styrol panel; 5, glass wool; 6, wood panel; 7, blower; 8, humidi®er; 9, guard heater; 10, copper tube with main heater; 11, test sample; 12, wet and dry bulb thermometer. (From Ref. 58.)

Thermal Properties

397

Fig. 4 Terada's apparatus. 1, Drying tank; 2, conductor (heat source); 3, test specimen layer; 4, thermocouples; 5, roots pump; 6, rotary pump. (From Ref. 56.)

Method of Kirk and Tatlicibaci These authors [28] devised the following method, which offers the advantage of speed and a special applicability to paper and other porous materials containing water or other volatiles. The device shown in Fig. 5 uses a ¯ash power pack and a ¯ash tube. The energy supplied by the former is radiated as a heat pulse to the sample. The sample, which is held in a special holder, receives heat on one side. On the other side of the sample a thermocouple is connected to an oscilloscope for rapid detection of the temperature change. From the measurements thermal diffusivity is obtained. The thermal conductivity k is determined from the density  and the speci®c that c of the sample as well as the diffusivity  by the relation k ˆ c. The measurements are made after conditioning the samples at 20 C and 50% RH. Kartovaara's Method Kartovaara et al. introduced a method that has given very interesting results [25]. They used a thermoacoustic method based on the propagation of a periodic temperature wave in the medium. In standard acoustic work the sample is placed at the bottom of a closed cell, and the signal is recorded with a microphone (Fig. 6a). In the present case the transmitted thermal wave was measured by generating a thermal wave on one side of a paper sample and detecting it at

Fig. 5 Heat pulse method for determining the thermal conductivity of paper. (From Ref. 28.)

398

Bùhmer

Fig. 6 Photoacoustic method for determining the thermal conductivity of paper (a) with standard photoacoustic cell and (b) with thermoacoustic cell with heater. (From Ref. 25.)

the opposite side with suitable thermal detectors (Fig. 6b). The thermal wave undergoes a phase shift that is directly proportional to the thickness of the sample. A conventional photoacoustic cell has a window that is optically transparent. Here, however, the window is made of thin aluminum foil, which means that it is thermally conductive. The thermal wave can therefore easily propagate through the aluminum window, enter the closed cell, and create a periodic pressure there that is then picked up with a microphone as sound. The thermal wave was generated by the resistance heater method. The details of the method and the theory behind it are given in Ref. 25. The phase shifts obtained with different basis weights were used to calculate the thermal diffusivity, D ˆ K=C, where K is the thermal conductivity, C is the speci®c heat, and  is the density. Typical Results Some reported values for the thermal conductivity of paper and board are given in Table 3. Some of the results are dif®cult to explain, and they apparently depend on the measuring methods and the ambient conditions. However, some conclusions may still be drawn. From samples 6 and 7 it is evident that an increase in the moisture content will increase the conductivity [17]. This conclusion is supported by samples 8±11 and 12±15, which also indicate that there is a modest increase in conductivity on increasing the temperature from 30 to 60 C. Because the paper in general is quite porous, some effect of the air might be expected. Terasaki et al. [59] showed that 90% of the effective thermal conductivity derives from the conductivity of the air present inside the paper, with the solid content of the sheet contributing only 10%. Kirk and Tatlicibaci [28] showed that the conductivity will increase with beating and on increasing the thickness of the sheet.

Thermal Properties

399

Table 3 Thermal Conductivity of Paper and Paperboard Sample No. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.

Material Paper Cardboard Blotting paper Handsheet, bleached sul®te Electrical insulating paper, 0.15 mm, 0.86 g/cm3 Sul®te pulp, 1500 g=m2 Sul®te pulp, 1500 g/m2 Fine paper, 0.074 mm, 0.86 g/cm3 Fine paper, 0.074 mm, 0.86 g/cm3 Fine paper, 0.074 mm, 0.86 g/cm3 Fine paper, 0.074 mm, 0.86 g/cm3 Copy paper, 0.06 mm, 0.84 g/cm3 Copy paper, 0.06 mm, 0.84 g/cm3 Copy paper, 0.06 mm, 0.84 g/cm3 Copy paper, 0.06 mm, 0.84 g/cm3

Thermal conductivity [W/(m K)Š  10 6 21 6.3 8.4±11.7 4.6±7.9 6.7 19 1.3 3.1 3.7 5.8 14.6 15.1 16.3 17.0

2

Condition

Ref.

Room temperature Room temperature 20 C 20 C, 50% RH 100 C vacuum

60 60 45 28 56

0% moisture content 110% moisture content 30 C, 40% RH 30 C, 60% RH 60 C, 40% RH 60 C, 60% RH 30 C, 60% RH 30 C, 60% RH 60 C, 40% RH 60 C, 60% RH

17 17 58 58 58 58 58 58 58 58

It should be noted that samples 8±11 should not be much different from samples 12±15, but the conductivity of the former is only a small fraction of that of the latter. The temperature and the moisture content were apparently the same, and the densities were very close. The difference in conductivity may therefore be due to some unknown parameter, for instance, the content of inorganic ®ller or other chemicals. This emphasizes the point that it is very important to de®ne the paper properly when conductivity measurements are made to enable a correct interpretation of the results. Sanders and Forsyth [43] used a steady-state technique to measure the conductivity and the contact resistance of paper and thin ®lmlike materials using varying pressure between the heat-conducting surfaces. According to their ®ndings there was surprisingly little variation between the bulk conductivities of different papers. This suggests that the variation in published values of the apparent thermal conductivity of paper is due mainly to variations in thermal contact resistance. In terms of an effective air gap, this might range from 8 m for an uncalendered sheet to 1 m for a smooth calendered paper. This may explain some of the discrepancies in Table 2 and other published results, and it means that the surface roughness should be compensated for. Kartovaara et al. [25] studied a variety of both mechanical and chemical pulps, and their data enable us to clarify some of the important effects on thermal conductivity. If density and thermal conductivity increase in the same proportion during calendering, thermal diffusivity should remain constant. Actually, the results showed that calendering caused a decrease in thermal diffusivity in the case of mechanical pulps whereas the opposite was true of chemical pulps. Figure 7 shows thermal conductivity as a function of density. It can be seen that the density explains most of the variation in thermal conductivity. As already mentioned, it is more dif®cult to evaluate the effect of humidity and temperature properly. Moisture variations introduce the problem of paper swelling,

400

Bùhmer

Fig. 7

Heat conductivity as a function of density. (From Ref. 25.)

and the compressibility will change because of both temperature and moisture variations. The data in Fig. 8 show, however, that higher relative humidity and temperature increase the conductivity signi®cantly, and these effects cannot be explained in terms of density differences. Unfortunately, the data do not permit a separation of the temperature and humidity effects, but the values in Table 3, notably for samples 8±11, indicate that the two effects may be of the same order of magnitude. Again, it must be emphasized that a copypaper with approximately the same density (samples 12±15) had a much higher conductivity, indicating that the conductivity is in¯uenced by some other factors.

III.

THERMAL EXPANSION

All materials expand and shrink with changes in temperature. It is, however, a special feature of paper and board that the effect of the moisture content will fre-

Fig. 8

The effect of moisture and temperature on heat conductivity. (From Ref. 25.)

Thermal Properties

401

quently overshadow the effect of the temperature in both magnitude and importance. It is therefore easy to understand why the thermal expansion has often been ignored and why so few data have been published on the subject. Examples have already been mentioned where paper might be used in moisture-free environments, and the thermal expansion may in these cases present a problem. New special applications of paper products may increase the interest in this problem, and a study of thermal expansion may therefore be warranted. The degree of expansion of a material is generally expressed as either a coef®cient of linear expansion, indicating a change of dimension, or a coef®cient of volume expansion, which is used to show a volumetric change. For ®ber products the linear expansion is normally the more important. Again, it should be emphasized that the dimensional change due to adsorption or release of moisture is much greater than the change due to heat, and the latter may well fall within the tolerance of measuring errors. Because paper and paperboard are hygroscopic products, a small amount of water may remain in the sample to be measured. The in¯uence of this water should be carefully assessed if thermal transition points are to be obtained from thermal expansion measurements. De®nitions The coef®cient of linear thermal expansion l and the coef®cient of volume thermal expansion v represent the percentages of change in length and volume, respectively, when the temperature rises by 1 C. They are de®ned by the equations l ˆ

@l=@t l

…4†

v ˆ

@v=@t v

…5†

and

where l ˆ length of sample v ˆ volume of sample t ˆ temperature In the vicinity of room temperature it may be assumed that the length is a linear function of the temperature, and the following equation may then be used: l ˆ where

l2 l1 …t2 t1 †l0

…6†

l0 ˆ length of sample at 273 K (0 C) l1 ˆ length of sample at t1 l2 ˆ length of sample at t2

A similar expression may be used for the volume expansion by replacing the length l in Eq. (6) with the volume v.

402

Bùhmer

The measurement is conducted under constant pressure, and the unit of expansion is K 1 . Measurement Coef®cient of Linear Thermal Expansion The following methods of measurement are generally available for the coef®cient of linear thermal expansion: A direct reading of the change in length Light interference Measurement of the movement of an electrode plate with the changes in the capacity of an electrode For the measurement of paper, the direct reading method of KubaÂt et al. [30] is probably the best available. The apparatus is shown in Fig. 9, and it can be described as follows. The measuring device consists of a glass tube containing a hollow stainless steel bar upon which the sample to be measured is placed. Dry nitrogen gas ¯ows through the glass tube to maintain an absolutely dry state in order to prevent any in¯uence of moisture. The cross section of the steel bar is square, and a liquid (glycerol or ethanol) whose temperature is controlled by a thermostat is circulated within it. The sample for measurement is placed on the steel bar and covered with a small U-section copper bar to eliminate temperature gradients through the sample. The copper bar is slit so that the position of one end of the sample can be followed and measured with an overhead traveling microscope. By means of a clamp, the other end of the sample is ®xed to a point on the stainless steel bar. The sample is dried at 105 C for 10 h in a stream of nitrogen gas passing through the glass tube. An initial length is determined, and the increase or decrease in the length due to a temperature change is then measured with the traveling microscope. A thermocouple is used for the temperature measurements. In the experiment of KubaÂt et al. [30] the speed of the temperature rise was 2:5 C/10 min, and the thermal expansion was

Fig. 9 Apparatus of KubaÂt et al. for determining the thermal expansion of paper. 1, Feed of gas (nitrogen); 2, traveling microscope; 3, thermocouple; 4, feed of thermostating liquid. (From Ref. 30.)

Thermal Properties

403

measured for both ascending and descending temperatures. From the change in length and the temperature variation, the linear expansion coef®cient could be obtained. Paper and paperboard will have several transition points between 0 and 100 C, and it is therefore essential that the accurate temperature be recorded together with each measurement of the linear expansion coef®cient. Klason and KubaÂt [29] used a dilatometer and a torsion pendulum for the measurement of the linear expansion coef®cient of a spruce sul®te pulp, cellulose II prepared by the viscose method, and sheets formed with ammonia-treated pulp ®bers. In these experiments the speed of the temperature rise was 0.2±0:5 C/min and the temperature varied between 163 C (110 K) and ‡197 C (470 K). Coef®cient of Volume Thermal Expansion Measurements of the coef®cient of volume thermal expansion have been carried out, but no data have been reported in the literature for paper or paperboard. To measure v the sample has to be formed into pellets or some similar form, which may be quite different from the original material. Accordingly, it is believed that the values obtained are of little signi®cance for paper and board. Using a dilatometer, pulp and cellulose were analyzed by Ramiah and Gording [39], and readers may wish to refer to this paper. Typical Results KubaÂt et al. [30] and de Ruvo et al. [40] both published linear expansion coef®cients for various paper and paperboard products. Some values are given in Table 4. From Table 4 it is apparent that the thermal expansion is smaller in the machine direction (MD) than in the cross-machine direction (CD). The expansion coef®cient therefore shows the same directional dependence as the modulus of elasticity or the extensibility, but the trend is opposite. If the product of the expansion coef®cient and the modulus of elasticity is calculated, the numerical values have been found to be approximately the same in both directions. This numerical value will, however, vary from grade to grade. The dimensional changes caused by the moisture in the ambient air (hygroexpansion) follow the same trend as the thermal changes, being less in the machine Table 4

Coef®cient of Linear Thermal Expansion K of Paper and Board (at 25 C) K

Sample No. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. Source: Ref. 30.

Material 2

Greaseproof, 40 g=m , 0.038 mm MG kraft paper, 50 g=m2 , 0.056 mm Sack paper, 70 g=m2 , 0.099 mm Newsprint, 52 g=m2 , 0.080 mm Groundwood printing, 60 g=m2 , 0.054 mm Fluting, 127 g=m2 , 0.234 mm Whitelined duplex, 400 g=m2 , 0.58 mm Solid bleached board, 240 g=m2 , 0.246 mm White chip board, 400 g=m2 , 0.55 mm Lined chip board, 400 g=m2 , 0.57 mm Bleached sulfate, 944 g=m2 , 0.92 mm

1

 10

6

MD

CD

7.5 6.4 6.1 5.6 3.7 7.4 5.9 4.2 3.6 2.0 6.1

15.5 13.4 16.2 13.6 10.1 12.1 12.3 8.7 15.2 13.7 7.9

404

Bùhmer

direction than in the cross-machine direction. The absolute values for hygroexpansion are, however, an order of magnitude greater than the linear expansion if one compares a change in humidity from 0% to 95% with a temperature change of 100 C [40]. From the description of the different paper grades given in Table 4 it is dif®cult to ®nd paper properties that correlate with the linear expansion. There is, for instance, no systematic effect of density although the degree of beating will have some effect [40]. The paper grades in Table 4 seem to have been selected at random, and it is therefore impossible to evaluate the effect of such variables as the content of ®ller, the amount of mechanical pulp in the furnish, the type of pulp, etc. Thermal transition points can be obtained from the change of the thermal coef®cient with temperature. KubaÂt et al. [30] discovered thermal transition points near 35 and 65 C, but it has been hypothesized that this phenomenon may be caused by small amounts of moisture remaining in the paper [29]. Ramiah and Goring [39] determined the volume expansion of sul®te pulp and cellulose, and it is noteworthy that the pulp swollen in water expands two to three times more than dry ®ber. Below the transition point, the dry ®ber expands less than above this point, but the opposite is true for water-swollen ®ber [38].

IV.

COMBUSTION

Paper and paperboard are generally ¯ammable. This property is advantageous in the sense that paper materials can be readily disposed of by incineration, but it is also a disadvantage in cases, for instance, when the paper product is used as wallpaper or in the form of building board. In these cases the ®ber product may pose a serious ®re hazard, and this problem has limited the use of paper products, for instance, for interior decoration. It has been a long-term goal of the building industry to improve the ®re resistance of paper, and the pressure to use ®reproof materials has triggered an enormous amount of work on ¯ame-resistant and in¯ammable materials. For those particularly interested in the subject, reference is made to the many reviews on the subject [23,35] and the patent surveys that can be found in scienti®c databases such as the ABIPC Catalogue or by using modern data retrieval techniques. In this connection several methods have also been developed to measure ¯ammability objectively. The results from these measurements serve well as an index of the ¯ammability of paper and board, but these data should be evaluated with the tacit understanding that the tests and the measurements are based on speci®c experimental conditions and therefore may have limited usefulness. De®nitions The reader should note that certain words related to combustion are often used erroneously. Several of these terms are de®ned here. Combustion: An oxidizing reaction in which heat and light are evolved. Ignition temperature (autoignition temperature): The lowest temperature at which a material when heated in air will spontaneously ignite from the heat of oxidation and in the absence of any other source of ignition energy [19].

Thermal Properties

405

Flash point: The lowest temperature at which ¯ammable gas given off from a material heated in air is ignited with external ignition energy when the density of the evolved ¯ammable gas has reached the combustion limit. Burning point: The lowest temperature at which the ¯ammable gas evolved at a temperature slightly higher than the ¯ash point is kept burning after being ignited (usually about 10 C higher than the ¯ash point). Oxygen index: The minimum concentration of oxygen expressed as volume percent in a mixture of oxygen and nitrogen that will just support the combustion of the material.

A.

Flammability

Measurement TAPPI Method A ¯ammability test is described in TAPPI T 461 [54]. The testing device consists of a metal cabinet, a holder, and a burner. A sample 210 mm in length and 70 mm in width is inserted in the holder and is suspended vertically on the knob in the center of the cabinet. A Bunsen or Tirrill gas burner 10 mm in inner diameter is used. The distance between the top of the burner tube and the lower edge of the sample is adjusted to 19 mm, and the ¯ame height is adjusted to 40 mm. After closing the cabinet door, the burner is ®red in such a way that the lower edge of the sample attached to the holder is exposed directly to the yellow ¯ame of the burner for 12 s, and then the ¯ame is removed. Flaming time, the length of time the sample glows after it has ceased ¯aming, and char length are both recorded. According to the TAPPI method it is further stipulated that a water immersion test be conducted for waterproof, ¯ame-resistant paper. UL Method Another ¯ammability test is described in Underwriters' Laboratories standard UL 94 [61]. This test is used to measure the ¯ammability of laminated board consisting of resin-saturated paper. The testing device is the same as the one described in the TAPPI method. The sample is 127 mm in length and 13 mm in width. Each set consists of ®ve samples. A ring stand with clamps that adjust for vertical positioning is used to hold the sample. The longer axis of the sample is aligned vertically. The sample is held by a clamp at a point 6 cm from the upper end. The height of the blue ¯ame of the burner is adjusted to 19 mm, and the center of the lower edge of the sample is exposed continuously for 10 s. The length of time the sample continues to burn after removal of the ¯ame is then recorded. After the ¯ame is extinguished, the sample is exposed for 10 s to the burner ¯ame, and the duration of ¯aming and glowing is again recorded. The ¯ame resistance of the sample is graded by the total ¯aming combustion time for the total of 10 ¯ame exposures for each set of ®ve samples. Typical Results Ordinary paper and board are consumed in a burning test. If the papers are treated with ¯ame or ®re retardants, the degree of burning differs depending upon the ef®ciency of the treatment. When 13 different commercial ¯ame-resistant wallpapers were tested according to the TAPPI method, the average values of 10 points of char length varied from 52 to 104 mm. Generally, if the char length exceeds 120 mm, it is very likely that the sample will burn completely. The measurements vary a great deal also with the UL method, and there is a tendency for the difference

406

Bùhmer

between replicates to become greater as the degree of ¯ame retardancy decreases. Results on laminated board for electrical insulation show that the total time of burning is 10±15 s for a material with good ¯ame resistance and 50±100 s for materials with a low degree of treatment. Since the reproducibility of ¯ammability tests in general is poor, it is advisable to have untreated samples as a blind for comparison with the treated samples. B.

Ignition Temperature

Measurement Several devices can be used to measure the ignition temperature of paper and board, but here the UL method stipulated in UL 94 [62] is described. A combustion chamber is put in a metal-melting bath and heated with a special electric furnace and with a temperature controller. The combustion chamber is a conical ¯ask made of heat-resistant glass with a ¯at bottom. The diameter of the bottom is 60 mm, that of the top 28 mm, and the height is 114 mm. A thermocouple protected with quartz is used for the temperature measurements. Pieces 6.4 mm square are dropped into the ¯ask at intervals of 5 C and are observed to see whether they ignite. The lowest temperature that results in ¯aming or glowing combustion within 2 min is taken to be the ignition temperature. Typical Results The ignition temperature of paper is about 450 C [19], but it is somewhat dependent upon the paper quality. The ignition temperature is 450 C for rayon ®bers, 475 C for cotton, and 550 C for ¯ame-resistant cotton (treated with Nmethyl-dimethyl-phosphonopropionamide) [22]. From the data published the ignition temperature of paper treated with ®re retardants seems to be about 100 C higher than that of an untreated sample. C.

Flash Point

Measurement The method devised by Ishii et al. [22] is described here. An electric furnace is used in this method. At the lower end of the furnace a chromel-alumel thermocouple is ®xed, and an aluminum foil containing a sample is placed on the thermocouple. A Ni-chrome wire is used as an ignition source. A 25 g powdered sample is used, and the rise of the furnace temperature is 30 C= min. The ¯ash point is the temperature at which a ¯ame starts to appear. The device can also be used to measure the ignition temperature if the Ni-chrome wire is removed as an ignition source. Typical Results No data seem to have been published on the ¯ash point of paper and board, and reference is therefore made to measurements made on various ®bers with cellulose as the origin. Ishii et al. [22] found that the ¯ash point for cotton was 361 C, for rayon 327 C, and for ¯ame-treated cotton more than 650 C. Based on these results, the ¯ash point of paper and paper board has been estimated to be around 350 C. It is not clear what the effect of mechanical pulp in the paper will be, that is, what effect the lignin in the ®ber will have on the ¯ash point. Other variables such as the content of inorganic ®llers and different organic chemicals are also thought to have an effect on the ¯ash point, but there are no data available that will help us to settle these questions.

Thermal Properties

D.

407

Oxygen Index

Measurement A method for the measurement of the oxygen index (OI) is given in ASTM D 2863 [4]. The testing device consists of a heat-resistant glass tube 75 mm in inside diameter and 450 mm in height, a sample holder, a gas supply, a ¯ow measurement and control device, and an ignition source. The size of the sample may vary from 70 to 150 mm in length, be 6:5  0:5 mm in width, and have a thickness of 3:0  0:5 mm. At least 10 replicates of each sample should be measured. The device was originally designed for plastics of moderate thickness, and a special device is therefore needed for thin sheet material such as paper in order to af®x the sample in the holder. The test is run by using an ignition source with oxygen±nitrogen mixtures with varying oxygen content. Gas of known composition is passed into the glass tube at a rate of 4  1 cm=s for at least 30 s. The upper edge of the sample is then ignited with a gas burner. The minimum amount of oxygen that will either keep the test piece burning for more than 3 min or consume more than 50 mm of the sample length must be measured to obtain an oxygen index by the equation OI…%† ˆ 100

‰O2 Š ‰O2 Š ‡ ‰N2 Š

…7†

where ‰O2 Š ˆ minimum ¯ow quantity of oxygen required for combustion (cm3 =s) ‰N2 Š ˆ ¯ow of nitrogen corresponding with the ¯ow above (cm3 =s) Typical Results Although there are no measurements of the oxygen index for paper and board reported in the literature, there are some results for other cellulosic materials. According to Ishii et al. [22], the index for cotton is 18%, for rayon 19%, and for ¯ame-retardant cotton 35%. From these ®gures the oxygen index of paper is estimated to be close to 20%. Generally, the oxygen index will be below 20% for ¯ammable materials and above 30% for ®re-retardant materials.

E.

Heat of Combustion

The heat of combustion is measured as the number of calories produced when a material burns. It is expressed as energy per unit mass, and it can be measured with the same calorimeter that is used to measure the heat of combustion of various solids. No special method has been published for paper and board, but values for the heat of combustion for various pulps varied from 17.4 to 18.0 kJ/g (4.2±4.3 kcal/ g) [8,13,53]. Paper and board are major components in municipal refuse, and in this connection they are important heat sources. Energy-ef®cient combustion requires suf®cient oxygen. With limited oxygen supply the ®ber material is converted to carbon monoxide with enthalpy H ˆ 96 kJ=mol. With suf®cient oxygen the organic substance is converted to carbon dioxide and H ˆ 370 kJ=mol [35].

408

V.

Bùhmer

THERMAL DECOMPOSITION OF CELLULOSIC MATERIALS

As paper is heated, it goes through a series of physical and chemical changes that affect the physical properties and ultimately lead to charring, pyrolysis, and complete destruction of the paper. The deterioration is ampli®ed not only by increasing temperature but also by the pressure of oxygen and water and the presence of reactive compounds that produce an autocatalytic effect. In particular, the effects of alkaline and acidic materials are well known [12]. It is also well known that acidic material is steadily formed during the storage of paper, because of both the hydrolysis of organic esters in the pulp and the adsorption of sulfur dioxide and nitrous oxides from the ambient air. The thermal stability of paper can therefore be improved both by removing inorganic material and by introducing products such as calcium carbonate that will neutralize the acids in the pulp that otherwise would be very detrimental [42]. At lower temperature it is dif®cult to distinguish between normal aging and oxidative thermal degradation, which is accelerated by heating. When rag paper is stored at 38 C for 6 months, the accelerated aging results in a 19% reduction of the fold strength. It should be noted that the moisture content during storage is equally important, and correct interpretation of aging experiments requires good control of and information about the moisture content during storage [52]. Thermal degradation precedes combustion of cellulosic materials and affects the ¯ammability, how the material burns, and how the combustion can be controlled. Thermal degradation also provides the basis for an analytical pyrolysis that is a powerful tool for the analysis of polymer materials, including pulp and paper. Thermal degradation is used to speed up the natural aging of paper. In view of the fact that our life span is less than 100 years and our records are supposed to last at least 500 years, it is essential that we be able to assess the permanence of paper within a period of months or at most a few years. This means that an accelerated aging procedure at higher temperature is required. With the growing interest in the conversion of biomass and waste paper to energy and chemicals, the thermal degradation of cellulosic materials has been extensively studied and discussed. The following summary provides a general understanding of the phenomena occurring during this process [47±49]. The thermal reactions are complex, and the substrates are generally inhomogeneous. It is therefore dif®cult to generalize on the subject, and some variations in the experimental data must be expected. This problem may be controlled by working with model substances, studying the thermal reactions under various experimental conditions. Because paper contains mostly cellulose and hemicellulose, it is logical that most basic studies have been focused on the thermal degradation of cellulose. It should be emphasized, however, that newsprint will contain almost 100% mechanical pulp, hence about 28% lignin. Recent studies by Oye [36], Lystad [31], and others have also shown that the lignin is as stable to heat or long-term storage as the cellulose itself, and this has led to a reevaluation of mechanical pulp in so-called permanent papers. As cellulose is heated, the activation energy of different reactions releases suf®cient heat to result in char formation and a variety of volatile products that can be measured as weight loss by thermogravimetry (TG). An Arrhenius plot with rate of weight loss of cellulose on the ordinate and the inverse absolute temperature on the abscissa is shown in Fig. 10. The measurements were made in both air and

Thermal Properties

409

Fig. 10 Arrhenius plot for the ®rst-order reaction in the isothermal degradation of cellulose in air (Ð) and nitrogen (---). (From Ref. 50.)

nitrogen. These data indicate a transition point around 300 C (573 K), which again points to the existence of two different pathways of pyrolysis. As shown in Fig. 11, the weight loss under isothermal conditions at moderate temperatures proceeds much faster in air than in nitrogen. As the temperature increases, the difference between the pyrolysis in nitrogen and air gradually diminishes, and it disappears at 310 C when pyrolysis by the second pathway takes over. Different pathways for the pyrolysis of cellulose are suggested in Fig. 12 [7,50]. A.

Low Temperature Reactions

As shown in Fig. 12, the dominating pathway at low temperature involves a reduction of the molecular weight or the degree of polymerization (DP) by bond scission; the appearance of free radicals; the removal of water; the formation of carbonyl, carboxyl, and hydroperoxide groups; the evolution of carbon monoxide and dioxide;

Fig. 11 First-order plot for the residual cellulose weight (normalized) versus time. Plots at 310 and 325 C for air and nitrogen are similar. (From Ref. 50.)

410

Fig. 12

Bùhmer

Competing pathways for the pyrolysis of cellulose. (From Ref. 50.)

and ®nally the production of charring residue. The kinetics of all these reactions that contribute to the degradation of the cellulose have been individually studied. The reduction of the degree of polymerization (DP) has, for instance, been measured in the temperature range 150±190 C by the viscosity method. The results are given in Fig. 13. It is generally assumed in the pulp and paper industry that a DP of 600 is the minimum value that might be acceptable. Below this value the strength of the pulp falls off very rapidly. The data in Fig. 13 may be used to determine the time necessary at different temperatures to degrade the cellulose to this level. According to the data, about 6.5 h is required at 190 C, whereas the time is increased to approximately 66 h at 150 C. Although the uncertainty is great in this process, it is tempting

Fig. 13 Viscosity-average degrees of polymerization (Pv ) of cellulose heated in air at 150± 190 C. (From Ref. 50.)

Thermal Properties

411

to extrapolate these values to lower temperatures, notably the storage temperatures in libraries and archives. It is then found that a DP of 600 based on these data corresponds to a storage time at 20 C of between 50 and 100 years. This is not very impressive compared with the papyrus found in Egyptian tombs and indicates that we have to be careful both in treating our raw materials and in planning our experiments. It should be noted that moisture will not be present at these high temperatures, but at lower temperature moisture will be an important factor in the degradation process. Arti®cial aging should therefore be done at controlled moisture levels. The data in Fig. 13 have also been correlated with rates of bond scission as given in Table 5, and they have been used to calculate the Arrhenius plot given in Fig. 14. It is apparent from both Table 5 and Fig. 14 that the effect of an inert atmosphere is reduced as the temperature increases, as already mentioned. An extrapolation of these data to 20 C (1=T ˆ 3:43  10 3 K 1 ) supports the conclusion that the depolymerization in nitrogen becomes negligible, but it should also be emphasized that the rate constant in air is so small that oxidation cannot be held responsible for the degradation of printed matter in archives and libraries. The culprit here is a process of hydrolysis, which, of course, will be dependent on both the humidity and the temperature. Figure 15 shows the rate of forming carbon monoxide and dioxide at 170 C in air and nitrogen. The carbon gases are formed much faster in air than in nitrogen, which is not unexpected, and the formation accelerates on continued heating. It is instructive to compare the initial rates of emission of these gases with the rates on bond scission for the depolymerization process discussed before. It can be seen from Table 6 that for air the rate of bond scission roughly equals the rates of carbon monoxide and dioxide emission in moles per glucose unit. In nitrogen, however, the rate of scission is signi®cantly greater than these combined rates, and this is not unreasonable in the absence of oxygen. The thermal degradation of cellulose in air apparently involves a free radical mechanism that is similar to that of many synthetic polymers, and also the formation of hydroperoxide groups. These groups are both formed and decomposed, and the concentration rapidly climbs until a steady state is reached. Figure 16 shows the development of a steady-state concentration in air at 170 C during a period of 100 min. It also shows the rate of decay in nitrogen, and this decay of the hydroperoxide

Table 5 Rate Constant kv for the Depolymerization of Cellulose in Air and Nitrogen Temperature ( C) 150 160 170 180 190 a

kv  107 mol=162 g mina In N2

In air

1.1 2.8 4.4 9.8 17.0

6.0 8.1 15.0 29.8 48.9

162 g represents 1 mol of monomer unit.

412

Bùhmer

Fig. 14 Arrhenius plot for the rate of bond scission in air (Ð) and nitrogen (---). (From Ref. 50.)

Fig. 15 Yields of CO and CO2 from heating cellulose at 170 C. (*) CO2 in N2 ; (&) CO in N2 ; (*) CO2 in air; (&) CO in air. (From Ref. 50.)

Table 6 Initial Rates of Glycosidic Bond Scission and Carbon Monoxide and Dioxide Formation at 170 C Rate  105 mol/162 g ha Reaction Bond scission CO evolution CO2 evolution a

In N2

In air

2.7 0.6 0.4

9.0 2.1 6.4

162 g represents 1 mol of monomer unit.

Thermal Properties

413

Fig. 16 Rate of formation and decay of hydroperoxide groups in cellulose at 170 C (443 K). Formation is in air, and decay is in nitrogen. (From Ref. 50.)

seems to follow a ®rst-order reaction with a rate constant of 2:5  10 2 min 1 at 170 C. If the rate of hydroperoxide decomposition is compared with the initial rate of bond scission in air (Table 6) it is apparent that the hydroperoxides could make a signi®cant contribution to the bond scission [50]. According to the preceding reasoning and in analogy with studies on the radiation of carbohydrates, it is suggested that the following three stages are involved in the low temperature pathway: initiation of the pyrolysis, propagation, and product formation. As shown in Fig. 17, the initiation process apparently involves the formation of free radicals triggered by the presence of oxygen and inorganic impurities as catalysts. The subsequent reaction of the free radicals

Fig. 17

Thermal autoxidation of cellulose in air. (From Ref. 50.)

414

Bùhmer

could lead to bond scission, oxidation, and decomposition of the molecule, producing char, water, and carbon monoxide and dioxide. Figure 18 gives one possible mechanism for the formation and decomposition of cellulose hydroperoxide formed thermally in air by the pathway indicated above. What are called ``low temperature reactions'' generally encompass the reactions at 150±300 C. This is of interest for some products that are used at very high temperature and also for some products that are exposed to a very high temperature during the process. For instance, autoignition has occurred when hardboard has been stacked with a product temperature of 170 C or higher [5]. The most important phenomenon with respect to the deterioration of ®ber products is, however, the relentless deterioration at normal storage temperature. This has brought about a dramatic development in our laboratories, and in the Library of Congress in Washington, DC, several million volumes have been taken out of circulation because the paper cannot be handled without becoming the proverbial dust. The reasons for this deterioration are today clearly understood. The main reaction is the hydrolysis of the cellulose catalyzed by acid components. Oxygen also plays a part, but under normal storage conditions the hydrolysis will be much more important. This means that both the humidity and the temperature will be important, and the pH value should be increased with calcium carbonate or a similar substance to give some buffer capacity. Arney and Jacobs [1,2] have made important contributions in this particular ®eld, and more recently Oye [36] and Lystad [31] showed that mechanical pulp and wood-containing paper have much better storage stability than their reputation would indicate. Mechanical pulp is well known to have poor color stability against the effect of UV radiation, but its heat stability is actually better than that of pure kraft or sul®te pulps.

B.

High Temperature Reactions

At temperatures above 300 C, cellulose is decomposed by an alternative route that gives a tarry substance of levoglucosan, other anhydroglucose compounds, randomly linked oligosaccharides, and glucose decomposition products. The complex mixture of the reaction products is known from gas-liquid chromatography (GLC), and the mechanism for the formation of these compounds has been established by

Fig. 18 Possible mechanisms of formation and decomposition of cellulose hydroperoxide formed thermally in air. (From Ref. 50.)

Thermal Properties

415

extensive investigation of the pyrolytic reactions of phenyl glucosides and related model substances [47±49]. The mechanism shown in Fig. 19 involves an intramolecular substitution of the glucosidic linkage in the cellulose by one of the free hydroxyl groups, and subsequent inter- and intramolecular substitutions give several anhydrosugars and randomly linked oligosaccharides, which can dehydrate and decompose on further heating to form a tar. The initial substitution requires some changes in the conformation of the sugar units and increases the ¯exibility of the molecule. These changes may eventually be achieved at elevated temperatures by a reduction of the molecular weight, some rupture of hydrogen bonds, or a glass transition, all of which would be expected to activate the molecule. Kinetic studies have shown that the tar-forming reactions accelerate rapidly at higher temperatures and will overshadow the production of char and gases. The data in Table 7 show the diminishing amounts of char and increasing amounts of tar on increasing the temperature from 300 C to 500 C. The chemical kinetics of the cellulose pyrolysis may be represented by several models, and one alternative is outlined in Fig. 20. In this model it is assumed that the initiating reaction mentioned before leads to the formation of an active cellulose that

Fig. 19 Pyrolysis of cellulose to anhydrosugars and other compounds by transglycosylation reactions.

416

Bùhmer

Table 7 Effect of Temperature on the Products from Pyrolysis of Cellulose Powder Under Vacuum Percent yield from cellulose Oven temp. ( C) 300 350 400 450 500

Pyrolysis time (min)

Char

180 30 5 3 3

21 8 5 4 3

Tar

Levoglucosan

1,6-Anhydrod-gluco furanose

Reducing sugar

60 70 77 78 81

34 38 39 39 38

4 4 4 4 4

47 52 60 57 57

is subsequently decomposed by two competing ®rst-order reactions, one yielding anhydrosugars and the other char and a gaseous fraction [7,8]. The detailed analysis of the pyrolysis tar shows the presence of levoglucosan, its furanose isomer, and their reaction products as the main components. The pyrolysate also contains minor amounts of a variety of products formed from the dehydration of the glucose units. Like aqueous reactions, the dehydration and charring reactions are strongly catalyzed by the presence of acid reagents. The GLC analysis has shown that the acid-catalyzed pyrolysis of cellulose at 350 C gives a pyrolysate containing levoglucosenone (instead of levoglucosan) as the major component and dianhydro-d-glucopyranose, furaldehyde, and hydroxymethyl furaldehyde as minor components [51]. The levoglucosenone formed by the dehydration reactions shown in Fig. 21 can be separated by fractional distillation and is a very reactive compound that can also be obtained by pyrolysis in the presence of mineral acids.

Fig. 20 Pyrolysis model for cellulose. For pyrolysis of pure cellulose under vacuum, the rate constants ki , kv , and kc were found to correspond to ki ˆ 1:7  1021 exp‰ …58:000=RT†Š min 1 , kv ˆ 1:9  1016 exp‰ …47:300=RT†Š min 1 , and kc ˆ 7:9  1011 exp‰ …36:000=RT†Š min 1 .

Thermal Properties

Fig. 21

417

Dehydration of cellulose and glucose derivatives to levoglucosenone.

On further heating, the ®ssion of the sugar units provides a variety of carbonyl, carboxyl, and ole®nic compounds as well as water, carbon monoxide, and char. The analyses of these products are quite similar to those obtained from the pyrolysis of levoglucosan. They also show the effect of acidic salts in producing char and water at the expense of tar and combustible volatile products. The effect of ¯ame retardants is described in Section V.D. Additionally these effects may be readily detected by thermal analysis of the cellulose, particularly thermogravimetry (TG), which shows the amount of volatiles and residues. Figure 22 shows that the pyrolysis of pure cellulose proceeds rapidly at 325±365 C, leaving little char, whereas the cellulose is gradually decomposed in the presence of ZnCl2 at a wider and lower temperature range (150±350 C), leaving substantial amounts of char. C.

Signi®cance

The preceding discussions are mainly based on studies in which cellulose was used as the model substance. In general, the conclusion is also easily adapted to paper and board and other products with cellulose as the main component. The thermal oxidation and the depolymerization of the cellulose together with the catalytic effect of acidic materials will undoubtedly account for the thermal degradation under different conditions. Some caution may still be warranted regarding the general conclusions. At higher temperatures the rapid breakdown of cellulose is likely to give an adequate explanation of the ¯ammability of paper, and the hemicelluloses will behave in a similar manner, as should be expected from their chemical structures. The lignin, however, will form mainly char during the pyrolysis, in contrast to the cellulosic structures, which liberate volatile products. These properties can be observed for TG analysis of wood and its components such as the example shown in Fig. 23. Accordingly, it is reasonable to expect that wood-containing paper will behave differently from paper containing chemical pulp. As pointed out before, it has also been demonstrated that paper containing lignin reacts differently toward long-term storage from chemical pulp, which may be considered to be pure cellulose.

418

Bùhmer

Fig. 22 Thermal analysis curves of (a) pure cellulose and (b) cellulose containing 5% ZnCl2 .

Accepting cellulose as the model, the existence of two different pathways leads to two different modes of combustion as shown graphically in Fig. 24. In the ®rst pathway, which is dominant at lower temperatures, the pyrolysis gives mainly a carbonaceous char and a gas mixture containing water and carbon dioxide, which is not ¯ammable. Oxidation of the resulting char then gives a glowing or smoldering combustion, which is a much slower process, proceeding as a front in the solid state as oxygen becomes available [10,38]. In the second pathway, which is operative at higher temperatures, the pyrolysis of the cellulosic material gives a mixture of combustible gases. These gases will be mixed with air and fuel, and the ¯aming combustion can spread rapidly in the gas phase. This combustion is inhibited by materials that reduce the production of combustible volatiles by enhancing the dehydration and the char-forming reactions. The smoldering combustion is inhibited by chemicals such as boric acid that prevent the oxidation of char to CO2 and produce more CO. The energy release is reduced from 370 kJ/mol for CO2 to 96 kJ/mol for CO.

Thermal Properties

Fig. 23

D.

419

Thermogravimetry of cottonwood and its components.

Theories of Fire Protection

With the preceding combustion processes in mind, it may be useful to survey the effects of various chemicals. By adding different chemicals it is possible to in¯uence several stages of this process. The effect of ®re retardants is often complicated and sometimes even unknown. Their effects may be divided into mechanisms of a physical or chemical nature [5,34,35]. Physical Mechanisms Thermal insulation. Thermal insulation restricts the heat ¯ow to the material to prevent the primary slow pyrolysis. Protective layer. A protective layer restricts the liberation of the pyrolytic gases from leaving the material and oxygen from penetrating it. Such a layer may be attained when added chemicals melt, forming a ®lm on the surface of the material. The chemicals may also crystallize and swell in the pores of the material when ¯ames are started to achieve the same effect. Dilution of gases. Chemicals releasing noncombustible gases may act as ®re retardants in two ways. The released gases may dilute the pyrolytic gases, thereby making them less ¯ammable, or they may lie as a protective layer on the surface of the material. Such gases may be water vapor, ammonia, nitrogen, sulfur dioxide, or carbon dioxide. To get the full effect of these gases it is important that they be released at a temperature below the starting level of active pyrolysis.

420

Bùhmer

Fig. 24

Graphic description of ¯aming and smoldering combustion.

Energy absorption. Flame inhibition may be achieved by using chemicals that will absorb a great amount of energy when heated. This may happen if the chemicals have a great heat capacity or will go through reactions that demand energy. Such reactions include melting, sublimation, dehydration, and chemical conversion. This mechanism may be very effective if it takes place during the slow pyrolysis. Chemical Reactions Addition of ¯ame inhibitors. Flame inhibitors are compounds that react with the free radicals in the ¯ame. In this way the combustion of the pyrolytic gases may be diminished or stopped. Such inhibitors should have a high

Thermal Properties

421

vapor pressure at the temperature prevailing in the ®re. The halogens are examples of such chemicals, and their effects increase with increasing atomic number. Control of the decomposing reactions. The ideal ®re retardant is a compound that would change the pyrolytic process to yield only carbon and water according to the general reaction …C6 H10 O5 †n ! 6n C ‡ 5n H2 O As pointed out, the cellulose is converted to levoglucosan at 180 C; at 270 C the levoglucosan is further decomposed to compounds that will burn. Effective inhibitors work by preventing the formation of glucosan. Protection against afterglowing. Useful chemicals that achieve this effect include products that release phosphoric acid or boric acid at high temperature. Small additions, less than 1%, may give the required protection. Combination of inhibitors. By combining different chemicals a synergistic effect may be obtained. This may reduce the total consumption of chemicals. A synergistic effect is often obtained when the chemicals interfere with different stages of the combustion process or when the ®re-retardant effects are due to different mechanisms. The ef®ciency of ®re retardants is frequently evaluated from their ability to form char and to reduce the formation of tar and volatile gases. A discussion of the way in which these chemicals should be added falls outside the scope of this text, and reference is made to the many reports dealing with this subject [23,35,48] E.

Analytical Methods

The physical changes taking place when paper is decomposed can be measured by thermal analyses, including differential thermal analysis (DTA) and thermogravimetry (TG). The decomposition products can be analyzed by spectroscopic or chromatographic methods including mass spectroscopy and gas-liquid chromatography. These methods are sometimes combined to provide more comprehensive data of the same phenomena. For instance, DTA and TG are combined to measure the heat and mass transfer during pyrolysis. A pyrolysis unit may also be combined with GLC and mass spectrometry to achieve fragmentation of the substrate, separation of the pyrolysis products by GLC, and identi®cation of the individual products by mass spectrometry. Several methods of analytical pyrolysis have been developed for rapid characterization of complex compounds and polymers such as cellulose that cannot be handled by the existing powerful analytical tools unless the sample is reduced to smaller, soluble, or volatile fragments. A detailed description of such equipment and methods is beyond the scope of this handbook, and information is available in related monographs and periodicals [20,21,24,63,64]. The following introduction should, however, make the reader acquainted with the basic principles. Pyrolysis Units The selection of the pyrolysis unit depends on the purpose of the experiment. It may consist of a simple aluminum boat that is heated in a

422

Bùhmer

quartz tube surrounded by a thermally controlled furnace. In this unit the pyrolysis products are swept by vacuum or by an inert gas into a series of condensers where they are recovered as tar or liquid condensate and analyzed separately. The pyrolysis products may also be swept directly into the injection port of a gas-liquid chromatograph for immediate analysis. To avoid overlapping it is essential that the products reach the injection port and start the separation process at the same time. In analytical experiments, the pyrolysis should be executed rapidly to achieve this purpose and also to prevent excessive fragmentation and secondary reactions. A rapid pyrolysis is particularly desirable when it is combined with mass spectroscopy. In these experiments, rapid fragmentation can be achieved by Curie point, Pyroprobe, ribbon probe, or laser beam pyrolyzers. In the Curie-point pyrolyzer, the sample is heated by ®laments of certain ferromagnetic alloys to attain the proper temperature in nanoseconds. At this temperature (the Curie point) a change in conductivity of the alloys automatically prevents further heating. In the Pyroprobe a platinum ®lament, electronically programmed, gives a rapid and linear heating rate, and this also serves as the heating source and sample holder. The probes are directly attached to the mass spectrometer or gas-liquid chromatograph [20,21,24]. Mass Spectroscopy A mass spectrometer is a very sensitive and powerful tool and as little as a few nanograms may be analyzed provided that the material can be vaporized. In this instrument the sample is evaporated, fragmented, and ionized with, for example, an electron beam. The positive ions formed are then separated according to their mass-to-charge (m=z) ratios. The quantitative recording of these mass ions provides a spectrum that is quite characteristic of the individual compounds. As mentioned above, wood and cellulosic materials can be analyzed with a combination of pyrolysis and mass spectrometry. A Curie point pyrolysis and ®eld ionization mass spectrometry are preferred because they minimize further fragmentation of the pyrolysis products. Integrated Curie point pyrolysis and ®eld ionization of hexosans give characteristic ions for polymeric units such as C6 H10 O5 (m=z 162), C6 H8 O4 (m=z 144), and C6 H6 O3 (m=z 126), which correspond with the anhydro sugar derivatives and other products discussed before. Cellulose is distinguished from other hexosans by the ratio of the major peaks and by its thermal stability, which requires a high energy transfer in order to get suf®cient amounts of volatile materials [46]. Xylan gives characteristic peaks at m=z 132, 114, and 96, whereas lignin has peaks at m=z 124, 138, and 150 for the phenolic fragmentation products. Figure 25 demonstrates the ability of this method to distinguish between different products. Gas-Liquid Chromatography Pyrolysis products may be fractionated to gases, liquids, and tars according to their volatility, and they may again be split into individual compounds by liquid chromatography [63,64]. The characteristic feature of this method is that the volatile components are passed through a steel column ®lled with an adsorbent material. The different components of the mixture are separated in the column and exit as separate bands. These bands are detected with special devices and recorded as a series of peaks, each peak corresponding to an individual component. The location of the peak or the time required to travel

Thermal Properties

Fig. 25

423

Pyrolysis and ®eld ionization mass spectroscopy of cellulose, lignin, and redwood.

through the column indicates the chemical structure; the area of the peak is proportional to the concentration. To achieve better separation, different columns are used for gases and liquids, and the tarry products may be made more volatile by converting them to their ether derivatives. Tarry pyrolysis products, which often contain a variety of sugar derivatives, can be treated with trimethylsilyl chloride, which reacts with free hydroxyl groups, before GLC analysis. This helps to separate the less volatile compounds containing hydroxyl groups and avoids interference with other products. The method is also used to analyze the sugars obtained from the hydrolysis of paper and board. Etheri®cation is then not used because the compounds involved are already volatile enough, and some also lack hydroxyl groups. When the compositions of the products are unknown, the pyrolysis and GLC can be combined with mass spectroscopy (PYGC-MS) to give quantitative information about the pyrolysis products [21,24].

424

Bùhmer

Differential Thermal Analysis Differential thermal analysis (DTA) is a method in which the thermal properties of a material are compared with a reference when the two products are heated or cooled at a constant rate and time. When the furnace is heated or cooled the test sample undergoes some transition different from that of the reference. The measuring devices are generally available commercially, and some reports have been published on paper and board [9,10,18,37,38,44,64]. A comprehensive treatise of thermal analysis is outside the scope of this handbook. Reference is instead made to the book by Wunderlich [64], which deals with this topic in detail, and also to the survey presented by Schneider and Tùppel [44]. The latter is particularly valuable because it gives many applications from the pulp and paper ®eld. The study of the thermal decomposition of paper reveals two endothermic peaks, one due to the loss of water at 100±150 C, the other due to the decomposition of paper at 300±370 C. According to an experiment performed in nitrogen atmosphere by Herbert et al. [18], the decomposition peak had no de®nite relation to the pulp species used but was closely related to the pH of the paper. As the pH was lowered, the endothermic peak was shifted to a lower temperature. It is notable that this peak appears at 340±370 C for cellulose, 306 C for xylan, and 430 C for lignin. This seems to support the opinion of Lystad [31], Oye [36], and others showing that lignin may have better stability against heat than was previously assumed. Ramiah [38] showed that the DTA curve in air or oxygen differs from the one obtained in nitrogen, which is not unexpected. Parks [37] also showed that the oxidation of cellulose leads to a slightly destabilization of the modi®ed cellulose, giving a slightly higher endothermic peak temperature. The oxycellulose can be stabilized with calcium ions. Thermogravimetric Analysis Thermogravimetric analysis (TGA) is a method of measuring the change in mass with a thermobalance when the temperature of a material is raised at a constant rate. With most devices on the market, both DTA and TGA can be carried out. Important information regarding the thermal decomposition of the material in question may be obtained by comparing the DTA curve with the TGA curve as shown by Schneider and Tùppel [44] with different polymers. The temperature at which the weight starts to decrease can be regarded as the initial temperature of thermal decomposition. Results published by Ramiah [38] showed that a softwood sul®te pulp started to lose weight at 280 C in an atmosphere of nitrogen, microcrystalline cellulose at 295 C, xylan at 195 C, and Klason lignin at 320 C. According to the results obtained on cellulose by Parks [37], the initial temperature of weight decrease is inversely related to the content of carboxyl groups. An apparent activation energy can be obtained from thermogravimetric analysis [10,37,38]. According to Ramiah [38] the activation energies of cellulosic materials range from 36 to 60 kcal/mol for cellulose, from 15 to 26 kcal/mol for hemicellulose, and from 13 to 19 kcal/mol for lignin. Cardwell and Luner [10] reported activation energies that are signi®cantly greater, varying from 33 kcal/ mol for rayon ®bers to 44 kcal/mol for bleached pine kraft pulp.

Thermal Properties

425

ACKNOWLEDGMENT This chapter is a revision and update of the ®rst edition chapter, ``Thermal Properties,'' written by S. Nakagawa and F. Sha®zadeh.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22.

Arney, J. S., and Jacobs, A. J. (1979). Accelerated aging of paper. The relative importance of atmospheric oxidation. Tappi 62(7):89±91. Arney, J. S., and Jacobs, A. J. (1980). Newsprint deterioration. The in¯uence of temperature on the relative contribution of oxygen-independent and oxygen-dependent processes in the total rate. Tappi 63(1):75±77. ASTM D 2766 (1971). Speci®c heat of liquids and solids. ASTM D 2863 (1970). Flammability of plastics using the oxygen index method. Back, E. L. (1981/82). Autoignition in hygroscopic organic materials. Especially forest products as initiated by moisture absorption from the ambient atmosphere. Fire Safety J. 4:185±196. Back, E. L. (1989). Understanding the effect of hot calendering. Das Papier 43(4):144± 154. Bradbury, A. G. W., Sakai, Y., and Sha®zadeh, F. (1979). A kinetic model for pyrolysis of cellulose. J. Appl. Polym. Sci. 23:3271±3280. Brandrup, J., and Immergut, E. H., eds. (1966). Properties of cellulose materials. In: Polymer Handbook, Vol. 1. Wiley, New York, p. 36. Browning, B. L. (1969). Analysis of Paper. Marcel Dekker, New York, pp. 205±206. Cardwell, R. D., and Luner, P. (1978). Thermogravimetric analysis of pulp: Kinetic treatment of dynamic pyrolysis of papermaking pulps. Tappi 61(8):119±123. Chiang, A. W. (1993). Molded ®ber dual ovenable containers. Tappi 76(5):103±104. Dixson, H. P., Jr., and Nelson, J. C. (1962). An accelerated aging study of several writing papers. Tappi 45(10):753±760. Eriksson, S. AÊ. I., and Back, E. L. (1966). Laboratory method for the determination of the auto-oxidative heat of organic board materials in a gas stream. Svensk Papperstidn. 69:300±304 (in Swedish). Gùtze, W., and Winkler, F. (1967). Calorimetric studies on textile ®brous materials. 1. Speci®c heat: Literature review. Faserforsch. Textiltechn. 18(3):119±123 (in German). Gùtze, W., and Winkler, F. (1967). Calorimetric studies on textile ®brous materials. 2. Measuring methods and apparatus. Faserforsch. Textiltechn. 18(5):222±227 (in German). Gùtze, W., and Winkler, F. (1967). Calorimetric studies on textile ®brous materials. 3. Adiabatic calorimeter AKM 1. Faserforsch. Textiltechn. 18(5):292±295 (in German). Han, S. T., and Ulmanen, T. (1958). Heat transfer in hot-surface drying of paper. Tappi 41(4):185±189. Herbert, R. L., Tyron, M., and Wilson, W. K. (1969). Differential thermal analysis of some papers and carbohydrate materials. Tappi 52(6):1183±1188. Iimure, N. (1958). Ignition temperature. In: Manual for Dangerous Article Treating (Kikenbutsu Toriatsukai Hikkei). Sangyo Tosho Co., Tokyo, pp. 62±63 (in Japanese). Irwin, W. J. (1979). Analytical pyrolysis: An overview. J. Anal. Appl. Pyrolysis 1:3±25. Irwin, W. J. (1982). Analytical Pyrolysis: A Comprehensive Guide. Chromatogr. Sci. Series, Vol. 22. Marcel Dekker, New York. Ishii, K., Sekigushi, T., and Takaya, T. (1972). Thermal properties and ¯ammability of various ®bers. J. Soc. Fiber Sci. Technol. Jpn. (Senni Gakkaishi) 28(9):359±367 (in Japanese).

426

Bùhmer

23. Johansson, I., and Johansson, S. (1982). Chemical aspects of wood technologies. STFI Commun. Ser. A, No. 772. STFI, Stockholm. 24. Jones, C. E. R., and Cramer, C. A., eds. (1977). Analytical Pyrolysis. Elsevier, Amsterdam. 25. Kartovaara, I., Rajala, R., Luukkala, M., and Sipi, K. (1985). Conduction of heat in paper. In: Papermaking Raw Materials. V. Punton, ed. Mech. Eng. Pub., London, pp. 381±412. 26. Keller, S. (1994). Heat transfer in a calender nip. J. Pulp Paper Sci. 20(1):J33±J37. 27. Kerekes, R. J. (1979). Heat transfer in calendering. Pulp Paper Mag. Can. Trans. Tech. Sect. CPPA 5(3):TR66±TR76. 28. Kirk, L. A., and Tatlicibaci, C. (1972). Measurement of thermal conductivity of paper by a heat pulse method. Tappi 55(12):1697±1700. 29. Klason, G., and KubaÂt, J. (1976). Thermal transitions in cellulose. Svensk Papperstidn. 79(15):494±500. 30. KubaÂt, J., Martin-LoÈf, S., and SoÈremark, C. (1969). Thermal expansivity and elasticity of paper and board. Svensk Papperstidn. 72(23):763±767. 31. Lystad, A. E. V. (1993). Ageing stability of high-yield pulps. Proc. Int. Mech. Pulping Conf., Oslo, pp. 340±349. 32. Mitsuya, T., and Kumasaka, T. (1992). Heat transfer and toner melting in an electrophotographic fuser system. J. Imaging Sci. Techn. 36(1):88±92. 33. Olsson, A.-M., and Back, E. L. (1989). On the speci®c heat of paper and wood products between 180±250 C. Nordic Pulp Paper Res. J. 4(4):258±262. 34. Ostman, B. A.-L. (1980). Status of ¯ame-protection of wood-based board. STFI Ser. A, No. 621. STFI, Stockholm, pp. 1±19 (in Swedish). 35. Ostman, B. A.-L. (1986). Flammability of cellulose. Mater. Sci. Monographs No. 36. Marcel Dekker, New York, pp. 269±279. 36. Oye, R. (1991). Wood Processing and Utilization. Wiley, New York, pp. 373±379. 37. Parks, E. J. (1971). Thermal analysis of modi®ed pulp. Tappi 54(4):537±544. 38. Ramiah, M. V. (1970). Thermogravimetric and differential thermal analysis of cellulose, hemicellulose and lignin. J. Appl. Polym. Sci. 14(5):1323±1337. 39. Ramiah, M. V., and Goring, D. A. I. (1965). Thermal expansion of cellulose, hemicellulose and lignin. J. Polym. Sci., C: Polym. Symp. 11:27±48. 40. de Ruvo, A., Lunberg, R., Martin-LoÈf, S., and SoÈremark, C. (1976). In¯uence of temperature and humidity on elastic and expansional properties and the constituent ®ber. In: Fundamental Properties of Paper Related to Its Uses. F. Bolam, ed. British Paper and Board Industry Foundation, London, pp. 785±810. 41. Salmen, N. L., and Back, E. L. (1980). Moisture-dependent thermal softening of paper evaluated by its elastic modulus. Tappi 63(6):117±120. 42. Samuelson, M. L. (1988). Study of copy papers. Governmental Testing Lab., Sweden, pp. 1±38 (in Swedish). 43. Sanders, D. J., and Forsyth, R. C. (1983). Measurement of thermal conductivity and contact resistance of paper and thin-®lm materials. Rev. Sci. Instrum. 54(2):238±244. 44. Schneider, R., and Tùppel, O. (1971). The use of thermoanalysis in the pulp and paper industry. Das Papier 25(12):849±860 (in German). 45. Scho®eld, F. H., and Hall, J. A. (1927). Thermal insulating materials for moderate and low temperatures. In: International Critical Tables of Numerical Data, Physics, Chemistry and Technology, Vol. 2. McGraw-Hill, New York, pp. 312±316. 46. Schulten, H. R., Bahr, U., and GoÈrtz, W. (1982). Pyrolysis ®eld ionization and mass spectrometry of carbohydrates. J. Anal. Appl. Pyrolysis 3:229±241. 47. Sha®zadeh, F. (1968). Pyrolysis and combustion of cellulosic materials. Adv. Carbohydr. Chem. 23:419±474.

Thermal Properties

427

48. Sha®zadeh, F. (1975). Industrial pyrolysis of cellulosic materials. Appl. Polym. Symp. 28:153±174. 49. Sha®zadeh, F. (1982). Introduction to pyrolysis of biomass. J. Anal. Appl. Pyrolysis 3:283±305. 50. Sha®zadeh, F., and Bradbury, A. G. W. (1979). Thermal degradation of cellulose in air and nitrogen at low temperature. J. Appl. Polym. Sci. 23:1431±1442. 51. Sha®zadeh, F., Furneaux, R. H., Stevenson, T. T., and Cochran, T. G. (1978). 1,5Anhydro-4-deoxy-D-glycero-hex-1-en-3-ulose and other pyrolysis products of cellulose. Carbohy. Res. 67:433±447. 52. Smith, R. D. (1970). The non-aqueous deacidi®cation of paper and books. Ph.D. Thesis, Univ. Chicago. 53. Sutermeister, E. (1948). Chemistry of Pulp and Papermaking. Wiley, New York, pp. 31± 34. 54. TAPPI T 461 (1989). Flammability of treated paper and board. 55. TAPPI T 1000 (1983). Thermal conductivity of materials by means of the guarded hot plate. 56. Terada, T., Ito, N., and Goto, Y. (1969). Effective thermal conductivity of insulating paper. Jpn. Tappi 23(5):191±197 (in Japanese). 57. Terasaki, K., and Matsuura, K. (1972). The study of heat properties of papers and consideration to use for heat exchangers. Jpn. Tappi 26(4):173±178 (in Japanese). 58. Terasaki, K., and Matsuura, K. (1972). The study of the effective thermal conductivities of papers for temperature and humidity, 2nd report. Jpn. Tappi 27(10):511±515 (in Japanese). 59. Terasaki, K., Matsuura, K., and Okada, M. (1973). The study of the effective thermal conductivity of papers, 3rd report. Jpn. Tappi 27(11):525±529 (in Japanese). 60. Tokyo Astronomic Observatory (1979). Physical Table 63. In: Annual Scienti®c Tables (Rikanenpyo), Maruzen, Co., Tokyo (in Japanese). 61. Underwriters Laboratories (1972). UL 94. Tests for ¯ammability of plastic materials. 3. Vertical burning test for classifying materials. 62. Underwriters Laboratories (1972). UL 94. Test for ¯ammability of plastic materials. 7. Ignition temperature test. 63. Willard, H. H., Merrit, L. L., Jr., Dean, J. A., and Settle, J. A., Jr. (1981). Instrumental Methods of Analysis, 6th ed. Van Nostrand, New York. 64. Wunderlich, T. (1990). Thermal Analysis. Academic Press, New York.

11 CHARACTERIZATION OF PAPER SURFACES USING OPTICAL PROFILOMETRY Ê GBERG PIA WA SCA Packaging Research Sundsvall, Sweden Ê KE JOHANSSON PER-A STFI, Swedish Pulp and Paper Research Institute Stockholm, Sweden

I. Introduction II. Surface Roughness

430 430

III. Surface Pro®le MeasurementsÐInstrumentation A. History B. Overview of Available Techniques C. Autofocusing D. Triangulation E. Interferometry F. Topographical Scanning Electron Microscope

432 432 433 435 437 438 440

IV. Data Analysis A. Imaging the Surface B. Numerical Analysis C. Two-Dimensional Versus Three-Dimensional Data Analysis D. Basic Descriptors E. Spectral Density Analysis F. Slope and Facet Statistics G. Fractal Descriptors

441 441 442 442 443 445 445 445

V. Applications A. Surface Roughening B. Gloss and Gloss Variation C. Printability

446 446 447 447 429

WaÊgberg and Johansson

430

D. Microstriations E. Miscellaneous VI. Conclusions References I.

447 447 448 448

INTRODUCTION

The paper surface plays an important role in ink transfer to the paper, which in turn is an important part of the printability of the paper. The surface has therefore always been very important for paper producers, but its signi®cance has increased in recent years because technological developments in the printing industry have been rapid. More colors are used in printing today and higher demands are put on papers of very different grades, not only on the typical graphic papers but also on rougher qualities such as linerboard and different types of paper containing recycled ®bers. Because the paper surface structure is so important to the papermaker, many attempts have been made to measure this property accurately. In the early days the most commonly used methods were based on air leak instruments or other indirect measurements [1]. These air leak methods are still the most commonly used roughness testers owing to their simplicity. However, as quality demands have increased, the demand for a better understanding of the nature of the surface structure has also increased, especially a need to understand the quality differences at different structural levels. This means that there is an evident need, at least in an R&D environment, for a more detailed characterization of the surface than can be achieved by air leak instruments. Several excellent books and review articles have been written in the area of surface characterization, especially in the ®eld of machine ®nishing and bioengineering. Thomas edited a book in 1982 that contains a detailed discussion of stylus instruments and how the surface data can be statistically evaluated [2]. Bennet, who has devoted an extensive part of her research to surface characterization, has written comprehensive books [3,4] on the subject as well as review articles [5]. A group in England at the University of Birmingham is working on the characterization of bioengineering surfaces, and their 1995 review focuses on a number of new measuring systems and their associated numerical characterization methods [6]. The ®eld of optical pro®lometry is not, however, fully covered in these publications, especially not in the ®eld of paper characterization. This chapter gives an overview of currently available pro®le measurements that are of interest for characterizing the paper surface. The emphasis is on non-contacting optical pro®le measurements. This chapter also provides a de®nition of surface structure and summarizes different ways of evaluating the surface roughness, in both two dimensions and three dimensions. Examples of applications are given in Section V. II.

SURFACE ROUGHNESS

Before giving an overview of different ways of characterizing surface structure, a de®nition of the nature of a surface is needed. In principle, the surface has two main attributes: the vertical height (depth) an the horizontal ``lateral'' dimensions. Figure 1 shows a schematic representation of three scales of surface structure: A, roughness;

Characterization Using Optical Profilometry

431

Fig. 1 A schematic representation of three scales of surface structure. Note: This is only a two-dimensional description. To understand the nature of the entire surface, a third dimension must also be taken into account. (From Ref. 6.)

B, waviness; and C, form. These three scales seem to be generally accepted, although their dimensions may differ between different authors and their particular applications. In an attempt to distinguish between the lateral dimensions of different machine ®nishing or optical surfaces, Bennet put a few examples together with a spatial wavelength scale in a diagram depicted in Fig. 2 [5]. In this diagram the different measuring techniques suitable for the different dimensions are also included. Not all the measuring techniques listed in this graph are covered in this chapter, but they are all described in Ref. 3. To understand what surface dimensions we are interested in for paper surface characterization, some important components have been added to the original graph by Bennet. As can be seen, paper in fact covers a large range of wavelengths. This means that it is dif®cult to characterize the paper surface. It is most challenging to characterize the effect of different pigments in the submicrometer to micrometer range and to see the effect of, e.g., the calendering process on the ®ber network covered by the millimeter range. There seems to be no optimal instrument, at least in Bennet's diagram, that could characterize every aspect of a paper surface. Another large problem in characterizing paper surfaces is, as all papermakers know, the matter of variability in the papermaking process. What is the smallest measured area that we can accept as being representative for the particular paper quality in question? Can we rely on a characterization made on measurements on an area of 1, 10, or 100 mm2 when we know that the property varies quite extensively over the cross direction of a paper machine several meters wide? In fact, many paper properties are governed by the raw materials, and it is thus often suf®cient to measure the surface in the millimeter range (compare the scales in Fig. 2). It is, however, important to establish the stability over larger ranges through statistical measurements. The fact that the results of a surface measurement can be quite different depending on the measured scale has been pointed out by Mangin [7]. He also pointed out that the surface compressibility plays an important role, but as the

WaÊgberg and Johansson

432

Fig. 2 Some examples of common surfaces compared with the spatial wavelength. The capability of some of the measuring techniques is also included as a comparison. The different dimensions that are encountered when measuring paper were added to the original graph by the authors. (Modi®ed from Ref. 5.)

aim of this chapter is to give an overview of non-containing pro®le measurements, this aspect will not be considered here. For machine-®nished surfaces, surface topography is so tied to the process that its measurement provides an excellent means of control. Although this is not altogether true of paper surfaces, a lot can be understood both about interactions in the printing press and about the in¯uence of different processes in the paper machine. The relationship between the surface and its functional performance has been studied for two-dimensional pro®les, but to gain a full understanding it is necessary to characterize the surface in three dimensions. Measurements of three-dimensional paper structure also generate data useful in modeling, e.g., in relation to interactions with liquids.

III.

SURFACE PROFILE MEASUREMENTSÐINSTRUMENTATION

A.

History

In the ®eld of precision engineering and the production of optical lenses, etc., the need for a detailed characterization of surfaces has been of major importance since the beginning of this century. According to Thomas [2], the ®rst modern pro®ling instrument was the Abbott pro®lometer, reported in 1933. The ®rst methods provided a

Characterization Using Optical Profilometry

433

simple graphical output from which peak-to-valley height information could easily be obtained. This early type of surface measurement included only two-dimensional tracks measured with different types of mechanical styluses. Quite a few measuring techniques have been reported since then, and most of them up to 1982 are summarized in a lucid way in the book by Thomas [2]. In the 1950s, interferometry was tried as a method of characterizing machined surfaces. The ®rst attempts at interferometric measurements were made on very smooth surfaces and were not very interesting for paper characterization purposes. However, Katagiri et al. [8] used a microinterferometer to characterize paper surfaces as early as 1961. They looked at features of the roughness of different papersÐvarnished art paper, brushed art paper, and castcoated paperÐbut the interpretation of the interferograms in those days was timeconsuming (without computers!), and the only conclusion was that the measured features seemed to be closely related to the visual smoothness. In the past 10 years or so, interferometric instruments that use white light as the light source have been applied to paper [9±13]. This type of technique enables detailed measurements of the paper surfaces to be made with high resolution. Because a microscope is used for the measurement, there is always a trade-off between the lateral resolution and the area that can be measured. This problem has been overcome by stitching together smaller areas measured with high resolution [14]. During the late 1960s surface measurement received a signi®cant boost due to the introduction of digital computers, but apart from some exceptions mentioned above, it was not until the 1970s that surface pro®le measurements (mainly mechanical styluses) were applied to paper surfaces. There are, however, several problems in using the mechanical stylus instruments for paper, because paper is such a soft and easily compressed material [15±17]. However, the ¯exibility of these instruments opened a new era for paper surface characterization. The optical autofocusing pro®lers were developed mainly during the 1980s. The prerequisite for all these devices was the rapid development of digital computers, especially for three-dimensional measurements. Most of the surface structure measurements have advantages and drawbacks. Some of these will be mentioned in this chapter. B.

Overview of Available Techniques

Table 1 is an attempt to summarize the different categories of surface pro®le instruments that are available for paper characterization. The resolution ®gures are approximate, because the manufacturers are constantly upgrading their instruments and because vertical and horizontal resolution are not always interpreted in the same way. The ®gures in Table 1 represent the maximum resolution in the x and the z directions, but they say nothing about the maximum area that can be measured at this resolution. Large area scales are of interest for paper characterization owing to the variability of the matrix. From the resolution data in Table 1, it is evident that each instrument has a range for which it is best suited. To obtain a better overview of the speci®c range each instrument covers, an amplitude±wavelength plot is presented in Fig. 3. Each block in this diagram represents the working area of a particular instrument, and it is easy here to obtain an approximate idea of the strengths and limitations of the different instruments.

b

These methods are addressed in Chapter 5. This is not an all-inclusive list.

SEM (2±4 detectors)

Electron-based Topographic

a

Laser in focus

Phase shifting White light int. White light int. Nomarsky

Inductive Inductive Conducting probe tip

Measuring principle

Triangulation Autofocusing

Optical Microscope (vertical scanning)a Microscope (interferometry)

Mechanical Low resolution High resolutiona

Instrument

> 6 nm

? > 0:3 m > 0:48 m > 0:36 m > 0:2 m > 1 m > 1 m > 2 m

CSLM,TSM WYKO Topo-3D WYKO RST Zygo Maxim 3-D MP2000 (Chapman) ESPRI, UNISCAN UBM Rodenstock ERA 3000

> 0:1 m 0.1 nm 0.1 nm

Lateral

> 1 nm

? > 1 AÊ > 1 AÊ > 0:5 AÊ > 1 AÊ > 0:2 m > 6 nm > 0:2 m

> 0:1 nm 2 pm 2 pm

Vertical

Approx. resolution

Talystep, Talysurf, Paperscape AFM STM

Examples of instruments

Table 1 An Overview of Surface Characterization Methods that Can Be Used for Paper

34,35

18,19 11 11,14 13 3 30,31 20,25 21

3,5 18,19 6

Ref.b

434 WaÊgberg and Johansson

Characterization Using Optical Profilometry

435

Fig. 3 Amplitude versus wavelength plot for a range of surface measurement systems. (Redrawn from Ref. 6.)

The emphasis in this chapter is on non-contacting pro®lometry, and newly developed optical non-contacting pro®le measurement techniques are probably more suitable than mechanical pro®ling methods for characterizing the paper surface topography. Three main categories will be covered: autofocusing, triangulation, and interferometry. A fourth technique, topographical scanning electron microscopy, will be mentioned only brie¯y, because its use as a pro®lometer is unusual. C.

Autofocusing

The autofocusing technique is based on the same technology as that employed in the pickup in a CD player and has been developed and re®ned over the past 15 years. The principle is described in Refs. 20±23. The general principle is similar to that of the mechanical stylus instrument, the difference being that instead of the diamond tip stylus (commonly 5±10 m in radius) there is a laser beam approximately 1 m in diameter that gives higher resolution. The system is schematically shown in Fig. 4. A laser diode 1 emits a laser beam with a wavelength of about 780 nm. The laser beam is sent through the beam splitters 2 and 3, through the collimator lens 9 and the objective lens 10, and focused onto the surface that is to be measured 13. Part of the light is re¯ected back through the lens system and the beam splitter, and further to the photodiode 5, where the focus of the laser beam is continually checked as the laser scans over the paper. To adjust the focus, the lenses are moved; this movement is recorded and represents the surface pro®le. The remainder of the re¯ected light goes through the beam splitter 3 and is recorded on a CCD camera 16 to provide an image of the surface re¯ectance at the same time as the pro®le is being recorded. Whereas the vertical movement in a mechanical pro®ler is recorded

436

Fig. 4

WaÊgberg and Johansson

A schematic overview over an autofocusing pro®lometer. (From Ref. 24.)

through an inductive transducer, the autofocusing pro®ler uses photodiodes that register whether or not the laser beam is in focus. This can be done by a critical angle prism, for example, as shown in Fig. 5. After passing through the objective lens, the re¯ected laser light becomes a parallel ¯ux if the surface under test is in focus at position B. A total re¯ection prism is placed to re¯ect the light at the critical angle. If the surface is out of focus, i.e., close to the lens, position A, the light ¯ux diverges slightly and the condition of total

Fig. 5

Principle of measurement of surface deviations. (From Ref. 23.)

Characterization Using Optical Profilometry

437

Fig. 6 A comparison between a surface measurement made by an autofocusing pro®lometer (UBM), to the right, and a scanning electron microscope (backscatter mode), to the left. The measured surface is an LWC paper covered with gold. In order to measure the same area, the paper has been marked with a pinhole, which can be seen in the upper right-hand corner in both images (black spot).

re¯ection is lost at the upper side of the optical axis, and vice versa. Therefore the differential output of two photodiodes gives the amount of surface deviation from the focal plane. Several comparisons between this autofocusing device and the mechanical pro®lers can be found in the literature [16,25,26]. Although this device is very ¯exible, does not damage the surface, and provides a more detailed measurement than most mechanical pro®lers, it also has its drawbacks. The measurement process is quite slow despite being non-contact optical, because for every measuring point the device mechanically adjusts the lenses so that the laser beam is in focus. Recent improvements of the autofocusing device, however, have made it possible to measure at higher speeds. The German company UBM [27], for example, can measure in a ``vibration-compensated mode'' with a measuring speed of 140 Hz as well as in a ``fast mode'' with a speed as high as 1200 Hz. In comparison, the mechanical pro®lers measure at a speed of approximately 100 Hz. An area with 512  512 measuring points will be measured in about 29 min if the slow mode is used but in less than 4 min if the fast mode is used. Figure 6 is an example of the result of a measurement made with the UBM autofocusing equipment in comparison with the result of a surface measurement made by a scanning electron microscope (SEM). Another drawback is that the focus detection system requires a ®nite amount of light, although that is usually no problem for paper characterization, because paper scatters light quite well. When there are a lot of steep slopes in a surface, however, the laser spot is unable to ®nd the focus. This may lead to incorrect spikes or sharp pits in the surface data. D.

Triangulation

Triangulation is a commonly used technique with a quite easily understood principle. Triangulation devices are often put together by the users themselves from components off the shelf, but it is also possible to buy a ready-made system. Figure 7 shows the basis of the measurement technique.

WaÊgberg and Johansson

438

Fig. 7 Principle of a triangulation pro®lometer with a laser sensor. A local height change (
Z) is converted via triangulation into a
X change on the matrix of light-sensitive pixels. The light-sensitive pixels convert the results into height data, allowing further processing. (From Ref. 28.)

A light source sends out a beam of light toward the surface of the substrate to be measured. The system is based on the general principle that a beam of light is focused on a small region of the surface, which is inclined at a speci®c angle. The light which is specularly or diffusely scattered (depending on the type of device used), is collected and focused onto a position-sensitive detector, normally a CCD array. The choice of device will be in¯uenced by the characteristics of the surface to be measured. Typically if the surface is highly re¯ective, then a specular device should be used; if not, then a diffuse device may well be the better option. Positions along the re¯ected image, as observed on the detector, depend on the heights from which they are re¯ected. The light source is usually a laser diode with a de®ned wavelength (depending on the resolution that needs to be obtained [24,28,30]). Lipshitz et al. [31,32] used a triangulation method for measuring coated surfaces where the light source was a noncoherent light rather than a laser diode light source. A drawback with the triangulation method using a laser diode as the light source is that the working range is limited by the wavelength of the laser. When a paper surface is measured there is also a limitation to the resolution of the measurement due to the light scattering of the ®brous material. Assoul et al. [33] reported a re®nement of the triangulation method using two detectors. The technique was used for skin surface characterization, but it could probably be used for paper applications as well. By using a two-detector system the effect of dissymmetry is reduced. E.

Interferometry

Interferometry is a technique that has been used for several years for the measurement of supersmooth surfaces. The principle is based on different types of interfe-

Characterization Using Optical Profilometry

439

rometers (Michelson, Mireau, etc.), where the path lengths of two beams of light are compared, one of which has been re¯ected from the specimen and the other from a reference surface. The interference patterns produced are digitally analyzed by the application of certain algorithms. The two most widely used techniques today are phase shift interferometry and vertical scanning interferometry [6,10,11,13,14]. Vertical scanning uses a narrowband light source with a relatively long coherence length that allows rough surfaces like paper to be measured, whereas the phase shift type uses a broadband light source with short coherence length to measure smooth surfaces. Scanning perpendicular to the surface is accomplished with a piezoelectric transducer whose position can be monitored via changes in the optical path length. An example of such an instrument is shown in Fig. 8. The measurement is quick, because the device can instantly acquire an image of the surface. It is based on a microscope with different lenses that can be interchanged. For measuring paper surfaces there is, as mentioned before, a trade-off between lateral resolution and the area of the surface that can be measured. This is overcome, however, by combining computer-controlled stages with interference microscopes and stitching together many small areas with high resolution to large ®eld-of-view areas with equally high resolution [14]. One example of a stitched area can be seen in Fig. 9. The measured area is the same gold-covered LWC paper as is depicted in Fig. 6. A major drawback with this type of instrument is that it requires surfaces with relatively high surface re¯ectance (higher than autofocusing). Rapid

Fig. 8 An example of an interferometric pro®lometer system based on a Mirau interferometer. (From Ref. 24.)

WaÊgberg and Johansson

440

Fig. 9 A gold-covered LWC paper measured with a white light interferometer. It is the same sample that is depicted in Fig. 6. Note that the mark from the needle can be seen in the upper right-hand corner. (Courtesy of AG Electro-Optics, Cheshire, England; Wyko interferometric pro®lometer.)

slope changes may sometimes cause a problem, and vibration isolation is also required, because the instrument is sensitive to vibrations in the environment. Manufacturers are aware of these problems and are re®ning their instruments accordingly. F.

Topographical Scanning Electron Microscope

The topographical scanning electron microscope (topo SEM) is not an instrument that usually belongs to the group of surface pro®lometers, but it may be interesting to consider this technique brie¯y together with the others to compare the differences. After all, this is a technique that provides a three-dimensional surface measurement with high resolution. Those who were working with the conventional SEM technique originally developed topo SEM for biological surfaces in the 1980s. The Japanese company Elionix Inc. was the ®rst developer and is still the biggest provider of these instruments. Enomae [34,35] was the ®rst to apply the technique to paper. The principle of the topo SEM system is easily understood by those who are familiar with the SEM technique (chapter 5). A conventional SEM has one electron detector for secondary electron images. To achieve a 2D image, two detectors are required; and to obtain a 3D image of the specimen, four detectors are required. Extra equipment is needed for the interpretation of the output data, such as a circuit for calculating or subtracting the outputs from the detectors. The principle for constructing the topographic image is illustrated in Fig. 10, where a system with only two detectors is depicted (for simpli®cation). The sum of the two detector outputs provides an image equivalent to a conventional SEM image derived from a single detector, and their difference produces an image sensitive to the surface topography of the specimen. From these images, it is possible to obtain a detailed topographical image with high resolution. The system measures quickly, but, as with the conventional SEM, the specimen needs to be gold

Characterization Using Optical Profilometry

441

Fig. 10 The principle for constructing the topographical image in a topo SEM is shown for simplicity with only two detectors, whereas in reality a four-detector system is used for a threedimensional image. (From Ref. 34.)

plated before the measurement. Another drawback of the system is the restricted area that can be measured. The usual measurement area, according to Enomae, is 240 m (x direction) by 180 m (y direction). The commercially available instruments (mentioned in Section III) all employ software that permits various roughness parameters to be calculated and used for pro®les as well as a graphical output. Figure 11 is an example of the latter.

IV.

DATA ANALYSIS

A.

Imaging the Surface

As far as the characterization of surface structure is concerned, it is not enough to be able to acquire nice colorful maps of the surface. The numerical analysis of the actual data is another area where the ¯exibility is great, and it can thus sometimes be dif®cult to select the most relevant parameter for a particular application. The result of a measurement of surface topography is normally a three-dimensional data set of height readings z at the lateral coordinates x, y: z ˆ z…x; y† This data set lends itself to visual rendering in different ways. One way is a straight perpendicular image in which gray values represent height. A more illustrative image can be made by an inclined 3D projection that can be enhanced with calculated shadows and highlights to give an intuitive feel for the topography of the surface (compare Fig. 11). A third possibility is the calculation of stereo-pair projections to be viewed using left±right eye separation.

WaÊgberg and Johansson

442

Fig. 11 An example of the surface image that can be obtained from a modern optical pro®ler thanks to the extensive software that is usually provided with the system. (Courtesy of AG Electro-Optics, Cheshire, England; measured with a Wyko interferometric pro®lometer.)

B.

Numerical Analysis

Besides visual evaluation, there is, of course, a need for a compact set of parameters, ultimately one number, to characterize the surface roughness in order to simplify comparisons of different surfaces. The numerical analysis of surface structure data is an area with great ¯exibility, and it is thus not always easy to choose the most relevant parameters for a particular application. In the machine ®nishing industry, quite a few standards have been developed over the years. The variation in nomenclature among different countries introduces further confusion [2]. Attempts have been made in both the two-dimensional and three-dimensional data analysis to bring order to the many roughness parameters. See also the relevant ISO standards [45±47]. C.

Two-Dimensional Versus Three-Dimensional Data Analysis

By two-dimensional data, we mean a height pro®le z ˆ z…x† as opposed to the threedimensional case z ˆ z…x; y†. Several surface-descriptive parameters have been developed for the two-dimensional case. Many of these, however, can easily be generalized to three dimensions, as will be shown later. The two-dimensional pro®les can be extracted from a three-dimensional data set, and three-dimensional data are in fact often acquired as successive pro®les from line scanning devices. In such cases it can be useful to make the pro®les with a greater separation in the y direction to be able to scan a greater area of the sample

Characterization Using Optical Profilometry

443

and to achieve better statistics without excessive data storage and computation. However, true three-dimensional data sets make it possible to make unbiased studies of textures and orientation of features in the surface structure. D.

Basic Descriptors

The following de®nitions are given for two-dimensional pro®les, but they can in most cases be generalized to three dimensions by performing a summation over both x and y coordinates. One factor of relevance to all statistical descriptors is the degree of pretreatment of the measured data to reduce irrelevant slopes or large-scale waviness. Normally some form of long-wavelength cutoff will be applied to reduce the in¯uence of such effects on the measured pro®le. The cutoff frequency used should be reported because it will in¯uence other parameters. Peak-to-Valley Height This primitive parameter, sometimes denoted Ry , is simply calculated as the vertical difference between the highest and lowest points of the pro®le: Ry ˆ zmax …x†

zmin …x†

…1†

This parameter is, of course, statistically unsafe and not very descriptive, because it relies on only two points. Ten-Point Height Ten-point height Rz is the difference between the mean of the ®ve highest peaks and the mean of the ®ve lowest valley points of the pro®le. This value is obviously somewhat more stable than the single peak-to-valley height, Ry . In many cases the de®nition of Rz includes the division of the scanned pro®le into ®ve equal segments in each of which the peak-to-valley height is determined. The Rz value is then taken as the average of these readings, which means that a compensation for waviness and slopes will take place, and this method thus tends to give lower values than the straight ten-point height. Average Deviation 1X z z ˆ N

Given a mean plane

the arithmetic average deviation is calculated as 1 X R2 ˆ zi z N

…2†

…3†

The average deviation Ra is less sensitive to single extreme height values than the peak-to-valley ®gures. Standard Deviation or Root-Mean-Square The root-mean-square RS is calculated as the square root of the average of squared deviations,  X 1=2 1 2 RS ˆ …zi z† …4† N

WaÊgberg and Johansson

444

RS , sometimes denoted Rrms (root-mean-square) is thus the standard deviation, , of the height readings. Because of the squared deviations in the formula, RS is more sensitive to large deviations than is the average deviation Ra . Higher Moments: Skewness and Kurtosis R2S is the variance of the surface pro®le and is also called the second moment. Higher moments than the second can also be used to characterize the pro®le.   1 1 X … zi R3s N   1 1 X Kurtosis ˆ 4 … zi Rs N

Skewness ˆ

z†3

…5†

z†4

…6†

The skewness expresses ``skewed'' deviations from a symmetrical distribution of peak and valley deviations in the pro®le. Positive skewness indicates that more peaks or higher peaks than valleys are present. The opposite, i.e., more or deeper valleys, is indicated by negative skewness, while a value of zero indicates a symmetrical pro®le. Kurtosis expresses the abundance of larger deviations relative to the Gaussian normal distribution. More extreme values give a kurtosis > 3, a normal distribution gives kurtosis ˆ 3, and values < 3 indicate a more concentrated or compact distribution of height values with fewer extremes than in the normal distribution. Some authors include the subtraction of 3 in the de®nition of kurtosis to achieve the value zero for a normal distribution. Bearing Ratio The bearing ratio is the amount of the cumulated pro®le that is reached by a line at a certain depth below the highest point of the pro®le. It can thus be seen as the distribution of porosity into the surface (see Fig. 12). A drawback of the bearing ratio description, especially for non-contact measurements, is its dependence on the highest detected point if that point is taken as the base for the depth values. A better choice is to count from the mean level where the bearing ratio is 50%.

Fig. 12 The bearing ratio is the cumulated amount of the pro®le that is reached by a line at a certain level.

Characterization Using Optical Profilometry

E.

445

Spectral Density Analysis

The wavelength of a periodic surface pro®le is simply the peak-to-peak distance along the surface pro®le. The corresponding spatial frequency is the number of wavelengths per length unit. A wavelength of 1 mm is thus equivalent to a frequency of 1000 m 1 . The aim of spectral analysis is twofold: 1. 2.

To ®nd the general size distribution of disturbances in the pro®le. To detect periodic structures in the pro®le.

A so-called power spectrum for the pro®le can be calculated by fast Fourier transforms (FFTs), which are available in most program packages for signal/image analysis. The power spectrum P…f † expresses the amount of variance as a function of the spatial frequency f : P…f † ˆ  2 …f † ˆ R2s …f †

…7†

The variance between certain limits of frequency can be achieved by integrating the power density function between these limits. Because the frequency (and thus the wavelength) range often covers several orders of magnitude, it is practical to plot the power spectrum on a logarithmic frequency or wavelength axis. By plotting the power (ˆ variance) integrated over relative bands instead of linear bands, the contributions from different parts of the spectrum become easier to grasp. A suitable set of wavelength bands may be an octave series (successive doubling of wavelength): 1, 2, 4, 8 mm in wavelength or 1000, 500, 250, 125 m 1 in frequency. The unit will then be variance per octave band, measured in square micrometers. As an alternative to the variance, the standard deviation ( ˆ RS ) can be plotted bandwise in the same fashion. The spectral information can be divided into narrower bands for better visibility of sharp peaks. In audio signal analysis, for example, one-third octave is a common division. Figure 13 gives an example of measurements of several different paper grades where the variance has been plotted over wavelength bands that constitute a series of octaves starting with 20 m. The same approach can be used for related structural features of paper such as formation or optical inhomogeneities [48]. F.

Slope and Facet Statistics

If a pro®le is divided into a chain of straight segments, it can be described by the distribution of slopes of these segments. In a three-dimensional case, ®tted planar ``facets'' correspond to the segment, and the orientation distribution characterizes the surface. Such descriptions are useful for investigations into gloss properties, because the specular re¯ection of light is directly in¯uenced by the abundance and slopes of such facets. G.

Fractal Descriptors

Fractal descriptions have been used for pro®le characterization [44]. The measure can brie¯y be described as the rate of increase in pro®le length with increasing lateral

WaÊgberg and Johansson

446

Fig. 13 An example of a spectral density plot of data for several different paper grades. (From Ref. 16.)

resolution. A prerequisite for a pro®le characterization by a fractal dimension is that the pro®le repeats its character at increasing resolution. This has not, however, been shown to be the case for paper surfaces [36].

V.

APPLICATIONS

All the above-mentioned pro®lometer systems provide useful information about the surface structure of different substrates that cannot be achieved by conventional air leak techniques. When studying the literature on surface measurement in general and optical pro®lometers in particular, it is striking how often new measuring techniques for surface evaluation are offered. Most of these published papers present a detailed measuring technique, but they often restrict themselves to comparing different measuring systems. It is more dif®cult to ®nd authors who describe the applicability of these new surface measurements. Below are a few examples of possible applications for paper characterization where surface topography measurement is addressed.

A.

Surface Roughening

Roughening of the surface generally occurs when paper comes into contact with water. It has been addressed by a number of authors, but the designation varies. Aspler and BeÂland called it ``®ber rising'' [37], whereas WaÊgberg called it ``®ber puf®ng'' [38]. This detrimental effect can arise in the coating or the printing process and can be characterized with any of the pro®lometric systems presented above.

Characterization Using Optical Profilometry

B.

447

Gloss and Gloss Variation

The correlation between gloss and surface structure has been established for many years, even with very simple measurements. The advantage of using pro®lometers is that they provide a better understanding of the mechanisms behind the correlation and of how gloss varies in the surface. From the three-dimensional surface, the number and size of facets can be calculated to understand how a glossy surface comes about. This topic has been addressed by, e.g., Lipshitz et al. [32], Hansebai and Morantz [39], and MacGregor et al. [40]. The latter authors pointed out that the facets in the surface can be divided into two sizes, the micro- and the macro-scale regions, and that the facets in the surface have an impact on the tonal reproduction in a printing press. C.

Printability

Surface topography and smoothness are often mentioned when printability and print quality are discussed. Hansebai and Morantz [39] measured the surface smoothness in a number of ways, both conventional and with an optical ``gloss meter'' they developed. They characterized the facets in a surface, divided the roughness into macro and micro components, and claimed they could thereby understand the ink transfer and printability of a particular surface. A correlation of 0:9 between the ink transfer (calculated from the Walker±Fetsko equation) and the microroughness was found. Although this work was not performed by an optical pro®lometer, it seems that the measuring technique presented here could well be used for this type of work. Other work where printability was addressed in correlation with surface topography is that performed by WaÊgberg and Wennerblom [41]. By using the results from an autofocusing surface topography measurement on linerboard printed by ¯exography, the impact of roughness on printability could be established. D.

Microstriations

Several types of defects in the surface have been detected using pro®lometry together with suitable data analysis. As an example, MacGregor and Connors [42] looked at microstriations. Their analysis was based on measurements via image analysis, but surface pro®lometry should also be suitable for characterizing this problem. E.

Miscellaneous

There are several other applications in the literature for surface topography. Two deserve to be mentioned in this context. Falkvall et al. [43] measured different fabrics for the paper machine. In this work they combined pro®le and microscopic measurements in order to be able to simulate and model new fabrics. Knapp and Lampman [9] used an interferometer to characterize the surface of different calender rolls. Corrosion and abrasive wear of calender rolls were evaluated. Interferometric pro®ling proved to be applicable to this type of characterization of surface degradation.

WaÊgberg and Johansson

448

VI.

CONCLUSIONS

In this chapter, an attempt has been made to give an overview of the different techniques that are available for non-contact characterization of the surface of paper. Three main types have been discussed: Interferometric measurements (white light interferometry) Triangulation measurements Autofocusing pro®lometers Another method mentioned was the topographical SEM, although this is not really an optical pro®ler. It was pointed out that the purpose of the measurement must be de®ned. Characterization at different structural levels may give quite a different understanding of the same surface. As of today, no single pro®le instrument can give a good characterization of all the features of a paper surface, from mineral pigments to cockles and streaks. It is dif®cult to give a single de®nition of surface structure, but it is important to realize that a surface consists of several features, both laterally and vertically. Laterally the surface can be divided into three scales: Roughness Waviness Form When it comes to data analysis, several standardized pro®le measurements, adopted from the ®eld of precision engineering, particularly for two-dimensional pro®les, are available. Some of these are also relevant for paper characterization. When three-dimensional surfaces are to be characterized, fewer standardized measurements are available, and the complexity of the interpretation is greater. However, three-dimensional measurements are instructive for the characterization for paper surfaces.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8.

Ballarati, A. (1980). Smoothness of paper and printability units and tests for the determination of smoothness. Annunciatore Poligra®co 193:13±15. Thomas, T. R. (1982). Rough Surfaces. Longman, London. Bennett, J. M., and Mattsson, L. (1989). Introduction to Surface Roughness and Scattering. Optical Society of America, Washington, DC. Bennett, J. M. (1992). Surface Finish and Its Measurement. Optical Society of America, Washington, DC. Bennett, J. (1992). Precision metrology for studying thin ®lms and surfaces. Thin Solid Films 220(1-2):227±233. Stout, K. J., and Blunt, L. (1995). Nanometres to micrometres; three-dimensional surface measurement in bio-engineering. Surf. Coating Technol. 71(2):69±81. Mangin, P. (1993). A structural approach to paper surface characterization. Process and Product Quality Conference (TAPPI), pp. 17±23. Katagiri, K., Kubota, M., and Kayaguchi, M. (1961). Application of microinterferometer to paper. Jpn. Tappi 20(6):394±398.

Characterization Using Optical Profilometry 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32.

449

Knapp, J. K., and Lampman, R. D. (1995). Characterisation of calender roll surface features using interferometric pro®ling. TAPPI Finishing & Converting Conf., pp. 15±24. Yamauchi, T., and Kishimoto, S. (1996). Application of vertical scanning interferometric pro®ler to paper surface. Paperi Puu 78(1-2):29±31. Wyant, J. C., Koliopoulos, C. I., Bhushan, D., and Basila, D. (1989). Development of a three dimensional non contact digital optical pro®ler. J. Tribol. Trans ASME 108(1):1±8. Caber, P. J. (1993). Interferometric pro®ler for rough surfaces. Appl. Opt. 32(19):3438± 3441. Zygo Corp. (1993). What is frequency domain analysis? Zygo R&D Tech. Bull. PdG 9. Zygo Corp., Middle®eld, CT. Wyant, J. C., and Schmidt, J. (1996). Computerized interferometric measurement of surface microstructure. WYKO corp., 2630 E Elvira Road, Tuscon, AZ, internal report. Bennett, J. M., and Dancy, J. H. (1981). Stylus pro®ling instrument for measuring statistical properties of smooth optical surfaces. Appl. Opt. 20(1):1785±1802. WaÊgberg, P., and Johansson, P.-AÊ. (1993). Surface pro®lometry: A comparison between optical and mechanical sensing on printing papers. Tappi J. 76(12):115±121. Enomae, T., and Lepoutre, P. (1995). Stylus pro®lometry on paper: Marking by the stylus. Tappi J 78(10):173±176. Conners, T. E., and Banerjee, S. (1995). Surface Analysis of Paper. CRC Press, Boca Raton, FL. BeÂland, M.-C. (1996). CLSM and AFM applied to pulp and paper research: A literature review. Proceedings of EU Conf. on Pulp and Paper Research, Stockholm, Sweden, Oct. 9±11. Breitmeier, U. (1993). Laserpro®lometrie: Messanlage fur biomedizinische Fragestellungen. Biomed. Technik 38(5):99±104. Gold, M. A., and Schuster, H. K. (1989). Papier-Topographie mit dem Laser-Scanner. Wochenbl. Papierfabrik. 7:271±274. Kohno, T., Ozawa, N., Miyamoto, K., and Musha, T. (1988). High precision optical surface sensor. Appl. Opt. 27(1):103±108. Miyamoto, K., Kobayashi, Y., Nakagawa, H., Kohno, T., Ozawa, N., and Musha, T. (1986). Non-contact high precision optical surface measuring instrument. Bull. Jpn. Soc. Precision Eng. 20(2):121±122. Efsen, J., and Hansen, H. N. (1993). Optisk ruhedsmaÊlning. Diploma work at Danmarks Tekniske HoÈjskole, Publ. No. m m 93.34. Mattsson, L., and WaÊgberg, P. (1993). Assessment of surface ®nish on bulk scattering materials: A comparison between optical laser stylus and mechanical stylus pro®lometers. Precision Eng. 15(3):141±149. Vogt, M., Praast, H., and GoÈttsching, L. (1995). Characterisierung der Rauheit von Papier mittels eines Lasersensors. Das Papier 11:663±671. UBM, Messtechnik GmbH, Ottostrasse 2, D-7505 Ettlingen, Germany. www.werple. net.au/tsi/UBM/htm Lloyd, M. D., Ashburn, M. J., and Chalmers, I. R. (1994). The use of a laser sensor to characterise the other paper grades. 48th Appita Annual Gen. Conf., pp. 103±108. Kapoor, S. G., Wu, S. M., and Pandit, S. M. (1978). A new method for evaluating the printing smoothness of coated papers. Tappi J. 61(6):71±74. Elgay, P., Keller, D. S., and Luner, P. (1995). Development of laser based surface pro®lometer. Part 1: Application to the paper surfaces. ESPRI Internal Report No. 103, pp. 85±90. Lipshitz, H., Bridger, M., and Derman, G. (1990). On the relationship between topography and gloss. Tappi J. 73(10):237±245. Lipshitz, H., Bridger, M., and Derman, G. (1990). Quantitative topographic analysis of coated papers. TAPPI Coating Conference, pp. 219±226.

450

WaÊgberg and Johansson

33. Assoul, M., Zahidi, M., Corcuff, P., and Mignot, J. (1977). Three-dimensional measurement of skin surface topography by triangulation with a new laser pro®lometer. J. Med. Eng. Technol. 18(1):11±21. 34. Enomae, T., Onabe, F., and Usuda, M. (1993). Application of new pro®lometry using topographic SEM to paper surface topography. Tappi J. 76(1):85±90. 35. Enomae, T., and Onabe, F. (1994). Evaluation of stylus pro®lometry using topographic scanning electron microscopy. Proceedings of Image Science and Technology's 47th Annual Conf., Vol. 2, pp. 841±844. 36. Kent, H. (1990). The fractal dimension of paper surface topography. 1990 TAPPI/CPPA Int. Printing and Graphic Arts Conf., pp. 73±78. 37. Aspler, J., and BeÂland, M.-C. (1994). A review of ®bre rising and surface roughening effects in paper. J. Pulp Paper Sci. 20(1):J27±J32. 38. WaÊgberg, P. (1994). A method for simulating ®bre puf®ng on LWC paper. 1994 TAPPI/ CPPA Int. Printing and Graphic Arts Conf. Proc., pp. 267±270. 39. Hansebai, A., and Morantz, D. J. (1988). Assessing various paper surface characteristics can be helpful in predicting ink transfer and printability. Am. Ink Maker (November):28± 34. 40. MacGregor, M. A., Johansson, P.-AÊ., and BeÂland, M.-C. (1994). Measurement of smallscale gloss variation in printed paper. 1994 TAPPI/CPPA Int. Printing and Graphic Arts Conf. Proc., pp. 33±43. 41. WaÊgberg, P., and Wennerblom, A. (1992). A correlation between results achieved with an optical pro®le tester, conventional paper testing and printability. 1992 TAPPI/CPPA Int. Printing and Graphic Arts Conf. Proc., pp. 187±196. 42. MacGregor, M. A., and Connors, T. E. (1989). MD microstriations in paper: A twosided shrinkage phenomenon? Tappi J. 72(4):177±181. 43. Falkvall, K., Martinsson, P., and Werner, A. (1996). Combination of image analysis and topography measurements in the creation of a 3D model. 3rd Eur. Res. Symp. I. A. for P&P Research and Pr., Stockholm, June, pp. 18.1±18.10. 44. Thomas, T. R., and Thomas, A. P. (1988). Fractals and engineering surface roughness. Surface Topogr. 1:143±152. 45. ISO. (1984). ISO 4287/1. Surface Roughness. Terminology. Part 1: Surface and its parameters. 46. ISO. (1984). ISO 4287/2. Surface Roughness. Terminology. Part 2: Measurement of surface roughness parameters. 47. ISO. (1985) ISO 4288. Rules and procedures for the measurement of surface roughness using stylus instruments. 48. Johansson, P.-AÊ., and Norman, B. (1996). Methods for evaluating formation, print uneveness and gloss variations developed at STFI. 1996 TAPPI Process and Product Quality Control Conf. Proc., pp. 139±145.

12 PAPER FRICTION ERNST L. BACK Feedback Consulting, E&E Back KB LidingoÈ, Sweden

I. Introduction II. Fundamentals of Sliding Friction A. Friction Between Equal Surfaces B. Friction Between Materials of Different Hardness C. Molecular Structure and the Effect of Adsorbed Layers

452 452 452 455 456

III. Measurement of Friction A. Instrumental Design B. Repeated Frictional Contact C. Oscillations in Kinetic Friction D. Test Parameters E. Sample Preparation and Test Variance F. Friction Between Single Fibers

457 457 461 462 464 464 467

IV. Interpretation of Paper±Paper Friction Measurements A. Paper Roughness B. Paper Strength and Moisture C. Paper Chemistry and Monolayers D. Fillers and Inorganic Impurities

467 467 468 469 470

V. Interpretation of Friction Between Paper and Other Materials A. Paper±Metal Friction Measurements B. Paper±Rubber Friction Measurements References

470 470 472 472

451

452

I.

Back

INTRODUCTION

Friction is the resistance to motion between two solid bodies in contact. It is the overall shear force needed to break the bonds formed in the geometrical nominal contact area between the two bodies. Only a very minor part of this nominal contact area is separated by distances where atomic bonds can be formed and broken when frictional movement occurs. Therefore the chemistry of the surfacesÐor monolayers on the body and the type of bonds formed thereÐgenerally determine the frictional force. Both physical and chemical aspects have to be considered in the testing procedure for the surfaces in sliding contact. In air, pure cellulose is a hydrophilic (water-friendly) high energy interfacial surface. It easily absorbs low energy, oleophilic (oil-friendly) and liphophilic (fat-friendly) molecules from the surroundings or from its interior to form a monolayer on its surface. Thus sampling for friction measurements requires speci®c care. In this chapter the fundamentals of sliding friction are ®rst reviewed for a better understanding of measurement principles. Three main procedures are then considered for the measurement of the kinetic and static friction coef®cients. Finally, important factors affecting paper±paper friction as well as paper±metal friction are reviewed.

II.

FUNDAMENTALS OF SLIDING FRICTION

A.

Friction Between Equal Surfaces

The following four empirical laws have been found to approximately describe the friction between equal, plane solid bodies: 1.

2.

The force required to slide one body on another is independent of the geometry and size of the apparent contact area. Thus in Fig. 1 the force is independent of how the little box is placed when it is being slid on the horizontal table. It is independent of the pressure in the geometrical nominal contact area. The force required is proportional to the load of the box.

Both of these laws appear in Leonardo da Vinci's notebook, from which Fig. 1 is taken [33]. The notes of Leonardo da Vinci were forgotten for 200 years. In 1699 Amontons rediscovered these laws [1], so they are named Amontons' laws. They mean that there is a constant proportion between the force F for sliding and the total load L. The proportional constant is the friction coef®cient f : F ˆf L

…1†

These laws are not valid at very low loads. 3.

The third law is attributed to Coulomb in 1785 [28]. It stipulates that the sliding friction between equal materials is independent of the differential speed over a considerable range, e.g., 10 10 ±1.0 m/s. This is true as long as friction can be separated from the effect of heat developed in the sliding contact area.

Paper Friction

453

Fig. 1 One of Leonardo da Vinci's illustrations, where he showed that the force required to slide a little box over a table was independent of the nominal geometrical contact area or of apparent contact pressure but proportional to the weight of the box. (From Ref. 33.)

4.

The fourth law was developed during the 20th century. It stipulates that the sliding friction between equal surfaces and especially between equal metals is independent of surface roughness. This was shown by the work of Hardy, Tomlinson, and, especially, Bowden and Tabor [21].

Together these four experimental laws have also developed some coherent understanding of frictional phenomena for elastic materials as summarized in the following. Metals have a well-de®ned plastic deformation pressure, P, above which the initial elastic deformation transfers into an irreversible plastic ¯ow. In only a few percent of the apparent nominal contact area of two metals is the distance between the solid surfaces less than a few angstroms (less than a nanometer). Only in this area, Ar , does metallic contact exist. Only there can surface adhesion develop, and it does so by spot welding, for example, under frictional movement. Here forces reach the shear strength of the metal. Plastic ¯ow occurs locally to a certain depth in this area owing to the high local pressure when the load in the friction zone is increased. Then, because of plastic ¯ow, the contact area increases proportionally. Thus the friction coef®cient, according to the ®rst law of friction, is f ˆ

F Ar S S ˆ ˆ L Ar P P

…2†

where S is the shear strength and P is the plastic deformation pressure. Whatever the corresponding nominal geometrical contact area is, the friction coef®cient is constant and equal to the ratio of shear strength to plastic ¯ow limit. Thus, between clean equal surfaces, the sliding friction coef®cient is dependent on only two mechanical material parameters. It is therefore independent of surface roughness, differential velocity, load, or applied pressure. Both parameters S and P vary with temperature, and so does the friction coef®cient. This is an approximation. With low loads or with low pressure the friction coef®cient increases. The frictional force then reaches a constant level, even in materials of low strength such as plastics [22,75]. According to Briscoe [22], the friction coef®cient of plastics is f ˆ

0 ‡ Pr

…3†

where 0 is the shear strength of the material under very low loads, Pr is the pressure in the real contact area, and  is a material constant. Usually   0 =Pr , so the

454

Back

friction coef®cient is constant and equal to . Yet at low loads Eq. (3) has to be considered. One example is feeding of paper sheets. The shear strength of the plastic increases with pressure according to the expression  ˆ 0 ‡ Pr

…4†

In low strength materials, the frictional forces are adsorbed over a greater depth of the material than in metals. They can also result in surface deformation or cracks, etc. Then Amontons' laws are not valid. The viscoelastic behavior of paper is a complication when applying Eq. (2). Yet a ¯ow limit can be de®ned in some arbitrary way to test the approach of Bowden and Tabor. For more extensible materials such as rubber and for very brittle materials such as silicates the four laws presented do not apply. Nor is the general approach of Bowden and Tabor for metals completely accepted. Even Tabor [75] states that there are experimental results that are dif®cult to explain on the basis of the general approach. For example, when surface irregularities hook into each other, this has an effect on friction. Plastics with a low cohesive strength or shear strength have a low friction coef®cient. Those with a high cohesive strength have a high friction coef®cient. This is indicated in Fig. 2 [34,35]. A friction coef®cient of 1.2 between two nylon 6 surfaces is the highest one shown in Fig. 2. Between pure cellulose surfaces a friction coef®cient at this level should be expected and is found in a ®rst frictional contact (see Section III.C). The four experimental laws mentioned earlier apply to the friction between equal surfaces of materials that are homogeneous to the depth affected by the shear forces. Additional monolayers of foreign material will affect the friction. The four

Fig. 2 The relationship between the kinetic friction coef®cient and the work of adhesion for different polymers. Dots that refer to friction between equal polymer surfaces are indicated by the polymer name. (From Ref. 34.)

Paper Friction

455

laws refer to the kinetic friction coef®cient only. Later approaches considered the occurrence of wear and how worn-off particles interfere with friction [73]. B.

Friction Between Materials of Different Hardness

When the body of a hard material, such as a metal, slides on a softer one, e.g., paper, the harder material with its surface irregularities demonstrates a ploughing action into the softer one. This is illustrated in Fig. 3. [29], where FN is the normal force and FF is the frictional force. The ploughing ``wear'' of the softer material adds frictional forces [29,30]. The overall friction coef®cient is then proposed as f ˆ AgPxy =Ar Pz

…5†

Here Pxy and Pz are the deformation pressures or elastic limits of the softer material in the direction of motion and in the direction perpendicular thereto, respectively. Ag and Ar are the contact areas perpendicular to and in the direction of the ploughing for the harder metal irregularities. This formula includes aspects of hardness and surface roughness. For a common angle of 84±85 against the normal axis of conical tops, this friction coef®cient then becomes [29] f ˆ 0:05Pxy =Pz

…6†

The friction coef®cient also increases with differential speed, because the work to throw the waste material outward increases [29,30]. The distance between various irregularities or tops of the harder material is important, especially at high speed. Work on the friction between metals and plastics has shown rather signi®cant effects of the hardness and surface roughness of steel on its friction against polymers [30,88]. Minor effects occur due to differential speed and pressure. Against lower strength plastics this ploughing friction of a metal is lower [30].

Fig. 3 The ploughing component of friction for a hard conical indenter sliding over a softer material. (From Ref. 29.)

456

Back

Stick-and-slip friction often occurs between two different surfaces in a certain range of pressure or temperature [34]. In this case the friction coef®cient oscillates with time with an amplitude of, e.g., 0.1 or 0.2 friction unit, even in a rather rigid measurement system [34]. The friction between a harder, rough surface material against a softer material is used, e.g., in automotive brakes. We hear the stickand-slip oscillating friction when the brakes screech. C.

Molecular Structure and the Effect of Adsorbed Layers

The cohesive strength and the deformation due to load in and around the relatively small ``atomic'' contact area both depend on the chemical structure of the surface material or surface layer and the bonding therein. This is important for the sliding friction. It is the background for lubrication or lubricated sliding friction. Egyptian mural paintings of the transport of statues of the pharaohs show an ``oiler'' pouring a lubricant in front of the wooden sledge as early as 2400 bc [33]. Low energy surfaces form easily, or automatically, on high energy surfaces by adsorption of oleophilic molecules from the surroundings. Therefore the observation of Leonardo da Vinci that most common friction coef®cients are within 0:3  0:2 is still valid. For example, the friction coef®cient between clean metallic surfaces is initially in the range of 2.0. After the adsorption of a layer of saturated fatty acid the coef®cient falls off depending on the chain length of the fatty acid. With more than 14 carbon atoms in the saturated fatty acid chain, the friction coef®cient at 20 C is as low as 0.04 [65] or in other measurements [46] below 0.10. Unsaturated fatty acidsÐ owing to a bend in the chain at the double bondÐresult in higher friction [40]. The adhesion of the fatty acid to the metal surface is due to the formation of a metal soap monolayer, a stable stationary layer. When the temperature exceeds the melting point of this metal soap, the sliding friction again increases to the original value of pure metals [50]. Fatty acids reduce the plastic±plastic friction in a similar way [79]. To some extent the long-chain alkyl compounds with a carboxylic group can become chemically bonded even to cellulose. Friction with multilayers of lubricating oil is more complicated. On cellulose, aliphatic molecules such as fatty acids often automatically form a low energy, low friction surface and so do alkanes in waxes and the alkyl chains of alkylketene dimer size. Oleophilic molecules originating from the wood extractives can occur naturally in paper. However, surface energy is not the predominant factor, as has been suggested [45]. Friction is a function of the structure of the monolayer molecules. Resin acids, with their aromatic structure, produce the highest friction known for organic molecules. This has been known for at least a thousand years. When the ®rst string instruments were introduced in Europe by the Arabs, pine rosin was introduced as an additive to be applied to the bow to increase friction [41,42]. Rosin powder is well known by baseball players and weight lifters to increase the ``stickiness'' when throwing balls or lifting weights, and a piece of rosin is in the pockets of ballet dancers for similar purposes when they are dancing on wooden ¯oors. Rosin sizing may increase paper±paper friction [24] so that it becomes equal to or slightly above that of pure cellulose materials [6]. Aromatic organic materials are used in hydraulic ¯uids to transfer high forces. In a comparison of the mutual friction between stearic acid surfaces and those of calcium stearate applied on a glass surface, the different effects of speed could not

Paper Friction

457

be explained adequately [23]. The calcium stearate showed stick-and-slip friction, but the stearic acid did not. This shows the complexity of interbonding and its effect on kinetic friction. As another example, the friction between equal plastic surfaces can decrease signi®cantly with increasing molecular weight [25], because the plastic ¯ow limit is thereby increased. Friction is higher for low density polyethylene than for high density polyethylene [34,35], which could be due to the higher crystallinity and lower bonding ef®ciency of the latter.

III.

MEASUREMENT OF FRICTION

A.

Instrumental Design

Rotating Platen In the evaluation of plastic±plastic or metal±plastic friction, it is common to use one circular rotating platen against a stationary platen or against a stationary rider at the platen circumference. The frictional force is measured against time on the stationary part and is usually evaluated over a range of differential speeds. The evaluation refers to a kinetic friction coef®cient. For different materials, wear is often measured simultaneously (see Chapter 13 of this volume). Such measurements use speeds of around 100 mm/s, where small oscillations in friction seldom

Fig. 4 Example of horizontal table friction tester. t, horizontal table; s, sled; e, elevator for the sled; l; load cell; c, means for connecting the load cell to the sled; g, guidance system for the sled, shown schematically; b, backing of foam rubber can be placed on either sled or table or both; ps , test piece for the sled; pt , test piece for the table; F, force between the sled and the load cell; h, distance between the table and the plane of action of the force F; ds , dt , driving mechanism for the sled or table, respectively. (From Ref. 53.)

458

Back

occur. In continuous rotating contacts between paper and metal, the problem is to keep the temperature and moisture in the contact area constant [58]. Horizontal Table The most common setup for measuring plastic ®lm±®lm or paper±paper friction is a horizontal table and a sledge (sled) with the samples attached. One of these parts is stationary, and the frictional force is usually evaluated on this part. Compare this to the setup of Leonardo da Vinci in Fig. 1. In many pieces of equipment the sledge is stationary. The moving part, the table, is set in motion, and a range of differential speeds and a range of loads can be used. Soft or hard supports (backings) can be used under the paper on the table, on the sledge, or on both parts. An example of a recently developed piece of equipment for horizontal plane friction measurements is shown in Fig. 4 [53,54]. It automatically deposits the sledge before frictional contact and automatically lifts and returns the sledge after the measurement procedure. Furthermore, measurements can be carried out in either direction by moving the sledge or the table. The sledge is square and is automatically placed parallel to the direction of movement. Several such pieces of equipment are commercially available [60,80±82]. The evaluation might refer both to the initial frictional force for a static friction coef®cient and to the force under motion for the kinetic friction coef®cient. The kinetic friction is plotted against contact time, i.e., distance of travel. Usually the traction contact is repeated after returning the sledge to its starting point on the table. Such measurements for plastic±plastic ®lm and paper±paper samples are illustrated in Figs. 5±7 [37,51,54,59,61]. The initial ``static'' friction coef®cient evaluated depends strongly on the acceleration rate relative to the speed employed, on the load applied, on the mass of the moving part, and on the stiffness of equipment parts where the force is measured.

Fig. 5 The friction force for polymer foils plotted against the distance of travel for a sledge on a table, with a static friction peak and the subsequent kinetic friction variation. (From Ref. 51.)

Paper Friction

459

Fig. 6 Repeated paper±paper friction measurements between the same contact areas. (From Refs. 37 and 59.)

Above a certain load and within a range of differential speeds, the kinetic friction is constant with rather little dependence on equipment design. There are two important exceptions. The sledge may be free to wiggle relative to its mean direction of motion on the table [66]. Such free sidewise motion appears to produce a lower friction coef®cient than a more controlled direction of motion in more recently designed equipment [54]. The other exception is the falloff of the kinetic friction with continuous distance of travel and with repeated contact in the same direction. In addition, there are several parameters to be settled somewhat arbitrarily for standard method development. With a heated table with a steel foil attached and a sledge with a paper sample, the horizontal table method is also employed to evaluate the friction of ¯uting against a corrugator in TAPPI T 828 [78]. Inclined Plane Various inclined plane types of equipment have been in use in the paper industry for decades [72,77]. Generally, the procedure is useful only for the static friction coef®cient. The inclination of a plane can be slowly tipped from a low starting level after a sledge is placed on the plane. With a variable angle, the normal force perpendicular to the sliding surface varies. The paper sample is attached to the plane and the sledge. The angle at which the sledge starts to move is measured. The static friction coef®cient is equal to the tangent of this angle. A recent example of an

Fig. 7 Paper±paper friction measurements for different acceleration rates (ramp times). (From Ref. 54.)

460

Back

Fig. 8 Example of inclined angle friction tester for both static and kinetic coef®cients. (From Refs. 43 and 72.)

inclined plane tester is shown in Fig. 8. The scale (a in Fig. 8) shows the static friction coef®cient directly. This type of equipment can also permit evaluation by suitable sensors of the time for the sledge to travel the distance to the end stops. In this case the kinetic friction can be evaluated from the force acting to accelerate the sledge [43,72]. Strip-on-Drum Method Recently the strip-on-drum method illustrated in Fig. 9 was proposed as an ASTM standard for the friction of a web material against a drum [26]. The method has been used for testing textiles and rubber webs [49,60]. It was recently tested for paper±paper friction with a ®xed paper strip under tension against another paper strip ®xed to the rotating drum [31]. The tension of the strip and the pressure between the sheets both vary around the contact area [70]. The mean friction coef®cient at a wrap angle of 90 is simply 2T2 =T1 , where T2 is the measured load for the applied load T1 (Fig. 9). For most papers this friction coef®cient decreases signi®cantly with increases in applied load T1 . Compared to the horizontal plane method, measurements with the strip-on-drum method showed signi®cantly higher friction [60]. This may be due to a lower paper±paper pressure. Continuous Fresh Contact Testing The reduced friction coef®cient between paper surfaces with continuous contact has prompted the development of continuous contact testing equipment. Here two strips are continuously drawn into a nip

Fig. 9

The strip-on-drum method used for paper±paper friction testing. (From Ref. 31.)

Paper Friction

461

Fig. 10 The Stora Feldmuhle friction tester. The paper strips A and B are drawn in different directions through a roll nip by lifting the axis upward in the direction of the upper arrow. (From Refs. 48 and 72.)

contact in different directions over rolls, as shown in Fig. 10. However, the exact normal force is dif®cult to calculate. Sample Variance Most papers do not have a homogeneous surface. There may be discrete local areas of ®llers, hydrophobants, or starch particles on the ®ber surface. They all are of colloidal size up to a few micrometers. There may exist local noncoherent monolayers of oleophilic materials, e.g., fatty acids. Also, such monolayers can be transferred to the sample surface during manual handling. Additionally, paper is an anisotropic viscoelastic material with consequences for surface properties such as surface shear strength. These characteristics present a challenge for equipment design and for the measurement procedure as discussed in the following. An example of sample variance between and within reams of paper is given in Ref. 44.

B.

Repeated Frictional Contact

The reduction in both the static and kinetic coef®cients with repeated or increased contact areas is a characteristic feature for paper. It has been dif®cult to interpret [6,17]. Nothing similar has been reported for other materials. The reduction in friction is smaller when the paper is preextracted to free it of oleophilic matter [6]. The total surface energy, or its polar or dispersive components, were not affected by repeated frictional contacts between hardwood handsheets [10]. A reconditioning time of 10 min between repetitions of frictional contact for conditioning did not change the characteristics (see Fig. 11). Thus, loss of moisture in the paper surface or frictional heat does not cause the effect. Ten repeats of frictional contact reduced the IGT surface strength by about 3% for a ®ller containing ®ne paper and by about 6% for a sack paper as evaluated with an IGT oil of 1470 poise. Yet neither paper lint nor ®ber dust could be collected from these surfaces [6]. It was thought that such bond breaking in the surface layers due to

462

Back

Fig. 11 Reduction in friction for a clay-®lled journal paper at repeated contacts. (&, &) Top side against top side; (*, *) wire side against wire side; (~) top side against top side, reconditioned at 50% RH for 10 min; (^) top side against top side after lint removal between contacts. Filled symbols are for extracted samples. (From Ref. 6.)

shear forces might be part of the explanation, because surface strength according to Eq. (2) is an important factor for the friction coef®cient (see also Section V.B). Recently it was shown that when the direction of sliding was reversed, both the static and kinetic coef®cients of friction remained at the original level [54]. An example from this work is shown in Fig. 12. So far a clear explanation is lacking. Because a reduction in kinetic friction coef®cient takes place with increases in the distance of movement, as is seen in Fig. 13, this distance has to be standardized in a measurement procedure. C.

Oscillations in Kinetic Friction

The kinetic coef®cients of friction in paper±paper friction can show minor oscillations, as can be seen in Fig. 5, e.g., with an amplitude of 0.05 unit. Here the con-

Fig. 12 Friction variation for a liner at repeated slidings. A-test: Sliding in the same direction. C-test: Sliding direction reversed for each repetition. (From Ref. 54.)

Paper Friction

463

Fig. 13 Friction variation for a copy paper at repeated slidings in the same direction for different traveling distances. (From Ref. 54.)

nection to the sledge was stiff [37,59]. This is not surprising per se. The very few points of atomic bonding distance where shear forces are acting change continuously during the sliding. New contact areas are created and disrupted while the surfaces are moving relative to each other. Also, ®llers, sizing agents, etc. occur locally on the ®ber surfaces in areas of micrometer size. Several publications have described these oscillations as a ``stick-and-slip'' friction [12,57,68,71,72]. Yet the amplitude of oscillation is much smaller than for stick-and-slip friction as described above. Some oscillations are due to the instrument design, as is evident from their damping after the start of movement [68]. They also occur with plastic ®lms [47]. The effect is reduced with a stiff connection between the sledge and the force sensor. The amplitude decreases with increasing differential speed, as illustrated in Fig. 14. The speed here was varied between 0.2 and 5.0 mm/s, and a spring was attached between wire and sledge. Part of the oscillations may be due to some wiggle of the sledge as well. To prevent wiggle, a central position of the connection between wire and sledge was found to be important [66].

Fig. 14 The effect of sliding speed on the so-called stick-and-slip phenomenon. (From Ref. 54.)

464

Back

There have been several approaches to an analysis and explanation of these minor oscillations [56,59]. There certainly also appear to be some differences for measurements on different papers [56], but they are of little practical importance. D.

Test Parameters

At nominal pressures above 2.5±5 kPa, the static and kinetic friction coef®cients of papers are reported to be fairly constant [37,54,59]. For a 25 kg ®lled paper sack placed on a horizontal plane, the nominal contact pressure is approximately 1 kPa, but it increases when additional sacks are placed on top. A pressure of about 7 kPa was recommended earlier as a standard [37]. The ASTM method for paper, ASTM D 4917 [4], requires a lower pressure of approximately 0.5 kPa. The recently issued ISO 15359 standard [53] requires 2.5 kPa. The Swedish and Finnish pulp and paper research institutes (STFI and KCL) now recommend a differential speed of 20 mm/s, and this is also that of the ISO standard [53]. This may still be low for practical paper sliding. The TAPPI T 816 [77], TAPPI T 549 [76], and ASTM D 4917 [4] methods require 2.5 mm/s (0.5 ft/ min). The German industrial methods (DIN methods) [32,82] require even slower speeds (0.04 mm/s). These are present examples of testing conditions for friction standards. The acceleration and mass of the moving part, e.g., the sledge or table determine the difference between the evaluated static and kinetic friction coef®cients. With a low mass of the moving part and a low accelerationÐa ramp time between 1 and 5 sÐthis difference is minor, as can be seen in Fig. 7. Moisture content has a signi®cant effect on the friction coef®cient and therefore must be controlled in friction testing [24,57] (see Section IV.B). It can be suspected that dried-in stresses have an effect as well, because surface strength does affect intersheet bonding. The means used to deposit the sledge on the table are important. The DIN 6723, 6724, and 6729 [32] methods require an automatic laydown of the sledge with a pick-up device as developed at the Federal Institute for Materials Testing (BAM) in Berlin [39,81]. So also does the ISO standard [53]. Such means are available in recently developed equipment [80±82]. Various types of hard or soft (e.g., foam rubber) backing have been proposed for application to one or both paper samples. It has been recommended that the length of frictional contact be kept short, e.g., 60 mm. In a repeated contact procedure the sledge should be automatically returned to the starting point. Table 1 summarizes test parameters for widely used national standards, including the ISO standard [53].

E.

Sample Preparation and Test Variance

Sampling is most critical for friction measurements. Touching paper samples can transfer oleophilic matter to their surface, thereby reducing friction. Also, the recirculating air in the conditioning room must be free of oleophilic material. The mean values in Figs. 11 and 15, where the friction variation is measured

Writing and printing Punched tapes Data, coated or uncoated Corrugated or solid ®berboard Writing and printing Writing and printing

ASTM D 4917 DIN 6729 DIN 6723 and 6724 TAPPI T 816

ISO 15359

d

c

b

a

Not speci®ed. No stick-slip allowed. Only the third static contact is evaluated. First and third contacts are evaluated.

TAPPI T 549

Paper or board material

Method

Table 1 Friction Standard Methods (Horizontal Plane)

3600

4000

4000 1125 4000 4000 0.5 optional (soft) 2.5 soft

0.5 soft 6.5 soft 1.5 soft 3.5 hard

Contact area Surface pressure of sledge and backing (mm2 ) (kPa)

60

130

130 1a 60 63

Sliding distance (mm)

20

2.5

2.5 0.04 0.04 2.5

Pulling rate (mm/s)

Three static and kineticd

One static and kinetic

One static and kinetic One static One static and kineticb Three staticc

Number of repeated contacts

Paper Friction 465

466

Back

Fig. 15 The static friction coef®cient for various commercial papers before and after repeated extraction with chloroform followed by acetone. (From Ref. 6.)

for various paper products [6], refer to ®ve repeated contacts over a length of 15 cm, at a speed of 5 mm/s and a pressure of 2.3 kPa at 45 to the machine direction of the paper after 20 s of contact before movement. They refer to sample handling with rubber gloves and measurements in a conditioned room at an early stage of equipment design [6,86]. As an alternative, automatic sample handling and placement in the equipment without touching has recently been developed [62,80]. The coef®cient of variation for the mean value of ®ve repeated contacts in Fig. 15 is shown in Fig. 16. After removal of oleophilic material by extraction, the variation was reduced and not signi®cantly larger than for other paper properties. The larger variation for papers containing resin components is to be expected. With an anisotropic paper structure there are numerous directional parameters possible, with their different surface strength characteristics. Even in a single direction the differential speed can be parallel or antiparallel to the machine direction (forward or backward).

Fig. 16 The variation coef®cient for static friction for various commercial papers before and after repeated extraction with chloroform followed by acetone. (From Ref. 6.)

Paper Friction

F.

467

Friction Between Single Fibers

Methods have been developed to measure the friction between yarns [74] and between single ®bers, wet or dry. For recent data and methods for pulp ®bers and how their properties affect this friction, see Refs. 2 and 3. IV.

INTERPRETATION OF PAPER±PAPER FRICTION MEASUREMENTS

Humans have depended on low friction for transport, on high friction for, for example, braking action, and on constant friction for feeding material in sliding transport. In a similar way the requirements for paper±paper friction can differ. A high kinetic friction coef®cient is usually desirable for good roll formation and roll handling. So also is the case when exterior braking action is applied on rolls in a pressroom. A high static friction coef®cient is necessary when rolls are being handled by forklifts to prevent ``telescoping.'' In the shipping of paper, where rolls can stand on each other, high static friction is required for the roll end covers to prevent movement in stormy weather. A high static friction coef®cient is generally also desired for packaging papers such as sackpaper and linerboard. Where the requirement is suf®cient local friction, e.g., in a converting operation, a higher load can compensate for a lower friction coef®cient. A low kinetic friction is desirable between the ¯uting (corrugating medium) and the steel surfaces of the corrugator. The ¯uting has to be pushed into the valleys of the corrugator roll with a minimum of load application to retain its compression strength [63,84]. A constant medium friction coef®cient is required for automatic feeding of copying paper and for papers in converting operations. In many cases, the friction coef®cient required is a compromise. For linerboard the friction should not be too high against the decker of the corrugating machine dryer. A very constant low friction level for punch cards was controlled in practice by balancing the sizing agent from a mixture of rosin soap and anionic wax dispersion. A.

Paper Roughness

Friction coef®cients for a range of commercial papers are shown in Fig. 15 plotted against the PPS (Parker print surface) roughness. Values refer to the static friction coef®cient on samples as received and handled with rubber gloves. The samples were also tested after repeated Soxhlet extraction with pro-analysis grade chloroform followed by acetone. This was to remove natural resin as well as the aluminum soaps formed in sizing [6]. After extraction the friction coef®cient for the various un®lled papers was about 0.65 and for the clay-®lled papers about 0.40±0.45, regardless of surface roughness. A signi®cant reduction in variation coef®cient also appeared due to the extraction of oleophilic material. Speci®cally, one example in Fig. 15 is a TMP-based, clay-®lled SC paper before and after mill supercalendering with eight nips, whereby the PPS roughness was reduced from 7 m to about 1:5 m. The extraction slightly increased the surface roughness of the calendered paper. The supercalendering had reduced the friction

468

Back

coef®cient slightly on the commercial unextracted paper but not on the extracted paper. Other data [8,16,24,36,83] indicate only minor effects of calendering on friction, leading to conclusions that may be based on side effects. Similarly, gloss brushing of calendered and uncalendered TMP-based clay-®lled SC paper with rotary horse tail brush equipment did not change the friction coef®cient signi®cantly [6]. For a full understanding of calendering, additional experiments are needed with papers free of oleophilic components. So far, effects of surface roughness can be neglected as long as surface strength is not changed and the following is considered. In commercial papers the wire side is often poorer in colloidal materialÐe.g., in lipophilic material or ®llersÐthan the top side. Thus surface energy and surface strength differ between these two sides of the paper, with consequences for the friction coef®cient.

B.

Paper Strength and Moisture

The approach of Bowden and Tabor resulting in Eq. (3) has not been tested for wellextracted cellulose or cellophane. For isotropic cellulose of density 1.5 g/cm3 the kinetic friction coef®cient should be in the range of that for nylon in Fig. 2, e.g., around 1.2. Friction coef®cients in this range have been reported for the wire side of handsheets of bleached low resin kraft pulp [10,36]. With increasing moisture content the paper±paper friction increases signi®cantly [24,36,57]. Several examples are shown in Fig. 17 [36]. Equation (3) may explain this effect. With increasing moisture content there is a signi®cant reduction in the deformation pressure at the onset of ¯ow, i.e., at the ¯ow limit [9]. The reduction of strength parameters is much less, so friction should increase according to Eq. (3). It is possible that the addition of other cellulose softeners such as glycerol or ethylene glycol may show a similar effect. In contrast, the friction coef®cient for nylon±nylon contact decreases with increasing moisture content [27].

Fig. 17 The effect of moisture equilibrium on the static friction coef®cient of different papers. (From Ref. 36.)

Paper Friction

469

Heat treatment of a dispersion-sized liner, carried out to illustrate the effect of auto-cross-linking, led to a slight increase in the static friction coef®cient, e.g., from 0.27 to 0.35 [6]. On the corresponding extracted liner, the static friction coef®cient increased from 0.54 to 0.61 and the kinetic coef®cient from 0.49 to 0.54. This effect can be due to a greater increase in surface strength than in ¯ow limit perpendicular to the sheet.

C.

Paper Chemistry and Monolayers

For handsheets made from extracted softwood and hardwood pulps, the static friction coef®cient in the ®rst sliding contact was found to be 1.2 and 1.1, respectively [10]. These values were measured with the equipment shown in Fig. 4 according to the ISO 15359 procedure (Table 1). They apply to the wire sides of handsheets dried against metal gloss platens. Traces of acetone-extractable material were still present in the sheets [10], but their effect was probably negligible on the wire-side sheet surface. These high friction coef®cients are in the range of that for nylon in Fig. 2. The lignin content had no signi®cant effect (kappa number 20±100) [10]. The effect of oleophilic matter in paper is illustrated in Figs. 11 and 15 for an extraction procedure. Another example is the effect of an activated oxidation of the paper surface by a corona treatment. A corona or ¯ame treatment produces a high energy surface on alkenyl succinic anhydride (ASA) and alkyl ketene dimer (AKD) sized paper with their surface layer of alkyl chains also on polymer ®lms. This results in a signi®cant increase of the friction coef®cient. Thereafter, with time, oleophilic material redistributes from the interior to the surface. A low energy surface is again formed, and the friction coef®cient again decreases [5±7]. Thus, because of redistribution of oleophilic material, the friction coef®cient of a paper might change with time. The effect of individual wood resin components has recently been investigated in some detail [40]. Both fatty acids and resin acids can produce a reasonable hydrophobization of paper. As long as fatty acids dominate the surface layer, the friction coef®cient increases with increasing oxidation or increasing oxygen-to-carbon ratio on the surface as evaluated by ESCA [45,52]. However, resin acids produce a high paper±paper friction close to that of cellulose materials whereas saturated fatty acids produce a low friction [6]. Unsaturated fatty acids produce higher friction than saturated ones of equal chain length (see Section II.C) [40]. Little is known about the effect of mixtures of fatty and resin acids as they occur in, for example, mechanical pulps as natural wood components. On storage of paper rolls, fatty acids can redistribute more easily to form monolayers. This would result in reduced friction. When wax is added to a coating color it reduces the friction coef®cient between the coated side and the other coated or uncoated paper side of the next sheet in a pile. When printed, the ink rub value to the next paper in contact falls off in some proportion to the friction coef®cient. Thus, a high friction coef®cient can more easily cause smearing of ink [69].

470

D.

Back

Fillers and Inorganic Impurities

Sack paper surfaces have been treated commercially with a suspension of colloidal silica or carborundum to produce high friction, and these materials have been added as ®llers to the paper furnish [38,67,87]. The effect is explained as the hard particles bonded in the paper surface ploughing into the other (soft) paper surface during the frictional contact, according to Eqs. (4) and (5). The form and maybe size of these hard ®ller particles in the paper surface are most important. For most papers the strength in the xy plane relative to that in the z direction is between 10 : 1 and 30 : 1. If this ratio is used for the linearity limit or deformation ¯ow pressure, the friction coef®cient can be raised according to Eq. (6) from about 0.5 to 1.5 by hard particles bonded in the paper. This is in agreement with experience [38,67,87]. Not only can hard ®llers bonded in the paper surface act in this way, but also hard impurities in the paper, which are apt to cause wear, may increase friction. Such impurities also can increase the variation coef®cient in friction measurements. Fillers bonded in polyethylene increase the friction in a similar way [79]. If a ®ller is not hard enough to plough into the paper surface or not bonded well enough in the surface layer, the ploughing action does not occur. Instead, the reduction caused in paper shear strength and in ¯ow limit by the ®ller will affect friction according to Eq. (2). Data for various pulps have shown that the addition of clay increases friction while beating the paper furnish reduces friction, especially for sul®te pulps [55]. In other work, calcined kaolin added as ®ller increased the friction while hydrous kaolin reduced it slightly [87]. The action of pulp extractives might perhaps explain both these results. Re®ning will free more fat from parenchyma cells and improve its dispersion and redistribution. The general comparison between the extracted SC paper and newsprint paper in Fig. 15 (both made from spruce TMP) indicates a lower friction when ®llers are added and the paper strength is thereby reduced. The clay-®lled SC journal paper with 23% overall ash content was formed on a paper machine with a single Foudrinier wire. By splitting the web, it was found that it had 32% clay on a 34% top-side weight of the sheet and 12% clay on a 28% wire-side weight. After extraction, the mean mutual friction coef®cient was higher on the rougher wire side with less clay than on the clay-rich smoother felt side. The effect of ®llers that absorb oleophilic extractives can be complex, e.g., when talcum is added for pitch control. Talcum per se gives a low friction and most often reduces paper±paper friction [11].

V.

INTERPRETATION OF FRICTION BETWEEN PAPER AND OTHER MATERIALS

A.

Paper±Metal Friction Measurements

Equations (4) and (5) are valid for paper±metal friction. The test methods used are similar to those described for paper±paper friction when instead of one paper sample a corresponding metal foil is introduced. So, to attain a low friction of paper against metal surfaces, a high polish of the metal surface is important. Naturally there is a signi®cant effect if oleophilic material is present on the paper surface. Because of this

Paper Friction

471

Fig. 18 The kinetic friction coef®cient between commercial NSSC-based ¯utings (corrugator medium) and a steel foil plotted against the surface temperature of the foil. (From Ref. 85.)

the contact time before movement starts affects the static friction coef®cient [16,39]. An alternative to a long-time contact is to condition the metal surface by repeated contact with fresh paper samples before the ®nal measurements, as for the measurements in Fig. 18 [85]. Figure 18 illustrates the kinetic friction coef®cient after repeated contact for commercial NSSC (neutral sul®te semichemical) ¯uting against steel foils as a function of foil temperature [85]. Here a low friction is desirable for the runnability of ¯uting in a corrugator [63]. The birch ¯uting holds more wax and fatty components than the oak ¯uting. With rising temperature the oleophilic matter becomes well redistributed onto the paper surface and the friction coef®cient thus falls off. Wastebased ¯uting most often contains some waxed corrugated OCC (old corrugated containers) and thereby usually has the lowest friction coef®cient [84]. Additives added to the ¯uting surface or a wax roll in contact with the web before the corrugating operation have been used commercially [63]. In the procedure illustrated in Fig. 18, the friction contact between the heated metal foil was carried out 15 times with fresh ¯uting samples. Thereafter the friction was measured with a fresh sample. This was to condition the metal foil in respect to ¯uting extractives. Also the friction coef®cient between moist papers and metal foils has been evaluated that way and found to increase with increasing paper moisture over a range of temperatures [84]. A similar effect was found for punch tape paper [16]. Rotational friction measurements between metal and ®ne papers hydrophobized with various types of size required a number of rotations to develop a constant friction, often at a higher level than the initial value [58]. In a printing press the braking action against the paper roll is carried out by copper bands. These copper bands should be rough and if possible free of oleophilic material to attain a high friction coef®cient. In data processing equipment, the friction of punched cards against steel increased with the surface roughness of the steel, e.g., from a friction coef®cient of 0.5 to 1.2 [14] and for punched tapes from 0.15 to 0.25 [16]. It also increased with

472

Back

Fig. 19 The steady-state friction coef®cient between hardwood kraft papers and natural rubber plotted against paper tensile strength. (From Ref. 19.)

differential speed [16]. In the impact printing process, the friction between a carbon ribbon and a print character (letter) is governed mainly by the interaction of the geometry of the character with the structure of the ribbon [13,14]. In the wet grinding process for the production of mechanical pulps, friction coef®cients have been calculated to be in the range of 0.1±0.4 [15]. B.

Paper±Rubber Friction Measurements

The friction characteristics between rubber and rubber and between rubber and other materials do not follow the laws presented in this chapter. They are more complex [64], as also is the case for the friction between paper and rubber. The kinetic friction coef®cient between paper and rubber increases for increasing paper surface smoothness and varies depending upon the load in an exponential way [18]. In repeated frictional contact (track length of 50 cm), the kinetic friction coef®cient falls off with track length to reach a steady level. Here, measured coef®cient values have been shown to range from 1 to 3 for natural rubber and to increase with paper surface strength [19]. They are usually highest against the wire side of a Foudrinier-made paper, as illustrated in Fig. 19. Therefore this friction has been proposed as a measure of paper surface strength [19]. The friction coef®cient between rubber materials and paper is of general importance for paper sheet feeding in printing and copying equipment [20].

REFERENCES 1.

Amontons, M. (1699). De la resistance causeÂe dans les machines. Des Sciences, Academie Royale Paris. December 19.

Paper Friction 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.

473

Anderson, S. R., Nordstrand, T., and Rasmuson, A. (2000). The in¯uence of some ®ber and solution properties on pulp ®ber friction. J. Pulp Paper Sci. 26(2):67±71. Andersson, S. R., and Rasmuson, A. (1997). Dry and wet friction of single pulp and synthetic ®bers. J. Pulp Paper Sci. 23(1):J5±J11. ASTM. (1994). ASTM D 4917. Standard test method for coef®cient of static and kinetic friction of uncoated writing and printing paper by use of the horizontal plane method. Back, E. L. (1988). The oxidative activation of paper surfaces and the friction between papers. Das Papier 42(10A):V31±V39 (in German). Back, E. L. (1991). Paper to paper and paper to metal friction. Proc. Int. Paper Physics Conf., Kona, HI, pp. 49±65. Back, E. L., and Danielsson, S. (1987). Oxidative activation of wood and paper surfaces for bonding and for paint adhesion. Nordic Pulp Paper Res. J. 2(BoÈrje Steenberg Special Issue):53±62. Back, E. L., and Olsson, A.-M. (1983). The effect of temperature on gloss calendering of paper board as evaluated in a press simulator. Svensk Papperstidn. 86(3): R31±R41. Back, E. L., and OÈstman, B. (1983). Hardboard stiffness and tensile strength over a moisture and temperature range. Forest Prod. J. 36(6):62±68. Backstrom, M., Fellers, C., and Htun, M. (1999). The in¯uence of kappa number and surface energy on paper-to-paper friction. Nordic Pulp Paper Res. J. 14(3) 204±208. Baumeister, M., and Kumeth, M. (1991). Talc/carbonate mixtures as alternative pigments for rotogravure papers. Wochenblatt Papierfabrik. 119(17):662±665 (in German). Baumgarten, H. L., and KlingelhoÈffer, H. (1979). Friction on paper surfaces and the transition from adhesion to sliding. Wochenblatt Papierfabrik. 107(23/24):941±946 (in German). Bayer, R. G., and Sirico, J. L. (1968). Comments on the frictional behaviour between a print character and a carbon ribbon. Wear 11:78±83. Bayer, R. G., and Sirico, J. L. (1971). The friction paper characteristics of paper. Wear 17:269±277. BergstroÈm, J., HellstroÈm, H., and Steenberg, B. (1957). Analysis of grinding process variables. Svensk Papperstidn. 60:409±411. Blume, P., and Stecker, A. (1967). Physical properties of punched tapes. Feinwerktechnik 71(6):262±271 (in German). Bolling, R. W., Jr. (1964). Measuring frictional properties of multiwall bag papers. Tappi 47(7):439±444. Borch, J. (1979). Effect of microroughness on rubber friction. J. Polym. Sci. Polym. Phys. 17:2241±2252. Borch, J. (1979). Measurements of surface strength of printing papers by rubber-paper friction degradation. Tappi 62(12):111±112. Borch, J. (1993). Surface characterization of communication papers. In: Products of Papermaking. C. F. Bolam, ed. Pira Int., Leatherhead, U.K., pp. 209±236. Bowden, F. P., and Tabor, D. (1964). The Friction and Lubrication of Solids, Vol. 2. University Press, Oxford. Briscoe, B. J. (1986). Interfacial friction of polymer composites. General fundamental principles. In: Friction and Wear of Polymer Composites. K. Friedrich, ed. Elsevier, Amsterdam, pp. 25±60. Briscoe, B. J., Evans, D. C. B., and Tabor, D. (1977). Saponi®cation on the sliding behaviour of stearic acid monolayers. J. Coll. Interface Sci. 61(1):9±13. Broughton, G., and Gregg, J. L. (1952). Some observations on the kinetic coef®cient of friction of paper. Tappi 35(11):489±493.

474

Back

25. Buckley, D. H. (1981). Surface Effects in Adhesion, Friction, Wear and Lubrication. Tribol. Ser. 5. Elsevier, Amsterdam. 26. Budinski, K. G. (1995). Standard test method for measurement of web/roller friction characteristics. ASTM Draft Method. ASTM, Philadelphia. 27. Cohen, S. C., and Tabor, D. (1966). The friction and lubrication of polymers. Proc. Roy. Soc. (Lond.) A291:186±257. 28. Coulomb, C. A. (1785). TheÂorie des machines simples, en ayant eÂgard de leurs partreÂs, et la roideur des cordages. MeÂm. Math. Phys., Paris X:161±342. 29. Czichos, H. (1978). Tribology. Tribol. Ser. 1. Elsevier, Amsterdam. 30. Czichos, H., and Feinle, P. (1984). On friction and wear of thermoplastics. Tribol. Schmierungstechnik 31(1):14±21 (in German). 31. De Silveira, G., and Hutchings, I. M. (1997). Determination of the friction of paper and board. In: The Fundamental of Paper Making Materials. C. F. Baker, ed. Pira Int., Surrey, U.K., pp. 1329±1353. 32. DIN 6723 (1981), DIN 6724 (1981), and DIN 6729 (1979) Norms for testing 90 g/m2 data paper, 80 g/m2 data paper, and punched tapes, respectively (in German). 33. Dowson, D. (1979). History of Tribology. Longmans, London. 34. Erhard, G. (1980). Friction and wear of polymer material. Ph.D. Thesis, Univ. Karlsruhe (in German). 35. Erhard, G., and Strickle, E. (1977). Machinery Parts from Thermoplastic Materials. Vol. 1. VDI-Verlag, DuÈsseldorf (in German). 36. Fellers, C., Backstrom, M., Htun, M., and Lindholm, G. (1988). Paper-to-paper friction: Paper structure and moisture. Nordic Pulp Paper Res. J. 13(3):225±232. 37. Finnish Pulp and Paper Research Institute (KCL). (1960). Internal Report. Helsinki (in Swedish). 38. Fletchter, C. H., Jr. (1973). Anti-skid treatments utilizing colloidal silica. Tappi 56(8):67± 69. 39. Franke, W., and Huhnt, D. (1976). Investigation on friction of information papers. Final Report ERP 2161. (in German). Bundesanstalt fur MaterialpruÈfung, Berlin. 40. Garoff, N., Jernberg, S., Nilvebrant, N. O., Fellers, C., and Backstrom, M. (1999). The in¯uence of individual wood extractives on paper-to-paper friction. Nordic Pulp Paper Res. J. 14(4):328±331. 41. Gerle, H. (1532). Musica Teusch. Printed in Germany. 42. Gerson, J. Ch. (1473). The Collectorium Supra Magni®catus. Esslingen, Germany. 43. Gockel & Co Gmbh. Equipment, MuÈnchen (1989). Inclined plane friction test. 44. Gunderson, D. E. (2000). Concerning coef®cient of friction. Tappi J. 83(6):39±41. 45. Gurnagul, N., Ouchi, M. D., Dunlop-Jones, N., Sparkes, D. G., and Wearing, J. T. (1992). Factors affecting the coef®cient of friction of paper. J. Appl. Polym. Sci. 46(5):805±814. 46. Hirano, F., Sakai, T., Kuwano, N., and Ohno, N. (1987). Chain matching between hydrocarbon and fatty acid as interfacial phenomena. Tribol. Int. 20(4):186±204. 47. HoÈlz, R. (1970). Measuring friction on ®lms. MaterialpruÈf. 12(4):109±148 (in German). 48. Horand, D., Wertschulte, F., and Altrogge, G. (1992). Method for measuring the kinetic friction of bendable sheet material and equipment herefore. German Patent 4040250 - C1 to Stora-FeldmuÈhle AG., DuÈsseldorf (in German). 49. Howell, H. G. (1953). The general case of friction of a string around a cylinder. J. Textile Ind. 44:T359±T362. 50. Iliuc, I. (1980). Tribology of Thin Layers. Tribol. Ser. 4. Elsevier, Amsterdam. 51. Imass Inc. (1988). Extended capability peel tester for pressure sensitive adhesives. Imass Inc., Box 134, Hingham, MA.

Paper Friction

475

52. Inoue, M., Gurnagul, N., and Aroca, P. (1990). Static friction properties of linerboard. Tappi J. 73(12):81±85. 53. ISO. (1999). ISO 15359. Paper and board: Determination of the coef®cients of static and kinetic frictionÐhorizontal plane method. 54. Johansson, A., Fellers, C., Gunderson, D., and Haugen, U. (1998). Paper friction: In¯uence of measurements conditions. Tappi J. 8(5):175±183. 55. Klemenschitz, W., HaÈner, A., and Dreier, F. (1972). The static friction coef®cients of pulps and paper. Das Papier 26(11):733±737 (in German). 56. KlingelhoÈffer, H., and Baumgarten, H. L. (1983). Paper to paper friction as vibration phenomenon. Allg. Papier Rundschau 18:6±7 (in German). 57. KlingelhoÈffer, H., and Proksch, A. (1961). Physical measurements of friction and adhesive forces in processing glazed papers. Das Papier 15(10):601±603 (in German). 58. Ko, P. L., and Amini, J. (1993). Friction characteristics between paper and polished steel surface. Proc. 6th Int. Congress on Tribology, Budapest, Vol. 5. 59. Kolhonen, E. (1988). Paper friction measurements for paper and board with the Tumila equipment. Paper presented in Munksund, Sweden, on September 1. 60. Levin, E. (1991). Friction experiments with a capstan. Am. J. Phys. 59:80±84. 61. Lindstrand, N. (1996). Reliable friction measurements lead to better paper products. Svensk Papperstidn./Nordisk Cellulosa 99(5/6):129±135 (in Swedish). 62. MarkstroÈm, H. (1994). Device for measuring friction of sheets. AB Lorentzen & Wettre. Swedish Patent C2-500155. (April 25). 63. McKee, R. C., Whitsitt, W. J., and Corbet, H. J. M. (1973). Investigation of solid lubricants as an aid in corrugating board. Paperboard Packag. 58(12):44±47. 64. Moore, D. F. (1972). Friction and Lubrication of Elastomers. Pergamon Press, New York. 65. Moore, D. F. (1975). Principles and Applications of Tribology. Pergamon Press, New York. 66. Payne, S. R., and van Hulle, N. J. (1969). The effect of sled design on the measured coef®cient of friction of wax based coatings. In: Technical of Petroleum Waxes. Tappi Special Tech. Assoc. Publ. 6, pp. 61±83. 67. Pellett, G. H. (1973). Fumed aluminia anti-skid: Properties and performance. Tappi 56(8):70±73. 68. Penner, A. P. (1995). Analysis of stick-slip motion of paper during friction testing. Proc. 3rd Research Forum on Recycling, Vancouver, BC, Canada, pp. 131±134. 69. Rieger, G. (1987). The in¯uence of matte coated papers on the mechanical resistance of offset colors. FOGRA-Forschungsbericht 4.025/2, MuÈnchen, Part 2 (in German). 70. Sato, J., de Silveira, G., and Hutchings, I. M. (1997). Measurement of friction of paper by the strip-on-drum method. Tribol. Int. 30(9) 633±640. 71. Schaffrath, H.-J. (1993). The behaviour of paper compressiveness and wear in respect to paper roll production. Ph.D. Thesis, Univ. Darmstadt (in German). 72. Schaffrath, H.-J., and GoÈttsching, L. (1993). Determination of coef®cient of friction paper against paper. Das Papier 47(9):539±547 (in German). 73. Suh, N. P. (1981). The genesis of friction. Wear 69:91±114. 74. Tabor, D. (1957/58). Friction, lubrication and wear of synthetic ®bers. Wear 1:5±24. 75. Tabor, D. (1981). Friction, lubrication and wear. Surface Colloid Sci. 28:245±312. 76. TAPPI. (1990). TAPPI T 549. Coef®cient of static and kinetic friction of uncoated writing and printing paper by use of the horizontal plane method. 77. TAPPI. (1992). TAPPI T 815 and T 816. Coef®cient of static friction of corrugated and solid ®berboard. Inclined plane method and horizontal plane method, respectively. 78. TAPPI. (1989). TAPPI T 828. Coef®cient of kinetic friction between corrugating medium and a heated steel surface. 79. Thompson, K. I. (1988). Coef®cient of friction testing and factors that affect the frictional behaviour of polytylene ®lms. Tappi J. 71(9):157±161.

476

Back

80. TJT Teknik. (a) (1996). Isofriction system. Instruction Manual. TJT-Teknik AB, Jarfalla, Sweden. (b) Same equipment. Mu measurements, Inc (1998) folder for Amontons II, Madison, WI. 81. TMI. (1997). Labmaster Slip and Friction. Testing Machines Inc., Islandia, NY. 82. Wazau, G. (1986). Measurement and testing instruments. Tribograph RMG 10.02.01 Friction Tester RPF, Berlin (in German). 83. Wennerblom, A. (1988). Low friction: A slippery problem. Nordliner Tech. News 23:1±7. 84. Whitsitt, W. J. (1987). Runnability and corrugating medium properties. Tappi J. 70(10):99±103. 85. Whitsitt, W. J. (1988). Private communication and measurements for the Swedish Pulp and Paper Research Institute (STFI) carried out at the Institute of Paper Chemistry (IPC). 86. Wiberg, R. (1972). Swedish Pulp and Paper Research Institute (STFI) internal recommendations. 87. Withiam, M. C. (1991). The effect of ®llers on paper properties. Tappi J. 74(4):249±256. 88. Ziemianski, K., and Capanidis, D. (1987). The in¯uence of the steel surface on the friction polymer to steel. Tribol. Schmierungstechnik 34(3):173±178 (in German).

13 PAPER ABRASIVITY RAYMOND G. BAYER Tribology Consultant Vestal, New York

I. Introduction II. Paper Wear Mechanisms

477 478

III. Methods to Evaluate Paper Abrasivity

483

IV. Characteristics of Paper Abrasivity

488

V. Design Considerations

494

VI. Similar Wear Situations

499

References

I.

501

INTRODUCTION

Paper abrasivity is the ability of paper to wear other materials. Although the wear caused by paper can generally be classi®ed as a mild form of wear, extremely hard materials, such as diamond, can be worn by paper. Wear by paper is generally of practical signi®cance in applications that involve the handling or processing of large amounts of paper or paper products. It is of most concern in those situations that involve relative sliding and are sensitive to small amounts of wear. Examples of applications in which there are such concerns are listed in Table 1. In many of these applications there is also concern with the wear caused by paper debris that becomes trapped between contacting and rubbing surfaces. The appearance of the wear caused by paper depends on the nature of the surface and the paper. Sometimes these surfaces appear to be polished. This is generally the case with hard materials and less abrasive papers. With more abrasive 477

478

Bayer

Table 1

Applications Exposed to Wear by Paper

Copiers Printers Scanners ATM machines Check sorters Newspaper presses Document readers Paper cutting and punching equipment Paper forms manufacturing equipment

papers and softer materials the surfaces often appear to be scratched. Figure 1 presents examples of wear scars produced by papers. The primary source of paper abrasivity is the hard particles contained in the paper. These particles can be present either as intentionally added ®llers or as impurities. A number of different wear mechanisms are possible, depending on the nature of these particles and the wearing surface. Many of the general characteristics of this type of wear and the mechanisms involved are similar to those occurring with other materials such as magnetic media, fabrics, and printer ribbons. To study this type of wear it has been found necessary to develop special techniques and test apparatuses. One reason for this is that paper is easily damaged and generally has much lower wear resistance than the material it is wearing. The second is that the wear rates are so small for most of the materials of interest. As a result, common wear test con®gurations, such as block-on-ring or pin-on-disk types of tests, would result in the paper being severely damaged or worn, with little or no wear to the wearing material. Figure 2 shows examples of some of the test apparatuses used to investigate wear by paper. Another test con®guration that has been used is that of a needle penetrating one or more layers of paper [27]. Slitting knives have also been used, as well as wear specimens inserted along paper paths in machines [19,23,26].

II.

PAPER WEAR MECHANISMS

Wear is damage to a solid surface that results from relative motion between that surface and another surface or substance and that generally results in the progressive loss of material from the surface [8]. Paper can wear all materials, ranging from the very soft to the very hard [13]. In all these cases, the wear is primarily caused by hard particles in the paper and is generally considered to be a mild form of abrasive wear [15]. Table 2 shows the range of wear coef®cients for several different abrasive situations. Abrasive wear is de®ned as wear due to hard particles or hard protuberances forced against and moving along a solid surface [8]. There are two general sources for these hard particles in papers. The hard particles may be essential components of the paper, such as when titanium oxide is used as a whitening agent or when clays composed of aluminum and silicon oxide particles are used as ®llers. They may also be contaminants associated with different

Paper Abrasivity

479

Fig. 1 Examples of the morphologies of wear scars produced by paper in wear tests. Wear specimens were hardened 52100 (Brinell 760 kgm=mm2 ) steel spheres. (a)±(d) Optical micrographs. (e), (f) SEM micrographs.

components of the paper or manufacturing steps. For example, wood and wood pulp often contain quartz particles and other minerals as contaminants [9,11,15,26,27, 29,34]. Even though wear by paper can be generally classi®ed as a form of abrasive wear, this does not imply that there is only one mechanism for the wear. There are several mechanisms by which hard particles can cause abrasive wear. One is by a cutting action, which results in the formation of chips. This mechanism generally

480

Bayer

Fig. 2 Wear test con®gurations used to study wear by paper. In (a) paper is wrapped around the surface of a drum. The design of one such apparatus is discussed in Ref. 36. The tests in both (b) and (c) use a continuous paper strip. In (b) the wear specimen is stationary; in (c) the wear specimen rotates over the surface of the papers. A version of the apparatus in (b), which can accommodate different contact conditions, has been designed by Falex Corporation, Aurora, IL. An example of the use of con®guration (c) can be found in Ref. 37.

Paper Abrasivity

481

Table 2 Abrasive Wear Coef®cients Wear situation Two-body abrasion with O…100 m) particles Two-body abrasion with O…50 m) particles Three-body abrasion with O…100 m) particles Three-body abrasion with O…50 m) particles Wear by paper Range Typical

Ka 10 1 10 2 5  10 10 3

3

1  10 9 ±5  10 10 6 ±10 5

5

a K ˆ (wear volume  hardness)/(sliding distance  load). Source: Refs. 11, 15, 33, and 34.

requires angular abrasives that are harder than the wearing surface. Even in these cases there is a critical angle of attack below which cutting will not occur. The critical angle is a function of material properties and the coef®cient of friction between the particle and the wearing surface [25,30,38,40,41]. If cutting does not take place, plastic deformation occurs, forming a grove in the surface, but no material is lost. With repeated deformation cycles, particles will eventually be broken off. This may occur in several ways. One is by fatigue processes, in which cracks are formed and propagated in the repeatedly deformed material. Another is by brittle fracture. A third possibility is a ratcheting process involving progressive elongation [21]. Even if plastic deformation does not occur, wear is still possible because of the high local stresses caused by these particles. Fatigue and other anelastic mechanisms associated with high stresses and strains can eventually result in the detachment of material from the surface. These are the likely mechanisms for wear when the wear surface is harder than the particles. References 2, 18, 20, 32, and 33 provide additional information about these wear mechanisms and others that are associated with particles as well as the general topics of abrasion and particle erosion. The nature of the abrasive wear that takes place is a function of many factors. One such factor is the hardness of the abrasive relative to that of the abraded surface. In abrasive wear situations there is a dramatic change in wear rate when the abraded surface becomes harder than the abrasive. Generally wear rates decrease by one to two orders of magnitude. Wear by paper exhibits the same effect when the hardness of the wear surface becomes greater than that of most minerals, i.e., greater than approximately 700 kg=mm2 [13]. Abrasive wear is also in¯uenced by the degree of freedom of movement that a particle has. Wear rates are highest when the particles are attached to a surface, such as occurs with sandpaper or other abrasive papers. This situation is generally referred to as two-body abrasion. When the particles are not attached and are free to roll between the surfaces or move with the abraded surface, wear rates tend to be lower, as shown in Table 2. This situation is referred to as three-body abrasion. With paper there is evidence of both modes being present [9]. There are two likely reasons for this. One is that many particles do not appear to be strongly attached to the surface [26]. The other is the tendency for loose material, such as paper debris, to remain in the interface for a period of time [11,15].

482

Bayer

Wear rates tend to increase with the sharpness or angularity of the particles. Rounder particles tend to be less abrasive than sharp particles. In addition, abrasive wear is also in¯uenced by the size of the abrasive. Wear rates increase with increasing particle size up to a certain size, typically in the range of 10±100 m, above which they become independent of size. Other properties of the particles that can be a factor in abrasive wear are their friability and tendency to agglomerate. The number of particles present is also a factor. Wear rates tend to increase with increasing amounts of abrasive. However, there is a saturation level at which the wear rate becomes independent of the amount of abrasive present. Surfaces worn by paper can have dramatically different appearances, ranging from highly polished to extremely grooved and scratched, as can be seen in Fig. 1. The highly polished surfaces also tend to exhibit grooves or striations in the direction of motion. However, in these situations these features are generally much ®ner and observable only under higher magni®cations. A polished appearance tends to occur with papers of low abrasivity, whereas the coarser appearance tends to be found with papers of high abrasivity. The hardness of the wear surface also seems to be a factor in the appearance. A polished appearance tends to be more likely with harder materials. This range in appearance is consistent with the appearance of wear scars produced by abrasion. Striations and grooves in the direction of motion are characteristics of surfaces worn by abrasive wear mechanisms. Very ®ne abrasives tend to produce polished surfaces, whereas coarse abrasives tend to produce highly grooved and scratched surfaces [11]. Abrasive wear between two surfaces is generally described by the equation V ˆ K 0 PS

…1†

where V is the wear volume, P is the load, and S is the amount of sliding between the two surfaces. K 0 is a constant and is sometimes called the speci®c wear rate. When the abrasives are harder than the abraded surface, K 0 is inversely proportional to the hardness of the wear surface for homogeneous materials. In this case this equation is generally written as V ˆ KPS=p

…2†

where p is the penetration hardness of the surface. The constant K is referred to as the abrasive wear coef®cient. K is dimensionless and a characteristic of the wear situation. K 0 , on the other hand, has the dimensions of wear rate per unit load and is a characteristic of the wear behavior of a material in a speci®c wear situation. K is generally used as a measure of the abrasivity of a material or situation. When the abraded material is harder than the abrasive, a different relationship with hardness is likely because of differences between the wear mechanisms in these two regions. It has been proposed that when the abraded material is harder, K is inversely proportional to pn , where n is greater than 1 [18]. When there is minimal paper damage, several wear studies have indicated that wear by paper can be described by Eq. (1) [10,11,19,23,28,35]. However, in situations where there is noticeable paper damage, the wear is still proportional to sliding distance but not necessarily proportional to load [11]. For hardness below approximately 800 kg=mm2 (Rc65), wear by paper tends to follow Eq. (2) [13]. Above that hardness, wear or K 0 is more sensitive to hardness.

Paper Abrasivity

483

For example, the data shown in Fig. 3 indicate that between 800 and 1000 kg=mm2 , K 0 is inversely proportional to p10 . Above 1000 kg=mm2 , K 0 is inversely proportional to p5 . This transition in wear behavior occurs in the hardness range of the harder minerals typically found in papers. Environmental and oxidation/corrosion effects can in¯uence the basic wear processes by modifying the properties of either the paper or the wearing surface. Examples of such behavior are discussed in Section IV.

III.

METHODS TO EVALUATE PAPER ABRASIVITY

Abrasivity is de®ned as the ability of a material or substance to cause abrasive wear [8]. The typical measure of abrasivity is the dimensionless wear coef®cient K of Eq. (2). Like the abrasivity of any other material, paper abrasivity is not an intrinsic material property. It can and often does vary with the wear situation. As a consequence, abrasivity, like most other wear properties, is best determined by wear tests that simulate the applications in which the material is used [1,3,5±7,17,18]. With paper there are two general features that are typically needed to provide this simulation. One requirement is that the ratio of the surface area of paper to that of the wear specimen must be very large. The second is that little or no damage should occur to the paper. Both conditions are characteristic of most applications in which there is concern with wear by paper. In these applications large amounts of paper are fed through the equipment with only brief contact with the wearing surface. Also, in these applications, with the exception of slitting and punching, little or no damage is allowed to occur to the paper. These same general characteristics are

Fig. 3

The effect of hardness on the resistance to wear by paper. (From Ref. 13.)

484

Bayer

also needed in wear tests used to evaluate the wear resistance of materials considered for use in these applications. Figure 2 shows test con®gurations that have been used to simulate these conditions. Although all three con®gurations can be used to evaluate paper abrasivity, the drum con®guration is the one that has been most often used for this purpose [11,12,14,16,24,31,34,36,37] (see also Ref. 18, pp. 498±576). With the drum con®guration, paper is wrapped around the surface of the drum and a spherical wear specimen is pressed against the paper surface [36]. A spherical wear specimen is used to eliminate alignment problems. As the drum rotates, the wear specimen moves axially across the drum. The wear path on the paper is helical. Because of the ®nite width of the contact area, which increases with wear, overlap can occur between adjacent wear paths, as shown in Fig. 4. The amount of overlap can be controlled by the ratio of drum speed to specimen speed. If the specimen is moved very rapidly, there will be no overlap. As the specimen speed is decreased, overlap will tend to increase. Studies with this apparatus have shown that the abrasivity of a paper is very sensitive to test conditions. Figure 5 shows the effect of load, drum speed, and specimen speed on abrasivity obtained with one apparatus [36]. These effects are the result of damage to the paper. Abrasivity tends to decrease with higher loads, higher drum speed, and speed ratios that result in more overlap. These are the same conditions that tend to cause more damage to the paper. Two effects have been identi®ed. One is a reduction in abrasivity because of loss of or changes to the abrasives on the surface of the paper. Scanning electron microscopic (SEM) examinations of paper surfaces before and after wear testing show that such changes do take place. Figure 6 shows some examples of this. Several mechanisms are considered possible for these changes. One is the detachment and removal of

Fig. 4 Example of overlapping of wear paths in paper wear tests with the drum apparatus shown in Fig. 2a. A similar situation occurs with the continuous feed, rotating specimen apparatus shown in Fig. 2c. Such conditions tend to reduce wear because of damage to the paper.

Paper Abrasivity

485

Fig. 5 The effect of test parameters on the apparent abrasivity of a paper for the drum test con®guration in Fig. 2a. (From Refs. 11 and 15.)

particles. Another is simple detachment, which allows the particles to move and create a three-body abrasive situation. Particles can be pressed further into the paper and no longer be able to contact the counterface. Large abrasive agglomerates can also be broken up into smaller, less abrasive particles. Because of the damage that occurs to paper, abrasivity tends to decrease with use, as shown in Fig. 7. The second reason for this decrease in abrasivity is the effect of paper debris. Paper debris tends to remain in the interface and modify the contact situation. Figure 8 shows an example of paper debris sticking to a wear specimen and the distorted wear scar that results. This effect tends to increase with higher humidity. To minimize such effects there should be no overlapping of wear tracks in tests used to determine abrasivity. Likewise, low surface speed and a light load should be used to minimize damage and the creation of wear debris. Values for these two test parameters tend to vary with the specimen shape used in the tests. With a drum apparatus and a 6.35 mm diameter spherical wear specimen, it was found that the effect of damage and wear debris is minimal for loads under 100 g and surface speeds less than 50 cm/s [11,15].

Fig. 6 Scanning electron micrographs of paper surfaces before and after wear testing. (a) shows a slightly damaged paper surface, (c) a moderately damaged paper surface, (e) a severely damaged paper surface. (b), (d), and (f) show the undamaged surfaces of the papers in (a), (c), and (e), respectively. (From Ref. 15.)

486 Bayer

Paper Abrasivity

487

Fig. 7 Abrasivity K of a paper decreases as wear tests are repeated with the drum wear test apparatus of Fig. 2a. ``Pass'' refers to the number of times the sample was tested. (From Ref. 15.)

Fig. 8 (a) An example of paper wear debris adhering to a wear scar obtained in tests with the drum apparatus of Fig. 2a. This could be observed by quickly separating the wear specimen from the surface of the paper while the drum was still rotating. (b) The distorted wear scar, which should be a circular ¯at spot, after the paper debris was washed off with a soap and water solution. (From Ref. 15.)

488

Bayer

Because the in¯uence of hardness tends to change above 800 kg=mm2 , Eq. (2) is no longer valid for hardnesses above that value. Therefore if this equation is to be used to determine abrasivity, a homogeneous material with hardness less than that value should be used as the wear specimen. A convenient material for this purpose is hardened 52100 steel (Rc62, 750 kg=mm2 ) spheres. This is a common ball bearing material and wear specimen. It is available in a variety of ball sizes with very well controlled size, ®nish, composition, and hardness. Other materials may also be used. However, if a harder material is used and Eq. (2) is used to determine abrasivity, the abrasivity value will be less than what would be obtained if a softer material were used. Under conditions for minimal paper damage, the amount of sliding that is typically necessary to produce measurable amounts of wear on hard specimens such as 52100 steel is in the range of 2  105 ±5  105 cm. Speci®c wear rate, K 0 in Eq. (1), and abrasion resistance, 1=K 0 , are also dependent on the wear situation. If the abrasivity of a paper is known, the general trends with hardness can be used to provide an estimate of these. However, there is suf®cient scatter about these trends to make it generally desirable for engineering purposes to determine these properties experimentally for the particular materials of interest. This is particularly true in the case of multiphase materials and thin coatings, which often are of interest for these applications. The same types of tests that are used to determine abrasivity can be used to determine the resistance of different materials to wear by paper. Using one paper as a reference, a useful way of comparing materials is in terms of the coef®cient K 0 of Eq. (1). Higher values mean lower wear resistance. One problem that is often associated with this type of evaluation is the variation in abrasivity that can occur between different samples of the same paper or with different environmental conditions, e.g., winter versus summer conditions. This is particularly true in situations where evaluations are performed over long periods of time. To minimize this effect, it is desirable to determine the abrasivity of the paper sample being used in the test to determine the wear resistance. This value can then be used as a scaling factor to eliminate the effect of paper and environmental variability in the comparison of wear resistance. IV.

CHARACTERISTICS OF PAPER ABRASIVITY

The abrasivities of a large number of papers used in business machines and computer applications have been measured. Values of abrasivity for these types of papers are typically in the range of 1  10 6 ±1  10 5 , but values as low as 3  10 9 and as high as 6  10 5 have been observed. A computer punch card stock, which is generally considered to be a low abrasivity paper, has a value of 1  10 6 [19,35]. Several attempts have been made to correlate these values with various attributes of the paper such as type of stock, weight, and ®llers [11,12,15,34]. Although some trends have been found, they are typically not predictive, i.e., good indicators of paper abrasivity. There tends to be signi®cant overlap in the abrasivities of papers of different types. One trend that has been observed is that clay-®lled papers tend to be more abrasive than carbonate-®lled papers, but are not always [34]. As can be observed in Fig. 9, both the highest and lowest abrasivities are for clay-®lled papers.

Paper Abrasivity

489

Fig. 9 The abrasivity of some printer and copier papers with clay and carbonate ®llers. Abrasivity was determined using a drum test apparatus similar to the one described in Ref. 36. Test parameters were 36 cm/s sliding speed, 0.25 mm/rev specimen speed, 100 g load, and a 52100 steel sphere (6.35 mm diameter, 760 kg/mm2 hardness) wear specimen. Environmental conditions were 20 C and 40% RH. (From Ref. 34.)

The majority of carbonate-®lled papers exhibit less abrasivity than most of the clay®lled papers. This lack of clear distinction in abrasivity of clay- and carbonate-®lled papers may be attributed to variations in the impurities contained in each of these two types of ®llers. In another study a trend was observed regarding the source of the papers. As a class, papers from Europe and the Far East tended to be more abrasive than papers from North America. However, there was considerable overlap in the abrasivity. Extremely low and high abrasivity papers were found in both groups [12]. Attempts also have been made to correlate abrasivity with the amount and type of hard particles contained in the papers [11,26]. However, there appears to be no correlation. In these studies the ash contents of papers were determined, and SEM analysis, energy-dispersive X-ray (EDX) analysis, and X-ray diffraction were used to characterize the particles. The results of ash content and EDX studies on a number of papers are shown in Figs. 10 and 11. As can be seen in Fig. 10, ash content is not a good indicator of paper abrasivity. In Fig. 11 it can be seen that the distributions of elements found in papers of high and low abrasivity were similar. These studies also failed to reveal any distinguishing characteristics regarding the surface distribution of particles on papers of high and low abrasivity. Papers of low and high abrasivity appeared to have similar distributions. The surfaces appear to be covered mainly by micrometer and submicrometer sized particles. The surface density of these particles was of the order of 1010 particles/cm2 . The density of particles greater than 10 m was much lower, on the order of 10 3 particles/cm2 . In most cases these larger particles appeared to be agglomerates of smaller particles, as shown in Fig. 12. As discussed in Section III, damage or wear to a paper tends to reduce its abrasivity. As a result the abrasivity of used paper tends to be lower than that of unused papers. For example, in the evaluation of check sorters for wear life, it has

490

Fig. 10

Bayer

A plot of wear test results with papers of different ash contents. (From Ref. 11.)

been found necessary to limit the use of check samples to maintain high wear rates [23]. Figure 7 shows this effect as well. Also the apparent abrasivity of a paper is usually lower in those situations in which paper debris accumulates at the interface. Because damage to the paper is normally small in most applications, this effect is usually not observed. One example of its occurrence is the case of electroerosion printing, where paper debris signi®cantly reduces the abrasive action of the paper surface [14]. Although it tends to reduce the abrasive action of a paper surface, paper debris is abrasive [10]. As a result, it tends to increase wear at other wear interfaces, where there is not direct contact with the paper. For example, paper debris can cause abrasive wear at internal interfaces of bearings that are exposed to it. There can be lot-to-lot variation in paper abrasivity. Typically, this is found to be by a factor of 2 or less [12]. However, larger variations can occur. The felt and screen sides of papers may also have different abrasivities. Even in cases where there is no intentional difference in surface treatments applied to the felt and screen sides, there can be differences because of the manufacturing process used. For example, papers made using Fourdrinier machines tend to have a higher concentration of ®ller on the felt side. Papers made by twin-wire formers tend to have the same ®ller concentration on both sides. Because of this there is also the possibility of a geographical trend in differences between felt and screen sides, because the tendency in the United States has been to use the former, whereas the tendency elsewhere has been to use the latter for the papers tested. In addition there is evidence that paper abrasivity can increase with paper density [26]. This is probably related to the tendency for particle bonding to increase with increasing paper strength and density. Abrasivity is also in¯uenced by relative humidity and the moisture content of the paper. Figure 13 shows the effect of humidity on abrasivity when paper debris is a factor and when it is not. The trends are opposite, and the effect is much more

Paper Abrasivity

491

Fig. 11 The results of elemental scans of the ashes of papers of different abrasivities, using EDX. (a) 11% ash, 7:4  10 7 cm3 wear; (b) 12% ash, 7:4  10 7 cm3 wear; (c) 5.2% ash, 9:8  10 7 cm3 wear; (d) 12% ash, 4:1  10 6 cm3 wear; (e) 20% ash, 2:1  10 5 cm3 wear; (f) 4.1% ash, 4:8  10 5 cm3 wear. (From Ref. 11.)

Fig. 12 Scanning electron micrograph of an example of the ®ne particle agglomeration found on paper surfaces. (From Ref. 15.)

492

Bayer

Fig. 13 The effect of relative humidity on the abrasivity of paper (*) under test conditions that resulted in little paper damage and wear debris generation and () under test conditions that resulted in signi®cant paper damage and wear debris generation. The effect tends to be greater in some papers than in others. (From Ref. 15.)

pronounced when paper debris is a factor. Figure 14 shows the effect of moisture content on abrasivity. Abrasivity tends to decrease with increasing moisture content. This is to be expected, because paper strength tends to decrease with increasing moisture content [26]. Inks can also in¯uence the abrasivity of paper surfaces. If the ink contains abrasive particles itself, the abrasivity will tend to increase. An extreme example of this is observed with magnetic ink, because the abrasive action of the magnetic ink

Paper Abrasivity

493

Fig. 14 An example of the effect of the moisture content of a paper on its abrasivity. (From Ref. 15.)

may completely mask the abrasivity of the paper [37]. Nonabrasive inks, containing only dyes, tend to decrease the apparent abrasivity by as much as an order of magnitude [11]. When inks are present there is also the possibility of oxidative or corrosive effects on the wear surface, which can in¯uence abrasive wear behavior [18, pp. 498±516]. This range of behavior for abrasivity and the general trends with paper characteristics are explicable in terms of the complex nature of abrasive wear. The following indicates some of the factors that may be involved and that may provide an explanation for the observed trends or lack of trends. One aspect is that only

494

Bayer

surface particles will cause wear [27]. Hence the total amount of abrasives in a paper may not be a good indicator of abrasivity; the surface density of particles is probably more signi®cant. However, this is much more dif®cult to determine than ash content. Another factor is that fewer large particles can cause the same amount of wear as a larger number of smaller particles. Therefore, a paper with a few large particles may be no more abrasive than a paper with a large number of smaller particles. Also, particles that are ®rmly ®xed to the surface can cause more wear than particles that are less ®rmly attached. As a result, papers with the same amount and types of abrasives may differ in their abrasivities because of the ways these particles are held on the surfaces. Since abrasive wear tends to be independent of particle hardness when the particles are harder than the wearing surface, minerals of different hardness may cause the same amount of wear. Studies have not been done to verify and evaluate the signi®cance of these and similar effects on the abrasivity of various papers.

V.

DESIGN CONSIDERATIONS

Because wear by paper is fundamentally an abrasive wear process, design approaches to minimize wear and increase life are the same as those used to control abrasive wear. For the mild abrasive wear processes involved with wear by paper, the primary material parameter is hardness. In general, the harder the material, the better it is in these situations As previously indicated, wear rates decrease with increasing hardness, with the rate of decrease being signi®cantly higher when the surface is harder than the abrasive than when it is softer. This trend is implied in Fig. 3. In addition, at this transition wear rates decrease by one to two orders of magnitude. With papers this occurs at approximately a hardness of 800 kgm=mm2 (Rc65). In mild wear situations, such as those involving light loads and limited sliding, it is sometimes possible to obtain adequate wear behavior with materials whose hardness is below Rc60. However, for most applications requiring long life and low wear rates it is generally necessary to use materials with a hardness of Rc60± Rc62 or greater. With composite materials, such as glass-®lled polymers or metals with carbide particles, the hardness of the hard particles is not necessarily a good indicator of the wear resistance in paper wear applications. With these types of materials wear resistance is in¯uenced by the relative amount of the particles and the matrix and the size of the hard particles. In abrasive wear situations the tendency is for the matrix to wear preferentially [37,39]. This results in more load on the particles, which tends to increase their wear rate. In some cases the preferential wear of the matrix can be so great that the particles are more easily removed from the surface than worn down. As a result the wear rate can be close to that of the matrix. Such effects are very sensitive to the nature of the abrasive wear situation. Therefore, when such materials are being considered for use in paper wear situations, it is recommended that their wear resistance be determined in abrasivity tests against paper.

Paper Abrasivity

495

The wear resistance of coatings can be in¯uenced by the properties of the substrate. As a result it is recommended that for use in paper wear applications the wear resistance of thin coatings be determined in abrasivity tests against paper. Because wear depth or the depth rate of wear is a concern in most engineering situations, an effective way of reducing wear or increasing wear life is by reducing the pressure in paper wear situations. For example, in the situation shown in Fig. 15, wear life could be extended by increasing the radius of the surface in contact with the paper or by reducing the tension in the paper. Both changes will result in lower pressure. Another way of reducing wear by paper is to reduce the amount of sliding associated with the contact. For example, the stationary surface in Fig. 15 could be replaced by a roller bearing to reduce the amount of sliding. In situations involving rolling contact, wear life can often be extended by reducing the amount of slip that occurs between the roller and the paper surface. Such slip can result from misalignment, deformation of the roller surface, or inadequate traction. An example of the latter would be the use of a rubber roller to feed the paper. Because of de¯ections in the rubber, the motion will not be pure rolling. As the rubber becomes softer, the amount of slip will increase. If the traction is not suf®cient, skidding will occur when the roller accelerates or decelerates. Equations (1) and (2) can be used to model wear in speci®c situations. In this manner the effects of different design parameters and other factors on wear life can be evaluated. There are several examples of the successful use of this approach in the literature. One involves the abrasive wear of type faces in an impact printer [24]. In this case a dynamic model describing the impact between the paper surface and the moving type was combined with Eq. (1) to provide the following relationships for type wear: V ˆ 0:5K 0 vt P0 fN   1 V ˆ 2K 0 vt P0 1 N f f  …mv=P0 †t

for f  2

…3†

for f  2

…4† …5†

Fig. 15 An example of a common situation of wear by paperÐthe use of a curved surface to control the motion of a paper web. In this case, a ®xed, circular capstan is used to illustrate this condition. The load between the paper and the ®xed surface is a function of the tension in the web and the geometry of the ®xed surface.

496

Bayer

where V is the volume of wear on the type element, N is the number of print cycles, v is the velocity of the type, t is the impact duration, P0 is the peak impact form, m is the effective mass of the type, and  is the coef®cient of friction between the type and the paper. f is called the slip factor and is a measure of the time period during which the type slides relative to the paper. This model provided good agreement with measured wear performance. It was found that the model correctly predicted the effect of type speed on wear. Using values of K 0 determined for the speci®c combinations of paper and type used in the printer test, measured and predicted values agreed to within a factor of 2. The theoretical values tended to be higher than the empirical values in this case. Equation (1) was also used to describe the wear of elastomer feed rolls used in a check sorter [31]. In this case Eq. (1) was directly used to assess the effects of various design changes that were being considered. Values for K 0 were again determined for the materials involved. Equation (2) was used for an investigation into the wear of slitter blades [10]. In this case an experimental apparatus was built to simulate slitter blade action. This is shown in Fig. 16. To avoid the problems of providing large amounts of paper, paper dust was supplied to the junction of the blades. Based on the appearance of worn blades, this was found to be an effective way of simulating actual behavior. Equation (2) was used to develop a wear model that correlated well with the observed wear behavior of the blades. The model resulted in the equation a2 2KLn ˆ p tan 

…6†

where a is depth of wear,  is the wear angle on the blade (see Fig. 17b), L is the load between the blades, and n is the number of revolutions. Equation (1) was also used to describe the wear behavior in electroerosion printing [14]. In this situation a head consisting of a number of electrodes is in continual sliding contact with the surface of a conductively coated paper. Printing is caused by providing a voltage to the electrode and burning away the conductive layer, exposing a dark underlayer. In this application a considerable amount of debris tended to adhere to the surfaces of the electrodes. As a result it was found necessary to modify Eq. (1) for the effect of debris. Tests indicated that with debris accumulation the effective abrasivity of the paper asymptotically decreased to approximately one-third or one-fourth of its initial value. This is shown in Fig. 18. K 0 was modi®ed by this factor to describe the abrasive wear behavior in this situation. Electrical discharge wear was also found to be affected by debris in a similar manner, and the relationship for it was modi®ed in a similar manner. It was found that the discharge created debris, which added to the accumulation on the electrodes. This contribution was related to the number of ®rings that occurred in a unit of travel. Combining these two wear modes, i.e., abrasive wear and electrical discharge wear, and including the effect of debris on wear behavior, a model was developed. This model provided an explanation for observed printer behavior. This is shown in Fig. 19. This model was used to describe the effects of load between the electrode, type of printer usage, and different levels of abrasivity on wear life.

Paper Abrasivity

497

(a)

(b)

Fig. 16 Apparatus used to study the effect of paper abrasivity on the wear of slitter blades. (a) The overall apparatus. (b) Details of the mechanism used to introduce paper debris into the contact region between the blades. (From Ref. 10.)

As previously indicated, paper properties such as type and ®ller content do not provide a good indication of the abrasivity of the paper. Therefore when Eqs. (1) and (2) are used to determine wear or wear life, it is generally necessary to experimentally determine values of K for the speci®c papers involved in the wear situation. Because the abrasivity can be in¯uenced by relative humidity, moisture content of the paper, ,and the presence of inks or other coatings, these determinations should be done under the same conditions as in the application. For example, in the case of type wear the abrasivity was determined using inked papers, because in this application ink was applied to the surface of the type before it impacted the surface of the paper. The method of applying the ink in the abrasivity tests is shown in Fig. 20. With some papers this was found to signi®cantly change the abrasivity, as shown in Table 3. The

498

Bayer

(a)

(b)

Fig. 17 Details of the contact situation between slitter blades. (a) General features of the contact. (b) The two parameters used to describe the wear of the blades. a is the recession of the blade's edge, and  is the angle at which it is worn. Both aspects can in¯uence cutting .(From Ref. 10.) Table 3

Effect of Ink on Abrasive Wear Coef®cient Paper

A B C

Without ink With inka Without ink With ink Without ink With ink

a Aqueous dye based ink. Source: Partially from Ref. 11.

K 1:5  10 8:2  10 1:6  10 9:8  10 6:6  10 1:4  10

5 6 5 7 7 6

Paper Abrasivity

499

Fig. 18 The effect of cleaning on the wear rate of electroerosion print heads. A cycle is 6:45  104 cm of sliding. The drum tests were conducted using the apparatus described in Ref. 36 with the normal spherical specimen being replaced by an electroerosion print head. The printer tests were conducted without printing; the head simply slid over the paper surface. Loads and speeds were approximately the same in both directions. (From Ref. 14.)

abrasivities of electroerosion papers were also determined under the same environmental and paper conditions associated with printer tests [14,16]. One other way of reducing wear by paper is to reduce the abrasivity of the papers. To do this it is necessary to reduce the amount and size of the hard particles exposed on the surface or to change the manner in which they are attached to the surface. Studies of paper abrasivity indicate that simply using less abrasive materials is not suf®cient (Section IV). Other factors appear to be involved. For example, these studies suggest that, by affecting the distribution of particles on the surface, ®nishing processes may be a way of reducing abrasivity in some cases. In others, it may be the reduction in the tendency to form large agglomerates of abrasive particles. It is therefore suggested that the approach used to reduce abrasivity be a system approach, involving all aspects of paper manufacturing.

VI.

SIMILAR WEAR SITUATIONS

Several other wear situations have characteristics similar to those of wear by paper. Wear caused by fabrics, threads, and polymer webs during their manufacture and processing and by printer ribbons and magnetic tapes in their use tend to be similar. In these cases these materials are easily worn or damaged by the counterface. The counterface is expected to withstand a very large amount of rubbing while these materials experience a negligible amount. Although other wear mechanisms may also be involved in some of these situations, abrasive wear by particles is generally the primary or predominant mode [11,12,22]. For example, in the case of printer ribbons, hard particles contained in the ink are primarily responsible for the wear [11]. However, there can be signi®cant oxidative or corrosive effects because of the che-

500

Bayer

(a)

(b)

Fig. 19 Variation in the wear rate of an electroerosion print head as a function of the print pattern. The number of ®res per unit of sliding distance, n, is a characteristic of different printing situations. It increases with the average amount of printing on a page. (a) Experimental data; (b) the model proposed for this behavior. (From Ref. 14.)

Paper Abrasivity

501

Fig. 20 Modi®cation of the drum test con®guration shown in Fig. 2a, which was used to investigate the effect of ink on paper abrasivity.

Table 4 Wear Volumes (in 10 Abrasivity (ASTM G56)a

Material 7C27MO2 420 SS 420F SS 17-7pH/CH900 416 SS 430 SS

6

cm3 ) for Different Materials in ASTM Test for Ribbon Ribbon

Hardness (Rc)

A

B

C

D

E

F

55 55 55 48 48 48

6.6 3.3 4.9 32.8 14.8 13.1

6.6 6.6 6.6 14.8 13.1 11.5

16.4 6.6 6.6 36.1 21.3 18.0

13.1 4.9 6.6 18.0 11.5 18.0

4.9 1.6 8.2 6.6 11.5 13.1

9.8 1.6 11.5 11.5 9.8 9.8

a Test duration 1 h. Source: Ref. 18.

mical nature of the ink. This can result in signi®cant difference in the wear resistance of materials of similar hardness, as shown in Table 4. For these situations, similar techniques are used to characterize the abrasivity of these materials and to determine wear resistance [4,18,22]. Solutions to wear problems are similar, such as the use of materials with hardness greater than Rc60.

REFERENCES 1. 2. 3. 4. 5.

American Society for Testing and Materials. (1976). ASTM STP 615. Selection and use of wear tests for metals. American Society for Testing and Materials. (1979). ASTM STP 664. Erosion: Prevention and useful applications. American Society for Testing and Materials. (1980). ASTM STP 701. Wear tests for plastics: Selection and use. American Society for Testing and Materials. (1982). ASTM Standard G56. Standard test method for abrasiveness of ink-impregnated fabric printer ribbons. American Society for Testing and Materials. (1982). ASTM STP 769. Selection and use of wear tests for coatings.

502 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29.

30. 31. 32.

Bayer American Society for Testing and Materials. (1988). ASTM STP 1010. Selection and use of wear tests for ceramics. American Society for Testing and Materials. (1993). ASTM STP 1199. Tribology: Wear test selection for design and application. American Society for Testing and Materials. (1995). ASTM Standard G40. Standard terminology relating to wear and erosion. Anstice, P. D., McEnaney, B., and Thornton, P. C. (1980). Wear of paper slitting blades: An examination of worn blades from paper mills. Tribol. Int. December:259±266. Anstice, P. D., McEnaney, B., and Thornton, P. C. (1981). Wear of paper slitting blades: The effect of slitter machine settings. Tribol. Int. October:257±262. Bayer, R. G. (1978). Wear by paper and ribbon. Wear 49:147±168. Bayer, R. G. (1978). Mechanism of wear by ribbon and paper. IBM J. Res. Dev. 22(6):668±674. Bayer, R. G. (1983). The in¯uence of hardness on the resistance to wear by paper. Wear 84:345±351. Bayer, R. G. (1983). Wear in electroerosion printing. Wear 92:197±212. Bayer, R. G. (1984). Aspects of paper abrasivity. Wear 100:517±532. Bayer, R. G. (1984). Abrasiveness of electrosensitive papers. Proc. 1984 Int. Printing Graphic Art Testing Conf. TAPPI Press, Atlanta, pp. 73±77. Bayer, R. G. (1985). Wear testing. In: Mechanical Testing. Volume 8. Metals Handbook. 9th ed. J. R. Newby, ed. American Society for Metals, Metals Park, OH, pp. 599±608. Bayer, R. G. (1994). Mechanical Wear Prediction and Prevention. Marcel Dekker, New York. Bayer, R. G., Baker, D., and Ku, T. C. (1968). Abrasive wear by paper. Wear 12:277± 288. Blau, P. J., ed. (1992). Friction, Lubrication, and Wear Technology. ASM Handbook, Vol. 18. ASM International, Metals Park, OH. Bower, A. F., and Johnson, K. L. (1989). The in¯uence of strain hardening on cumulative plastic deformation in rolling and sliding line contact. J. Mech. Phys. Solids 37(4):471±493. Broese van Groenou, A. (1983). The sphere-on-tape: A quick test on wear of materials used in magnetic recording. Proc. Int. Conf. Wear Materials. ASME, pp. 212±217. Cole, G. F. (1972). The prediction of wear by paper. Wear 21:141±154. Engel, P. A., and Bayer, R. G. Abrasive impact wear of type. J. Lubrication Technol. SME Trans. Ser. F 98(2):330±333. Hokkirigawa, H., and Li, Z. (1987). The effect of hardness on the transition of abrasive wear mechanisms of steels. Proc. Int. Conf. Wear Materials, Vol. II. ASME, pp. 585±593. Klinga, L. P., and Back, E. L. (1964). Fiber building board variables in¯uencing the wear of cutting tools. Svensk Papperstidn. 67:309±316. Kurrle, F. L. (1980). Correlation of IPC needle penetration with guillotine trimmer knife life. Tappi 63(7):100±102. Larsen-Basse, J. (1969). On abrasive wear by paper. Wear 14:133±136. Laufmann, L., Brautigam, G., Gerteiser, N., and Rapp, H. (1985). Natural ground CaCO3 in alkaline and neutral papermaking: Synthetic wire abrasion; CaCO3 as copy paper ®ller. 1985 Alkaline Papermaking/TAPPI Seminar Notes. TAPPI, Atlanta, pp. 61± 72. Moore, M., and Swanson, P. (1983). The effect of particle shape on abrasive wear: A comparison of theory and experiment. Proc. Int. Conf. Wear Materials. ASME, pp. 1±11. Payne, N. G., and Bayer, R. G. (1991). Friction and wear tests for elastomers. Wear 150:67±77. Peterson, M. B., and Winer, W. O., eds. (1980). Wear Control Handbook. American Society of Mechanical Engineers, New York, NY.

Paper Abrasivity

503

33. Rabinowicz, E. (1965). Friction and Wear of Materials. Wiley, New York. 34. Raski, J. Z., and Borch, J. The abrasiveness of commercial printer and copier papers. Unpublished report (1988). IBM Corporation, Endicott, NY. 35. Richardson, R. C. D. (1969). Abrasive wear by paper. Wear 14:423±430. 36. Roshon, D. D. (1974). Testing machine for evaluating wear by paper. Wear 30:93±103. 37. Roshon, D. D. (1978). Electroplated diamond-composite coatings for abrasive wear resistance. IBM J. Res. Dev. 22(6):681±686. 38. Suh, N., Sin, H.-C., and Saka, N. (1980). Fundamental aspects of abrasive wear. In: Fundamentals of Tribology. N. Suh and N. Saka, eds. MIT Press, Boston, pp. 493±518. 39. Veronesi, V., Engel, P. A., and Ferrari, A. (1981). The application of electro-deposited composite coatings to parts worn by cardstock: A case history. Proc. Int. Conf. Wear Materials. ASME, New York, NY, pp. 747±752. 40. Zum Gahr, K. (1981). Formation of wear debris due to abrasion. Proc. Int. Conf. Wear Materials. ASME, New York, NY, pp. 396±405. 41. Zum Gahr, K., and Mewes, D. (1983). Severity of material removal in abrasive wear of ductile metals. Proc. Int. Conf. Wear Materials. ASME, New York, NY, pp. 130±139.

14 TESTING THE TACTILE PROPERTIES OF TISSUE AND NONWOVENS SUEO KAWABATA* Kyoto University Kyoto, Japan

I. The Tactile Property II. The Objective Evaluation of Fabric Hand A. Primary Hand and Total Hand B. The Standardization Process for Primary Hand and Total Hand C. The Objective Evaluation System D. Conversion Equations E. The Mechanical Parameters

507 507 509 511 512 513

III. Measuring System

518

IV. Predictive Ability of the Objective Evaluation

519

V. Hand Chart VI. Application to Nonwovens A. Agreement in the Subjective Assessment B. Measuring Conditions for the Mechanical Parameters C. Objective Measurement Procedures VII. Application to Facial Tissue Paper VIII. Other Materials IX. Importance of the Objective Method

*

506

520 521 522 522 522 525 525 526

Appendix 1. The Constants Applied to the Suiting Equations, Eqs. (1)±(3)

527

Appendix 2. The KESF-B Testing Machine System

528

References

530

Professor Emeritus, Kyoto University. 505

506

I.

Kawabata

THE TACTILE PROPERTY

Many materials are used in close contact with the human body, creating interactions between the material characteristics and the human senses. Although utility performance such as the strength of the material and durability are important properties for these materials, consumers seek better quality, that is, more comfort, in materials that satisfy performance requirements to some extent. Material performance in relation to ``better ®t to the human body and senses'' is an essential requirement for materials that interact with humans. However, conventional methods for evaluating material ®tness for human use have relied on subjective inspection of the tactile properties of the material using the human hand. This subjective method is referred to as ``hand evaluation.'' Because of the importance of the tactile property, much effort has been invested in establishing an objective method for evaluating the tactile property of many human-interactive materials. Some people, however, misunderstand the objective method and believe that the evaluation of tactile properties is based on the sense of the individual and that there are no common criteria for such evaluation. Although the evaluation of the tactile property is related to human senses, it is not based on artistically motivated emotion; rather, it is based on a more physiological response related to the human sense of comfort. This sense of comfort is common to all humans; therefore, common criteria may exist between individuals. A typical example of human-interactive materials is clothing fabric. Although the primary function of clothing fabric is to protect the human body from the environment and mechanical injury, its tactile property is also an important performance aspect. For example, when a nonwoven material is used to construct clothing, its tactile property is of primary importance, in addition to its utility performance. If its utility performance is within a standard range, the quality of the nonwoven material depends mainly on the tactile property. When a nonwoven material is, for example, used as cushioning material in a machine that is completely isolated from human contact while in use, the quality of the part is ®rst evaluated by its utility performance, such as cushioning properties, and the tactile property of this part is of little importance. This material is, however, still a human-interactive material, because there are many times when people handle it during the process of cushion manufacture and assembly, and workers would prefer a material that feels nice over one that does not if there is not much difference between the utility performance of the two. Thus, suitability to human senses is an important aspect for nearly all materials used in human-interactive environments. The quality of clothing fabric depends on a comfortable tactile property rather than mechanical strength. People have long examined clothing properties by means of a subjective method, the so-called hand evaluation of fabric. The manufacture of textiles is now a major industry with modern technologies, and an objective evaluation method is necessary for the industrial production of textiles. An objective system for evaluating the tactile property of clothing fabrics was developed by Kawabata around 1975 on the basis of his fundamental research on fabric mechanics and many other pioneering works in this ®eld. This method is now being applied widely in the textile industry and also to other sheetlike and human-interactive materials [1±4]. In this chapter, this objective method is introduced for application to papers and nonwovens.

Tactile Properties of Tissue and Nonwovens

507

Paper is also a human-interactive material. Important investigations aimed at creating a system for the objective evaluation of paper quality are being conducted by the paper industry. When consumers are purchasing facial tissues and paper towels in the marketplace, they, of course, prefer better quality. Although quality is related to the function of these papers, consumers seek comfortable papers and judge them by hand touch, by how they feel. Many scientists and engineers in the paper industry have come to recognize this aspect of paper quality and have made great efforts to discover what consumers want and what physical properties of paper are related to its quality as judged by hand touch. This includes not only the quality of comfort related to touch, but also the functional properties of papers, for example, the printing performance of paper, which is subjectively evaluated by printers, paper producers, etc. [5±8]. The interaction between material properties and human senses is also important for papers, and this is another aspect investigated by the paper industry. A preference-difference model has been developed and introduced in detail by Ramsay [9] and Lyne et al. [10]. A method applied in these investigations is multidimensional scaling (MDS), the basic principle of which was developed by Torgerson [11,12]. Lyne et al. made extensive application of MDS to solve the relationship between consumer preference and the physical properties of paper [10]. They found that surface softness is a basic property that is related to the preference of consumers for papers, especially for facial tissue papers, in addition to rigidity and resilience. They also obtained the interesting result that the surface geometry of paper is also an important factor in determining consumer preference. An open formation is preferred. They concluded that in choosing such papers consumers prefer higher surface softness, lower rigidity, and high or moderate open structure. II.

THE OBJECTIVE EVALUATION OF FABRIC HAND

A.

Primary Hand and Total Hand

Clothing fabrics must ®t human needs and expectations. For many years, people have subjectively evaluated this ``®tness'' by physically touching fabrics. A common idea of a ``®t'' fabric is one in which the fabric's mechanical properties provide comfortable wear and a soft surface, and one that provides beautiful static and dynamic garment silhouettes during wear. ``Hand evaluation'' is the inspection of these comfort-related and aesthetic properties of fabric by a subjective method. As mentioned earlier, the sensations that are evaluated during hand evaluation are not purely physiological but are also based on experience and training. The ®rst step we took in approaching an objective evaluation method was to get reliable experts who could judge fabric hand based on rich knowledge and experience in fabric hand evaluation. We found these experts among professionals in fabric ®nishing and weaving in textile mills, especially in the ®nishing mills, which work in close proximity to ®nal fabric products. Ten was selected as the appropriate number of experts for this study. It was important that the experts belong to different mills or organizations to avoid localism in the judgment. It is also true that consumer preference is a source of judgment standard for fabric hand; however, individual consumers do not necessarily have vast experience in judging fabric quality. The experts who work close to the ®nal stages of the textile manufacturing

508

Kawabata

process may judge the fabric hand based on consumer preference by using information feedback from the consumers and consequently have relatively common criteria for judging fabric hand. The initial discussion between the experts focused ®rst on ®nding important and frequently used terms for fabric hand and second on the criteria for the hands de®ned by these terms. It was discovered that there are several important expressions related to hand that are commonly used by the experts. For winter/autumn suiting, they are Stiffness This is not equal to simple stiffness but describes a complex expression of fabric property related mainly to bending stiffness. A springy property promotes this feeling. A fabric that has a compact weave density and is made from springy, elastic yarn yields an intense feeling of stiffness. This property is necessary for suiting to create a moderate space between the body and the suit and to allow comfortable body motion. This may be understood if we imagine a suit wetted by water, which sticks to the body and hinders body motion. Smoothness This term denotes a mixed feeling coming from a combination of smooth, supple, and soft feelings. A fabric woven from cashmere ®ber exempli®es this feeling. This smoothness is especially important for highly quality fabrics. Fullness A feeling coming from a combination of bulky, rich, and wellformed impressions. A springy property in compression and thickness, accompanied by a warm feeling, is closely related to this property. Fullness increases the deformability of fabric in the low load region. For midsummer suiting, Crispness is used instead of Smoothness, and Antidrape is added as follows. Stiffness The same Stiffness as in winter/autumn suiting. Crispness A feeling coming from a crisp and ridged fabric surface. This is found in a woven fabric made from a hard and strongly twisted yarn. Crispness creates a cool feeling. Fullness The same Fullness as in winter/autumn suiting. Anti-drape The opposite of limp conformability, whether the fabric is springy or not. This property is necessary to create a large space between the body and clothes to maintain air circulation inside the clothing. There are many other expressions of fabric hand that express fabric characteristics. However, the hands listed above are primary hands that are commonly used by experts for expressing fabric characteristics. In addition, each of these hands may be graded according to its intensity such as very strong, strong, above average, etc. There are many other hands; however, they are not commonly used as primary hands, they can be expressed by a combination of the primary hands, and their intensity is expressed by only two grades of intensity: ``present'' or ``absent.'' For these reasons, we have grouped these common and primary hands as ``Primary hand.'' Primary hand expresses the physical characteristics of a fabric. There is another hand that expresses fabric quality such as good or poor. This is the overall hand, which evaluates fabric performance relating to the ®tness of a fabric for use by humans as suiting. Consumers evaluate this performance based on fabric tactile

Tactile Properties of Tissue and Nonwovens

509

properties and experience. Experts evaluate this quality based on the Primary hands as an overall hand and grade the fabric quality as excellent hand, good hand, poor hand, etc. We called this ``Total hand.'' We also derived an objective hand evaluation system based on this expert system. Figure 1 shows the process of subjective hand evaluation by experts. Usually, the total time spent for this judgment was 10± 20 s for one sample.

B.

The Standardization Process for Primary Hand and Total Hand

Approximately 500 samples of commerical men's suiting were collected, and their primary hands and total hand were judged by the experts using the following subjective method. Primary Hand The experts con®rmed the importance of the selected three (or four) primary hands and discussed the de®nition of each of the primary hands to coordinate their individual understanding of Primary hand de®nitions. Then each expert judged all samples piece by piece on the basis of his own criteria. First, the expert divided the fabric samples into three groups according to the degree of conformance to each primary hand: a strong feel, weak feel, and neutral feel. Each of these groups was again divided into the three subgroups in the same way as the ®rst sorting. All samples were eventually divided into nine groups by this procedure, the samples that possessed especially strong feeling were separated from the highest feeling group, and the samples that possessed especially weak feeling were separated from the lowest feeling group. As a result, all samples were graded into 11 groups in order of feeling intensity for each Primary hand by each expert, as shown in Fig. 2. Standard samples were selected for each grade by considering the agreement between the experts' judgments. The feeling intensity of the group was expressed by numbers

Fig. 1

The sequence of subjective judgment of fabric hand by experts.

510

Kawabata

Fig. 2

Sampling of the standard samples for Primary hand and the Hand values.

ranging from 10 for the highest feeling intensity to 0 for the lowest. These numbers were labeled ``Hand value'' or ``HV''. Total Hand The Total hand value was also standardized by the experts for men's suiting in the same way as the standardization of Primary hand, but with ®ve grades. Samples were ®rst graded into three groups: good quality, poor quality, and midrange quality. Excellent quality samples were selected from the samples in the good group and low quality samples from the poor group; then the samples were graded into the ®ve grades shown in Fig. 3. Standard samples were selected for each of these ®ve grades in the same manner as with Primary hand, and these were also expressed by numbers ranging from 5 (excellent) to 1 (poor). This grading number was labeled ``Total hand value'' or ``THV.'' Although the experts came close to agreeing on criteria for hand judgment, there were still individual differences. However, the standard samples enabled us to coordinate the criteria of the experts. The experts could numerically evaluate Hand value and Total hand value by comparing the hand of a sample with that of the standard samples. It became possible to express the fabric hand of a fabric numerically, for example, Hand value of Stiffness ˆ 5:8, Smoothness ˆ 6:5, Fullness ˆ 7:0, and THV ˆ 3:8. Decimal values are occasionally used to express a feeling between

Fig. 3

Sampling of the standard samples for Total hand and the Total hand value.

Tactile Properties of Tissue and Nonwovens

511

grades and also as the average of scores from two or more. Although this evaluation method is still subjective, it is based on standardized criteria. This is already a semiobjective method. Another important point of this method is that nonexpert people can evaluate fabric hand on the basis of professional criteria.

C.

The Objective Evaluation System

Fabric mechanical properties, including fabric surface properties such as surface friction and geometrical contour, are closely related to fabric hand and can be detected by hand touch. Instead of hand touch, these mechanical properties are measured objectively and expressed by parameters in the objective system, the parameters are converted into Primary hand values. Then these Primary hand values are converted into THV, as shown in Fig. 4. There are two important procedures in the objective system. The ®rst is the selection of the mechanical properties, the properties related to fabric hand, and the parameters that express each of these properties. These parameters must express the fabric properties accurately. First, in selecting the fabric properties, a basic and simple property is more desirable than a complex property because it facilitates fabric structural design for control of the fabric mechanical properties. Because fabric mechanical properties are generally nonlinear, we have to ®nd the region most related to deformation or load with hand evaluation. The fabric hand is related to a relatively low load region, and the full property range up to the breaking load is generally not necessary. From these viewpoints, basic research on fabric properties is required for the suitable selection of these parameters. Second, the principle of this objective system is based on the subjective evaluation of experts, and numerical expression of the subjective evaluation is necessary to connect the subjective evaluation with the fabric hand. As already mentioned, we had standardized the Hand values and Total hand value of suiting. Such preparation is also necessary to apply this method to other materials. The subjective evaluation must be conducted by approximately 10 experts from different mills or organizations, and the results must be averaged in order to obtain generality in the objective evaluation.

Fig. 4

The sequence of objective hand evaluation.

512

D.

Kawabata

Conversion Equations

The ®rst conversion equation converts the fabric mechanical parameters into Primary hand values. The 16 mechanical parameters Xi (i ˆ 1; 2; . . . ; 16) were selected for expressing fabric mechanical properties as shown in the next section, and a linear equation was applied to derive Primary hand values Yk (k ˆ 1; 2; 3 for winter suiting; see Section II.A) from Xi as follows. Yk ˆ c0 ‡

16 X iˆ1

cki xi

…1†

where c0 and cki are constant values and xi is the value of Xi normalized by the population mean mi and standard deviation i of Xi , as follows. xi ˆ

Xi

i

mi

…2†

Equation (1) is a linear equation; however, it does not necessarily express a linear relation between the mechanical parameters and Yk . Some of the parameters are transformed into logarithms so that the distribution of each parameter in the population becomes a symmetrical form with nearly Gaussian distribution. The second conversion equation for converting Primary hand values into Total hand value is a nonlinear equation that was designed considering the nonlinear effect of Primary hand on Total hand, as follows [1]. THV ˆ C0 ‡

3 X

Zk

…3†

kˆ1

where Zk ˆ Ck1

Yk

Mk1

k1

‡ Ck2

Yk2

Mk2

k2

…4†

and C0 , Ck1 , and Ck2 are constant coef®cients; Mk1 and Mk2 are the population means of YK and Yk2 , respectively; and k1 and k2 are the standard deviations of YK and Yk2 , respectively. Constant coef®cients were derived by applying a multivariable regression method as follows. The coef®cients for conversion equation (1) were derived by using stepwise block regression. This is because, ®rst, fabrics generally possess nonlinear mechanical properties. Also, one propertyÐfor example, the tensile propertyÐis expressed by a set of three parameters, which we call a ``block.'' The fabric property consists of six properties, that is, six blocks, and each block consists of two or three parameters, for a total of 16 parameters. Each of these blocks expresses a property of fabric, and it is not desirable to decompose a block by the process of formulating the regression equation. Second, there are some correlations between parameters, which makes it dif®cult to understand the meaning of the derived coef®cients. Stepwise regression [13] was applied to avoid this dif®culty. Consider a multivariables regression equation y ˆ f …x1 ; x2 ; x3 †, where x1 , x2 , and x3 are variables. First, the three regression equations y ˆ f1 …x1 †, y ˆ f2 …x2 †, and y ˆ f3 …x3 † are derived independently, and the equation having the highest predictive ability is selected, for example, y ˆ f2 …x2 †. Second, the residual value, E ˆ Y y, is

Tactile Properties of Tissue and Nonwovens

513

regressed with the remaining parameters x1 and x3 to get equations E ˆ g1 …x† and E ˆ g3 …x3 †, where Y is the experimental value of y. The equation that has the highest predictive ability is selected. Then an equation, for example, E ˆ g3 …x3 †, is added to the ®rst equation y ˆ f2 …x2 † to get the second step equation, y ˆ f2 …x2 † ‡ g3 …x3 †. The residual of this equation is again regressed with the other remaining parameters and applied to the second step equation. This procedure is repeated until the last parameter is reached. The stepwise ``block'' regression is a kind of stepwise regression in which the variables are replaced with a set of parameters of a block and the blocks are regressed stepwise. If necessary, the parameters in each block can be redetermined by applying stepwise regression within the block. The second conversion equation converting Primary hand value YK into THV was derived by applying ordinary linear multivariable regression. However, it includes squares of the variables. This is due to the consideration that an optimum value of the Primary hand value may exist that will yield the highest THV. The coef®cients of Eqs. (1) and (4) are shown in Appendix 1. As will be shown in Section VI, it has been found that these equations may be applicable to predicting similar hand properties for nonwovens, papers, etc., without any change of the constant coef®cients. That is, we may use Eqs. (1) and (3), replacing only the population mean and standard deviation with those of the population of the new materials. E.

The Mechanical Parameters

The mechanical properties and parameters are listed in Table 1. Tensile Property The test specimen is 20 cm in width and 5 cm in length, and extension is applied in the lengthwise direction (Fig. 5) under a constant rate of tensile strain. This is a type of biaxial extension mode called ``strip biaxial extension.'' This deformation mode creates a simpler relation between the fabric structure and the tensile property than does uniaxial extension. The linearity, resilience, and tensile energy are parameters that describe the tensile property. Bending Property Pure bending is applied to a specimen (Fig. 7) at up to 2:5 cm 1 of bending curvature under a constant rate of bending curvature. The bending stiffness and hysteresis of bending moment are measured as the parameters of the bending property. Shearing Property Fabric specimens of the same size as those used in tensile testing are subjected to a constant tension, 10 gf =cm, in the lengthwise direction; then shear deformation is applied (Fig. 6) at up to 8 of the shear angle under a constant rate of shear displacement. Shear rigidity and hysteresis of the shear force at 0:5 and 5 of the shear angle are measured as the shear parameters. Compression Property The specimen is compressed in the thickness direction by up to 50 gf =cm2 of pressure under a constant rate of compression displacement (Fig. 8). The load±thickness curve is similar to that of the tensile property. Linearity, resilience, and compression energy are measured as the parameters of the compression property.

514

Kawabata

Table 1 Mechanical and Surface Properties Standardized for the Objective Hand Evaluation for Suiting Parametersa Tensile LT WT RT EMc

Description

Unit

b

Bendingb B 2HB Shearingb G 2HG 2HG5 Compression LC WC RC Surfaceb MIU MMD

Linearity of load±extension curve Tensile energy Tensile resilience Extensibility, strain at 500 N/m (gf =cm of tensile load)

None N/m (gf cm=cm2 ) % None

Bending rigidity Hysteresis of bending moment

10 10

Shear stiffness Hysteresis of shear force at 0:5 of shear angle Hysteresis of shear force at 5 of shear angle

N(m deg) [gf =…cm deg†Š N/m (gf =cm†

4 2

N m …gf cm2 =cm† N …gf cm=cm)

N/m (gf =cm†

Linearity of compression±thickness curve None Compressional energy N/m (gf cm=cm2 ) Compressional resilience % Coef®cient of friction None Mean deviation of coef®cient of friction None (frictional roughness) Geometrical roughness m

SMD Construction T Fabric thickness W Fabric weight per unit area

mm 10 g=m2 …mg=cm2 †

a The parameters are applicable to other material such as nonwovens and tissue papers with some changes of measuring conditions. b Average of the values in warp and weft directions is applied. The warp and weft directional values are identi®ed by 1 and 2, respectively, such as in MMD-1, B-2, etc. c EM is not used in the equation for conversion to HV. Source: Refs. 1 and 3.

Surface Property Surface friction and geometrical contour are measured using the special contactors shown in Figs. 9 and 10. The friction is measured with a 5  5 mm surface consisting of parallel steel wires placed perpendicular to the sweep direction under a constant contact force of 50 g and a constant sweep velocity of 1 mm/s. The frictional force varies because of the surface structure of the fabric. The mean frictional coef®cient and mean deviation of the frictional coef®cient are measured as the parameters of surface friction. The surface contour is measured by a Ushape contact that is given a constant contact force of 10 gf . The mean deviation of the thickness variation is measured under the same sweep velocity as that of the friction measurement. For the measurement of woven materials and papers, the friction is measured more sensitively by this U-shape contactor.

Tactile Properties of Tissue and Nonwovens

515

Fig. 5 The tensile property. Parameters: LT (Linearity of load-extension curve† ˆ WT=‰area of triangle Q1 Q2 Q3 ], WT (Tensile energy† ˆ ‰area Q1 A Q2 Q3 ]. RT (Tensile resilience† ˆ ‰area Q1 B Q2 Q3 Š=WT; EM Extensibility, strain at 500 N/m.

Fig. 6 The shear property. Parameters: G (Shear stiffness); 2HG (Hysteresis of shear force at 0:5 of shear angle); 2HG5 (Hysteresis of shear force at 5 of shear angle).

Fig. 7 The bending property. Pure bending is applied on a specimen. Parameters: B (Bending rigidity) Measured from mean slope in the range K ˆ 0:5  1:5 cm 1 ; 2HB (Hysteresis of bending moment) Measure at K ˆ 0:5 cm 1 .

Fig. 8 The compression property. Parameters: LC (Linearity of compression curve) ˆ WC=[area of triangle Q1 Q2 Q3 ]; WC (Compressional energy† ˆ ‰area Q1 A Q2 Q3 ]; RC (Compressional resilience† ˆ ‰area Q1 B Q2 Q3 Š=WC; T0 (Initial thickness used as ``Thickness,'' de®ned as the fabric thickness at pressure P ˆ 0:5 cm2 ). 516

Fig. 9 The surface property. Sweep distance 2 cm. (a) Contact element for friction. (b) Contact area for geometrical roughness. (c) Mean deviation is the average deviation of the hatched area.

Fig. 10 The contact surface of the friction sensor is a ®ngertip surface simulated by parallel steel wires 0.5 mm in diameter. 517

518

Kawabata

The frictional force and geometrical contour are converted into electrical signals by transducers consisting of linear transformers, and then their signals are processed by a second-order high-pass ®lter to cut the frequency components that are lower than 1 Hz. When we touch and sweep across a fabric with a ®nger, the sweep velocity is around 5 cm/s, and 1 Hz in the testing sweep corresponds to 50 Hz in an actual sweep, as shown in Fig. 11. The frequency zone from 50 Hz to 300± 400 Hz is the most sensitive frequency range for human senses. Fabric Structure: Thickness and Weight of Fabric Tensile, bending, and surface properties are measured in both the warp direction (machine direction) and weft direction (the transverse direction), averaged, then substituted into the conversion equation.

III.

MEASURING SYSTEM

The mechanical parameters described in Section II, except fabric weight, are measured with newly developed instruments. These instruments were developed in parallel with the development of the standardization of fabric hand. The tensile and shear properties, the bending property, the compression property, and the surface property are measured by four different machines. These four machines have been commercialized as the KESF-B-1,-2,-3, and -4, respectively. These machines are now popular, mainly in the textile industries, and their use is gradually expanding to nonwovens, papers, and other ®elds such as the cosmetics industry. Figure 12 illustrates the principles of these systems. A specimen 20  20 cm2 is used in each of the four machines. In 1997, a fully automated model of the KESF was developed, and the operation was greatly simpli®ed.

Fig. 11

A high pass ®lter attenuates the low frequency component of the surface signal.

Tactile Properties of Tissue and Nonwovens

Fig. 12

IV.

519

The measuring system KESF-1±4. This system has recently been fully automated.

PREDICTIVE ABILITY OF THE OBJECTIVE EVALUATION

The ability of the objective evaluation system to measure a sample's THV is high, and the predicted Hand value and THV are within the individual variation range of the experts' subjective evaluations, close to the average value of the scattered values. Figure 13 shows the predicted values using the objective system for new samples that were not used in formulating the equation. Although the scattering of the experts' evaluation values of THV is usually over a wide range for this judgment, the equation can yield a prediction relatively close to the average value of the experts' evaluation. Figure 14 shows curves for Z1 , Z2 , and Z3 in Eq. (3) for men's autumn/ winter suiting. These curves show the contribution of each Primary hand value to the THV. This is a powerful tool for the design of high quality fabrics.

520

Kawabata

Fig. 13 The correlation between the predicted THV derived by the objective system and the subjective THV obtained by the average score of 10 experts.

V.

HAND CHART

The Hand chart shown in Fig. 15 is convenient for expressing hand properties of a fabric. Hand values and THV are plotted on this chart. The shadowed zone in the chart is the good zone, and the hand values of fabrics with higher THV are normally within this region. This zone was derived on the basis of the statistical analysis of many commercial fabrics. Data for the three samples shown plotted on this chart are given in Table 2. Sample 1 is a good hand sample, sample 2 is a good but borderline sample, and sample 3 is outside the high THV zone, with a THV lower than those of

Fig. 14 The contribution of Primary hand to the THV of winter/autumn suiting. The original expressions in Japanese are given in parentheses.

Tactile Properties of Tissue and Nonwovens

521

Fig. 15 Distribution of the hand values of three fabric samples on the hand chart. The primary hands were originally de®ned and standardized with Japanese names. (*) Sample 1; (~) sample 2; (&) sample 3. Screened area is the high THV zone.

Table 2 Example of THV and Hand Values for Three Samples of Winter/Autumn Suiting

THV Primary hand values HV of Stiffness HV of Smoothness HV of Fullness

Sample 1

Sample 2

Sample 3

Remarks

4.5

3.7

2.7

Max. value 5

4.0 7.7 7.0

3.6 6.7 5.3

7.7 3.7 4.1

Max. value 10 Max. value 10 Max. value 10

the other two samples. Figure 15 conveniently provides a full overview of fabric hand characteristics and quality. It clearly indicates which primary hand values must be improved to increase THV, that is, to improve quality.

VI.

APPLICATION TO NONWOVENS

The method for developing the objective evaluation system that has been introduced here for suiting material may also be applicable to other human-interactive materials. We have developed an evaluation system for nonwovens [2]. Because nonwovens are relatively new human-interactive materials compared with traditional woven fabrics, the expressions used for fabric hand are not as clearly de®ned, nor as common, as the hand expressions related to suiting. It is therefore dif®cult to ®nd experts like those we found for the hand evaluation of men's suiting. Nonwovens, however, are used close to human skin, such as in apparel and sanitary and medical products, so they are also a human-interactive material. Nonwovens are commonly used in everyday products, and consumers seriously consider the tactile feel of these materials in their use. Consumers usually compare products made by different companies to select the product that has the best touch. In this circumstance, consumers

522

Kawabata

can judge the quality to some extent, but it is dif®cult to ®nd professional experts to provide a hand evaluation. Following the development of the objective evaluation system for woven fabrics, we extended the investigation of the hand evaluation system to nonwovens. The subjective hand evaluation of nonwovens was conducted by nonprofessionals, mainly graduate students majoring in clothing science, who are also consumers of these nonwoven products. Some professional people also participated in the evaluation, but they were not necessarily professional in nonwovens. As there are no common and clear expressions corresponding to those used for the primary hand of suiting, we assessed only the quality resulting from touch, that is, good touch, poor touch, and midrange touch, to obtain ®ve grades of THV in the same manner as the grading of the THV for suiting. The samples were commercial products with a fabric weight of 40 g=m2 to 70 g=m2 , and some controlled fabrics were included. A.

Agreement in the Subjective Assessment

Seventy-eight nonwoven samples were assessed. They were all commercially produced nonwovens, except for 18 controlled samples. There were 21 judges for THV, including graduate students and experts in the assessment of fabric hand for suiting. The correlation coef®cient between THV evaluated by these judges is shown in Table 3. Group I (18 samples) consisted of controlled samples with different ®ber denier. Group II (38 samples) were commercial fabrics and popular spunbonded fabrics; and group III (22 samples) were a mixture of various types, mostly spunbonded. All of the samples used for this trial were collected with no consideration given to fabric type. The criterion for the collection was that the fabrics had a range of thickness similar to that of suiting and shirt fabrics. We found that there was good agreement in the judgments between these three evaluation groups. We used the average value of THV from these three groups for further analysis. B.

Measuring Conditions for the Mechanical Parameters

The same measuring equipment as those used for suiting may be applied with conditions modi®ed as given in Table 4. A new modi®ed contactor for measuring surface friction was also examined. The contact surface for friction testing using the standard KESF-B4 surface tester is made of 10 parallel steel wires; however, it was found that the single wire contact is more sensitive in the case of nonwovens for detecting delicate frictional properties, especially MMD, which is the most important parameter related to THV for nonwoven fabrics. If this new sensor is not available, the MMD and MIU measured with a standard sensor can be applied, although the predictive ability of the THV decreases slightly. See Ref. 3 for more details. C.

Objective Measurement Procedures

The procedure is the same as the one used to derive the THV of suiting. The Primary hand values are derived ®rst, then the THV is derived from the Primary hand values. It was con®rmed that this two-step method is more effective than deriving the THV directly from the mechanical parameters. The same equations as those used for suiting were applied to derive the Primary hand values. The normalized variables

b

0.899 0.955 1 0.975 0.970 1 0.992 0.876 1 0.974

1 0.970 0.993 1 0.876 0.963

Female students

0.938 1 0.955 0.985

Male students

Number of judges. The mean value of the correlation coef®cients between individuals and the mean scores within groups. Source: Ref. 2.

a

1 0.938 0.899 0.970

Experts

Between groups

Correlation Coef®cients Between and Within Groups of Judges in the Assessment of THV of Nonwoven Samples

(a) Sample I, number of experimental samples N ˆ 18 Experts (7)a Male students (6) Female students (8) All (21) (b) Sample II, number of experimental samples N ˆ 38 Male students (10)a Female students (9) All (19) (c) Sample III, number of experimental samples N ˆ 22 Male students (9)a Female students (9) All students (18)

Table 3

0.716 0.808 0.762

0.826 0.849 0.715

0.938 0.908 0.913 0.898

Within groupsb

Tactile Properties of Tissue and Nonwovens 523

524

Kawabata

Table 4

New Measuring Conditions for Nonwoven Fabrics

Property Tensile Maximum load Shearing Maximum shear angle Measurement of G Measure of 2HG Measurement of 2HG5 Surface measurement MIU MMD and SMD

Unit 50 N=m …gf =cm† 0:4 Between 0 and 0:4 of shear angle At 0:2 of shear angle Not measured Value in the rough direction Values in the transverse direction

in Eq. (3) were derived using the mean and standard deviation of the nonwoven population. We call this method ``population transfer'' in this chapter. For the derivation of THV, we investigated two methods: Method 1 The THV equation that was derived for nonwovens was applied to the THV. That is, the equation is the same type as Eq. (2) except that the coef®cients in the equation are those that were derived for nonwovens. Method 2 The same equation as that for suiting is applied using the ``population transfer'' for the THV, that is, with the same coef®cients as Eq. (2) but using new variables for every Primary hand value normalized by the mean and standard deviation of the nonwoven population. A direct method for deriving the THV from the mechanical parameters was also examined. The results of the predictive ability of the THV are shown in Table 5. The new equation yields the best predictive results and can predict THV with nearly the same level of correlation as subjective evaluation. It is also interesting to note that the application of the THV equation for suiting with the population transfer method yields nearly the same predictive accuracy as the new equation. This fact suggests that there is a common criterion for good touch in human-interactive materials. This may be con®rmed by the plot of Primary hand values on the Hand chart for suiting in Fig. 11. In Fig. 16, the hand values are plotted for the good nonwoven samples having higher values of THV (shown by open circles) and the Table 5 The Correlation Coef®cient Between Predicted and Subjective THV (Mean Score of 18±21 Judges) and the Root-Mean-Square of Prediction Errora Equation New equation for nonwovens Winter/autumn suiting equation Direct equationb a

Correlation coef®cient R

Prediction error (rms)

0.854 0.701 0.503

0.549 0.767 2.876

These R and rms values were examined by using new samples that were not used for the regression. Equation converting mechanical parameters directly into THV. Source: Ref. 2.

b

Tactile Properties of Tissue and Nonwovens

525

Fig. 16 The hard values of the nonwoven samples plotted on the hand chart for suiting. Circles indicate the good quality samples selected by consumers, and crosses indicate poor quality samples. The primary hand values were derived by the objective method; THV was obtained by subjective judgment.

poor samples having lower values (shown by crosses). The circles representing good samples are completely within the good zone for suiting, and the crosses representing poor samples are outside the zone. This plot indicates a trend in which poorer samples possess excessively high stiffness but low smoothness and fullness. VII.

APPLICATION TO FACIAL TISSUE PAPER

The objective evaluation method developed for the quality evaluation of suiting was applied to the quality evaluation of facial tissue paper by Kawabata and Nima [3]. The mechanical parameters were measured under the same conditions as nonwovens, and the population transfer method was applied. That is, we did not need new equations for the derivation of both Hand values and THV but applied the same equations as those used for suiting. For inspection of the prediction of tissue paper quality, a subjective assessment of THV was conducted in parallel with an objective evaluation. The group of judges consisted of 10 specialists working in hair salons and 10 ordinary consumers, including male and female students. In the objective evaluation, the Primary hand values were ®rst derived, then the THVs were derived by applying the population transfer method. Table 6 shows that the Primary hands of tissue papers are closely related to the THVs of the papers and that the THV of the paper can be derived by this objective method. In Fig. 17, the hand values of three different papers are plotted on the hand chart used for suiting. There are high, middle, and low THV samples. It can be seen that the hand chart for suiting is applicable to the evaluation of the quality of tissue paper. VIII.

OTHER MATERIALS

Application may be extended to other human-interactive materials. Quality evaluation of leathers and automobile upholstery materials [3] was investigated by applying the population transfer method, and it has been con®rmed that the same hand chart as that used for suiting (Fig. 11) can be commonly applied to the quality evaluation of these materials. There are some cases, however, in which it is dif®cult to measure

526

Kawabata

Table 6 Correlation Coef®cient Between the Subjective THV and Objective Values of Each Primary Hand and THV of Facial Tissue Papersa Value

Correlation coef®cient b

Stiffness (koshi ) Smoothness (numerib ) Fullness (fukuramib ) THV

0:24 0.95 0.85 0.84

a The subjective values of THV were evaluated by a joint group of professionals in hair salons and ordinary consumer groups. The objective values of HV and THV were derived based on the equations for winter/autumn suiting. b Original expression of standardized Primary hand.

all of the mechanical properties sued for suiting; for example, boardlike material is too stiff or the bending property to be measured, and the bending property is not important for evaluating the tactile property of such materials. It is not necessary to use all the mechanical properties to apply the equation developed for suiting to the derivation of Primary hand values. For leather materials, the same conditions as those used in nonwoven measurement were applied, but the tensile property was omitted. For automobile panel sheets, only the surface, compression, and construction (thickness and weight) parameters were applied. The step rank of these parameters is higher than those of other parameters used in the stepwise regression, and we may eliminate other parameters from the regression equation with only a slight reduction in the accuracy of the predictive ability of the Primary hand values This is a strong point of stepwise regression.

IX.

IMPORTANCE OF THE OBJECTIVE METHOD

The primary use of paper has been as a printing material. The use of paper is expanding, but it is still a member of the human-interactive materials family. People have loved the tactile feel of paper, even as a printing material, and prefer

Fig. 17 The hand values and THV of the three samples of facial tissue papers are plotted on the hand chart. (*) Excellent quality sample; (~,X) poor quality samples.

Tactile Properties of Tissue and Nonwovens

527

paper with a better touch. The similarity in the quality judgments of suiting and tissue paper was discussed in Section VI. This suggests that there is a common criterion for quality stemming from the tactile property of human-interactive materials. This criterion is based on whether or not a material is comfortable. Humans do not like materials that injure their skin when they are touched. They generally like a soft touch. ``Soft'' and ``pleasant'' have the same meaning for most people. In addition to this basic criterion, experience with respect to comfortable use of a material leads to a clearer criterion for hand judgment. In addition to the utility performance of paper, the hand property of paper, which interacts with the human senses, will be an essential performance aspect for paper in future. We must therefore develop an objective method for evaluating the hand properties of paper. APPENDIX 1.

THE CONSTANTS APPLIED TO THE SUITING EQUATIONS, EQS. (1)±(3)

Table A1 Hand Value Equation Parameters for Evaluating the Primary Hand Value for Men's Winter Suitinga Ci Mechanical parameters Tensile LT log WT RT Bending log B log 2HB Shear log G log 2HG log 2HG5 Compression LC log WC RC Surface MIU log MMD log SMD Construction log T log W

Smoothness (numeri) C0 ˆ 4:7533 (5) 0:0686 0.0735 0:1619 (4) 0:1658 0.1083 (3) 0:0263 0.0667 0:3702 (2) 0:1703 0.5278 0.0972 (1) 0:1539 0:9270 0:3031 (6) 0:1358 0:0122

Stiffness (koshi) C0 ˆ 5:7093

Fullness (fukurami) C0 ˆ 4:9799

(4) 0:0317 0:1345 0.0676 (1) 0.8459 0:2104 (2) 0.4268 0:0793 0.0625 (5) 0.0073 0:0646 0:0041 (6) 0:0254 0.0307 0.0009 (3) 0:1714 0.2232

a Equation KN101-W series. Numbers in parentheses indicate order of importance. Source: Refs. 1 and 4.

(3) 0:1558 0.2241 0:0897 (6) 0:0337 0.0848 (4) 0.0960 0:0538 0:0657 (1) 0:2042 0.8845 0.1879 (2) 0:0569 0:5964 0:1702 (5) 0.0837 0:1810

Population parameters, men's winter suitings …n ˆ 214† Mi

i

0:6082 0.9621 62.1894

0:0611 0.1270 4.4380

0:8673 1:2065

0.1267 0.1801

0:0143 0.0807 0.4094

0.1287 0.1642 0.1441

0.3703 0:7080 56.2709

0.0745 0.1427 8.7927

0.2085 1:8105 0.6037

0.0215 0.1233 0.2063

0:1272 1.4208

0.0797 0.0591

528

Kawabata

Table A2 3:1466

Suiting THV Equation Parametersa for Men's Winter Suiting, where C00 ˆ

k

Yk

1 2 3

Smoothness Stiffness Fullness

Ck1 0:1887 0.6750 0.9312

Ck2 0.8041 0:5341 0:7703

Mk1

Mk2

k1

k2

4.7537 5.7093 4.9798

25.0295 33.9032 26.9720

1.5594 1.1434 1.4741

15.5621 12.1127 15.2341

a Equation KN301-W. Source: Refs. 1 and 4.

APPENDIX 2.

THE KESF-B TESTING MACHINE SYSTEM

As we have seen in this chapter, it is possible to measure several mechanical and surface properties of a fabric or sheetlike material by modifying conventional basic testers such as tensile testers. These measurements are actually dif®cult, however, because they must be made in the very low load regions and also because mechanical properties are very delicate and behave nonlinearly. In the early stage of the development of an objective evaluation system, four testing machines were developed for fabric mechanical properties. The measurement principles of these machines have been described in this chapter. This system has been commercially available throughout the world for 20 years, sold mainly to textile manufacturers and textile research laboratories. In recent years the application of these four testers to other materials such as papers and cosmetic materials has become very common. The four testers are usually sold separately for these applications. The most popular one in the paper

Fig. A1

The KESF-AUTO model.

Tactile Properties of Tissue and Nonwovens

529

industry, especially in the United States, is the surface tester. The bending tester is also used in the paper industry. This is understandable, considering the results of Lyne et al.'s investigation [10]. The design of the system has been revised once, and now we are close to completing a second set of design revisions. The third model offers completely automated measurement operations of all stages from specimen chucking to data processing. For more information and technical questions, please contact the manufacturer at Kato Tech Co., Nishi Kujyo Karato cho, Minami-ku, Kyoto 601, Japan. Figures A1 and A2 illustrate the third generation KESF-B system.

(a)

(b)

530

Kawabata

REFERENCES 1. 2. 3. 4. 5. 6.

7. 8. 9. 10. 11. 12. 13.

Kawabata, S., and Niwa, M. (1989). Fabric performance in clothing and clothing manufacture. J. Textile Inst. 80(1):19±50. Kawabata, S., and Niwa, M. (1994). Objective hand measurement of nonwoven fabrics. Part 1. Development of the equation. Textile Res. J. 64(10):597±610. Kawabata, S., and Niwa, M. (1996). Recent progress in the objective measurement of fabric hand. In: Macromolecular Concept and Strategy. Okamura, RoÃnby, and Ito, eds. Springer-Verlag, Berlin, pp. 81±103. Kawabata, S., and Niwa, M. (1996). Objective measurement of fabric hand. In: Modern Textile Characterization Methods. M. Raheel, ed. Marcel Dekker, New York, pp. 329± 353. Lyne, M. B. (1979). Tappi 62(11):103±107. Lyne, M. B., Parush, A., Richer, F., Jordan, B. D., Donderi, D. C., and Ramsay, J. O. (1983). A multidimensional analysis of paper-related factors in the subjective evaluation of print quality. In: The Role of Fundamental Research in Papermaking. I. Brander, eds. Mech. Eng. Pub., London, pp. 655±683. Parush, A. (1980). The relationship among perceived contrast noise and content in printed images. M.A. Thesis, McGill University. Lyne, M. B., and Parush, A. (1983). A survey of North American newsprint print quality. J. Pulp Paper Sci. 9(5):TR117±TR123. Ramsay, J. O. Psychometrika 45(2):149±165. Lyne, M. B., Whiteman, A., and Donderi, D. C. (1984). Multidimensional scaling of tissue quality. Pulp Paper Can. 85(10):T259±T265. Torgerson, W. D. (1952). Psychometrika 17(4):401±419. Torgerson, W. S. (1958). Theory and Method of Scaling. Wiley, New York. Draper, N. R., and Smith, H. (1966). Applied Regression Analysis. Wiley, New York.

INDEX

Abbott pro®lometer, 432 Abrasivity, 477±503 abrasion, three-body, 481 abrasion, two-body, 481 abrasive wear, 478±494 ash effect on, 489 business papers, 488±494 characteristics, 488±494 coef®cient, abrasive wear, 478, 482, 488, 496±497 density effect on, 490 electroerosion printing, wear, 490, 496, 499 evaluation, 483±488 ®ller effect on, 478, 488±490, 497 humidity effect on, 490, 497 impact printing, wear, 495±496 inks, 492, 497, 499±501 mechanisms, 478±483 moisture effect on, 491±492, 497 paper debris effect on, 490 slitter blades, wear, 478, 496 strength effect on, 490±492 test, drum con®guration, 484±485 tests, wear, 478, 483±488 wear, of paper, 482±488, 489±490 wear behavior, 494±501 wear rate, speci®c, 481, 482±483, 488, 495±496 wear resistance, of coating, 495 x-ray analysis, 489 Absorptance intensity of light, 98

[Absorptance] Kubelka±Munk theory, 97±98, 101, 116, 123, 127±133 Absorption, light coef®cient, 98, 127±129, 131±133 power, 127 Absorption, liquid (see Penetration, liquid and Wetting) Air permeability, 282, 290±294 Bekk tester, 293 Bendtsen porosity tester, 293 coef®cient, 282±287, 289±290 constant, 282 Emanueli porosity tester, 293 Emiel Greiner porosity meter, 293 Gurley densometer, 290±293 Gurley±Hill S-P-S tester, 293 method, ¯owmeter, 293±294 method, volumetric, 293 Oken-type denso-asperometer, 293 permeation time method, 290±293 pulp, 240, 284±285, 294 resistance, 290 routine control, 290±294 Schopper air permeability meter, 293 Shef®eld porosimeter, 293 Smooster, 293 speci®c, 282 Williams tester, 293 Air resistance, 282, 290 Amontons' laws, 452, 454 Anisotropy elastic constants, 403 531

532 [Anisotropy] electrical, 334±335, 346±347, 351±355 friction, 466 mechanical properties, 403±404 thermal expansion, 403±404 Ash (see also Filler) on-line measurement, 38 Assman psychrometer, 57 Atmosphere, ISO standard, 52, 314 Atmospheric conditioning, environment cabinet, 67±68 Atmospheric conditioning, laboratory, 5, 47±93, 52 air supply, 4±5 controls, 64±67 international standards, 52 standards by country, 52 Atmospheric mixing ratio, 49 Atomic force microscope (AFM), 150, 152, 167±168, 216±246 analysis, qualitative, 224±237 applications, 224±246 chemical force, 223, 245±246 contact mode, 220±221 detection mechanism in, 220 electric force, 223 environmental conditions for, 223±224 feedback electronics in, 220 ®ber characterization by, 226±232 (see also Fiber) force modulation, 222 intermittent contact mode, 221±222 lateral force, 222, 245±246 magnetic force, 223 material compatibility analysis by, 242±246 material interactions analysis by, 242±246 morphology application, 237±238 nanomechanical properties determined by, 242 noncontact mode, 221±222 operation, 217±223 optical, near-®eld scanning, 223 phase detection, 223 phase imaging, 244±245 principles, 217±219 probe, 219 scanner, 220 scanning thermal, 223 surface feature quanti®cation by, 237±242 surface roughness evaluation by, 238±242 (see also Roughness, surface)

Index [Atomic force microscope (AFM)] topogra®ner, 217 variants, 222±223 Autoignition temperature de®nition, 404 measurement, 406 Basis weight, on-line measurement, 14, 25±29 Bekk tester, 293 Bendtsen porosity tester, 293 Beta-ray, back scatter, 38 Beta-ray, transmission, 118 Breakdown strength, electrical (see Dielectric strength) Bridge methods for dielectric properties tests Schering bridge, 381 transformer bridge, 381 Wheatstone bridge, 381 Brightness, 96, 99±102 on-line measurement, 38±39 standard methods, 99±103 Bristow test, 309, 317±320, 329 Brunauer±Emmett±Teller (BET) equation, 271 Burning point, de®nition, 405 Bursting test, 3 Calibration, 5±7, 9 Caliper (see Thickness, sheet) Calorimetry, 391±393 Capillarity, 272±282, 284±285, 304±317, 325 capillary imbibition, 304±313 capillary pressure, 272, 307 capillary radius, 272, 275, 284±285, 307, 317 capillograph, 314±316 Hagen±Poiseuille equation, 273, 284 Laplace equation, 307 Lucas±Washburn equation, 307, 316 Capacitance, electrical, 335±336, 366±367, 382 Capillography, 314±316, 330 Cell wall delamination (see Fibrillation, internal) Cellulose biodegradability, 199 chain structure, 233±237 cotton, 152 (see also Cotton) depolymerization, 408±418 dielectric constant, 347±349, 370

Index [Cellulose] electrical conductivity, 338±344 ®brils, (see Fibril and Fibrillation) friction, 456 friction coef®cient, 454, 468 liquid penetration, 319 loss tangent effects, 368±369 phase detection, 223 staining, 225 (see also Staining) structure, 161 swelling by interchain bond breaking, 304 thermal decomposition, 408±424 thermal expansion, 403±404 Charge decay, 380±381 decay time, 380 maximum potential, 380 Chromaticity coordinates, 105±110, 123 diagram, 105±106 Clausius±Mosotti relation, 347 Coating analysis, 153, 156±157, 165±166, 176, 193, 195±197, 203, 216, 278 components, phase imaging, 244 ¯uorescence, 195 optics, 129±133 particles, microscopy analysis, 156 penetration, 329 pore size distribution, 272±273, 281±282 thickness, 166, 195 wear resistance, 495 weight, microscopy analysis, 156 weight, on-line measurement, 37±38 Cobb test, 304, 325±326 Coef®cient, abrasive wear, 478, 482, 496±497 Coef®cient, friction, 452±472 ash effect on, 470 cellulose, 454 fatty acid effect on, 456, 469 handsheets, 468±469 kinetic, 452, 454, 457±458, 461, 463±464, 467±468, 470, 472 liner, dispersion-sized and extracted, 468 moisture effect on, 464 nylon 6-to-nylon 6, 454 oscillations, 462±463 plastics, 454 pulp, 468±469 static, 452, 457±459, 467±468, 470 Color (see also Colorimetry) charts, 108 difference, 109±110

533 [Color] dominant wavelength, 106 excitation purity, 106 gloss effect on, 138 hue, 105, 108, 111 intensity, 104 lightness, 104 matching, 108 metamerism, 108, 110 object, 104 on-line measurement, 38±39 saturation, 106 space, 108±109 strength, 108 temperature, 106 wavelengths, 97±98, 103±105 Colorimetry, 103±112, 118±125, 138 charts, 108 chromaticity coordinates, 105±112, 123±125 CIE system, 103±112 illuminants, 106±108 instrumentation, 118±127 metameric matching, 108 Munsell system, 108 scales, 108±109 standard observer, 104 trichromatic coef®cients, 105±106 tristimulus values, 103±110, 138 visual color matching, 108 visual ranking, 111±112 Combustion, 404±407 (see also Decomposition, thermal) burning energy, 407 burning point, 405 ¯aming, 405, 418 ¯ash point, 405, 406 heat, 407 ignition temperature, 404, 406 oxygen index, 405, 407 smoldering, 418 Compound microscope, 153 Conduction, thermal, 390±400 conductivity (see Conductivity, thermal) speci®c heat, 391±393 Conductivity, electrical, 334±335, 338±346 AC (alternating current), 335, 338 cellulose, 338±344 DC (direct current), 335, 338±346 ®ber, 346, 377 ionic, 341±344, 369 liquid penetration test for, 321±322

534 [Conductivity, electrical] moisture in¯uence, 338±340 static charge, 374 surface, 335, 374, 379 volume, 334, 338±346 Conductivity, thermal, 390, 393±400 ASTM steady state method, 394±396 density effect on, 399±400 electrical insulation, 396 ®ller effect on, 399 heat pulse method, 397 Kartovaara's method, 394, 397±398 measurement, 394±400 moisture effect on, 399±400 pulp, 399 steady state methods, 394±396 surface roughness effect on, 399 Terada's method, 396 Terasaki's method, 396 thermoacoustic method, 397±398 unsteady state methods, 396±398 Contact angle, 304±309, 314±317, 329 Controls, humidity, 60±64, 69±79 multistage, 62 on-off, 60±62 pressure vessel, 60 proportional, 62±64 tables/graphs, 69±79 two-temperature, 64 Convective transfer theory, 58 Cotton (see also Cotton ®ber) dielectric constant, 348±349 electrical conduction, 338±339 ¯ash point, 406 ignition temperature, 406 oxygen index, 407 Cotton ®ber artifacts, 211 cell wall structure, 161 cross-section, 161 ®brillar network, 231±235 surface, 225 surface roughness, 240 Curl ®ber, 153 sheet, on-line measurement, 37 Darcy's law, 282±284, 326 Data acquisition accuracy, 5 precision, 5 sampling, 5

Index Decomposition, thermal, 408±424 analytical methods, 421±424 bond scission, 409±414 charring, 406±421 DP (degree of polymerization) changes, 408±418 energy output for, 408±409 high-temperature reactions, 414±418 inhibition, 419±421 isothermal degradation, 408±409 lignin, 424 low-temperature reactions, 408±414 molecular weight changes, 409±414 oxidative degradation, 408±409 protection against, 419±421 pyrolysis, 408±423 tar formation, 414±421 weight loss in, 408±409 Defects, paper, on-line detection, 42 Densitometry, optical, 118, 120, 126±127 Density caliper method, 269 determination, 269±270 mercury method, 269 microscopy methods, 155, 164, 193 surface roughness effect on, 269 Depolarization currents, 352±355 Depolymerization, thermal, 408±418 Dielectric breakdown strength (see Dielectric strength) Dielectric constant (relative permittivity), 337, 346±351 anisotropic, 346±347, 351±352 ash in¯uence, 347, 349, 369 capacitor paper, 366 cellulose, 347±349, 370 complex, 337 density in¯uence, 369 electrical grade papers, 369±370 electrostatic transfer, 363 frequency effect on, 348±349 measurement, 381±383 moisture in¯uence, 339±344, 347±348, 350±351, 369 pulp, 346, 370 salt effect on, 349±350 thermal depolarization technique for, 355 water effect on, 350±351 Dielectric loss, 334, 336±337, 346, 365±369, 381±383 (see also Dielectric loss tangent) angle, 336

Index [Dielectric loss] breakdown strength (see Dielectric strength) current, 336±337 electrical grade papers, 365±369 factor, 337, 346 measurement, 381±383 Dielectric loss tangent, 336±337, 366±369, 382±383 carboxyl effect on, 368 cellulose (crystalline fraction) effects, 368±369 hemicellulose effect on, 367±368 ion effect on, 369 lignin effect on, 367±368 measurement, 382±383 Dielectric strength, 334, 337, 365±366 AC (alternating current) test, 383±385 air impermeability, 371 barrier effect, 371 DC (direct current) test, 383 density effect on, 366, 371 dipole oscillation (rotation), 367 electrical grade papers, 365±366, 370±371 ®nes effect on, 371 impulse test, 383 measurement, 383±385 nondielectric papers, 341 oil impregnation, 366 process effect on, 371 thickness effect on, 365, 370±371 time effect on, 383 uniformity, 371 Differential scanning calorimetry, 281, 391±393 Differential thermal analysis, 421, 424 Diffraction gratings, 152 x-ray, 489 Diffuser, perfect re¯ecting, 125 Diffusion, gas coef®cient, 286±287, 289±290 permeation, 286±288, 289±290 solid-solution, 288 Diffusion, liquid, 313 Dissipation factor, electrical (see Dielectric loss tangent) Drop penetration test, 304, 321±322 Elastic modulus, on-line measurement, 40±41

535 Electrical properties, 333±355, 361±385 anisotropic, 334±335, 346±347, 351±355 capacitance, 335±336, 366±367, 382 charge, static, 372±374 (see also Charge decay) Clausius-Mossotti relation, 347 conductivity (see Conductivity, electrical) density effect on, 366 depolarization, 352±355 dielectric parameters (see Dielectric constant, Dielectric loss, and Dielectric strength) dipole moment, 202, 231, 368 dissipation factor (see Loss tangent) electrical grade papers, 365±371 ®bers, 346±347 frequency effect on, 348±351 hardboard, 349 insulation resistance, 335, 365±366, 375±380 ion concentration effect on, 347±348, 369 loss parameters (see Dielectric loss) measurements, 375±380 moisture in¯uence, 339±344, 346±348, 350±355 orthotropic, 334±335 permittivity (see Dielectric constant) polarization, 338, 346±347, 367, 372 relaxation time, 344 resistance (see Resistance, electrical) salt effect on, 349±350 surface conductivity, 335, 374, 379 surface resistivity, 335, 363±365, 375±380 temperature effect on, 341±344, 347±349, 352±355, 375±376 time effect on, 344±345, 377, 380±381, 383 V-t test, 383 volume conductivity, 334, 338±346 volume resistance, 335 volume resistivity, 335, 343±344, 363, 375±379 water effect on, 350±351 wood, 346, 349 Electrical grade papers, 365±372 Electricity, static, 372±375 charge, static, 372±374 conductivity, 374 neutralizer, static, 374±375 triboelectric series, 372 triboelectricity, 372 Electroerosion printing, wear, 490, 496, 499

536 Electron microscope (EM), 151, 159±167, 203±216 Electron spectroscopy for chemical analysis, 330 Electrophotography, 361±365 paper requirements, electrical, 362±365 relaxation time constant, charge, 363 (see also Resistivity, electrical) thickness, dielectric, 363 Emanueli porosity tester, 293 Emiel Greiner porosity meter, 293 Environment, cabinet, 67±68 Environment, laboratory air supply, 4±5, 67 conditioning, 5, 47±93 controls, 60±64 health/safety, 9, 67 in¯uences, inside, 5 in¯uences, outside, 4±5 lighting, 9, 67 noise, 67 testing rooms, 64±67 vibration, 4 Environment, testing rooms, 64±67 Expansion, thermal, 400±404 anisotropic, 403±404 density effect on, 404 linear coef®cient, 401±404 measurements, 402±404 moisture effect on, 400±401, 403±404 pulp, 403±404 volume coef®cient, 403±404 Fiber analysis (identi®cation), 153, 176±178 bagasse, 162 bonding, 160±162, 164±165, 188, 190±193, 198, 228, 238±240, 246, 304 cell wall structure, 155±157, 164, 166 coarseness, 153 collapse, 155, 158, 164, 188, 194, 208 conductivity, electrical, 346, 377 conformability (¯exibility), 154, 160 cotton (see Cotton ®ber) critical point drying, 163 cross-section, 155±162, 166±167, 172, 175, 183±192, 194±195, 212±213 curl, 153, 186 deformation, 166, 194 delamination, 157, 209, 213±216 density, 191 distribution, 191

Index [Fiber] electrical properties, 346±347, 355 ®brillation, 154, 164, 188±191, 212±213, 213±232 (see also Fibril) ¯exibility, 190±191 fractions, 153 fracture, 166 friction, 246, 467 hygroexpansion, 194 length, 153±155, 158, 186 lignin in (see Lignin) loading, 166 lumina, 157, 166 orientation, 101, 134, 193, 267, 304, 346, 351 orientation, on-line measurement, 34 penetration, liquid, 288, 304, 319, 324 phase imaging, 244±245 roughness, 238±242 scattering, light, 134 shrinkage, 188, 194, 208 spruce, 161±162 staining, 176±178, 188 stiffness (see ¯exibility) strength, 155, 158 stress, 157 structure, 152, 157, 161, 163±64, 211±216 surface, 152±155, 188±194, 198, 223±245, 304, 330 swelling, 150, 161, 167, 208, 213±216, 224, 288, 304, 313, 319 voids, 166 wall thickness, 172, 188, 194, 197 weight, 153 wetting, 304, 314±316, 318 width, 153±154, 186, 188, 198 wood, 154, 157, 164, 185, 232, 235, 238, 240, 242 wrinkling, 242 Fibril angle, 156±157, 187 orientation, 157, 187 Fibrillation, 188±192 (see also Fibril) beating effect on, 154 bonding, 193 cotton ®ber, 231±235 external, 188±190, 209±212 index, 154, 164. 190 internal, 161, 190±192, 212±216 network, 231±232 Fick's laws, 286, 313

Index Filler (see also Ash) analysis, 153 characterization, 164±166 distribution, 156, 193, 195, 216 examination, 174, 176, 193 ¯uorescent properties, 176, 195 light scattering, 130 in micropores, 274 optical properties, 97 Fines 192±193, 195, 212 analysis, 175±176, 188, 192±193, 195, 212 characterization, 153, 158, 166 classi®cation, 211 detection, 158, 191, 195 in micropores, 274 Flash point de®nition, 405 ®ller effect on, 406 measurement, 406 Flotation (dye) test, 322 Flow free molecular (Knudsen), 283 turbulent, 283 Flowmeter, capillary (rotameter), 289, 293±294 Fluorescence, 96, 110±111, 118, 123, 174 coating, 195 lignin, 186, 191, 195 microscope, 151, 158±159 quanti®cation, 178 Formation, 97, 101, 138±139 ¯oc frequency, 117, 139 microdensitometers, 120 on-line measurement, 33±34 testers, 116±118, 120 theory, 138±139 wavelength spectra, 139 Frequency, ¯oc, 117, 139 Fresnel re¯ection (scattering), 98, 112±115, 117±118, 131, 136±137 Friction, 451±476 Amontons' laws, 452, 454 ash effect on, 470 clothing fabric, 514±518 coef®cient (see Coef®cient, friction) continuous contact testing equipment, 460 ®bers, 246, 467 ®ller effect on, 462, 468±470 horizontal table setup, 457±459 inclined plane equipment, 459 instrumentation, 457±461

537 [Friction] kinetic, 456, 458±459, 462±463, 467 (see also Coef®cient, friction) lignin effect on, 469 measurement, 457±466 moisture effect on, 464, 468 nonwovens, 522 paper-to-metal, 452, 470±471 paper-to-paper, 452, 457±458, 460, 462, 467±470 paper-to-rubber, 472 pulp, 470 repeated contact, 461±462 rotating platen evaluation, 457 sample preparation, 464±466 sample variance, 460 standard methods, 464 static, 467 (see also Coef®cient, friction) stick and slip, 455±456, 463 strength effect on, 453±454, 456, 461, 464, 466, 468, 470, 472 strip-on-drum method, 460 surface energy effect on, 456, 461, 468±469 surface roughness effect on, 453, 455, 467±468, 471 surface topography effect on, 244±246 test variance, 464±466 Friele±MacAdams±Chickering color formula, 109 Gas adsorption, 270±271, 272±273 gas drive method for pore size distribution, 272±273 Hagen±Poiseuille equation, 273 speci®c surface measurement by, 270±271 Gas±liquid chromatography, 414, 421±423 Gas permeability, 282±297 ¯ow, 87±88, 283 ¯owmeter, capillary (rotameter), 289, 293±294 Hagen±Poiseuille equation, 273, 284 Knudsen ¯ow, 284 mean free path, 283±284 measurement, 288±294 molecular ¯ow, 283±284 Reynolds number, 283 turbulent ¯ow, 283 Gloss, 96, 112±115, 137±138 bloom, absence, 113 contrast (luster), 113±115 distinctness-of-image 113, 115

538 [Gloss] glossmeter, 114, 119±120, 123, 126±127 goniophotometers, 114, 118, 120, 126±127 interpretation, 137±138 measurement, 113±115 on-line measurement, 39±40 pro®lometry, 447 roughness effects, 138 sheen, 113 specular, 113, 127 Goniophotometry, 118, 120, 126±127 (see also Gloss) Gurley densometer, 290±293 Gurley±Hill S-P-S tester, 293 Hagen±Poiseuille equation, 273, 284 Heat pulse method for conductivity, 396±397 (see also Thermal properties) Hercules size test, 322 Horizontal table test, friction, 457±459 Hoyland's apparatus, 313 Humidity absolute, 49±50 Assman psychrometer, 57 calibration, 68±69 conditioning, 47±93 convective transfer theory, 58 cycle, 64 dew point tables, 69±78 electronic psychrometer, 56 enclosures, 64±68 hygrometer, dew point, 55±56 psychrometer, wet/dry bulb, 56±60 psychometric coef®cient, 57±58 relative, control, 5, 60±64 relative, de®ned, 49±52 relative, measurement, 52±60 reliability, 52±53 salt solutions, saturated, 68±69, 79 saturation, 51 sensors, 49, 52±60, 62, 68±69, 79 speci®c, 49±50 standard atmospheres, 52 tables/graphs, 78±79 vapor pressure table, 69 Hydrophobicity, 314, 318 lignin, 304 sizing agents, 318±319, 329 Hygroexpansion, 403±404 ®ber, 194

Index Hygrometer, dew point, 55±56 dew-point tables, 69±78 Hygrometry, 49, 52 Ignition temperature de®nition, 404 measurement, 406 IGT print tester, 320±321, 461 Illuminants, 106±108 Image analysis, 97, 118, 120±123, 152, 179, 184, 247 coatings, 165 contact angle measurements, 314 defect detection, on-line measurement, 42 density distribution of pulp, 155 ®ber orientation, on-line measurement, 34 ®ber properties, 153, 193 ®brillation, 190 ®nes, 153, 192 formation, on-line measurement, 33±34 ink particles, 199 instrumentation for, 118, 120±123 for quantitative sterology, 273 sheet surface, 42, 445, 447 surface roughness, 138, 198 Immersion test, 325±326 Inclined plane test, friction, 459 Infrared absorption test, water penetration, 321±324 Interference microscope, 156 Interferometry, 432±435, 438±440, 448 elastic modulus analysis, 41 phase shift, 439 for pro®lometry, 432±435, 438±440, 448 vertical scanning, 439 Inverse gas chromatography, 330 Kubelka±Munk theory, 97±98, 101, 123, 127±136, 138 equations, 62, 128±130, 271 heterogeneous sheets, 129±130 homogeneous sheets, 127±129 limitations, 130±133 Kozeny-Carman equation, 272, 328 Laboratory, testing (see also Environment, laboratory) academic, 4 accuracy, 5±6 automation, 9±10 certi®cation, 6±7

Index [Laboratory, testing] customer, 4 data acquisition/processing, 9±10 electrical stability, 4 equipment, 3±6, 8±9 insulation, 66 interlaboratory reference systems, 10±11 mill, 4 precision, 5±6 quality control, 6±7 research, 3±4 size, 67 statistics, 10±11,18±19 test equipment, 7±8 types, 3±4 water/air supply, 4±5 Laplace equation, 307 Laser scanning microscope, confocal (CLSM), 150±151, 159, 168±203 applications, 185±203 artifacts, 180±184 beating effects characterized by, 188±193 contamination, microbial, examined with, 199 deinking studied by, 199 dual (multiple) wavelength of, 175 extended focus image, 179 ®ber properties studied by, 185±187 (see also Fiber) ®brillation examination by, 212 (see also Fibrillation) ®ller examination by, 174, 176, 193, 195 ®nes examination by, 212, 216 ¯uorescence mode, 169, 174±175, 178 imaging modes, 174±175, 178±180 ink transfer studied by, 198±199 Leica CLSM, 168 network structure investigated by, 193±198 objective, 173±174 optical section rotation for, 180 pinhole size, 172±173 polymers, synthetic, studied by, 199 principles, 168±174 pulp characterization by, 188±198 pulp examination by, 172±173, 175, 185 raw materials identi®ed by, 185±187 reconstruction, three-dimensional, for, 179±180 re¯ection mode, 169, 174 sheet structure investigated by, 193 simulated ¯uorescence process for, 179

539 [Laser scanning microscope, confocal (CLSM)] slide preparation for, 175±176 specimen preparation for, 175±179 staining for, 176±179 step size for, 174 stereo images, 180 topographical image, 179, 197 transmission mode, 175 Leica CLSM, 168 Light absorptance, 98±99, 127±133 absorption coef®cient, 127 color wavelengths, 103±105 diffraction (see scattering) dominant wavelengths, 105±106 illuminant standards, 106±108 intensity, 98, 117 interactions with paper, 95±148 luminance factor, 108, 110 luminosity, 97, 104, 110 re¯ectance, 96±115, 123±139 scattering (see Scattering, light) transmittance, 96±99, 115±118, 127±136 Lightness (color intensity), 104 Lignin dielectric losses for, 367±368 distribution in cell wall, 156, 158±159, 163, 166, 185±186, 191, 228, 304 ¯uorescence, 186, 191, 195 glass transition temperature, 228 heat stability, 408, 417, 434 hydrophobicity, 304 identi®cation on ®bers, 244 labeling with bromine gas, 166 mass spectrometry, 422 phase detection, 223 pyrolysis, 417 removal, 159 staining, 162±163, 178, 191 thermal decomposition, 424 Lucas±Washburn equation, 307, 316 Luminosity, 97, 104, 110 Luster, 113±115 Marangoni effect, 313 Mass spectroscopy, 421±423 MD/CD/ZD properties (see Anisotropy) Mercury intrusion porosimetry, 274±280 Micro®bril (see also Fibril and Fibrillation) quanti®cation, 224 structure, 151, 161

540 Microscopy, 149±265 atomic force (AFM) (see Atomic force microscope) calibration of magni®cation diffraction gratings, 152 latex spheres, 152 compound, 153 electron (EM), 151, 159±167, 203±216 ®ber analysis by, 153±168, 172, 175±199, 204±217, 222±247 (see also Fiber) ®brillation examination by, 151±152, 154, 160±164, 187±190, 209±216, 224±232, 235 (see also Fibril and Fibrillation) ¯uorescence, 151, 158±159 ¯ying spot, 151 interference, 156 laser scanning, confocal (CLSM) (see Laser scanning microscope, confocal) optical, 151±159, 168±203 phase contrast, 156 polarizing, 151, 157±158 pulp analysis by, 153, 166±167, 177, 204 re¯ection electron (REM), 163 relative bonded area, measurement, 153±154 (see also Relative bonded area) scanning electron (SEM), 151, 160, 163±167 scanning electron, environmental (ESEM), 167 scanning electron, low-temperature (LTSEM) (see Scanning electron microscope, low-temperature) scanning electron, topographical (topo SEM), 165, 435, 437, 440±441 scanning probe, 167±168, 216±246 scanning transmission electron, 167 scanning tunneling, 152, 167 stereoscopic, 153 techniques, 152±168 transmission electron (TEM), 160±163 ultraviolet (UV), 151, 158±159, 163 Mie theory, 98, 130, 134 Moisture air, laboratory standards, 52 air, measurement, 48±60 controls, 60±64 gravimetric method, 49 hygroscopic sensors, 53±55 on-line measurement, 14, 29±32, 125

Index [Moisture] thermodynamic methods, 55±56 Mottle, 115 Mullen tester, 3 Munsell color charts, 108 Neutralizer, static, 374±375 Nitrogen adsorption (see Gas adsorption) Nuclear magnetic resonance spectroscopy, 281±282 Oken-type denso-asperometer, 293 On-line testing, 13±45 ash, 38 basis weight, 25±29 beta-ray, back scatter, 38 brightness, 38±39 caliper, 32±33 CD resolution, 21±22 characteristics, 19±22 coating weight, 37±38 color, 38±39 defect detection, 42 delay, 17±18 direct measurements, 19 dynamic measurements, 22, 28 elastic modulus, 40±41 environment, 18 ®ber orientation, 34 formation, 33±34 frequency, 17±18 full-width measurements, 22 gloss, 39±40 inferred measurements, 19 instrumentation, 8 measurement, 17±18 moisture content, 29±32 opacity, 39 optical, 28, 38±40, 125 porosity, 34 pro®le testing, 24±25 reel-based measurements, 19±21 scan average, 20±21 scan-based measurements, 19±21 sensor calibration, 22±24 sensors, 14±15, 22±43 sheet thickness, 32±33 smoothness, 35±36 static measurements, 22 statistics, 19±21 strength, 14, 16, 40±42 structural, 25±38

Index [On-line testing] surface strength, 41±42 web ¯atness (curl), 37 Opacity, 96, 99±103, 116, 131, 138±139 factor, 131 formation effect on, 138±139 on-line measurement, 39 paper backing, 102±103 printing, 103, 116 standard methods, 99±103 white backing, 102±103 Optical microscope, 151±159, 168±203 Optical properties, 95±148 absorptance (see Absorptance) brighteners, 110±111 brightness (see Brightness) character recognition, 115 coatings, 96, 128±134 colorimetry (see Colorimetry) densitometry, optical, 118, 120, 126±127 ®bers, 97, 136 ®llers, 97 ¯uorescence, 96, 110±110, 118, 123, 151, 158±159, 174, 178 formation (see Formation) gloss (see Gloss) instrumentation, 118±127 on-line control, 125 opacity (see Opacity) re¯ectance (see Re¯ectance) standardizing laboratories, 125±126 surface texture perception, 112±115 tint, 110 transmittance (see Transmittance) transparency, 116 whiteness (see Whiteness) yellowness, 110 Oxygen index de®nition, 405 measurement, 407 Penetration, liquid, 303±330 Bristow test, 317±320, 329 capillary imbibition, 304 cellulose, 319 Cobb test, 304, 325±326 conductance test, 321±322 Darcy's law, 326 diffusion, 313 drop test, 304, 321±322 ®ber (see Fiber, penetration, liquid) Fick's laws, 286, 313

541 [Penetration, liquid] ¯otation (dye) test, 322 Hercules size test, 322 Hoyland's apparatus, 313 immersion test, 325 infrared absorption, 321±324 Kozeny±Carman equation, 328 Laplace equation, 307 Lucas±Washburn equation, 307, 316 measurement, 314±329 multiple internal re¯ection spectroscopy, 322±324 pH in, 314 polarity test, 321 pressure, 319±320, 326±329 in printing, 309±310, 320±321, 327±329 rate, 326±329 spectroscopy, 322±324 surface energy, 303 surface tension effect, 319 swelling, 304 test methods, 304, 314±329 theory, 305±313 ultrasonic propagation test, 321, 324±325 uptake, total, 325±326 Permeability (see also Air permeability, Gas permeability and Water vapor permeability) diffusion-type, 282, 286±290 (see also Fick's law) diffusion-type measurement, 289±290 ¯ow-type, 282±283, 288±289 (see also Darcy's law) ¯ow-type measurement, 288±289 Permeation (see Permeability) Permittivity, electrical, 335±337, 346±351 complex, 335±337 Phase contrast microscope, 156 Polarity, surface, 306, 321, 372±373 Polarizing microscope, 151, 157±158 Pore size distribution, 268, 272±282 capillary pressure method, 272 coating, 272±273, 281±282 de®nition, 268 differential scanning calorimetry (DSC), 281 gas drive method, 272±273 measurement, 272±282 mercury intrusion, 274±280 nuclear magnetic resonance (NMR), 281±282 qualitative stereology, 273

542 [Pore size distribution] sorption, 272±273 Stone-Scallan plot, 279 swelling, 317 x-ray small angle scattering method, 274 Pore space, 267±268 (see also Pore size distribution) effective, 268 ®ber, 268 sheet, 267±268 structure, 274 volume ratio, 268 Pores, internal (see Pore space, Pore size distribution, Porosity, and Voids) Pores, surface, 279±280 analysis by mercury intrusion, 277±279 liquid penetration of, 309 truncated cones model, 329 Porosimetry, mercury intrusion, 274±280 Porosity, 34, 267±282 (see also Pore size distribution and Speci®c surface area, sheet) bulk density measurement, 269 caliper method, 269 de®nition, 267±268 gas displacement method, 269 liquid displacement method, 269 measurement, 34, 269±270 mercury method, 269 on-line measurement, 34 solid density measurement, 269 Printing dielectric, 333, 365±372 (see also Dielectric constant, Dielectric loss, and Dielectric strength) drying mechanism in, 199 electroerosion, wear behavior, 490, 496 electrophotographic, 333, 361±365 (see also Electrophotography) impact, friction in, 471 impact, wear in, 495±496 ink transfer in, 198±199 liquid wetting and penetration in, 195, 305, 309±310, 313, 320±321, 327±329 nonimpact, heat transfer for, 390 opacity, 103, 116 surface roughening effect in, 194, 446 surface topography effect in, 447 Pro®lometry, 429±450 Abbott pro®lometer, 432 applications, 446±447 autofocusing, 433, 435±437

Index [Pro®lometry] bearing ratio, 444 3D projection, 441 data analysis, 442±444, 448 fractal descriptors, 445±446 gloss analysis by, 447 interferometry (see Interferometry) kurtosis, 444 measurement, 432±441 microstriations measured by, 447 numerical analysis, 442 peak-to-valley height in, 443 printability correlation with, 447 pro®ler, optical autofocusing, 433, 448 scanning electron microscopy, topographical, 165, 435, 437, 440±441 skewness, 444 spectral density analysis, 445 statistics, facet, 445 statistics, slope, 445 stereo-pair projection, 441 stylus, mechanical, 432±433, 435, 437 surface form, 430, 448 surface roughening, 446 surface roughness, 430±432, 448 surface waviness, 430, 448 ten-point height, 443 topo SEM system, 440±441 triangulation, 433, 437±440, 448 Prufbau print tester, 320±321 Psychrometer, wet/dry bulb, 56±60 Pulp (see also Fiber) air ¯ow through, 240, 284±285, 294 analysis, 153, 166±167, 177, 204 bonding potential, 153 (see also Fiber, bonding) characterization, 188±198 cross-section, 204±206 (see also Fiber, cross-section) delamination, 213 dielectric constant, 346, 370 dielectric properties, 346, 348±349 examination, 160±161, 163, 172±173, 175, 185, 208 expansion coef®cient, 403±404 ®nes detection, 158, 191, 195 friction, 470 friction coef®cient, 468±469 heat of combustion, 407 lignin distribution, 185±186, 228 lignin identi®cation, 244

Index [Pulp] lignin removal, 159 microbial contamination, 199±203 papermaking properties, 164 photoyellowing, 185 pore structure, 277±279 re¯ectance, standard methods, 101, 103 speci®c heat, 393 staining, 176±178 (see also Fiber, staining) surface area, 271±272 swelling, 157 (see also Fiber, swelling) thermal conductivity, 399 Pyrolysis, 408±423 (see also Decomposition, thermal) Pyrolyzers, 422 Radius, capillary, 272, 275, 284±285, 307, 317 Rate, speci®c wear, 482±483, 488, 495±496 Re¯ectance, optical, 96±116, 123±139 angular, variation, 112±115, 137±138 cell wall, 135 cellophane, 136 coating, 129±133 color effects, 138 diffuse, 99±103, 112, 136±137 diffuse blue factor, 101±102 diffuser, perfect re¯ecting, 125 directional blue factor, 102 formation effect on, 138±139 Fresnel re¯ection (scattering), 98, 112±115, 117±118, 131, 136±137 gloss, 112±115, 123, 137±138 Kubelka±Munk theory (see Kubelka± Munk theory) pulp, standard methods, 101, 103 re¯ectometry, 115±116, 125±126 roughness effect on, 138 scattering (see Scattering) spectral, 103 specular, 112±115, 137±138 standards, reference, 125±126 surface ®nish, variation, 115 Re¯ection electron microscope (REM), 163 Re¯ectivity, sheet, 136, 139 Refractive index, 98, 112, 128, 131, 134 Relative bonded area (RBA), Relaxation time constant, charge, 363 (see also Resistivity, electrical) Repellency, liquid (see Wetting) Resistance, air, 282

543 Resistance, electrical AC (alternating current), 337 DC (direct current), 375±380, 381±382 insulation, 335 volume, 335 Resistivity, electrical measurement, 375±380 moisture in¯uence, 375±377 surface, 335, 363±365, 375±380 temperature in¯uence, 375±377 volume, 335, 343, 363, 375±379 Reynolds number, 283 Rotating platen evaluation, friction, 457 Roughness index, 317 Salt solutions, saturated, table, 79 Saturation, color, 106 Scanning electron microscope (SEM), 151, 160, 163±167 Scanning electron microscope, environmental (ESEM), 167 Scanning electron microscope, lowtemperature (LTSEM), 150, 167, 203±216 applications, 211±216 cell wall structure examined by, 212±216 coating structure examined by, 216 cryopreparation artifacts in, 209±211 cryosystem, 204±205 ®ber structure comparison by, 211 ®brillation examination by, 212 (see also Fibril) ®ller examination by, 216 (see also Filler) ®nes examination by, 212 (see also Fines) sheet structure investigated by, 216 specimen preparation for, 206±209 Scanning probe microscope, 167±168, 216±246 Scanning transmission electron, 167 Scanning tunneling microscope, 152, 167 Scattering, light, 127±136 coef®cient, 127 ®ber, 134 ®ller, 130 heterogeneous sheets, 129±130 homogeneous sheets, 127±129 Kubelka±Munk theory (see Kubelka± Munk theory) Mie theory (see Mie theory) particulate, 133±136 sheet simulation, 134±136 small particle, 134

544 [Scattering, light] speci®c surface area, determined by, 134 Schering bridge, 381 Schopper air permeability meter, 293 Sensors, humidity, 49, 52±60, 62, 68±69, 79 calibration, 68±69 reliability, 52±53 Sensors, on-line measurement, 14±15, 22±43 calibration, 22±24 Shef®eld porosimeter, 293 Sizing agents, 304, 313 alkenyl succinic anhydride (ASA), 469 alkyl ketene dimer (AKD), 304, 469 Slitter blades, wear, 478, 496 Smooster, 293 Smoothness, on-line measurement, 35±37 Solubility coef®cient, 287 Speci®c heat, 391±393 de®nition, 391 measurement, 391±393 pulp, 393 Speci®c surface area, sheet, 134, 268, 270±272 Brunauer±Emmett±Teller (BET) equation, 271 de®nition, 268 gas adsorption method, 270±271 Kozeny±Carman equation, 272 measurement, 134, 270±272 optical method, 134, 271 silvering-catalytic method, 272 solution adsorption method, 271 Speckle, 115 Spectrophotometry (see Spectroscopy) Spectroscopy electron, 119 ¯uorescence, 110±111, 118 multiple internal re¯ection, 115±116 spectrophotometers, 96, 118±119, 127 Spreading, liquid (see Wetting) Staining, 176±178 (see also Fiber, staining) ¯uorochromes, 177 lignin, 162±163, 178, 191 pulp, 176±178 silver labeling, 178 Stereoscopic microscope, 153 Stone-Scallan plot, 279 Strength, color, 108 Strength, electrical (see Dielectric strength) Strength, mechanical bursting test, 3 ®ber, 155, 158

Index [Strength, mechanical] ®ber bonding, 228, 238±240 (see also Fiber, bonding) ®ber orientation effect on, 34 ®nes effect on, 191 Mullen tester, 3 on-line testing, 14, 16, 40±42 properties, 8 shear, 453±454, 461, 470 surface (see Surface strength) tear, humidity effect on, 5 tensile, delamination effect, 213±216 Strip-on-drum method, friction, 460 Structure, ®ber (see Fiber, cell wall structure) Structure, internal (see Porosity) Surface abrasivity (see Abrasivity) chemistry, 304, 330 conductivity, 335, 374, 379 energy (see Surface energy) ®ber (see Fiber, surface) friction (see Friction) imaging by atomic force microscopy (AFM), 217 pro®le measurements, 432±441 (see also Pro®lometry) pulp, 271±272 (see also Fiber, surface) resistivity, 335, 363±365, 375±380 roughness (see Surface roughness) softness, 507 speci®c (see Speci®c surface area, sheet) strength (see Surface strength) structure, 430 tension (see Surface tension) texture, 115, 138 topography (see Surface topography) Surface energy acid-base interactions, 305±308 friction, 456, 461, 468±469 sorption, 303 wetting, 303 work of adhesion, 305±306 Surface polarity test, liquid penetration, 321 IGT tester, 320±321 Prufbau tester, 320±321 Surface roughness analysis, 138, 165±166, 174, 180, 193±194, 197±199, 224, 238±242 (see also Pro®lometry) coating effect, 156 ®ber, 238±242

Index [Surface roughness] moisture effect on, 198 on-line measurement, 35±36 Surface strength IGT test, 320±321, 461 linting test, 41±42 Surface tension in capillary pressure method, 272 fountain solution, 310±313 liquid penetration affected by, 319 Marangoni effect, 313 mercury, 278 surfactants, 310±313 Surface topography (see also Pro®lometry) analysis, 165, 197, 219, 222±224, 227, 230±231, 237±238 Surfactants, 306±308, 310±313 Swelling ®ber, 150, 161, 167, 208, 213±216, 224, 288, 304, 313, 319 ®ber network, 319 pore size distribution, 319 Tactile properties, 505±530 anti-drape, 508 bending property, 513 clothing fabric, 506, 507±521 compression property, 513 crispness, 508 facial tissue paper, 525 fullness, 508, 510 hand chart, 520±521 hand evaluation, 506, 507±512, 520±522, 525 measuring system, 518 mechanical parameters, 513±518, 522 multi-dimensional scaling (MDS), 507 nonwovens, 521±525 objective method, 506, 507±518, 519, 522, 526±527 population transfer, 524±525 preference-difference model, 507 primary hand, 508±512, 519, 522±526 regression analysis, 512±513 shearing property 513 smoothness, 508, 510 softness, surface, 507, 527 stiffness, 508, 510 surface friction, 514±518, 522 surface property, 514±518 tensile property, 513 total hand (THV), 509±512, 521±525

545 Temperature, color, 106±108 Temperature, control, 5 Temperature effects (see Thermal properties) Testing (see Laboratory, testing and Online testing) Testing room control, 4±5, 7, 47±93 Thermal analysis, 417, 424 Thermal depolarization current technique, 352±355 Thermal properties, 389±427 anisotropic, 403±404 burning energy, 407 burning point, 405 capacity, 391 combustion, 404±407 conduction, 390±400 conductivity, 390, 393±400 decomposition (degradation), 408±424 depolymerization, 408±418 dimensional change, 400±404 expansion, 400±404 ®bers, 392±393, 404, 406, 424 ¯ame inhibitors, 420±421 ¯ammability, 404±406, 417±418 ¯ash point, 405, 406 free radical formation, 409±414 ignition temperature, 404, 406 linear expansion coef®cient, 401±404 measurement methods, 391±392, 394±398, 402±403, 405±407, 421±424 moisture effect on, 393±400, 403±404 oxygen index, 405, 407 pulp, 393, 399, 404, 407, 414, 417, 424 pyrolysis, 408±423 speci®c heat, 391±393 transition points, 404 volume expansion coef®cient, 403±404 wood, 393, 414, 417, 422 Thermogravimetry, 408, 417, 421, 424 Thickness, coating, 166, 195 Thickness, sheet on-line measurement, 32±33 surface roughness effect on, 288 Tint, 110 Topo SEM system, 440±441 Transformer bridge, 381 Transient current, electrical, 380 Transmission electron microscope (TEM), 160±163 Transmittance, optical, 96±101, 115±120, 127±136

546 [Transmittance, optical] cell wall, 135 coating stock, 129±133 contrast ratio, 116 intensity of light, 98, 117 Kubelka±Munk theory (see Kubelka± Munk theory) printing opacity, 116 transparency ratio, 116 Transparency, 116 Triboelectricity, 372 Ultrasonic propagation technique, 321, 324±325 Ultraviolet (UV) microscope, 151, 158±159, 163 Uptake, liquid (see Penetration, liquid) Vapor, water concentration in air, 49±52 degree of saturation, 51 density in air, 50 dew point tables, 69±78 effective pressure, 51 measurement, 52±60 mole fraction, 50 partial pressure, 51 pressure tables, 69 sensors, (see Sensors, humidity) Volume conductivity, 334, 338±346 Volume resistance, 335 Volume resistivity, 335, 343, 363, 375±379 V-t (Voltage-time) test, 383 Water vapor permeability, 282, 287±290, 294±297 closed cell method, 296 comparison method, 296±297 diffusion, solid±solution, 288 gravimetric measurement, 295

Index [Water vapor permeability] hygrometric measurement, 296±297 pressure gradient measurement, 289±290, 295 sweep gas method, 296±297 test conditions, 294 transmission rate, 294±295 Wavelength dominant, 105 spectra, 139 Wear by paper (see Abrasivity) Wear of paper, 482±488, 489±490 Web ¯atness (see Curl, sheet) Wet-bulb depression, 56 Wetting, 303±320, 329±330 capillograph, 314±315 contact angle, 304±309, 314±317, 329 ®ber (see Fiber, wetting) heat, 330 Laplace equation, 307 pressure, 305 in printing, 305, 309±310, 320±321, 327±329 repellency, 320±321 sizing agents, 304 surface free energy, 303, 305±308, 314 test methods, 314±321 theory, 305±313 times, 318 Young±Dupre equation, 306, 309 Wheatstone bridge, 381 Whiteness brighteners, 110±111 ¯uorescent, 110±111 formula, 110 Williams tester, 293 X-ray analysis, energy-dispersive (EDXA), 163 Yellowness, 110