Aerosol Science for Industrial Hygienists

Aerosol Science for Industrial Hygienists

Related Titles

Books

B R A D E N and KOLSTAD Measuring the Demand for Environmental Quality M A S U A D A and T A K A H A S H I Aerosols, Proceedings of the Third International Aerosol Conference WILLIAMS and L O Y A L K A Aerosol Science: Theory and Practice Z A N N E T T I et al. Air Pollution

Journals

Journal of Aerosol Science Aerosol Science and Technology Annals of Occupational Hygiene Applied Occupational and Environmental Hygiene Atmospheric Environment Environmental Health and Pollution Control Occupational Health and Industrial Medicine Full details of all Elsevier Science publications and free specimen copies of any Elsevier Science journals are available on request from your nearest Elsevier office.

Aerosol Science f o r I n d u s t r i a9l H y "g l e n l "s t s JAMES H. VINCENT

Professor, Division of Environmental and Occupational Health, School of Public Health, University of Minnesota, Minneapolis, U.S.A

@ Pergamon

U.K.

Elsevier Science Ltd, The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, U.K.

U.S.A.

Elsevier Science Inc., 660 White Plains Road, Tarrytown, New York 10591-5153, U.S.A.

JAPAN

Elsevier Science Japan, Tsunashima Building Annex, 3-20-12 Yushima, Bunkyo-ku, Tokyo 113, Japan Copyright 9 1995 Elsevier Science Limited

All Rights Reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means: electronic, electrostatic, magnetic tape, mechanical photocopying, recording or otherwise, without permission in writing from the publisher First edition 1995

Library of Congress Cataloging in Publication Data Aerosol science for industrial hygienists/James H. Vincent. p. cm. Includes index. 1. Aerosols-Toxicology. 2. Indoor air pollution. 3. Industrial hygiene. I. Title. RA577.5.V56 1995 615.9' 02-dc20 95--4232

British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library

ISBN 008 042029X

Printed and bound in Great Britain by Bookcraft, Bath

This book is dedicated, with love, to my mother Elvina Vincent

who celebrated her 87th birthday as I was putting the finishing touches to the work

This Page Intentionally Left Blank

Contents xiii

PREFACE Chapter 1.1 1.2 1.3 1.4

1. I N T R O D U C T I O N TO A E R O S O L S What is an aerosol? 'Good' versus 'bad' aerosols Workplace aerosols and occupational health Aerosols and gases

Chapter 2. T H E P R O P E R T I E S OF AIR AND GASES Introduction 2.1 Basic nature of gases 2.2 Pressure, volume and temperature Mean free path Diffusion Viscosity Gas mixtures Phase transitions: solids, liquids and gases Elementary fluid mechanics 2.3 Basic physics Streamlines Boundary layers Similarity Potential flow Stagnation Separation Turbulence

11 11 11 11 15 16 18 20 21 24 24 25 27 28 31 32 32 34

Chapter 3. P R O P E R T I E S OF A E R O S O L S 3.1 Aerosol generation in workplaces Mechanical generation of dry aerosols Mechanical generation of liquid droplet aerosols Formation by molecular processes

37 37 37 40 41

vii

Contents

3.2

3.3

3.4 3.5 3.6 3.7 3.8 3.9

The evolution of aerosols Coagulation, agglomeration and coalescence Condensation and evaporation Particle morphology Particle shape classification Fibres Aerosol concentration Particle size Elementary particle size statistics Electrical properties Mineralogical and chemical properties Biological properties

42 43 43 46 46 51 51 52 56 62 66 67

Chapter 4. THE M O T I O N OF A I R B O R N E PARTICLES 4.1 Introduction 4.2 Drag force on a particle Drag 4.3 Particle motion Equations of motion Motion under the influence of gravity Motion under electrical forces Motion in thermal gradients Motion without external forces 4.4 Similarity in particle motion 4.5 Particle aerodynamic diameter 4.6 Impaction 4.7 Elutriation Aspiration 4.8 4.9 Diffusion Molecular diffusion Coagulation Turbulent diffusion

72 72 72 72 77 77 77 83 85 88 89 91 95 99 103 107 107 111 113

Chapter 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 5.9 5.10

116 116 116 118 122 125 127 129 131 132 134

5. THE OPTICAL P R O P E R T I E S OF A E R O S O L S Introduction Physical basis Concept of extinction or transmittance Particle extinction coefficient Experimental measurements of extinction Light scattering Particle scattering coefficient Mass concentration aerosol photometry The visual appearance of aerosols Optical microscopy

viii

Contents

Chapter 6.1 6.2 6.3

6. THE I N H A L A T I O N OF A E R O S O L S Introduction The human respiratory tract Aerosol inhalation Basic definitions and terminology Concept of inhalability Experimental measurements of inhalability Physical basis of inhalability 6.4 Experiments to investigate aerosol deposition in the respiratory tract 6.5 Extrathoracic deposition Physical basis Experimental results 6.6 Thoracic deposition Physical basis Tracheobronchial deposition Alveolar deposition Thoracic deposition and penetration Total respiratory tract deposition 6.7 6.8 Deposition of fibrous aerosols Electrostatic respiratory tract deposition 6.9 6.10 Mathematical modelling of lung deposition

Chapter 7.1 7.2 7.3

7.4 7.5 7.6 7.7 7.8

7. THE FATE OF I N H A L E D PARTICLES Introduction Biological mechanisms of clearance and re-distribution Experimental methods The use of animals in inhalation research Experimental approaches Studies of clearance and build-up Experimental studies of dust accumulation in lung-associated lymph nodes The significance of 'overload' Kinetics of clearance Dosimetry

147 149 149 150 151 151 152 153 155 156 158 159 162

166 166 166 170 170 171 174 182 185 187 197

Chapter 8. 8.1 8.2

S T A N D A R D S FOR H E A L T H - R E L A T E D A E R O S O L M E A S U R E M E N T AND C O N T R O L Introduction Progress towards criteria for measurement of coarse aerosols

136 136 136 140 140 143 143 146

ix

204 204 207

Contents

8.3

8.4 8.5

Progress towards criteria for the measurement of finer aerosol fractions Thoracic aerosol Tracheobronchial and alveolar aerosol Harmonisation of criteria for aerosol standards Fibrous aerosols Standards The 'traditional' approach The new criteria and health effects Strategies for exposure assessment Limit values

Chapter 9. A E R O S O L SAMPLING IN W O R K P L A C E S Introduction 9.1 9.2 Sampling by aspiration Background Identification of aerosol sampler performance indices 9.3 Aspiration efficiency of thin-walled sampling probes in moving air Qualitative physical picture for a thin-walled probe facing the wind Simple impaction model Effects of orientation 9.4 Aspiration efficiency of blunt samplers Sampling from calm air 9.5 9.6 Physical factors which can complicate sampler performance Effects of freestream turbulence on aspiration efficiency Effects of external wall interactions on aspiration efficiency Effects of electrostatic interference on aspiration efficiency Transport losses of particles after aspiration Overall sampler performance 9.7 Sampling in stacks and ducts 9.8 Sampling for coarse aerosols in workplaces Static (or area) samplers for 'total' or inhalable aerosol Static (or area) samplers intended primarily for finer aerosol fractions Personal samplers for 'total' or inhalable aerosol Personal samplers intended primarily for finer aerosol fractions

211 211 212 214 223 225 225 226 227 230 238 238 238 238 239 241 241 243 247 251 256 261 261 262 263 264 267 268 269 269 274 275 280

Contents

9.9

9.10 9.11 9.12 9.13

9.14

9.15

Sampling for respirable aerosol in workplaces Static samplers for the respirable fraction Personal samplers for the respirable fraction Sampling for 'respirable' fibres Practical sampling for thoracic aerosol Sampling for more than one fraction simultaneously Aerosol spectrometers Sampling of bioaerosols Criteria for bioaerosol sampling Sampling Sampling system components Pumps Filters Quantitation of collected samples

282 282 284 287 289 289 291 295 296 296 297 297 298 299

Chapter 10. 10.1 10.2

10.3 10.4 10.5 10.6 10.7 10.8 10.9 Chapter 11.1 11.2 11.3 11.4 11.5 11.6 11.7 11.8 11.9 11.10

D I R E C T - R E A D I N G M O N I T O R I N G OF WORKPLACE AEROSOLS Introduction General characteristics of optical monitoring Extinction monitoring Light scattering photometry Light scattering photometers Optical particle counters Electrical particle measurement Condensation nuclei particle counters (CNC or CPC) Mechanical aerosol mass monitors Nuclear mass detectors Overview

304 304 305 306 307 307 313 320 323 325 327 328

11. C O N T R O L OF W O R K P L A C E A E R O S O L S Introduction Adjustments to industrial processes Behaviour of aerosols in the workplace atmosphere Extraction of workplace aerosols by exhaust systems Transport of aerosols in ventilation ducts Particle removal systems Gravitational separation Inertial separation Cyclone separation Wet scrubbers Spray towers Venturi scrubbers Energy considerations in wet scrubbers

332 332 333 334 336 339 346 347 350 353 356 356 359 361

xi

Contents

11.11 Filtration

11.12

11.13 11.14 11.15 Chapter 12.1 12.2 12.3 12.4 12.5 12.6 12.7

361 362 365 368 371 373 374 376

Macroscopic picture of a filter Single fibre collection efficiency Practical filters Electrostatic precipitation Corona discharge Particle charging Particle electrical migration velocity Simple model of precipitator collection efficiency Practical precipitator systems Comparison between air cleaning systems Containment of aerosols Personal respiratory protection

377 379 381 383 386

12. AEROSOLS AND VAPOURS Introduction The transport of gases Inhalation of gases The sampling of gases Direct-reading instruments for gases Air cleaning for the control of gases Industrial hygiene importance of aerosols and vapours

389 389 389 392 392 397 398 398

POSTSCRIPT

401

INDEX

403

, o

Xll

Preface The idea for this book began in December 1989 when I visited the Republic of China, Taiwan at the invitation of the Council of Labor Affairs to give a series of lectures on aerosol science to occupational health professionals in that country. That experience provided the stimulus for thinking about how the subject of aerosol science in its widest sense could be presented as a coherent and relevant body of knowledge. Further stimulus was added soon afterwards when I joined the School of Public Health at the University of Minnesota. It has been said so many times as to be almost r e d u n d a n t - but it is worth repeating anyway ~ that there is nothing so powerful as teaching itself in helping the teacher better understand his or her own subject. Airborne contaminants in workplaces are the sources of a high proportion of occupational illness. Such airborne contaminants may appear either as gases or aerosols. Aerosols in workplace atmospheres have been ~ and continue to be ~ a major focus of industrial hygiene. So aerosol science remains an important component in the graduate-level education of industrial hygienists and other occupational health professionals. Several excellent texts already exist which lay out the elements of aerosol science in a way which is accessible to such students. I especially admire the work of Professor William C. Hinds, 'Aerosol Technology: Properties, Behavior, and Measurement of Airborne Particles' published by John Wiley and Sons in 1982, and continue to recommend it as a reference text to my own students. My new book sets out to be complementary to this and other such texts. In particular, worker exposure is taken as the central concept, around which most of the subject matter revolves. This has resulted in a structured approach which draws together wide-ranging aspects of aerosol science within the occupational health framework. This text deals specifically with aerosols, and provides a broad introductory overview of modern aerosol science as it relates to industrial hygiene and to occupational health in general. It is intended to provide the basis of graduate-level coursework which deals with the properties, behaviour and effects of airborne contaminants. The introductory chapters are concerned with the nature and properties of aerosols, and how they are generated in the occupational environment. The book then goes on to provide a description

xiii

Preface of the fundamental mechanical properties of aerosols, in particular those mechanical properties associated with the motion of airborne particles (since it is these properties which govern how particles are transported in the workplace air, are inhaled and are deposited in the respiratory tract, in addition to how exposure to them might be effectively measured and controlled). The next chapter is devoted to the optical properties of workplace aerosols since these are important in the visual appearance of aerosols and in many aspects of measurement. This leads to the core of the book which deals with the processes which govern the nature of exposure to and the subsequent fate and effects of airborne particles, building up to a rational framework for standards, measurement and control. In this core are chapters describing: (a) the inhalation of airborne particles and how they can reach the interface between the workplace environment and the biological system of interest (i.e., the worker); (b) the clearance, re-distribution and storage of particulate material in the respiratory tract and elsewhere in the body; (c) the development of scientifically-based standards; (d) the development of health-related sampling and measurement; and (e) the technical means by which workplace aerosols (and particularly worker exposures) are controlled to within demonstrably safe levels. Included here also is a chapter which deals with the direct-reading instrumentation for aerosols which provides an increasingly important tool in the technical armory of the modern industrial hygienist. Finally, a chapter is added which relates what has been said about aerosols to gaseous and vapour contaminants. Here, there are of course important differences. However, there are also some similarities, and I have felt that a discussion of these is useful in helping to dispel the notion that aerosols (and so aerosol scientists) and contaminant gases (and so chemists) are totally disparate and 'never the twain shall meet'. The overriding goal is to provide a means for u n i f y i n g - not d i v i d i n g - the science of industrial hygiene. As already stated, the book is targetted at graduate students in industrial hygiene and other occupational (and environmental) health disciplines. As such it differs from and is broader than ~ my previous book, 'Aerosol Sampling: Science and Practice' published by John Wiley and Sons in 1989, which was aimed primarily at researchers. This new text recognises that graduates entering any of these fields come from a very wide range of starting disciplines, but acknowledges that all such graduates should satisfy the prerequisites of a broad undergraduate education in the physical and life sciences and mathematics (including calculus). The approach is therefore intentionally somewhat rudimentary. However, it is hoped that sufficient pointers are g i v e n - and references c i t e d - to enable more advanced students (and, indeed, experienced practitioners and researchers) to move into more challenging realms. As such, therefore, it is my intention that the book should also be seen as a framework for more advanced study and reflection. The book derives from research, experience and time spent thinking about

xiv

Preface the subject over many years. It therefore goes without saying that I owe a debt of gratitude to a very large number of people. So, firstly, thanks go to all my colleagues past and present. For the most part, they are too large in number to list individually. However I do wish to mention specifically the contributions of David Mark for the countless hours of discussion which provided the early stimulus for much of what is in this book, and of Sidney Soderholm for his careful reading and critique of the chapter on aerosol standards. I also wish to thank my many students who, inevitably, have acted as 'guinea pigs' for the ideas and approach which are contained in these pages. I especially wish to thank Perng-Jy Tsai for taking the time to check the worked examples. Finally, my everloving thanks go to my wife, Christine, for her patience and support during the period of writing, not least for her cheerful acceptance of the fact that I failed to keep my promise that my previous book was to have been "my last" . . . J A M E S H. V I N C E N T Minnetonka, Minnesota December 1994

XV

A NOTE ON UNITS AND DIMENSIONS Industrial hygiene is a field which, by virtue of the wide range of disciplines from which it draws, is plagued by confusing combinations of units. For example, it is the most c o m m o n convention to talk of concentration of an airborne contaminant in terms of [mg m-3], whilst at the same time be thinking of linear dimensions in terms of [feet], of air velocities in terms of [feet per minute, or fpm], or air flowrates in terms of [cubic feet per minute, or cfm]. Meanwhile, in aerosol science in general, the old ~ and somewhat unwieldy ~ 'centimetre-gram-second' (or 'cgs') system of units is still commonly used, especially in the United States. In an attempt to simplify the system of units used in the physical sciences and engineering, the 9th General Conference of the Weights and Measures in 1948 proposed the new 'Syst6me Internationale d'Unit6s' (or 'SI') system. This is the chosen approach in this book. In the SI system, the following so-called 'base' units are relevant here. These are :the metre [m] for length the kilogram [kg] for mass the second [s] for time the ampere [A] for electric current the degree Kelvin [~

for t e m p e r a t u r e

the mole for amount of a given substance. In addition, there is a range of primary units which are derived from the base units. Relevant ones are:the litre [1] for volume [ - m 3 • 10 -3 ] the micrometer [l~m] for size [ -- m x 10 -6

]

the Pascal [Pa] for pressure [ - kg m -1 s -2 ] the Newton [N] for force [ - kg m s -2 ] the Joule [J] for energy [ - kg m 2 m s -2 ] the watt [W] for power [ - kg m 2 s -3 ] the Coulomb [C] for electrical charge [ - A s -1 ] the volt [V] for electrical potential difference [ - J C -1 ,

C-1 C1].

xvi

or kg m 2 s -2

A note on units a n d d i m e n s i o n s

Finally there are some physical quantities which appear frequently in the book and whose dimensions are as follows" viscosity [ N s m -2, or kg m -1 s -1 diffusivity [ m 2 s -1

]

]

density [ kg m -3 ]. Working within the SI system of units is usually very straightforward, and so its use is strongly recommended. While most of the equations given in this book apply directly in other systems of units, the reader who wishes to do so is advised always to check the dimensional consistency of the relationship used. This is particularly important when constants are involved which themselves have dimensions.

xvii

This Page Intentionally Left Blank

CHAPTER 1

Introduction to aerosols 1.1 W H A T IS AN A E R O S O L ? 'Aerosol' is a scientific term which applies to any disperse system of liquid or solid particles suspended in a gas usually air. It therefore is much wider than suggested by the narrow popular use of the term in relation to pressurised spray-can products. The term 'aerosol' applies to a very wide range of particulate systems encountered terrestially. First there are naturally-occurring aerosols. These include, for example, snowstorms, duststorms, sandstorms, volcanic ashes, smokes (from fires), mists, clouds, and so on. They also include biologicallyactive systems such as airborne pollens, fungi, viruses and bacteria. In terms of overall global amount (e.g., mass or particle number), such naturallyoccurring particulate clouds are thought to represent the majority of overall terrestial aerosol. However, we also have man-made aerosols, resulting from a wide range of anthropogenic ~ including industrial activity. Such aerosols may be found either outdoors or indoors, and become important in relation to the natural aerosol background because they are usually quite inhomogeneous (with relatively high local concentrations). Many aerosols, either natural or man-made, may contain toxic substances which can lead to adverse effects when they come into contact with plant and animal life. Such aerosols are regarded as pollutants, hence the need is clearly identified to: understand their physical, chemical and biological nature and behaviour; understand their interactions with ecological and biological systems; and develop means for monitoring and controlling their presence. It is clear from these introductory remarks that aerosols feature very widely in nature and in human experience. So it comes as no surprise to find that the study of aerosols ~ their characteristics, effects and applications ~ has grown into a major field of scientific enquiry, touching on many individual disciplines, including physics, chemistry, mathematics, engineering, biology and medicine.

Aerosol science for industrial hygienists Many, indeed most, industrial processes produce aerosols, releasing airborne particles into the air so that, in the absence of control measures, they may be discharged either into the outdoor environment or into the workplace atmosphere. The former contributes to the pollution of the ambient atmosphere and so impacts on populations and the environment at large. This is 'air pollution' in its widest sense. In the primary context of this book, contamination of a workplace atmosphere impacts primarily on the working population in question. Here the local concentrations of airborne particles tend to be higher than in the ambient atmosphere, leading in some cases to high prevalences of aerosol-implicated health effects. Because specific aerosol types and properties tend to be associated with specific industries, then so too do those occupational health effects. For example, it is well-known that pneumoconiosis, a disease of the alveolar region of the lung, has been associated with the mining and extraction industries in which workers are exposed to relatively insoluble mineral dusts; asbestosis, lung cancer and mesothelioma have been associated with exposure to asbestos fibres in industries where asbestos is mined or processed; nasal cancer has been associated with some hard wood dusts found in the furniture industries; and so on. It is fair to surmise that a significant proportion of the working population is exposed occupationally to aerosols of one sort or another. Workplace aerosols therefore require the attention of all those concerned with quality of the workplace environment, none more so than the industrial hygienist. Indeed, over the years, aerosols of widely-varying types in a highly diverse range of industrial settings have been the focus of a major part of industrial hygiene research and practice. Most such interest in aerosols stems from the ability of particles to be inhaled into the respiratory tract. To a lesser extent, there is interest in toxic particles which can be deposited onto and absorbed through the skin (e.g., from pesticide sprays). There is also some interest in very large, gritty dust particles from certain industrial processes which are not actually suspended but, rather, are projected into the air and can cause discomfort or damage upon impaction onto the skin or (particularly) the eyes. A summary classification of a range of typical aerosols is given in Figure 1.1. It contains not only examples of the workplace aerosols with which this book is primarily concerned but also, for the sake of comparison, some naturally-occurring and man-made aerosols found in the outdoor atmospheric environment. At the top of this figure, a number of types of aerosol are shown which relate primarily to the material from which the aerosol is formed and manner of its formation. These are:

Dust, an aerosol consisting of solid particles made airborne by the mechanical disintegration of bulk solid material (e.g., during cutting, crushing, grinding, abrasion, transportation, etc.), with sizes ranging

Introduction to aerosols from as low as sub-micrometer (~m) to over 100 ~xm. In principle, there is no upper limit in this category. But whether or not 'particles' of 'rock-like' dimensions (which are certainly produced during many of the industrial processes mentioned) can be considered to be part of an aerosol depends on the extent to which they are truly 'airborne'.

Spray, an aerosol of relatively large liquid droplets produced by the mechanical disruption of bulk liquid material, with sizes upwards of a few micrometres. Here the upper limit is more clearly defined than for dusts because, above a certain size, droplets become unstable (due to the relationship between surface tension and gravitational and shearing forces) and so break up. Thus, for example, an 'aerosol' consisting of large, centimetre-sized, water droplets is unheard of. Mist, an aerosol of finer liquid droplets produced during condensation or atomisation, with sizes up to a few micrometres. Fume, an aerosol consisting of small solid particles produced by the condensation of vapours or gaseous combustion products. Usually, such particles are aggregates made up from large numbers of very small primary particles, with the individual units having dimensions of the order of a few nanometres (nm) and upwards. These primary particles are difficult to discern under the optical microscope. Aggregate sizes are usually less than 1 p~m. Smoke, an aerosol of solid or liquid particles usually resulting from incomplete combustion, again usually in the form of aggregated very small primary particles. The aggregates themselves have extremely complex shapes, frequently in the forms of networks or chains, having overall dimensions that are usually less than 1 p~m. Bioaerosol, an aerosol of solid or liquid particles consisting of, or containing, biologically-viable organisms (viruses, bacteria, allergens, fungi, etc.), with sizes ranging from sub-micrometre to greater than 100 ~m. In the main body of Figure 1.1, the examples of aerosols given are classified according to their ranges of particle size, a property which is highly influential in nearly all aspects of aerosol behaviour. The bottom part of the figure includes an indication of the particle size ranges of some health-relevant fractions ~ inhalable, thoracic, tracheobronchial and alveolar. These refer to particles which, by virtue of their size, can be inhaled through the nose and/or mouth of an exposed subject during breathing and can subsequently penetrate to and be deposited in progressively deeper parts of the human respiratory tract. A more detailed account of these fractions will be given in Chapters 6 and 8.

Aerosol science for industrial hygienists 0.01

Liquid

Physical definitions

0.1

1.0

10

'

I

I ~---. I Dust I Fly ash

~ M i s t

I I Fume ~ ~ I I -q---Oil smokes--~ ~1

Solid

Tobacco

~ I smoke _. -.. I I I

I00

~-- Cement dust--~ n dust I Coal a I

]

, -----~

Atmospheric dust I

Typical

aerosols and aerosol particles

Viruses

I l

-~ I Airborne asbestos ~ (diameter) I I

~ _/ I .4--Bacteria-~ | -I ~

Pollens I I

I Airborne' asbestos --~ (length)

I Respirable _ I particles Tracheobronchial I I I particles U ~ q:horacic particles i ~ I , I I inhalable particles I I I I I I I I j I I

Health-related deft n it ion s/fraction s

0.01

O. I

I000

' Spray i 1 I I

! .0

I0

I I I I I I I ~ I I I I O0

1000

of aerosols (from Vincent, J.H., 1989, a d a p t e d b y p e r m i s s i o n Wiley and Sons Limited).

Aerosol

Particle d i a m e t e r (l~m) Figure

1.1.

Summary

classification

Sampling." Science and Practice, C o p y r i g h t

of John

1.2 ' G O O D ' VERSUS 'BAD' A E R O S O L S 'Aerosol science' in general covers all aspects of airborne particulate matter. Many aerosols are considered to be 'good' insofar as they make positive contributions to the earth and to man. For example, a balanced climate requires cloud and droplet formation in the atmosphere. Many industrial processes require aerosols as components vital to their effectiveness (e.g., photo-reproduction, chemical reactors, materials synthesis, etc.). Many therapeutic substances need to be aerosolised before they may be delivered to the site in the body where they can be effective (e.g., as in inhalers for the treatment of respiratory complaints such as asthma, etc.). Elsewhere, industry usefully employs sprays for the efficient delivery of pesticides and paints, etc. to their desired sites. There is clearly plenty of scope for aerosol science to continue to make direct 'positive' contributions to the earth's environment and ecology and to man's health and industrial activities. Modern aerosol research is therefore increasingly motivated in these directions.

Introduction to aerosols However, there is just as wide a range of aerosols which are clearly perceived as 'bad'. An unwanted aerosol is, by definition, a 'pollutant' and particles once they are airborne are capable of rapid dispersal and may be difficult to recapture. One aspect which is growing in importance in terms of its potential cost to industry is the problem of the contamination of products (e.g., electronics components, pharmaceuticals, food, etc.) by unwanted particle deposition. A whole branch of aerosol science has grown out of this need for 'clean technology'. However, the area where interest has consistently remained high since the birth of aerosol science is that concerned with the ill-health that can arise from human exposure to aerosols by inhalation, whether it be in the home, in the outdoor environment, in public buildings, or in the workplace. Such exposure needs to be eliminated or, at least, kept to within limits considered to be 'safe'. Historically, workplace aerosols have long been a major source of concern. It therefore comes as no surprise to find that much of the pioneering work carried out in aerosol science during the latter half of the twentieth century has been driven by the needs of industrial hygiene.

1.3 W O R K P L A C E A E R O S O L S A N D O C C U P A T I O N A L H E A L T H Depending on the nature of a particular industry and its working materials, workplace aerosols can occur in almost any of the categories listed in Figure 1.1, sometimes of one type only but often in the form of complex aerosol combinations. It is therefore part of an industrial hygienist's job to: recognise the nature and magnitudes of the hazards associated with particular types of workplace aerosol; assess the levels of worker exposure to aerosols of the types suspected of being associated with ill-health; and observe standards and, where appropriate, to initiate appropriate control measures. Therefore, for an enquiring industrial hygienist, a grasp of the relevant fundamentals of aerosol science is essential. This book sets out to provide an appropriate framework. The scope of that framework, and hence of this book, is summarised in Figure 1.2. Here the primary component is identified as worker exposure to aerosols (E) and the way in which this can lead to ill-health. The central 'tree' of Figure 1.2 therefore links the processes of exposure, dose, response and effect. In the first instance, exposure leads to the arrival and accumulation of particulate material in the body, leading to a cumulative 'dose' at vulnerable

Aerosol science for industrial hygienists

Figure 1.2. Framework for an integrated approach to the study of aerosols in the industrial hygiene context.

sites of some quantity representing 'harmfulness'. This leads in turn to a biological response which, depending on the body's recovery-and-repair defence mechanisms, might ultimately be the precursor to substance-related ill-health in a form that can be identified clinically or pathologically. Statistical analyses of the relationship between exposure and the incidence of ill-health have sometimes been used as the basis for the setting of health-based exposure limit values ~ that is, quantitative guidelines for controlling the quality of the workplace environment. This is largely a 'black box' approach and, in principle, the process may be improved by incorporating into the modelling process the kinetics and dynamics of the various components of the disease processes describing the fate and biological effects of the inhaled particulate material. With the improved knowledge that we now have from inhalation toxicological and other research, this may now be feasible for some hazardous substances. In addition, however, working exposure limit values need to incorporate pragmatic considerations of what is achievable under practical

Introduction to aerosols industrial conditions. Inevitably, health-based limit values are usually more stringent than actual, pragmatic ones. It is, of course, an ultimate goal to close the gap between them. The right 'arm' of Figure 1.2 concerns the measurement of aerosol exposure. Here, ideally, a criterion for measurement needs to be established so that the index of exposure which is provided is valid in relation to the disease or ill-health in question. Only then can appropriate instruments be designed, built and made available to industrial hygienists. In turn, these instruments may be used in measurement strategies designed to provide reliable exposure histories of exposed workers. Such data can then be used as the basis of epidemiological research to determine the true relationship between aerosol exposure and ill-health (where that relationship is not yet known or known sufficiently well), or for evaluating risk. More immediately, however, the results can be matched against existing exposure standards so that an assessment can be made of the control measures that need to be carried out in order to achieve compliance. The left 'arm' of Figure 1.2 closes the loop by introducing the engineering control of the workplace environment that is carried out to ensure that worker exposure is kept to within safe limits. Appreciation of the nature and behaviour of aerosols is involved at all stages of the integrated industrial hygiene philosophy embodied in Figure 1.2. These aspects are summarised in Figures 1 . 3 - 1.6, underlining the fact that, in industrial hygiene, the term 'aerosol science' refers not just to aerosol physics but also to engineering, chemistry and biology and beyond. In turn we see that industrial hygiene is not only highly multi- and inter-disciplinary but also ~ arguably ~ central to the study and practice of occupational health.

Figure 1.3. Summaryof the aerosol science considerations that occur in relation to aerosol generation and exposure in the workplace.

Aerosol science for industrial hygienists

Figure 1.4. Summary of the aerosol science considerations that occur in relation to linking exposure, dose and response - - leading to the setting of health-based standards for workplaces.

Figure 1.5.

Summary of the aerosol science considerations that occur in relation to the measurement of workplace aerosol exposures.

Introduction to aerosols

Figure 1.6.

Summary of the aerosol science considerations that occur in relation to controlling workplace aerosol exposures.

Later chapters in this book will address many of the scientific areas touched on in these figures. Finally, returning to Figure 1.2, it is relevant to note that the scope of this description can be widened to cover almost every aspect of occupational and environmental health, not just in relation to aerosols but also to gaseous contaminants and other agents or factors which may cause illness or injury.

1.4 A E R O S O L S A N D GASES In considering airborne contaminants in workplaces, the industrial hygienist is concerned not only with aerosols but also with gases and vapours. Although the latter lie outside the primary scope of this book, it is however important to recognise the similarities and differences between them. Gaseous and vapour contaminants are similar to aerosols in the sense that they too consist of disperse systems of unconnected entities. That is, both classes of contaminant are contained in and transported by the air and are therefore

Aerosol science f o r industrial hygienists

readily available for worker exposure by inhalation. The behaviour of gas molecules is governed by the gas and fluid laws, so they can exhibit motions in the form of convection and diffusion. In turn they may be inhaled and deposited selectively in the respiratory tract. In addition, depending on the environmental (physical and chemical) conditions, they may undergo chemical changes, or become excited, ionised or dissociated. However, unlike relatively macroscopic aerosol particles, they do not experience gravitational, inertial or electrical forces (unless ionised), nor do they undergo other mechanical processes of coagulation, agglomeration, break-up, etc. In general, relative to aerosol particles, individual molecules are better-defined by virtue of their unique molecular structures. So, from the industrial hygiene perspective, they are usually simpler to recognise and evaluate (but not necessarily control).

REFERENCES The reporting of the science and applications of aerosols is widely distributed throughout the peer-reviewed literature. In the first instance, there are the premier aerosol science journals such as: Journal of Aerosol Science m

Aerosol Science and Technology Journal of Aerosols in Medicine.

Other journals containing significant amounts of aerosol-related material directly relevant to occupational health include:m

Annals of Occupational Hygiene The American Industrial Hygiene Association Journal Applied Occupational and Environmental Hygiene Scandinavian Journal of Work Environment and Health Journal of Occupational Hygiene Atmospheric Environment

as well as a wide range of medical, environmental and engineering journals.

10

CHAPTER 2

The properties of air and gases 2.1 I N T R O D U C T I O N This book is concerned, as its title suggests, with aerosols. However the discussion of particulate material ~ liquid or solid ~ suspended in air cannot take place without understanding the nature of the air itself. It is therefore appropriate to provide a rudimentary outline of the main physical properties of air as they relate to aerosol science and industrial hygiene. The discussion is widened further to include brief description of the properties of airborne contaminant gases and vapours as a basis for recognising the similarities and differences between aerosol and gasesous contaminants (see also Chapter 12).

2.2 BASIC N A T U R E OF GASES Much of industrial hygiene is concerned with the transport of pollutants of various kinds in the vicinity of human subjects, both through and by the workplace atmospheric air. Air is a mixture of gases, the main constituents being nitrogen (about 78% by volume) and oxygen (21%), with a variety of other trace gases including argon, carbon dioxide, water vapour, etc. (amounting to about 1% in total). It is a colourless, odourless gas of density 1.29 kg m -3 at a standard temperature and pressure (STP) of 20~ (293~ and pressure of 1.01 • 105 Pa, respectively. The basic physics of matter, in particular of gases and about the phase changes that can take place between gases and liquids (and vice-versa), are covered very widely in the many excellent physics texts that are available.

Pressure, volume and temperature The dynamical behaviour of a gas may be based on what we refer to as the 'kinetic theory of gases'. This treats a gas as a 'billiard ball' system; that is, one

11

Aerosol science for industrial hygienists

Figure 2.1.

Schematic diagram to indicate the basis for developing the kinetic theory of gases.

where the molecules are small rigid spheres which undergo perfect collisions with one another in which no energy is lost. Starting from the molecular concentration n, mass rn, effective diameter d m and random velocity u between collisions, kinetic theory provides a basis for relating the properties of temperature, pressure and mean free path (the mean distance between inter-molecular collisions), and the phenomena of diffusivity, viscosity and thermal conductivity. Consider a box of volume V containing N molecules, so that the number concentration, n = N / V (see Figure 2.1). As molecules near the walls collide with the walls, they exchange momentum and hence a force is experienced. For an analogy, consider a room full of people throwing tennis balls at one of the walls. The impact of a single ball is experienced by the wall as a single impulse, and is associated with the momentum which is exchanged during impact. A succession of such impulses becomes equivalent to a series of such individual forces. When the number of impulses becomes very large, then the series of impulsive forces becomes indistinguishable from a continuous steady force acting on the wall. Likewise for the molecules in the box, the net effect of many molecular collisions with the wall, averaged over time, appears as a steady force which in turn may be represented as the pressure P (i.e., force per unit area). From the mechanics of the 'billiard ball' collisions and the

12

The properties of air and gases

application of some statistics, it may be shown that the pressure on the walls of the box may be expressed as mNums (2.1)

Force/Area - P = 3V

leading to mNums

P V =

(2.2)

In this expression, Urns is the mean-square (ms) random velocity of the gas molecules ~ in other words, the arithmetic mean of the squares of the randomlydirected molecular velocities. This comes from the Maxwell-Boltzmann statistical distribution of the magnitudes of the molecular velocities (u) arising from the inter-molecular collisions. This distribution is given by

f ( u ) -- 4rru 2

exp

(2.3)

2zrR T

2R T

where M is the molecular weight (29 g mole -1 for air). From Equation (2.3) we have oo Urms

"-- { fa bl2 f ( u )

1/2 du

}1/2 --

(2.4)

M

0

where Urms is the root-mean-square (rms) molecular velocity. A typical distribution, describing the probability of finding molecules moving at the velocities indicated in air at a temperature of 20~ is illustrated in Figure 2.2. It shows, as described by Equation (2.3), that molecules may experience a very wide range of random velocities, in principle from zero to greater than 1000 m s -1. Figure 2.2 shows not only Urms for the distribution in question but also the related ~ but mathematically different ~ mean random velocity (urn) given by oo blm --

fur(u)du

8R T ~ 1/2 =(

~rM )

(2.5)

0 By comparing Equations (2.4) and (2.5) we can see that Urms

13

>

Um.

Aerosol science for industrial hygienists Urns (at STP)

am

"X

f (u)

,,4 500

0

I000 u

I

1500

(m/s)

Figure 2.2. Statistical distribution of random velocities of air molecules under standard atmospheric conditions, indicating the mean and root-mean-square molecular velocity. For the gas contained inside the box shown in Figure 2.1, the ideal gas law relates P, V, N and the temperature T by means of the familiar expression

PV-

nmRT

(2.6)

where

mN nm =

(2.7) M

is the number of moles of air in the box. Here R is the universal gas constant, and is usually given in SI units as 8.314 J ~ mole -1 (although it is sometimes useful to give it as 82 atmosphere cm 3 oK-1 m o l e - l ) . Also in Equation (2.6) n m is the number of moles of gas present, and T is the absolute temperature expressed in degrees Kelvin (~176 where ~ refers to temperature in degrees Centigrade). When temperature is held constant, then this reduces to the simple form

PV= constant

(2.8)

known as Boyle's law. From Equations (2.4) and (2.5), it is clear that, for a given gas, t e m p e r a t u r e and rms molecular velocity are physically equivalent. Thus, for example, we have

umsM T =

(2.9) 3R 14

The properties of air and gases Example 2.1. W h a t are the r o o t - m e a n - s q u a r e and m e a n velocities, respectively, for air at 35~ at atmospheric pressure T (in ~

= 273 + 35 = 308~

Note that M = 29 g mole -1 = 0.029 kg mole -1 From Equation (2.4) 3 x 8.314[J ~ Um S

mole -1] x 308[~

-"

0.029[kg mole -1] so that Urns ~ 26.49 X 104 m 2 s-2 and hence *

Urms = 515 m s -1

Similarly, from Equation (2.5) we get *

Mean

u m = 474 m s -I

free path

T h e mean free path (mfp) is t h e m e a n d i s t a n c e t h a t a gas m o l e c u l e t r a v e l s b e t w e e n c o l l i s i o n s w i t h o t h e r gas m o l e c u l e s . A g a i n f r o m t h e ' m o l e c u l e s - i n the-box' m o d e l and the application of s o m e m o r e statistics, we get V

mfp

1

-

= ~/~N

1r d 2

(2.10) ~/~n

~r d2m

T h i s s h o w s t h a t , f o r a g i v e n gas, mfp d e p e n d s i n v e r s e l y o n gas d e n s i t y In t u r n , f r o m E q u a t i o n ( 2 . 1 ) , this m e a n s t h a t mfp is also i n v e r s e l y p r o p o r t i o n a l to gas p r e s s u r e . A d d i t i o n a l l y , f o r c o n s t a n t v o l u m e it is d i r e c t l y p r o p o r t i o n a l to t e m p e r a t u r e .

15

Aerosol science for industrial hygienists

Example 2.2. What is the mean free path for air molecules at STP? What does its value become at the top of Mount Everest (where the air pressure falls to about 50% of that at sea level)? Note that d m = 0.00037 Ixm = 37 X 10-11 m n = 25 x 1024 molecules m -3 at STP

From Equation (2.10) *

mfp = 0.066 ~m.

At the top of Mount Everest, Equation (2.1) gives n = 12.5 x 1024 molecules m -3 so that *

mfp = O.13 jxm

As we shall discuss in a later c h a p t e r , the m a g n i t u d e of m f p b e c o m e s very i m p o r t a n t in relation to the a e r o d y n a m i c b e h a v i o u r of particles in s o m e w o r k p l a c e aerosols w h e r e t h e r e are high p r o p o r t i o n s of particles w h o s e sizes are of the s a m e o r d e r of m a g n i t u d e as m f p (or less).

Diffusion D i f f u s i o n relates to the mass transfer associated with the r a n d o m m o l e c u l a r m o t i o n s . F o r a gas species which is u n i f o r m l y d i s p e r s e d in air (for e x a m p l e , a c o n t a m i n a n t trace gas well-mixed in w o r k p l a c e air), t h e r e is no net t r a n s f e r of gas due to these m o t i o n s , m e r e l y the m u t u a l e x c h a n g e of individual m o l e c u l e s each way across any defined b o u n d a r y . If, h o w e v e r , t h e r e is a g r e a t e r c o n c e n t r a t i o n of the trace m o l e c u l e s on o n e side, t h e n t h e r e is an i m b a l a n c e in the e x c h a n g e process, resulting in a net flux of m o l e c u l e s f r o m regions of high to low c o n c e n t r a t i o n . This scenario is illustrated in Figure 2.3. Kinetic t h e o r y p r o v i d e s the e x p r e s s i o n ( k n o w n as Fick's law) On

N e t flux = J = - D m

(2.11) Ox

16

The properties of air and gases

Figure 2.3.

Schematic diagram to illustrate the phenomenon of molecular diffusion.

w h e r e On/Ox is the local c o n c e n t r a t i o n g r a d i e n t which is driving the flux a n d w h e r e the m i n u s sign indicates that the net flux is t o w a r d s the r e g i o n of l o w e r c o n c e n t r a t i o n . D m is the m o l e c u l a r diffusion coefficient given by

mfp Dm

=

u m

(2.12)

which d e p e n d s on b o t h t e m p e r a t u r e a n d p r e s s u r e . In fact, we see f r o m the p r e c e d i n g section that D m increases as the gas t e m p e r a t u r e i n c r e a s e s b u t d e c r e a s e s as p r e s s u r e increases. Example 2.3. What is the diffusion coefficient for air molecules at STP? For STP, Equation (2.4) gives um

"

-

470 m s-1

Also, note that

mfp = 0.066 p~m = 6.6 • 10 -8 m Equation (2.12) gives *

D m ~- 1 0 - 5

m 2 s-1

17

Aerosol science for industrial hygienists As we shall see later, the diffusivity of gas molecules u n d e r n o r m a l atmospheric conditions is many orders of m a g n i t u d e greater than for even very small aerosol particles.

Viscosity Viscosity is associated with a moving gas, in particular with the forces acting between regions moving at different velocities. This physical scenario, k n o w n as 'shear', is illustrated in Figure 2.4 which shows a pair of plates of area A i m m e r s e d in the gas, separated by a distance y, and moving at velocity U relative to one another. The law of m o t i o n for a real fluid requires that the velocity of the gas at the surface of each plate must be zero, hence there is 'shear' associated with their relative motions. The internal m o t i o n of the molecules in the gas between the plates is such that there is a net force resisting this shear (and hence this relative motion). This is r e g a r d e d as a 'friction' force, and kinetic theory again comes into play to give ~A U Friction force, F v =

(2.13)

where the coefficient o f viscosity is given by

mfp Ix - n m u m

Figure 2.4.

(2.14)

Illustration of the phenomenon of viscosity and how it relates to the kinetic theory of gases.

18

The properties o f air and gases Here we see by application of Equations (2.5) and (2.10) that the coefficient o f v i s c o s i t y is i n d e p e n d e n t o f p r e s s u r e , a n d - - f o r a g i v e n g a s - - is d e p e n d e n t only on temperature.

Example 2.4.

W h a t is the viscosity of air at S T P ?

Note again that n = 25 x 1024 molecules m -3 at STP mfp = 6.6 x 10 -s m u m = 470 m s-1 Note also that the molecular mass is given by

m = M/Na where N a is the number of molecules per mole (referred to as Avagadro's number) and is 6 x 1023 molecules per mole. Thus 0.029 [kg mole -1] m

-

6 x

10 23

[molecules mole- l]

= 4.8 x 10-26kgmolecule -l Equation (2.14) now gives Ix = 25 x 1024[molecules m -3] x 4.8

x 10-26[kg

molecule -l]

x 470 [m s-'] x 6.6 x 10 -s [m]

*

Ix = 12.4 x 10 -6 kg m -1 s -1 (or N

s m -2)

N o t e , h o w e v e r , t h a t this r e s u l t h a s b e e n o b t a i n e d f r o m a c a l c u l a t i o n b a s e d on kinetic theory. In actual fact, thisas f o r t h e o t h e r q u a n t i t i e s c a l c u l a t e d above is o n l y a n e s t i m a t e s i n c e t h e i n i t i a l i d e a l i s i n g a s s u m p t i o n s d o n o t h o l d strictly. It is i m p o r t a n t t o n o t e t h a t , in c a l c u l a t i o n s w h i c h f o l l o w l a t e r in this b o o k , t h e t r u e v i s c o s i t y o f air at S T P is 1 8 . 3 2 6 x 10 - 6 N s m - 2 ( w h e r e 18 • 10 - 6 N s m - 2 is u s u a l l y a g o o d e n o u g h a p p r o x i m a t i o n ) .

19

Aerosol science for industrial hygienists Gas mixtures

The gas laws can be applied to systems of mixed gases. Extending Equation (2.6), Dalton's law gives

PV + P1V1 + P2V2 + P3V3 + ' ' " --(nml

+ nm2 + r i m 3 ' '

(2.15)

")RT

where the nm'S are the numbers of moles present for each gas component in the overall volume V, and P~, V2, etc. are their partial pressures and volumes, respectively. For fixed pressure (e.g., for contaminant gases in workplace air), this becomes V

--

Vl

-+- V 2

-.~- V 3

-Jr- 9 9 9

(2.16)

Here, therefore, the total volume is equal to the sum of the partial volumes of the individual gas components, where it is assumed that the molecules of each component have been brought together so that the pressure of that component is equal to the overall pressure P. For a single gaseous contaminant in air (where air for this purpose is treated as a single gas), this gives V-

Vai r -F Vcontaminan t

(2.17)

In turn this can be written as the concentration of the contaminant gas. Expressed in terms of relative volume, this is Vcontaminant c = (Vai r -+- Vcontaminant) Vcontaminant

(2.18) for Vcontaminan t <~ Vai r

Vair

In the usual industrial hygiene parlance, c is expressed in units of parts per million (ppm), such that Vcontaminant

c [ppm] -

x 106

(2.19)

V ir For many contaminants, including gases and vapours, concentration is also expressed in terms of its mass per unit volume of air (e.g., mg of contaminant 20

The properties of air and gases p e r m 3 of air). So it is u s e f u l to b e a b l e to c o n v e r t f r o m o n e to t h e o t h e r ( e . g . , p p m to m g m - 3 ) . T h e s t e p s a r e s t r a i g h t f o r w a r d . Example 2.5. For methanol vapour (molecular weight, 32 g mole -1) the mass concentration in air at STP is 200 mg m -3. W h a t is this concentration expressed in ppm? Note that the concentration may be expressed as 200 [mg m -3] C

=

32 [g mole-I] = 6.25 • 10 -3 mole m -3 Volume occupied by one mole (V1) is given by Equation (2.6), placing n m

RT V1

82[atm cm 3 ~

--

1. Thus

-~1 • 293[~

---

P

l[atm]

= 24,026 cm 3 mole -1 = 24 x 10 -3 m 3 mole I So the fractional volume occupied by 200 mg m -3 is c = 6.25 X

10 - 3

[mole

m -3] •

24 • 10 -3

[m 3

mole -1]

= 150 x 106 *

c = 150 ppm

More generally, the expression [ p p m ] • M [ g m o l e -1] [mg m-3] -

(2.20) 24 [litre m o l e - l ]

is u s e f u l so l o n g as it is r e m e m b e r e d t h a t t h e figure in t h e d e n o m i n a t o r d e p e n d s on t e m p e r a t u r e ( a n d t h a t t h e figure s h o w n is for 20~ o n l y ) .

Phase transitions: solids, liquids and gases M a t t e r is u s u a l l y a c k n o w l e d g e d to exist in t h r e e p h a s e s - solid, l i q u i d o r gas. It consists of s m a l l p a r t i c l e s c a l l e d atoms, w h i c h in t u r n m a y c o m b i n e to f o r m

21

Aerosol science for industrial hygienists larger entities known as molecules. Whether or not atoms or molecules come together to form solids, liquids or gases depends on combinations of pressure, volume and temperature. The most familiar example is water which, over reasonable ranges of familiar terrestial conditions, can exist either as solid ice, liquid water or gaseous water vapour. In the solid state, molecules of a material are located in fixed positions, held together in this ordered way by interatomic forces which are electrostatic in nature. When we speak of atoms or molecules occupying fixed positions inside a solid material, we are referring to their mean positions. In fact, for any temperature above 0~ they are in oscillatory motion about their mean locations and energy is continually being exchanged between the kinetic form (associated with velocity) and the potential form (associated with displacement). Averaged overall, energy is shared equally between the two energy forms. If extra internal energy is given to the solid matter in the form of heat, then the molecules perform greater excursions about their mean locations. If enough energy is supplied, the solid melts and enters the liquid phase. At that point, bonds may be broken and remade, and individual molecules can move around in the lattice, changing places with one another. The state of the material has now become 'fluid'. It is of particular interest to industrial hygienists to consider what happens near the surface of a liquid. Molecules there are 'bound' to other molecules only in the general direction of the body of the liquid, so, unlike molecules deeper in the body of the liquid, they experience a net inwards-seeking force. This accounts for the well-known phenomenon of surface tension. However, there is a statistical probability that a given molecule located instantaneously near the surface of the liquid may escape from the surface as a free entity and so enter the gaseous vapour phase. Thus, we have the phenomenon of evaporation. Conversely, molecules in the vapour phase may enter the liquid through the surface, and so contribute to condensation. For a given liquid-gas interface, transport of molecules will be taking place in both directions simultaneously. So whether there is net evaporation or condensation depends on which flux is dominant. This in turn depends on the local thermodynamic conditions. If enough energy is supplied to a liquid, then a temperature is eventually reached at which all the interatomic bonds can be broken permanently. All the atoms or molecules now become free to move at random, and the material becomes a gas in which all of the internal energy exists as kinetic energy. The preceding scenario for the transition from the liquid to the gaseous or vapour phase applies in principle to all substances. For example, under extreme thermodynamic conditions (e.g., very low temperature), even a gas such as helium can become a liquid. In relation to industrial hygiene, however, it is the convention to refer to as gases all substances which, under practical workplace conditions, are always found in the free molecular phase (e.g., air, sulphur dioxide, etc). On the other hand, vapours are regarded

22

The properties of air and gases as the free molecular phase of some other substance (e.g., water, organic solvents) which can, in the workplace, also be found in the liquid state. The saturation vapour pressure (P~) represents the pressure that would be exerted by vapour molecules in equilibrium with the same material in liquid form inside a closed container. This quantity is unique for the molecules in question and for a given temperature. It is also only unique for a flat liquid surface, and so is not exactly the same for a curved liquid surface (e.g., as in the case of a small water droplet). This physical picture can now be extended to enable discussion of the important environmental question of humidity. This relates to the presence in the air of free water molecules. The mass of water vapour per unit volume of air is referred to as the absolute humidity. Its partial vapour pressure (Pw,v) cannot exceed its saturation vapour pressure (Pw,~v) if the temperature and pressure are kept constant. It reaches a pressure of 1 atmosphere (1.01 x 105 Pa) at the temperature at which water boils (393~ The water vapour in air is considered to be saturated when Pw,v becomes equal to Pw,~v at the same temperature. At lower values of Pw,v it is unsaturated, and relative humidity (RH, expressed as a percentage) is given by Pw,v ) RH =

x 100

(2.21)

PW,SV In general, for any liquid-gas system, the saturation ratio is given by

PV SR =

(2.22) Psv

As already stated, for fixed temperature and pressure conditions, S R cannot exceed unity. But S R > 1 can be achieved by cooling a vapour that is already saturated. This is referred to as 'supersaturation'. Here, the reduction in temperature causes the saturation vapour pressure to be reduced while the partial pressure remains the same. Referring back to water, for a given mass concentration of vapour in the air, R H can therefore be raised by lowering the temperature (and hence raising SR). Conversely, raising the temperature lowers RH. For cooling, the temperature at which the water vapour becomes saturated is known as the dew point. Below this, condensation may take place, hence the appearance in the air of water droplets which are visible as mist or fog. The preceding discussion provides an i m p o r t a n t mechanism for the formation and evolution not only of water droplet aerosols but also aerosols of other substances where phase changes can take place (see Chapter 3).

23

Aerosol science for industrial hygienists 2.3 E L E M E N T A R Y F L U I D M E C H A N I C S Since, by definition, an aerosol consists of particles suspended in air, it is inevitable that the behaviour of an aerosol will be highly dependent on the properties and behaviour of the air itself ~ including the nature of its motion. This therefore requires a discussion of elementary fluid mechanics. Here the general term 'fluid' is used to refer to what in industrial hygiene is 'air'. But it also applies to gases other than air (and, to a considerable extent, liquids as well). There are two fluid mechanical aspects relevant to aerosol behaviour. At one end of the scale ~ the microscopic level ~ we are concerned with the flow of air over and around a small airborne particle of matter, and therefore to the fluid mechanical drag (and lift) forces acting on it. At the macroscopic level, interest is focused on the behaviour of larger-scale moving air systems (e.g., in workplace atmospheres at large, around samplers, in ducts, in filtration devices, etc.) within which aerosols are contained and dispersed.

Basic physics The starting point for describing the dynamical behaviour of a fluid is the familiar 'second law' of Sir Isaac Newton, the great English physicist from the 1600s. This states that the product of mass and acceleration of a 'body' is equal to the sum of all the forces acting. For a fluid, this law is applied to each small elemental packet and the forces include body forces (e.g., buoyancy, gravity), pressure forces (associated with local gradients in static pressure) and shearing forces (associated with viscosity and local gradients in velocity). Applying this to each of the three available spatial dimensions (x, y and z) provides three equations of motion. These, in themselves, are not sufficient to completely describe the motion of the fluid system because it is also necessary to ensure that the amount of fluid in the system is conserved. That is, no fluid is gained by ~ or lost from ~ the system. This is taken care of by the addition of a fourth equation which ensures that continuity is maintained. Altogether, therefore, we have a system of four equations, and this is the well-known set of Navier-Stokes equations from which all else in fluid dynamics is derived. In general, these equations include the possibility that the fluid density may vary (i.e., the flow may be compressible). However, as far as industrial hygiene is concerned, this feature may usually be neglected and flows may be treated as incompressible. The qualitative physical basis of the Navier-Stokes equations is deceptively simple. In practice, however, the resultant equations themselves are quite complex when it comes to applying them to realistic systems, and analytical solutions for the behaviour of flowing systems are available only for the very simplest cases. So it is not appropriate to go into further details in this book.

24

The properties of air and gases But there is ample literature available elsewhere for the interested reader. The ageless classic text of Schlichting (1968) is a good starting point for an in-depth reading of the subject. In terms of applications, the advent of m o d e r n c o m p u t e r s and appropriate software m e a n s that numerical solutions are now b e c o m i n g increasingly accessible.

Streamlines

O n e i m p o r t a n t set of solutions that can be provided by the N a v i e r - S t o k e s equations is the pattern of streamlines (or, perhaps m o r e a p p r o p r i a t e l y in three dimensions, streamsurfaces). It is this pattern which most graphically characterises the flow. Described most simply, it is equivalent to the flow visualisation that would be obtained by m a r k i n g the fluid with a suitable

Figure 2.5. Some streamline patterns typical of flows of interest to industrial hygienists: (a) flow about a particle falling under gravity; (b) flow into the hood of a local exhaust ventilation system and then through the ventilation ducting; and (c) flow about a worker (facing the wind) and into a small aspirating personal aerosol sampler attached to his body.

25

Aerosol science for industrial hygienists visible tracer. To illustrate this point, Figure 2.5 shows sketches of a number of flow systems relevant to aerosols in industrial hygiene situations. Figure 2.6 is another, describing the flow near an idealised single-orifice blunt aerosol sampler of axially-symmetric geometry, placed facing into the wind. Figure 2.6a is a sketch based on an actual photograph of an experimental flow pattern in a wind tunnel, where visualisation was achieved using fine balsa dust which was viewed in a two-dimensional slice through its axis under intense 'slit' illumination provided by an ordinary domestic slide projector. Figure 2.6b shows the approximately equivalent streamline pattern obtained by solution of the flow equations. From such comparisons, it is clear that experiment and theory are in good agreement. One feature in particular is worth noting in Figures 2.5 and 2.6. Based on the fact that there can be no transfer of fluid across streamlines, it follows directly that the convergence of the streamlines represents an increase in fluid velocity (as the fluid is being 'squeezed' into a smaller volume). The converse is true for diverging streamlines.

Figure 2.6. Streamline pattern for the flow near a simple idealised blunt aerosol sampler: (a) sketch based on a photograph from flow visualisation (from Vincent and Mark, 1982); and (b) corresponding streamline pattern obtained from potential flow theory.

26

The properties of air and gases In the preceding, the concept of laminar flow is invoked to convey how the fluid motion involves the smooth sliding over one a n o t h e r of layers of fluid contained within streamlines.

Boundary layers One physical constraint applicable to real fluids is that the velocity of the fluid must be zero everywhere at solid boundaries to the flow. This means that there must be a strong velocity gradient b e t w e e n the ftow very close to the wall and the faster-moving fluid further outside. The region over which the fluid is sheared in this way is referred to as the boundary layer. F r o m solutions of the N a v i e r - S t o k e s equations, we would find that this is where the effects of viscosity are most strongly felt. It is from the effects of such boundary layers that, for aerosol particles and for other bodies i m m e r s e d in fluids, the forces of drag and lift are derived.

Figure 2.7. Illustration of boundary layer flows: (a) for plane flow over a thin flat plate; and (b) for the flow into a channel between a pair of plates.

27

Aerosol science for industrial hygienists By way of illustration, Figure 2.7a illustrates the nature of a typical boundary layer for a thin plate immersed in and parallel to ~ a moving air stream. The approaching flow is uniformly distributed, but once it strikes the leading edge of the plate, the zero velocity boundary condition immediately comes into effect. As shown, the boundary layer close to the leading edge (see hatched area) is very thin; but it increases in thickness with distance downstream. Figure 2.7b shows the equivalent flow inside a channel made up of two such plates. Here the boundary layer on each plate grows until it merges with the other, at which point the flow is considered to be 'fully developed'. Now the velocity profile across the channel takes on a classic parabolic shape which is well-known for laminar channel flows.

Similarity One important consequence of boundary layers which is important in all applications of fluid mechanics is the concept of dynamic similarity. In general, the concept of similarity relates to the way in which the characteristics and behaviour of physical systems can be scaled as the variables (e.g., physical dimensions, velocities and other physical quantities) change. In fluid mechanics, in particular, similarity provides an important link between the behaviours of the microscopic and macroscopic systems described above. The scaling of the dynamical nature of geometrically-similar flows can be derived directly from solutions of the Navier-Stokes equations. This is achieved mathematically by non-dimensionalising the physical dimensions, velocities and pressures in the basic equations (e.g., by dividing local velocities by a characteristic main velocity U, and local dimensions by a characteristic main dimension D), and re-organising the terms in the basic equations. This is a standard approach applied widely throughout physics and engineering. To demonstrate this here would require us to set out the Navier-Stokes equations in greater detail than is desirable for present purposes. A simpler physically-based approach which ultimately comes down to the same thing requires consideration of the relative magnitudes of inertial forces (reflecting mass • acceleration of elements of the fluid) and viscous forces (reflecting the forces on fluid elements associated with shearing). For the inertial forces (IF), their magnitude for a flow represented by local velocities and scale of u and x, respectively, and characterised by the overall characteristic velocity and scale U and D respectively, may be written in the form

Ou I IFI~

Pair U

U2 ~

(2.23)

Pair

Ox

D

28

The properties of air and gases where Pair is the density of the air. Such forces are responsible for the main flow in the body of the fluid (i.e., away from the effect of solid surfaces). For viscous forces (VF), their magnitude may be written as

O2U

I VFI

[.I,U

(2.24)

---> OX2

D 2

where, as defined earlier, Ix is the viscosity of the air. Such forces occur close to flow boundaries where the fluid is undergoing 'shear', and are the forces responsible for what we refer to as 'friction' associated with energy loss. Such energy loss occurs in the form of heat. From Equations (2.23) and (2.24), the ratio of these forces is

IIFI

Pair

-

UD -= Re

(2.25)

IVf[ and this dimensionless quantity is known as the Reynolds' number (Re, after Sir Osborne Reynolds, a pioneering British fluid dynamicist of the late 1800s). For the microscopic flow of air about a small airborne particle, for example, D would be the particle diameter and U the relative velocity between the particle and the surrounding air. For the macroscopic flow of air in a pipe, D now becomes the internal diameter of the pipe and U the mean air velocity inside the pipe. The same result as Equation (2.25) can be achieved by purely dimensional arguments, and it is instructive to present this alternative approach. This follows the law of dimensional analysis which states that, for any physical system, any solution for its behaviour must be independent of the system of units using for describing it. Here we begin by making a list of all the possible basic independent physical variables that can influence the behaviour of the system in question, and noting their dimensions. For most fluid flow systems, the variables in question are the dimension D and characteristic velocity U as already defined. Now we assign the dimensions M - time, L -- length and T - time. To achieve similarity by the dimensional argument requires D a U b Pair c p.d =

dimensionless in M, L and T

For the variables indicated we have D in [LI U in [L T -1] [3air in [M L -3] I~ in [M L - 1 T - I ]

29

(2.26)

Aerosol science for industrial hygien&ts From the dimensional statement in Equation (2.26), we must now ensure that the dimensions of M, L and T are each zero. This results in the three equations ForM, c + d = 0 ForL, a+ b + 3c-d-0 For T , - b - d = O By setting a --1, we may solve these three equations simultaneously to determine b, c and d. It is a simple matter to obtain b = 1, c - 1 and d = - 1 . Thus, the dimensionless group which achieves the dynamic scaling required is again D U Pair =- Re

(2.27)

as in Equation (2.25). The Reynolds' number concept identified in this discussion embodies some of the most important ideas in the whole of fluid dynamics. Physically, its meaning is rooted in the boundary layer concept. As described in the first analysis, it is effectively the ratio of the magnitude of inertial forces in the main body of the flow (e.g., near the axis of the pipe) to the magnitude of viscous forces close to the flow boundaries (e.g., near the pipe wall). For large Re (i.e., large U and D), it means that the boundary layer part of the flow has relatively small effect on the overall character of the flow system as a whole. On the other hand, for small Re (i.e., small U and D), the boundary layer flow and associated viscous effects dominate. Many properties of the flow (e.g., streamline pattern, drag force, etc.) vary quite sharply with changes in Re at small values of Re, but become much less sensitive to changes at large values. In general, for dynamical similarity to occur in the shape and nature of fluid flows having geometrically similar boundaries, Re should be kept constant. Thus, for example, for a pair of such flows (say System 1 and System 2) where the dimensional scale of System 1 is 10 times greater than that for System 2, then the characteristic velocity in System 1 should be 10 times less.

Example 2.6. Consider the flow of air at atmospheric pressure moving through a pipe of diameter 10 cm at a mean velocity of 2 m s -1. What is Reynolds' number and what is the nature of the flow? Pair -- 1 . 2 9 3

kg m -3

Ix -- 1 8 . 3 x 1 0 - 6 N s m - 2

30

The properties of air and gases so that 0.1[m] x 2[m s-11 X 1.293[kg m -31 R E -18x10-6[N s m -2]

14,250 Here Re is large, so inertial forces dominate.

Example 2.7. Consider from the same point of view the flow of air around a small particle of diameter 10 txm where the relative velocity between the air and the particle is 1 mm s -1. 10-5 x 10-3 x 1.293 Re = 18 x 10 - 6

0.0007

*

Here Re is very small, so viscous forces dominate.

The following easy-to r e m e m b e r expression is useful and accurate enough for most workplace or laboratory estimates" Re-

7 x 104DU

(2.28)

where D is in [m] and U is in [m s - l ] .

Potential flow

For large enough Re, a given flow system may be treated for many purposes as if it were inviscid (i.e., having zero viscosity). Then the N a v i e r - S t o k e s equations become much simpler and amenable to mathematical treatment. In fact, they become equivalent to equations familiar to physicists and engineers working in other fields ~ for example, the flow of heat in a t e m p e r a t u r e field or the flow of electric charge in an electric field. Hence we refer to potential flow. Despite the idealised nature of its underlying assumptions, potential flow solutions have many practical applications relevant to industrial hygiene. For e x a m p l e , they have been widely (and successfully) used to describe the flow around aerosol sampling devices and the hoods of local exhaust ventilation systems.

31

Aerosol science for industrial hygien&ts

Stagnation In addition to the features already described, a moving fluid is also characterised by static pressure and velocity pressure. The first of these is associated with the potential energy of the flow, and the second with its kinetic energy. The sum of these two pressures is the total pressure, associated with the total energy in the system. These are concepts which are familiar in relation to the design and characterisation of ventilation systems. They relate in particular to what happens along a streamline of the fluid motion. One important property of a streamline is that, everywhere along its length, the sum of the local static pressure and the local dynamic pressure remains constant. This is the well-known Bernouilli theorem. In a given streamline pattern, it is not inconsistent that an individual streamline may intersect with a solid flow boundary. This does not imply that there is any flow into the surface. Rather it describes the limiting case where the fluid comes to rest at that point. Here, therefore, all the energy in the flow is converted into potential energy in the form of static pressure and so the dynamic pressure falls to zero. The value of the static pressure therefore becomes a maximum, and so the point where the streamline intersects with the surface is known as the stagnation point. To an industrial hygienist, this has two aspects of practical interest. The first is that stagnation points on surfaces (e.g., near the entrances to samplers or to ventilation ducts) are where there tends to be maximum aerosol deposition. The second ~ and incidentally as far as this book is concerned ~ is that it provides the principle of operation of the pitot-static tube which is widely used in velocity measurement. A typical example of the phenomenon of stagnation is given in Figure 2.8.

Separation The phenomenon of boundary layer separation is particularly important for flow about a body and is associated with the distribution of static pressure

Stagnation point (velocity falls to zero) /

Dividing streamline (or streamsurface) Figure 2.8.

Illustration of the phenomenon of stagnation.

32

The properties of air and gases

over the body surface. For idealised frictionless inviscid flow, there is no net loss of energy along a streamline as it approaches the body, diverges to pass around it, and then recovers on the downstream side. This is because it is assumed that there is no velocity shear (and hence no energy losses due to viscosity) near the boundary. There is just the transformation of potential energy to kinetic energy (as the streamlines get closer together and, hence, the velocity increases) and back to potential energy again. However, for a streamline which passes close to the surface and where there are friction losses due to viscosity (the real situation), not all of the initial potential energy is recovered. The result is that the fluid in the boundary layer close to the surface of the body (on its downstream side) comes to rest prematurely. When and where this occurs, the flow breaks away from the body, enclosing a negative-pressure recirculating region which may or not be turbulent (depending on the characteristic Re-value for the flow, as discussed below). Some typical separated flows are illustrated in Figure 2.9.

(a) Sphere with stable wake cavity

(b) Surface-mounted block

///////////

////

/7// Q

|

-------Stagnation points

Region of separated flow

Figure 2.9. Illustration of some typical separated flows: (a) sphere with a relatively stable wake 'cavity' ('doughnut-shaped' region of recirculating flow); and (b) surface-mounted block (e.g., as for a building in the atmospheric wind). (From Vincent, J.H., Aerosol Sampling: Science and Practice, Copyright 1989, adapted by permission of John Wiley and Sons Limited)

33

Aerosol science for industrial hygienists Turbulence

No discussion of fluid mechanics relevant to practical applications, no matter how elementary, would be complete without mention of the p h e n o m e n o n of turbulence. This was first identified by the aforementioned Sir Osborne Reynolds in the nineteenth century during his classic experiments with fluid flow in pipes (which led directly to the formation of the Reynolds' number concept). This complex and fascinating phenomenon is certainly relevant to many aspects of air flows in the context of industrial hygiene. The starting point for a qualitative discussion is the ideal, non-turbulent case known as laminar flow, in which ~ as stated earlier ~ the layers of the fluid (i.e., between the streamlines) slide smoothly over one another. However, from the Navier-Stokes equations it can be shown theoretically that, if inertial forces are large enough in relation to viscous forces (i.e., Re is large enough), then a disturbance in velocity of sufficient amplitude introduced into the flow can lead to overall instability. Such a disturbance might arise, for example, due to the passage of the flow around some sort of flow blockage and the resultant flow separation. Fluid mechanical amplification of the resultant triggering disturbance leads to a state of overall instability which appears as randomly-fluctuating motions superimposed on the mean flow. Two of the consequences are: (a) an apparent increase in the viscosity of the fluid (which, in the example of flow through a pipe, would appear as an increase in resistance to the flow); and (b) an accompanying sharp rise in its mixing properties. In the light of the latter, it is necessary to qualify the earlier comment about the inability of fluid to be transported across streamlines. Strictly speaking, this applies only to laminar flow. For turbulent flow, the resultant turbulent mixing provides a mechanism for the exchange of fluid elements across streamlines ~ in both directions. However, it remains that there can be no net transfer of fluid when transport is averaged over time. Turbulence itself is an extremely complex phenomenon and cannot be treated in detail in this book. However, industrial hygienists need to be aware of its existence and recognise where it can play a role, not just in aerosol aspects but also in relation to ventilation and air movement in general. The macroscopic properties of turbulence are especially important. For a given turbulent flow, if it is visualised at an instantaneous point in time, its streamline pattern will appear to be highly chaotic. However, averaged over time, the more-familiar streamline pattern is restored. Once a disturbance has occurred and turbulence has been established, it is important to recognise that the consequences referred to are associated with the properties of the turbulence and are not intrinsic properties of the fluid per se. So the apparent increase in fluid viscosity is just that ~ apparent. It is an externally observed property of the bulk fluid motion. Meanwhile the basic internal viscosity of the fluid, determined by the individual molecular motions of the gas as 34

The properties of air and gases described by kinetic theory, remains the same. From practical observation, it becomes clear that, in a given flow system, turbulence usually occurs when Re exceeds a certain value. That value depends on the flow system in question, whether it be the flow in a pipe or duct, flow in a boundary layer, or around a bluff flow obstacle (e.g., a small particle), etc. For each such system, Re can be calculated from the characteristic dimension and velocity, using either Equation (2.27) or the easy-to-remember Equation (2.28). Typically, for most practical systems like those listed, the threshold value of Re for the onset of turbulence is around 2000-3000. At the microscopic level, the origin of the turbulence has already been described as a small initial disturbance. Clearly, this in itself cannot provide the energy which drives the intense turbulent motions like those usually encountered. This must derive from the mean flow itself. Once the flow has been rendered unstable by the action of the mean flow, the initial and relatively large-scale ~ fluctuating motions (i.e., eddies) created by the triggering disturbance are stretched and distorted, and so are broken down successively into smaller and smaller eddies. The energy from the mean flow which goes into the turbulence is therefore handed down ~ in a so-called 'cascade' process ~ through the resultant range of eddy sizes until, eventually, the eddies become very small. At that point, when the Re-values for the eddies themselves become very small, viscous forces take over inside the eddies and the energy is therefore finally dissipated as heat. From this qualitative description, there emerges a physical picture of turbulence as an irreversible spectral phenomenon, characterised by a continuous distribution of eddy sizes and associated velocity fluctuations (superimposed on the mean flow). One important feature of this picture is that the energy that goes from the mean flow into the turbulence can never be recovered as useful kinetic energy. This is a consequence of the second law of thermodynamics, one of the most important laws of physics. For many working purposes relevant to industrial hygiene, the picture of turbulence may be simplified so that the phenomenon can be described in terms of two 'bulk' properties; the characteristic intensity of the fluctuations (U', the root-mean-square value of the fluctuating velocity) and the characteristic mean length scale of those fluctuating motions (l). The degree of mixing increases with both, and a fair estimate of the diffusivity of a turbulent fluid (Dft) is given by Dft ~ lU'

(2.29)

which, as can be seen by inspection has the correct dimensions of [L2T-1], or [m 2 s -1] in the SI system of units. The scale and intensity of turbulence can vary greatly from one situation to another.

35

Aerosol science for industrial hygienists

Example 2.8. Estimate the diffusivity of turbulent air moving in a pipe of diameter 10 cm at mean velocity 1 m s -1 and where the intensity of turbulence is given as 3%. Note that the intensity of turbulence is given by U

t

- 0.03 U where, as before, U is the mean velocity. The characteristic length scale (l) cannot exceed the diameter of the pipe. A rough estimate is 3 cm. From Equation (2.29), we get Oft ~ 3 X 10 - 2 [m] • 0.03 [m s -1]

*

Dft ~ 1 0 - 3 m 2 s - 1

It is i n t e r e s t i n g to n o t e t h a t Dft as c a l c u l a t e d h e r e is t w o o r d e r s of m a g n i t u d e g r e a t e r t h a n the classical diffusivity of gas m o l e c u l e s q u o t e d e a r l i e r ( = 10 -5 m 2 s - l ) . In m o s t practical s i t u a t i o n s in w o r k p l a c e s , the scale a n d i n t e n s i t y of t u r b u l e n c e will be c o n s i d e r a b l y g r e a t e r still.

REFERENCES Schlichting, H. (1968). Boundary Layer Theory, 6th Edn. McGraw-Hill, New York. Vincent, J.H. and Mark, D. (1982). Application of blunt sampler theory to the definition and measurement of inhalable dust. In: Inhaled Particles V (ed. W.H. Walton). Pergamon Press, Oxford, pp. 3-19.

36

CHAPTER 3

Properties of aerosols 3.1 AEROSOL GENERATION IN WORKPLACES Many industrial processes generate aerosols in one form or another, usually as a side effect of the process itself and by a wide variety of physical and chemical means. The list of possibilities outlined below is far from exhaustive, but serves to indicate the range of types of aerosol that need to be recognised and dealt with in the field. Mechanical generation of dry aerosols Mechanical generation of dry aerosols in the form of airborne dust particles occurs as a consequence of many industrial processes. For example: In the mineral extraction industries (e.g., coalmining, iron ore mining, quarrying, etc.), during the excavation and cutting of rock and during subsequent crushing, grading, sieving, riddling, transportation, handling, bagging, etc; In the textiles industries (e.g., cotton, flax, asbestos) during bale and bag opening, and subsequently during various processes such as carding, spinning, winding, weaving, etc; In the chemical industries (e.g., silica gel, fertilisers) during the processing, handling and transportation of bulk powders and aggregates, etc; In foundries during moulding, sand blasting, etc; In the wordworking industries (e.g., furniture) during the cutting, working and sanding of hard and soft timbers, etc; and many others. In these, the principl~ physical processes of aerosol generation are associated with the fracture, breaking and cutting of large pieces of bulk material objects into smaller ones (and the creation of new 37

Aerosol science for industrial hygienists surfaces), abrasion, agitation, the breaking of adhesive forces (e.g., van der Waals, electrostatic) holding primary particles together in aggregates, entrainment into the air, etc. Of course, the amount of aerosol generated by such means varies with the nature and condition of the bulk material in question. But, for a given material, it tends to increase with the energy given to the dispersal process. It is also strongly sensitive to whether the material is hydrophilic (i.e., can absorb water), influencing the surface binding forces which hold the bulk material together. In turn, it is dependent on the previous history of the material (i.e., processing, additives, etc.) and on the humidity of the atmosphere in which it is held. Other factors include the friability of the bulk material ~ that is, its ability to be broken or crumbled. The actual process of aerosol generation and dispersal is also ultimately dependent on the movement of the surrounding atmosphere. As indicated in Chapter 2, there is in principle no upper limit on the size of individual particles produced in this way. The industries given as examples above are well-known to be 'dusty' and have long been the subject of industrial hygiene interest. For materials like those described, it is therefore relevant to discuss the question of 'dustiness'. This term refers to the dust-generating capacity of certain types of bulk material when handled mechanically under specified conditions. Over the years, a variety of technical laboratory methods have been proposed which enable quantitative assessments to be made of relevant (and consistent) indices of dustiness, with the aim of enabling materials to be tested and placed in rank order with respect to this property. Two of these are illustrated in Figure 3.1. The one shown in Figure 3.1a is the gravity dispersion method, involving the dropping of known masses of sample bulk material from a given height into an enclosed space under defined conditions, and assessing the resultant dust cloud (e.g., by gravimetric sampling or optically). This is intended to simulate an industrial process by which bulk aggregated material undergoes a single drop (e.g., as in emptying a bag into a hopper). The version shown in Figure 3.1b is the mechanical dispersion method whereby the sample is dispersed by agitation. This is intended to simulate an industrial process where the aggregated material undergoes multiple drops or continuous agitation (e.g., as in a mixer, or during conveying). The version shown employs a rotating drum for this purpose. Another approach (not shown) is the gas dispersion method, involving the passage of a stream of compressed gas through the sample. In these (and other devices not shown), the concentration of the airborne dust resulting from the mechanical agitation is measured, either gravimetrically by weighing samples captured by aspiration and collected by impaction or filtration, or by remote optical sensing methods (the physical means of which are are described in later chapters). Inevitably, such dustiness estimation methods differ in important physical respects, and so it is not surprising to find that they tend to give differing results. Indeed, they do not necessarily even

38

Properties of aerosols

Figure 3.1. Illustrations of technical methods that have been proposed for measuring indices of 'dustiness' for bulk solid materials" (a) the gravity drop method; and (b) the rotating drum method.

place materials in the same rank order of dustiness ~ although, it might be argued, each might be a valid test for a particular class of industrial dust generation process. For many industries, and hence for many industrial hygiene situations, dustiness is a useful index upon which primary control strategies might be based. Therefore some move towards standardisation is required ~ and, in fact, is taking place. In recent years, a working party of the Technology Committee of the British Occupational Hygiene Society has carried out important work on dustiness estimation (Hammond et al., 1985). Based on this work Table 3.1 gives, by way of illustration for each of the methods shown in Figure 3.1, the relative dustiness index (using the average for all the 'dustiness' values obtained for each method as the basis of normalisation) and the rank order of dustiness. In this table, for each estimation method indicated, the original raw data (e.g., the figures obtained from the optical detector in Figure 3.1a, or from the aerosol sampled mass in Figure 3.1b) have been normalised by

39

Aerosol science for industrial hygienists Table 3.1. Relative 'dustiness' for a range of c o m m o n industrial materials as o b t a i n e d using the gravity dispersion and rotating d r u m m e t h o d s (see Figure 3.1) (based on data r e p o r t e d in H a m m o n d et al., 1985).

Material

Method

Rubber 'E' Rubber 'F' Rubber 'G' Sulphur Oil absorber Chalk Silica Charcoal

Gravity drop

Rotating drum

0.04(1) 0.36(3) 0.72(4) 0.20(2) 0.95(5) 1.39(6) 1.41 (7) 2.92(8)

0.02(1) 0.04(2) 0.14(4) 0.20(5) 0.05(3) 0.22(6) 2.81(7) 4.55(8)

Note: The figures given are tor the ratio of the dustiness for the material in question and the average of the values measured for all eight materials shown. The figures in parentheses are the rank orders of 'dustinbss' as measured by the two methods.

dividing each by the average for that method. It is seen that not only are the magnitudes of the relative dustiness different but so too (in parentheses) is the rank ordering between the two methods. Thus, the methods described are very empirical. Work is continuing on this important problem. Lyons and Mark (1994) have further extended the rotating drum method and determined the dustiness of a wide range of aggregated materials in terms of their abilities to generate aerosol falling within defined health-related particle size fractions (described in Chapter 8). In addition, at the time of writing this book, laboratories from a number of European countries are planning Pan-European research to investigate factors such as humidity, electrostatic charge, wall losses, adhesion, friability, particle size, and so on (C.M. Hammond, Michelin Tyre plc, U.K., personal communication). Similar efforts have been taking place in the United States (e.g., Heitbrink et al., 1990).

Mechanical generation of liquid droplet aerosols Mechanical generation of droplet aerosols can be achieved for any liquid provided that enough mechanical energy is provided in disrupting the surface of the original bulk liquid. Liquid droplet aerosols are less common in workplaces than dry dusts, but are found in some types of industry; for example: Processes where liquids are involved as the primary working material (e.g., paint spraying, crop spraying, etc.);

40

Properties of aerosols Processes where liquids are involved as the auxiliary material (e.g., machining fluids, electroplating, etc); and so on. They may also be found, perhaps ironically, as a side effect of certain technical measures (e.g., wet scrubbers in the underground mine environment) which are sometimes employed for the control of dry dusts. Physically, the droplet formation process involves the mechanical breaking of the surface tension forces which hold the bulk liquid together. This may be achieved, for example, by simply splashing or otherwise mechanically agitating the liquid surface; or more vigorously by virtue of the shearing forces associated with passing the liquid under pressure through a nozzle, by mixing the liquid with a compressed gas jet, or by the action of centrifugal forces, electrostatic forces, etc. The process is variously described as nebulisation, atomisation and spraying, the actual terminology employed usually reflecting the size of the resultant droplets (e.g., from small to large for the processes in the order indicated). The size of droplets produced is dependent on the energy that goes into aerosol generation, with the smallest particles usually being generated by the most energetic mechanisms. As mentioned in the previous chapter, the ultimate upper limit of size of droplet sustainable once the aerosol has been formed is governed by the balance between surface tension forces (tending to hold the molecules of the droplet together) and gravitational and shearing forces (tending to distort and, ultimately, disrupt the droplet). Formation by molecular processes

The other class of basic aerosol generation processes involves the formation of particles by the aggregation of molecules, by physical condensation and/or chemical reactions. Physical formation by this means is usually known as nucleation, and involves the kinetics of molecular transfer from the gas to the liquid phase and thereafter, possibly, the solid phase. Heterogeneous nucleation refers to the most common version of the process, involving the condensation of molecules onto pre-existing surfaces in the form of very small (sub-micrometre) aerosol particles known as condensation nuclei. Normal atmospheric air contains sufficient numbers of such particles to enable this process to occur readily if the physical conditions are favourable. In particular, the atmosphere needs to be supersaturated so that the saturation ratio (SR) is greater than 1 (see Chapter 2). This requires the cooling of air that is already saturated, as occurs for example in the atmosphere prior to the formation of the water droplets that go to make up clouds and fogs. Here, therefore, the surface of any insoluble solid nucleus will have a layer of adsorbed vapour molecules, thus providing the starting point for the arrival of more molecules, in turn leading to droplet formation and growth. Whether

41

Aerosol science for industrial hygienists or not that growth can proceed ~ and how far it can proceed ~ will depend on the physical conditions required for such evolution. This is discussed later in this chapter. For a soluble solid nucleus, the process is somewhat different. For such substances which have a strong affinity for water, the conditions for the condensation of water molecules is more favourable than for insoluble nuclei, and so droplet formation can take place at lower values of S R. Homogeneous nucleation is less common, and refers to the spontaneous formation of liquid droplets by the transfer of molecules directly from the gas phase. It does not require the pre-existence of small particulate nuclei. Here, however, the conditions for nucleation are less favourable. So considerably higher values are required for the supersaturation ratio, with typically S R > 4. Much of the preceding discussion is focused on the common example of the formation of water droplets in humid air. But the same processes can occur with other materials; for example, in the condensation of hot metal vapours to form the primary particles which are the 'building blocks' of metal fume (e.g., during the cooling of the plasma in arc welding). Here the processes leading to the formation of aerosol particles recognisable as being associated with the processes in question are very complicated, involving the production of very fine primary particles of nanometre dimensions and their subsequent aggregation into complex structures which, although they are much larger than the original primary particles, are still relatively fine (micrometre-sized). Such processes are characterised by the involvement of heat energy, and so conditions may sometimes be favourable for chemical reactions also to occur. This, therefore, leads to another version of the molecular formation process, chemical formation. This occurs when certain types of free radical species present in the air act as nuclei upon which other molecules may interact and react chemically. Examples include the formation of soot particles during the combustion of hydrocarbons (e.g., as in the burning of fossil fuels) and the formation of oxides in a metallic arc (e.g., as in welding). The complicated kinetics of the processes described require detailed discussion which is beyond the scope of this book. A more rigorous discussion appears in several texts, including that by Friedlander (1977).

3.2 T H E E V O L U T I O N OF A E R O S O L S It should not be assumed that an aerosol, once it has been dispersed, will necessarily retain the properties with which it began. Depending on the material in question, the initial generation process and the concentration of the aerosol, and other conditions in the surrounding air, a number of possibilities exist for evolutionary changes. These include" (a) growth by coagulation, agglomeration, coalescence and condensation; and (b) diminution by disintegration and evaporation. 42

Properties of aerosols Coagulation, agglomeration and coalescence If particles in an aerosol can come into contact with one another and, in so doing, remain joined together, then the total number of airborne entities in the aerosol is reduced whilst the effective entity size is increased. At the same time, however, the mass concentration remains the same. One mechanism by which the particles come into contact is thermal in origin, derived from their random (or Brownian) motions governed by collisions with gas molecules whose own random velocities are a function of gas temperature as described by the classical kinetic theory of gases. Another mechanism is so-called kinematic coagulation, where particles are brought together as a result of their different migration velocities under the influence of an externally-applied force field. One example of this mechanism involves gravity, in which particles of different sizes in still air fall at different velocities. Here the slowly-falling small particles are overtaken by the faster-falling larger ones, and so become engulfed by them. Whatever the actual mechanism, this process is commonly referred to in general as coagulation, although it is also sometimes called agglomeration or more appropriately for liquid droplets where the joining of two or more spherical droplets results in a single larger spherical droplet ~ coalescence. For thermal coagulation, the dynamics of the process is continuous and its full description complicated. However, a simplified collision model can shed useful light, indicating that the rate at which the process occurs depends on the size of the particles, their thermal velocities and their instantaneous airborne concentration. The process is discussed more fully in Chapter 4. The converse to the process of coagulation is disintegration. This occurs when a system of particles which has previously combined together to form a single particle is subjected to external forces such that the adhesive and cohesive bonds which hold its individual elements together are broken. In particular for systems where the individual elements are small, such binding forces are strong and they are not readily overcome. For example, therefore, fine fume or smoke aggregates of nanometre-sized primary particles are not prone to break-up. However, in strongly sheared flows (e.g., in some types of sampling device such as an impactor) or in strong turbulence, shear forces can be encountered which are sufficient in magnitude to cause the break-up of dusty agglomerates and the shattering of large droplets. Such particles can also be broken up during high velocity impaction onto surfaces.

Condensation and evaporation We have already discussed the subject of condensation in connection with the initial aerosol formation out of the vapour phase. Depending on the ambient conditions, condensation may continue after formation, with molecules 43

Aerosol science for industrial hygienists entering the particle from the vapour phase, leading to particle growth, and hence changes in the particle size distribution of the aerosol (towards larger particles) and an increase in its airborne mass concentration. Alternatively, the ambient conditions may be such that evaporation predominates, with molecules leaving particles and entering the vapour phase. Now particles become smaller and the aerosol mass concentration decreases. The basis for a physical discussion of these phenomena has been given in Chapter 2. However, it should be noted that the description there assumed the liquid surface to be flat. For the case of the curved surface of a liquid droplet, the kinetics of the transport of molecules inside and outside of the surface is modified by the change in geometry. Specifically, for given pressure and temperature, the smaller the droplet the greater the partial pressure which is required to maintain equilibrium at the liquid surface. This is known as the Kelvin effect, and is important to the discussion about the growth of droplets by condensation ~ or conversely diminution by evaporation. For the droplet to maintain a constant size, the equilibrium value of the saturation ratio (S R = eqSR) is given by eqSR = exp

(

}

pRdeqT

(3.1)

deq is the equilibrium droplet size, 13 the coefficient of surface tension for the liquid in question, M is its molecular weight, p is its density, and R is the universal gas constant. Note here that S R = Pv/P~v as before (Chapter 2). But now, whereas p~ is the equilibrium partial pressure at the liquid droplet surface, p~,, is the saturation vapour pressure corresponding to a flat liquid surface. where

Example 3.1. For a water droplet of diameter 0.01 txm in air at 20~ saturation ratio required to maintain it at that size?

what is the

Note that 13 = 7 2 . 7 x 1 0 - 3 k g s - 2 M = 18 g m o l e - 1 = 0 . 0 1 8 k g m o l e - 1 p = 103 k g m - 3 R = 8 . 3 1 4 k g m 2 s - 2 ~ -1 m o l e - 1

From Equation (3.1) we get (4 x 72.7[kg S -2]

X

O.O18[kg mole-l])

ln{eqSR} = (103[kg m -3] x 8.314[kg m2s -2 o K - l m o l e -1] x 10-8[m] x 293[~ = 0.2148

44

Properties of aerosols *

eqSR = 1.24

Note that, similarly, for a water droplet in equilibrium at diameter 1 I~m, we obtain eqSR about 1.002 F o r a d r o p l e t of g i v e n d i a m e t e r a n d for a n y v a l u e of S R less t h a n t h e e q u i l i b r i u m v a l u e c a l c u l a t e d f r o m E q u a t i o n ( 3 . 1 ) , t h e d r o p l e t will s h r i n k by e v a p o r a t i o n . H o w e v e r , if S R is g r e a t e r , t h e d r o p l e t will g r o w b y condensation.

1.008

~R

0.1

1.0 d (p,m)

1.008

/

/

~R 1.000

0.996 0.1

1.0

I0

d (p,m)

Figure 3.2. (a) Equilibrium saturation r a t i o (eqSR) as a function of droplet diameter (d) for pure water (at 20~ The curve derives from the Kelvin equation (Equation (3.1)) so that a droplet described by conditions in the hatched area above the curve will grow while one in the hatched area below will shrink. (b) Equilibrium saturation r a t i o (eqSR) as a function of droplet diameter (d) for sodium chloride solution (at 20~ where the figures on the curves relate to the mass of sodium chloride present in the droplet. (Note that the concentration of sodium chloride decreases as the size of the droplet increases.)

45

Aerosol science for industrial hygienists From Equation (3.1), it is seen that each equilibrium droplet size (for a given liquid) is uniquely associated with a given value of eqSR. The effect is illustrated for droplets of pure water in Figure 3.2a. In this figure, the region above the curve represents conditions (S R, d) where droplet growth will occur; the region below the curve is where droplets will shrink. One important practical result is that, in a polydisperse system consisting of different-sized droplets and for a given value for SR, larger particles may grow by condensation while the smaller particles shrink by evaporation. So we have a situation where larger droplets can become even larger at the expense of smaller droplets. In many practical cases of heterogeneous nucleation, the starting nucleus is made up of a solid material which is soluble in the liquid in question (e.g., salt in water). Here, as the droplet grows by condensation, the composition of the droplet changes, and this in turn influences the kinetics of molecular motion and hence of particle growth. This is shown for sodium chloride in Figure 3.2b, where it is seen that the effect is towards the potential for enhanced growth (i.e., the region above the curve is enlarged). The effect is the most marked for smaller droplets. An excellent fuller description of the evolution of droplets by condensation and evaporation is given by Hinds (1982). Such considerations are certainly important in extreme situations (e.g., inside some industrial processes) and are relevant to aerosol long-term behaviour in the atmosphere. But they may also be relevant to some workplace exposure situations where droplet aerosols are formed. They can be particularly important in relation to health effects for some types of particle which, when inhaled into the lung where the air is further humidified, can subsequently grow by condensation during transport inside the respiratory tract. This in turn can lead to significant effects on regional deposition.

3.3 P A R T I C L E M O R P H O L O G Y Particle shape classification Particle shape can have a significant bearing on effects relevant to industrial hygiene; for example, on the way particles behave in the air, and how they behave after they have been deposited in the respiratory tract. Particle shape falls into a number of categories, some of which are illustrated in Figure 3.3. Firstly there are aerosols in which particles are spherical, including liquid mists, fogs and sprays (unless, of course, the droplets are so large that they become distorted by the effects of gravity and shearing forces) and some dry aerosols (e.g., glassy spheres condensing out of some high temperature 46

Properties of aerosols

Figure 3.3. Illustrationof typical shapes of particles found in the industrial work environment.

processes). Idealised spheres (e.g., of polystyrene latex) are also artificially produced in the laboratory for some aerosol research applications. Secondly, there are non-spherical, angular particles which have no preferred dimension or whose aspect ratio cannot be said to be substantially different from unity. These are referred to as regular or isometric. Thirdly, there are essentially two-dimensional, flat platelet particles. Finally there are long, thin, rod- or needle-shaped particles, referred to as fibrous or acicular. Particles typical of aerosols at mineral extraction industry workplaces (such as coalmining or quarrying) are usually angular but tend to be isometric in shape, not exhibiting any obvious difference between length and breadth. But there are some exceptions. For some dusts arising, for example, during extraction of anthracite or minerals from mica-bearing rock, there are substantial populations of platelet-like particles. Particles from aerosols in the atmospheres of textiles factories exhibit a variety of morphological properties, ranging from isometric like those just described to elongated ones (see the examples in Figure 3.4). For the cotton, flax and jute industries, most of the finer particles tend to be irregular but not obviously fibrous. These tend to be fragments which may not necessarily be intrinsic to the basic textile material itself but which may have become attached to the bulk material during its previous history (e.g., dirt, detritus, biological entities, etc.). On the other hand, there do tend to be fibrous entities amongst the coarser particles, and these are more likely to be of the basic textile material. In the case of asbestos aerosols, most of the particles tend to be both very fine and elongated, and are therefore markedly fibrous. The same is true for aerosols found in the man-made mineral fibres industries.

47

Aerosol science for industrial hygienists

Figure 3.4a

Figure 3.4b

48

Properties of aerosols

Figure 3.4c

Figure 3.4d

49

Aerosol science for industrial hygienists

Figure 3.4e Figure 3.4.

Photographs of actual dust particles typical of those actually found in workplace atmospheres in the textile industries.

The welding fume example illustrated in Figure 3.5 reveals another morphological property exhibited by some aerosol particles. Here, as already mentioned, the particles of the aerosol exist in the form of complex structures, each one made up of a large number of very much smaller (nanometre-sized) primary particles formed during condensation from a hot metallic vapour. This property leads to the concept offractal geometry which has only relatively recently been applied to aerosol particle morphology. As first introduced by Mandelbrot (1983), the main property of an object exhibiting fractal properties is the fact that its detailed structure repeats itself when viewed at progressively larger magnifications, a phenomenon sometimes referred to as 'self-similarity'.

Figure 3.5. Picture of a typical welding fume particle.

50

Properties of aerosols Fibres

No discussion of workplace aerosols can be conducted without drawing special attention to the morphological properties of fine fibrous aerosols (of which asbestos is the most widely-discussed example). For such particles, significant adverse health effects are known to be associated with inhalation exposure, and the two dimensions of particle length and diameter, respectively, are both independently relevant to the risk. On the one hand, it is the particle's diameter which (largely) governs its aerodynamic motion and hence its ability once inhaled ~ to penetrate far down into the deep lung (see Chapter 4). On the other hand, it is the particle's length which mainly influences how well the lung's defence mechanisms can cope with the particle after it has been deposited (see Chapter 7). The morphology of asbestos fibres is further complicated by the fact that, for some types (e.g., chrysotile), the fibres are not straight, but tend to be flexible and curly. In addition, closer inspection of chrysotile fibres under the electron microscope reveals that they are made up of bundles of even finer fibrous elements (or fibrils). Under certain conditions (e.g., under the action of lung fluids after deposition in the respiratory tract), such particles may be broken down into their much finer fibril constituents. Although serious interest in the health effects associated with the inhalation of fibres began with asbestos, attention in recent years has also turned to other materials from which airborne fibres can be generated, including glass and carbon fibre. From this has emerged the concept of a 'durable fibre': that is, one that can persist for a long time in human tissue. 3.4 A E R O S O L C O N C E N T R A T I O N Interest in aerosol measurement by industrial hygienists is stimulated by the practical need to assess the exposures of people at work to potentially harmful particles, to use the information gained to assess risks to health, and to provide a basis for the setting and maintenance of standards. For workplace aerosols, such concentration is usually expressed in terms of particulate mass per unit volume of air, most commonly in units of milligrams per cubic metre (mg m-3), in contrast to atmospheric aerosols for which concentrations tend to be orders of magnitude lower and so are usually expressed in terms of micrograms per cubic metre of air (Ixg m-3). A related property of practical interest for some particles (e.g., durable fibres, bacteria) is the number concentration (particles m-3). It is fair to say that, although such concentration indices might not always be the best in relation to a given aerosol-related health risk, the choice in practice is often dictated by practical measurement considerations. A good example of this dilemma is the case of fibres. Here, because the health risk is known to be associated with very small

51

Aerosol science for industrial hygienists fibre concentrations which are usually present with much larger amounts of other non-fibrous particles, the selection and counting of such particles when viewed under the microscope is the only realistic practical approach currently available. Generally speaking, aerosol concentration is analogous to a gas density. Thus, it may be described in terms of its spatial distribution, implying in turn that it can be described mathematically by a continuous function. In reality, this view needs to be qualified since each 'point' can only be defined in the limit as the volume in space with dimensions no smaller than the mean distance between particles. For most practical purposes, it is indeed a reasonable starting assumption that the spatial distribution of concentration is continuous. But caution is required when the concentration of the aerosol of interest is very low ~ for example, in the case of an aerosol which, because of the known extreme hazard to health, is controlled to very low levels by stringent technical measures (e.g., radioactive aerosols, certain biological aerosols, asbestos, etc.). In such cases, in interpreting results for measured aerosol concentration from sampling exercises, statistical problems could arise from the lack of continuity in the aerosol's spatial distribution. This is a potential trap for the unwary industrial hygienist. How should he (or she) interpret the result in the case, for example, where an apparently large concentration may be due to the collection during an 8-hour sampling shift of a single very large, but potentially dangerous, particle?

3.5 P A R T I C L E SIZE Particle size is a property which has already been mentioned as being extremely important in virtually all aspects of aerosol behaviour. But it is a property whose definition is not always as simple as might at first appear, and can be elusive. The simplest case is that of a particle which is perfectly spherical. By definition, this has only one dimension ~ its true geometric diameter (say, d). This is what would be obtained if the spherical particle were to be sized under a microscope. But in nearly all practical situations, particles are not spherical, as shown in most of the examples given in Figures 3.4 and 3.5. For these, because of the particle geometric non-uniformity, no single geometric dimension can be assigned. Therefore, another index must be found to enable 'size' to be defined. This leads to definitions described in terms of one or more 'effective' or 'equivalent' diameters not true diameters as such but dimensions derived from knowledge of some other property (or combination of properties) of the particle. Firstly, if we were to examine a two-dimensional picture of a non-spherical particle, a 'characteristic' effective geometric diameter could in principle be identified. There are a number of ways of achieving this, as shown in Figure 3.6. For example, the Feret diameter (dr) is the width of the particle 52

Properties of aerosols

Figure 3.6.

Illustration of how the Feret and Martin diameters are defined as indices of particle size for a non-spherical particle.

contained within a pair of parallel tangents to its image when viewed in two dimensions. The Martin diameter (dM) is the length of the chord which divides the two-dimensional picture of the particle into two equal areas. Such measures of particle size cannot be unique for an individual particle, since the choice of orientation of the defining line in each instance is arbitrary. But provided that the measurement ~ notably with respect to the placement of the microscope cross-wires ~ is always made in the same direction, useful results can be obtained for ensembles of particles. However, these approaches are rarely used nowadays in industrial hygiene. In general, for more quantitative purposes, other definitions are generally more appropriate. For example, the equivalent projected area diameter (dp) is the diameter of a fictitious sphere which, in two dimensions, projects the same area as the particle in question. This too, however, is dependent on the orientation at which the particle is viewed, and so is most appropriate for ensembles of particles. But there are two other definitions which are unique for a given individual particle. First we have the closely-related equivalent surface area diameter (dA). Both dp and d A a r e relevant to many aspects of the visual and optical appearance of aerosols. Then there is the equivalent volume diameter (dv), the diameter of a sphere that has the same volume as the real particle in question, and is relevant to the drag force the particle experiences as it moves through the air. These three concepts are illustrated in Figure 3.7.

53

Aerosol science for industrial hygienists

Figure 3.7. Illustration of how the equivalent projected area, surface area and volume diameters are defined as indices of particle size for a non-spherical particle. E x a m p l e 3.2. Calculate the e q u i v a l e n t surface a r e a d i a m e t e r for a 10 I~m • 10 txm • 10 txm cube. The cube has 6 sides, each of area 100 Total surface area = 600

ixm 2

Surface area of the equivalent sphere By definition,

7rdA2 --

p~m 2

600

= ~dA2

p~m 2

600 So d A = V'

) = 13.8 ixm

7/"

E x a m p l e 3.3. Calculate the e q u i v a l e n t v o l u m e d i a m e t e r for a straight cylindrical fibre o f l e n g t h (L) e q u a l to 50 Ixm and d i a m e t e r (d) e q u a l to 2 txm.

7rd2L The fibre has volume =

= 157 ixm 3

54

Properties of aerosols rrdv3 V o l u m e of the e q u i v a l e n t s p h e r e -

By definition, 7rdv3/6 = 157 txm 3 *

So d v = 3~/' (942/rr) = 6.7 Ixm

For embodying many aspects of the airborne behaviour of particles, however, none of the above definitions of particle size is sufficient. More appropriately for many applications, d v may be combined with knowledge of particle density and shape in order to arrive at the aerodynamic diameter (dae). As will be seen as this book progresses, this last definition is one which is the most widely used in the industrial hygiene context. There are some types of particle where further considerations need to be invoked. This is the case, for example, for fibres where, for a full description of particle size, both diameter and length should be defined. Complex aggregates such as those (e.g., smokes) formed during combustion also pose special problems. As already mentioned, these are made up of large numbers of very small primary particles and the degree of complexity is such as to render difficult the definition of size in relation to any of the measurable geometrical properties like those described above. So, although aerodynamic diameter can be usefully applied to describe aerodynamic behaviour, and a geometrical diameter can be applied to describe aspects of visual appearance of individual particles or aerosols as a whole, these do not always properly convey the full nature of the particles. Here, therefore, Mandelbrot's concept of fractal geometry can convey additional information. As already stated, this applies to the properties of some types of particle which reflect the tendency to exhibit self-similar structure. This means that, when viewed at increasing magnifications, the structure of the particle appears in greater and greater detail ~ but that at each scale the structure appears to be geometrically similar. Application of fractal geometry to a particle of complex shape leads to a relationship of the form N ~- Rgf

(3.2)

where N here is the number of primary particles making up the aggregate and f is the fractal dimension. In this expression, Rg is the radius of gyration of the particle, another measure of particle size (not used elsewhere in aerosol science), given by Moment of inertia

) 1/2 (3.3)

Rg =

Mass

55

Aerosol science for industrial hygienists Typical particles in smoke aerosols generated during combustion may contain from 102 tO 105 primary particles, and f may take values ranging from 1.4 upwards. As already stated, fractal considerations are relatively new in aerosol science. However, many aerosols of the type having fractal properties (e.g., combustion aerosols) also have important occupational health implications. So it is quite likely that, one day, such fractal properties might become of more direct interest to industrial hygienists. Meanwhile, the general fractal concept is finding increasing application not only in relation to particle morphology in aerosol science but also elsewhere to the structure of turbulence, filters, the lung, and so on.

3.6 E L E M E N T A R Y P A R T I C L E SIZE STATISTICS Only rarely in practical situations ~ usually under controlled laboratory conditions ~ do aerosols exist that consist of particles of all one size. Such aerosols are referred to as 'monodisperse'. More generally, however, in workplaces and elsewhere, aerosols consist of populations of particles having wide ranges of sizes, and so are termed 'polydisperse'. For these, particle size within an aerosol needs to to be thought of in statistical terms. The importance of this, as will be seen later, is that particles falling within specific size ranges (or aerosol size fractions) may be associated with different types of health effect. The first assumption that is usually made is that, for a polydisperse aerosol, its particle size characteristics can be described in terms of a distribution function that is continuous. That is, all particles sizes within the overall range of interest are possible, although not necessarily with equal probability. For most practical purposes, a rudimentary outline of the statistics is sufficient. Consider an ensemble of particles whose sizes can be represented in terms of a single dimension (say, d). The fraction of the total number of particles with size falling within the range d to d + d d may be expressed as

dn = n(d)dd

(3.4)

where oo

f

n(d)dd = N

(3.5)

0

in which n(d) is the non-normalised number frequency distribution function sometimes referred to as the number probability density function ~ and N here is the total number of particles in the ensemble under consideration.

56

Properties of aerosols Alternatively, we may consider the particle size distribution in terms of the mass frequency distribution function, m(d), where

dm = m(d)dd

(3.6)

and O0

f m(d)dd 0

= M

(3.7)

where M is now the overall mass of particles contained in the ensemble. Similar relationships may be written for other forms of the frequency distribution (e.g., in terms of surface area). They are all interrelated. Which form is actually used in practice depends on how particle size is measured. For example, if we were to count and size particles under a microscope, then the distribution would be obtained in terms of n(d). On the other hand, if we weighed samples which have been classified and collected according to size, then the distribution would be obtained in terms of m(d). It is often helpful in particle size statistics to plot distibutions in the alternative cumulative form. For example, for the distribution of particle number this is given in terms of the number with diameter less than d, thus d

Cn(d ) - j n(d)dd

(3.8)

0 In terms of mass, the

mass

with diameter less than d is given by d

Cm(d) - I m(d)dd

(3.9)

0 Either of these can be expressed fractionally; so, for example, the of mass with diameter less than d is given by

fraction

d

f m(d)dd 0 M

57

(3.1o)

Aerosol science for industrial hygienists A typical mass distribution for a workplace aerosol is shown in Figure 3.8, both in the frequency and cumulative forms. Note here that the cumulative distribution describes the mass (e.g., in units of [mg]) contained in particles below the stated size. Since the cumulative distribution is obtained by integrating the frequency distribution, it follows conversely that the frequency distribution derives from differentiating the cumulative distribution. Thus, it is seen that the frequency distribution represents the mass fractions of particles contained within narrow size bands, and so may be expressed as d M / d d (e.g., in units of [mg ~m-1]). Figure 3.8 contains a number of important features. Firstly the mass median particle diameter (dmm), at which 50% of the mass is contained with smaller particles and 50% is contained within larger ones, can be read off directly from the cumulative plot. Secondly, the frequency distribution shown exhibits a strong degree of asymmetry such that the peak lies at a value of d which is substantially smaller than dram, and there is a long 'tail' in the distribution that extends out to relatively large particles. This characteristic is very common in practical polydisperse aerosol systems like those found in workplace

1.0r 0.8 m(d)

I 106 0.4 0.2 0

5

10

15

20

25

30

35

40

45

50

12-

Cm(d) [mg]

dm~ (or MMD) 0

5

l0

I

15

I

I

20

25

I

30

I

35

I

40

I

45

I

50

d (l~m)

Mass-based particle size distribution for a typical workplace aerosol, shown in: (a) the non-normalised frequency distribution form; and (b) the non-normalised cumulative distribution form. Note that the normalised versions of these curves are obtained if their vertical axes are divided by the total mass sampled. Figure 3.8.

58

Properties of aerosols e n v i r o n m e n t s . Very often, the overall distribution may be r e p r e s e n t e d to a fair first a p p r o x i m a t i o n by the log-normal m a t h e m a t i c a l function M

{

m(d) =

exp

(lnd-lndmm) 2 ) -

(3.11)

d V(2r;) ln~rg

2(lnerg)2

w h e r e O'g is the geometric standard deviation, reflecting the width of the distribution. This is given by d84% O'g - "

dmm =

dmm

(3.12)

d16%

Example 3.4. For the particle size distribution shown in Figure 3.8, estimate by graphical methods the median particle diameter and the geometric standard deviation. Firstly, the cumulative distribution function is normalised and plotted on log-probability axes. This is shown in Figure 3.9 Here we see that the plot is linear, confirming that the particle size distribution is indeed log-normal *

It may be read directly from the graph that Also from the graph,

d84 %

= 21 p~m and

dram --

dl6 % =

9 ~m

4.1 i~m.

From Equation (3.12) 21 - 2.30

O'g

9 or

9 = 2.20

O'g -"

4.1 indicating an accuracy which, from such graphical methods, is usually sufficient for practical purposes.

For a perfectly m o n o d i s p e r s e aerosol, O'g = 1. M o r e typically for aerosols f o u n d in the workplace e n v i r o n m e n t , O'g ranges from a b o u t 2 to 3. T h e

59

Aerosol science for industrial hygienists

E

,.,,,.,,i

l

d84~, = 21 I~m

9

dmm = 9 I~m di6 % = J [ 4.1 I~m / o-

o~ I

~1~1 i i/ 2

5

! 11 1

10 20 50 80 90 95 98 Percentage of mass with diameter less than stated d.

Figure 3.9. Cumulative particle size distribution (by mass) for a typical workplace aerosol, plotted on log-probability axes to illustrate the property of log-normality.

property of log-normality (or even a reasonable approximation to it) which is so often found in practice in workplace aerosols provides some additional useful aspects. In particular, it enables conversions between relationships for distributions based on particle number, mass, surface area, and any other aerosol property. As long ago as 1929, Hatch and Choate developed a set of equations for this purpose, each equation having the form qMD - NMD exp(qln2org)

(3.13)

where NMD is the number median particle diameter and qMD is the median diameter weighted by dq. We note, for example, that the link between particle number and volume (and hence mass) for particles of diameter d is d3, based on simple geometrical considerations. This, therefore, leads to the choice of q = 3 if we wish to use Equation (3.13) to convert distributions from number to mass . . . . . . . . . . , for particle surface area, q = 2. Note that, in such conversions, O'g does not change. The appearance of a log-normal particle size distribution is usually associated with a single aerosol generation process. But in many workplaces there may be more than one type of aerosol. In such cases, therefore, it is not unusual to find two or more particle size distributions superimposed. These are referred to as multi-modal. Figure 3.10 shows an example of a clearly bimodal aerosol like that found in an underground mining environment where there may be both relatively coarse dust (generated by the extraction process itself) and relatively fine diesel particulate (generated by underground transportation) (e.g., Cantrell and Rubow, 1990). We have already mentioned that fibrous aerosols represent a special

60

Properties of aerosols 0.3 I m(d)

Dust

0.2 -

l 0.1

0 0.01

0.1

1.0 d (Izm)

I I00

I0

Figure 3.10.

Illustration of the frequency distribution (by mass) for a typical bimodal aerosol (e.g., like that found in underground mining where diesel vehicles are being used).

case in o c c u p a t i o n a l health terms. B e c a u s e of the i n t r o d u c t i o n of the additional d i m e n s i o n , particle size statistics for such aerosols n o w r e q u i r e the c o n s i d e r a t i o n of three distributions; d i a m e t e r (d), length (L) and aspect ratio (d/L). E x a m p l e s for a typical s a m p l e of w o r k p l a c e a i r b o r n e asbestos dust are s h o w n in Figure 3.11, where each of the two c u m u l a t i v e distributions s h o w n is seen to be log-normal.

|.0

100

'

IO

E :L 0.1

0.01

I

I

I

5

20

50

I

i

80 95 Percentage of particles less than stated size

E :L

l

~9

Figure 3.11. Illustration of the cumulative size distributions (by number) for both length and diameter for airborne asbestos fibres. The ones shown are typical of an amosite asbestos fibre generated in a laboratory chamber during animal (rat) inhalation experiments (e.g., reference aerosol as defined by the Union Internationale Contre le Cancer [UICC]).

61

Aerosol science for industrial hygienists

3.7 E L E C T R I C A L P R O P E R T I E S In aerosol science, both in industrial hygiene applications and elsewhere, the electrical properties of aerosols have frequently tended to be ignored, or ~ at b e s t occasionally invoked to provide qualitative explanations of unexpected, or otherwise implausible, observations. However, in recent years, a growing body of experimental work has indicated that the state of static electrification (i.e., particle charge) in an aerosol may be of significant practical relevance in a number of industrial hygiene areas. For example, it has now been established that it can affect the behaviour of particles in the lung after inhalation (leading to enhanced deposition in some cases). It can also influence sampling and filtration. In recent years, instrumentation has been developed which has enabled the measurement of the electrical properties of workplace aerosols, in terms of both the magnitude of the charge carried by individual particles and how that charge is distributed between positive and negative polarity and over populations of particles of different sizes (e.g., see Figure 10.17 in Chapter 10). An example of a particle charge distribution from an aerosol typical of those found in workplaces is shown in Figure 3.12. From comprehensive measurements made on workplace aerosols of widely-varying types (Johnston et al., 1985), two main features were found which appear to be common to all the workplaces surveyed, irrespective of the type of aerosol generated and how it was generated. Firstly, each particle is charged either net positive or negative and, for the aerosol as a whole, the charges on individual particles are distributed almost symmetrically between positive and negative

I:I

-100

I

0

!00 Plus

Minus

qle Figure 3.12. Illustration of the electric charge distribution for particles of a given size in a typical workplace aerosol, in which the vertical axis represents the probability of finding a particle carrying charge q/e (where e is the charge per electron, 1.6 • 10 -19 C).

62

Properties o f aerosols 100o

oy IOO

Z/

10-

1

0.1

J

l

1.0

.,

I

10

d (~m) Figure 3.13. Typical data for the median magnitude of particle charge, expressed in terms of Iqm/el as a function of particle size (d) for a typical workplace aerosol (where e is the charge per electron, 1.6 • 10-19 C). The data shown are based on results obtained at a rock crushing operation in a quarry (from Johnston, A.M. et al., Annals of Occupational Hygiene, Copyright 1985, adapted by permission of the British Occupational Hygiene Society). polarity. Secondly, the median magnitude of charge per particle for each given workplace aerosol (Iqml) may be represented by the simple empirical relation (see Figure 3.13) qlTl = Ad n

(3.14)

e where n is a constant coefficient and, if particle diameter (d) is in [~m], A is the number of charges equivalent in magnitude to one electron (e) carried by a 1 ~m-diameter particle. From the above, a given aerosol might appear to be of neutral polarity even though individual particles might be very highly charged. In many aspects of particle behaviour (including effects of particle deposition in the lung), it is individual particle charge which is important. But this notwithstanding, in practice there is no such thing as an aerosol where all the particles are truly neutral. For an aerosol exposed to the naturally-occurring airborne charges of both polarities (e.g., air ions, cosmic rays, etc.), such 'neutrality' is defined in terms of Boltzmann equilibrium, for which, in Equation (3.14), A ~ 3 and n ~ 1/2 (Liu and Pui, 1974). This is what would be achieved for an aerosol that has been in existence for a long time (which, under normal conditions, would in practice be upwards of about 1 hour). Thus it is seen that, for most long-lived aerosols encountered in the ambient outdoors atmosphere, they are likely to be in a state of Boltzmann neutrality. In workplaces, however, due to the more confined environment together with the turnover of air due to ventilation, aerosols to which workers are exposed are 'fresher'. So 63

Aerosol science for industrial hygienists

particles will usually carry charge above that corresponding to Boltzmann equilibrium. The level of charge is associated with the process with which they were made airborne. For dry aerosols produced by mechanical dispersion, charging occurs as the result of the creation of new surfaces, contact charging (the making and breaking of contacts) and triboelectric charging (associated with friction). For liquid droplets, charging may be the result of the reorganisation of the surface energy of the original bulk liquid. For Table 3.2. Typical summary data for particle charge for a range of typical workplace aerosols, based on the coefficients A and n in E q u a t i o n (3.14) for aerosols with isometric particles and on the coefficient or for fibrous asbestos aerosols in E q u a t i o n (3.15). Industry

Location

Jute

Cotton Flax

Glass fibre

Chemicals

Rubber Batteries Quarry Coal mine Asbestos 1

Asbestos 2

A

n

Batching Spreading Carding Drawing Spinning Winding Weaving Weaving Hackling Carding Weaving Spinning Winding Slivering Spinning Winding Weaving Silica A Silica C Silica D Mixing Rolling Oxide production Soldering Primary crusher Secondary crusher Return roadway

22.3 27.8 28.6 20.7 13.2 13.0 35.3 33.2 22.6 49.8 29.2 10.3 13.7 3.6 4.8 4.3 4.4 11.2 10.1 24.1 6.2 5.1 2.9 2.1 22.0 4.0 25.0

1.03 0.80 1.19 1.18 1.44 1.16 1.25 1.24 1.77 1.12 1.23 1.84 1.84 1.34 1.04 1.19 1.44 1.01 1.21 0.72 1.69 1.91 0.90 0.98 1.50 1.43 1.20

Carding Spinning Weaving Carding Spinning (dry) Spinning (wet) Weaving

13.0/~m length 10.1 9.6 8.4 11.0 4.9 6.0

64

Properties of aerosols

such aerosols, it is reasonable to expect that A in Equation (3.14) will be substantially greater than 3. In fact, values as high as 40 have been found for workplace aerosols, tending to be higher the greater the amount of energy that went into the aerosol generation process. In addition, since the ability of a particle to retain electrical charge should be approximately proportional to its surface area, we might expect n ~ 2. In practice, values were found ranging from 1.2 to 2.5, tending on the whole to be less than 2. By way of illustration, Table 3.2 shows some results for the states of static electrification of some typical workplace aerosols, most of them expressed in terms of the quantities A and n contained in Equation (3.14) (from the Johnston et al. study). Example 3.5. For particles made airborne during the carding process in the manufacture of flax textiles, estimate the median magnitude of charge (in Coulombs) carried by particles of diameter 7 Ixm. C h a r g e p e r p a r t i c l e is given by E q u a t i o n (3.14)

F r o m T a b l e 3.2, n o t e t h a t A = 49.8 a n d n = 1.12. So

qm = 49.8 x 71-12 e

N o t e that e l e c t r o n i c c h a r g e , e = 1.6 x 10 -~9 C

T h u s Iqm]

= 49.8 x 71-12 •

1.6 x 10 -19 C

= 7.04 x 10-17 C

Once again, fibres (e.g., asbestos) present a special case which is worthy of mention. Results have shown that, as a result of their large overall physical dimensions, the level of charge carried by long, thin fibres can be significantly larger than for isometric particles of corresponding aerodynamic diameter. For such particles, charge per particle has been found to be relatively independent of fibre diameter, but increases approximately linearly with fibre length. Thus qm -

cr L

(3.15)

e

where cr here is the charge per unit length of fibre. This is a feature which could have significant effects on some aspects of airborne behaviour, including deposition in the lung after inhalation (see Chapter 6). Typical results for workplace aerosols are shown in Table 3.2. 65

Aerosol science for industrial hygienists 3.8 M I N E R A L O G I C A L A N D C H E M I C A L P R O P E R T I E S As already described, the raw material from which aerosols are derived, and the processes by which many are formed and subsequently evolve, are often very complicated. So too, therefore, are their chemistry and mineralogy. Particle composition features significantly in the cellular reactions that are stimulated once an aerosol is inhaled and particles subsequently come into contact with biological tissue. Therefore, it clearly has an important bearing on possible toxic effects to exposed workers. Discussion of particle composition brings us close to the question of toxicity, a subject which is usually considered to lie outside aerosol science itself. The mechanisms by which inhaled particles of a given substance can provoke biological responses are varied. Even a relatively insoluble particle can appear to the cells of the lung as a undesirable 'foreign body' and so stimulate a whole range of defence processes which relate to mineralogical composition as well as to other factors such as particle size and morphology. Here the surface properties are particularly relevant so that one insoluble mineral particle might appear to the cells as more toxic than another. For example, on the one hand, crystalline silica (e.g., quartz) has been shown to be relatively toxic to cells, although uncertainty remains as to the exact mechanisms. Here, biological responses to the presence of a quartz particle can be so severe as to lead to permanent pathological changes in the lung (e.g., fibrosis). On the other hand, other insoluble materials such as titanium dioxide and fused alumina are considered to be relatively inoccuous. Other substances may be soluble in the lung environment. For these, when they come into contact with the lung, material be transferred into the blood whereupon it may be transported to other parts of the body, possibly leading to adverse health effects in regions remote from the lung itself (e.g., kidneys, liver, bone, etc.). Some metal-containing aerosols (e.g., lead, cadmium) come into this category. In workplaces, aerosols are often made up of complex mixtures of compounds and mineral types. Therefore, even though the mass concentration, size distribution and morphology of particles in the aerosol as a whole may be known, it often remains difficult to assess exposure to such aerosols in a truly meaningful way. Ultimately, the choice of a limit value which is assigned to a particular workplace aerosol as a basis for controlling exposures to within 'safe' limits comes down to considerations of composition. Here, therefore, we have the concept of an aerosol fraction in terms not just of its particle size distribution but also of its chemical and mineralogical composition. Again, the case of quartz is a good example to illustrate some of what is involved since it is frequently present ~ together with other minerals in workplace dusts, in particular those found in the extraction industries. Even in relatively small proportions (e.g., 10% or less) it can present a significant risk to the health of exposed workers if the overall dust level is high enough. So, in 66

Properties of aerosols

assessing exposure in a manner which reflects the magnitude of that health risk, it is important to be able to measure not just the overall dust in the fine particle size fraction relevant to particle deposition in the alveolar region of the lung (the region at risk) but also the quartz content within that fraction. Another complicating factor, however, is that the quartz content of the dust does not behave in isolation. Evidence from epidemiology and animal inhalation studies has suggested that the response may be modified by the presence of certain other minerals. Metal-containing aerosols represent another interesting class of problem. Whereas it is currently common practice to determine exposure in terms of the airborne concentration of the metal atoms that are present (e.g., by atomic absorption spectrophotometry), it is known that certain molecular forms are more harmful than others. For example, in the production of nickel, epidemiology has suggested that although water soluble, sulphidic and oxidic forms might be associated with lung and possibly nasal cancer, there is no such evidence for the metallic form (Doll et al, 1990). In such cases, the question of which chemical species is the most relevant to ill-health is therefore an important issue. It is a significant challenge to analytical chemistry to develop quantitation techniques that can be used for aerosol speciation in the industrial hygiene setting.

3.9 B I O L O G I C A L P R O P E R T I E S The science of aerobiology ~ the study of airborne particles of biological o r i g i n - is one of the oldest branches of aerosol science. But it is perhaps fair to say that, in the development of what we might now call 'mainstream' modern aerosol science (in particular physics and chemistry), aerobiology has tended to be excluded. Or perhaps it was not specifically included. So it has continued over the last two decades to develop somewhat independently. Now, however, in the mid-1990s, there is sharply renewed interest in the whole range of what we now term as 'bioaerosols' ~ particularly by those concerned with occupational and environmental health. This is stimulating increased involvement in aerobiology by aerosol scientists and by industrial hygienists and other health professionals, bringing to the subject a wider range of backgrounds and expertise. So a new multidisciplinary approach to the problems associated with bioaerosols is emerging. This in turn is bringing the subject of aerobiology back into the mainstream of aerosol science. Bioaerosols are complex and highly diverse. They include particles in the forms of bacteria, viruses, fungal spores, endotoxins, allergenic and toxic substances of plant and animal origin, protein aerosols, and others. Their airborne concentration may be expressed in more than one way, depending on the type of particle. Bacteria and fungal spores, for example, may be

67

Aerosol science for industrial hygienists expressed in terms of the number of bacterial entities of a given type per unit volume of air (i.e., number m-3). On the other hand, such particles, if they are viable, may be expressed in terms of their ability to reproduce that is the number of 'colony-forming units' per unit volume of air (i.e., cfu m-3). For endotoxin and allergenic material, it is appropriate to express concentration in terms of the mass of the active component per unit volume of air (i.e., >g m-3). Such diversity presents many difficulties in sampling and measurement methodology for exposure assessment. This is an area which has not been well developed in relation to industrial hygiene. However, the new awareness of the importance of bioaerosols in many occupational settings is leading to new efforts to develop appropriate procedures. This is discussed in Chapter 9. A comprehensive review of these aerosols, their properties and their health effects in occupational settings like those found in agriculture, sawmills, textiles manufacture, meat and other food processing, biotechnology, research laboratories, waste disposal, construction, health care, etc. has been given by Lacey and Dutkiewicz (1994). A short summary is presented in Table 3.3. Some examples are shown in Figure 3.14.

Table 3.3. Some bioaerosols found in occupational environments. Category found (particle size)

Type/description

Occupational setting

Viruses etc. (<0.1 ~m)

Anthropogenic Zoonotic

Health care, education, Agriculture (livestock)

Bacteria (0.5-5 ~m)

Zoonotic gram-negative

Agriculture (livestock, contamination by rodents)

Non-zoonotic

Agriculture (grain), textiles manufacturing (as sources of endotoxin)

gram-negative Gram-positive

Agriculture (livestock)

Mycobacteria

Health care

Fungi ~1-50 ~m)

Filamentous

Agriculture (crop harvesting and storage), municipal waste, cotton manufacture, wood manufacture, office settings

Endotoxins

Lipopolysaccharides (LPS) (associated with bacteria)

(0.03-0.05 ~m)m

Agriculture (grain)

Pollens (10-200 ~m)

Agriculture, outdoor settings, etc.

68

Aerosol science for industrial hygienists

Figure 3.14a

Figure 3.14b

69

Properties of aerosols

Figure 3.14d Figure 3.14. Photographs of some bioaerosols typical of those experienced by agricultural workers: (a) Aspergillus sporaphore collected on stage 1 of a cascade impactor (see Chapter 9) exposed to aerosol generated from mouldy hay; (b) Actinomycete spores collected on stage 2 of a cascade impactor exposed to aerosol generated from mouldy hay; (c) typical summer dry daytime air spores (Cladosporiurn and Epicoccum); (d) typical grain dust spores (Epicoccum and Alternaria). Note that the scales shown are approximate. (Photographs courtesy of John Lacey, Rothamstead Experimental Station, Harpenden, Herts, England, U.K.).

70

Properties o f aerosols REFERENCES Cantrell, B.K. and Rubow, K.L. (1990). Mineral dust and diesel exhaust aerosol measurements in underground metal and non-metal mines. In: Proceedings of the Seventh International Conference on the Pneumoconioses (Pittsburg, August 1988), NIOSH Publication No. 90-108, pp.651-655. Doll, R. and 20 co-authors (1990). Report of the International Committee on Nickel Carcinogenesis in Man. Scandinavian Journal of Work Environment and Health, 16 (special issue), pp. 1-82. Friedlander, S.K. (1977). Smoke, Dust and Haze. John Wiley and Sons, New York. Hammond, C.M., Heriot, N.R., Higman, R.W., Spivey, A.M., Vincent, J.H. and Wells, A.B. (1985). Dustiness estimation methods for dry materials. British Occupational Hygiene Society (BOHS) Technical Guide No. 4, Science Reviews Ltd., Northwood, Middlesex. Hatch, T. and Choate, S.P. (1929). Statistical description of the size properties of non-uniform particulate substances. Journal of the Franklin Institute, 207, 369. Heitbrink, W.A., Todd, W.F., Cooper, T.C. and O'Brien, D.M. (1990). The application of dustiness tests to the prediction of worker dust exposure. American Industrial Hygiene Association Journal, 51,217-223. Hinds, W.C. (1982). Aerosol Technology. John Wiley and Sons, New York. Johnston, A.M., Vincent, J.H. and Jones, A.D. (1985). Measurements of electric charge for workplace aerosols. Annals of Occupational Hygiene, 29, 271-284. Lacey, J. and Dutkiewicz, J. (1994). Bioaerosols and occupational lung disease. Journal of Aerosol Science, 25, 1371-1404. Liu, B.Y.H. and Pui, D.Y.H. (1974). Equilibrium bipolar charge distribution of aerosols. Journal of Colloid and Interface Science, 49, 305-312. Lyons, C.P. and Mark, D. (1994). Development and testing of a procedure to evaluate the dustiness of powders and dusts in industrial use. Report LR 1002, Environmental Technology Executive Agency of the U.K. Department of Trade and Industry, London. Mandelbrot, B.B. (1983). The Fractal Geometry of Nature. Freeman, New York.

71

CHAPTER 4

The motion of airborne particles 4.1 I N T R O D U C T I O N The physical processes governing the motion of airborne particles are highly relevant to workplace aerosols. Therefore, a working appreciation of, for example, the mechanisms underlying the transport and deposition of particles in ventilation ducts, deposition onto workplace surfaces, inhalation into and deposition inside the human respiratory tract, sampling and filtration, and so on, is clearly important to the industrial hygienist. This chapter sets out to review the important elements of the mechanics of individual particle motion. It begins with the interaction by a moving particle and its surroundings, leading to the concept of a drag force. This force represents the resistance to particle motion which in turn is driven by external and inertial forces. It is the competition between these two types of force (i.e., 'driving force' and 'resistance force') which governs the outcome of the motion in terms of transport and deposition. These concepts will be developed in this chapter, and in later chapters will be extended to practical industrial hygiene situations. For a seminal treament of the subject of the mechanics of aerosols, readers are strongly recommended to refer to the work of Fuchs (1964).

4.2 D R A G F O R C E ON A P A R T I C L E

Drag When a body immersed in a fluid moves relative to the fluid, it experiences a force associated with the resistance (by the fluid) to its relative motion. The picture shown in Figure 4.1 applies directly to a small aerosol particle moving in air. The drag force acting on the particle may be derived from solutions of the Navier-Stokes equations for the air flow in the particle-fluid system in question. Mathematically this is achieved by determining the distributions of the local static pressure (normal) and viscous (tangential) forces over the surface of the particle. For very slow ('creeping') flow approaching at velocity

72

The motion of airborne particles

Figure 4.1. Schematic to show the forces acting on a spherical particle moving

in air.

v and passing over a sphere of diameter d, integration of these forces yields the well-known Stokes' law for the overall drag force F D = - 37rd ~v

(4.1)

where the Reynolds' n u m b e r for the particle dvpair Rep =

(4.2)

is very small (Rep < 1). Particles for which this applies are sometimes referred to as 'Stokesian'. In Equation (4.1), the minus sign indicates that the drag force is acting in the direction opposing the particle's motion. Strictly, this expression should be modified by three factors. The first derives from the fact that, in reality, the air surrounding the particle is not continuous but is made up of individual gas molecules which are in

Figure 4.2.

Diagram to illustrate the phenomenon of 'slip', showing the contrasting cases where (a) d > > mfp, and (b) d < mfp.

73

Aerosol sciencefor industrial hygienists random thermal motion. On the one hand, if the particle is large enough (i.e., much greater than the mean free path between gas molecules, mfp), it experiences the air as a continuum. This is because it cannot 'recognise' individual collisions with gas molecules. On the other hand, for particles which are so small that d is of the same order of magnitude or less than mfp, the nature of particle motion may be envisaged as 'slip' between successive collisions with the gas molecules. These contrasting scenarios are illustrated in Figure 4.2. To account for particle slip, it is now necessary to modify the expression for the drag force by the introduction of a correction factor know as the Cunningham slip correction factor (Ccun). Thus

- 3~rd~xv FD =

(4.3)

Ccun where, according to Cunningham (1910)

2.52mfp ) Ccun

-

1 +

(4.4)

d

(.)

(055d)

for d down to 0.1 ~m. The more complicated empirical expession

Ccu n = 1 +

{2.514 + 0.8exp

d

}

(4.5)

mfp

has been shown experimentally to work well for d down to 0.01 ~m. Typically, as stated in Chapter 2 for air at STP, mfp = 0.066 p,m. So it is seen that consideration of the effects of the slip correction starts to become important for aerosol particles with d less than about 1 ~m. This means that, for many practical industrial hygiene situations involving mineral dusts and other such relatively coarse aerosols, it may be neglected. But for other, finer, aerosols (e.g., welding fume, diesel fume), it must be taken into account. The second modifying factor concerns departures from Stokes' law at Rep values exceeding about 1. To discuss this, it is useful here to introduce the concept of the dimensionless drag coefficient (C D). Thus for a spherical particle

CD

(

projected area of the sphere

) (4.6)

( 1/2 Pair V2) 74

The motion of airborne particles

where the denominator on the right-hand side serves to normalise the force per unit frontal area of the sphere with respect to the velocity pressure in the approaching flow. For Stokesian particles, Equation (4.6) gives 24 (4.7)

C D --

Re P From Equations (4.4) or (4.5) - (4.7), the general expression for drag force is therefore FD =

( coRep) ( 24

)

(4.8)

Ccu n

where the term inside the first bracket on the right-hand side is the correction to allow for 'non-Stokesian' behaviour. The general shape of the relationship between C D and Rep over a wide range of Rep for spherical particles is shown in Figure 4.3. The Stokesian region, corresponding to Equation (4.7) is described by the straight portion (on the log axes shown) to the left. The curve tends to become less steep for Re -> 1, levelling out for Re -> 2000. In the intermediate range, the empirical expression CD =

(24) ( RERJ3) ~ Re

{1 +

}

6

has been found to agree with experimental data to within 2%.

Figure 4.3. Curve showing the relationship between the drag coefficient (CD) and Reynolds' number (Rep) for a spherical particle. Note the linear region (on the log axes) for the Stokes region where Rep < 1. (From Vincent, J.H., Aerosol Sampling: Science and Practice, Copyright 1989, adapted by permission of John Wiley and Sons Limited).

75

(4.9)

Aerosol science for industrial hygienists So far we have considered only spherical particles. However, experience has shown (as described in Chapter 3) that particles like those encountered in the industrial hygiene context are rarely so idealised. This, therefore, leads to the third modifying factor. It is reasonable to expect that particle shape will influence particle motion, and so a further correction to the drag equation is needed. Now the drag force is written in the form CDRe p ) D

m

3~rdv~xV+

__

24

(4.10) Ccun

where d is now replaced by d v, the equivalent volume particle diameter as described earlier, and + is the dimensionless dynamic shape factor. The latter is usually determined experimentally from observations of the falling speeds of particles settling in air under the influence of gravity. It should be noted that + for an irregular particle will depend on the particle's orientation with respect to the flow during its motion. Since this is usually difficult to define in an experiment, then d~ is usually obtained for motion which is averaged uniformly over all possible orientations. Some values of + for typical workplace aerosol particles are shown in Table 4.1, where it is seen that + ranges from 1.0 for perfect spheres to about 1.4 for non-spherical particles of quartz, and 2.0 for platelet-shaped particles of talc. In Equation (4.10), Ccun should also be adjusted for particle shape, but the value of Ccu ~ for a spherical particle serves reasonably well for most practical situations.

Table 4.1. Values for the dynamic shape factor (+ of typical aerosol particles. Shape/type

Shape factor, +

Sphere Fibre (L/d=4)

1.00 1.32 (axis perpendicular to motion) 1.07 (axis parallel to motion) 1.36 1.04- 1.49 2.04

Quartz dust Fused alumina Talc (platelet)

In summary, Stokes' law describing the aerodynamic in air. However, a number be important under certain

has been identified as the basic starting point for drag forces acting on an aerosol particle moving of corrections have been identified which could conditions. These are:

76

The m o t i o n o f airborne particles

'slip' for d <_ m f p , non-Stokesian flow for Rep >_ 1, and non-spherical particle shape, respectively. These corrections should never be completely ignored. But in many industrial hygiene situations, they may be considered small enough that to a first approximation ~ Stokes' law is a reasonable working approximation.

4.3 P A R T I C L E M O T I O N

Equations of motion

As in the case of fluid mechanics, the starting point for all considerations of particle transport is again Newton's Second Law (mass • acceleration = net force acting). This leads to a general equation of particle motion. For the forces acting, the drag force describing the resistance of the fluid to the particle's motion has already been described. In addition, there may be an external force (e.g., gravity, electrical, etc., or some combination of forces), the effect of which is to generate and sustain particle motion. So long as the particle is in motion relative to the fluid, the drag force will remain finite. The proper relationship for describing the particle motion is a vector equation, embodying the relative motion of the air and the particle and the forces acting for each of all three available dimensions. The 'shorthand' vector notation used for expressing this mathematically appears deceptively simple. But the resultant set of equations which need to be solved for particle motion in specific cases can become quite complicated. However, the important principles involved can be illustrated by reference to some simple - - but nonetheless relevant - - examples.

Motion under the influence of gravity

The first example is for a particle falling still air. Here, as shown in Figure 4.4, the linear motion in the single vertical (y) illustration we neglect the effects of slip, particle non-sphericity. Then the equation

under the influence of gravity in problem reduces to one involving dimension. For the purposes of departures from Stokes' law and of particle motion is given by

dpy m

= -

3TrpMVy + ( m g -

dt 77

FB)

(4.11)

Aerosol science for industrial hygienists Drag force

Bouyancy force ('upthrust')

elocity, v

I y-direction

Gravitational force

Figure 4.4.

Schematic to illustrate the problem of a particle falling in air under the influence of gravity, indicating the forces acting.

where Vy is the particle's velocity in the downwards y direction, m is its mass and g is the acceleration due to gravity. The term F B is included to allow for the upwards buoyancy force arising from the mass of air that is displaced by the particle (as described by Archimedes' principle). But since the density of air (1.29 kg m -3 at standard temperature and pressure) is so much less than the density of a particle (usually greater than 103 kg m-3), this term may usually be neglected for practical purposes. With this in mind, Equation (4.11) may be re-organised to give dvy + dt

(Vy) ~

-g=0

(4.12)

T

where d2p -r =

(4.13) 18Ix

78

The motion of airborne particles in which p is particle density. In E q u a t i o n (4.13), closer inspection of its units reveals that -r has dimensions of time, the significance of which we shall see shortly. E q u a t i o n (4.12) is a simple first-order linear differential e q u a t i o n of the type familiar in almost all areas of science and engineering. Its solution is very straightforward. In terms of the particle velocity at time t, it has the well-known exponential form

Vy-- g-r

[

1 - exp

(

(4.14)

- -T

for the case where the particle starts from rest (Vy = 0) at time t - 0. This shows that particle velocity under the influence of gravity tends exponentially towards a terminal value, as shown in Figure 4.5. The sedimentation or falling speed is given by v ~ - g-r

(4.15)

for which the corresponding mechanical mobility (B), defined as the velocity per unit of applied mechanical force, is Vs

B -

(4.16)

mg As stated above, E q u a t i o n (4.15) relates for illustrative purposes to the simple case where slip, non-Stokesian conditions and particle non-sphericity are neglected. More generally, we should write

lYs = g T

v y

0.63

9

vs

I

/ 0

T t (secs)

Figure 4.5.

Form of the function describing a particle's settling (or falling) speed (Vs) as a function of time (t), starting from rest at t = 0.

79

Aerosol science for industrial hygienists

v~ -

(Ccun)(24) +

g-r

(4.17)

CDRep

R e g a r d l e s s of particle size (but again u n d e r the b r o a d g e n e r a l a s s u m p t i o n of S t o k e s i a n c o n d i t i o n s ) , particle velocity rises to 6 3 % of its final t e r m i n a l value at t = -r. T h e q u a n t i t y -r is t h e r e f o r e r e c o g n i s e d as b e i n g the t i m e c o n s t a n t (or, inversely, the rate) of the e x p o n e n t i a l process. F o r the e x a m p l e given, it is s m a l l e r the faster the particle r e a c h e s its t e r m i n a l velocity. M o r e g e n e r a l l y , f r o m E q u a t i o n (4.13), it is seen to be a f u n d a m e n t a l p r o p e r t y of the particle itself, relating to the rate at which the particle c o m e s into d y n a m i c a l e q u i l i b r i u m with its s u r r o u n d i n g s a n d the forces acting (i.e., w h e r e it is n e i t h e r a c c e l e r a t i n g or d e c e l e r a t i n g ) . A e r o s o l scientists t h e r e f o r e r e f e r to it as the particle relaxation time. It is a v e r y i m p o r t a n t q u a n t i t y which occurs time a n d t i m e again in c o n s i d e r a t i o n s of particle m o t i o n . Example 4.1. Calculate the falling speed of a spherical particle of diameter 0.01 txm and density 1 x 103 kg m -3. Assume that the viscosity of air is 18 x 10-6 N s m -2 (as mentioned in Chapter 2), and that the mean free path for molecular motions is 0.066 txm Also note that, for a spherical particle, the dynamic shape factor, r = 1 Also note that the Cunningham correction factor for a particle of the size indicated is given by Equation (4.5), thus

C = 1+

(O.O66)(

I).55 x 0.01

2.514+0.8e(

0.066

0.01 = 22.45 Also note that Stokes' law is obeyed for such a small particle Now from Equation (4.17) 22.45 x (10-8)2[m 2] x 103[kg m -3] • 9.81[m s -2] VS ~-

18 X 18 X 10-6[N s m -2] in SI units, so that *

vs = 6.80 x 10- S m s -1

80

The motion of airborne particles E x a m p l e 4.2. C a l c u l a t e t h e falling s p e e d of a q u a r t z p a r t i c l e w i t h e q u i v a l e n t v o l u m e d i a m e t e r 150 Ixm a n d d e n s i t y 2.7 • 103 kg m -3. .....

Note here that, for such a large particle, we may neglect the slip correction, so that Ccu n = 1. But it is likely that we may have to make a correction for non-Stokesian conditions First, for a particle of quartz, Table 4.1 gives the dynamic shape factor cb = 1.36 In Equation (4.17) we see that we cannot assign Rep until we know the falling speed. Therefore, since vs depends on Rep while Rep depends on vs, the calculation is less straightforward than for the low Rep case. First, begin by assuming that Stokes' law applies. Then, initially, we get (150 x 10-6)2[m 2] x 2.7 x 103[kg m -3] • 9.81[m s -2] k'sl

=

18 x 18 x 10 -6 x 1.36[N s m -2] = 1.35 m s -1 Now check Rep, recalling that this is given quite accurately for workplace conditions by 7 • 104dVsl . We get Rep = 7 x 104 x 150 x 10-6[m] x 1.35[m s -1] = 14.18 = 14 which is well outside the range for Stokes' law to apply So now recalculate a new value for v s using Equation (4.17) with this value for Rep. H e r e

(143)

we need to calculate a value for the drag coefficient from Equation (4.9), thus

CD =

{1 +

~

} = 3.37

6

(24)

We can use this to calculate a new value of v S, giving

Vs2 = Vsl x

=1.35 x 0.5

(3.37 x 14) = 0.68 m s-1

81

Aerosol science for industrial hygienists Now we have to go back and check Rep again, thus R e p - - 7 • 104 • 150 • 106[m] • 0.68[m s -1] = 7.08 -~ 7 Now

CD=

~

{1 +

~

} = 5.52 6

So

Vs3

= Vsl •

(24)

= 1.35 • 0.62

5.52 • 7 = 0 . 8 4 m s -1 At this point, we are beginning to see that the process of converging towards the final m and correct m value for vS is iterative. So we may continue with the cycle of calculating vs,

Rep, calculating

calculating

CD, then recalculating v~, etc. until we get c o n v e r g e n c e -

that

is, where the calculated value of v~ does not change from one cycle of the calculation to the next. Conducted in the manner indicated above, it is clear that this can become laborious. However it becomes relatively simple with modern spreadsheets. For the present example, it so happens that the number of iterations to achieve this is quite small; namely (following the above nomenclature): Vs4

=

0.79 m s -1

Vs5 = 0.80 m s-1 Vs6 =

0.80 m S-1

We see that convergence has been achieved. So, for the example given, the final answer is *

vs = 0.80 m s -1.

Full calculations for vs for particles covering wide ranges of size and shape have

been

droplets

reported

with

p -

in t h e

literature.

103 k g m - 3

are

Some

shown

for falling speed shows the values calculated l a w is a s s u m e d ,

as g i v e n b y E q u a t i o n

examples in T a b l e

water

The

first c o l u m n

if slip is n e g l e c t e d

and Stokes'

(4.15). The second

82

for spherical

4.2.

column

shows the

The motion of airborne

particles

Table 4.2. Calculated values for the falling speed of a spherical particle of water, firstly assuming no slip and Stokesian conditions, then allowing for both slip and non-Stokesian conditions. v~ ( m s - ' )

Particle diameter (for spheres with p = 10 3 kg m -3)

Stokes' law, no slip

0.01

3.03 x 10 -'~

0.1

3.03 •

1

3.03 x 10 . 5

10

3.03 •

100 1000

10 -7

Actual

6.74 x l O - a 8.68 •

10 -7

3.50 x 10-5

10 -3

3.05 •

10-3

3.03 x 1 0 - '

2.48 •

10-~

3.03

3.86

actual value when the Cunningham slip correction is included and when the correction for non-Stokesian conditions is applied. Inspection of these results show that there is quite a wide range (--- 2 - 50 I~m) over which Stokes' law provides v~ to within about +10%. This is sufficient for many practical industrial hygiene considerations. Motion under electrical forces The same basic ideas may be applied to particles moving under the influence of other types of force, the only difference being in the way that the force itself is specified. Electrostatic precipitation is another example which is relevant to industrial hygiene. Here the electrical force requires that the particle carries some electric charge, either that which it has acquired during the process of being made airborne (as discussed in Chapter 3) or

Charge, q ,,

FE~

FD Velocity, v E

Figure 4.6.

Schematic to illustrate the problem of a charged spherical particle moving in an externally applied electric f i e l d .

83

Aerosol science for industrial hygienists d u r i n g s u b s e q u e n t artifical c h a r g i n g (as w o u l d be the case in an e l e c t r o s t a t i c p r e c i p i t a t o r air cleaning device). F o r a particle c a r r y i n g c h a r g e q, the p h y s i c a l p r o b l e m is s h o w n s c h e m a t i c a l l y in F i g u r e 4.6, w h e r e the electrical force (FE) is given by

F E - Eq

(4.18)

w h e r e E is the e x t e r n a l l y a p p l i e d electric field. U s i n g this e x p r e s s i o n to r e p l a c e the g r a v i t a t i o n a l force (rng) in E q u a t i o n (4.11), the result for the t e r m i n a l velocity (VE) of a S t o k e s i a n s p h e r i c a l particle ~ this t i m e k n o w n as the electrical drift velocity is

Eq VE =

(4.19)

3rrlxd a n d the c o r r e s p o n d i n g electrical mobility ( Z ) , d e f i n e d n o w as v e l o c i t y p e r unit a p p l i e d electric field, is VE

Z =

(4.20) E

Example 4.3. Calculate the electrical drift velocity for a spherical particle of diameter 5 Ixm, carrying a charge equivalent in magnitude to 500 electrons, and moving in a uniform electric field of intensity 10 kV cm -1. Note that E = 10 kV c m - 1 = 106 V m -1. Compare this with the electric field required to cause ionisation and breakdown of air at STP ~ 30 kV cm -1 (SO w e see that 10 kV cm -1 is not unusually high) Again, note that air viscosity (Ix) is 18 • 10-6 N s m -2 Since the magnitude of charge per electron is 1.6 • 10-19 C, particle charge is given by q = 500 [electrons] • 1.6 x 10 -19 [C electron -1] = 8 • 10 -17 C

From Equation (4.19), assuming Stokes' law applies, we get 10 6 [V m - 1 ] • 8 x 10 -17 [C] VE --

3 • 3.142 • 18 • *

10 -6

[N

VE = 0.094 m s -1

84

s

m -2] •

5

X 10-6 [m]

The motion of airborne particles I 0 0 --

I0

1.0

0.1

0.01 103

! 04

105

106

E (Vim) Figure 4.7. Calculated data for values of the electric drift velocity (VE) for charged particles moving under electric fields (E) typical of those which might be found in workplaces under certain conditions.

As indicated in Chapter 3, the natural charge on workplace aerosols tends to increase somewhat more rapidly than linearly with particle size. The same is true for artificial charging in, for example, the corona discharge of an electrostatic precipitator (see Chapter 10). For such levels of particle charge, Equation (4.19) indicates that the electrical drift velocity of particles will increase with particle size. The implication of this is that collection efficiency of particles by electrostatic forces will be greatest for the largest particles. Some results for the electrical drift velocity of typical workplace particles (using data for median particle charge taken from Table 3.2) are given in Figure 4.7. They are shown for a range of electric fields reasonably typical of what might be encountered in practical situations, where electrostaticallygenerated fields might locally be as high as 105 V m -1 (or even higher). With possible particle drift velocities of the order of a few centimetres per second, it can be seen that significant particle motion associated with electrical forces is possible. Motion in thermal gradients

The relevance to industrial hygiene of particle transport of aerosol particles under the influence of thermal forces should not be ignored. The visual appearance of the soiling of walls and other surfaces close to central heating radiators is clear enough evidence of the phenomenon of thermophoresis, a physical process derived from the thermal motion of the gas molecules surrounding a suspended particle (and hence based on the kinetic theory of gases). 85

Aerosol

industrial hygienists

science for

cold

Hot d >>

o

o

0 0 0

o

L~

o

oo

o

0

o o

0 0

0 0 o

0 _

0 0 0

0

O__o

mfp 0

o

o

~

~ o

o

O0 0

o o

O~ 0

~0~'~~0 ~

0

o

o

o

o

0

0

o

o ~

000 o

I

0 0

o

0

oWl/l/l/l,

o

0 dP 0

o

Net force

~176 0

0 0

I~

~ 0

0

0

d < 0 0 0

0 o

0

0

0

0

0

0

o

0

0 0

0

0 0

O0 0

0

0

0

~. X ~ ~

mfp 0 0

o

~0

0

0

O0

0

0 o

o

0 o

o

o

OI

I

~

Net force

0 0 0 0 00 0 0 O0 O0 0 0 0 000 I 0 0 0 0 0 0 0 0 0 0 0 0 0 O0 0 0 0 0 0 0 ~3 0 0 0 0 0 0 0 0 0 0 O0

I

Figure 4.8. Schematic to illustrate the problem of a particle moving under the influence of thermophoretic forces. Note that the net force arising from collisions of the particle with thermally-moving molecules acts in the direction from hot to cold.

The physical scenario is shown schematically in Figure 4.8. For a particle which is present in a thermal field where there is a temperature gradient, it receives more m o m e n t u m from collisions with gas molecules from the 'hot' side (the more energetic gas molecules) than from the 'cold' side (the less energetic gas molecules). The result is a net force in the direction down the temperature gradient (i.e., towards the cold region). For very fine particles which are small compared to the mean free path b e t w e e n gas molecules (d < < mfp), the problem is relatively simple, based on classical kinetic theory considerations. Here, the thermophoretic velocity (vT) has the form

vT

-

- klOT

(4.21)

where OT is the temperature gradient and the minus sign indicates the direction of the t h e r m o p h o r e t i c m o t i o n towards the colder side. The coefficient k a depends on the actual local temperature (T) but, significantly, not on particle size or composition.

86

The m o t i o n o f airborne particles

For larger particles where the particle 'sees' the gas as a continuum (d ~> m f p ) , the basic idea is the same, but the problem is now complicated by the

fact that there is a temperature gradient set up within the particle itself which, in turn, influences the temperature gradient in the region immediately outside the particle. In this region, therefore, the thermal conductivity of the particle (O'p) and of the surrounding air ( O ' a ) become relevant. Now (4.22)

v T -- - k20 T

where the new coefficient, k 2, contains the ratio ~p/ty a, as well as the local temperature. This time there is some dependence on particle size, but it remains relatively weak. Realistic order-of-magnitude calculations are difficult for thermal precipitation in practical situations because the temperature gradient is usually hard to specify or measure. But some hypothetical results are given in Figure 4.9 for some aerosols with contrasting thermal conductivity, typical of those found in industry, using equations for k I and k 2 collected from the literature by Hinds (1982). The results shown here are for hard wood dust and granite. But the calculations can be performed for any substance for which reasonable estimates of thermal conductivity are available. The results in Figure 4.9 are expressed in terms of thermophoretic drift velocity per unit temperature gradient. Here, for a very small particle (d < 0.1 ~m) at T 293 ~ vT ~ 2 x 10 - 4 cm S - 1 per 1 ~ c m - 1 , independently of what the particle is made of. But for larger particles with d > 1 I~m, vT falls markedly, the more so the greater the ratio Crp/~a. The magnitudes of the thermophoretic velocities (per unit of temperature gradient) are small enough to suggest that, in most practical situations, thermal contributions to particle motion would be much less than for other co-existing forces such as gravitational settling or electrostatic precipitation. Nevertheless, it can be an effective particle

_ H a r d w o o d dust O'P~ 5

"7 E 10 - 4 o

"7ra~

dust

E

Orp/tra -- 1O0

[--

~ 10-5 0.01

I 0.1

I 1.0

I 10

d (l~m)

Figure 4.9. Calculated results for the thermophoretic drift velocity (VT) as a function of particle size (d) for aerosols typical of those found in industrial workplaces.

87

Aerosol science for industrial hygienists

deposition mechanism in small confined spaces where high temperature gradients can be maintained. In particular, the fact that vT tends to be constant over quite wide ranges of small particle sizes is a very useful property in relation to certain applications. Indeed, it was exploited very effectively in some of the aerosol samplers based on thermal precipitation which ~ for a time ~ found favour in the British coal mining industry.

Motion without external forces

The concept of particle motion without the application of an external force is also important. For example, consider the simplest case where the air is stationary and a spherical particle is projected into it with finite initial velocity in the x direction. Motion is described by the equation db' x

= - 3zrlxdv x (4.23) dt where the particle now experiences only the drag force. As in the case of a particle falling under the influence of gravity, this differential equation also has a standard form. This leads to the simple exponential solution m

vx

= VxoeXp

(') - ~

(4.24)

T

where Vxo is the initial particle velocity relative to the fluid at time t = 0. This model portrays the situation where a particle is projected into a fluid with a finite initial velocity, analogous for example to a bullet shot from a gun into a water tank. Equation (4.24) has particular relevance to moving air since it describes the fact that, although a particle injected into the flow with zero velocity at first lags behind the flow, it is progressively pulled along by the drag force exerted by the fluid until it eventually 'catches up' with it. At this point, the particle is then transported along at the same velocity as the air itself, and may thereafter be considered to be 'airborne'. This state of being airborne is therefore seen to stem directly from the particle drag force. Further manipulation of Equation (4.24) yields a further important result. Simple integration provides the distance travelled by the particle relative to the air before it comes to rest ~ or, for the converse moving air case, catches up with it. Thus s -

where s is the

stop

Vxo 9

distance.

88

(4.25)

The motion of airborne particles

4.4 SIMILARITY IN P A R T I C L E M O T I O N We have already discussed the question of dynamical similarity in relation to fluid motion, and arrived at the Reynolds' number scaling concept. From the general equation of motion for a particle in the fluid, as expressed by the three-dimensional version of Equation (4.11), we may also examine the conditions under which its behaviour may be scaled. In this way it may be shown that, for systems which are geometrically and fluid-dynamically alike, similar particle motion (i.e., in terms of relative trajectories) occurs provided in the first place ~ that d2pU

St =

(4.26) 18~D

is constant. Here, as in the earlier fluid mechanical discussion, D and U are the characteristic dimensional and velocity scales, respectively. The dimensionless quantity in Equation (4.26) is known as the Stokes' number. For small enough particles, this definition needs also to include the Cunningham correction term to allow for the phenomenon of slip. Similarly, for nonspherical particles it also needs to include the dynamic shape factor. In addition, solutions for the motion of non-Stokesian particles will also depend on Rep. However, the simple form shown in Equation (4.26) is a fair working assumption for many workplace aerosols and the use of St alone for scaling purposes is usually adequate. The physical significance of St becomes apparent for situations where the flow is distorted (i.e., divergent or convergent); for example, near a bluff obstacle in the workplace, inside a bent tube or duct, or in the vicinity of a sampler. Equation (4.13) with (4.26) gives T

T

St =

= (D/U)

(4.27) "rd

where, if D is the dimensional scale of the physical system which is responsible for the distortion (e.g., the width of the bluff flow obstacle), it is also equivalent to the dimensional scale of the distortion itself. It follows that D / U reflects the length of time (Td) for a 'packet' of fluid to pass through the distorted flow region. St is therefore the ratio of the particle relaxation time to the timescale associated with the flow distortion, and so is a direct indication of how well the particle is able to respond to changes in the flow velocity and direction. Note, for example, that a very small particle with correspondingly small 9 will yield a small value of St in many flow systems. This indicates that the particle will tend to respond quickly to changes in the 89

Aerosol science for industrial hygienists flow and so tend to 'follow' the flow closely. A large particle, having large -r and a correspondingly larger St, will tend to respond less effectively to the changing flow. A very large particle will therefore tend to continue along in the direction of its original motion, and not to 'see' the changes in flow direction and velocity. The same concept can be viewed slightly differently. By combining Equations (4.25) and (4.26), we get another relation

St =

(4.28) D

where St is now expressed as the ratio of particle stop distance to the dimensional scale of the flow distortion. Similarly to the preceding argument, the particle will tend to follow the air flow when s is small compared to the flow distortion; and vice-versa when s is of the order of it or larger. From the preceding discussion, it is clear that St is an important measure of the ability of an airborne particle to respond to the movement of the air around it, and that particle trajectory patterns may differ to an extent dictated largely by the magnitude of St, the extremes being St < < 1 and St > > 1, with St ~ 1 representing some intermediate situation. Thus we have the concept of particle 'inertia', which is a function both of the particle itself and of the flow in which it is moving. This is illustrated in Figure 4.10, and embodies one of the most important concepts in the whole of aerosol particle mechanics. So far, we have assumed the absence of gravity. Of course, except in highly specialised working environments (e.g., space stations), gravity will

Air streamline

Particle St = 0 ~ ~ 0 ,

_

~ - , t . . . ~ ..~D- ' ~ ~ s O " ~ ~)..~r..O.- -0 s , . ,

D tjajectories

St increasing

Figure 4.10. Schematic diagram to illustrate the phenomenon of inertial motion of a particle in a flow field which is changing in direction. Note the role of Stokes' number (St).

90

The motion of airborne particles always be present and so must be borne in mind. In order to assess its possible contribution, we may introduce another dimensionless quantity, the

gravitational parameter

Vs G -

(4.29) U

which, in effect, compares the rate at which a particle moves towards a boundary by gravitational settling with that for it to leave the system by airborne convection. This is particularly relevant to particle transport in, for example, elutriators, filters, and the alveolar region of the human lung. In situations where inertial and gravitational forces are acting simultaneously, then to help assess which of the two mechanisms is predominant, we have the further dimensionless parameter, the Froude number St Fr -

U2 =

G

(4.30)

gD

which expresses the ratio of the magnitude of inertial forces to the magnitude of gravitational forces. The larger the value of Fr, the smaller the effect of gravity in relation to inertial effects. It should not be ignored so long as Fr is less than ~ or of the order of ~ unity.

4.5 P A R T I C L E A E R O D Y N A M I C D I A M E T E R For two spherical particles having different diameters (d 1 and d2) and different densities (Pa and P2), their falling speeds in air will be the same provided, from Equations (4.13) and (4.15), that

d21 [31- d22 P2

(4.31)

where for simplicity at this point the slip, Reynolds' number and particle shape corrections have again been neglected. Equation (4.31.) leads directly to a new definition of particle size based on falling speed; namely, the particle aerodynamic diameter (d~e). This is the particle size parameter which, as we shall see as this book progresses, is the most widely appropriate in relation to particle motion. It is defined as the d i a m e t e r of a spherical particle of density p - p* = 1 g c m - 3 -- 10 3 kg m - 3 (equivalent to that of water) w h i c h has the s a m e falling speed in air as the particle in question. T h u s for a g i v e n spherical particle w e h a v e

91

Aerosol science for industrial hygienists

p ) 1/2 dae = d

~

(4.32)

p*

So far, as stated, attention has been focused only on spherical particles. For non-spherical particles, and also restoring the corrections for 'slip' and non-Stokesian (Rep > 1) behaviour described above, the general form for aerodynamic diameter becomes

dae=dv{0CcunCRep))

112 (4.33)

(p* Cc*nCDRep+) where the terms marked * refer to the spherical water droplet. The overall concept of particle aerodynamic diameter is shown schematically in Figure 4.11.

Figure 4.11. Schematic diagram to illustrate the concept of particle aerodynamic diameter for a non-spherical particle.

92

The motion of airborne particles Example 4.4. For the quartz particle in Example 4.2, with equivalent volume diameter 150 Ixm, density 2.7 x 103 kg m -3 and dynamic shape factor 1.36, calculate its equivalent aerodynamic diameter. Again, neglect the slip correction for such a large particle From Equation (4.33) 2.7 X 103[kg m -3] ) dae = 150 x

12(x coRepC )12 Re

103[kg m3] x 1.36 ) 1/2

C;Rep = 211 x CDRep

Note from the calculation in Example 4.2 that the falling speed of the particle is 0.80 m s -1. So we can calculate its Rep and CD directly. Thus Rep = 7 x 104 x 150 x 10-6[m] x 0.80[m S-1] = 8.40 and

(24)(8.42,3)84 6 ,

CD

= 4.82 respectively Using Equation (4.9), we may now reduce dae to the form 211

}

dae =

• 24 (8.40•

1/2

That is (Re;)2/3 }

dae =

162

1+ 6

93

1+

(Rep)2/36 }

Aerosol science for industrial hygienists in which Rep* = 7 • 10 4 • dae[i~m ] • lO-6[m Ixm -1] x 0.80[m s -1] = 0.056 dae where dae is in [Ixm]. As in the earlier example for this particle, we again have a calculation problem since Rep* on the right-hand side of the above equation for dae is itself dependent on the unknown size of the hypothetical water droplet. Again this can be solved by iteration. An alternative is the graphical approach, in which both sides of this equation are plotted as a function of dae; that is y _. d a e

and

y=

162

{ o056dae,2/3 ( )} 1 +

6

with dae in [Ixm] From this plot, the aerodynamic diameter of the quartz particle of interest is estimated from the point at which the two lines intersect. Thus it is easy to show that

*

dae = 370 fxm

600

y = 162 { 1+(

500

O'056dae'~/ 2/36}

400 300 200

y=d~

100

0

k,"

I

I O0

!

I

I

200

300

I

1

I

400

500

I

600

Particle aerodynamic diameter, d~ (l~m) Note: This is clearly much larger than would be expected under the assumption that Stokes law applies. This is because the equivalent water droplet, being much bigger than the original particle of interest, lies even further outside the Stokesian regime, and so experiences a much greater drag force

94

The motion of airborne particles

Particles of extreme aspect ratio, notably long and thin fibres (e.g., asbestos), deserve special mention. The aerodynamic diameter of a fibre depends strongly on its orientation during motion. A number of theoretical models have been developed. In one of them, Cox (1970) derived equations for a cylindrical fibre with its axis perpendicular to and parallel to its direction of motion, respectively

(dae)2(d ) (9) (2L) =

and

-8

{ln

(dae)2(d ) (9) (2L) -

+ 0.193 )

(4.34)

+ 0.807 }

(4.35)

d

-4

{ln

d

where L is the fibre length and d is its diameter. For the long, straight fibres of amosite asbestos settling out in a spiral centrifuge (analogous to settling under gravity), St6ber (1972) obtained by experiment the empirical expression

(dae)2(d )

0.232

(4.36)

which by inspection is seen to be quite close to Equation (4.34). This suggests that, during settling under gravity, such fibres would fall with their axes horizontal.

4.6 I M P A C T I O N We now move on to discuss some aspects of particle motion which have particularly important implications in the industrial hygiene context (e.g., in particle deposition onto surfaces, in filtration devices and samplers, etc.). Consider again what happens in distorted flows, where the bluff body case serves as a useful illustration. Figure 4.12 shows for example a disc-shaped flow obstacle facing a wide aerosol-laden airstream. The air itself diverges to pass around the outside of the obstacle. The flow of airborne 'inertialess' particles would do the same. However, as described above, real particles exhibit the features of inertial behaviour, in particular the tendency to continue to travel in the direction of their original motion. This tendency

95

Aerosol science for industrial hygienists Limiting particle trajectory Streamlines

\

o.~

Impacting Non-impacting

particle trajectories

particle trajectories

Enclosed area, b"

Disc area, b'

Figure 4.12. Illustration of the phenomenon of impaction of particles onto the leading surface of a fiat disc placed normal to the airstream. Note that the efficiency of impaction is given by the ratio b"/b', which in turn is primarily a function of Stokes' number, St.

is greater the more massive the particle, the greater its approach velocity and the more sharply the flow diverges. Figure 4.12 shows a typical pattern of streamlines for the axially-symmetric flow in question, together with corresponding trajectory patterns for particles of given aerodynamic diameter (dae). Some trajectories intersect with the disc, which means that such particles will 'impact' onto its surface. Since particle trajectories, like streamlines in laminar flow, cannot cross one another, we may define a limiting trajectory (or dividing trajectory) surface, inside which all particles will impact onto the disc and outside which all will pass by. Impaction efficiency (E) is then defined as number of particles arriving by impaction E --

(4.37) number geometrically incident on the obstacle

which, for particles uniformly distributed throughout the oncoming flow, is equivalent to b

t!

b

!

E =

(4.38)

where b" and b' are the areas projected upstream by the limiting trajectory surface and the obstacle itself, respectively, as shown in Figure 4.12. 96

The motion of airborne particles If all the particles that impact onto the obstacle in the manner indicated actually stick and so are removed from the flow, then E is also equivalent to the collection efficiency. However, for real particles under real conditions, the situation is less ideal, and it is wise to be aware of the possibilities of particle bounce, rebound or blow-off, especially where dry, gritty particles are involved. This has been studied by several workers, including Vincent and Humphries (1978) who showed that, for gritty particles of fused alumina, the ability of particles to be retained following impaction onto a solid disc facing the wind is strongly influenced by particle size and microscopic morphology (especially in terms of the small asperities which go to make up the particle's surface), the size of the disc, the freestream air velocity approaching the disc, and the nature of the boundary layer flow over the front surface of the disc. The larger the particles, the more gritty and the greater the freestream air velocity, the greater the likelihood that such particles will be re-entrained after impaction. From the earlier discussion, it is a reasonable approximation that particle trajectories are determined largely by the Stokes number for particle flow about the obstacle (St) and, to some extent, by the particle Reynolds number (Rep). Also they will be dependent somewhat on the Reynolds number for the flow about the obstacle (Re) since this governs the shape of the streamline pattern. Thus, for a particle which is small compared to the size of the obstacle, E is a function of St, Rep and Re. That is E = f(St, Rep, Re)

(4.39)

if it is assumed that re-entrainment is negligible. It is obvious that E will take values close to zero for St close to zero and rise steadily as St increases, eventually levelling off towards E = 1 at large values of St somewhere exceeding unity. A large amount of experimental information is available in the literature, and a typical S-shaped trend exhibited by most of these is shown in Figure 4.13. Although the physics of what has been described is basically quite simple, there are no analytical solutions for impaction efficiency. Instead, calculations have to be carried out numerically, based on the determination of particle trajectories in a series of small steps from some starting point sufficiently far upstream of the obstacle where the particle is in dynamical equilibrium with the flow and so where the particle velocity is well-known. By such numerical models, provided that the correct flow field is used in the calculation, good agreement may be achieved between theory and experiment. Consider now what happens for a particle whose trajectory, as traced by the motion of the particle's centre of gravity, passes by outside the obstacle. If this trajectory passes close enough to the surface of the obstacle and if the particle is geometrically large enough, it may be collected by interception, as illustrated in Figure 4.14. Although for d < < D this effect on E is negligible,

97

Aerosol science for industrial hygienists

1.o

E

0.5

0 0.01

0. I

I 1.0

St

I I0

Figure 4.13. Typical S-shaped trend for the efficiency of impaction (E) as a function of Stokes' number (St) for a bluff flow obstacle like that shown in Figure 4.12.

Limiting particle trajectory for interception

Figure 4.14.

X~ d

Illustration of the phenomenon of interception by a bluff flow obstacle. Note that it is a function of the ratio d/D.

it becomes a significant influence if d becomes of the same order of magnitude as D. This might occur, for example, during particle collection in a filtration device made up of thin fibrous collecting elements. It is also relevant at this point to mention the contributions to particle deposition on the obstacle associated with other physical mechanisms. For example, particles may arrive at the surface of the obstacle under the influence of external forces such as gravity or electrostatic forces. They may also arrive under the influence of diffusion (see below). Although these are not based on inertia per se, they may be operating simultaneously with impaction and interception. When a number of deposition mechanisms (1, 2, 3 , . . . , n) are acting in such a situation, they combine in a very complicated way. But, to a fair working approximation, the overall efficiency of collection may be written in the form Eoveral 1 = 1

-

(1-E1)(1-E2)(1-E3)...(1-En) 98

(4.40)

The motion o f airborne particles Nozzle

/

/ Impactor collection plate Figure 4.15.

Illustration of the concept of an impactor.

where El... E n are the individual collection efficiencies associated with each of the mechanisms acting independently. If each of these individual efficiencies is small ( < < 1 ) , then Equation (4.40) reduces to Eoveral I = E 1 + E 2 + E 3 + . . .

+ E n

(4.41)

This picture is particularly relevant to the deposition of particles on the collecting elements of a filter where there is usually more than one physical collection mechanism operating. Again, although the example given has been concerned with deposition from a wide airstream onto a bluff flow obstacle, the same basic ideas apply in other distorted flow situations (e.g., at bends, expansions and contractions in pipes and ducts, etc.). In particular, much of the preceding provides the basis of a special class of device which finds application in particle size-selective aerosol sampling ~ the impactor. Here the particle laden flow is passed through a nozzle and the resultant jet is directed towards a plate (see Figure 4.15). The flow is deflected by the plate and so ~ by virtue of the mechanisms described above ~ particles may be deposited onto the plate. The nozzle may be cylindrical or slot-shaped. As expected, collection efficiency is governed in the first instance by a Stokes' number, this time one based on the air velocity in the jet nozzle and the width of the nozzle. With careful design with respect to nozzle width and depth, and nozzle-to-plate spacing, impactors can be achieved with very sharp and well-specified particle size 'cut' characteristics. The principles of such devices have been described in full by Marple and Willeke (1979) and others.

4.7 E L U T R I A T I O N As stated above, impaction refers to particle deposition under the influence of 'internal' inertial forces, without involving any external forces, gravity or

99

Aerosol science for industrial hygienists

otherwise. In this book, the general term 'elutriation' is used to refer to another mode of particle deposition relevant to industrial hygiene ~ from a moving air stream under the influence of an externally applied force. Traditionally, the term has been used mainly to describe the gravitational separation of particles carried along by smooth laminar flow through a narrow horizontal channel where particles are deposited onto the floor of the channel. A direct extension of this idea is the gravitational elutriation that occurs during aerosol flow vertically upwards (e.g., through a vertical tube, or into an inverted sampling device). Both aspects of gravitational elutriation were first described by Walton (1954). The general principle is the same if some other force (e.g., electrostatic) is the main agency of deposition. The process is relevant to aerosol behaviour not only in sampling devices but also in the airways of the lung after inhalation. The principle of elutriation is shown in Figure 4.16 where, for the purpose of illustration, the force bringing about deposition is that due to gravity. Particles are transported through a two-dimensional channel, conveyed along by the moving air. The flow is laminar so that, after a relatively short distance from the entrance, the boundary layer at each plate has grown to the stage where the flow is fully developed, at which point the velocity profile across the duct has stabilised and become parabolic. From simple solution of the Navier-Stokes equations (where, for the two-dimensional channel flow, an analytical solution is accessible), this velocity profile is given by

(y)2

ux=U~o{1-

~

(4.42)

}

b

Air flow velocity profile

~////////////////ff/////////////////////////////////////////////~////// t~'~~~~_~,o~_.__

- / / / / / /

/ / / ~ /

__ ~

/ / / / / / / / / / / / / / / / / / / / / / / / F I , , - / / / / / / / / / / / / / / / / / / / /

trajecto~_...

/

All particles contained within the shaded region are collected

Schematic to illustrate the phenomenon of elutriation for the transport of aerosol through a horizontal rectangular channel.

Figure 4.16.

100

The motion of airborne particles

where u x is the air velocity in the longitudinal x-direction at a distance y from the central axis of symmetry of the channel (plate spacing 2b) and Uxo is the maximum value at the centre. Particles entering the elutriator come under the influence of the drag force (acting longitudinally along the channel) and vertically (towards the floor). The result is that particles follow trajectories governed by both components. If the relaxation time (-r) of particles is small compared to the time that the particle would spend inside the channel, then it may be assumed that the particle is always in dynamical equilibrium with its surroundings. That is (a) its longitudinal velocity is always the same as the local air velocity corresponding to its position, and (b) its vertical velocity is always equal to its terminal settling velocity (Vs). Particle motion, in terms of its component velocities in two dimensions, may therefore be described by the pair of simple equations dx Px

-=

H x

--

dt

(4.43) dy Vy

"-

V s

--

dt which can be solved very easily to give the family of trajectories corresponding to various particle aerodynamic sizes and positions of particles (relative to the axis) on entry into the elutriator. Some are shown in Figure 4.16. Example 4.5. For air flowing at 1 1 min -1 through a channel of width 10 cm and height (2b) 0.5 cm, calculate the length of the channel which will be sufficient to prevent the penetration of particles with aerodynamic diameter 5 txm. For the channel, the mean air velocity is given by

10_3 ) [m3 s -1] 60

Uave -(0.1[m] • 0.005[m]) = 0.033 m s -1 so that Re = 7 x

10 4

• 0.005[m] X 0.033[m S-1] -- 12

from which we confirm that the flow through the channel is certainly laminar (since Re << 2000)

101

Aerosol science for industrial hygien&ts The falling speed of these particles is given by Equation (4.15), so that (neglecting slip) (5 x 10-6)2[m 2] x 103[kg m -3] x 9.81[m s -2] VS ~-

18 • 18 • 10-6IN s m -2] =7.6x

10-4 m s -1

Equation (4.43) can be solved to provide particle trajectories. By simple integration, the length of the channel (X) for all the particles entering to just be collected X

(y)

+b

-v~

;dy

b

0

3 v~

-b

At this point it is useful to relate the average velocity to the velocity distribution as given by Equation (4.42). This is achieved again by simple integration, so that +b U~v~ =

~ 2b

{1-

(y)2 ~ b

2Uo

} dy--*

-b

and hence for the example given Uave Uo

From this we see that

X=

( Ua~e)

(0.033[m s -1] • O.O05[m])

2b=

Vs

0.00076[m s -1]

The required length of the elutriator, X = 0.22 m T h e p r e c e d i n g t h e o r y a n d w o r k e d e x a m p l e r e v e a l a n u m b e r of i m p o r t a n t f e a t u r e s e x h i b i t e d by t h e p r o c e s s of e l u t r i a t i o n . T h e first is t h a t , for e a c h g i v e n p a r t i c l e size, t h e r e is a d i s t a n c e i n t o t h e c h a n n e l b e y o n d w h i c h p a r t i c l e s c a n n o t p e n e t r a t e . L o o k e d at a n o t h e r w a y , this m e a n s t h a t , for a g i v e n l e n g t h of c h a n n e l , t h e r e is a ' c u t - o f f ' p a r t i c l e a e r o d y n a m i c d i a m e t e r w h e r e n o t h i n g

102

The motion of airborne particles larger can penetrate. Furthermore, although we have treated the case for ideal parabolic laminar flow, closer inspection reveals that the penetration characteristics of the system are independent of the actual shape of the velocity profile, and are determined only by dae , channel dimensions and volumetric flowrate. Finally, it was recognised many years ago that such penetration characteristics as exhibited by horizontal elutriators are governed to a considerable extent (although not exclusively) by the same physical factors as and so are broadly similar in form to ~ the penetration of inhaled particles into the alveolar regions of the human respiratory tract. These features have important implications for health-related aerosol measurement in workplaces, where elutriation has played an important historical role.

4.8 A S P I R A T I O N

Aspiration concerns the process by which particles are withdrawn from ambient air through an opening in an otherwise enclosed body. It is therefore relevant to aerosol sampling systems. It is also relevant to the inhalation of aerosols by humans through the nose and/or mouth during breathing. Development of physical models and practical aspects of aspiration will be described in a later chapter of this book (see Chapter 9). Here, a largely qualitative introduction to the basic aspiration process is given. In order to identify the nature of the process of aspiration and to enable some generalisations, Figure 4.17 shows a sampler comprising a body of arbitrary shape placed in a moving airstream. It has a single orifice located

i

Aroaao

Ns . . . . . . . .

Limiting particle trajectories Streamlines

as

Figure 4.17. Schematic to illustrate the process of aspiration of particles from moving air (as might be applied, for example, in aerosol sampling). (From Vincent, J.H., Aerosol Sampling: Science and Practice, Copyright 1989, adapted by permission of John Wiley and Sons Limited).

103

Aerosol sciencefor industrial hygienists at arbitrary orientation with respect to the wind, through which air is drawn at a fixed volumetric flowrate. There are two competing flow influences on particle transport ~ the external wind which diverges to pass around the outside of the body, and the convergent flow into the orifice, respectively. The interaction between these two gives rise to the complex distorted overall flow pattern shown. It may be thought of as having two parts, the external divergent part and the internal convergent part. Particles moving in this flow system respond to the changes in flow velocity and direction in the ways described earlier. Generally in moving air, it is the wind that brings particles into the region of influence of the aspirating body and it is the inertial forces that provide the dominant influence on aerosol transport in that region. In fact, the system shown may be regarded as just a more complicated version of the impaction of particles onto a bluff body. This time, however, particles may be thought of as having to undergo t w o successive impaction processes, the first onto the surface of the body as governed by the external part of the flow and the second onto the plane of the orifice as governed by the internal part. As we shall see later, this picture provides the basis for constructing a quantitative physical model for the efficiency with which particles are aspirated from the ambient air and into the sampler body through the orifice. Aspiration efficiency (A), as the term suggests, defines the effectiveness of the aspiration process. It needs to be expressed initially in terms of the flux of particles entering the orifice(s) of the sampler and so requires definition of the limiting particle trajectory inside which particles are aspirated and outside which particles either impact onto the sampler surface or pass by. Thus, referring to Figure 4.17 we have

us A =

(4.44)

No where N s is the number of particles per unit time passing directly through the plane of the sampling orifice and N o is the number originally contained with the aspirated air volume. This leads directly to

f cvda as A =

(4.45)

f cvda ao

where c and v are local values of particle concentration and air velocity, respectively, over elementary areas da, assumed to be distributed continuously

104

The motion of airborne particles everywhere throughout the region in front of the sampler. This may be re-written as

f cvda a rt

A =

(4.46)

f cvda a o

since, by the definition of the limiting particle trajectory, the particle flux through a" is the same as that through a~. For uniform distribution of the approaching particle concentration and air flow velocity (specifically, cv = constant), this reduces to a" A =

(4.47) ao

This shows that the basic definition of aspiration efficiency needs to be expressed in terms of particle trajectories. This, indeed, forms the basis for some theoretical and experimental determinations of A. However, for most practical purposes, a definition based on particle concentration is more appropriate. Integrating Equation (4.45) with c v = constant leads to

CsU~as A =

(4.48)

coUao where U is the windspeed and Us the mean velocity of aspiration through the plane of the sampling orifice. Since, by continuity, U~a~ = Ua o, then Cs

A -

(4.49) Co

That is Concentration of particles in the air actually entering the orifice A

Concentration of particles in the undisturbed upstream air

105

(4.50)

Aerosol science for industrial hygienists Aspiration efficiency defined in this way is the most useful description of performance for an aerosol sampling system. Starting from Equation (4.50), and from considerations of particle impaction from one region of the flow to another, a system of physical equations may be developed which can, in principle, provide estimates for A as a function of particle aerodynamic diameter (dae), body and orifice geometry and dimensions (D and 8, respectively), orientation with respect to the wind direction (O), external windspeed (U) and mean aspiration velocity (Us). For present purposes, it is useful to express some of the generalisations which arise. In the first instance we therefore have

A -f

St,

,

Us

(4.51)

,O,B

o

where St (= dae2p*U/18txD) is a characteristic Stokes' number for the aspiration system and B is an aerodynamic shape ('bluffness' or 'bluntness') factor. Secondly

'Shadow' region (due to of sampler body)

presence

Only particles falling under gravity within this shaded region are aspirated

~~ -~

Limiting particle trajectory

l Sampler body

?

Figure 4.18. Schematic to illustrate the process of aspiration of particles from calm air (where gravity is the main agency responsible for bringing particles into the vicinity of the sampler). (Figure from Vincent, J.H., Aerosol Sampling: Science and Practice, Copyright 1989, adapted by permission of John Wiley and Sons Limited).

106

The motion of airborne particles

A ~

--

cosO as S t - + m

(4.52)

Us indicating that A levels off for large particles approaching the body at high windspeeds. All the above assumes that inertia is the predominant mechanism, an assumption which usually holds for moving air. But for aerosols under conditions where the Froude number (Fr) is small, then gravitational effects become significant, and the view needs to be modified. In the limit, we talk about sampling from 'calm air'. Here there is no external wind that can bring the particles into the region of influence of the sampler, as shown in Figure 4.18. This role is played by gravitational settling. Now the main functional relationship is

A =f

{ vs} Stc,

(4.53)

where St c (= d2e P*Us/18txS) is a modified Stokes' number more relevant to this new system. Orientation with respect to up or down is also important.

4.9 D I F F U S I O N

Molecular diffusion

Particle motion has so far been assumed to be well ordered and ~ in theory at least deterministic. In reality, however, even in apparently smooth air flows, aerosol particles exhibit random movement associated with their collisions with gas molecules which themselves are in thermal motion (as described by the classical kinetic theory of gases outlined in Chapter 2). Such movement is independent of any convection associated with the air itself, and is known as molecular diffusion ~ or 'Brownian' motion, named after Robert Brown who first observed the phenomenon as long ago as 1827. As a result of this type of random motion, there is a net migration of particles from regions of high concentration to regions of low concentration. That is, although individual particles may diffuse in either direction, a greater number end up travelling down the concentration gradient. The resultant local net flux of particles by this process is described by the well-known Fick's law of classical diffusion, which is described for the simple one-dimensional (for the x direction) by

107

Aerosol science for industrial hygienists Oc L o c a l net flux = - D B

(4.54)

Ox w h e r e c is the local c o n c e n t r a t i o n a n d D B is t h e coefficient of B r o w n i a n diffusion. F r o m classical k i n e t i c t h e o r y for a small s p h e r i c a l p a r t i c l e in t h e S t o k e s ' r e g i m e , the l a t t e r is given by

kBTCcun DB =

(4.55) 3rrtxd v

w h e r e T is t h e air t e m p e r a t u r e (in ~ a n d k B is B o l t z m a n n ' s c o n s t a n t (1.38 • 10 -23 J ~ H e r e t h e n u m e r a t o r r e p r e s e n t s t h e t h e r m a l e n e r g y of t h e gas m o l e c u l e s t h a t is b e i n g t r a n s f e r r e d to t h e p a r t i c l e s a n d t h e d e n o m i m a t o r r e p r e s e n t s the loss of p a r t i c l e e n e r g y d u e to viscous effects. T h e r e f o r e , D B e m b o d i e s t h e c o n t i n u a l i n t e r c h a n g e of t h e r m a l e n e r g y b e t w e e n t h e gas m o l e c u l e s a n d particles, a n d vice-versa. Example 4.6. Calculate the coefficient of diffusion for spherical particles of diameters 0.01 and 1 ~m, respectively, in air at STP. Note that, for air at STP, mfp = 0.066 ~m and ix = 18 x 10-6 N

s m-2;

also T = 293~

For the 0.01 Ixm particle, the Cunningham correction factor is given by Equation (4.5), so that

e( Ccu. =

1+

.55x.01

{2.514+0.8

.o66

}

= 22.45 (as already given in Example 4.1) Now Equation (4.55) gives 1.38 x 10-23[j ~

x 293[~

x 22.45

OB 3 x 3.142 x 18 *

x

10-6[N

S

m 2] • 10-8[m]

D B = 5.35 x 10.8 m 2 s -1 .

This is m a n y o r d e r s of m a g n i t u d e s m a l l e r t h a n t h e v a l u e of 10 . 5 m 2 s -1 g i v e n in C h a p t e r 2 for air m o l e c u l e s . F o r a l a r g e r p a r t i c l e of d i a m e t e r 1 ~ m in air, D B is e v e n s m a l l e r ~ o n l y of t h e o r d e r of 10 -11 m 2 s -1.

108

The motion o f airborne particles E q u a t i o n (4.54) leads directly to the g e n e r a l diffusion e q u a t i o n describing the local rate of change of c o n c e n t r a t i o n OC

02C

= OB Ot

(4.56) OX2

w h o s e solution for the simple o n e - d i m e n s i o n a l case of N o particles r e l e a s e d initially at x = 0 at time t - 0 gives the G a u s s i a n f o r m

c (x,to)

t=0

C (X,ti)

0

I

t=

t2

x

c (x,tz)

~

t=t

X

0

Figure 4.19.

Illustration of the change in aerosol spatial distribution over time as influenced by diffusion.

109

Aerosol science for industrial hygienists Noex p (- xZ/4DBt)

C (X,t) --

(4.57)

2 (~rDBt)l/2 for the c o n c e n t r a t i o n d i s t r i b u t i o n along the x direction at time t. This p r o c e s s is illustrated in F i g u r e 4.19. T h e r o o t - m e a n - s q u a r e d i s p l a c e m e n t of particles f r o m their origin at time t is Xrm s --

(2DBt)l/2

(4.58)

This is an i m p o r t a n t result in the c o n t e x t of this b o o k b e c a u s e it p r o v i d e s a basis for m a k i n g q u a n t i t a t i v e e s t i m a t e s of the m a g n i t u d e of the effect of m o l e c u l a r diffusion in given situations. A case of g r e a t e r practical r e l e v a n c e is that for diffusion in t h r e e d i m e n s i o n s , w h e r e the e q u i v a l e n t result is Xrm s --

(6DBt)I/2

(4.59)

Example 4.7. For a particle of diameter 0.01 ~m in air at STP located at the centre of a spherical cavity of diameter 300 p~m, estimate the most probable time it will take to reach the wall of the cavity by diffusion. Note that D B = 5.35 • 10-s m2 s -1 (from Example 4.6) The distance the particle needs to diffuse is 150 p~m The time taken is obtained from Equation (4.59), thus (150 •

1 0 - 6 ) 2 [ m 2]

t --

6 x 5.35 x 10- s [ m 2s -l] = 0.07 s *

t = 70 ms

Note that this problem is relevant to the deposition of particles inhaled into the alveolar region of the lung (see Chapter 6)

A e r o s o l diffusion in a flowing gas s y s t e m is r e f e r r e d to as 'convective diffusion', and this is p e r h a p s the aspect which is m o s t r e l e v a n t to industrial h y g i e n e , especially insofar as d e p o s i t i o n is c o n c e r n e d . In simple t e r m s this m a y be e n v i s a g e d by s u p e r i m p o s i n g the possible e x c u r s i o n d u e to diffusion on the trajectories that w o u l d o t h e r w i s e result in the a b s e n c e of diffusion,

110

The motion of airborne particles

streamlines rticle from

Figure 4.20.

Illustration of the phenomenon of convective diffusion.

combining the deterministic and probablistic aspects of particle motion (as illustrated in Figure 4.20). Equation (4.59) could be useful in this respect. But, more rigorously, Equation (4.56) may be used to examine how its solutions scale for particle and flow systems of different sizes. This approach is analogous to these already described for scaling inertial and gravitational particle behaviour. This time, it is revealed that the governing parameter is the inverse of the dimensionless diffusion coefficient UD

Pe =

(4.60) DB

where, as before, D and U are dimensional and velocity scales, respectively. The new scaling quantity (Pe) is known as the Peclet number. It is smaller the more pronounced the contribution due to diffusion. Using our earlier estimate for D B, it is clear from Equation (4.60) that diffusion is likely to contribute significantly only for very small particles in systems of small geometry where characteristic air velocities are low. However, such conditions are found in some types of filter and, as already stated, in the alveolar region of the human lung. Therefore, molecular diffusion is relevant in the industrial hygiene context, and so Pe is a dimensionless parameter which should take its place alongside the important Reynolds (Re), Stokes (St), Froude (Fr) and gravitational (G) numbers already identified as being useful in thinking about the behaviours of workplace aerosols.

Coagulation The phenomenon of coagulation was mentioned in Chapter 3 during the discussion about aerosol evolution and stability. In particular, thermal

111

Aerosol science for industrial hygienists coagulation is strongly associated with the random, diffusive motions of the airborne particles. Directly from Fick's law as stated in Equation (4.54), a simple classical model can be constructed. For a monodisperse aerosol consisting of spherical particles of diameter d, it provides the number of times per second that a given particle experiences collisions with other particles dn (4.61)

= 8rrdOBc2

dt leading to the rate of change of airborne number concentration dc

- 47rdDBc2

(4.62)

dt where the product (d DB) is referred to as the coagulation coefficient which, from Equation (4.55), can be seen to be independent of particle size (at least for particles which are large enough for slip to be ignored). Integration of Equation (4.60) leads finally to

c(0) c(t) =

(4.63)

( 1 + c(O)dDBt ) where c(0) is the concentration at time t = 0. Inspection of this equation shows that c(t) falls monotonically with t, but tends to level out (see Figure 4.21). For small values of c(0), c(t) remains of the order of c(0), so

Co .o t.. 4,)

o I=i

.o u .o

0

t

Figure 4.21. Graph showing how the particle number concentration of particles falls with time during coagulation as controlled by Brownian diffusion.

112

The motion of airborne particles

coagulation can be ignored. As far as the industrial hygienist is concerned, this is a reasonable situation for most workplace aerosols.

Turbulent diffusion

The phenomenon of turbulence was discussed earlier. It is a property of fluid mechanical systems which, under certain circumstances, can become very prominent and so is expected to be a significant factor in many areas of aerosol science. It is particularly relevant to workplace environments where almost all flows encountered are turbulent. Turbulent mixing associated with the chaotic motions in a turbulent flow may be thought of as a form of diffusion over and above the molecular variety. Indeed, to a fair approximation in most cases, the flux associated with turbulent diffusion may be described in terms of an expression which is directly analogous to Fick's law as given in Equation (4.54). The only difference is that the turbulent diffusivity of the particles, Dpt , is added to D B. A related dimensionless quantity appropriate for scaling purposes is

Dpt D* =

(4.64) DU

As stated earlier, the simplified characterisation of turbulence in terms of an intensity and a length scale is a convenient way of thinking about many aspects of turbulent fluid flows, in particular for obtaining a measure of the diffusivity of the fluid itself (Dft), as suggested by Equation (2.29). It also provides a means of applying rudimentary aerosol mechanical ideas to the behaviour of particles in turbulent flow situations, and so of saying something about overall particle diffusivity. Here, the central consideration concerns the ability of a particle to respond not only to the changes in the streamline pattern of the mean flow, but also to the eddying motions of the turbulent flow. For an inertialess particle that is able to follow all of the fluid motions faithfully, its turbulent diffusivity will be identical to that for the fluid itself. However, a relatively massive particle of the type more representative of aerosols encountered in reality responds to each turbulent eddy as if it were a steady flow distortion of the type already described. Therefore, the particle's ability to respond to the motion of an individual turbulent eddy must be related to an inertial parameter something like the Stokes' number already defined. This new inertial parameter may be represented by a-U'

gpt --

(4.65)

113

Aerosol science for industrial hygienists

Dpt Drt

N \ \

0

i

0.1

I

1.0

10

I

100

Kpt Figure 4.22. Graph to illustrate the change in turbulent diffusivity of a particle relative to that of air (Dpt/Dft) as a function of the turbulent inertial parameter (Kpt) (not to scale).

where, as before, "r is the particle relaxation time, and U' and l are the characteristic turbulence rms velocity and length scale defined earlier. Here it is noted that it is more appropriate formally for the denominator of this inertial parameter to be represented by the Lagrangian microscale of the turbulent motions. However, to a reasonable first approximation, it is sufficient to write

Dpt = f( Kpt )

(4.66)

Dft where, as before, Dft is the turbulent diffusivity for the fluid. As Kpt increases (i.e., larger particles, greater turbulence intensity, smaller length scale), the particle responds less and less well to the individual fluctuations. In addition to the particle's response to individual turbulent eddies, it is also necessary to consider its overall response to encounters with a succession of eddies. This depends on how the fluid velocities it meets in each successive eddy correlate with one another (in terms of magnitude and direction). The outcome of this aspect of the turbulent diffusion of particles is very difficult to describe in simple terms, but a good summary has been given by Soo (1966). The net result of all these factors, however, remains that particle diffusivity falls as Kpt increases. The shape of this trend (not to scale) is shown in Figure 4.22. Eventually a stage is reached where the particle is so large, or the turbulence itself is such (i.e., I is very large or U' is very small), that the particle does not 'see' the turbulence motions as it travels with the mean flow. In that limit, Dpt --~ O.

114

The motion of airborne particles It is r e a s o n a b l e to expect that, for aerosols in m o s t w o r k p l a c e a t m o s p h e r e s , Dpt will be of the same o r d e r of m a g n i t u d e as Dft. B a s e d on c o n s i d e r a t i o n s like those in E x a m p l e 2.8, it b e c o m e s a p p a r e n t that Dpt s h o u l d be at least of the o r d e r of 10 -3 m 2 s -1. Clearly this is m a n y o r d e r s of m a g n i t u d e g r e a t e r than the values for the m o l e c u l a r diffusion of air m o l e c u l e s and m i c r o m e t r e - s i z e d aerosol particles q u o t e d earlier (10 -5 a n d 10 -8 to 10 -11 m 2 s - l , respectively).

REFERENCES Cox, R.G. (1970). The motion of long slender bodies in a viscous fluid: Part 1, General theory. Journal of Fluid Mechanics, 44, 791-810. Cunningham, E. (1910). On the velocity of steady fall of spherical particles through fluid medium. Proceedings of the Royal Society, A83, 357-365. Fuchs, N.A. (1964). The Mechanics of Aerosols. Pergamon Press, Oxford. Hinds, W.C. (1982). Aerosol Technology. John Wiley and Sons, New York. Marple, V.A. and Willeke, K. (1979). Inertial impactors. In: Aerosol Measurement (Eds. D.A. Lundgren et al.). University Presses of Florida, Gainesville, FL, pp. 90-107. Soo, S.L. (1966). Fluid Dynamics of Multiphase Systems. Blaisdell, New York. St6ber, W. (1971). A note on the aerodynamic diameter and mobility of non-spherical aerosol particles. Journal of Aerosol Science, 2, 453-456. Vincent, J.H. and Humphries, W. (1978). The collection of airborne dusts by bluff bodies. Chemical Engineering Science, 33, 1147-1155. Walton, W.H. (1954). Theory and size classification of airborne dust clouds by elutriation. British Journal of Applied Physics, 5 (Supplement 3), $29-$40.

115

CHAPTER 5

The optical properties of aerosols 5.1 I N T R O D U C T I O N Whereas most of the properties of aerosols outlined in Chapters 2-4 can be directly linked ~ in one way or another ~ with health effects or environmental control, optical properties may appear to be somewhat peripheral. However, there are two aspects which are particularly relevant to occupational health. The first concerns the visual appearance of a workplace aerosol. The fact that it is visible at all is usually a first indication that worker exposure is high enough to demand attention. Furthermore, its visible intensity is a direct indication of the level of exposure. In addition, other qualitative features of the aerosol's appearance (e.g., colour) can provide some information about its physical nature. From such considerations, therefore, an experienced and enlightened industrial hygienist can learn a great deal from the visual appearance of a workplace aerosol. At the more quantitative level, however, the optical properties of aerosols can form the basis of direct-reading aerosol instrumentation for measuring not only aerosol concentration but also particle size characteristics. This chapter therefore sets out to provide an appreciation of the physical basis of how light interacts with clouds of airborne particles.

5.2 P H Y S I C A L BASIS The basic physics associated with the optical properties of aerosols is concerned with the interaction of electromagnetic radiation with individual suspended particles and with ensembles of such particles. If a particle has different dielectric properties to those of the surrounding medium, as reflected in their respective refractive indices, then it represents a dielectric inhomogeneity. As a result, interactions with incident light can be detected from outside. On the other hand, if the particle and medium have the same dielectric properties, then there will be no such interactions. So the particle would not be visible to an outside observer or detector.

116

The optical properties of aerosols Physically, the whole problem may be treated in terms of a plane electromagnetic wave incident on a particle whose geometric surface defines the boundaries of the inhomogeneity and whose dielectric properties are described by the refractive index for the particle medium. Mathematically, it is based on the electromagnetic theory of J.C. Maxwell, another great British physicist of the 1800s. In this theory, the refractive index (or dielectric constant) is expressed in the form m =x+

iy

(5.1)

containing mathematically real and imaginary parts (of magnitudes x and y, respectively). The outcomes of the mathematical solution for this interaction are summarised schematically in Figure 5.1. They include all the well-known phenomena such as reflection, diffraction and refraction which, lumped together, constitute light scattering. In addition, the solution also contains the absorbed component, describing how some of the incident energy goes into increasing the vibrational energy of the molecules inside the particle. Such absorbed energy appears in the form of heat, raising the temperature of the particle. This part of the interaction is associated with the imaginary component of the refractive index. Thus, on the one hand, for a highly transparent particle (e.g., of glass), the coefficient y is very small and so absorption is correspondingly weak. On the other hand, for an opaque black particle (e.g., of carbon), y is relatively large and so absorption is much stronger. There is one further process that deserves mention; namely, the physical mechanism by which radiation incident at one wavelength can be scattered at another. This occurs by virtue of so-sailed 'inelastic' interactions involving the absorption and re-emission of radiation energy by the individual molecules of ~....~

Diffraction

Incident Light

Refraction

Reflection

Figure

5.1.

Summary of outcomes of the mathematical solution for light interacting with a particle. 117

A e r o s o l science f o r industrial hygienists

the particle. However, such interactions do not have much direct relevance to workplace aerosols. So attention here will be focussed on the simpler cases where the wavelength of the incident and scattered radiation are the same. Such interactions are referred to as 'elastic'. The first theory of light scattering was given by Lord Rayleigh in the late 1800s, and applies to very small particles and molecules of size much less than the wavelength of the radiation )~. In effect, for visible light (with)~ from about 0.4 - 0.7 I~m), this means particles with geometrical diameter (d) less than about 0.05 ~m. Under these conditions, the particles may be treated as 'point scatterers', and the resultant mathematical treatment is relatively simple. But the most significant advance, in terms of its relevance to aerosols, came when Mie (1908) extended Maxwell's theory to particles of dimensions not too different from the wavelength of the radiation (d ~,). In this region the mathematical treatment of the phenomenon becomes very complicated. Things become simpler again, however, for particles which are large compared to the wavelength of the radiation (d > > ~). Now the scattering problem reduces to one that can be thought of in terms of classical optics. The physics of light scattering has been described very comprehensively in the texts of Van de Hulst (1957) and Kerker (1963). For a beam of light energy incident on a system of many suspended particles (e.g., an aerosol), the fraction of energy which interacts in the manner indicated is either scattered or absorbed. This energy is effectively removed, so that the beam itself may be regarded as having been attenuated or undergone extinction. The energy that remains in the beam is transmitted. From this picture, the interaction of light with an aerosol may be considered in one of two ways; either in terms of the e x t i n c t i o n of the beam (or, conversely, its t r a n s m i t t a n c e ) or in terms of the scattered component. Both approaches have been reviewed by Hodkinson (1966), and will be examined in the following.

5.3 C O N C E P T OF E X T I N C T I O N O R T R A N S M I T T A N C E Consider an aerosol where there is an incident beam of light for which the radiation flux (or intensity) is I 0. This is shown schematically in Figure 5.2. As a result of the removal of light energy from the beam by interactions like those described above, the intensity of the beam which passes straight through is I, where I < I 0. For a homogeneous aerosol where light is removed at a constant rate K (expressed as a fraction lost per unit length of beam) as the beam passes through it, then the change in intensity from I x to I~ - dI over a portion dx (equivalent to x to x + dx of the overall path length X, as measured from the light source) is given by dI = - Ix K dx

118

(5.2)

The optical properties of aerosols

Io

sLight

o

o

,-

o ~

o"

~

[.i~" ~..i:i .i~.~ii.!.I "i II"'~I'~I"I'I~':"t ...

..

".-

.-

,0-qi I 9 ~": " ~. .

~

Figure

9

9

. .

..

.-

.

9 .. . -t.]

9

i . . i

.

dOtPetictalr

.

9

.

~1

9

.

.

X

* ~J

5.2. Schematicto describethe phenomenonof light extinctionby an ensembleof particlesin an aerosol.

where the minus sign indicates that the intensity of the transmitted beam is being reduced. Integrating this expression over the complete path length, we get

= exp { - K X }

(5.3)

/0

where I/I o is the transmittance of the aerosol and K is the overall extinction coefficient. The combined product inside the brace on the right-hand side of Equation (5.3) is often referred to as the turbidity of the aerosol. This in turn is directly related to visibility. The general exponential form of Equation (5.3) is one which is very familiar in many areas of physics and chemistry, applying to any system involving the transmission of radiation energy (of any sort) as it passes through a scattering or absorbing medium. This law has been attributed variously to a number of people, including Bouguer and Lambert, both physicists from the 1700s. To understand this process more fully, the basic physical problem of how the light interacts with each individual particle needs to be considered. The particle extinction coefficient (Q) is defined as flux scattered or absorbed Q =

(5.4) flux geometrically incident

119

Aerosol science for industrial hygienists so that, for an aerosol consisting of n spherical particles (each of diameter d) per unit volume, we have

Ird2

)

CAQ

K = n

Q

--

cpQ

(5.5)

=

4

4

where Cp is the projected area concentration of the aerosol and c A is the corresponding surface area concentration. For non-spherical particles, Cp and c A may also be related as shown (c A = 4Cp) provided that the particles are reasonably isometric and convex in shape and are randomly oriented. Both concentration quantities are expressed in units of [area of particulate matter per unit volume of air]. Further, it can easily be shown from geometrical considerations that

c M = cA

~

(5.6) 6

where c M is the corresponding mass concentration and, as before, p is the particle density. From Equations (5.3) and (5.5), we may write an equation which may be used as a working basis for aerosol concentration measurement. Thus -

41n(I/Io)

cA =

(5.7)

OX and

- 2pdln(I/Io) =

(5.8)

3QX Inspection of Equations (5.7) and (5.8) underlines one of the problems associated with the use of light extinction as a means of measuring the concentration of an aerosol ~ that the measurement (by either definition shown) is dependent on particle size to an extent determined by how the extinction coefficient (Q) depends on particle size. In fact, c M is independent of particle size only under conditions where Q is proportional to d.

120

The optical properties of aerosols E x a m p l e 5.1. A h y p o t h e t i c a l a e r o s o l c o n s i s t s o f s p h e r i c a l p a r t i c l e s , w i t h n u m b e r c o n c e n t r a t i o n s in d i a m e t e r r a n g e s as f o l l o w s ' -

Range (~m)

Midpoint (~m)

n (m -3)

0- 5

2.5

45 • 106

5 - 10

7.5

125 • 106

1 0 - 15

12.5

106 x 106

1 5 - 20

17.5

87 • 106

2 0 - 25

22.5

34 • 106

2 5 - 30

27.5

12 x 106

3 0 - 35

32.5

5 • 106

What is the surface area concentration of this aerosol? If the extinction coefficient is 2 (see below), what is the percentage extinction for a light beam whose path length is 1 m? Note that the total surface area of n spherical particles of diameter d is

n(~cl2) Create the following table:

Midpoint

n

Surface

(Ixm)

(m -3)

area concentration in range (m 2 m -3)

2.5

4 5 x 106

883 x 10 -6

7.5

125x 106

22092 x 10 -6

12.5

106x106

52039x 10 -6

17.5

87 x 106

83714x10-6

22.5

34 x 106

54082 x 10 -6

27.5

12x 106

28514x10 -6

32.5

5• 106

16594 x 10 -6 Sum

*

c A = 0.258 m 2 of aerosol per m 3 air.

From Equation (5.7), we have

121

257918x 10 . 6

Aerosol science for industrial hygienists

In

(') ~

=-4c

A

Qx

Io

= - 4 x 0.258 [ m 2 m - 3 ] = - 2.064

So transmittance,

X

2 x 1 [m]

= 0.127 /o

That is, only 12.7% of the initial light intensity is transmitted *

Extinction = 87.3%

In E q u a t i o n s (5.7) and (5.8), I/I o and X can be m e a s u r e d relatively easily. But before we can use that information to m a k e useful s t a t e m e n t s a b o u t concentration, we also need to know something about Q. This is less straightforward. The extinction coefficient is one of the most i m p o r t a n t optical properties of an aerosol and its complexity represents one of the m o r e serious limitations insofar as the optical m e a s u r e m e n t of aerosols is concerned. For idealised aerosols where particles all have the same k n o w n size, shape and refractive index, then it is possible to define a single value of Q which can be calculated (using Rayleigh or Mie theory, as indicated below) or d e t e r m i n e d experimentally. T h e n we can devise a simple e x p e r i m e n t a l set-up to enable the direct m e a s u r e m e n t of c A or c M. U n f o r t u n a t e l y , the real situation is rarely so straightforward. For polydisperse aerosols having non-spherical shape and (frequently) poorly-defined refractive index, it is realistic only to be able to obtain c A (and possibly CM) by m a k i n g some assumptions upon which to base a rough estimate of Q. 5.4 P A R T I C L E E X T I N C T I O N C O E F F I C I E N T T h e r e are a n u m b e r of a p p r o a c h e s to how Q might be d e t e r m i n e d , d e p e n d i n g on the relative magnitudes of particle size and the wavelength of light. To begin with, it is useful to define the particle size parameter zrd et =

(5.9)

The simplest case is for particles which are large c o m p a r e d to the wavelength of light (or > > 1), where an e l e m e n t a r y m o d e l based on H u y g e n s wave optics (as described in any e l e m e n t a r y text on classical optics)

122

The optical properties of aerosols may be used to estimate Q. Although somewhat removed from true reality, this approach enables the illustration of some of the important features of extinction. Consider a light beam of cross-sectional area S interacting with a single particle of projected area A. From the basic definition of Q, attenuation of the beam due to this one particle is expressed by -

QA

(5.10)

s If a detector is now placed as shown in Figure 5.3, and if the wave amplitude of the radiation per unit area of the beam is B then, without the particle present, the wave amplitude at the detector is BS and so the corresponding intensity is proportional to B2S 2. When the particle is present, the amplitude at the detector falls to B ( S - A ) and the corresponding intensity is proportional to B 2 ( S - A ) 2. Therefore I

{S 2 - ( S - A ) 2}

Io

S2

(5.11)

which, when A < < S, reduces to -

2A

(5.12) /o

s

Comparing Equations (5.10) and (5.12), it is seen that Q = 2. This is equivalent to saying (as stated in Babinet's principle of diffraction) that the amount of light removed from the beam by diffraction is equal to that which is lost by being geometrically incident on the particle. This is an important ,,,,.Area, S

r./ II Incident Light

i

I-.,Area. A f

!1

"J

tl

i!;,

v

I -

I I I_ I-v

I= Detector plane F i g u r e 5.3.

Schematic to describe a simple model for the phenomenon of light extinction by a single particle.

123

Aerosol science for industrial hygienists

result, but possibly ~ to some readers ~ a somewhat curious one. Why should the extinction coefficient as defined by Equation (5.4) exhibit a value which is greater than unity? This does not seem to be our experience from viewing 'every-day' large objects. However, the above model contains the implicit assumption that the detector must be a very long way away from the object in relation to the size of the object and the wavelength of the light. This can happen for a tiny aerosol particle at a relatively short distance, in which case we may feel comfortable with the result Q --~ 2 for aerosols. On the other hand, Q ---> 1 is more appropriate for much larger objects viewed under 'every-day' conditions. For particles which are small compared to the wavelength of the light (so that oL < < 1), Rayleigh obtained the following analytical expression for particle extinction by scattering

(8~ { 1,)2

ascatt

=

(m2+2)

3

(5.13)

where m is the particle refractive index as already defined. In addition, for absorption (m2_ 1) } Qabsorb

--

4~x

(5.14)

-

(m2+2)

Mathematically, in a manner consistent with Equation (5.1), the scattering result represents the real part of the light-particle interaction and the absorption result represents the imaginary part. From the above, it is seen that the extinction coefficient decreases with decreasing particle size or increasing wavelength of the light. The scattering component changes particularly sharply (as governed by or4). The important range between the two extremes of low and high ~ is most relevant to most aerosols found in workplace environments. This is where the Mie theory applies and for which, because of the mathematical complexity of the solutions, calculations for Q have been performed only for particles of relatively simple shape. Some typical results of such calculations over a wide range of ot are given in Figure 5.4. For the Mie range in the middle, the {Q,e~} curve, after its initial rise out of the Rayleigh region, exhibits strong oscillatory patterns. Physically, these oscillations may be regarded as the consequences of the 'constructive' and 'destructive' wave interference that takes place between the diffracted and refracted radiation. It is for the same reason that, for particles which are strongly absorbing (and where the refracted component is reduced), the oscillations are much less marked. Indeed, they may be damped out altogether.

124

The optical properties of aerosols 4

2 (b) m = 1.55 - 0.66i 0

I 5

1. 10

! 15

.

I 20

et = ' t r d l k

Figure 5.4. Typical results from calculations (using Mie theory) of the extinction coefficient (Q) as a function of particle size (d) and light wavelength (h) as embodied together in the particle size parameter ot = rrd/h (from Hodkinson, J.R. in Aerosol Science (ed. C.N. Davies), Copyright 1966, reproduced by permission of Academic Press Ltd, London). 5.5 E X P E R I M E N T A L M E A S U R E M E N T S O F E X T I N C T I O N Experimental extinction measurements are made in order to obtain values for Q, either for the purpose of validating theoretical calculations or for determining Q for aerosols where there are no calculated values. For workplace aerosols, for example, such information would be needed in order to enable the design and calibration of a suitable monitoring instrument based on the principles of light extinction. On the face of it, extinction measurements should be quite simple. To set up an appropriate optical system would involve a parallel light beam and a detector, between which the aerosol would be interposed. From measurements of I, I o and X we could proceed to determine Q (if we know CA) or CA (if we know Q). But, for the determination of Q, the provision of a test aerosol with well-defined concentration and appropriate particle size characterisation presents considerable difficulties. Then there are problems associated with the optics themselves. It is an in-built assumption that the light reaching the detector is only that which has not interacted with the aerosol and so has not been scattered or absorbed. However, for a light beam and detector of finite physical dimensions, the angular acceptance of the light measuring system will inevitably be such that some of the light scattered in the forwards direction will enter the detector and so be included in the measurement of I. As a result, I will be overestimated, suggesting an apparently low value of Q. The greater the angular acceptance of the detector system, the greater this error. For many practical instruments which have been built to operate on the principles of light extinction, there is a resultant uncertainty in characteristics such that calibration may be unreliable and performance may be inconsistent from one application to the next. In the best practical systems, this problem has been recognised and its effect minimised by careful design of the optics; for example, by the use of pinholes and lenses which prevent light leaving the aerosol at angles

125

Aerosol science for industrial hygienists Pin-holes

Light s ~ c e ~ l ~

..~.-_ ~..~:.,~2,.,~i~,,~~.Photodetector """ "" ' ~ : A:ero:o:l

lenses Figure

5.5.

Typical experimental set-up of the type required for the accurate measurement of extinction of light by an aerosol.

other than very close to the forwards direction from reaching the detector (shown schematically in Figure 5.5). Some good determinations of Q, obtained using an experimental system like that s u g g e s t e d in Figure 5.5, are shown in Figure 5.6. Here, for experimental convenience, Q is plotted as a function of the parameter 13 = 4ot(rn-1) which brings together both c~ and the particle refractive index m. Results are shown for non-spherical but reasonably isometric ~ particles randomly oriented with respect to the incident light. Such particles are quite representative of those found in many workplace aerosols. Figure 5.6 shows that the extinction coefficient tends to be quite constant at Q ~ 2 for 13 greater than about 15. But at smaller 13, Q exhibits only a single peak, higher for non-absorbing than for absorbing particles. From theory, backed up by good experimental data, there is a reasonable starting point for practical aerosol instrumentation based on light extinction principles. One area where there is particular interest is in the tomographic determination of aerosol in workplace atmospheres. This involves mapping the aerosol spatial distribution in a horizontal plane by means of a light

4

Irregular-shaped, 'real' particles

Q 2 ///

0

absorbing

5I

I I0

[3 = 4a (m-l)

I 15

I 20

Figure 5.6. Experimental data for the particle extinction coefficient (Q) as a function of the modified particle size parameter 13 - 4oL(m-1), obtained using apparatus like that described in Figure 5.5. The results are for non-spherical but reasonably isometric particles (from Hodkinson, J.R. in Aerosol Science (ed. C.N. Davies), Copyright 1966, reproduced by permission of Academic Press Ltd, London). 126

The optical properties of aerosols beam which is scanned through a workplace area and is transmitted to a succession of periphally-located detectors. Ramachandran and Leith (1992) have demonstrated that the extinction of light of multiple wavelengths (e.g., white light) for the various optical paths defined in such a system may allow the extraction of both the concentration and particle size distibution over the horizontal plane swept out by the scanning light beam. This is considered practicable because the optical path lengths can be made long. More generally, however, for extinction measurement where the path length is short or where the aerosol concentration is very low, the overall aerosol extinction coefficient (K) is likely to be low, so that I ~ I 0. Here therefore, since we are looking for small changes in light intensity, the sensitivity of measurement may be poor. For this reason, light extinction is not widely used today in workplace aerosol measurement instrumentation. Light scattering the converse of extinction ~ has many advantages, and so is much more widely used.

5.6 L I G H T S C A T T E R I N G Consideration of scattered light is based on the same physical principles as just outlined for extinction. However, because we now have the angular distribution of the scattering to take into account, the problem becomes yet more complicated. But, in terms of possibilities for instrument development, it does create more options. Now, instead of the extinction coefficient (Q) which was employed in the preceding section, the related particle scattering coefficient, S(O), is defined as flux scattered per unit solid angle at angle (O) from the forwards direction S(O) =

(5.15) flux geometrically incident on the particle

At the outset, it is assumed that the incident light is contained within a beam which is plane parallel and unpolarised. The physical scenario is shown schematically in Figure 5.7. Here, as before, I o is the intensity of the light incident on the aerosol. Now, for a detector which is being used to observe the aerosol through a slit detector placed at the angle O, it sees light coming from a scattering plane within the transmission path where the actual incident light, Ii, will be less than I o ~ due to extinction in the part of the beam before it reaches that plane. However, for simplicity, it is assumed here that the aerosol concentration is low enough such that I i = I o. From Equation (5.15), the total light flux received by the collecting optics in Figure 5.7 may be shown to be

127

Aerosol science for industrial hygienists

to

/

~.

" ~ Scattered light

Co

slit, w x h (Solid angle within which light is received, whlj~)

Figure 5.7. Physical scenario for assessing angular light scattering by an aerosol (from Hodkinson, J.R. in Aerosol Science (ed. C.N. Davies), Copyright 1966, reproduced by permission of Academic Press Ltd, London).

i:(

)( wh )

cploS(O )

(5.16)

sin 0

where w and h are the width and height, respectively, of the collecting slit which is placed at the focal point of a collecting lens of focal length f, and Cp as before ~ is the projected area aerosol concentration. In this expression, the first bracketted term on the right-hand side represents the volume (hatched area in Figure 5.7) from which the light is received by the collecting optics and the second represents the geometrical solid angle within which it is received. From this, as in the previous discussion about extinction, it is seen that there exists a basis upon which aerosol concentration measurement might be made. This time, however, instead of requiring knowledge of the extinction coefficient (Q), it is S(O) that needs to be defined and quantified. The application of light scattering principles in an instrument aimed at measuring aerosol concentration is referred to as 'aerosol photometry'. 128

The optical properties of aerosols 5.7 P A R T I C L E S C A T T E R I N G C O E F F I C I E N T As for the simpler case of extinction, the problem of determining the particle scattering coefficient may be broken down into three regimes ~ very fine particles, intermediate-sized particles and very large particles. For the reasons discussed earlier, there is no need to discuss the case of very large objects since these are not relevant to workplace aerosols. Again, for very small particles, Rayleigh's theory is applicable. For Rayleigh scattering, the total normalised scattered light flux from a particle into the angle O, as defined generally by Equation (5.15), is

s(o)

=

--

27r

{

( 1 "k-COS20) ~ ~) ( 1 + COS20)

(5.17)

(m2+

where the coefficient r replaces the two bracketted terms in the equation for S(O). Built-in to this equation, there is an additional concept, involving the degree of polarisation of the light. According to electromagnetic theory, light energy contains vibrations in the electric and magnetic vectors which are contained within the plane of the wavefront (i.e., transverse to the beam). For light where the vibrations are uniformly distributed throughout that plane, the light is regarded as unpolarised. But for light where the vibrations have a preferred direction, it is polarised. The degree of polarisation depends on the extent of this non-uniformity in the vibrations. Light which is initially non-polarised may undergo changes in degree of polarisation after it has interacted with some dielectric inhomogeneity. For example, reflected light may become polarised preferentially in a given direction, a phenomenon familiar to anyone who has used polarising sunglasses. In light scattering by aerosols, the same basic physics comes into play. Therefore, in Equation (5.17) which represents the total light scattered by a particle into the angle 0, there are two components Sa(O) = r and 32(0 ) ~-~ + COS20

(5.18)

where SI(O ) and $2(O ) are the components for scattered light polarised in and normal to the plane containing the incident and scattered light beams and the detector (see Figure 5.8), and

S(O) = 81(0 ) + 32(0 )

(5.19)

from Equation (5.17). The first component, $1(O ), gives an angular scattering pattern that is circular (and hence uniform in all directions). The second, $2(O ), gives a symmetrical figure-of-eight pattern indicating enhanced scattering in the forward and backward directions and zero scattering at 90 ~, 129

Aerosol science for industrial hygienists

•

Electromagnetic vibrations in the

tion of light beam (x)

Figure 5.8. Schematic to illustrate the two-component nature of polarised light. independently of particle size or wavelength of the light. The angular distributions of the two components of S(O) for Rayleigh scattering by fine particles are shown in Figure 5.9. As for extinction, the situation becomes more complex for larger particles of the order of or greater than the wavelength of the light. Some typical scattering patterns are shown in Figure 5.10, where S(O) is now replaced by the Mie intensity parameter, s(O) - 2rretzs(o). Here the general shape is $1

$2

s2(o)

Incident light

sl(0)

Small particle (d << k)

Figure 5.9. Angular distribution of the two polarised light components undergoing Rayleigh scattering by fine particles (from Hinds, W.C., Aerosol Technology, Copyright 1982, adapted by permission of John Wiley and Sons Inc). 130

The optical properties of aerosols 10 4 --

10 ~

s2 (0) 10 2

~"

i0 ~

I~

10 o

III ,~

10 -~

r

r

s~ (o)

10 .2

10 .3

10 -~ 0

t 20

t 40

1 60

I 80

I 100

I 120

I 140

I 160

I 180

0~

Figure 5.10. Lightscattering patterns for water droplets, showing both polarised components, calculated from Mie theory (from Hinds, W.C., Aerosol Technology, Copyright 1982, reproduced by permission of John Wiley and Sons Inc). seen to have much more fine structure, with many more maxima and minima than for Rayleigh scattering. This is entirely consistent with the oscillations noted earlier for the extinction coefficient (Q). Not shown in this diagram is the fact that, as expected for absorbing particles and consistent with the behaviour of Q, the oscillations are substantially damped out. The same is true for particles of non-spherical shape and for polydisperse aerosols. One important feature is the asymmetry in the pattern with a marked tendency for scattering preferentially in the forwards direction. This tendency increases with increasing particle size. 5.8 MASS C O N C E N T R A T I O N A E R O S O L P H O T O M E T R Y We have already touched on the problems associated with the particle size dependence of optical aerosol measurement in relation to the industrial hygienists' usual desire to obtain aerosol concentration in terms of mass per unit volume. This has recently been summarised concisely by Gebhart (1993). A volume scattering functionmay be identified as the amount of light scattered per unit volume (and hence, if density is know, the mass) of aerosol. Thus, for a given refractive index and a given angle scattered light flux Sv =

(5.20) (rid3/6)

131

Aerosol science for industrial hygienists

" regioOniWahtlrlrlSOpl ee sfiSlfla1 I00

~o

I0I

a = ~d/A

I IO0

J IOOO

Figure 5.11. Graph showing the trend between the volume scattering function (Sv) and the parameter e~ = 7rd/?~ (from Gebhart, J. in Aerosol Measurement (eds. K. Willeke and P.A. Baron), Copyright 1993, adapted by permission of Van Nostrand Reinhold, New York).

where the ideal practical system for industrial hygiene purposes would be one where S v is constant for all aerosols. In Equation (5.20), the scattered light flux can be inferred from inspection of equations like those given earlier (for both the extinction and scattering modes). The actual trends for S v are shown in Figure 5.11 as a function of c~ ( - 7rd/~, see Equation (5.9)). This shows two regions. The first comes from the fact that, for particles considerably larger than the wavelength of the light, Q ~ 2. Thus from Equation (5.4), we see that S v is proportional to d -1. For the other region, for particles much less than the wavelength of the light, Equation (5.4) with Equations (5.9) and (5.13) show that S v is proportional to d 3. Figure 5.11 reveals that there is an intermediate range of particle size (about 0.5~-5~) where S v does not vary so strongly with particle size. Depending on the choice of light wavelength used, this is a range which contains some aerosols of interest to industrial hygienists. So there is some scope for achieving the desired practical goal.

5.9 T H E V I S U A L A P P E A R A N C E OF A E R O S O L S It follows that much of what has been discussed in the preceding paragraphs can be related to the qualitative visual appearance of aerosols. Consider first the Rayleigh region (c~ < < 1). This is particularly relevant to very small particles (including air molecules themselves, as well as tiny clusters of water molecules) high up in the ambient atmosphere. Here the small value of extinction coefficient (Q) accounts for the long visibility. The fact that Q increases with increasing cx (and hence, for constant particle size, decreasing light wavelength) means that, for white light, the blue components will be

132

The optical properties of aerosols

~.bhm < kred

r V V

Q blu~

~176 /Otred

~blue

ot ( = "rrd/k) << 1 Figure 5.12. Illustration of the effect of different wavelengths on the a m o u n t of light scattered or a t t e n u a t e d (the basis of u n d e r s t a n d i n g red sunsets and blue sky).

attenuated more than the red ones. This phenomenon is illustrated by reference to the section of the {Q,ot} graph shown in Figure 5.12. It explains why red light is transmitted preferentially and hence why sunsets, where the sun is viewed in essentially the transmission mode, appear to be reddish. Conversely, scattering is preferential to the shorter wavelengths. Hence, for the atmospheric example, this provides an explanation for why the sky, where the light from the sun is viewed in the scattering mode, appears blue. Indeed, were it not for the presence of small scatterers in the atmosphere, no light would be scattered at all and the sky would look black. This, in fact, is the case for an observer in outer space. A similar line of reasoning can be applied to very fine aerosols encountered in workplaces (e.g., smokes and fumes). Any aerosol is visible as such only because light is scattered by it. That is, we view the scattered light. An ultrafine aerosol will therefore appear bluish, the more so the finer it is. Perhaps the most familiar illustration of this effect is the bluish visual appearance of cigarette smoke when viewed under strong illumination. Another is the appearance of welding fume. For aerosols containing larger particles, there are many more possible options. First of all, a coarse, highly absorbing aerosol, when observed in the transmission mode ~ for example, by looking at the aerosol in a chimney plume against a light background such as the sky ~ will appear black. This is because Q is uniformly large (at around 2) for all a, so that overall extinction over the range of a is large. By contrast, for transparent particles (e.g., water droplets) the oscillations in Q permit reasonable transmission

133

Aerosol science for industrial hygienists at some values of c~ (and hence some wavelengths). So such aerosols look brighter. The angular distribution of scattering can be observed visually and provide information about the nature of some aerosols. With the naked eye, for example, we may observe the 'rainbow' effect for aerosols of transparent liquid droplets large enough for the principles of geometrical optics to be applicable. Here, from considerations of an incident horizontal light beam (e.g., corresponding to the sun being low on the horizon) with refraction and total internal reflection of light inside each droplet, the dispersion of the light is the predominant mode of interaction, each droplet scattering a given colour in the visible spectrum at a particular angle. For the cloud of droplets as a whole, the rainbow represents the angular dispersion of the various colours in the incident light. For smaller droplets, where Mie theory applies, the matter is made more complicated by the added contribution from diffraction. Other visual effects are less obvious to the naked eye. However, it is possible to observe an aerosol under laboratory conditions using a telescope in conjunction with a bench-top optical spectrometer table. Here, for an incident beam of white light, the telescope can be used to scan through the angles from 0~ to 360 ~ As indicated by figures like Figure 5.10, the maxima and minima in the scattering patterns of the various wavelengths occur at different angles. When they are superimposed, they form a series of red-coloured bands which are observed as the telescope is scanned. These are referred to as higher-order Tyndall spectra (named after John Tyndall, who discovered the effect during the 1800s). This essentially-qualitative visual effect provides a means by which the mean size of particles in the test aerosol (usually previously-collected dust which has been re-dispersed) can be estimated. This can be achieved, for example, by noting the angular position of the red bands and referring to a calibration table for monodisperse test particles of known refractive index.

5.10 OPTICAL M I C R O S C O P Y Strictly speaking, airborne particles per se cannot be viewed directly under the microscope. But microscopy is an important tool by which industrial hygienists may examine aerosol particles after they have been deposited onto a slide or a filter. In many respects, the visual examination of a small particle in this way involves much of the same physics as has been discussed already. In particular, the ability to see a particle through the microscope depends, as before, on the interaction between incident light waves and an object which is dielectrically different from its surroundings. That ability is controlled by the way in which light is scattered and absorbed by the particle. For example, for a particle which is small, say d -< )~, its appearance is dominated by the 134

The optical properties of aerosols

scattered ~ in particular the diffracted ~ light. Resolution of a microscope is defined as the ability to distinguish between two point objects in spite of their two diffraction patterns. This in effect defines the smallest object that can be properly 'seen'. It has been shown that this is roughly of the order of 0.5 ~xm, although improvements can be achieved by adjustment of the numerical aperture of the microscope (which in turn is a function of angle at which the light enters the objective lens of the microscope and the refractive index of the medium from which the light enters the lens). From all this, it may be estimated that, by optical microscopy in the visible region, particles cannot be observed of size below about 0.2 Ixm. Of course, that limit can be lowered by viewing particles in radiation at shorter wavelengths, either using electromagnetic radiation or, in particular, using a high-energy electron beam (as in electron microscopy). It has already been indicated earlier in this chapter that, if a particle has the same refractive index as the medium which surrounds it, it will be invisible. In microscopy, the difference in the two refractive indices is referred to as the index o f visibility. Here, for particles which are not readily visible in air (i.e., are relatively transparent and colourless), then the appropriate choice of immersion or mounting medium can improve matters. So too can modifications to the method of illumination (e.g., dark field, oblique, polarised light, phase contrast). These are subjects which are special to microscopy and which are therefore fully described elsewhere.

REFERENCES Gebhart, J. (1993). Optical direct-reading techniques: light intensity systems. In: Aerosol Measurement (Eds. K. Willeke and P.A. Baron). Van Nostrand Reinhold, New York, pp. 313-344. Hodkinson, J.R. (1966). The optical measurement of aerosols. In: Aerosol Science (Ed. C.N. Davies). Academic Press, London, pp. 287-357. Kerker, M. (1963). Electromagnetic Scattering. Pergamon Press, Oxford. Mie, G. (1908). Beugung an leitenden Kugeln. Annal der Physik, 15, 377--445. Ramachandran, G. and Leith, D. (1992). Extraction of aerosol size distributions from multispectral light extinction data. Aerosol Science and Technology, 17, 303-325. Van de Hulst, H.C. (1957). Light Scattering by Small Particles. John Wiley and Sons, New York.

135

CHAPTER 6

The inhalation of aerosols 6.1 I N T R O D U C T I O N The routes by which aerosol can reach human biological systems and so place them at risk can involve skin deposition, the food chain (after deposition on plants, etc.) and inhalation. Aerosol exposure via the inhalation route is of considerable interest in occupational health, representing a major potential source of hazard for workers in many occupational environments. The nature and magnitude of the hazard in a given situation depend on a complex combination of many factors, including: (a) particle size distribution (which governs how the aerosol enters the body by inhalation, and how it penetrates into and is subsequently deposited in the respiratory tract); (b) airborne concentration (which governs how much is deposited); and (c) morphology, mineralogy and chemical composition (which govern the subsequent fate and biological responses to the presence of the particles in contact with vulnerable tissue). It is therefore clear that many aspects of basic aerosol science touched on in previous chapters are important to understanding the nature of inhalation and subsequent events. This chapter sets out to review the relevant background, much of it driven by the need to develop biologically-appropriate criteria for health-related aerosol measurement in workplaces. In turn this is important in providing part of the scientific framework needed for the assessment of overall level of risk in practical situations.

6.2 T H E H U M A N R E S P I R A T O R Y T R A C T The human respiratory tract is a complex system, its primary objective being to supply the body with oxygen (during inhalation) and to eliminate carbon dioxide (during exhalation). For most of the purposes of this book, a rudimentary outline of the form and function of the respiratory tract, and how it handles particles which are deposited there, is sufficient. Greater details may be found in any of the many physiological texts dealing with the

136

The inhalation of aerosols human respiratory system. The book of medical paintings by Netter (1980) provides an excellent graphic description of the relevant detailed anatomy and physiology. The simple model used as the basis of present discussion is shown in Figure 6.1, where the inhaled air enters the respiratory tract through either the nose or the mouth. The inhaled air first encounters the upper region of the respiratory tract, comprising the airways of the nose and mouth which, together, form the nasopharynx. The overall 'head' region is assumed in the present context to include both the nasopharynx and the larynx. The nose in particular is a complex system of cavities (or turbinates) in which the air, as it passes through, becomes humidified and warmed prior to entering the lung. The airflow through this system is highly non-uniform, consisting of a combination of jets and relatively quiescent 'backwater' regions and, depending on the rate at which air is being inhaled, the flow may be either laminar or turbulent. Parts of the surface of the nasal cavity are ciliated, being covered with tiny hairlike protrusions (cilia) which 'beat' such that, aided by a mucous layer, any foreign particles which are present are transported away. For particles depositing in the posterior region towards the back of the nasal passage, they are swept by this ciliary action backwards towards the throat (or pharynx) whereupon they are swallowed and eliminated through the gastrointestinal tract. For particles which deposit in the anterior region towards the front, they are swept forwards towards the unciliated parts from

Figure 6.1. Simplerepresentation of the human respiratory tract (from Vincent, J.H., Aerosol Sampling: Science and Practice, Copyright 1989, reproduced by

permission of John Wiley and Sons Limited). 137

Aerosol science for industrial hygienists where they may be removed by nose-blowing, etc. Particles deposited in the mouth are carried by the saliva and eliminated by swallowing or by spitting. The whole region below the larynx is the lung, referred to as the thoracic region. Air enters through the larynx into the trachea which in turn leads to a branching system of conducting airways referred to as bronchi and, eventually, conducting and terminal bronchioles. Beyond the terminal bronchioles lie the respiratory bronchioles, alveolar ducts and alveolar sacs. Models have been proposed for describing the complex morphometry of the lung, the most widely applied one being that proposed by Weibel (1963). His 'Model A' version is summarised in Table 6.1, indicating the classes of conducting airway ducts in the tracheobronchial (TB) region and how they branch all the way down to the alveolar (alv) region. Data are also given about duct dimensions at each branching generation as well as ~ for an instantaneous breathing volumetric flowrate of 60 1 min -~ the mean air velocity at and residence time within each. Table 6.1 shows that there are 16 generations of branching in the conducting airways such that, by the time

Table 6.1. Summary of data on the human lung, describing branching, dimensions of successive branching dimensions, and air velocities and residence times at each (from Lippmann, 1977; H i n d s , 1982). Region

Airway

Gc

T-B~

Trachea Bronchi

0 1

11

N(; d

D(; ~

L(; t

1

18

12(I

2.5

3.9

31)

2

12

48

2.3

4.3

11

2x l03

1.1

3.9

X(;g

20

V(; h

11.52

t(; i

7.4

13

Bronchioles

Ab

Respiratory bronchioles Alveolar ducts Alveolar sacs

14

16x 103

0.74

2.3

69

(I.14

16

16

6 6 x 103

0.60

1.6

180

tt.(15

31

18

2 6 x 104

0.50

1.2

530

(I.(13

6(1

21

2 x 106

0.43

0.7

3 x 103

0.003

210

23

8 x 106

0.41

0.5

1 • 104

0.001

550

17

aT-B = tracheobronchial region. bA

= alveolar region.

cG

= generation.

number per generation.

dN G

=

eD G

= diameter of duct at generation G ( m m ) .

eL G

= length of duct at generation G (mm).

gX G

= cross-sectional area of duct at generation G (cm2).

hV G

= air velocity at generation G (m s - l ) .

itG

= residence time at generation G (ms).

138

The inhalation of aerosols the terminal bronchioles are reached, there are over 60,000 such branches. As the air enters the trachea below the larynx, the sudden expansion of the flow leads to flow separation and associated turbulence. This could persist for one or more branching generations. Beyond about the third or fourth generations, however, the effective cross-sectional area of the flow has increased to the point where the Reynolds number for the inspiratory airway flow falls below that which can sustain turbulence (Re less than about 2000), beyond which the flow becomes laminar. The surfaces of these conducting airways are, like parts of the nasal region, covered in cilia and mucous such that particles deposited there can be swept along on the 'ciliary escalator' ~ this time upwards towards the larynx so that such particles are eliminated, again through the gastrointestinal tract. Below the conducting airway region, we have the alveolar region. This comprises the respiratory bronchioles, alveolar ducts and alveolar sacs, the latter consisting of up to 8 million small closed-ended cavities (i.e., alveoli) each of dimension about 300 txm. This is the region where the gas exchange takes place, where oxygen is dissolved into the blood and carbon dioxide is released. By the time this region has been reached, the mean air velocity has fallen to the point where the contribution of convection to the transport of gas molecules becomes very small. At this point it is useful to distinguish between the tidal air volume (i.e., that which is convected in and out during the breathing cycle) and the reserve air volume (i.e., that which is retained in the deepest part of the lung where there is insignificant actual convection). Now, therefore, we have to address the question of how mixing takes place between the tidal and reserve air volumes so that inhaled air molecules (including essential oxygen) ~ as well as other entities (including small particles) ~ can be transported into the alveolar spaces. This transport at the interface of the tidal-reserve interface is mainly by diffusion. Firstly this occurs by direct mixing associated with basic molecular diffusion. Secondly there is mixing by convective diffusion which occurs as a consequence of the combined effect of radial molecular diffusion and the velocity profile. In addition, Taulbee and Yu (1975) identified a further 'apparent' diffusion-like contribution to mixing arising from the statistical distribution of velocities in different tubes of the same generation. In the alveolar region, there are no cilia. So ciliary action for the elimination of deposited particles is not available in the way it is for the ciliated conducting airways higher up in the respiratory tract. Here, however, a different mechanism of clearance comes into effect, involving the free alveolar macrophage cells which are present on the alveolar surfaces. These are cells which have evolved primarily for the defence of the deep lung. They move about freely on the alveolar surface, scavenging and engulfing foreign particles with which they come into contact, and thereafter carrying them by chemotaxis towards the conducting airway system, whereupon ciliary clearance takes over in the final stage of elimination.

139

Aerosol science for industrial hygienists

For healthy adult humans, flow rates and breathing patterns for the respiratory tract system described can vary greatly, depending on individual factors as well as on the type of activity undertaken. Typically for someone 'at rest', the inhaled volumetric flow rate (over a breathing cycle) is of the order of 7 1 min -1 (the minute volume) and the breath rate is about 10 breaths min-~. For someone carrying out moderate to hard work, the minute volume can rise beyond 20 1 min -1. 'Tidal' volumes can range from less than 1 1 to greater than 2 1, and about 1 1 of reserve air is retained. Although humans have always been exposed to aerosols (e.g., naturallyoccurring, smoke from fires), substantial long-term exposure at levels like those which may be found in workplaces have come comparatively recently. The respiratory defence mechanisms summarised above have evolved to deal with inhaled substances of a wide range of types. But it is perhaps arguable whether we have yet evolved to the stage where the respiratory tract can deal effectively with some of the aerosols to which humans have been exposed in workplaces since the relatively recent industrial revolution.

6.3 A E R O S O L I N H A L A T I O N

Basic definitions and terminology In the following sections, we shall examine the efficiency with which aerosol particles may enter the human body from the ambient workplace air through the nose and mouth, and subsequently penetrate through ~ or be deposited in ~ the various regions of the respiratory tract. To aid this discussion, consider a much simplified schematic picture of the respiratory tract, as shown in Figure 6.2. In this figure, Nin h is the number of particles per unit time inhaled through the nose and/or mouth and N O is the number that would be inhaled if the concentration in the workplace and inhaled air, respectively, are the same. So the efficiency of inhalation is

Ninh A =

(6.1)

No where the major difference from the aspiration picture presented earlier (see Chapter 4) is that, for human inhalation, the sampling flow rate is not steady but, rather, cyclical. At the next level, DET is the number of particles per unit time depositing in the nasopharnynx ~ or, more commonly, the extrathoracic (ET) region

140

The inhalation of aerosols

Extra-thoracic region (ET)

Tracheobronchial region (TB) horacic

/~7.

4

~__~~

[ regi~ (th~

Alveolar region (alv)

Figure 6.2. Schematic of the human respiratory tract on which to base a quantitative discussion of the regional deposition of inhaled aerosols.

and Nthor is the number penetrating down to the entrance to the thoracic region. So the efficiency of deposition, EET , in the nasopharynx is given by

OET

-

E E T --

=

( Nthor)

1-

Ninh

(6.2)

Ninh

Next, if DTB is the number of particles per unit time depositing in the tracheobronchial region and Nalv is the number of particles penetrating down to the entrance to the alveolar region, then the efficiency of deposition in the tracheobronchial region may be defined in the first instance in the form

E'

TB --

OTB

= 1-

Nthor

( av)

(6.3)

Nthor

As will be seen below, this form has apparently been a convenient definition for the experimental determination of tracheobronchial deposition. However, for industrial hygiene purposes, it is more appropriate to define it as a fraction of what is inhaled. That is 141

Aerosol science for industrial hygienists __ (

Ntho r - Nal v ETB --

Nth~

Nalv

nh

Ninh

Ninh

)

(6.4)

- ( 1 - EET ) E' TB Now, if Dal v is the number of particles per unit time depositing in the alveolar region, then it is convenient again to express the efficiency of alveolar deposition as a proportion of what is inhaled. That is Dalv

Eal v =

(6.5)

Uinh At this point it should be noted that not all the particles entering the alveolar region are deposited. Some remain airborne for long enough to be carried back up through the respiratory tract during the exhalation half of the breathing cycle. If the number undeposited per unit time (i.e., exhaled) is N~xh then, as we shall see later, it is useful also to define the efficiency of penetration to the alveolar region ~ the so-called efficiency of respiration ~ as (D~lv + Ne• gresp "--

(6.6) Ninh

Similarly it is useful, again for industrial hygiene considerations, to consider the efficiency of aerosol deposition in the thoracic region as given by (DTB 4- Dalv)

(6.7)

Ethorde p --

Ninh or the efficiency of penetration into the thoracic region as given by (DTB + Dal v + Nexh)

(6.8)

Ethorpe n =

Ninh For the whole respiratory tract, we have the efficiency of total respiratory tract deposition (DET + DTB H- D~lv)

(6.9)

Etotdep =

Ninh 142

The inhalation of aerosols Finally, the efficiency of penetration into the respiratory tract is given by (DET + DTB -4- Dal v + Nexh)

(6.10)

Etotpen --

Ninh which, by reference to Figure 6.2, may be seen to be identical to A as given by Equation (6.1).

Concept of inhalability The first part of the overall process of aerosol exposure is the entry by inhalation (i.e., the aspiration) of aerosols from the ambient air and into the top of the respiratory tract. It is perhaps surprising to find that this part of the exposure process has received so little consideration over the years. In fact, it was only as recently as the late 1970s that the idea was first proposed that the only particles which could pose a potential risk to health by inhalation are those which are actually capable of entering the body during breathing (i.e., are inhaled), and that not all particles will necessarily have the same probability of entering. Thus emerged the concept of the human head as an aerosol sampler and, hence, of inhalability. Such thinking has led to a fresh approach to the establishment of relevant guidelines for the sampling of 'total' aerosol. This will be discussed in greater detail in Chapter 8.

Experimental measurements of inhalability Once it was decided that inhalability is an important part of the exposure process, experiments were conducted to obtain quantitative data for the aspiration efficiency of the human head over representative ranges of breathing, aerosol and external wind conditions. Whilst the original main driving force was the practical need to derive criteria for sampling healthrelated fractions in workplaces, it has always been recognised that the case of the aerosol exposure of humans in the ambient atmosphere is also important. Here an extended range of environmental conditions applies. In particular, whereas the windspeed in workplaces is usually less than 1 m s -1 and rarely exceeds 4 m s -1, it could reach 10 m s -1 ~ or even higher ~ at ground level in the outdoor environment. The first experimental studies were performed in Britain and Germany in the late 1970s and early 1980s. In these, experiments were carried out using life-sized human models (in the form of tailor's mannequins) in wind tunnels, generating comprehensive sets of data for the aspiration efficiency of the human head. A typical experimental set-up, for the large wind tunnel

143

Aerosol science for industrial hygienists

Figure 6.3. Typical wind tunnel set-up of the type suitable for investigating the aspiration efficiency of the human head (example shown is the large wind tunnel at the Institute of Occupational Medicine, Edinburgh, Scotland, U.K.).

at the Institute of Occupational Medicine (IOM) in Edinburgh, is shown in Figure 6.3. As discussed earlier in Chapter 4, aspiration efficiency (A) is a fundamental property describing the entry of particles into aerosol sampling devices. Here it is applied to the human head acting as an aerosol sampler. For present purposes, A may be described for each particle size and given external wind and breathing conditions by Equation (6.1). The experiments reported all employed the same basic methodology for obtaining A, with the test aerosol generated in the wind tunnel upstream of the life-sized mannequin and the upstream reference aerosol concentration obtained from samples collected using isokinetic thin-walled sampling probes (as will be described in Chapter 9). This was compared with the concentration of aerosol as sampled by the mannequin itself (where the 'inhaled' aerosol was collected onto filters mounted just behind the nose and/or mouth). The earliest such experiments were reported by Ogden and his colleagues (Ogden and Birkett, 1977; Ogden et al., 1977), followed by Vincent and Mark (1982) and Armbruster and Breuer (1982). These covered particle

144

The inhalation o f aerosols

aerodynamic sizes (dae) up to 100 Ixm, windspeeds in the range from close to zero to as high as 8 m s-~, and mannequin breathing flow rates corresponding to people at work. The results from these three early sets of experiments were notable for the consistency which they showed in terms of the broad tendencies exhibited by the data. The main trend was found to be the one between aspiration efficiency (A) and dae , with additional dependencies on breathing flowrate and windspeed which were found to be relatively weak (except at the higher end of the windspeed range). It was this broad consistency between data sets which first suggested the feasibility of applying the overall data towards new health-related definitions of 'total' aerosol based on the inhalability concept. Indeed, these earlier data were applied towards the first quantitative inhalability criteria which emerged during the early 1980s (see Chapter 8). The major trends found in the earlier studies were repeated in the results of the later work reported by Vincent et al. (1990). Those results are summarised concisely in Figure 6.4 in terms of A versus dae , where the envelopes shown enclose over 90% of the actual data reported. Here the clear relationship between A and dae is shown, with A first decreasing from unity as dae increases from zero, then tending to level off at around 0.5 for larger particle sizes. Importantly, there is no evidence of any 'cut-off' anywhere in the particle size range indicated. There is, however, a clear effect at the higher windspeeds (U) tested, notably a sharp upturn in A for larger particle sizes.

1.5

1.0

<

0.5

0

50 Particle

I00

aerodynamic diameter, dae (l~m)

Figure 6.4. Summary of recent experimental results for the aspiration efficiency of the human head (A) as a function of particle aerodynamic diameter (dae) (envelope of data in Vincent et al., 1990). Note that the hatched envelopes contain over 90% of the actual experimental data points for the ranges of conditions indicated.

145

Aerosol science for industrial hygienists

Physical basis of inhalability It is desirable to find a quantitative physical explanation for the observed trends. As already stated in Chapter 4, aspiration efficiency for sampling systems may be described in general as a function of particle aerodynamic diameter, windspeed, sampling flow rate, sampler body and orifice shapes and sizes and sampler orientation. However, as described later in Chapter 9, quantitative modelling of aspiration efficiency has only been achieved for very simple sampler configurations facing the wind. We are not yet close to a rigorous theory that fully accounts for the inhalability observations. Some progress has been made by assuming highly simplistic flow models (including, for example, two-dimensional geometry, perfectly spherical head, unidirectional sampling flow rate, etc.) (e.g., Dunnett and Ingham, 1988; Erdal and Esmen, 1995). But, whilst such mathematical models are instructive and represent progress towards the ideal goal, they are currently of limited practical value. The more realistic human head problem may be simplified somewhat by assuming fixed shape and dimensions (i.e., a 'standard' human head) and by imposing fixed cyclical breathing conditions (i.e., 'at work'). Then, blunt sampler theory based on the semi-empirical so-called 'impactionmodel' approach (Vincent and Mark, 1982; Vincent, 1989; Vincent et al., 1990) permits the functional statement s

A =f{

K, U )

(6.11)

where K -

d2 e U

(6.12)

is an inertial parameter which reflects the ability of particles to follow the streamlines of a distorted air flow just outside the nose and mouth. It is, in effect, a stop distance for particles moving near the human head acting as an aerosol sampler; or, alternatively, a Stokes' number in which the dimensional scale is fixed. To explore how the experimental data relate to Equation (6.11), the aspiration efficiency data from the later experiments are re-plotted in Figure 6.5 in a form that appears to give a fair collapse. In the absence of a quantitative physical model, we have to fall back on an empirical expression for describing the main observed trends. The form A* = A exp{ a ( 1 - U) } = [ exp{bK c} + d K ]

(6.13)

is consistent with Equation (6.11) but has physical meaning only in that A is a function of K and U as suggested by aspiration theory. Best-fit values for the coefficients are

146

The inhalation o f aerosols

!.0

I

0.5 O

II

C I 01

I 10 2

I !03 dae 2 U ( ~ m Z m

i 10 4

I 10 5

s -I)

Figure 6.5. New plot of the experimental results summarised in Figure 6.4, this time in terms of the modified aspiration efficiency, A* ( = A exp{a(1-U)} ) as a function of the inertial parameter dZeU (where dae is particle aerodynamic diameter and U is the windspeed) (envelope of data as presented in Vincent et al. 1990). This form of representation aims to draw attention to the role of particle inertia in aerosol inhalation.

a - 0.0267, b - - 0 . 0 3 9 3 , c - 0.370 and d - 1.04 • 10 -5 so long as dae is expressed in [l~m] and U in [m s -1].

6.4 E X P E R I M E N T S T O I N V E S T I G A T E A E R O S O L D E P O S I T I O N IN THE RESPIRATORY TRACT The second part of the process of aerosol exposure involves the arrival of particles at relevant sites inside the respiratory tract, as s u m m a r i s e d earlier. To investigate the first part ~ the inhalation phase ~ the use of m a n n e q u i n s , as described above, is an a p p r o p r i a t e e x p e r i m e n t a l m e t h o d since the physical factors involved are all essentially external to the h u m a n body. T h e only 'internal' factor is the b r e a t h i n g cycle and, whilst this can vary greatly from one person to the next and from one situation to another, such changes have b e e n shown to have relatively little effect on inhalability. H o w e v e r , for aerosols once they have arrived inside the body, b e h a v i o u r is influenced much m o r e strongly by h u m a n physiological factors. T h e r e f o r e , idealised laboratory model simulations are less a p p r o p r i a t e . A l t h o u g h some valuable information has been achieved using plaster cast models of airways, especially for the nose and the larger airways, the best results have b e e n o b t a i n e d using living volunteer h u m a n subjects. Such experiments began in the 1950s in the U n i t e d States with the pioneering work of T h e o d o r e H a t c h and co-workers

147

Aerosol science for industrial hygienists (e.g., Brown et al., 1950), and have been continued in m o r e recent years by workers in the United States and E u r o p e , notably by M o r t o n L i p p m a n n at New Y o r k University and Willi Stahlhofen at the Gesellschaft ffir Strahlenund U m w e l t f o r s c h u n g (Frankfurt) and their respective colleagues. This work has been driven largely by the need to u n d e r s t a n d the dosimetry of inhaled particles to the various parts of the respiratory tract as the basis for risk assessment. The experimental m e t h o d s for investigating respiratory tract deposition will not be described in detail here. T h e r e are several approaches, all involving volunteer subjects who are asked to inhaled idealised test aerosols of spherical particles of non-toxic material and known size u n d e r controlled specified breathing conditions. In most of the e x p e r i m e n t s r e p o r t e d in the literature, the test aerosols have been electrically neutralised prior to administration to eliminate one source of otherwise-uncontrolled variability or bias (unless, as in some cases, electrostatic forces were themselves the subject of enquiry). In one approach, the concentration of aerosols entering during inhalation is m e a s u r e d (e.g., using detectors based on the principles of light scattering) along with the concentration of those exiting during exhalation. For particles of given size, the difference b e t w e e n total

Figure 6.6. Typical experimental set-up of the type used at the Gesellschaft ftir Strahlen- und Umweltforschung (Frankfurt) for investigating respiratory tract deposition of inhaled particles in human volunteer subjects. During the inhalation cycle, aerosol and dilution air are mixed (top left) and pass through a pneuomotachograph flow meter, and then through an optical sensing region. The subject inhales aerosol through the mouthpiece (bottom right). Particles in the exhaled air are detected by the same optical system, and exit (by means of a non-return valve) through the expiration channel located at the aerosol-air mixer. (Diagram by courtesy of Josef Gebhart).

148

The inhalation of aerosols inhaled and exhaled aerosol provides the efficiency of total respiratory tract deposition. By the use of a bolus (i.e., a short 'burst' of aerosol injected at a given point in the inhalation part of the breathing cycle), additional information may be obtained about where in the respiratory tract deposition has taken place. A typical experimental set-up for this approach is shown in Figure 6.6. Another approach is to label the inhaled aerosol particles with some marker such that those which are retained in the respiratory tract can be detected non-invasively by external means. Most such studies in the past have involved radioactively labelling the particles (e.g., with a gamma emitter) and measuring the emissions from regions of the body using sensitive (e.g., gamma camera) techniques. More recently, sophisticated new detection methods for minute amounts of magnetic material (e.g., based on superconducting quantum interference devices (SQUIDS)) have provided an alternative method for some types of aerosol. In the following sections dealing with aerosol deposition in the various regions of the respiratory tract, results will be given which have been obtained using human subjects and a variety of such experimental methods.

6.5 E X T R A T H O R A C I C D E P O S I T I O N

Physical basis Extrathoracic deposition refers to particles which are deposited in the nose, mouth and throat. Since these regions are anatomically so different, it is reasonable to expect that the amount of aerosol that is deposited will differ greatly between nose and mouth breathing. In both, however, it is reasonable to predict that the principal mechanism of particle deposition there will be impaction, possibly augmented by gravitational settling. Some studies have been carried out to investigate whether, for electrically charged aerosols, there is likely to be any significant contributions from electrostatic deposition, but such contributions were shown to be small (Fry, 1970). For this region, therefore, it is suggested that an appropriate controlling parameter will be an inertial parameter along the lines of a Stokes' number. Thus, for a flow whose boundaries and dimensions are fixed, this can be expressed in the form of the combination, d2eQinh, where Qinh is the inhalation flow rate averaged over a breathing cycle. Qinh is directly proportional to the temporally and spatially-averaged airway velocity during inhalation.

149

Aerosol science for industrial hygienists

Experimental results The efficiency of extrathoracic deposition (EET) is defined for each particle size and set of breathing parameters by Equation (6.2). Experiments to determine EET have been reported by a number of groups of workers and some of the results are summarised together in Figure 6.7 for nasal breathing and mouth breathing, respectively. They are plotted in the form of EET versus da2eQinh suggested above, and cover a reasonably representative range of breathing patterns from 7.5 1 min -1 and 7.5 breaths min -1 to 20 1 min -1 and 15 breaths min -1. The hatched envelopes shown are representative of over 90% of the actual data assembled in the 1977 paper of Lippmann. Although many new such data have been published more recently, these envelopes remain quite representative and, indeed, have been influential in establishing criteria for regional respiratory tract deposition. The scatter in the results, as reflected in the width of the envelopes, is considerable. This is not surprising in view of the inter-subject variability and differences in experimental techniques from one laboratory to another. Nevertheless the results support the initial hypothesis that the main trend is indeed that between EET and d2e~inh . As expected, we see that EET increases steadily with daeQinh 2 (and hence as both dae and Qinh are increased independently), eventually levelling off at 100%. This confirms that impaction is the predominant deposition mechanism for both nose and mouth breathing. 2 The magnitude of EET for given daeQinh is clearly substantially greater for nose

1.0-

M o u t h br

Nose breathing

~,,~~~

EET 0.5

0 i0 !

102

-

103

I 104

dae 2 Qinh(p'm21 min "1)

Figure 6.7. Efficiency of extrathoracic deposition of aerosols in human subjects (EET) as a function of the inertial parameter, dae2Qinh (envelope of data originally collected and summarised by Lippmann, 1977).

150

The inhalation of aerosols

than for mouth breathing. This means that the initial dose of particles to the nose is much greater than to the mouth region. Conversely, it also means that the penetration of particles below the larynx (as reflected in the magnitude of Nthor ) is much less for nose breathing than for mouth breathing. That is, the nose is a more effective filter for protecting the lung from inhalation exposure.

6.6 T H O R A C I C D E P O S I T I O N

Physical basis In the large airways of the lung just below the larynx ~ the tracheobronchial region ~ it is again a reasonable starting point to assume that particle deposition will be by impaction in the relatively high-velocity air flow there. Again, as for extrathoracic deposition, possible contributions from other mechanisms, especially electrostatic forces for charged aerosols, have been considered but have been shown to be generally small except under extreme conditions of particle charge and concentration (Chan et al. 1978; Yu, 1985). Particles which fail to be filtered out in the larger airways of the tracheobronchial region arrive at the alveolar region of the deep lung. Here, the physical scenario changes significantly. Now, because of the very low ~ and eventually vanishingly small ~ air velocities, particle impaction is no longer an effective deposition mechanism. Instead, for particles which are large enough, deposition will be by gravitational settling. Obviously this effect will be greatest for the particles penetrating with the largest aerodynamic diameter and for the longest residence time of particles in the lung. It should be noted that although such small particles (no more than a few micrometres) have very small terminal settling velocities, they have very short distances to fall in the small (order of 300 Ixm diameter) alveolar sacs before they reach the lung wall. But for very small particles, diffusion dominates particle transport and so controls deposition. As discussed in Chapter 4 and shown in Equation (4.55), the coefficient of diffusion ~ and hence the efficiency of deposition ~ increases as the geometrical particle size decreases in this region of particle size. In the intermediate range of particle size, it therefore follows that deposition efficiency goes through a minimum. Particles which fail to be deposited in the alveolar region remain airborne and so are expired during the exhalation phase of the breathing cycle. For the alveolar region, unlike for the other regions, a significant contribution to deposition from electrostatic forces has been observed in some experiments. This has some important potential implications, particularly for certain types of aerosol, and will be discussed separately below.

151

Aerosol sciencefor industrial hygienists Tracheobronchial deposition By virtue of the experimental methods used, the efficiency of particle deposition in this region for particles of given aerodynamic diameter and breathing parameters has usually been defined in the first instance by E'TB as given by Equation (6.3). Typical experimental data are plotted in this way, again as a function of the impaction parameter d2eQinh, in Figure 6.8 for ranges of breathing parameters similar to those for the extrathoracic deposition described in the previous section. Again, the envelope shown is based on relatively early ~ but still representative data assembled by Lippmann (1977). Such results confirm that E'TB v e r s u s d2eQinh is indeed the main trend, and hence that impaction is the predominant deposition mechanism. E'TB increases steadily as d2eQin h increases, eventually levelling off at 100% for large enough particles or high enough flow rates. The original results embodied in Figure 6.8 were reported for volunteers who were non-smokers. Lippmann also reported similar data for cigarette smokers (not shown) which exhibit not only even wider scatter but also a significant tendency towards enhanced tracheobronchial deposition. This suggests some narrowing of the airways for those volunteers who were smokers. We return to the fact that E'TB as plotted in Figure 6.8 reflects only the deposition in that region as a fraction of what enters it. But it would appear

1.0--

Non-smokers

~TB

0.5 --

0 I0 !

10 2

10 3

dae 2 Qinh

I 10 4

($1'm21min -t)

Figure 6.8. Efficiency of tracheobronchial deposition of aerosols in human 2 h (envelopeof subjects (E'TB) as a function of the inertial parameter, daeQi~ data collected and summarised by Lippmann, 1977).

152

The inhalation of aerosols 1.0

0.5

0 102

103

104

dae2 Qinh (~m2 ! rain -I)

Figure 6.9. Efficiency of tracheobronchial deposition of aerosols in human subjects as a proportion of inhaled aerosol (Exl3), combining the results for EET and E'xB as shown in Figures 6.7 and 6.8 according to Equation (6.4).

that this form of presentation is a result of the way in which the experiments were performed. If such results are to be subsequently used for assessing dose, it is more useful to also relate tracheobronchial deposition to the inhaled aerosol concentration. So the modified definition, ETB, given by Equation (6.4) is more appropriate. Now it is clear that the trends would be very different, with ETa tending towards zero at large values of d2eOinh, reflecting the non-availability of particles due to their removal by deposition higher up in the extrathoracic region. It is worth noting again that, for tracheobronchial deposition, there is a substantial difference between nose and mouth breathing due to the effect of the differences already referred to in EET. The resultant trend in ETB, based on the results for EET and E'TB shown in Figures 6.7 and 6.8, is illustrated in Figure 6.9 for mouth breathing non-smokers. Here the effect of the scatter in the original data as presented by Lippmann is very marked.

Alveolar deposition The efficiency of alveolar deposition is usually defined for each particle size and set of breathing parameters by Equation (6.5). Typical data (again collected by Lippmann, 1977) are summarised in Figure 6.10. Here, because impaction no longer makes a significant contribution to deposition, Eal v is

153

Aerosol science for industrial hygienbsts 1.0-

0.5

0.1

I I

I

d (Is,m)

I

I !0

dac (Is,m)

I Figure 6.10. Efficiency of alveolar deposition of aerosols in human subjects as a proportion of inhaled aerosol (Ealv) as a function of particle diameter (dae) (envelope of data collected and summarised by Lippmann, 1977). The envelope is plotted as a function of d~e for the larger particle size range and particle geometrical diameter (d) for the smaller particle size range.

now plotted simply as a function of particle size (dae for particles larger than about 0.5 txm where gravitational settling dominates and geometrical diameter d for smaller particles where diffusion takes over). The data plotted in this way show that Eal v falls to zero for particles with d~e larger than about 10 Ixm, simply as before for ETB ~ reflecting the non-availability of larger particles due to the filtration effect of the upper respiratory tract. Eal v exhibits a minimum for particles with diameter around 0.5 txm, then begins to increase again due to increasing deposition by diffusion as particles thereafter get progressively smaller. In recent years, experiments have been carried out to investigate the effect on Eal v of different breathing parameters. Figure 6.11 indicates how Eal v differs for contrasting breathing conditions corresponding broadly to 'at rest' and 'at work' (from Vincent and Mark 1984 based on results originally reported by Stahlhofen et al., 1981). Here it is seen that Eal v takes greater values for the lower breathing rate, directly as a result of the longer residence time of particles in the deep lung. It does not follow that the actual amount deposited will be greater because the increase in deposition efficiency will be offset to some extent by a corresponding decrease in the volume of air inspired. Whether or not the amount deposited is actually greater or less will depend on these factors together with the details of the particle size distribution of the inhaled aerosol.

154

The inhalation of aerosols

,0[ / ' a t rest'

// 0.5

~~

'at work'

,

,'v----

5

o

!o dae (l~m)

Figure 6.11. Efficiency of alveolar deposition of aerosols in human subjects (Eaiv) as a function of particle aerodynamic diameter (dae) for contrasting breathing patterns corresponding to the different work rates indicated (from Vincent, J.H. and Mark, D., Annals of Occupational Hygiene, Copyright 1984, reproduced by permission of the British Occupational Hygiene Society).

Thoracic deposition and penetration T o t a l efficiency of p a r t i c l e d e p o s i t i o n in t h e t h o r a c i c r e g i o n , E t h o r d e p , m a y be o b t a i n e d f r o m t h e s u m of the i n d i v i d u a l t r a c h e o b r o n c h i a l a n d a l v e o l a r 1.0

o=

0.5

o

0.I

I

I

5

I0

15

dae (p,m)

Figure 6.12. Efficiency of thoracic penetration of aerosols inhaled by mouthbreathing humans (Ethorpen) as a function of particle aerodynamic diameter (dae), adjusted to an inhalation flowrate of 43.51min -1 flowrate (envelope of data summarised in ACGIH, 1985). Note that these data were obtained by reference to measurements of the efficiency with which particles are deposited in the extrathoracic region.

155

Aerosol science for industrial hygienists deposition efficiencies as described by Equation (6.7) and data like those summarised in Figures 6.9 and 6.10. However, in relation to their potential application in criteria for health-related aerosol measurement, it might be considered more appropriate to consider the efficiency of thoracic penetration, Ethorpen, as given by Equation (6.8). This is simpler and may be obtained using Equation (6.2) from knowledge only of the efficiency of extrathoracic deposition (as shown in Figure 6.7). This, for example, was the approach taken in the 1985 Report of the Air Sampling Procedures Committee of the American Conference of Governmental Industrial hygienists (ACGIH) as part of the progress towards new particle sizeselective criteria for health-based aerosol sampling (see Chapter 8). The results are summarised in Figure 6.12. They relate to a mouth-breathing human subject with an average flowrate during the inhalation half of the breathing cycle adjusted to 43.5 1 min -1 (corresponding to a minute volume of about 20 1 min-1).

6.7 T O T A L R E S P I R A T O R Y T R A C T D E P O S I T I O N From all the preceding, it is useful to draw together all the information and to examine the efficiency of total aerosol deposition in the respiratory tract.

1.0

--

0.5

m

0

I 1

0.

I 10

dae (~m)

Figure 6.13. Efficiencyof total lung deposition of aerosols inhaled by humans (Etotdep) as a function of particle aerodynamic diamter (dae) (envelope of data collected and summarised by Lippmann, 1977).

156

The inhalation of aerosols D a t a d r a w n f r o m t h e s a m e e x p e r i m e n t s as for t h e d a t a a l r e a d y p r e s e n t e d a r e s u m m a r i s e d in F i g u r e 6.13 ( a g a i n see L i p p m a n n , 1977), w h e r e Etotdep ~ as given by E q u a t i o n (6.9) ~ is n o w p l o t t e d as a f u n c t i o n of dae. T h i s p l o t s h o w s clearly t h a t , for p a r t i c l e s g r e a t e r t h a n a b o u t 10 txm, v i r t u a l l y all t h e p a r t i c l e s w h i c h are i n h a l e d a r e d e p o s i t e d s o m e w h e r e in t h e r e s p i r a t o r y tract. B u t for p r o g r e s s i v e l y s m a l l e r p a r t i c l e s , an i n c r e a s i n g p r o p o r t i o n r e m a i n s a i r b o r n e a n d so is s u b s e q u e n t l y e x h a l e d . This e x h a l e d p r o p o r t i o n r e a c h e s its m a x i m u m w h e n Etotdep r e a c h e s its m i n i m u m . Example 6.1. Assume that the particle size distribution shown in Figure 3.8 (Chapter 3) represents the mass of aerosol inhaled over a given time interval. Estimate the mass of inhaled aerosol which is subsequently exhaled. Refer to the plot of Etotdep versus dae shown in Figure 6.13, using the line which bisects the envelope containing the experimental data. Also refer to the cumulative particle size distribution as given by Figure 3.8b. All the information needed can be read off (to a fair approximation) from these. Note that the efficiency of exhalation, Eexh, is given by Eex h -" 1--Etotdep From the data on these two graphs, create the following table"

Range of dae (l~m) 012345678910-

1 2 3 4 5 6 7 8 9 10 50

TOTALS

*

Midpoint (txm)

Mass in range (mg)

Eex h

Mass

(mg)

exhaled

0.5 1.5 2.5 3.5 4.5 5.5 6.5 7.5 8.5 9.5 30.0

0.1 0.3 0.7 0.9 0.9 0.9 0.8 0.7 0.7 0.6 5.5

0.70 0.60 0.50 0.40 0.30 0.10 0.05 0.03 0.02 0.00 0.00

0.07 0.18 0.35 0.36 0.27 0.09 0.04 0.02 0.01 0.00 0.00

12.0 (mass inhaled)

The fraction exhaled = 1.39/12.0 ~ 12%

157

1.39 (approx.) (mass exhaled)

Aerosol science for industrial hygienists

In Figure 6.13, it is not surprising that the scatter is wide since, although Etotdep versus dae is a main trend, there are other significant relationships. These include Qinh and also the period of the breathing cycle. It is reasonable to expect a better collapse of the data (and hence less scatter) if Etotdep were to be plotted in terms of a physically appropriate combination of these variables. Heyder et al. (1980) sought such a parameter and proposed the empirical expression X M - (lnQinh- 1.43)In(daZe T24V'Qinh)

(6.14)

where T is the inhalation time (equivalent to half the period of the breathing cycle). They plotted measured values for Etotdep against this new quantity for three volunteer human subjects for particles in the range of d~e from about 0.5 to 8 Ixm, for Qinh in the range 7 . 5 - 60 1 min -1, and for T in the range 1 - 8 s. A very good collapse was obtained.

6.8 DEPOSITION OF FIBROUS A E R O S O L S Already in this book, fibrous aerosols have been identified as a special case in relation to occupational health. Because of the known serious health risks associated with inhaling such aerosols, it is important to know something about how and where they are deposited in the respiratory tract. At the same time, because of those health risks and because a 'safe' fine fibrous material has not yet been discovered, it has not so far been possible to conduct inhalation experiments with human volunteer subjects with such aerosols. Therefore, it is necessary to try to infer the nature of respiratory tract deposition for long, thin fibres from what is known about the aerodynamic behaviour of fibres obtained in other aerosol experiments and from animal inhalation studies. Because deposition in most respects is governed by aerodynamic diameter, then much of what has been presented so far with regard to regional and total deposition may apply to fibres. This means that, in the alveolar region, gravity remains the dominant influence, along with diffusion (although a long fibre, by virtue of its large geometrical size, has a smaller diffusion coefficient than for the corresponding isometric particle). However, for fibres of long aspect ratio in the sorts of flows and geometrical boundary conditions pertaining to the respiratory tract, other effects not expected to be significant for isometric particles may contribute significantly to deposition. These include orientation effects on particle motion (e.g., Yu et al., 1986) and deposition by interception (Timbrell, 1970). Such contributions to fibre deposition occur particularly at or near bifurcations in the conducting airways. Electrostatic deposition can also be important for fibres, as discussed in the following section.

158

The inhalation of aerosols 6.9 E L E C T R O S T A T I C R E S P I R A T O R Y T R A C T D E P O S I T I O N As indicated earlier, electrostatic forces do not play a significant role in particle deposition in the extrathoracic and tracheobronchial regions of the respiratory tract. However, there may be some situations where they could be influential in alveolar deposition. It i s worth noting that most of the deposition data summarised in the graphs of preceding sections were obtained for aerosols which had been electrically neutralised (to close to Boltzmann equilibrium) prior to inhalation. Therefore, they reflect deposition only by the mechanical mechanisms of impaction, interception, gravitational settling and diffusion. From those data, it is clear that, for particles which are large enough for there to be efficient deposition by those mechanical mechanisms, any scope for further contributions (or enhancement of deposition efficiency) by other forces is limited. But in the region where the combined effects of the various mechanical deposition mechanisms are relatively weak ~ that is, near the minimum in the total deposition efficiency curve (see Figure 6.13) ~ there is scope for enhancement by such other means. Since that minimum in the total deposition curve is associated mainly with the deposition characteristics of the alveolar region, it is therefore in the alveolar region where there is particular scope for electrostatic enhancement of deposition. In general, for an aerosol of charged particles, electrical forces can arise from the effects of space charge. That is, charged particles within an aerosol may interact with one another by virtue of the Coulomb-type forces which bring about mutual attraction or repulsion depending on whether particle charges are of the same or opposite polarities (i.e., like charges repel, opposite charges attract). In a population of particles, the net result of such individual interactions integrated over the whole cloud is an overall internal electric field by which particles can be transported from one region to another and, ultimately, be deposited (or precipitated) onto solid surfaces. However, for this space charge effect to become significant requires relatively high aerosol concentration and charge level. It has been argued that, although this could occur in the upper respiratory tract under extreme exposure conditions, such conditions are not likely to be found for aerosols under conceivable workplace conditions (Yu, 1985). In any case, it is not likely to operate at the alveolar level. The main reason here is that, by the time the aerosol has reached the alveoli, many particles will have been filtered out by the upper respiratory tract. Furthermore, by virtue of the sheer number of alveoli (as many as 8 million), it is unlikely that sufficient particles would be found together in a single alveolar sac for multi-particle space charge effects to become significant. Image forces provide a more feasible mechanism for electrostatic deposition, operating on an individual particle basis. The mechanism is shown schematically in its simplest form in Figure 6.14a, where a charged particle close to a flat 159

Aerosol science for industrial hygienists

/ / I I I / /

Charged aerosolparticle \,•

[+q] -/Q~I~

Imageofcharged

aerosolparticle \1/

/

-to,. I-q]

(a) Conductingsurface

Charged

aerosol[+q]particle ~ ~ ~ ~ . ~ / •

aerosoilm ~age ofparticle charged [-Rqlx]

1 I

Conductingspherical

[

[

[

*-- R---q

I I R21X

=

., ]

Figure 6.14. Schematic to show the nature of electrostatic deposition by image forces of charged particles (a) near a conducting flat plate, and (b) inside a closed conducting sac (reprinted from Vincent, J.H., 'On the practical significance of electrostatic lung deposition of isometric and fibrous aerosols'. Journal of Aerosol Science, 16, 511-519, Copyright 1985, with kind permission from Elsevier Science Ltd, The Boulevard, Langford Lane, Kidlington OX5 1GB, UK).

conducting surface forms an image of itself, equal in magnitude but opposite in polarity. The Coulomb force between the particle and its image (charges +q and - q respectively) is always attractive, regardless of particle charge polarity, so that the particle is accelerated towards the conducting wall, whereupon it may be deposited. The magnitude of the force is given by

160

The inhalation of aerosols

q2 (6.15)

FCoulom b --

47r%(2x) 2 where eo is the permittivity of a vacuum (a universal constant given by 8.85 • 10 -12 As g -1 m -1) and x is the distance of the particle from the conducting surface. For the alveolar region of the lung, the lung wall becomes the conducting surface of interest, this time approximately in the form of a conducting spherical shell. The image force model for this modified picture is shown in Figure 6.14b, and the force itself may be shown from classical electrostatic image theory to be xRq 2

(6.16)

Fcoulom b -41r~o(R2-x2)2

where R is the radius of the spherical alveolar sac and x is now the distance of the particle from the centre of the sac. From Equations (6.15) and (6.16) it is seen that the electrostatic force increases sharply as the particle gets closer to the wall. In fact, it becomes substantial only at very close range, and so is an effective overall deposition mechanism only in very small confined spaces (such as a small alveolar sac). In Chapter 3, the electrical charge properties of workplace aerosols were discussed and it was described how all such aerosols observed were found to be charged to levels well above Boltzmann equilibrium (Johnston et al., 1985), some more so than others. Whether or not the actual levels of charge are sufficient to bring about significant enhancement of alveolar deposition depends on the relationship between electrical and gravitational forces (Vincent, 1985). The balance is dependent on particle size, on

Alveolar volume (last 10 generations)

L

~~

Airway length x (generation)

Figure 6.15. Schematic of one of the earliest lung deposition models, indicating the increasing cross-sectional area of the lung for progressively deeper airway generations (Taulbee and Yu, 1975).

161

Aerosol science for industrial hygienists

gravitational forces as contained in the definition of aerodynamic diameter, and on electrical forces by virtue of the charge-per-particle relationship as contained in Equation (3.14). The crucial point is whether a particle which is small enough to remain airborne in the lung (i.e. with d~e = 0.5 txm, close to the minimum in the alveolar deposition curve) can carry enough charge for electrostatic forces to become significant. From experimental and theoretical evidence on lung deposition, from experiments with both humans and animals, together with knowledge of charge levels on workplace aerosols, it would appear that alveolar deposition is not significantly enhanced for workplace aerosols consisting of particles of non-extreme shape (i.e., isometric dust particles, platelets, etc.). On the other hand, for long, thin fibres (such as asbestos particles), there can be significant contributions to lung deposition from electrostatic forces. Indeed, experiments with rats inhaling amosite and chrysotile asbestos charged to levels very similar to those found in actual asbestos textile industry workplace aerosols have revealed electrostatic enhancement of lung deposition by a factor of up to two times (Jones et al., 1983). Such a marked effect for fibres can be explained on physical grounds. Most asbestos fibres are very thin (with diameter less than 1 Ixm) and so may be shown to be very small in terms of their aerodynamic diameter (see Equations (4.34) and (4.35)). Such fibres can therefore penetrate very effectively down to and remain airborne in the deep lung. But they are geometrically large by virtue of their length and so in relation to isometric particles of equivalent aerodynamic diameter ~ can carry a disproportionately high amount of electrical charge. For such aerosols, therefore, we should be aware of the possible role of significant electrostatic deposition in assessing the initial lung dose of exposed workers, especially in view of the well-known risk to health.

6.10 M A T H E M A T I C A L M O D E L L I N G OF L U N G D E P O S I T I O N Another approach to understanding the nature of particle deposition in the respiratory tract, complementing experiments with humans and animals, is the use of mathematical and/or numerical models. The ultimate goal is to achieve a computer-based simulation that can be demonstrated to work effectively over all breathing and morphological conditions and for all types of aerosol. When that has been achieved, then experiments with human subjects will no longer be necessary for aerosol dosimetry. Early modelling attempts were based on relatively simple simulations of lung morphometry. For example, the model proposed by Taulbee and Yu (1975) treated the airway system as analogous to a single 'horn-shaped' channel of increasing cross-sectional area, using Weibel's model to provide the functional form relating channel length (equivalent to generation number) 162

The inhalation of aerosols

and cross-section and allowing for the expansion and contraction of the alveolar volume during the breathing cycle. This simple representation is shown in Figure 6.14. Taulbee and Yu mathematically described the airflow in this lung simulation, also including the mixing associated with the 'apparent' diffusion coefficient arising from the distribution of velocities in different tubes of the same generation. They then set up a mass balance for successive elements of the simulated lung system, and modelled the loss of particles to the walls by inertial impaction, gravitational settling and diffusion. Despite the apparent simplicity of the simulation chosen, the results of the calculations were found to be in quite good agreement with the available human data. Since then, the increased availability of computers has enabled more complicated models based on more representive application of lung morphology information like that described in the lung model morphometry model proposed by Weibel (1963). In such models, not only is deposition (by inertial impaction, interception, gravitational settling and diffusion) determined for particles in the lung ducts themselves at each generation but also at the branching bifurcations. This has led to significant progress. for example, the calculated results of Yu and Diu (1983) based on this improved approach were in quite good agreement with the experimental data of Stahlhofen et al. (1980). But new and even more refined models continue to appear. One, for example, is the stochastic model of Koblinger and Hofmann (1990) in which the histories of a large number of particles passing through a complex system of airways is analysed by the application of Monte Carlo techniques. Other studies have identified specific features of interest. Gradon and Orlicki (1990) used their model of particle deposition in the tracheobronchial tree to identify deposition 'hot spots' at bifurcations. Yu (1985) incorporated electrostatic forces and, again, obtained good agreement with the available experimental data. The balance between the factors influencing deposition shifts somewhat for fibrous particles, and Timbrell (1965) was the first to note the relatively stronger role of direct interception. Since then, a number of mathematical models have emerged for the special case of the lung deposition of fibres (e.g., Asgharian and Yu, 1989). In their model, Podgorski and Gradon (1990) dealt with the interesting case of a flexible fibre whose shape during transport in lung airways can be deformed. This might be particularly relevant to chrysotile asbestos.

REFERENCES American Conference of Governmental Industrial Hygienists (ACGIH). (1985). Particle size-selective sampling in the workplace. Report of the Technical Committee on Air Sampling Procedures, ACGIH, Cincinnati, OH.

163

Aerosol science for industrial hygienists Armbruster, L. and Breuer, H. (1982). Investigations into defining inhalable dust. In: Inhaled Particles V (Ed. W.H. Walton) Pergamon Press, Oxford, pp. 21-32. Asgharian, B. and Yu, C.P. (1989). Deposition of fibres in the rat lung. Journal of Aerosol Science, 20, 355-366. Brown, J.H., Cook, K.M., Ney, F.G. and Hatch, T. (1950). Influence of particle size upon the retention of particulate matter in the human lung. American Journal of Public Health, 40, 450--458. Chan, T.L., Lippmann, M., Cohen, V,W. and Schlesinger, R.B. (1978). Effect of electrostatic charges on particle deposition in a hollow cast of the human larynx-tracheobronchial tree. Journal of Aerosol Science, 9, 463-468. Dunnett, S.J. and Ingham, D.B. (1988). The human head as a blunt aerosol sampler. Journal of Aerosol Science, 19, 365-380. Erdal, S. and Esmen, N.A. (1995). Human head model as an aerosol sampler: calculation of aspiration efficiencies for coarse particles using an idealised human head model. Journal of Aerosol Science, 26, 253-272. Fry, F.A. (1970). Charge distribution of polystyrene aerosols and deposition in the human nose. Journal of Aerosol Science, 1, 135-146. Gradon, L. and Orlicki, D. (1990). Deposition of inhaled aerosol particles in a generation of the tracheobronchial tree. Journal of Aerosol Science, 21, 3-19. Heyder, J., Gebhart, J., Rudolf, G. and Stahlhofen, W. (1980). Physical factors determining particle deposition in the human respiratory tract. Journal of Aerosol Science, 11,505-515. Hinds, W.C. (1982). Aerosol Technology. John Wiley and Sons, New York. Johnston, A.M., Vincent, J.H. and Jones, A.D. (1985). Measurements of electric charge for workplace aerosols. Annals of Occupational Hygiene, 29, 271-284. Jones, A.D., Johnston, A.M. and Vincent, J.H. (1983). Static electrification of airborne asbestos dust. In: Aerosols in the Mining and Industrial Work Environment (Eds. V.A. Marple and B.Y.H. Liu). Ann Arbor Science, Ann Arbor, MI, Chapter 46, pp. 613-631. Koblinger, L. and Hofmann, W. (1990). Monte-Carlo modelling of aerosol deposition in human lungs. Part I: simulation of particle transport in stochastic lung structures; Part II: deposition fractions and their sensitivity to parameter variations. Journal of Aerosol Science, 21, 661-688. Lippmann, M. (1977). Regional deposition of particles in the human respiratory tract. In: Handbook of Physiology; Section IV, Environmental Physiology (Eds. D.H.K. Lee and S. Murphy). Williams and Wilkins, Philadelphia, PA, pp. 213-232. Netter, F.H. (1980). The CIBA Collection of Medical Illustrations: Volume 7, Respiratory System, 2nd Edn. CIBA-GEIGY Corporation, Summit, NJ. Ogden, T.L. and Birkett, J.L. (1977). The human head as dust sampler. In: Inhaled Particles IV (Ed. W.H. Walton). Pergamon Press, Oxford, pp. 93-105. Ogden, T.L., Birkett, J.L. and Gibson, H. (1977). Improvements to dust measurement techniques. IOM Report No. TM/77/ll, Institute of Occupational Medicine, Edinburgh, U.K. Podgorski, A. and Gradon, L. (1990). Motion and deposition of fibrous flexible particles in laminar gas flow through a pipe. Journal of Aerosol Science, 21, 957-968. Stahlhofen, W., Gebhart, J. and Heyder, J. (1980). Experimental determination of the regional deposition of aerosol particles in the human respiratory tract. American Industrial Hygiene Association Journal, 41, 385-398. Stahlhofen, W., Gebhart, J. and Heyder, J. (1981). Experimental determination of the regional deposition of aerosol particles in the human respiratory tract. American Industrial Hygiene Association Journal, 41, 385-398. Taulbee, D.B. and Yu, C.P. (1975). A theory of aerosol deposition in the human respiratory tract. Journal of Applied Physiology, 38, 77-85.

164

The inhalation of aerosols Timbrell, V. (1970). Inhalation of fibres. In: Proceedings of the 3rd International Conference on the Pneumoconioses (Ed. H.A. Shapiro), held in Johannesburg, April-May 1969, pp. 3-9. Vincent, J.H. (1985). On the practical significance of electrostatic lung deposition of isometric and fibrous aerosols. Journal of Aerosol Science, 16, 511-520. Vincent, J.H. (1989). Aerosol Sampling: Science and Practice. John Wiley and Sons, Chichester, U.K. Vincent, J.H. and Mark, D. (1982). Application of blunt sampler theory to the definition and measurement of inhalable dust. In: Inhaled Particles V (Ed. W.H. Walton). Pergamon Press., Oxford, pp. 3-19. Vincent, J.H. and Mark, D. (1984). Inhalable dust spectrometers as versatile samplers for studying dust-related health effects. Annals of Occupational Hygiene, 28, 117-124. Vincent, J.H., Mark, D., Miller, B.G., Armbruster, L. and Ogden, T.L. (1990). Aerosol inhalability at higher windspeeds. Journal of Aerosol Science, 21,577-586. Weibel (1963). Morphometry of the Human Lung. Springer, Berlin. Yu, C.P. (1985). Theories of electrostatic lung deposition of inhaled aerosols. Annals of Occupational Hygiene, 29, 219-227. Yu, C.P. and Diu, C.K. (1983). Total and regional deposition of inhaled aerosols in humans. Journal of Aerosol Science, 14, 599-609. Yu, C.P., Asgharian, B. and Yen, B.M. (1986). Impaction and sedimentation deposition of fibers in airways. American Industrial Hygiene Association Journal, 47, 72-77.

165

CHAPTER 7

The fate of inhaled particles 7.1 I N T R O D U C T I O N Although, strictly, an aerosol ceases to exist as such when its particles have been deposited (and so are no longer airborne), the interest of the industrial hygienist in inhaled particles does not end there. In understanding the overall picture of the risk associated with exposure to airborne particles by the inhalation route, the arrival of particles at their initial site of deposition in the respiratory tract is only the first stage the exposure stage ~ in a series of events which might, under a given set of circumstances unfavourable to the worker, lead to ill-health. Those events include re-distribution, clearance, storage and biological response. These may lead in turn to pathological changes and manifestations of ill-healIh that can be identified clinically. This complex chain of processes is summarised in Figure 7.1 for a type of aerosol consisting of insoluble particles capable of producing a fibrotic response or emphysema (as, for example, in exposure to quartz-containing mineral dusts like those found in some mineral extraction industries). The picture might look somewhat different for aerosols capable of producing other types of health outcome (e.g., carcinogenic or allergenic). Such a description is far from being complete or general. But it does serve to give some sense of the complexity of the overall system of responses linking exposure with ill-health.

7.2 B I O L O G I C A L M E C H A N I S M S OF C L E A R A N C E A N D RE-DISTRIBUTION In Chapter 6, the form and function of the human respiratory tract was outlined as the basis for discussing particle deposition. Now that discussion is expanded to facilitate a further description about what happens to particles after deposition. Mucociliary clearance takes place in parts of the nose and in the tracheobronchial region of the lung. In the tracheobronchial region, particles are

166

The fate of inhaled particles [ Exposure[

l

Inhalation Breathing parameters Redistribution "= = Lung deposition

= Clearance

Cumulative lung burden

i Dose of 'harmfulness' J"1 /

of the inhaled material ,]

Biological response ~ ~

Leukocyte stimulation [

~

Cell damage

= Leukocyte recruitment -

]

1

Leukocyte accumulation (inflammation) Removal of stimulus

Persistence of stimulus

Return to normal

Chronic inflammation

!

1

/ Repair

!

[Fibrosis ]

,,,,.

Destruction

1

[ Emphysema l

Figure 7.1. S u m m a r y of chain of processes linking exposure, dose and response for insoluble particles capable of causing a fibrotic response.

deposited onto the mucous layer which moves under the 'beating' action of the cilia. This mechanical clearance process is shown schematically in Figure 7.2. The action of the cilia is such that particles are transported in the upwards direction in the respiratory tract. When they reach the epiglottis, they are swallowed and so ultimately end up being cleared to the gastrointestinal tract. Such clearance is quite fast, with linear ciliary transport rates of magnitude up to a few millimetres per second. As a result, the majority of particles deposited in the tracheobronchial region are removed within a few hours. The small proportion of particles which do not deposit directly onto the mucous layer can penetrate the layer of cilia and become attached to the respiratory tract wall (Gehr et al., 1990). Here they may be retained for significantly longer. Although the overall proportion of particles thus retained is thought to be very small, such long-term retention could be important especially for highly-toxic substances. For particles depositing in the alveolar region, clearance takes place via the scavenging and engulfment of particles by free alveolar macrophage cells

167

Aerosol science for industrial hygienists

Figure 7.2. Simple picture of the process of ciliary action for eliminating particles deposited in the conducting airways of the lung. Here the particle is conveyed out of the lung on the mucous layer which is being driven by the 'beating' action of the cilia.

(i.e., by phagocytosis), as shown schematically in Figure 7.3. Such cells, of dimension of the order of 15 txm, are always present in the alveolar spaces, even in a lung which is not being challenged by the inhalation of airborne particles. However, the arrival of foreign particles at the lung wall serves to stimulate the recruitment of additional macrophages, thus increasing the overall capacity for phagocytosis as part of the lung's defence mechanism. Phagocytosed particles are subsequently transported (carried by and the macrophages) towards the mucociliary escalator by chemotaxis then on to the gastrointestinal tract. Clearance by this route can be quite fast, with the major proportion eliminated within a few days. H o w e v e r , there are particles which, although they may be phagocytosed relatively rapidly after deposition, become 'fixed', ' e m b e d d e d ' or 'sequestrated'. These are

Figure 7.3. Simple picture of the process of phagocytosis of particles deposited in the alveolar region of the lung. Here the particle is engulfed by the free alveolar macrophage, after which it may be carried towards the mucociliary escalator and thence removed from the lung.

168

The fate of inhaled particles eliminated much more slowly, on a timescale which might be up to several hundred days or even longer. Lymph nodes are located throughout the body. Those associated with the lung play an important role in what happens after particles have been deposited in the lung. These particular lymph nodes are small organs located close to the lung to which they are connected via lymphatic ducts (see Figure 7.4 which shows the location of a typical tracheobronchial lymph node). Material from the lung (i.e., fluid, cells, foreign material) can drain into the lymph nodes through these ducts. The nodes are effectively closed receptacles, originally evolving as sinks for bacteria and infection. They are part of the body's immune response system and, as such, are primary initiators of the body's defence mechanisms. The lung disposes of foreign particles of matter in much the same way as it would biological organisms. So particles deposited in the lung, in addition to being eliminated towards the gastrointestinal tract, may also be translocated to the lymph nodes where, unless the particles are soluble in lymphatic tissue, they will remain (since there is no significant pathway for further removal). For some fibrous particles, although they may be fine enough (in terms of diameter) that they can penetrate aerodynamically down to the alveolar region, they may be of such physical dimensions (i.e., in terms of length) that phagocytosis ~ and in turn clearance ~ may be impaired. For example, an asbestos fibre of diameter less than 1 ~m and length 40 ~m may be quite capable of reaching and depositing in the alveolar region. Yet an alveolar macrophage with dimension only of the order of 15 ~m has considerable difficulty in engulfing such a particle. This might explain the preferential

Figure 7.4. Simple picture of a lung-associated lymph node bronchial lymph node).

169

(e.g., a

tracheo-

Aerosol science for industrial hygienists

long-term retention of the longer fibres as observed by some workers (e.g., Morgan et al., 1978). Long, thin, needle-like fibres in the alveolar region also have the capacity to pierce the interstitium of the lung wall and thereafter to migrate to other parts of the body, close to, or even remote from, the lung itself. It is well-known, for example, that fine asbestos fibres can penetrate outside the lung in this way and may reach the pleural cavity. This is the site of the serious disease of mesothelioma, thought to be uniquely associated with exposure to asbestos and, possibly, other fine, durable fibres. Finally, for many types of aerosol (e.g., some metal compounds), particles may be cleared from the lung by dissolution. Here, in addition to the (essentially) mechanical clearance processes already referred to, material can be eliminated by chemical clearance. This means that material may readily enter the blood and so may be dispersed much more widely around the body than if particles were retained within the lung or eliminated via the gastrointestinal tract. For example, lead is found in the blood following exposure to airborne lead. Similarly, cadmium may be found in the kidneys.

7.3 E X P E R I M E N T A L M E T H O D S

The use of animals in inhalation research

In order to investigate what happens to particles in the long term after inhalation and deposition, experiments with humans are usually not acceptable. This relates to the difference between 'real' aerosol particles and idealised, safe particles like those used in the human lung deposition research described in Chapter 6. For 'real' particles, some of their properties which may affect behaviour after deposition, in particular composition or extreme aspect ratio, are the very ones which are directly implicated in adverse health effects. Here, therefore, inhalation experiments with animals have been the favoured approach for obtaining the desired information concerning the fate of particles after inhalation. This, however, raises several issues. Apart from broad question of ethics and conscience, there are significant scientific questions about the general applicability of results from animal inhalation research to humans. Although it is likely that primates would provide the animal model most closely matching human responses in specific exposure instances, the initial cost and the longevity of the animals and the associated time scale of the biological responses resulting from exposure ~ are prohibitive. Beagle dogs have been extensively used in some laboratories, but even then some of the same limiting factors apply. It is inevitable, therefore, that most of the reported inhalation research with 'real' aerosols has been conducted using small mammals. Rats, hamsters and 170

The fate of inhaled particles mice have been the most popular. The rat with a typical lifespan of about 3 years ~ has featured most frequently, and the identification and use of a relatively small number of strains (e.g., 'Wistar' or 'Fischer', 'inbred' or 'outbred', usually 'pathogen-free') have enabled a degree of consistency to be assumed between results from different laboratories. Mauderly (1994) has recently reviewed the relevance and role of the rat in assessing human risk arising from chronic exposure to solid airborne particles. He concentrated on the area of coal dust exposure, the one where perhaps the greatest body of animal and corresponding human research has previously been conducted. Even there, however, he identified the need for extensive further research, involving yet more animal inhalation studies, further epidemiology and further investigation of basic mechanisms. It is reasonable to expect that similar conclusions would be reached for almost every other type of aerosol exposure imaginable. So, once again, sharp attention is drawn to the difficulty in making general extrapolations from rats to humans. Meanwhile, experience has shown that, by the use of the (largely empirical) scaling relationships that have been derived from long experience, inhalation research with rats can provide useful empirical information about toxicological effects in humans. Nonetheless, since all the experimental data presented in this chapter have been derived from rat inhalation studies, some caution should be exercised in discussing how closely the individual observed effects and postulated mechanisms might actually relate to humans and so to practical industrial hygiene. Finally on the general question of animal inhalation research, there undoubtedly remain reservations in some quarters about the use of laboratory animals in health-related research in general and inhalation toxicology. However, experiments like those described below are performed in respected laboratories under carefully regulated protocols and procedures.

Experimental approaches Animal inhalation experiments require that the rats are exposed to the desired test aerosol in some appropriate way ensuring a well-defined exposure regimen (i.e., aerosol type, concentration and duration). There are two basic approaches. One is the whole-body approach, where the rats are placed together in small cages ~ usually in groups of 12 ~ which are then placed inside rectangular chambers where the desired test atmospheres are generated. Those at the Institute of Occupational Medicine in Edinburgh (see Figure 7.5) can hold up to four such cages, and hence up to 48 rats. The other approach is nose-only, where the aerosol is generated in a smaller volume and the animals are placed outside in cylindrical confinement tubes and constrained in position so that their noses project into the test atmosphere of interest. A typical such system is shown in Figure 7.6. A typical exposure 171

Aerosol science for industrial hygienists

Figure 7.5. Pictureof a typical whole-body exposure facility of the type used for inhalation studies involving rats (photograph courtesy of Alan D. Jones, Institute of Occupational Medicine, Edinburgh, Scotland, U.K.). regimen for ongoing exposure is 7 hours per day, for 5 days per week, for 40 weeks per year. Aerosol exposure levels are usually given in terms of mg m -3 of respirable particles. In experiments to investigate the fate of the inhaled material, the exposed animals are usually killed at intervals during the exposure and post-exposure periods so that their lungs and contents can be analysed and recorded as functions of the exposure and post-exposure times. Analysis might include for example: preparation of sections of lung (or other tissue) for subsequent examination under the optical or electron microscope (i.e., by histology) to enable the determination of (a) pathological changes, (b) the locations of particle deposition or subsequent retention, and (c) changes in particle properties after residence in the lung; and

172

The fate of inhaled particles

Figure 7.6. Picture of a nose-only exposure facility of the type used for inhalation studies involving rats (photograph courtesy of Rudolph Jaeger, CH Technologies (USA) Inc. and Steve Burns, Westwood, NJ). quantitative determination of amounts (mass or particle number) of particulate material in tissue (e.g., by infrared spectrophotometry, atomic absorption, X-ray diffraction, scanning or transmission electron microscopy, etc). There have been essentially two types of experiment. The first may be referred to as the 'clearance' type (sometimes referred to as 'sub-chronic'), in which groups of animals are exposed for relatively short periods ranging from a few minutes up to a few days, in order to build up an initial burden of particulate material in the lung. Sub-groups of animals are then removed and killed serially at subsequent intervals so that the clearance of the deposited material can be assessed as a function of the time post-exposure. Such assessment usually involves removing the lungs and other relevant tissue post mortem and carrying out quantitative analysis of their contents. In the assessment process, either the full lung contents or some 'marked' fraction (e.g., radioactively or fluorimetrically-labelled) may be analysed. The second type of experiment is the 'build-up' type, where groups of animals are exposed for prolonged periods, and sub-groups are removed

173

Aerosol science for industrial hygienists at intervals during the overall exposure period so that the lung and tissue contents can be assessed as a function of time for (effectively) ongoing exposure.

7.4 S T U D I E S O F C L E A R A N C E A N D B U I L D - U P Many experiments have been carried out over the years in several countries to investigate the fate of particles in the deep, alveolar region of the lung after deposition. This has been achieved in practice by removing the rats from exposure a n u m b e r of days prior to killing, thus ensuring that all the particles that remained were those that had been deposited originally in the alveolar region. All those particles that had originally deposited in the faster clearing airway regions would by then have been eliminated. The results which have a p p e a r e d in the literature are too n u m e r o u s and diverse to describe fully in this wide-ranging text. However, because there have been considerable discussions and varying interpretations about the results of such studies, it is considered useful here to depart from the practice adopted in Chapter 6 and to present some actual experimental data. That way, readers may form their own individual opinion about the nature of the trends which are exhibited. For this purpose, therefore, a representative selection of data is presented which embodies some of the more important aspects of the behaviour of aerosol particles in the lung after inhalation and deposition. Attention is focused on particles which are

Quartz 1.0

"~ ~ -9

9

10.0 mg

Ii

~e 5-4

L

'*~e0.'~.~e

~._

I x|- .=~g

= 0.5

- ~ _

"6 u~

1.9 ~e 0.2

0

I

!oo

I

I

200 300 Days post-exposure (AT)

I

400

Figure 7.7. Results from clearance studies with rats inhaling quartz, showing the mass fraction retained as a function of time post-exposure (AT in days) for various amounts of aerosol deposited initially in the deep lung (masses in mg as indicated on the plots shown). The figure is based on results from Klosterk6tter and Bunemann (1961) (as summarised by Vincent, J.H., Annals of Occupational Hygiene, Copyright 1990, reproduced by permission of the British Occupational Hygiene Society).

174

The fate o f inhaled particles considered to be insoluble and so which are mainly confined to the lung itself or in lung-associated tissue. Some results from clearance-type inhalation studies for insoluble materials e m b o d y i n g properties typical of m a n y w o r k p l a c e aerosols are shown in Figures 7 . 7 - 7.9. T h e y are for SiO a (crystalline quartz) (from d a t a r e p o r t e d by K l o s t e r k 6 t t e r and B u n e m a n n , 1961), amosite asbestos (from d a t a r e p o r t e d by Bolton et al., 1983; Vincent et al., 1985) and diesel particulate (from d a t a r e p o r t e d by C h a n et al., 1984). In addition, Figure 7.10 shows some results for

K mt~e;t~

1.0

O o,-1

= 0.5

O .,,.q r

[x.

0

100

I

200

I

300

400

Days post-exposure (AT)

Figure 7.8. Results from clearance studies with rats inhaling amosite asbestos fibres, showing the mass fraction retained as a function of time post-exposure (AT in days) for various amounts of aerosol deposited initially in the deep lung (masses in mg as indicated on the plots shown). The figure is based on results from Bolton et al. (1983) and Vincent et al. (1985) (as summarised by Vincent, J.H., Annals of Occupational Hygiene, Copyright 1990, reproduced by permission of the British Occupational Hygiene Society).

Diesel exhaust 1.0 .~.. t~

11.8

particulate ---o-~.-_

- """""-.-.~6.5

mg

"~ 0.5 t__

0

I

I00

0

I

200

Days post-exposure

I

300 (AT)

Figure 7.9. Results from clearance studies with rats inhaling particulate, showing the mass fraction retained as a function exposure (AT in days) for various amounts of aerosol deposited deep lung (masses in mg as indicated on the plots shown). The on results from Chan et al. (1984).

175

I

400

diesel exhaust of time postinitially in the figure is based

A e r o s o l science f o r industrial hygienists

1.0

"~ o

I~e~~ e~e

Toner

14.0 mg

2.0 mg

o 0.5 ~" .o

0.3 mg

o

0

I

I

1O0

200

I

300

Days post-exposure (AT) F i g u r e 7.10.

Results from clearance studies with rats inhaling photoreproductive toner particulate, showing the mass fraction retained as a function of time post-exposure (AT in days) for various amounts of aerosol deposited initially in the deep lung (masses in mg as indicated on the plots shown). The figure is based on results from Muhle et al. (1988).

1000

Quartz

x/x,,Xxr

'7 100 -3 tN

/ I

I

I

10

100

500

Days exposure (T) F i g u r e 7.11.

Mass accumulation of quartz in the lungs of rats as a function of exposure time (T) for post-exposure time (AT) equal to 3 days, results presented in the form of the ratio (Z) of the lung burden to the respirable aerosol exposure concentration (in units of/xg in the lung/mg m-3). The eye-drawn line indicates the main trend. The figure is based on results from Vincent et al. (1987), Jones et al. (1988a) and Muhle et al. (1988) (as summarised by Vincent, J.H., Annals of Occupational Hygiene, Copyright 1990, reproduced by permission of the British Occupational Hygiene Society). V O 9

A x

0.1 mg m -3 1 10 30 1

(Vincent et al., 1987; Jones et al., 1988a)

(Muhle et al., 1988)

176

The fate o f inhaled particles

toner, a particulate material somewhat similar to polystyrene latex, c o m m o n l y used in the photo-reproductive process (from data reported by Muhle et al., 1988). For all these results, it is important to note the different lung burdens at the start of the post-exposure periods. Figures 7.11 - 7.15 show some corresponding results from build-up studies for crystalline quartz, amosite asbestos, titanium dioxide, diesel particulate and toner (from data reported by W a g n e r et al., 1974; Davis et al., 1978; Vostal et al., 1982; Bolton et al., 1983; Vincent et al., 1985, 1987; Wolff et al., 1987; Jones et al., 1988a,b; Muhle et al., 1988; Strom et al., 1988). The experimental data portrayed in Figures 7 . 7 - 7.15 were obtained by research teams at several different laboratories using different experimental techniques and a range of inhaled particulate materials relevant to occupational health. So it is encouraging to be able to identify a n u m b e r of key c o m m o n features. In the results from the clearance studies (Figures 7 . 7 - 7.10), lung burden is seen to fall post-exposure for all the aerosol types examined. The

/

Amosite

I000

oV~

.~ 100 ~1.

~

,,l /e

9 9

Days exposure (T)

Figure 7.12. Mass accumulation of amosite asbestos in the lungs of rats as a

function of exposure time (T) for post-exposure time (AT) equal to 3 days, results presented in the form of the ratio (Z) of the lung burden to the respirable aerosol exposure concentration (in units of ~g/mg m-3). The eye-drawn line indicates the main trend. The figure is based on results from Wagner et al. (1974), Davis et al. (1978), Bolton et al. (1983), Vincent et al. (1985) and Jones et al. (1988b) (as summarised by Vincent, J.H., Annals of Occupational Hygiene, Copyright, 1990, reproduced by permission of the British Occupational Hygiene Society). V

0.1 mg m -3

0 A II

1 5 10 30

V

90

9

177

Aerosol science for industrial hygienists Titanium d i o x i d e

/

v 0 , ~ ~176

1000

e~

E

100

e~

b4

10

I

I

I

10

100

I 500

D a y s e x p o s u r e (7")

Figure 7.13. Mass accumulation of titanium dioxide in the lungs of rats as a function of exposure time (T) for post-exposure time (AT) equal to 3 days, results presented in the form of the ratio (Z) of the lung burden to the respirable aerosol exposure concentration (in units of ~g/mg m-3). The eye-drawn line indicates the main trend. The figure is based on results from Vincent et al. (1987) and Jones et al. (1988a) (as summarised by Vincent, J.H., Annals of Occupational Hygiene, Copyright 1990, reproduced by permission of the British Occupational Hygiene Society). O

1 mgm -3

9 V

10 15

/~

30

I? I,

50 90

rate of fall is quite rapid during the first few days, but then progresses m o r e slowly at longer intervals. H o w e v e r , for high initial lung burdens above a certain threshold, overall clearance is m a r k e d l y slower. That threshold appears to be of the order of 0 . 4 - 2 mg of particulate material per rat irregardless of aerosol type. For the purpose of relating such data to other species (including possibly h u m a n s ) , this range of lung b u r d e n might usefully be expressed as 0.2 - 1 mg of particulate material per g r a m of rat lung tissue (where the typical rat lung weighs about 2 g). The observed reduction in the rate of clearance associated with high lung b u r d e n has been referred to as the 'overload effect' (Bolton et al., 1983). It is now considered to be a very significant p h e n o m e n o n in relation to adverse effects, and so will be discussed m o r e fully below. A l t h o u g h some broad primary trends have been identified, it is nonetheless i m p o r t a n t to note that other dependencies (e.g., on particle size and shape) may also be involved. For example, as already mentioned, other experiments for fibrous aerosols have shown that there is

178

The fate o f inhaled particles Diesel e x h a u s t

particulate

. /

•215 / •

1ooo

•

/

• E

IOO-

.1 N

10

1

/

*'v

A/

A

4-

~'v^ " + +

V

A

-~

+

: I

....

1

10 100 Days exposure (T)

I

500

Figure 7.14. Mass accumulation of diesel exhaust particulate in the lungs of rats as a function of exposure time (T) for post-exposure time (AT) equal to a few days, results presented in the form of the ratio (Z) of the lung burden to the respirable aerosol exposure concentration (in units of txg/mg m -3) (based on results from Vostal et al., 1982; Wolff et al., 1987; Strom et al., 1987). The eye-drawn line indicates the main trend. Although the trends are broadly similar for the three studies cited, note the differences in the magnitude of Z, thought to be associated with differences in the aerosol generated. O

0.25 mg m -3 (Vostal et al., 1982)

9

0.75

A I,

1.5 6

x

6

(Strom et al., 1988)

V

0.35

(Wolff et al., 1987)

+

3.5

A

7

a m a r k e d preference for longer fibres to be retained for longer than s h o r t e r fibres ( M o r g a n et al., 1978). For the build-up studies (Figure 7.11 - 7.15), the results are plotted in terms of the 'standardised' lung b u r d e n , Z, the ratio of actual lung mass b u r d e n to the respirable exposure concentration. H e r e it should be n o t e d that the results are displayed on logarithmic axes, a necessary r e q u i r e m e n t for portraying data which cover such wide ranges but an a p p r o a c h which can suppress differences at larger lung b u r d e n values. T h e lines d r a w n on the graphs are eye fits to the data, i n t e n d e d to draw attention to the main trends. The results indicate that lung b u r d e n increases progressively with exposure time and that the increase appears to be a p p r o x i m a t e l y linear. In addition, for m a n y of the experiments r e p o r t e d and for considerable portions

179

Aerosol science for industrial hygienists Toner I000

//"

p/

E

I00 ::1. N

I I 1o 100 D a y s e x p o s u r e (T)

I 500

Figure 7.15. Mass accumulation of photoreproductive toner particulate in the lungs of rats as a function of exposure time (T) for post-exposure time (AT) equal to a few days, results presented in the form of the ratio (Z) of the lung burden to the respirable aerosol exposure concentration (in units of/xg/mg m -3) (based on results from Muhle et al. 1988). The eye-drawn lines indicate the main trend. @

1 mg m-3

9

4

A

16

of the exposure periods studied, lung burdens appear to scale directly in proportion to the exposure concentration. That is, the data tend to collapse into a single relationship when plotted in the form of Z versus exposure time. One exception is in the results for diesel particulate (see Figure 7.14), for which the clear differences between the data sets may well be associated with differences in the test aerosol and how it was generated (e.g., type of diesel engine used). The other exception is for the toner particulate (see Figure 7.15), for which there is no obvious explanation. The general tendency exhibited by most of the build-up data is somewhat surprising since one consequence of the change in clearance rate between low and high lung burden (as reflected in the results of the clearance studies) might be expected to manifest itself as a more rapid increase in lung burden after the overload threshold has been reached. Does this represent an inconsistency? It is possible that different kinetics could apply for chronic exposure than for post-exposure, since ~ for the latter ~ the cessation of the arrival of new particles is likely to change the pattern of the recruitment of macrophages, the main agent of clearance. However, the overall body of data, acquired from experiments carried out in several different laboratories using different assay techniques and exposure methods, do not permit such a firm generalisation. Therefore, although it can be instructive to try to identify and rationalise some broad trends, the temptation to oversimplify is cautioned against.

180

The fate of inhaled particles

In workplaces, unlike for the rats in the laboratory experiments summarised here, people are rarely exposed to a single type of aerosol. Mixtures are usually the norm. For example, exposure to pure quartz is very uncommon, whereas exposure to rock dust (e.g., in mining and quarrying) frequently results in exposure to quartz in the presence of less-toxic other minerals (i.e., often referred to as 'nuisance' dust), with the quartz usually in low proportions (typically less than 10%). The same is also true for most asbestos dust exposures. A number of additional rat inhalation studies with mixed aerosols have been carried out and reported, for which the results are complex and, again, often contradictory. One set of results, for example, suggests that the build up in the rat lung of quartz and asbestos in the presence of large amounts of non-toxic 'nuisance' dusts occurs in a manner which appears to be independent of the presence of the nuisance dust (McMillan et al., 1986). This interesting, albeit rather surprising, result has not yet been confirmed elsewhere. So further work is needed on the important subject of mixed dust exposures. Meanwhile, however, results for the various aerosol types inhaled singly (as was the case for all the results in Figures 7 . 7 - 7.15) can continue to provide valuable insight into the fate of aerosols encountered by workers in the occupational environment. The results from inhalation experiments involving amosite asbestos have already been described. One other form of asbestos is particularly relevant to human exposure in view of its widespread usage, namely, chrysotile. The fibres of this material are more complex than for amosite, being curly in shape and prone to dissolution and leaching, and to being broken down into very fine fibril elements when immersed in surfactant substances (including lung fluids). It is believed that such properties may have a significant bearing on the fate of such particles in the lung. Inhalation studies of the 'build-up' type have been reported for this material (Wagner et al. 1974; Jones et al. 1989, 1994). They show that the mass of fibre increases during the early part of the chronic exposure regime. But, unlike for the other substances described above, it does not go on increasing. Rather, it tends to level off. This confirms that there is an additional removal mechanism for mass, which was not present for the other substances described above. This is consistent with results from histological studies of lung sections taken p o s t m o r t e m from people who had been exposed to chrysotile asbestos and which revealed the presence of very few fibres. In the Jones et al. studies, electron microscope determinations of the size distributions (length and diameter) of the fibres in the lung showed in addition that the proportions of long and very thin fibres increased as the exposure time progressed. These results are consistent with the further hypothesis that some chrysotile fibres break down into their fibril constituents during their time in contact with lung fluids. As stated earlier, most of the particulate material mentioned so far in this chapter is relatively insoluble. But, of course, there are many aerosols of soluble material (e.g., some species of metals such as lead and cadmium)

181

Aerosol science for industrial hygienists w h e r e chemical dissolution plays a p r e d o m i n a n t role in the fate of the i n h a l e d material. F o r such substances, the lung itself is n o t a significant e v e n t u a l target as far as risk is c o n c e r n e d . In turn, t h e r e f o r e , l o n g - t e r m a c c u m u l a t i o n in the lung itself is of relatively little interest.

7.5 E X P E R I M E N T A L S T U D I E S O F D U S T A C C U M U L A T I O N LUNG-ASSOCIATED LYMPH NODES

IN

T r a n s f e r of inhaled particulate m a t e r i a l to the l u n g - a s s o c i a t e d l y m p h n o d e s has b e e n e x p e r i m e n t a l l y studied less extensively. This is not surprising in

Quartz

104

o/

10 3 -

10 2

, - ,

10

-

/

-

-,1 0

~

','02 104 --

I /o 10 3

Titanium

I

I

!04

i05

dioxide

.,-,

/

103 -

o:

~

102 -

10-

! 102

oo

10 3

I

10 4

I

10 5

Mass in deep lung (l~g) Figure 7.16. Data for the mass accumulation of quartz and titanium dioxide in the lymph nodes of rats during chronic exposure, shown as a function of the mass of dust present in the alveolar region of the deep lung. The lines are shown to indicate the main trends. The figures are based on results of Vincent et al. (1987) (as summarised by Vincent, J.H., Annals of Occupational Hygiene, Copyright 1990, reproduced by permission of the British Occupational Hygiene Society).

182

The fate of inhaled particles the light of the difficulty of identifying, excising and assaying small-scale lymphatic tissue in excised rat lungs. But the m a t t e r of particulate material in the lymph nodes is widely considered as i m p o r t a n t because of its strong correlation with indices of the toxicity of the inhaled particulate material. Some useful data on lymph node accumulations have b e e n o b t a i n e d , and some typical ones are shown in Figure 7.16 for quartz and titanium dioxide (from Vincent et al., 1987). They are plotted here in terms of lymph n o d e b u r d e n as a function of the particulate b u r d e n of the lung itself. Again, the lines on the graph are eye fits to the data points. T h e results reveal an i m p o r t a n t lung b u r d e n d e p e n d e n c e , indicating that lung b u r d e n is the primary 'driving force' for the t r a n s p o r t of particulate material to lymph nodes. A l t h o u g h there is little accumulation in the lymph nodes at low lung burdens, transfer of particles to lymph nodes takes off quite sharply once lung b u r d e n has r e a c h e d a certain threshold. It is p e r h a p s interesting to note from Figure 7.16 that the o r d e r of m a g n i t u d e of that threshold is close to that for the onset of clearance overload. H e r e , h o w e v e r , there is some evidence that that threshold is s o m e w h a t lower for dusts k n o w n to be toxic (e.g., quartz) than for relatively non-toxic dusts (e.g., titanium dioxide). It is interesting to c o m p a r e this interpretation with the idea that m e a s u r e m e n t of lymph n o d e b u r d e n s in rats might be a useful m e a n s for assessing the toxicity of insoluble dusts (e.g., Le Bouffant et al., 1988 in their studies of coalmine dusts). Example 7.1. A coalminer has worked underground in a dusty atmosphere for 30 years. The respirable coal dust exposure levels in the mine where he worked were as follows for successive 5 year time intervals: 1950-1955 1955-1960 1960-1965 1965-1970 1970-1975 1975-1980

15 mg m -3 13 mg m -3 9 mg m -3 7 mg m -3 6 mg m -3 4 mg m -3

He retired in 1980. From the information given in this chapter, estimate the total dust burden in his lung at retirement, making whatever assumptions are necessary. At what point, if at all, did his lung burden reach the overload threshold? Based on the knowledge we have at this time, this question cannot be answered with great confidence. At best, we can make a 'back-of-the-envelope' calculation based on a number of assumptions 1.

The information about clearance of insoluble dusts in rats can be applied directly

2.

In the absence of clearance or build-up data for coal dust, use the results for titanium dioxide

183

Aerosol science for industrial hygienists 3.

The coalminer worked for 5 days/week, 48 weeks/year

4.

The inspiration rate for the miner is 20 1 min -1, compared to 0.1 1 min -1 for rats

5.

The human lung weighs about 1 kg, compared to about 2 g for the rats

Number of days worked in each 5 year period = 5 • 48 • 5 = 1200 days From Figure 7.13, the slope of the graph gives the rate of accumulation of dust in the rat lung. We find by inspection that this is 2 (Ixg/mg m -3) Rate for rat day of exposure Assume that the corresponding rate for the worker will increase in proportion to the amount of air breathed. So 20

)

(400 Ixg/mg m -3)

Rate for miner ~ 2 x O. 1

day of exposure.

From this, we can estimate the cumulative lung burden for the first 5 year period (1950-1955), as follows:15 [mg m -3] x 1200[days] x 400[(ixg/mg m -3) days -1] Thus the total amount of dust retained in each time interval is

*

1950-1955 1955-1960 1960-1965 1965-1970 1970-1975 1975-1980

7200mg 6240mg 4320mg 3360mg 2880mg 1920mg

Total

25,920 mg

Total lung burden at retirement = 25.9 g

To examine the question of overload, re-tabulate the above in terms of the cumulative mass of dust per unit mass of lung tissue

184

The fate of inhaled particles 7.20 mg g-1 13.44 mg g- 1 17.76 mg g-I 21.12 mg g- 1 24.00 mg g-1 25.92 mg g-1

Tissue burden after 5 years 10 years 15 years 20 years 25 years 30 years

Recall that overload in rats occurred for lung tissue burdens exceeding about 1 mg g-1. So it is clear from the above calculation that the coalminer's lung will become overloaded very early in his working life, probably within the first year of exposure Note that, although this calculation was been no more than a rough approximation, the magnitude of the final lung burden estimate is consistent with measurements of the contents of actual miners' lungs taken at autopsy (Ruckley et al., 1981)

7.6 T H E S I G N I F I C A N C E O F ' O V E R L O A D ' As introduced above, the term 'overload' was first coined in the paper by Bolton et al. (1983) to describe the i m p a i r m e n t of clearance associated with accumulation of heavy burdens of insoluble particles in the lung. H o w e v e r , as was pointed out at the time, the p h e n o m e n o n in question had been noted even earlier by Ferin (1972). Since then the p h e n o m e n o n has been e x a m i n e d in greater detail, in particular in a series of papers by Paul M o r r o w and various co-workers in which consideration was confined to those particles thought to be relatively inert (thus excluding demonstrably harmful dusts such as quartz and asbestos). This work is summarised in M o r r o w (1994) where the point is made that, even for such relatively-inert particles, the occurrence of overload ~ and hence build up of excessive dust burdens in the lung ~ can result in significant pathological changes (including tumours). Morrow, in his 1988 version of the overload hypothesis, stated that impaired clearance would occur if lung burden in the rat (he specified the F344 strain) exceeds 1 mg of particulate material per gram of lung tissue. This is generally consistent with the estimate given above. Yu et al. (1989) inspected the data from a n u m b e r of rat inhalation studies like those described, and showed a strong correlation b e t w e e n the effective rate of clearance (expressed in terms of the fraction of lung b u r d e n cleared each day) and lung burden. As summarised in Figure 7.17, the data suggest that the overload p h e n o m e n o n is i n d e p e n d e n t of the type of particle and so is generic in nature (at least for relatively non-toxic aerosol). This idea is extended in Table 7.1 which summarises the extent to which, based on the

185

Aerosol science for industrial hygienists 0.2

O u

g.,

ej

0

1

2 3 4 Mass in the deep lung (mg)

F i g u r e 7.17. S u m m a r y of c l e a r a n c e d a t a f r o m i n h a l a t i o n e x p e r i m e n t s with rats for v a r i o u s t y p e s o f i n s o l u b l e a e r o s o l , s h o w i n g t h e e s t i m a t e d r a t e o f c l e a r a n c e as a f u n c t i o n of lung b u r d e n ( b a s e d on d a t a c o l l e c t e d t o g e t h e r a n d p r e s e n t e d by Y u et al., 1993).

T a b l e 7.1. S u m m a r y o f k e y g e n e r i c r e s p o n s e s o b s e r v e d in rat i n h a l a t i o n s t u d i e s with i n s o l u b l e p a r t i c u l a t e m a t e r i a l ( b a s e d on M o r r o w , 1994). Slower clearance

Higher retention

Chronic inflammation

Fibrosis

Tumours

Titanium dioxide

*a

9b

,.

+c

+c

Volcanic ash

+c

+c

+c

+c

9b

Fly ash

+~

+.

Petroleum coke

*a

.~,

+c

+c

+c

Diesel particulate

+c

+c

+c

+c

+c

Material

.

.

.

.

.

.

.

.

.

PVC

+c

-[-c

-[-c

')b

__d

Toner

+c

+c

+c

+c

_d

Carbon black

+~

.~

+c

..

+~

aNot reported, bUnsure, cPositive. ONegative.

available experimental evidence, a range of primary overload-related effects may be generalised. Morrow (1994) has further suggested that the actual mechanism of overload involves the inhibition of macrophage mobility by the 'volumetric' presence of particulate material in the lung. He proposed the following empirical expression k - 0.0210 - 0.0052

In

Vpart

(7.1)

where the clearance rate, k, is expressed in terms of the proportion of particles removed per day, and the particulate volume, V p a r t , is expressed

186

The fate of inhaled particles in nanolitres (nl). For the latter, we see that, for particles of density 103 kg m -3, a 1 mg lung burden converts to 1000 nl.

7.7 KINETICS OF C L E A R A N C E Data like those described above have formed the basis of pharmacokinetic models for describing the fate of particulate material in the lung. The simplest such models are those based on compartmental simulations of the lung. The flows of particles in and out of the various 'black box' compartments are described mathematically by first-order linear differential equations. Such 'box' models are largely empirical, but purport to relate to plausible biological mechanisms and contain plausible coefficients. Other more complicated models refer more explicitly to the physiological and biological processes taking place in the lung. It is instructive to begin with the simplest such model which formed the basis of much of the earlier thinking on the kinetics of deposition and clearance of insoluble particles. As shown schematically in Figure 7.18, the lung is represented in this model in terms of three compartments, broadly described as fast-clearing, medium-clearing and slow-clearing. These compartments are said to obey 'linear kinetics', meaning that the instantaneous rate at which particulate material is eliminated from each is directly proportional to the mass present in that compartment. Although this model is largely notional, primarily aimed at providing a mathematical description

Inhaled aerosol ,,,

TB region (fast)

Slow

Medium

.,

Alveolar region

Figure 7.18.

Simplest kinetic model for the clearance of insoluble particles from the lung.

187

Aerosol science for industrial hygienists of the main observed trends for lung burden as a function of time during exposure and post-exposure (the period following the cessation of exposure), it is useful that each compartment can be related to widely-accepted biological ideas. For instance" the fast-clearing compartment is linked to the ciliary clearance process in the tracheobronchial region; the medium-clearing compartment is linked to the first-phase macrophage clearance action in the alveolar region; and the slow-clearing compartment is linked to the second-phase (or inhibited) macrophage action for 'fixed' or 'embedded' particles in the alveolar region. Inspection of the available experimental data have suggested that these three compartments have clearance time constants (for rats) which are typically of the order of 0.5 days, 10 days and 100-200 days, respectively. In order to formulate a mathematical picture of what happens in this simplest kinetic model, consider just the alveolar region. This is relevant to a range of possible disease outcomes, including, for example, pneumoconiosis and emphysema. Here we drop the fast-clearing compartment and concentrate on the two deep lung compartments, medium- and slow-clearing, into which particles are deposited at the mass rates e~m and % (in mass per day) and removed at the rates k m and ks (in fraction per day, the direct inverse of the time constant of clearance already referred to). If the instantaneous mass of particulate material contained in each of these two compartments is M m and M s, respectively, then the changes in mass in a small time interval dt are dM m(t) = otmdt - { M m(t) k m} dt (7.2) dMs(t ) = ets d t - { Ms(t ) ks) dt which may be integrated to give the total deep lung burden after an exposure time T followed by a post-exposure time AT. Thus Mtotal(T, AT) = /

Otm

(

km

+{

[1 - e x p ( - T km) ] } e x p ( - A T km) (7.3) [1 - e x p ( - T ks)]} e x p ( - A T ks)

ks

188

The fate of inhaled particles In the first instance, when this simple linear model is applied to the situation following the cessation of exposure, it predicts that the total deep lung burden will fall to zero when post-exposure time AT becomes large (see Figure 7.19a). The model does not contain anything which relates to the overload phenomenon. Next, when the model is applied to predicting what happens to lung burden during a prolonged, constant-level chronic exposure regimen, it suggests that lung burden should eventually level off. This reflects the equilibrium which is eventually achieved between the mass rates at which particles are deposited into the various compartments and the corresponding rates at which particles are eliminated (see Figure 7.19b). This is consistent neither with the overload hypothesis nor with the data from 'build-up' studies (in rats), both of which point to a continued increase in deep lung burden, without levelling off (see especially Figures 7.11 - 7.15). If, as appears reasonable, the lung burden may be taken as rising more or less linearly during uniform chronic exposure (as is seen for a substantial part

(a) Clearance

Time post-exposure (AT)

,-,.s

r

(b) Build-up

Time of exposure (T) Figure 7.19.

Graphs showing the predicted trends for clearance and build-up of insoluble inhaled particles resulting from the simplest kinetic model shown in Figure 7.18. Note in (a) that the lung burden post-exposure falls eventually to zero; also in (b) that, during ongoing exposure to the same respirable aerosol concentration, the lung burden eventually levels off.

189

Aerosol science for industrial hygienists

Inhaled aerosol

Tracheobronehial region (fast)

L

I-

Lymph nodes

Alveolar region

I

!

Medium / slow compartments

Sequestration compartment Figure 7.20. Revised kinetic model for the transfer within and clearance from the lung of deposited insoluble particles, including both the sequestration and lymph node compartments.

of the exposure duration for many of the rat experiments described), then this may be simulated mathematically by invoking an additional compartment from which material clears either very slowly or not at all (Soderholm, 1981). Such thinking has led to a modified version of the above model, and this has become known as the 'sequestration model'. It is shown schematically in Figure 7.20. Again, it is instructive to show how a simple version of this can lead to a set of equations by which lung burden can be calculated (and hence compared with measured experimental data). Here, the working equations are (Vincent et al., 1985, 1987) dMm(t)

--

dMs(t)

= c~sdt- ( k s M s ( t ) } d t - {kosMs(t)}dt

dMo(t)

otm dt - { kmM m ( t) } dt - { komMm ( t) } dt (7.4)

{komM m(t) } dt + ( kosMs (t) }dt

where M o refers to the mass of particles in the new sequestration compartment, and kom and kos are the rates at which particles are transferred from the two clearing compartments into ~he non-clearing sequestration compartment. Integration of this equation is only marginally more complicated 190

The fate of inhaled particles than for the simplest model described above, leading to the final expression for the total lung mass

Mtotal(T, AT) -

{Om

--- [1 - e x p ( - T km*)]

}

e x p ( - A T km* )

km*

[1 - e x p ( - T ks*)] } e x p ( - A T ks*)

(7.5)

ks* T+AT

+ko{ f [Mm(t + Ms(t)]dt } in which km* - km+k o and ks* = k~+ k o

(7.6)

and where kom ~ ko~ ~ k o. A typical calculated curve is shown in Figure 7.21, and is seen to be qualitatively consistent with data like those shown in Figures 7.11 - 7.15. By fitting the experimental data to the model (e.g., by least-squares regression), values for the o~'s and k's may be found which give good quantitative agreement between theory and experiment (see Table 7.2).

Linear region

I:::

o,.q t~

z~

f

Non-linear region at 'start-up'

Figure 7.21.

Graph showing the build-up trend for the lung burden of insoluble inhaled particulate as predicted by the revised kinetic model shown in Figure 7.20. Note that the rate of build up eventually becomes linear so that the lung burden continues to increase steadily.

191

Aerosol science for industrial hygienists Table 7.2. Summary of coefficients obtained by fitting the sequestration model in Figure 7.23 to the data contained in Figures 7.11, 7.13 and 7.16 (Vincent et al., 1987). Dust type

r

r

Titanium dioxide Quartz

3.8 3.8

3.8 3.4

kmb

ksb

kob

Mthreshc

kLb

0.0625 0.0714

0.0050 0.0091

0.0167 0.0050

1800 900

0.0016 0.0016

al~g of dust/mg m -3 of air/day, bDays --l. clxg of dust.

The model needs to be extended further to include transport to lungassociated lymph nodes. Here, based on the experimental observations (see Figure 7.16), the rate of mass accumulation in the lymph nodes is directly proportional to the mass in the deep lung over and above some threshold. This too is shown schematically in Figure 7.20, and leads to an expression for the mass M L accumulated in the lymph nodes

T+AT ML(T, AT ) - k L { f [Mtotal(t) - Mtotal,thresh]dt }

(7.7)

0

where k L is the rate of transfer to lymph nodes (fraction per day) which is equal to zero for Mtota i < Mtotai,thresh , the latter being the threshold lung burden above which transport to lymph nodes takes off. In addition, a further adjustment to the sequestration model is required, and this is achieved by setting k i n * -- k m +

ko

+ kL

and k~* = k s + k o

+ kL

(7.8)

For the complete sequestration model shown in Figure 7.20, a full set of coefficients obtained by fitting the model to the data for the rat inhalation experiments for titanium dioxide and quartz shown in Figures 7.11, 7.13 and 7.16 is summarised in Table 7.2. These coefficients are quite plausible based on what is known about the biological clearance and redistribution mechanisms implied in the model. One consequence of the sequestration model is worthy of particular note in relation to insoluble mineral dusts. Consider a situation where there is a period of exposure followed by a period of no exposure (i.e., post-exposure). The sequestration model shows that, the longer the pre-exposure period, the greater the proportion of the lung burden which becomes locked up and so is unavailable for clearance (i.e., is sequestrated). This means that the overall rate of clearance, as reflected in the rate of fall of the lung burden post-exposure, will appear to decrease for increasing prior exposure

192

The fate of inhaled particles durations. This effect is illustrated in Figure 7.22 where clearance has been calculated for a range of values for the exposure period (T) prior to the post-exposure phase. For example, for T = 256 days, over 80% of the dust in the lung at the end of the exposure period (i.e., at AT = 0) will in theory never be cleared. It is important to note, however, that this does not imply zero clearance for ongoing (or chronic) exposure. For any inhaled dust deposited in the deep lung, the majority is still cleared. For example, in the earlier E x a m p l e 7.1, whereas the estimated rate of sequestration in the lung of a h u m a n worker was found to be 400 p~g/mg m -3 per day, the mass of respirable dust actually inhaled was about 10 mg/mg m -3 per day (for an inhalation rate of 20 1 min -1 over a period of 8 hours). That is, only about 4% of the inhaled respirable dust was sequestrated. In addition, however, on retirement the vast majority of the dust present in his lungs will be that

1.0 256 128

O

.=. 0.5

32

.o O s,d

T=2

u.,

0

.,

I

100

,

I

200

I

300

Days post exposure [AT] 1.0

~r

0.5

I

0

I

100

200

300

Days post exposure [AT]

Figure 7.22. Calculated clearance and build-up curves for insoluble inhaled dusts from the sequestration model shown in Figure 7.20. This shows how the fraction retained post-exposure levels off at a higher level the longer exposure time. This is because a larger proportion of the deposited material has already been sequestrated by the time exposure ceases.

193

Aerosol science for industrial hygienists

which ~ over the years ~ has become sequestrated. So, according to the sequestration hypothesis, there will be virtually no further elimination of dust from his lungs after retirement. It is noted that the sequestration model in Figure 7.20 does not embody any direct, lung burden-dependent simulation of the process which was referred to earlier as 'clearance overload'. However, it should be stressed that the linear mathematical model outlined above is only one of a number which have been described in the literature in recent years. In fact it is one of the simplest and most empirical. In another relatively simple approach taken in the non-linear model of Yu et al. (1988), the rate of clearance of insoluble particles from the alveolar region is assumed to vary non-linearly with the mass in the lung in such a w a y a s to account simultaneously for sequestration and clearance overload. Other models have been proposed elsewhere that are more complicated and more explicitly involve many of the biological processes which take place in the lung. For example, the model of Strom et al. (1988) for insoluble particles embodies the kinetics of free (unphagocytosed) particles on the alveolar walls as well as in mobile macrophages, whilst also explicitly embodying sequestration and transport to lymph nodes. The latter two processes are both assumed to be exponential, sequestration as a function of particulates in mobile macrophages and lymphatic transport as a function of the sequestrated mass. This is in contrast to the simpler sequestration and lymphatic clearance model outlined earlier. The model of Strom et al. therefore appears to have the phenomemon of clearance overload directly built in. More recently, St6ber et al. (1994) have proposed an even more comprehensive physiology-oriented compartmental kinetics (so-called 'POCK') model. It is summarised in Figure 7.23. It involves a particle deposition-activated alveolar macrophage recruitment process which leads to a quasi-steady-state macrophage population on the alveolar epithelial surface. The exposure-dependent distribution of particles contained in the macrophage population is used to determine the particle load in mobile and immobile macrophages, where the conditions for immobility depend on the volume uptake capacities of individual macrophages. The immobilised, particle-filled macrophages are ~ in effect ~ sequestrated in the interstitial spaces of the deep lung. Thus, this model is seen to mechanistically link the two important phenomena of overload and sequestration. Although the particulate material which is 'bound' in this way is no longer available for macrophage-mediated clearance in the 'classical' sense described at the beginning of this chapter, transfer to the lung-associated lymph nodes does take place. St6ber and his colleagues found good agreement between this model and available experimental lung burden data for diesel particulate, carbon black and photo-reproductive toner. But they were careful to avoid testingthe model against data for the much more toxic quartz. For the latter, the properties of quartz and other similarly-cytotoxic particles would need to

194

The fate of inhaled particles Inhaled Tracheobronchial

particles

]

tract

t

.,

~

surface

J

0 g.q c~ 0

<

Lung-associated lymph nodes Figure 7.23. Schematic describing the more complicated physiologically-oriented

compartment kinetics ('POCK') model. It shows how the phenomena of sequestration and overload may both be linked to the mobility of the alveolar macrophages. (From St6ber, W. et al., Alveolar clearance and retention of inhalable insoluble particles in rats simulated by a model inferring macrophage particle load distributions. Journal of Aerosol Science, 25, 975-1002, Copyright 1994, with kind permission from Elsevier Science Ltd, The Boulevard, Langford Lane, Kidlington OX5 1GB, U.K.).

be dealt with specially. They were also careful to point out that, at this stage at least, the model should not be generalised beyond rats. In addition to these models for dust where the particles are reasonably isometric, Yu and his colleagues have extended their own ideas towards describing the fate of durable fibrous dusts (e.g., amosite asbestos), in which the rate of clearance is a decreasing non-linear function of fibre length (Yu and Asgharian, 1990). They have also developed a model specifically for chrysotile asbestos which attempts to account for the changes in particle morphology during retention, in particular the splitting of individual fibres into their constituent fibrils (Yu et al., 1990). The resultant predictions are in fair agreement with the experimental data for chrysotile asbestos published by Jones et al. (1989). In most of what has been described so far, it has been assumed that the aerosol particles themselves are insoluble. This applies to many aerosols of relevance to occupational health. However, there are other aerosols which, to a greater or lesser extent, are soluble when they come into contact with lung fluids. For these, the lung is just the first site of the inhaled particulate material. Dissolution of the particles in the lung fluids leads to the entry of

195

Aerosol science for industrial hygienists

material into the blood and thence throughout the body to other organs, including the brain, kidney, liver, bone, etc. A linear first-order mathematical approach is still appropriate, but inevitably becomes more complicated. By way of illustration, Figure 7.24 illustrates the type of model which has been proposed for describing the fate of inhaled particles containing lead.

PARTICLE INHALATION

LUNG

DISSOLUTION

VESSEL-RICH GROUP GI tract, heart, kidneys, brain, spleen, pancreas, thyroid, thymus, adrenals

MUSCLE GROUP Muscles and skin

FAT GROUP Fatty tissue and marrow

LIVER

METABOLISM Figure 7.24.

Schematic of a kinetic model for describing the fate t h r o u g h o u t the body of inhaled soluble particles.

196

The fate of inhaled particles Most of the kinetic models described here do not in themselves shed m u c h new light on the detailed biological processes which take place in the lung. This is true even for the more sophisticated physiologically-based models (like that of St6ber and his colleagues). H o w e v e r , such models and their comparison with experimental data from rat inhalation studies can provide useful guidance in the design of e x p e r i m e n t s to further investigate those processes. For the present, however, their main justification in the b r o a d e r field of occupational health is in providing a basis for thinking a b o u t dose and, in turn, assessment of the risk associated with aerosol inhalation.

7.8 D O S I M E T R Y It is a natural extension of the preceding ideas to move towards a rationale for risk assessment and i m p r o v e d epidemiology. In relation to radioactive aerosols (in the area which is referred to as 'health physics'), the sort of knowledge contained in the preceding sections has been applied for many years. But it does not yet a p p e a r to have been extensively applied to workplace aerosols in the wider context. In the context of aerosol inhalation, the term 'dose' has been used in a n u m b e r of different ways. Some researchers use it to refer to the rate at which material is deposited in the lung or to the actual a m o u n t delivered, others to the cumulative exposure of a subject over a defined period of time, and others to the cumulative tissue burden. R a a b e (1989) has a t t e m p t e d to clarify matters. H e states To the toxicological purist, (the term) 'dose' refers specifically to the total quantity of chemical, biochemical, or radiative energy or specific structural and/or biochemical alteration transferred to or effected in a known mass of sensitive tissue. The 'dose rate' for this transfer or alteration is essentially proportional to the concentration of toxicant in the tissue . . . (So) the time integral of dose rate yields the 'total dose' delivered to the tissue . . . The broader use of the term 'dose' . . . can lead to confusion and is best avoided in inhalation toxicology. The philosophy expressed by R a a b e is e n d o r s e d here. So we use the t e r m 'dose' in relation to the integrated 'harmfulness' i m p a r t e d to lung tissue over a period of time by the aerosol particles retained in the lung. It may be described as c o m p o s e d of three essential ingredients, as follows: (1) The exposure history, expressed in terms of the exposure function, E,,.

This is the time-weighted average worker exposure concentration of a relevant aerosol fraction during the n th day since exposure began. It can be m e a s u r e d using aerosol sampling i n s t r u m e n t a t i o n like that which will be described in C h a p t e r 9.

197

Aerosol science for industrial hygienists

(2) The time-dependent retention of particles in the lung, expressed in terms of the retention function, R m . This is the proportion of aerosol material remaining at the end of the m th day since it was deposited (where m can take all values up to and including n). It can be derived from kinetic models like those described earlier in this chapter.

(3)

The ability of particles to transmit the potential to damage the lung, espressed in terms of the 'harmfulness' function, G m . For a given aerosol material at the end of the m th day since it has been deposited, this ingredient relates to toxicity, and hence to the chemical and physical nature of the individual aerosol particles. In toxicological terms, G m may be regarded as the rate of transfer (per unit mass of retained material) of some effect to a known mass of sensitive tissue. Allowance is made for the fact that, during the time that an individual particle remains in the lung, it may become more or less harmful. Thus, G m may change with time.

A simple example serves to illustrate the concept of cumulative dose based on these ideas. Consider the simple case of 3 days of exposure followed by 2 days post-exposure. For this we may construct the array

Day

Incremental dose received

1

E1R1G 1

2

E1R2G 2

+

E2R1G 1

3

E1R3G 3

+

E2R2G 2

+

E3R1G 1

4

E1R4G 4

+

E2R3G 3

+

E3R2G 2

5

E1RsG s

+

E2R4G 4

+

E3R3G 3

where the overall cumulative dose is the sum of all the individual contributions contained in this array. In general, for N days altogether, the cumulative dose D ( N ) is given by the simple summation N

N-n+ 1

n=l

m=l

D(N) =

(7.9)

where E,, = 0 on days where there is no exposure (e.g., at weekends, during vacation) and, of course, after exposure has ceased (e.g., for a worker upon retirement). Such an expression might be used as a relevant index of the risk associated with aerosol inhalation. Recent studies have shown that the cumulative dose

198

The fate of inhaled particles calculated in this way relates well with inflammation in the lungs of rats as measured in terms of neutrophil cell populations (obtained for lung samples obtained by bronchoalveolar lavage) in rats which have been inhaling mineral dusts (Vincent and Donaldson, 1990). The results for pure quartz dust, known to be a toxic mineral, are particularly interesting since, as shown in Figure 7.25, they show the neutrophil population continuing to increase steadily even after exposure has ceased. In this figure, the lines are based on calculations using the above dosimetric model under the assumption that the biological response is directly proportional to the integrated dose. The results indicate that the 'harmfulness' of the quartz particles is highly persistent, a finding which, if it can be validated for humans over longer timescales, may have a significant bearing on how we think about risk for workers who have been removed from exposure to dusts containing quartz. This is particularly important when considered in the light of the consequences of the sequestration hypothesis. Similar experiments for other dusts (e.g., titanium dioxide, coal dust) showed their 'harmfulness' to be smaller in magnitude and less persistent. With the exception of radioactive aerosols, models of the form described above have not been widely used in assessing risks associated with aerosol inhalation. A more common index is cumulative exposure, as has been applied in some epidemiological studies. This is given by

200 -

J~ ~'

Quartz

I I

f

J

"

Z 105+PE ...... 50

100

150

Days elapsed

Figure 7.25. Neutrophil population (representing the inflammatory response) for rats as a function of exposure (including post-exposure) to quartz. The solid symbols are for a respirable dust exposure concentration of 10 mg m -3, the open symbols for 50 mg m -3. The curves indicate the fitted dosimetric model given by Equation (7.9), the dashed portions representing the periods following the cessation of exposure. The numbers on each curve indicates the exposure time plus the post-exposure time (AT). (From Vincent, J.H. and Donaldson, K., 1990, British Journal of Industrial Medicine, 47, 302-307, reproduced by permission of the BMJ Publishing Group, London).

199

Aerosol science for industrial hygienists N

C(N) =

] En

(7.10)

tl-- ]

In general, Equations (7.9) and (7.10) give very different results. But it should be noted that they do become identical for aerosols which are either cleared very quickly or become inoccuous very soon after they have been deposited in the lung. This would be true, for example, for a particle consisting of a very short-lived radionuclide or for a non-persistent dust like titanium dioxide. However, the evidence suggests that Equation (7.9) should be preferred to Equation (7.10) for persistent toxic dusts such as those containing quartz. In addition to the points that have already been made, it should be noted that Equation (7.9) takes into account the shape of the exposure history of the subject in question, including such properties as:

Biological studies (e.g., bronchoalveolar cell iavage after inhalation or instillation)

Jv E_ {~/Zlm~__+l ! D(N) = ~n - ' ' = ! m

] m

~

Validation by reference to epidemiology

/ Pharmacokinetic models of deposition, ~ clearance, etc. from ~ results of animal studies

]Exposure history]-

Design of suitable]_ sampling strategy J -

Figure 7.26.

I Results of lung autopsy investigations to validate animal models

Retrospective assesment (where necessary)

]Choice of sampling ]criteria and instruments

Rationale for thinking about future research towards developing a working dosimetric model for dust exposure. 200

The fate of inhaled particles

a general t e m p o r a l decline in exposure over a working lifetime (as has been the case for most workplace c o n t a m i n a n t s ) ; and 'peaky' exposures characterised by relatively short periods of high intensity of exposure interspersed with periods of low intensity. Whilst it is widely believed that such properties should be i m p o r t a n t in relation to dose and risk, they do not feature in E q u a t i o n (7.10). W h a t has been outlined above therefore is a hypothesis or possible f r a m e w o r k for the d e v e l o p m e n t of practical and realistic dosimetric models. Much m o r e w o r k needs to be p e r f o r m e d before such a model can be ultimately realised. Figure 7.26 summarises a rationale for thinking about future directions. It includes: the choice of sampling criteria, i n s t r u m e n t a t i o n and strategies for assessing En; d e v e l o p m e n t of p h a r m a c o k i n e t i c models for the assessment of R m and the validation (and, if necessary, adjustment) of those models with respect to h u m a n s (e.g., by relating t h e m to data on lung contents in h u m a n s with k n o w n exposure histories obtained p o s t m o r t e m ) ; biological studies (e.g., b r o n c h o a l v e o l a r d e t e r m i n e G m for different dusts; and

lavage of rat lungs) to

validation with respect to k n o w n epidemiology.

REFERENCES Bolton, R.E., Vincent, J.H., Jones, A.D., Addison, J. and Beckett, S.T. (1983). An overload hypothesis for pulmonary clearance of UICC amosite fibres inhaled by rats. British Journal of Industrial Medicine, 40, 264-272. Davis, J.M.G., Beckett, S.T., Bolton, R.E., Collings, P. and Middleton, A.P. (1978). Mass and number of fibres in the pathogenesis of asbestos-related lung disease in rats. British Journal of Cancer, 37, 673-688. Chan, T.L, Lee, P.S. and Hering, E.E. (1984). Pulmonary retention of inhaled diesel particles after prolonged exposures to diesel exhaust. Fundamental and Applied Toxicology, 4, 624-631. Ferin, J. (1972)., Observations concerning alveolar dust clearance. Annals of the New York Academy of Sciences, 200, 66-72. Gehr, P., Schfirch, S., Geiser, M. and Im Hof, V. (1990). Retention and clearance mechanisms of inhaled particles. Journal of Aerosol Science, 21 (Supplement 1), $491-$496. Jones, A.D., McMillan, C.H., Johnston, A.M., McIntosh, C., Cowie, H., Bolton, R.E., Borzucki, G. and Vincent, J.H. (1988b). Pulmonary clearance of UICC amosite fibres inhaled by rats during chronic exposure at low concentration. British Journal of Industrial Medicine, 45, 300-304. Jones, A.D., Vincent, J.H., McIntosh, C. and Addison, J. (1994). The fate and effect of inhaled chrysotile asbestos fibres. In: Inhaled Particles Vii (Eds. J. Dodgson and

201

Aerosol science f o r industrial hygienists R.I. McCallum), Annals of Occupational Hygiene, 38, Supplement 1, Elsevier Science, Oxford, pp. 619--629. Jones, A.D., Vincent, J.H., Mclntosh, C., McMillan, C.H. and Addison, J. (1989). The effect of fibre durability on the hazard potential of inhaled chrysotile asbestos fibres. Experimental Pathology, 37, 98-102. Jones A.D., Vincent, J.H., McMillan, C.H. et al. (1988a). Animal studies to investigate the deposition and clearance of inhaled mineral dusts. Institute of Occupational Medicine (Edinburgh, Scotland, U.K.), Technical Report TM/88/05. Klosterk6tter, W. and Bunemann, G. (1961). Animal experiments on the elimination of inhaled dust. In: Inhaled Particles and Vapours I (Ed. C.N. Davies). Pergamon Press, Oxford, pp. 327-341. Le Bouffant, L. et al. (1988). Compared in vitro and in vivo toxicity of coal mine dusts: relationship with mineral composition. In: Inhaled Particles VI (Eds. J. Dodgson, R.I. McCallum, M.R. Bailey and D.R. Fisher). Pergamon Press, Oxford, pp. 611-620. Mauderly, J.L. (1994). Contribution of inhalation bioassays to the assessment of human health risks from solid airborne particles. In: Toxic and Carcinogenic Effects of Solid Particles in the Respiratory Tract (eds. D.L. Dungworth, J.L. Mauderly and G. Oberd6rster), ILSI Press, Washington DC, pp. 355-365. McMillan, C.H., Jones, A.D., Vincent, J.H., Johnston, A.M., Douglas, A.N. and Cowie, H. (1989). Accumulation of mixed mineral dusts in the lungs of rats during chronic inhalation exposure. Environmental Research, 48, 218-237. Morgan, A., Talbot, R.J. and Holmes, A. (1978). Significance of fibre length in the clearance of asbestos fibres from the lung. British Journal of Industrial Medicine, 35, 146-153. Morrow, P.E. (1988). Possible mechanisms to explain dust overloading of the lungs. Toxicology and Applied Pharmacology, 113, 1. Morrow, P.E. (1994). Mechanisms and significance of "particle overload". In: Toxic and Carcinogenic Effects of Solid Particles in the Respiratory Tract (Eds. D.L. Dungworth, J.L. Mauderly and G. Oberd6rster). ILSI Press, Washington DC, pp. 17-25. Muhle, H., Bellmann, B., Creutzenberg, O., St6ber, W., Kilpper, R., MacKenzie, J., Morrow, P.E. and Mermelstein, R. (1988). Pulmonary deposition, clearance and retention of test toner, TiO 2 and quartz during a long-term inhalation study in rats. Toxicologist, 8, 270--275. Raabe, O.G. (1989). Experimental dosimetry: introduction. In: Extrapolation of Dosimetric Relationships for Inhaled Particles and Gases (Eds. J.D. Crapo, E.D. Smolko, F.J. Miller, J.A. Graham and A.W. Hayes). Academic Press, San Diego, CA, pp. 81-90. Ruckley, V.A., Chapman, J., Collings, P., Douglas, A.N., Fernie, J.M., Lamb, D. and Davis, J.M.G. (1981). Autopsy study on coalminers' lungs. Institute of Occupational Medicine (Edinburgh, Scotland, U.K.), Technical Report TM/81/18. Soderholm, S.C. (1981). Compartmental analysis of diesel particle kinetics in the respiratory system of exposed animals. Paper presented at the EPA Diesel Emission Symposium, October 1981, Raleigh, CN. St6ber, W., Morrow, P.E., Koch, W. and Morawietz, G. (1994). Alveolar clearance and retention of inhaled insoluble particles in rats simulated by a model inferring macrophage load distributions. Journal of Aerosol Science, 25, 975-1002. Strom, K.A., Chan, T.L. and Johnson, J.T. (1988). Pulmonary retention of inhaled submicron particles in rats: diesel exhaust exposures and lung retention model. In: Inhaled Particles VI (Eds. J. Dodgson, R.I. McCallum, M.R. Bailey and D. Fisher). Pergamon Press, Oxford, pp. 645-657. Vincent, J.H. and Donaldson, K. (1990). A dosimetric approach for relating the biological response of the lung to the accumulation of inhaled mineral dust. British Journal of Industrial Medicine, 47, 302-307.

202

The fate of inhaled particles Vincent, J.H., Johnston, A.M., Jones, A.D., Bolton, R.E. and Addison, J. (1985). Kinetics of deposition and clearance of inhaled mineral dusts during chronic exposure. British Journal of Industrial Medicine, 42, 707-715. Vincent, J.H., Jones, A.D., Johnston, A.M., McMillan, C., Bolton, R.E. and Cowie, H. (1987). Accumulation of inhaled mineral dust in the lung and associated lymph nodes: implications for exposure and dose in occupational lung disease. Annals of Occupational Hygiene, 31,375-393. Vostal, J.J., Schreck, R.M., Lee, P.S., Chan, T.L. and Soderholm, S.C. (1982). Deposition and clearance of diesel particles from the lung. In: Toxicological Effects of Emissions from Diesel Engines (Ed. S. Lewtas). Elsevier, New York, pp. 143-159. Wagner, J.C., Berry, G., Skidmore, J.W., and Timbrell, V. (1974). The effects of the inhalation of asbestos in rats. British Journal of Cancer, 200, 252-269. Wolff, R.K., Henderson, R.F., Snipes, M.B., Griffiths, W.C., Mauderly, J.L., Cuddihy, R.G. and McClellan, R.O. (1987). Alterations in particle accumulation and clearance in lungs of rats chronically exposed to diesel exhaust. Fundamental and Applied Toxicology, 9, 154-166. Yu, C.P. and Asgharian, B. (1990). A kinetic model of alveolar clearance of amosite fibers from the rat lung at high lung burdens. Journal of Aerosol Science, 21, 21-28. Yu, C.P., Asgharian, B. and Abraham, J.L. (1990). Mathematical modelling of alveolar clearance of chrysotile fibers from the rat lungs. Journal of Aerosol Science, 21, 587-594. Yu, C.P., Chen Y.K. and Morrow, P.E. (1989). An analysis of alveolar macrophage mobility kinetics at dust overloading of the lungs. Fundamental and Applied Toxicology, 13, 452. Yu, C.P., Morrow, P.E., Chan, T.L., Strom, K.A. and Yoon, K.J. (1988). A non-linear model of alveolar clearance of insoluble particles from the lung. Inhalation Toxicology, Premier Issue, 97-107.

203

CHAPTER 8

Standards for health-related aerosol m e a s u r e m e n t and control 8.1 I N T R O D U C T I O N The purpose of much of what was contained Chapters 6 and 7 is aimed in part at providing information to enable the risk of aerosol exposure in the workplace (and elsewhere) to be assessed realistically and in a manner relevant to health. Part of that assessment involves making appropriate measurements of aerosol exposure. The three possible routes of exposure dermal, ingestion and inhalation ~ were mentioned in Chapter 6. For the inhalation route, attention is focused in particular on: the health-related particle size-selective criteria that have evolved from knowledge about inhalation and deposition and that are being proposed for use as the basis of such exposure measurement; and the kinetics governing the subsequent fate and effects of the deposited material and how these bear upon measurement strategy. The ultimate purpose is to provide a rational framework for the setting of standards by which aerosol exposures in workplaces can be controlled to levels consistent with maintaining the health of the workers. In general, an ideal standard may be described as containing five primary main ingredients: (1) Criteria: relating to the scientific basis upon which, for a given contaminant, a measurement or assessment procedure is chosen. It includes reference to the nature of the health effect which can result, and the portion of overall airborne contaminant relevant to the internal exposure which leads to that health effect. That portion might be a fraction based on chemical composition (e.g., the quartz content of a mine dust, the nickel content in the atmosphere of an electroplating shop, coal tar pitch volatiles in coke oven emissions, etc.) or, specifically in the case of airborne particulate matter, particle 204

Standards for health-related aerosol measurement and control

size (e.g., fine particles capable of penetrating deep into the lung) or shape (e.g., fine fibrous particles exceeding a certain aspect ratio). (2) Sampling instrumentation: the technical means by which a given relevant airborne contaminant, as identified in the specific criteria, can be extracted from the atmosphere of interest (e.g., the general workplace air or the so-called 'breathing zone' of a worker). In particular, it should be a particle size-selective sampler which aspirates and selects an appropriate portion of the workplace aerosol.

(3)

Analytical instrumentation: the technical means by which the collected sampled may be quantitated and so used to determine the exposure concentration of the fraction of interest. In the simplest case, this might take the form of an analytical balance capable of measuring the mass of material collected on a filter. Or in some cases it might consist of a more sophisticated apparatus ~ or methodology which may permit determination of relevant chemical or mineralogical subfractions or species (e.g., atomic absorption spectrophotometry for metals, infrared spectrophotometry for asbestos, X-ray diffractometry for quartz, etc).

(4) A sampling strategy: providing a framework within which sampling instrumentation is used in practice to assess the exposures of individuals (or groups of individuals), embodying considerations of where to sample, what duration, how many times, etc. In the first place, the strategy should take into account the nature and dynamics of exposure, uptake of the contaminant into the body, its redistribution, storage, metabolisation and elimination, and its toxic effects. Secondly it should take account of the inter- and intra-individual variabilities in exposure.

(5)

A limit value: defining the upper end of the range of intensity or magnitude of permissible exposure to the fraction identified in the criterion for the substance in question. For airborne contaminants, it is usually described in terms of an airborne concentration (e.g., mass of substance or number of entities per unit volume of air), derived from a measurement made by sampling over an appropriate period of time. The usual underlying rationale is that this represents the 'threshold' level of exposure at ~ and below which, according to current knowledge, there is no evidence of injury to workers if the substance is inhaled day after day. Ideally, this should derive directly from toxicological or epidemiological considerations. For many substances, a 'full-shift' 8-hour time-weighted average (TWA) is appropriate. For others, a shorter reference period might be defined.

In the minds of many, a 'standard' is synonymous with the last of these, the limit value. However, it is an important principle that, strictly speaking, a standard cannot be complete without first considering each of the other

205

Aerosol science for industrial hygienists four elements. It simply would not be logical to prescribe a limit value without reference to sampling and analytical methods and, in turn, to the criteria underpinning those methods and the strategies by which they should be applied. In addition, there is a sixth important ingredient which features frequently in the consideration of practical working limit values that is, actual workplace exposure data. A good such data set will provide characterisation of the nature of exposure in the industry in question (i.e., its magnitude and variability between work areas, job tasks and workers, etc). This in turn will enable the practical feasibility of a given proposed limit value to be assessed and provide guidance towards a workable pragmatic policy. As already mentioned, exposure to all hazardous substances (including aerosols) may occur by a number of possible routes. Exposure by inhalation applies entirely to airborne contaminants, and is a widespread primary mode of exposure in workplace situations. Here it is relatively easy to identify the target system in exposed subjects (i.e., the respiratory tract) and the physical and physiological parameters (e.g., physical properties of the contaminant, breathing rates, airway dimensions, etc.) which govern how ~ and how much contaminant arrives at given regional sites in the respiratory tract. This, therefore, provides a clear scientific means of identifying criteria upon which a relevant health-related measurement might be made and, subsequently, a basis for dose estimation. For exposure by the ingestion and dermal routes the situation is less clear-cut, so that truly scientific options for criteria are limited. For these routes, therefore, any approach to standards needs to be developed based on information about job tasks, work practices, the nature of the potentially harmful substances, and assessment of the industrial process in question. Whilst recognising the potential importance of such exposures in some situations, the setting of standards in relation to the inhalation route of exposure remains a major focus for the setting of health-related standards. It is, therefore, a primary emphasis for this book. A major part of what follows is a description of the evolution of health-related particle size-selective criteria which can, in turn, be used as 'yardsticks' for the performance characteristics of 'ideal' practical aerosol sampling instruments. This process began as a series of often-fragmented but related ~ ideas which emerged from various standards organisations. However, over a period of about 15 years from the late 1970s onwards, there has been a quickening of the pace towards an internationally-harmonised set of criteria. The following sections therefore outline the historical development of the individual criteria, and then go on to describe the harmonised set of criteria that have emerged. Although this description relates to the status as of 1995, and the situation is expected to evolve still further, the basic principles outlined below will continue to be relevant to future generations of aerosol standards.

206

Standards for health-related aerosol measurement and control

8.2 P R O G R E S S T O W A R D S C R I T E R I A F O R M E A S U R E M E N T OF COARSE AEROSOLS As described earlier, the first part of the overall process of aerosol exposure is the entry by inhalation of particles from the ambient air and into the respiratory tract. For some types of aerosol, particles constitute a risk to health immediately upon entering the body in this way, regardless of where they are eventually deposited. So, in the first place, inhaled aerosol is a fraction which is an objective for measurement in itself in many occupational situations. In the past the recommendations for the health-related sampling of coarse particles in most countries have been and continue to be ~ based on the concept of so-called 'total' aerosol. This concept is intended to relate to all particulate matter which might be considered airborne. Practical sampling instruments for 'total' aerosol have been developed with this in mind and sold commercially, and have been extensively deployed in industrial hygiene. However, most such instruments were originally developed without particular regard to specific quantitative criteria or performance indices. Now, upon closer inspection, it is becoming apparent that their performance characteristics have varied greatly from one to the other. It follows, therefore, that 'total' aerosol has been effectively defined in each particular situation by the particular sampling instrument chosen to do the job. In addition, switching from one instrument to another in a given practical situation might well produce different measurements of exposure, even though the level of exposure itself might not have actually changed. This situation brings into even sharper focus the previous argument about the need to integrate sampling criteria, instrumentation and limit values into the standards concept. In the 1970s, the idea first emerged of the human head as an aerosol sampler, and hence of inhalability as a quantitative definition of what had previously been known as 'total' aerosol. The way in which the subject has since evolved has led to some confusion on terminology. Early use of 'inhalable' in the early 1980s gave way to 'inspirable' in order to avoid a clash with other uses of the term (e.g., in the United States in relation to the definition of a finer fraction for the purpose of air pollution measurement). Now, however, there is increasing international harmonisation on methodology and terminology in this whole area, and there is general agreement on the term 'inhalable'. As will be related below, this is now providing the basis for departures from the inconsistency of the old 'total' aerosol approach and is the starting point for a more scientific sampling rationale. Since the concept of inhalability was first suggested, a number of important experimental studies have been conducted in laboratories in Britain and Germany to determine the efficiency with which particles enter the human head during breathing through the nose and/or mouth. These experiments,

207

Aerosol science for industrial hygienists involving life-sized human mannequins in large wind tunnels, provided data for a range of windspeeds relevant to workplace exposures and for particles with aerodynamic diameter up to 100 txm. The results are summarised in Chapter 6. They suggest that the efficiency of inhalation of a particle can, for practical workplace purposes (where ambient windspeeds are usually less than about 4 ms-l), be described in terms of a single function of particle aerodynamic diameter (dae), starting off at 100% for very small particles and falling to about 50% for particles with da~ around 30 Ixm and above. For cases (e.g., outdoors) where windspeeds are greater than 4 m s-i, an additional windspeed dependence is required in order to describe the observed trends. Data from experiments like these have formed the basis of recommendations for replacing the old 'total' aerosol concept with a quantitative sampling convention based on human inhalability. The first ~ and historically important for that reason ~ was by the International Standards Organisation (ISO, 1983). During the meetings of its ad hoc working group in the late 1970s, the only data available for consideration were those from the experiments of Ogden et al. (1977) and Ogden and Birkett (1977) for d ~ up to about 30 Ixm. These data were extrapolated to larger particle sizes and used to arrive at the curve described by the purely empirical expression l(dae ) -

1 - 0.15 {log]0(l + d a e ) } - 0 . 1 0 l o g ] 0 ( 1 + d a e )

(8.1)

where dae is expressed in [Ixm] and the aspiration efficiency for the human head (A) used in Chapter 6 is now replaced by inhalability (I) to indicate that it has now been applied to a convention for aerosol sampling. This curve is shown in Figure 8.1, where the dotted line indicates the part of the curve based on extrapolation from the actual experimental data. In the 1983 ISO recommendations it was stated that, when this curve is used as a criterion for practical sampling purposes, a tolerance band may be applied where da~ may be allowed to vary by +15% at each prescribed value. Further latitude is permitted if it can be demonstrated that 67% of mass samples obtained in

4

extrapolation beyond data available

j

at the time

,.0

~

50

~

~

I 100

Particle aerodynamic diameter, dae (p~m) Figure 8.1.

First inhalability curve as proposed by ISO in 1983.

208

Standards for health-related aerosol measurement and control

practice fall within _+10% of the result that would be achieved had Equation (8.1) been followed exactly. Somewhat later, the more comprehensive data set provided by the addition of results from subsequent experimental studies (Vincent and Mark, 1982; Armbruster and Breuer, 1982) provided the basis for the proposal of a more representative curve (Vincent and Armbruster, 1981). The main difference in the new, modified curve from the earlier ISO curve is the absence of a cut-off at large particle sizes and ~ instead ~ the levelling-off of I at around 0.5. In the recommendations of the American Conference of Governmental Industrial Hygienists ( A C G I H , 1985), this curve was included as a convention for defining the inhalable (known as 'inspirable' at the time) fraction and described by the empirical expression (8.2)

/(dae ) -- 0.5 {1 + exp ( - 0 . 0 6 dae ))

for dae (again expressed in [Ixm]) up to and including 100 Ixm. Beyond 100 Ixm the A C G I H recommendations acknowledge that there is no information on which to base a firm recommendation. It should be noted, however, that this does not imply a 'cut-off' (i.e., I ~ 0), as the A C G I H report is sometimes misunderstood as saying. The curve given by Equation (8.2) is shown in Figure 8.2. In its original 1985 convention, the A C G I H also recommended that, as a performance band for practical sampling instruments, I may vary by +0.1 for each value of dae. That is, as with the ISO recommendations, there is a working envelope which is representative of a high proportion of the original data and so is a reasonable basis for assessing the performances of practical devices. In the ISO and A C G I H proposals referred to, attention appears to have been focused primarily on particle size-selective criteria for sampling in indoor workplaces where windspeeds even as high as 4 m s-1 are uncommon. For most workplace conditions, therefore, curves like those in the two conventions

~ o,N

0

!

50

!

100

Particle aerodynamic diameter, dae (l~m) Figure 8.2.

Inhalability curve in the form proposed by Vincent and Armbruster (1981) and recommended by ACGIH in 1985.

209

Aerosol science for industrial hygienists

described above may be regarded as generally adequate. However, there are some outdoor workplaces where conditions might sometimes lie outside these wind conditions. Furthermore, at least as far as the ISO is concerned, sampling in the ambient atmosphere for the purpose of evaluating the risk to the community at large is an important part of its general remit. From the more recent experimental evidence shown earlier in Figures 6.4 and 6.5, it is clear that the existing ISO and A C G I H conventions do not properly represent what happens at those higher windspeeds. In particular, the experimental data suggest that using samplers with performance based on either of these curves could lead to a significant underestimation of the exposure of humans to large particles. This could be important in situations where there are large particles containing potentially hazardous substances (e.g., radioactive nuclides, heavy metals, polycyclic aromatic hydrocarbons, etc.). It is therefore appropriate to consider how the definition of inhalability might be extended. As shown earlier, Equation (6.13) represents the experimental data for the aspiration efficiency of the human head quite well over the range of conditions examined. But it is substantially more complicated than both the existing conventions shown above in Equations (8.1) and (8.2) and so is not really suitable for direct application as a practical criterion for the performance of sampling instruments. In the search for a more suitable expression, the much simpler A C G I H curve provides perhaps the best starting point. With this in mind, a new convention has been proposed, based on an empirical modification of the 1985 A C G I H curve of the form (Vincent et al., 1990) I(dae ) = 0.5 {1 + exp (-0.06 dae)} + B(dae, U)

(8.3)

where the new, additional term (B) is given by B(dae, U) = p Uq exp {rdae }

(8.4)

in which dae is again in [p~m] and U is in [m s-l]. The coefficients p, q and r are constant, for which best fit with the experimental data are achieved for p = 1 x 10- 5 , q = 2.75 a n d r = 0.055 The new convention is shown graphically in Figure 8.3 where it is seen at low windspeed (U < 4 m s-1) to be almost identical to the A C G I H curve, and so is applicable to most workplace situations. At higher windspeed, however, it also gives good agreement with the trends exhibited by the latest experimental data obtained under those conditions. In general, therefore, the new expression as described by Equation (8.3) corresponds well with the experimental data over the full range of conditions. Recently this has been incorporated into a revised ISO set of proposals (ISO, 1992).

210

Standards for health-related aerosol measurt .'~ent and control

:

/ 4j

,.,,,,

~

0

,.~

-

I

0

I 100

50 Particle aerodynamic diameter, d~ (~m)

Figure

Inhalability curves adjusted for high windspeed effects, as proposed by Vincent et al. (1990) and adopted by ISO in 1992.

8.3.

8.3 P R O G R E S S T O W A R D S C R I T E R I A F O R T H E M E A S U R E M E N T OF FINER A E R O S O L F R A C T I O N S

Thoracic aerosol Thoracic aerosol deposition was defined formally in Chapter 6 as the sum of the tracheobronchial and alveolar deposition subfractions of the inhalable fraction, for which data are available by combining the separate measurements of tracheobronchial and aNeolar lung deposition. H o w e v e r , the conventional curve for thoracic aerosol proposed by both ISO (1983) and A C G I H (1985) was based on the efficiency with which particles penetrate to the thoracic region (see Figure 6.12 for a s u m m a r y of the data for such penetration suitably adjusted to a 'reference' h u m a n subject). This approach therefore implies that ~ if the conventional curve is m e a n t to relate to actual exposure ~ the mass contained in the aerosol fraction that is exhaled should generally be negligible. The extent to which this is true depends on the aerosol particle size distribution, but is probably a fair working assumption for most workplace aerosols. In any case, for the purpose of a convention that is to be used as the basis for routine health-related aerosol sampling and standards, the use of a curve based on penetration to ~ as opposed to deposition in the region of interest errs on the side of tzeing consevative. The basic convention in its original form proposed by both ISO and A C G I H is shown in Figure 8.4. It follows a curve based o n a cumulative log-normal function with its median at dae = 10 Ixm and having a geometric standard deviation (~rg) of 1.5. That is, it has the mathematical form

y-

1-

ie[ 1 0

exp

{ l(,n,x,lO,)2}] -

x X / ~ lnl.5

dx

2

211

lnl.5

(8.5)

Aerosol science for industrial hygienists 1.0 ISO (1983) and ACGIH (I 985) version

O O

r

0.5

u O

0

I0

20

30

Particle aerodynamic diameter, dae (~m)

Figure 8.4. Conventional curve for the thoracic fraction, as recommended by ISO in 1983 and ACGIH in 1985.

Comparing Figure 8.4 with the data in Figure 6.12, it is seen that the conventional curve lies somewhat to the right. This too was justified on the grounds that m e a s u r e m e n t s based on this criterion would err on the side of the 'worst-case' situation. Again, tolerance bands are suggested in both the ISO and A C G I H recommendations. For ISO, it is the same as stated above for the inhalable fraction. That is, dae may be allowed to vary by + 1 5 % at each prescribed value but that further latitude is permitted if 67% of mass samples obtained in practice fall within + 1 0 % of the result that would be achieved had the convention been followed exactly. For the A C G I H version, the geometric median value of dae is allowed to vary by +_1 txm, while O'g may vary by +0.1.

Tracheobronchial and alveolar aerosol

Of the thoracic larynx, a further tracheobronchial region. Most of

aerosol, representing particles that penetrate below the subdivision takes place between what is deposited in the region and that which penetrates down to the alveolar the effort over the years has gone into defining the finer

212

Standards for health-related aerosol measurement and control

--o

1.0-~CGIH

(1968)

o'~ --~ .~. o"

BMRC (1952) 0.5 ACGIH (1985)

~

I

5

IO

Particle aerodynamic diameter, dae (l~m) Figure 8.5.

Various conventional curves of historical interest which have been proposed for the respirable fraction.

alveolar penetration fraction. Indeed, the history of setting particle sizeselective criteria for aerosol measurement began with identification of the need to measure fine aerosol fractions in workplaces, with special reference to lung diseases of the deep lung such as pneumoconiosis. That fine alveolar penetration fraction became widely known as the respirable fraction, and is still widely referred to as such. In Figure 8.5, a number of the curves are shown which have been adopted over the years as conventions for the respirable fraction. These include the historically-important British Medical Research Council (BMRC, 1952) curve and the first A C G I H curve ( A C G I H , 1968), both of which were contained as options in the early ISO (1983) recommendations. Also shown is the more recent 1985 A C G I H version. ISO specifies tolerances for respirable aerosol in the same terms as for the inhalable and thoracic fractions. For the 1985 A C G I H version, the geometric median value of dae is allowed to vary by +0.3 Ixm, while O'g may vary, as before, by +0.1. Here it should be recognised ~ as already discussed for thoracic fraction that the approach to such respirable conventions does not purport to reflect deposition p e r se. Rather, as stated, the respirable fraction is intended to be representative of the penetration of particles to the alveolar region, and hence the availability of particles for deposition. So, again, in relation to actual human alveolar deposition (and hence exposure at that region), there is a mass discrepancy resulting from the effect of the exhaled aerosol. But again, this is likely to be small for many workplace aerosols. For both the thoracic and respirable aerosol fractions, based on the penetration of particles to the thoracic and alveolar regions respectively, it is important to note that, although such curves are considered sufficient for routine health-related sampling and standards, non-routine investigative exposure assessment in relation to specific aerosol-related ill-health (e.g., in

213

Aerosol science for industrial hygienists

epidemiologic research), would best be carried out using sampling instrumentation that follows more closely the deposition as opposed to penetration data. As discussed later in Chapter 9, this identifies a role for versatile aerosol spectrometers such as cascade impactors. In both the ISO and A C G I H documents, the need to define the tracheobronchial fraction separately is not given high priority since, it is felt, the thoracic fraction usually meets the practical needs adequately. If desired, a rough working convention for the tracheobronchial fraction may be derived simply by subtracting the respirable curve from the thoracic curve. However, caution is urged in doing so since both the thoracic and respirable curve are merely conventions, reflecting the original experimental inhalation data but not always being strictly faithful to them.

8.4 H A R M O N I S A T I O N OF C R I T E R I A F O R A E R O S O L S T A N D A R D S As can be seen from the above, there have been considerable variations in the way the experimental inhalation and lung deposition data have been interpreted as bases for criteria for aerosol exposure assessment. Although this has worked reasonably well in the past in relation to national standards, it has become clear in recent years that an international concensus on criteria and ultimately standards is desirable. So harmonisation has been increasingly considered to be a worthwhile and useful goal. The proposals for new criteria for health-related aerosol standards include recognition of the various processes by which aerosols come into contact with the human respiratory system, reflecting: aerodynamic processes outside the body by which particles are inhaled through the nose and/or mouth; and processes inside the body whereby those inhaled particles are deposited. Thus, inhaled aerosol is seen to be a fraction of the true total workplace aerosol. The latter is the aerosol whose concentration would be measured using a sampler whose sampling efficiency is unity (or 100%) for all relevant particle sizes. In turn, thoracic and respirable aerosol are subfractions of the inhalable fraction. The 1983 ISO recommendations explicitly included such a philosophy. In practical terms, this requires that a sampling instrument for a fine aerosol subfraction should first aspirate the inhalable fraction and then select the desired subfraction. The 1985 A C G I H recommendations did not embody this philosophy, opting for a simpler approach in which each of the inhalable, thoracic and respirable fractions is defined independently. Real progress towards harmonisation of these criteria gained momentum during the period 1989-1992 with coordination of the activities of ISO, CEN

214

Standards for health-related aerosol measurement and control

and A C G I H working groups and committees, and was underlined at the ACGIH-sponsored International Symposium on Air Sampling Instrument Performance which was held at Research Triangle Park, NC, in October 1991. Discussions dealt not only with the issues of philosophy and principle but also with the quantitative descriptions of the various aerosol fractions. Firm agreement has since been achieved between ISO (1992), A C G I H (1993-1994) and the Comit6 Europe6n Normalisation (CEN, 1992) on the following harmonised criteria: Health-related sampling should be based on one or more of the three, progressively-finer, particle size-selective fractions; inhalable, thoracic and respirable. The inhalable fraction is described empirically by a single curve relating inhalability (/) and particle aerodynamic diameter (dae). Thus

I(dae ) - - 0 . 5 {1 + e x p ( - 0 . 0 6 d a e ) }

(8.6)

reiterating what was already stated in the original 1985 A C G I H definition, as expressed by Equation (8.2). Again, it is intended that this should replace the ill-defined so-called 'total' aerosol on which most current practice is still based. For high windspeed situations, ISO recommends the additional modifications noted in Equation (8.3). The thoracic fraction takes the form (8.7)

T(dae ) = /(dae ) { 1 - gT(dae ) }

where FT(dae ) is a cumulative log-normal function with its median at dae - 11.64 txm and having a geometric standard deviation (Crg) of 1.5. That is

F T (dae) -

i[ 0

1 x X / ~ lnl.5

exp

{

. . . . . . 2 lnl.5

dx

(8.8)

Expressed in the manner given in Equation (8.7), it is clear that thoracic aerosol is a subfraction of the inhalable fraction, consistent with the physical nature of actual human exposure. This means that, as a fraction of inhalable aerosol, the thoracic fraction falls to 50% at dae -- 11.64 ~m. It also means that, as a fraction of true total aerosol, it falls to 50% at dae = 10 Ixm as described in the original 1985 A C G I H recommendations. The proposed adjustment is achieved whilst still retaining fair agreement between the conventional curve and the experimental deposition data.

215

Aerosol science for industrial hygienists Similarly, the respirable fraction takes the form R ( d a e ) -- I ( d a e )

{1-

(8.9)

FR(dae)}

where FR(dae ) is a cumulative log-normal function with its median at dae = 4.25 ~m and having a geometric standard deviation (O-g) of 1.5. This means that, as a fraction of inhalable aerosol, the respirable fraction falls to 50% at dae = 4.25 p~m; or, as a fraction of total aerosol, it falls to 50% at dae --- 4 ~m. This definition of respirable aerosol represents a significant change from both the earlier ISO and A C G I H recommended curve. The new curve shown here was originally proposed by Soderholm (1989) in order to achieve a workable compromise between the original A C G I H curve (median dae = 3.5 p~m) as applied widely in the United States and the B M R C curve (median dae = 5 v~m) as applied widely in Europe.

Table 8.1. Table of values for the health-related fractions identified in the new particle size-selective criteria for aerosol exposure assessment (see also Figure 8.6). (Values given are rounded to 2 significant figures.) Inhalable (fraction of total)

Thoracic (fraction

0

1.00

1

0.97

2 3 4 5 6 7 8 9 10 12 14 16 18 20 25 30 35 40 45 5O 100

0.94 0.92 0.89 0.87 0.85 0.83 0.81 0.79 0.77 0.74 0.72 0.69 0.67 0.65 0.61 0.58 0.56 O.55 0.53 0.52 0.50

dae (~m)

Thoracic (fraction of total)

Respirable (fraction of I)

Respirable (fraction of total)

1.00

1.00

1.00

0.97

1.00 1.00 0.99 0.98 0.95 0.89 0.82 0.74 0.65 0.47 0.33 0.22 0.14 0.09 0.03 0.01

0.94 0.92 0.89 0.85 0.80 0.74 0.66 0.58 0.50 0.35 0.23 0.15 0.09 0.06 0.02 0.01

1.00 1.00 0.95 0.80 0.56 0.35 0.20 0.11 0.06 0.03 0.O2

1.00 0.97 0.90 0.73 0.50 0.30 0.17 0.09 0.05 0.03 0.01

of/)

216

Standards for health-related aerosol measurement and control 1.00.9 0.8

Inhalable

0.7 0.6 O

9 0.5 r-" [a., 0.4 Thoracic 0.3 0.2 Respirable

0.1 | 100

10 Particle aerodynamic diameter dae (p.m)

Figure 8.6. Summary of the conventional curves that have been internationally agreed as of 1993-1994 (by ISO, ACGIH and CEN) for the inhalable, thoracic and respirable fractions. This figure shows thoracic and respirable aerosol as subfractions of the inhalable fraction. It is expected that these criteria will form the basis of future standards. T h e s e h a r m o n i s e d criteria are given in T a b l e 8.1 a n d s h o w n g r a p h i c a l l y in F i g u r e 8.6. T h e s e are i n t e n d e d for use as ' y a r d s t i c k s ' for the p e r f o r m a n c e s of practical a e r o s o l s a m p l e r s . Example 8.1. For an aerosol which, in the workplace air, has a concentration of 12 mg m -3 and a log-normal particle size distribution with a mass median aerodynamic diameter (MMAD) of 15 Ixm and geometric standard deviation (Crg) of 2.1, calculate the mass concentration contained in the inhalable, thoracic and respirable fractions. Assume that m(dae) is the non-normalised particle size distribution of the total aerosol, and so is given by Equation (3.11) in the form 12

{ exp

m (dae) =

1 (ln(dae/dm))2} -

dae N / ~ lntrg

2

lntrg

The mass concentration contained within the total aerosol is given by

Ctot

"

-

f m(dae ) ddae = 12 m g m -3

0

217

Aerosol science for industrial hygienists For the inhalable fraction O0

tin h =

f l(dae) m(dae)ddae 0

where the combined function inside the integral sign on the right is the non-normalised particle size distribution of the inhalable fraction of total aerosol. Likewise for the thoracic fraction

Cth~

T(dae) m(dae) dda e 0

and for the respirable fraction

Cresp =

f

n(dae) m(dae) ddae

0 The non-normalised particle size distributions for the four fractions are shown in the following figure. The concentration of aerosol contained in each is the area under each curve.

0.6

aerosol 0.4 E

aerosol

tall

racic aerosol ,_ Respirable aerosol

.-1

E

0.2

0

10

20

30

Particle aerodynamic diameter

40

50

dae(v,m)

Unfortunately the form of m(dae ) does not permit the necessary integrations to be carried out analytically. So a numerical approach is required, and this is facilitated by the availability of modern spreadsheets. For most practical purposes, the simplest approach to the integration

218

Standards for health-related aerosol measurement and control is to assume that the curves shown in the figure are histograms. So the area u n d e r each segment of the histogram can be calculated very easily (as height x width), and the sum of such areas provides the desired total a r e a -

and hence the concentration. H o w e v e r ,

it should be noted that the assumption that each s e g m e n t in the histogram is rectangular does introduce some errors, these becoming m o r e significant the g r e a t e r the width chosen for the segments. So for m o r e accurate calculations, o t h e r numerical integration m e t h o d s are available. A simple one involves treating each section of the curve as a trapezoid (the 'trapezoidal rule'). C o m p a r i s o n between these options is illustrated as follows:

SIMPLEST OPTION

Add these, and the shaded areas

"7 E =1.

represent an overestimate of the integral

t~

BETTER OPTION t~

Adding these gives a better estimate of the integral

dae (l~m)

In the present e x a m p l e , the reader is e n c o u r a g e d to try these m e t h o d s . T h e results, using the trapezoidal rule, are" *

Cinh = 8.11 mg m -3

*

Cthor

= 3.68 mg m -3

*

Cresp

= 0.72 mg m -3

219

Aerosol science for industrial hygienists Example 8.2. A s s u m e the (hypothetical) proposal that, for the purpose of assessing the acceptability of practical samplers in relation to the respirable aerosol criteria, a tolerance band should be defined as the envelope b e t w e e n a pair of curves R(dae)4.5 and R(dae)4.0. These are r e p r e s e n t e d in E q u a t i o n (8.9) by FR(dae)4.5 and FR(dae)4.0 for which the median values for dae are shifted to 4.5 and 4.0, respectively. Two samplers, one with overall particle size selective p e r f o r m a n c e matching one curve (the top side of the tolerance band) and the other matching the o t h e r curve (the lower side of the tolerance band) are used to sample the workplace aerosol referred to in the previous example. W h a t is the difference in mass concentration sampled? Again this problem involves numerical integrations, for which a spreadsheet is suitable. By this means, create the following table"

dae (p~m)

Mass4.5 Mass4.0 < dae < dae

m

R4. 5

R4. 0

mR4.5

mR4.0

0 1 2 3 4 5 6 7 8 9 9 10

0. 000 0. 008 0.081 0.205 0.330 0.431 0.502 0.544 0.563 0.566 0.566 0.555

1.00 0.97 0.91 0.76 0.55 0.35 0.21 0.12 0.07 0.04 0.04 0.02

1.00 0.97 0.89 0.69 0.45 0.26 0.14 0.07 0.04 0.02 0.02 0.01

0. 000 0. 008 0.074 0.156 0.181 0.151 0.104 0.064 0.037 0.020 0.020 0.011

0. 000 0. 008 0.072 0.142 0.149 0.112 0.070 0.040 0.021 0.011 0.011 0.005

0.000 0. 004 0.045 0.160 0.329 0.495 0.622 0.706 0.760 0.784 0.784 0.799

0. 000 0. 004 0.044 0.150 0.296 0.426 0.517 0.571 0.601 0.617 0.617 0.625

14

0.459

0.430

0.00 0.00

0.001 0.000

0.000 0.000

0.810 0.811

0.632

15

0.00 0.00

0.632

where

m R45 R4.01

-

the non-normalised frequency particle size distribution function for the total workplace aerosol (in dimensions of [mg i~m-1]) the calculated enveloping curves for the respirable fraction (upper and lower)

220

Standards for health-related aerosol measurement and control

mR4"5/ = mR4.0 Mass45/ Mass4.0

the non-normalised frequency distribution functions for the aerosol sampled by the two samplers, respectively, given by rn x R (also expressed in [mg ixm-l]) the cumulative mass concentrations for each sample, given by the area under the curves for mR (both in [mg m-3])

Note again that the cumulative mass concentrations were calculated by integrations performed using the trapezoidal rule. From the preceding, the two sampled concentrations are 0.81 and 0.63 mg m -3, respectively. *

The two sampled concentrations differ by about 30%.

At this point, as seen in Example (8.2), the definition of a curve for defining a particular aerosol fraction raises a difficult question. H o w do we decide whether or not the performance of a given instrument in relation to the convention is acceptable? In practice, it becomes necessary to define a test protocol for deciding w h e t h e r or not an instrument's p e r f o r m a n c e matches the appropriate particle size-selective criterion. In turn, 'acceptable performance' may be defined by requiring that an instrument's aspiration efficiency curve falls within the specified tolerance band. H o w e v e r , herein lies a potential problem. Bartley and D o e m e n y (1986) were the first to point out, based on numerical studies for a range of respirable aerosol samplers with performance curves lying within tolerance bands prescribed in the 1985 A C G I H r e c o m m e n d a t i o n s , that the sampled respirable mass of a given aerosol can ~ for typical workplace dusts but depending strongly on particle size distribution ~ differ considerably. Such biases can occur for samplers whose performances might, on the basis of the sampling efficiency data alone, appear to match the criterion in question. In extreme cases (e.g., for coarse aerosols with a relatively small portion of their mass in the respirable range), Bartley and D o e m e n y found that the bias in respirable mass sampled could be as large even as x 5 in some cases! Therefore we are alerted to the need, w h e n testing aerosol samplers in relation to specific particle size-selective criteria, to be cognisant not only of the proximity of data points to the 'target' curve but also of the implications to errors in mass m e a s u r e m e n t d e p e n d e n t on particle size distribution. A l t h o u g h the original 1983 ISO r e c o m m e n d a t i o n s did go some way towards acknowledging this problem, it is clear that more effort is needed. As a primary goal, a well-defined experimental protocol is required in order to test individual samplers against specific conventions. Otherwise, a poor test might claim to show that an instrument conforms adequately to the criterion

221

Aerosol science for industrial hygienists

in question when, in fact, experimental errors could make the claim doubtful. From the work carried out so far, considerable progress has been achieved in understanding the nature and magnitude of the problem. The so-called 'bias-map' approach appears suitable, and a series of recently-published works (e.g., Liden and Kenny, 1992, 1993) provide a good insight of how this would work in practice. In the first instance, this approach needs a detailed experimental characterisation of the particle size-selective performance of a given sampler (as has been described, for example, for a number of respirable aerosol samplers by Liden and Kenny, 1991), leading to a definitive, practical performance curve. This curve can be inspected directly with respect to the extent to which it falls within the tolerance bands of the criterion in question (respirable aerosol in the case of the experiments of Liden and Kenny). For an aerosol with a given particle size distribution, it is then straightforward to numerically compute and compare the masses that would be collected respectively by the given sampler and by a 'perfect' sampler (i.e., one whose performance curve matches exactly the curve for the specified 'target' criterion). This provides the bias ~ the fractional difference in mass collected between the actual and perfect samplers. The bias map referred to is then a plot of the contours of equal mass bias on axes representing the mass median particle aerodynamic diameter (MMAD) and geometric standard deviation (~g) for ranges of aerosol particle size distributions. A typical example is given in Figure 8.7. The CEN has proposed a draft standard for the assessment of the performances of sampling instruments (Liden, 1994). In it, it is suggested that a given sampler may be classified on the basis of the proximity of

3o

2O

<

10%

1.5

2.0

2.5

3.0

3.5

O'g

Figure 8.7. Example of a typical bias map which enables assessment of the acceptability of a given sampler for achieving mass determination performance consistent with a given particle size-selective criterion.

222

Standards for health-related aerosol measurement and control its sampling efficiency curve to the criterion of interest and its accuracy in collecting the 'correct' mass. This requires extensive laboratory testing of the given sampler and numerical analysis, by which it may be placed in one of three three classifications. In descending order, these are Class 1 for the most accurate, Class 2 and Class 3. In addition, Class 4 may be applied to a sampler which is not tested according to this laboratory-based protocol but which fares acceptably well in statistically-viable, side-by-side workplace comparisons with a designated 'reference' sampler. At the time of writing, this sampler assessment standard is still under discussion. In the meantime, the best that can be recommended at present is that we should continue to judge sampler performance on the basis of its proximity to the specified curve, but that we should be aware of the errors that can arise and the conditions under which they can become significant.

Fibrous aerosols

Fibrous aerosols, as has been mentioned several times, can pose a particularly serious risk to health, and so represent a special problem for the industrial hygienist. Asbestos-containing dusts come into this category. Because of their unusual morphological characteristics touched on earlier, such fibres are specifically not included in the above conventions. Particle size selection by aerodynamic diameter alone is not sufficient for such particles. Instead, there are different sets of criteria for particle size-selective measurement, based on an appreciation of both: the aerodynamic characteristics of particle motion which govern fibre deposition in the alveolar region of the lung; and the biological effects that can then ensue and influence the subsequent fate of the particles. It is also. a relevant factor that exposure to asbestos and other fibrous workplace dusts is usually very low in terms of mass concentration. Furthermore, it usually occurs in the presence of other particulate matter whose mass concentration far outweighs (and so 'masks') the fibrous material. So mass sampling for fibrous aerosols is rarely an acceptable option for workplaces. With this practical issue in mind, along with the scientific considerations mentioned above, it has become a common convention ~ since the 1960s to assess 'respirable' fibres in terms of the airborne number concentration of relevant particles. Here it is very important to recognise that this use of the term 'respirable' is not the same as that described earlier for non-fibrous particles. But it has become widely-used in relation to fine fibrous aerosols. So, to avoid confusion, it is given here in inverted commas.

223

Aerosol science for industrial hygienists Inevitably, the above scenario leads to microscopy as the primary tool for quantitative assessment of sampled aerosol, since this enables individual fibres to be identified visually. Criteria have therefore evolved which provide a basis for deciding (a) whether a given observed particle qualifies to be called 'a fibre'; and, if so, (b) whether that particle's dimensions are such that it is likely to be able to penetrate down to the deep lung and then be harmful when it has been deposited there. Taken together, these factors determine whether a given particle may be regarded as a 'respirable' fibre. Again, it is emphasised that the use of the term 'respirable' is quite different from that used earlier to describe non-fibrous particles. Routine assessment for airborne asbestos fibres is carried out using phase contrast microscopy (PCM), an optical microscopy approach which enables improved contrast for the visualisation of relatively fine fibres. Although this technical approach stretches the range of conventional optical microscopy, it is ultimately limited by fundamental optical considerations to fibres of diameter no less than about 0.25 ~m. Criteria for the optical identification and counting of fibres have been proposed by a number of bodies, two of the most prominent being the European-based Asbestos International Association (AIA) and the U.S. National Institute for Occupational Safety and Health (NIOSH). A I A (1979) require that the particle must satisfy the following basic counting rules: aspect ratio greater than 3; length greater than 5 ~m; and diameter less than 3 ~m. Rules published by other bodies (e.g., the Australian National Health and Medical Research Council, the World Health Organisation, etc) are similar. Those contained in the NIOSH Method 7400 (NIOSH, 1979; see also Carter et al., 1989) are basically similar, but do not impose any constraint on fibre diameter. Apart from the difference with respect to diameter highlighted between the A I A and NIOSH methods, other differences between the various sets of rules are concerned primarily with the details of how to deal with fibres which are attached to other particles, split fibres, fibres which lay partly outside the microscope field of view, etc. Although such criteria were originally developed for the assessment of airborne asbestos fibres, there remains the question of identification. It is not always easy to determine whether a fibre observed by PCM is indeed asbestos. So most of the criteria that have been proposed relate only to geometrical considerations, and do not involve identification of mineralogical species. That is, if the particle is fibrous and looks like asbestos, it should be treated as if it were abestos. Meanwhile, however, there has long been concern about the possible health effects of non-asbestiform materials which can give rise to fine airborne fibres

224

Standards for health-related aerosol measurement and control

(e.g., man-made mineral fibre). So similar criteria are being applied to the measurement of 'respirable' fibres of all such materials. Although the above geometrical criteria are well established for the cost-effective and routine assessment of most exposures to fine aerosols, a number of proposals have been made for variants reflecting more directly the specific health effects that can arise. For example, Walton (1982), Pott (1978) and Lippmann (1988) have all suggested that, as far as lung cancer or mesothelioma are concerned, an increased emphasis on even longer fibres should be considered. On the other hand, Timbrell et al. (1988) argued that, for asbestosis, a more appropriate index of fibre concentration might be one based on the effective surface area of 'respirable' fibres.

8.5 S T A N D A R D S

The 'traditional' approach Particle size-selective sampling as outlined above has two main aims. The first is to provide a basis for biologically-relevant aerosol measurement for epidemiological research leading to the setting of more meaningful exposure standards in the future. The second is to provide a basis for routine aerosol measurement in relation to controlling the quality of the workplace atmospheric environment in relation to such standards. Until now, the 'traditional' approach to aerosol standards for workplaces has involved the concepts of 'total' aerosol and respirable aerosol. Here, the 'total' aerosol approach is the default. So it is applied in the first instance for all aerosols which may be described as a 'nuisance'. For this, no specific health risk is defined but the standard relates to dust levels beyond which discomfort can be experienced or where the level of aerosol could present an undesirable level of challenge to the body's defence mechanisms. The A C G I H (1993-1994) is more specific, defining such aerosol as particulates not otherwise classified (PNOC). As specified by A C G I H , such aerosols are 'inert' in that they do not produce any significant clinical, pathological or toxic effect (including changes to lung structure, scarring or fibrosis) when exposures are kept under 'reasonable' control. They do, however, produce some biological responses, but these are reversible when the challenge is removed. The second instance for the 'total' aerosol approach occurs for aerosols which present a risk as soon as they come into contact anywhere in the body. This refers to, for example, any substances which may be carcinogenic (e.g., nickel) or soluble substances (e.g., lead) which can produce adverse health effects systematically. Regarding the respirable aerosol approach, this is applied when the health effect in question produced only in the alveolar region of the lung (e.g., pneumoconiosis, emphysema).

225

Aerosol science for industrial hygienists

During the 1970s, it began to be realised that a more detailed particle size-selective breakdown of aerosols is desirable, and one or two exceptions to the traditional approach began to emerge. For example, in cotton industry workplaces in the United States, the 1975 N I O S H dust standard is based on an intermediate size fraction, defined in terms of the performance of a sampler based on the vertical elutriator principle which allows particles to be sampled with da~ up to around 15 I~m. This might be taken as acknowledgement of the fact that cotton workers' byssinosis is a complaint of the conducting airways (i.e., the tracheobronchial region) of the lung. Elsewhere, in the 1980s US Environmental Protection Agency (EPA, 1984) proposed new standards for suspended particulate matter in the ambient atmosphere based specifically on a fraction closely identified with particles which may penetrate to the thoracic region. This fraction is referred to as 'PM10', reflecting a penetration curve which falls to 50% at d ~ = 10 p~m.

The new criteria and health effects

The adoption of the new particle size-selective criteria by I S O / A C G I H / C E N have given renewed impetus towards a new set of standards based on a more scientifically-relevant rationale; that is, on the three fractions, inhalable, thoracic and respirable. These replace the old 'total' and respirable aerosol scenario. To determine which aerosol types might best be placed into a given category, it is required to consider the types of health effect with which they might be associated. Some biologically-active particles (e.g., bacteria, fungi, allergens) may, if they deposit in the extrathoracic airways of the head, lead to inflammation of sensitive membranes in that region, such as symptoms of 'hay-fever' (e.g., rhinitis). Other types of insoluble and persistent particle (e.g., radioactive material, wood dust) depositing in some parts of the same region may lead to more serious local conditions, such as ulceration or nasal cancer. For the health-related measurement of all such aerosols, it is appropriate ideally to sample according to a criterion based on the extrathoracic fraction. But it is argued that, for practical purposes in relation to such health effects, the inhalable convention is the most appropriate of the three, and is sufficient. The next class of aerosols is for particles which may lead to adverse health effects after deposition in the trachobronchial region of the lung. In this category are substances which may provoke local responses leading to such effects as bronchoconstriction, chronic bronchitis, etc. For the health-related measurement of all such aerosols, it is appropriate to think in terms of sampling according to a tracheobronchial criterion. But it is generally agreed, again for practical purposes, that the thoracic convention is appropriate and sufficient. The third class of aerosols is for particles which deposit in the alveolar 226

Standards for health-related aerosol measurement and control

region of the lung. Here there is a range of local effects, including pneumoconiosis (e.g., silicosis, asbestosis, etc.), emphysema, alveolitis and pulmonary carcinoma. In relation to these, the respirable convention represents the most appropriate sampling criterion. Finally, as a general rule for aerosol substances that are suspected or confirmed carcinogens, or are soluble and known to be associated with systemic effects (where toxic material can enter the blood after deposition in any part of the respiratory tract and be transported to other organs), standards should be specified in terms of the inhalable convention. For the latter, it is interesting to consider, however, that the rate of transfer to the blood may be different for different parts of the respiratory tract. Therefore, if it is desired to use the results of sampling to assess risk based on detailed dosimetric considerations, it might be more appropriate to make measurements of the inhalable, thoracic and respirable fractions simultaneously. The health effects associated with exposure to aerosols are varied and complex, and detailed description lies far beyond the scope of this book. Many alternative texts are available (e.g., Rom, 1992).

Strategies for exposure assessment

Health-related aerosol measurement needs to be carried out with regard not only to such particle size-selective criteria but also to the kinetics of the processes by which the particles can cause harm after inhalation and the variability of exposure. Rappaport (1991) has recently addressed these factors in relation to the general problem of occupational exposure to airborne toxic substances. In Chapter 7, the nature of the problem is illustrated for aerosols using a simple dosimetric model, where it is seen that the aim of measurement is to build up a picture of the exposure histories of workers in given occupational groups over their working lives, from which the risk to their health may be assessed. From such considerations, it is clear that, for aerosols, we should be concerned not only with choices about the particle size-selectivity of measuring instruments but also with the duration and frequency of sampling. For aerosol types where the dispersal of inhaled material around the body and the subsequent biological effects of exposure for short periods at high concentration may be severe and rapid, sampling should be of correspondingly short duration and at relatively frequent intervals. On the other hand, for most of the aerosols encountered in workplaces, such processes are relatively long term so that the effects of short-term high exposures are heavily damped out by the body's defence mechanisms. For such aerosols, a sampling strategy based on 8-hour T W A measurements is appropriate. However, the frequency of sampling should be sufficient that,

227

Aerosol science for industrial hygienists 99.99 0 ~

r

o 95-

o

0

o

80-

69 8~ 1 7 6 1 7 6

50205--

o

o

oo o

oo

r

1--

-0.1

I

0

100

E x p o s u r e ( m g Ni m "3 o f air)

Example of a log-normal occupational aerosol exposure situation, based on individual data records obtained for workers in a nickel smelter. Note the excellent linearity of the data as plotted on the log-probability axes shown.

Figure 8.8.

when allowance is made for the variability b e t w e e n samples, a reliable measure of exposure is obtained. The question of exposure variability touches on an i m p o r t a n t p r o p e r t y of workplace exposure (to any airborne c o n t a m i n a n t ) . In a given plant, w o r k e r exposures within a given working area may be defined as ' h o m o g e n e o u s ' if it can be assumed that the space is contiguous, the a t m o s p h e r e is well-mixed and the workers are all engaged in similar tasks. If that is the case, then it should be the case that the exposures of the workers in that area ~ if many m e a s u r e m e n t s are taken, including repeat m e a s u r e m e n t s for the s a m e workers ~ will be log-normally distributed. Such a distribution is s o m e t i m e s referred to as 'monomorphic'. This is illustrated in Figure 8.8 for exposures of nickel smelter workers to nickel-containing aerosols. H e r e , the a p p e a r a n c e of the straight-line tendency of the data when plotted as a cumulative distribution on log-probability axes is an indication of the log-normality of the data. The geometric standard deviation (Crg) for these data of a b o u t 2.8 is typical of workplace exposures m o r e generally. The essential m e a n i n g to this is that a single m e a s u r e m e n t of exposure gives only limited information. It follows that any strategy for exposure assessment of a workforce, using sampling criteria and technical i n s t r u m e n t a t i o n like those described in this book, must be designed statistically to reduce the uncertainty in the quantities sought to an acceptable level. In any given sampling exercise, there remains the further question a b o u t how best to reflect the true exposures of individual workers (or of groups of workers); either:

228

Standards for health-related aerosol measurement and control static (or area, or fixed-position) measurement, where the chosen instrument is located in the workplace atmosphere and provides a measurement of aerosol concentration which is, it is hoped, relevant to the workforce as a whole; or personal measurement with the chosen instrument mounted on the body of the exposed subject and moving around with him (or her) at all times.

When choosing one or other of these alternatives, some important considerations need to be taken into account. For a few workplaces (e.g., some working groups in longwall mining, as reported by Hadden et al., 1977), it has been shown that a reasonably good correlation may be obtained between suitably-placed static instruments and personal samplers. More generally, however, static samplers have been found to perform less well, usually tending to give aerosol concentrations that are consistently low compared to those obtained using personal samplers. One advantage with static samplers is that a relatively small number of instruments may be used to survey a whole workforce. If this can be shown to provide valid and representative results, it is a simple and cost-effective exercise. Furthermore, the high flow rates that are acceptable for static samplers mean that, even at very low aerosol concentrations, a relatively large sample mass can be collected in a short sampling period. The use of personal samplers is more labour intensive, requiring more instruments hence greater effort in setting them up and in recovering and analysing the samples afterwards. Furthermore, it involves the direct cooperation of the workers themselves. Also, for such samplers, it is inevitable that the capacities of the pumps used will be limited by their portability. So flow rates will usually be low (rarely greater than 4 1 min-1). However, personal aerosol sampling is widely accepted by professional industrial hygienists as the only reliable means of assessing the true aerosol exposures of individual workers. So this is by far the most common mode of aerosol measurement in workplaces. Finally, in discussing personal aerosol sampling, the 'breathing zone' concept is frequently invoked as a practical guide as to where the sampler should be placed on the worker's body. For example, some published reference methods recommend that the sampler should be located on the body not more than 30 cm from the nose-mouth region. However, such guidance can lead to a false sense of security since it does not follow that, just because a sampler is mounted in this manner, it will necessarily collect the same amount of aerosol in the fraction of interest as is inhaled by its wearer. Studies of the physics and aerodynamics of the aspiration process have taught us that simple proximity alone does not ensure a representative sample.

229

Aerosol science for industrial hygienists Limit values

The terminology for limit values varies between countries, or even between standards bodies or regulating agencies in the same country. Perhaps the system used by A C G I H is the most widely recognised and used. Here, A C G I H defines a threshold limit value-time weighted average ( T L V - T W A ) as the time-weighted average concentration of exposure for a normal 8-hour working day and a 40-hour working week. Below this almost all workers may be repeatedly exposed, day-after-day, without adverse effect. Excursions above the T L V - T W A are permitted provided that they are compensated by corresponding excursions below during specified periods (dependent on the substance in question). In the United States, the equivalent limit value assigned by the Occupational Safety and Health Administration ( O S H A ) is defined as a permissible exposure limit (PEL). N I O S H defines it as a recommended exposure limit (REL). In Britain, the equivalent terminology used by the Health and Safety Executive is the occupational exposure standard ( 0 ES). A C G I H also applies a threshold limit value-short-term exposure limit ( T L V - S T E L ) to some substances and this is intended to supplement (and not replace) the T L V - T W A . It is applied when toxic effects arising from short-term exposures are likely. It is defined as a 15-minute T W A exposure level which should not be exceeded at any time during the working day, even if 8-hour T W A exposures are below the T L V - T W A value. In Britain, the HSE also defines O E S - S T E L s , but here the reference period is 10 minutes. The A C G I H applies a threshold limit value-ceiling (TLV-C) to some substances. This is an exposure level that should not be exceeded during any part of overall exposure. This implies that instantaneous monitoring is being carried out. However, if this is not feasible, it is acceptable to make a time-weighted average measurement based on a reference period of up to 15 minutes. The British Health and Safety Executive also defines a related maximum exposure limit (MEL) for some substances which are carcinogenic or for which no threshold effect can be identified. This is a timeweighted average which should not be exceeded under any circumstances. The reference period for time-averaging may be 8 hours or 10 minutes or both in the case of a particularly hazardous substance (e.g., cadmium oxide fume). The setting of numerical values for such limits has traditionally been based on considerations of epidemiological and toxicological information about given types of aerosol, as well as on available data about past and current worker exposures in relevant industries. The A C G I H list of TLVs is perhaps the most extensive list of scientifically-based limits, and this list is updated as an when additional information about a compound becomes available (see A C G I H 1993-1994 for a recent version). As a result it is highly influential, not only in the United States but also in other countries throughout the

230

Standards for health-related aerosol measurement and control

world. In that list, limits are suggested for a wide range of substances which can appear as aerosols. But, as stated earlier, the list currently recommends limits for 'total' and respirable aerosol only. A rationale is proposed in the 1985 A C G I H report for the setting of new aerosol limits based on the new particle size-selective criteria. This is outlined in Figure 8.9. The process first involves the identification of substances constituting a potential risk to health, together with the possible effects at 'target' tissues, either in the respiratory tract itself or elsewhere in the body. The next stage is to examine whether the substance appears in the workplace atmosphere as an aerosol and, if so, to assess its particle aerodynamic size distribution. It is this which determines the mass fractions that are first inhaled and are then deposited in the thoracic and alveolar regions, respectively. Having established the mass proportions arriving at each region, the third part of the process is to assess the kinetics of the health-related dose (of 'harmfulness') to target tissue. Taken together with other information (e.g., from epidemiology, toxicology, clinical observations, pulmonary physiology, biological monitoring, etc.), this

rata l

Experimental toxicology] Epidemioiogy | Clinical observations | Pulmonary physiology |

Physical and chemical

ofpa-i , s i

i'No-a,a

- ,,,,

Industrial/environmental hygiene studies

I

I

.!

]

1

I Thoracic I

Respirable

Previous TLV documentation

Si,esofar

in

respiratory tract o r

! aoadairwa, 1 other

target organs and

clinical disease~

-.

~ ~ J

"<'.--IN

[

-2 .....

.o_.,-

I IPM-~. ,~ I TLV I V

I (~ritical dose to I ! target organs , Data review and 12

ITrach~

[[Yes

"~

r No

I X'voo,ar [

deposition

_

__

I T P M - I ~ - - ~ . ,~ I TLV I V

Yes

I

I deposition i NO I RPM-I~--~..~ I TLV I

"y"

lYe |

determination of critical IZ mass concentrations for [particulate fractions

e

I

* Desposition in or transfer to target organ?

Panic size selective

TLVs

Figure 8.9. Rationale proposed by ACGIH (1985) for the setting of new aerosol TLVs (where IPM is inhalable particulate mass, TPM is thoracic particulate mass, and RPM is respirable particulate mass) (from American Conference of Governmental Industrial Hygienists Inc., Particle Size-Selective Sampling in the Workplace: Report of the A CGIH Technical Committee on Air Sampling Procedures, reprinted with permission).

231

Aerosol science for industrial hygienists semi-quantitative rationale aims to generate new particle size-selective TLVs for workplace aerosols which should represent a consistent improvement over the old ones. At this stage, however, progress towards implementation of such a rationale is in its infancy, and considerable work is still required to bring such a scheme to full fruition. It is confidently expected that, in the future, such a revised list of limit values based on the new particle size-selective sampling criteria will be adopted. When it does, appropriate sampling instrumentation will be required. The question of sampling instrumentation is discussed in Chapter 9. But it is fair to say here that, at the time of writing, there is not a wide range of commercially available devices which demonstrably meet the new criteria. However, a small number of instruments have emerged which appear to match the performance requirements of those criteria quite well. It is of interest to note that, in the British HSE list of EOSs (see HSE 1993 for the latest version), the old 'total' aerosol concept has already been replaced by the inhalability criterion and suitable sampling instrumentation clearly identified (HSE, 1989). In the meantime, it is of interest to compare current aerosol limit values as they appear in the American TLV and the British OES, lists respectively. Table 8.2 contains a representative, non-exhaustive selection, the main bulk of which shows limit values for aerosols which are non-fibrous. The table shows that, in both the USA and Britain, limit values are specified in terms of either a coarse ('total' or inhalable) or a fine (respirable) fraction or sometimes both. For the actual numbers there are many similarities, not surprisingly since most of them will have been based on the same documentation. At the bottom of the table, limit values are also given for fibrous asbestos dusts, based on the size selection criteria for fibre diameter, length and aspect ratio, as described above. It is noted that a limit value of 10 mg m -3 is a frequent entry for both the 'total' and the inhalable categories. This illustrates an interesting problem since, although such numbers were derived from the same source, they are now subject to different measurement criteria. Recent studies in a range of industries (reviewed by Vincent, 1995) have been conducted to compare individual worker aerosol exposures using two different personal sampling instruments side-by-side, one of the type traditionally used for 'total' aerosol (e.g., see Figure 9.24) and the other a new sampler with performance conforming to the inhalability criterion (see Figure 9.26). The results show that the inhalable aerosol sampler consistently collects significantly more than the 'total' aerosol sampler. Such a difference is strongly dependent on the particle size distributions of the aerosol studied, and so may be expected to vary greatly from one workplace situation to another. But the observed trends are (at least) qualitatively consistent with what can be deduced from knowledge of the physics of the sampling process for the samplers in question. So it is likely that the broad trend will be the same elsewhere. This in turn

232

Standards for health-re&ted aerosol measurement and control Table 8.2. C o m p a r i s o n b e t w e e n limit values (8-hour T W A ) for some typical workplace aerosols, all figures for non-fibrous dusts given in [mg m -3] and for fibrous asbestos dusts in [fibres cm -3] ( A C G I H 1993-1994; H S E , 1993). Notes: 1. Figures in parentheses are under review. 2. For the H S E - O E S s , those which are in fact M E L s are m a r k e d as such. .

Substance

ACGIH-TLV 'Total'

Aluminum (as AI) Metal Welding fumes Aluminum compounds Pyro powders (AI) Oxide (AI) Ammonium chloride Fume Barium compounds Soluble (Ba) Barium sulphate Beryllium & compounds (Be) Boron oxide Cadmium & compounds (Cd) Cadmium sulphide pigments (Cd) Calcium carbonate Calcium oxide Calcium silicate Calcium sulphate Caprolactam dust Carbon black Chromium & compounds (Cr) Chromium compounds VI (as Cr) Coal dust Cobalt & compounds

(Co) Copper Dusts and mists Fume Cotton dust Emery Fibre glass dust Glycerin mist Grain dust Graphite

HSE-OES

Respirable

Inhalable

Respirable

10

5

5 10

10

5

10

10

10 5

0.5

0.5 10

(0.002) 10

0.002 10 0.01

0.002

0.05 (MEL)

10 2 10

10 2 10 10 1 3.5

0.04 (MEL) 5

1 3.5

(0.5)

0.5

(0.05) 2

0.05 (MEL)

(0.05)

(0.1)

1 0.2

1 0.2 0.5 10

0.2* 10 10 10 4 2

10 10 (MEL) 10

5

*American limit value for cotton dust based on an intermediate ('thoracic') particle size fraction.

233

A e r o s o l science f o r industrial hygienists T a b l e 8.2.

Cont.

Notes: 1. Figures in parentheses are under review. 2. For the HSE-OESs, those which are in fact M E L s are marked as such. Substance

ACGIH-TLV 'Total'

Indium & compounds (In) Iron Oxide fume (Fe) Salts (Fe) Kaolin Lead Dust (Pb) Fume (Pb) Lead & compounds (Pb) Lead chromate (as Pb) (as Cr) Magnesite Magnesium Oxide fume (Mg) Manganese & compounds Dust (Mn) Fume (Mn) Mica Mineral wool fibre Molybdenum/compounds Soluble (Mo) Insoluble (Mo) Nickel & compounds Metal Soluble (Ni) Insoluble (Ni) Paraquat Platinum & compounds Metal Soluble salts (Pt) Portland cement Potassium hydroxide Rhodium & compounds Metal Soluble (Rh) Insoluble (Rh) Rouge Silica (amorphous) Silica (fused) Silica (crystalline) Cristobalite Quartz Tridymite

HSE-OES

Respirable

Inhalable

Respirable

0.1

0.1

5 1

(0.15) 0.15 0.15 0.05 0.012 10

10

5

10

10

5

(5) (1)

5 1

3 10

10 5 (MEL)

5 10

10

(1) (0.1) (1) 0.5

5

0.5 (MEL) 0.1 (MEL) 0.5 (MEL) 0.1

0.1

1 0.002 10 2

5 0.002 10

1 0.01 1 10 10

0.1 0.001 0.1 6 6

234

1

0.01

3 3 0.1

0.05 0.1 0.05

0.4 (MEL) 0.4 (MEL) 0.4 (MEL)

Standards f o r health-related aerosol m e a s u r e m e n t a n d control T a b l e 8.2.

Cont.

Notes: 1. Figures in parentheses are under review. 2. For the HSE-OESs, those which are in fact M E L s are marked as such. Substance

ACGIH-TLV 'Total'

Silicon Silicon carbide Silver & compounds Metal Soluble (Ag) Starch Sucrose Sulphuric acid mist Talc Tantalum & compounds Metal Oxide dusts (Ta) Tin & compounds Metal Oxide (Sn) Inorganic (Sn) Organic (Sn) Titanium dioxide Tungsten & compounds Soluble (W) Insoluble (W) Uranium compounds (U) Vegetable oil mists Welding fume (NOC) Wood dust Hard wood Soft wood Zinc chloride fume Zinc oxide fume Zinc oxide dust

HSE-OES

Respirable

Inhalable

10 10

10 10

0.1 0.01 10 10 1

(0.1) 0.01 10 10

Respirable 5 5

1

2

10

1

5 5 2 2 2 0.1 10

2 2 2 0.1 10 1 5

0.2

0.2 10 5

5 (MEL) (51

1 5 1 5 10

Asbestos Amosite Crocidolite Chrysotile Other forms

1

5

(0.5) (0.2) (2) (2)

235

see see 0.5 0.5

Other forms Other forms (4-hr TWA) (4-hr TWA)

Aerosol science for industrial hygienists carries the i m p l i c a t i o n t h a t it m a y be s o m e w h a t p r o b l e m a t i c a l ~ for s o m e s u b s t a n c e s ~ to assign the s a m e n u m e r i c a l limit v a l u e in the A m e r i c a n a n d British lists. F o r e x a m p l e , a l t h o u g h this m i g h t be a c c e p t a b l e for the relatively i n n o c u o u s ' n u i s a n c e ' a e r o s o l s ( P N O C ) , such d i f f e r e n c e s m i g h t h a v e m o r e serious h e a l t h - r e l a t e d i m p l i c a t i o n s for m o r e h a r m f u l s u b s t a n c e s . F o r c o n s i s t e n c y , it c o u l d be a r g u e d t h a t e i t h e r the British O E S s h o u l d be l o w e r e d , or the A m e r i c a n T L V s h o u l d be raised. B u t this m i g h t raise s o m e u n c o m f o r t a b l e policy issues.

REFERENCES American Conference of Governmental Industrial Hygienists (ACGIH) (1968). Threshold limit values of airborne contaminants. ACGIH, Cincinnati, OH. American Conference of Governmental Industrial Hygienists (ACGIH) (1985). Particle size-selective sampling in the workplace. Report of the Technical Committee on Air Sampling Procedures, ACGIH, Cincinnati, OH. American Conference of Governmental Industrial Hygienists (ACGIH) (1993-1994). Threshold limit values for chemical substances and physical agents and biological exposure indices, ACGIH, Cincinnati, OH, pp. 4245. Armbruster, L. and Breuer, H. (1982). Investigations into defining inhalable dust. In: Inhaled Particles V (Ed. W.H. Walton). Pergamon Press, Oxford, pp. 21-32. Asbestos International Association (AIA) (1979). Airborne asbestos fibre concentrations at workplaces by light microscopy (membrane filter method). AIA Health and Safety Publication RTM1, AIA, Paris. Bartley, D.L. and Doemeny, L.J. (1986). Critique of the 1985 report on particle size selective sampling in the workplace. American Industrial Hygiene Association Journal, 47, 443--447. British Medical Research Council (BMRC) (1952). Recommendations of the BMRC panels relating to selective sampling. From the Minutes of a joint meeting of Panels 1, 2 and 3 held on March 4th. Carter, J., Taylor, D. and Baron, P. (1989). Asbestos fibres, method 7402, revision No. 1: 5/15/89. In: NIOSH Manual of Analy6cal Methods, DHHS/NIOSH, Cincinnait, OH. Comit6 Europ6en de Normalisation (CEN) (1992). Workplace atmospheres: size fraction definitions for measurement of airborne particles in the workplace, CEN Standard EN 481, Brussels. Environmental Protection Agency (EPA) (1984). National ambient air quality standard, proposed rule. Federal Register, 49, 10408-10462. Hadden, G.G., Jones, C.O. and Thorpe, H.L. (1977). A comparative assessment of dust surveillance procedures, including the use of personal and fixed-position sampling instruments. Institute of Occupational Medicine (Edinburgh, Scotland, U.K.), Technical Report TM/77/15. Health and Safety Executive (HSE) (1989). General methods for the gravimetric determination of respirable and total inhalable dust. HSE, MDHS14 (revised), HMSO, London. Health and Safety Executive (HSE) (1993). Occupational exposure limits 1993. HSE, EH40/93, HMSO, London. International Standards Organisation (ISO) (1983). Air q u a l i t y - particle size fraction definitions for health-related sampling. Technical Report ISO/TR/7708-1983 (E), ISO, Geneva.

236

Standards for health-related aerosol measurement and control International Standards Organisation (ISO) (1992). Air q u a l i t y - particle size fraction definitions for health-related sampling, ISO CD7708, ISO, Geneva. Liden, G. (1994). Performance parameters for assessing the acceptability of aerosol sampling equipment. The Analyst, 119, 27-33. Liden, G. and Kenny, L.C. (1991). Comparison of measured respirable dust sampler penetration curves with sampling conventions. Annals of Occupational Hygiene, 35, 485-504. Liden, G. and Kenny, L.C. (1992). The performance of respirable dust samplers: sampler bias, precision and inaccuracy. Annals of Occupational Hygiene, 36, 1-22. Liden, G and Kenny, L. (1993). Optimisation of the performance of existing respirable dust samplers. Applied Occupational and Environmental Hygiene, 8, 386-391. Lippmann, M. (1988). Asbestos exposure indices. Environmental Research, 46, 86-106. National Institute for Occupational Safety and Health (NIOSH). (1975). Criteria for a recommended standard - - occupational exposure to cotton dust. DHEW (NIOSH) Publication No. 75-118, USGO, Washington, DC. National Institute for Occupational Safety and Health (NIOSH). (1979). USPHS/NIOSH membrane filter method for evaluating airborne asbestos fibers. Criteria for a recommended s t a n d a r d - occupational exposure to cotton dust. NIOSH Technical Report. Ogden, T.L. and Birkett, J.L. (1977). The human head as a dust sampler. In: Inhaled Particles IV (Ed. W.H. Walton). Pergamon Press, Oxford, pp. 93-105. Ogden, T.L., Birkett, J.L. and Gibson, H. (1977). Improvements to dust measurement techniques. IOM Report No. TM/77/ll, Institute of Occupational Medicine, Edinburgh, U.K. Pott, F. (1978). Some aspects of the dosimetry of the carcinogenic potency of asbestos and other fibrous dusts. Staub-Reinhalt. Luft, 38, 486. Rappaport, S.M. (1991). Assessment of long-term exposures to toxic substances in air. Annals of Occupational Hygiene, 35, 61-121. Rom, W.M. (Ed.) (1992). Environmental and Occupational Medicine (2nd Edn.). Little, Brown and Co. Soderholm, S.C. (1989). Proposed international conventions for particle size-selective sampling. Annals of Occupational Hygiene, 33, 301-321. Timbrell, V.A. and 8 other authors (1988). Relationships between retained amphibole fibres and fibrosis in human lung tissue specimens. In: Inhaled Particles VI (Eds. J. Dodgson, R.I. McCallum, M.R. Bailey and D. Fisher). Pergamon Press, Oxford. Vincent, J.H. and Armbruster, L. (1981). On the quantitative inhalability of airborne dust. Annals of Occupational Hygiene, 24, 245-248. Vincent, J.H. (1995). Progress towards implementation of new aerosol industrial hygiene standards: with special reference to the aluminium industry. Science of the Total Environment, 163, 3-10. Vincent, J.H. and Mark, D. (1982). Application of blunt sampler theory to the definition and measurement of inhalable dust. In: Inhaled Particles V (Ed. W.H. Walton), Pergamon Press, Oxford, pp. 3-19. Vincent, J.H., Mark, D., Miller, B.G., Armbruster, L. and Ogden, T.L. (1990). Aerosol inhalability at higher windspeeds. Journal of Aerosol Science, 21,577-586. Walton, W.H. (1982). The nature, hazards, and assessment of occupational exposure to airborne asbestos dust: a review. Annals of Occupational Hygiene, 25, 115-247.

237

CHAPTER 9

Aerosol sampling in workplaces 9.1 I N T R O D U C T I O N Aerosol measurement in workplaces is carried out by industrial hygienists for a variety of different ~ but related ~ reasons. The first is for the assessment of the workplace atmosphere so that aerosol concentrations there can be compared against the health-related exposure limits that have been laid down, thus providing a basis for decisions on control action. The second is for the provision of valid measurements of worker exposure for the purposes of epidemiology. The third is for the measurement of aerosol concentrations in exhaust ventilation systems so that the effectiveness of emission control equipment can be assessed. The technical aspects associated with aerosol measurement include: the aspiration of particles into sampling devices; the aerodynamic selection and subsequent assessment (e.g., mass determination) of desired fractions or subfractions; the direct-reading detection of aspirated (and selected) aerosol; and remote, non-aspirating direct-reading aerosol monitoring. In this chapter, the current state of aerosol science underpinning the first two of these is reviewed. This is followed by a discussion of some recent technical developments which have arisen, in particular in response to the recent emergence of the health-related particle size-selective criteria (as described in Chapter 8). Direct-reading aerosol monitoring will be discussed in Chapter 10. 9.2 S A M P L I N G BY A S P I R A T I O N

Background The first important aspect of the performance of an aerosol sampler is the efficiency with which particles are transferred from the air outside the sampler

238

Aerosol sampling in workplaces and into the sampler through its one or more entry orifices. Obviously a sampler with zero aspiration efficiency for aerosol particles in the size range of interest would be totally useless as a practical device. In practice, the aim is to find a sampling device whose efficiency matches some pre-determined selection criterion. For example, for sampling in relation to health effects, we might wish to simulate the entry of particles into the human respiratory tract. In such cases, sampling should be based on the inhalability criterion (as described in Chapter 8). In cases where we wish to sample with respect to particles depositing at locations inside the respiratory tract, it should be carried out based on either the thoracic or respirable criteria. However, there are other industrial hygiene situations where it is desired to sample true total aerosol; that is, particles of all sizes with 100% efficiency. This is frequently the case, for example, in the measurement of aerosol concentration in ducts (e.g., of ventilation and exhaust systems) for the purpose of assessing the performances of aerosol separation equipment. Because of the nature of aerosols and their potential health-related effects, it is ~ as has been argued earlier ~ scientifically unsound to use an aerosol sampler without first having identified, and then applied, a criterion relevant to the reason for the measurement. Therefore, it is important to know whether the chosen sampler matches that criterion; and if not, to know the extent of any biases in the measurement introduced by deviations of performance from the ideal. For these reasons, quantitative sampler performance is central to most aerosol measurement methods and devices. In Chapter 4 the aerosol mechanical framework was established upon which to base the development of physical models for determining the sampling characteristics of aerosol sampling devices. Here, this will be extended into some simple quantitative examples, leading up to a discussion of some samplers of practical relevance. In much of what will be described, experimental data for the performances of various aerosol samplers will be presented. These will be taken from a variety of sources involving different experimental methods. These methods have been described in detail elsewhere and are reviewed by Vincent (1989). They will not be discussed at length in this book.

Identification of aerosol sampler performance indices Discussion of sampler performance can begin by referring to Figure 9.1 which represents the air flow near a single-orifice sampling device of arbitrary shape placed at arbitrary orientation with respect to the external wind. This enables the discussion at this point to be completely general, without regard to any particular sampler configuration. It is noted that the case of moving air is, perhaps, the most relevant to industrial hygiene since this is indeed the case in most workplace situations. It encompasses fixed-position samplers

239

Aerosol science for industrial hygienists

in workplace air which is moving even as slowly as 0.2 m s-1 (quite typical). It also includes the relative motion between a personal sampler mounted on the body of a worker and the surrounding air when the worker moves around during his/her work. In Figure 9.1, the sampler is thus seen as an entry located on the surface of a 'bluff' (or 'blunt') flow obstacle and an inner transition section leading to a 'sensing' region, shown here (as in most samplers) in the form of a filter on which the sampled particles are collected. The flow pattern indicated represents the motion of air near the sampler. In particular the limiting (or dividing) streamline is the invisible surface in the flow which separates the sampled air from that which is not sampled. Thus, the shaded area inside the limiting streamline (or rather, for three-dimensional flow, the limiting streamsurface) represents the volume of air which is aspirated. All the air outside that surface passes by outside the sampler. So too would any airborne particles small enough to be considered inertialess. This clear distinction between the sampled and non-sampled flows is a direct consequence of the fact that there can be no net flow of air across streamlines (see Chapter 2). For given particle size and flow conditions, N o is the number of particles per unit time flowing in the undisturbed sampled air volume contained within the limiting streamsurface upstream of the sampler. It represents the particle concentration in the undisturbed upstream air. In addition, N~ is the corresponding number of particles passing directly (whilst remaining airborne at all times) through the plane of the sampling orifice, and N~ is the number passing through the orifice plane after rebounding from the external walls of the sampler. Once particles have crossed the orifice plane and so entered the body of the sampler, some may be deposited on the internal walls of the transition section and so fail to reach the filter. The number reaching the filter is therefore N F = P ( N s + Nr)

(9.1)

where P is the fractional penetration of particles through the transition region. In some sampling devices this is regarded as an unwanted loss of aerosol. Later, however, we shall see how this part of sampler performance can be employed effectively in simulating what happens to inhaled particles after they have entered the human respiratory tract. From the above definitions, a number of performance indices of sampler can be identified. The first is the aspiration efficiency as already described in Chapter 4. Reiterating Equation (4.44), this is given in terms of the quantities defined above as

Us A =

(9.2)

N0 240

Aerosol sampling in workplaces This is the purely aerodynamic part of sampler performance. Next the entry (or apparent aspiration) efficiency is (N s + Nr) (9.3)

Aap p =

No which allows for the fact that, under some conditions, there may be a tendency towards oversampling associated with particles entering the sampler after rebounding from the external wall of the sampler. Such contributions to sampler performance are regarded as unwanted interferences and are frequently not easy to diagnose in practical sampling situations. Finally, the overall sampling efficiency (or sampling effectiveness) is

NF A o v e r a l I --

P(Ns +

Nr)

(9.4)

=

No

No

where the loss of particles inside the sampler before they can reach the filter is seen as a tendency towards undersampling. Discussion of sampler performance needs to bear these factors in mind. In particular, departures from ideal performance ~ as embodied in aspiration efficiency (A) will depend on the balance in individual cases between the oversampling associated with particle rebound from external surfaces and the undersampling associated with particle losses inside the sampler. This balance will depend on many factors, including flow conditions and particle size and type, and so is usually difficult to quantify.

9.3 A S P I R A T I O N E F F I C I E N C Y OF T H I N - W A L L E D S A M P L I N G P R O B E S IN M O V I N G A I R

Qualitative physical picture for a thin-walled probe facing the wind The thin-walled sampling probe facing a moving airstream represents the simplest limiting case of the generalised picture portrayed in Figure 9.1. This is a sampler whose performance characteristics have been the subject of a great deal of research, stimulated mainly by the need to measure true total aerosol concentration in stacks and ducts as a basis for regulating emissions. In such situations, statement of the sampling problem is straightforward because the sampling criterion is clearly specified and the airflow velocity and direction are usually well defined and can be measured independently. The nature of the airflow in the vicinity of a cylindrical thin-walled

241

Aerosol science for industrial hygienists

Figure 9.1. Air and particle transport near a blunt aerosol sampler of arbitrary shape and orientation, on which to base considerations of sampler performance (from Vincent, J.H., Aerosol Sampling: Science and Practice, Copyright 1989, adapted by permission of John Wiley and Sons Limited.) probe facing the wind is shown in Figure 9.2 for three contrasting sets of aerodynamic conditions. The picture here is somewhat simplified in that the limiting (or dividing) streamsurface is shown to terminate at the leading edge of the probe. In reality, this will not be case as the stagnation points where the limiting streamsurface arrives at the sampler body will be located at different places on the surface of the body ~ inside or outside ~ depending on the flow conditions. Nevertheless, the simple picture shown is adequate for present purposes. Figure 9.2a shows what happens when the mean air velocity over the plane of the sampler entry (Us) is less than the freestream air velocity upwind of the sampler (U). Here, from the discussion about streamlines in Chapter 4, it is clear that the pattern of the sampled air, as contained within the limiting streamsurface, must be divergent. Aerosol particles encounter this flow distortion and, due to their inertia, those approaching from outside the sampled air volume may 'impact' into it, and hence onto the plane of the entry orifice. In this way, the aerosol which enters the orifice is enriched with particles in relation to the aerosol upstream of the probe. In Figure 9.2b, where Us is greater than U, the converse happens. Here the sampled flow is convergent so that particles in the aerosol can 'impact' out of the sampled air volume. Now the result is for there to be a reduction in aerosol concentration in the sampled air. In Figure 9.2c, we have U s = U, in which case the sampled flow is perfectly matched to the wind so that the flow is

242

Aerosol sampling in workplaces (a)

'

.

Limiting streamline

(b)

(C)

-

v

Qwp--

OIP-w

Figure 9.2. General form of the streamline pattern and superimposed particle

trajectories for a thin-walled sampling probe facing the wind: (a) sub-isokinetic; (b) super-isokinetic; and (c) isokinetic. (From Vincent, J.H., Aerosol Sampling: Science and Practice, Copyright 1989, reproduced by permission of John Wiley and Sons Limited.) undistorted. Here, there will be no gains or losses due to impaction, and so the aerosol concentration in the sampled air will the same as that upstream. The three conditions illustrated in Figure 9.2 may be described as subisokinetic, super-isokinetic and isokinetic, respectively. The latter is the ideal sampling condition for thin-walled probes. Such probes used in this way are frequently referred to as isokinetic samplers.

Simple impaction model

A simple, s e m i - q u a n t i t a t i v e m o d e l m a y be d e v e l o p e d to describe the aspiration efficiency for a thin-walled sampling probe facing the wind. For

243

Aerosol science for industrial hygienists the sampler shown schematically in Figure 9.3, aspiration efficiency, from Equations (4.49) and (9.2), is Cs

A =

(9.5) CO

where c s is the sampled aerosol concentration and c o is the aerosol concentration in the undisturbed air (i.e., the true concentration). In Figure 9.3, the flow is seen to be sub-isokinetic, but the following arguement could be applied just as readily for super-isokinetic flow. In addition to indicating the quantities c~, c 0, U~ and U already defined, the figure also identifies the nozzle diameter (~), its cross-sectional area (a~) and the projected area upstream containing the sampled flow (ao). The volume of air geometrically incident from far upstream onto the tube entrance per unit time is as U. The volume actually aspirated by flowing through the tube entrance ~ and, by continuity, through the upstream plane contained within the limiting streamsurface ~ is a~U~. Therefore, the volumetric flow rate of air from the original a~U which now diverges to pass by outside the tube is the difference, as(U-Us). For the flow system shown, it is clear that all particles originally contained within the sampled air volume (a~U~) will enter the tube. Therefore, c~ must be at least as large as c o. However, in addition, there will be some particles from the air outside the limiting streamsurface that will spill over by impaction. Since c~ must exceed c o by an amount equivalent to this spillover, the mass flow of particles into the tube may therefore be expressed as

C~asU~ = CoaoU + oLCoas(U-U~) Cross-sectional area as Sampled air ao

Co streamline

,,.,.,.,.............. ,

Aerosol concentration Cs

Figure 9.3. Schematic of a thin-walled sampling probe facing the wind, on which to base development of an impaction model for aspiration efficiency (from Vincent, J.H., A e r o s o l S a m p l i n g : S c i e n c e a n d P r a c t i c e , Copyright 1989, reproduced by permission of John Wiley and Sons Limited.)

244

(9.6)

Aerosol sampling in workplaces where r is a dimensionless impaction parameter. Since continuity requires that asUs = aoU (i.e., since there can be no net transfer of air across streamlines), Equation (9.6) reduces to Cs

A =

= 1 + er { R - l }

(9.7)

Co

where R = U/U~ is the sampling velocity ratio. The spill-over of particles is described by the second term on the right-hand side of this expression and cannot include particles which had not originally been incident on the tube entrance. Therefore r can take values only between 0 and 1, and is thus seen to be an impaction efficiency directly analogous to that referred to in Chapter 4 for describing the efficiency of inertial deposition onto a solid bluff body. For this reason, we may write - f {St)

(9.8)

where d 2

aeP 9 U

St =

(9.9) 18~

is now a Stokes' number for the sampler. Experimental investigations into the relationship between r and St have given rise to a number of suggested empirical relations. The best, based on extensive work carried out in Russia (Belyaev and Levin, 1974), is

1{

1

}

(9.10)

(1 + G St)} with the coefficient G given by 0.62 G = 2 +

) (9.11)

R Equation (9.7) is the basic working expression for describing the performances of thin-walled probes. It shows that, for the isokinetic condition given by R = 1, A is equal to unity (100%) for particles of all sizes. In addition, for large St, corresponding to a combination of large dae , large U and small ~, A ~ R. These trends are illustrated in Figure 9.4.

245

Aerosol science for industrial hygienists

dQ~ .

A

]

dae= 0

R = UIUs

1.5

R = UIUs increasing

R=UIUs=I

1.0

0.5

0

I

I

5

I0

I

15

Particle aerodynamic diameter dae (~m)

Figure 9.4. Illustration of the trends in aspiration efficiency for a thin-walled sampling probe facing the wind: (a) A as a function of velocity ratio R (from Vincent, J.H., Aerosol Sampling: Science and Practice, Copyright 1989, reproduced by permission of John Wiley and Sons Limited), and (b) A as a function of particle aerodynamic diameter, dae.

S u b s e q u e n t to the work of Belyaev and Levin, Lipatov et al. (1986) in U k r a i n e carried out very elegant experiments in a small wind tunnel in which they directly observed (though a telescope) the trajectories of particles in the flow approaching a thin-walled p r o b e facing the wind. These e x p e r i m e n t s were designed specifically to eliminate the ambiguities that could arise from sampler wall effects (including the p h e n o m e n o n of blow-off or r e b o u n d from

246

Aerosol sampling in workplaces

the external tube wall, as will be discussed later) and which are now thought to have afflicted many earlier experiments. The results confirmed that the Belyaev and Levin model embodied in Equations (9.7), (9.10) and (9.11) is an excellent predictor of aspiration efficiency for thin-walled sampling probes facing the wind. Nonetheless, the search for refinements has continued. Liu et al. (1989) developed an empirical expression for aspiration efficiency based on correlation with exact numerical calculations, and obtained

(R-l) A=I+

forR> 1 0.418

)

1+ St (9.12) (R-l) A=I+

forR < 1 0.506R1/2 ) 1+ St

for wide ranges of St and R. Rader and Marple (1988) took into account the effects of particle interception at the tip of the sampling probe, and obtained

A = 1 + (R-l)

{ (

1

1-

)}

(9.13)

1+3.77STO.883 Effects of orientation

The situation becomes more complicated when the thin-walled probe is placed at an angle with respect to the approaching aerosol (i.e., is 'yawed'). This was first studied scientifically by Durham and Lundgren (1980), and later by Davies and Subari (1982) and Vincent et al. (1986). The physical picture is shown schematically in Figure 9.5, where the tube-shaped probe is placed at an angle 0 to the forward-facing direction. Here, as for the forward-facing sub-isokinetic case, the flow again diverges as it approaches the tube entrance. But, for the same value of the sampling ratio R, the divergence is less than for the forward-facing direction because the tube entrance projects upstream a smaller cross-sectional area. Now, the body of the sampler no longer imposes an infinitesimally small obstruction, but appears 'blunter' as more of the body of the tube presents itself to the flow, the more so as the angle 0 increases. In addition to these complications, the sampled flow has to change direction in order to enter the sampling 247

Aerosol science for industrial hygienists Vena c o n t r a c t a

Limiting streamsurface Figure 9.5. Schematic on which to base development of an impaction model for a thin-walled sampling probe aligned at an angle to the wind (from Vincent, J.H., Aerosol Sampling: Science and Practice, Copyright 1989, reproduced by permission of John Wiley and Sons Limited).

orifice. Therefore, the sampled flow not only undergoes divergence (or convergence for super-isokinetic flow) but also 'turns'. The net result of all these considerations is a more complicated expression for aspiration efficiency. Thus A = 1 + e~'(Rcos 0 - 1)

(9.14)

where the modified impaction parameter is

~x'= 1 -

(9.15) {1 + G(O)St(cos 0 + 4R1/2sin1/20)}

The coefficient G is now described as a function of orientation, although G - 2.1 has been found to provide a fair working approximation (Vincent et al., 1986). Physically, Equation (9.15) embodies considerations of the change in time associated with the passage of a particle through the distorted flow region as the contribution due to turning increases as 0 increases. It is now seen that A ~ Rcos0 for large St. Some typical trends for aspiration efficiency obtained using this model are shown in Figure 9.6. For the probe placed facing the flow, this expression reduces ~ as it should to the form of Equation (9.7). When Rcos0 = 1, then A = 1 for all particle sizes. So this is, in effect, the isokinetic condition for the yawed probe. Finally, when the probe is placed at 90 ~ to the flow, it yields the interesting result that

248

Aerosol sampling in workplaces 1.0 0=0

0 increasing

I

0

I

5

10

I

15

Particle aerodynamic diameter d~ (Ixm) Figure 9.6. Typical trends in aspiration efficiency for a thin-walled sampling probe facing at various angles to the wind (fixed R, varying dae and 0).

A - f { S t R 1/2}

(9.16)

which, on closer inspection, reveals that, for fixed volumetric flow rate, A is independent of the probe dimensions (Stevens, 1986). This is also consistent with the form suggested by Davies and Subari (1982) and Tufto and Willeke (1982). Finally, it is important to note that the preceding discussion applies for yaw angles only up to (and including) 90 ~. For backwards-facing orientations, the picture becomes complicated by the wake properties of the airflow about the sampler body, possibly accompanied by turbulence. Under such conditions, and on the basis of present knowledge, it is difficult to predict sampler performance. Example 9.1. A thin-walled tube of diameter 1 cm and sampling flow rate 20 1 min -1 is used to sample particles of aerodynamic diameter 25 txm from an airstream moving at 2 ms -1. Calculate the aspiration efficiency. The sampling velocity is given by ( 2 0 x 10-3/60) [m 3 s-1] • 4

Us =

= 4.24 m s-1 3.142 • 10 -4 [m 2]

From Equation (9.9), Stokes number is (252x 10-12) [m 2] • 103[kg m -3] • 2[m s-1] St = 18 • (18• 10-6 ) [N s m -2] • 10-2[m] = 0.386

249

Aerosol science for industrial hygienists The velocity ratio is

R =

- 0.472 4.24

From Equation (9.11), the coefficient G to be used in the impaction efficiency equation is 0.62 G=

)

2+ 0.472

= 3.32 So, from Equation (9.10), impaction efficiency is

1

)

1 + (3.32 • 0.386) = 0.561 This is now used in Equation (9.7) to give the aspiration efficiency A = 1 + 0.561 ( 0 . 4 7 2 - 1) = 0.704 *

Aspiration efficiency is 0.704 (or 70%)

E x a m p l e 9.2. F o r the p r e v i o u s e x a m p l e , c a l c u l a t e t h e a s p i r a t i o n efficiency w h e n the t u b e is p l a c e d at 90 ~ to the flow. Note that cos 0 = 0 and sin 0 = 1 Here we need a new impaction efficiency, oL', as given by Equation (9,15). Usiiag G(0) = 2.1 (the recommended working approximation), we get

{

1

)

or' = 1 (1 + (2.1 x 0.386 x 4 x V/0.472)) = 0.690

250

Aerosol sampling in workplaces Using this in Equation (9.14), we get A = 1 + 0.690(0-1) = 0.31 *

Aspiration efficiency now falls to 0.31 (31%)

9.4 A S P I R A T I O N E F F I C I E N C Y OF BLUNT SAMPLERS As already stated, thin-walled sampling tubes represent the simplest case of the general aspiration problem outlined in Chapter 4. But in reality, true thin-walled probes do not exist since the walls must have finite thickness, no matter how thin. Therefore, there must always be some aerodynamic blockage, even for forwards-facing 'thin-walled' probes, and the overall nature of the flow will contain features associated with both the convergence of the sampled air to enter the sampling orifice as well as the divergence to pass around the solid body of the sampler. In practice, however, it is possible to design thin-walled probes with walls sufficiently thin that such effects are negligible. This notwithstanding, thin-walled sampling probes enjoy only very limited application in industrial hygiene. Indeed, in spite of the considerable research which has gone into understanding their performance, application is still confined largely to aerosols in stacks and ducts. For most types of sampler used in practical aerosol measurement (and including, of course, the human head itself), the 'blocked-off' area projected by the body of the device is very much greater than the area projected by the sampling orifice itself. Under such conditions, sampler 'bluntness' becomes a very important factor. Attempts have been made to model the performances of so-called 'blunt' samplers along the same lines as described above for the idealised thin-walled probe. The physical picture for a sampler of simple shape facing the wind (e.g., an axisymmetric disc with a central sampling orifice) is shown in Figure 9.7. It is based on the flow pattern for the same system obtained by flow visualisation shown in Chapter 2 (see Figure 2.6a). The main goal in developing a physical model for the aspiration efficiency of this system is to evaluate the expression A = A 1A2

(9.17)

in terms of the relevant system variables. Here, A~ and A 2 represent the 'efficiencies' with which aerosol particles are conveyed towards the sampling orifice through the two distinct characteristic parts of the flow (assigned the subscripts 1 and 2, respectively). Each of these parts may be described in turn by an expression of the type shown in Equation (9.7), but with modified

251

A e r o s o l science f o r industrial hygien&ts

Limiting streamline Figure 9.7. Schematic of a simple blunt sampler facing the wind, on which to base development of an impaction model for aspiration efficiency (from Vincent, J.H., Aerosol Sampling: Science and Practice, Copyright 1989, reproduced by permission of John Wiley and Sons Limited).

Stokes' n u m b e r and velocity ratio terms derived from the local characteristics of the flow. The result is a set of working e q u a t i o n s which, for the simple axisymmetric case of a disc-shaped sampler facing the wind, is (Vincent, 1987, 1989)

St

--

2 dae p,

R

=

U/Us

(9.18.2)

r

=

8/D

(9.18.3)

r

=

r2/R

(9.18.4)

S

=

B ~1/3 D for 8 < <

St 1

-

-

(9.18.1)

U/18Fa8

D

(9.18.5)

(9.18.6)

St y 02/3

e,1

=

1-

A1

=

l + c x 1(07~/3 - 1)

(9 . 18 . 8)

St 0 U3

(9.18.9)

St 2

-

{1/(1 + GaStl)}

(9.18.7)

cx2

=

1 - {1/(1 + GzSt2) }

(9.18.10)

a 2

=

1 + e~z{(r/0~3) 2 - 1}

(9.18.11)

A

=

A1A 2

(9.18.12)

w h e r e ~ and D are the d i a m e t e r s of the sampling orifice and the s a m p l e r body, respectively, and B is the a e r o d y n a m i c 'bluntness' of the s a m p l e r which

252

Aerosol sampling in workplaces

"5F "OI ot

I

I

I

I

I

I

1.5

I

1

I

1 100

I 1000

[

! .0

~i= 2 mm

100 nm I

0

I

4 0 / ~ ~ 6 0

I

I

t

0.1

!.0

10

i 100

I 1000

0

I

I

I

0.1

1.0

10

St

Figure 9.8. Some typical trends in aspiration efficiency calculated from blunt sampler theory for a blunt sampler facing the wind (from Vincent, J.H., Aerosol Sampling: Science and Practice, Copyright 1989, adapted by permission of John Wiley and Sons Limited). derives from its geometry. The new quantity S is the width of the region over the sampler surface contained within the stagnation ring formed by the limited streamsurface (and within which the flow is convergent towards the sampling orifice). Some typical trends calculated from this model are shown in Figure 9.8. It is seen that, even for the axisymmetric system as simple as that shown in Figure 9.7, behaviour is very complicated. For this case, quite good agreement is achieved with the small amount of available experimental data. A more complex version of this model incorporates orientations other than facing the wind (0 up to 90~ For this we may express aspiration efficiency in the general functional form

A-f

{

St, R,

,B,O D

)

(9.19)

Again, as for the thin-walled probe, A --~ R cos 0 for large St. A physical model, based on extending the one described in Equations (9.18), has proved

253

Aerosol science for industrial hygienists

only moderately successful when compared to experimental results obtained for the aspiration efficiencies of single-orifice blunt samplers of simple shape at orientations other than forwards-facing. Most of the aspiration efficiency models described above are semiempirical, firmly based on the physical transport of particles in flows whose broad features are described but with coefficients which are fitted with respect to experimental data. This type of approach may be broadly referred to as the 'impaction model'. Other approaches have been proposed based on more rigorous mathematical descriptions of the air flow field around the blunt sampler. The problem is not simple because the flow with the required boundary conditions appropriate to aerosol samplers is difficult to solve in the required detail using existing computers. The approaches that have been tried involve modelling aspiration efficiency on the basis of potential air flow (e.g., Ingham, 1981; Dunnett and Ingham, 1986; Erdal and Esmen, 1995), on viscous flow (e.g., Ingham and Hildyard, 1991; Chung and Dunn-Rankin, 1992), and on turbulent flow (e.g., Ingham and Wen, 1993). Dunnett and Ingham (1988) also developed a mathematical model to describe the flow near a blunt sampler (in particular to locate the positions of flow stagnation on the surface of the sampler) and then applied it directly into an 'impaction model' of the type described above. This, therefore, represented an attempt to link the mathematical and semi-empirical 'impaction model' approaches. It is an ultimate goal that mathematical studies like those cited will eventually

Figure 9.9. Picturesto illustrate some features of the airflow near a simple blunt sampler when placed with its sampling orifice at large angles with respect to the wind (from Vincent, J.H., Aerosol Sampling: Science and Practice, Copyright 1989, adapted by permission of John Wiley and Sons Limited).

254

Aerosol sampling in workplaces

lead to a rigorous universal model that can be run on a personal computer, capable of predicting the aspiration efficiency of any sampler configuration or of designing a sampler that will have pre-determined performance. Although the work so far looks promising, the goal still seems a long way off. Meanwhile, the 'impaction model' approach provides the best opportunity for achieving workable predictive models that will be useful in the short or medium term. One particular problem not addressed in any of the work cited above is how to deal with the air flow when the orientation of the sampler with respect to the wind is large (i.e., 0 greater than 90~ Here the flow becomes extremely complicated, hard to visualise, and even harder to predict. To illustrate the nature of the difficulties involved, Figure 9.9 suggests some of the features of the air flow expected for blunt samplers at such orientations. At present, no mathematical solutions have been obtained for such flows. Recently, Tsai and Vincent (1993) examined the limited amount of available experimental data for thin-walled samplers and blunt samplers for 0 = 90 ~ and 180 ~ and proposed extensions of the 'impaction model' to describe what happens under these conditions. Those extensions took empirical account of the effect of R and r on the shape of the flow and particle transport near the sampling orifice for large orientations, and yielded A9o --

[1 + 4 (2.21 St)

(R/r) 1/2] (9.20)

A18 0 =

[1 + 4 (4.5 St) +1/3 F-0.29] where R, r and + are as defined earlier. These gave quite good agreement with the available data. These equations were subsequently incorporated into a model for the aspiration efficiency of a blunt sampler when orientation with respect to the wind is averaged uniformly over the range 0-360 ~ (Tsai et al., 1994). This is the situation most appropriate to most industrial hygiene sampling. Firstly, in particular, it applies directly to the human head studies which led to the I S O / A C G I H / C E N definition of inhalability. Secondly, it applies to some of the new rotating-head omnidirectional samplers which have been developed for area sampling in the workplace and the ambient astmosphere. By correlating the orientation-averaged impaction model with data for the human head (from the mannequin studies reported by Vincent et al., 1990) and for two rotating head samplers (Mark et al., 1985, 1990), Tsai et al. obtained (A9o_Ao) A = 0.5A o +

] + 0.5A18 o

(181R-2-31r 1.01 + 2)

255

(9.21)

Aerosol science for industrial hygienists

for uniform averaging of A over the range 0 ~< 0 ~< 180 ~ where A 0 is as calculated from Equations (9.18) and A90 and A180 are as calculated from Equations (9.20). Equation (9.21) is expected to apply for samplers which are essentially symmetrical in terms of shape and of the location of the sampling orifice on the blunt sampler body. For personal sampling, the situation is more complex still since not only is the body of the wearer asymmetric (i.e., close to elliptical in cross-section) in shape but so too is the position of the sampler on the body. Tsai et al. (1995) have very recently extended the preceding model in order to reflect the effect of these asymmetries, and obtained the complicated expression Aav e =

(0.4 Ao + 0.2

A90 +

0.4 A180)

0.4 (A0-A90)

] (9.22)

(32 St'-~176

'0-69

+

1)

0.1 (A0-A90) (85 St'17-1~

]

'-1-12 + 1)

Here, St', R' and r' are versions of the previously-used St, R and r which have been modified to account for the asymmetries identified. This model has been found to be in good agreement with experimental data like those described later in this chapter for a range of practical personal samplers.

9.5 S A M P L I N G F R O M C A L M A I R All the sampling scenarios described so far refer to moving air. This relates not just to the motion of the air but, more specifically, to the relative velocity between the sampler and the surrounding air. This distinction becomes important in thinking about personal aerosol samplers which workers wear on their bodies. Even in calm air, the motion of the workers as they move around the workplace introduces the relative motion referred to. So, in such cases, the moving air model remains the most relevant. However, there are some cases where the relative velocity between the sampler and the surrounding air is low enough that the air may be regard as 'calm'. So it is appropriate here to discuss the limiting case of calm air sampling. In an ideal calm air sampling system, the only movement of air is that induced by the aspirating action of the sampler itself. Whereas in the case of moving air the particles are brought to the vicinity of the sampler by

256

Aerosol sampling in workplaces

Figure 9.10.

Picture of the scenario for the aspiration of particles by a blunt sampler in calm air (from Vincent, J.H., Aerosol Sampling: Science and Practice, Copyright 1989, adapted by permission of John Wiley and Sons Limited). Note here that, unlike for the moving air case, it is gravity that brings the particles into the vicinity of the sampler.

convection, particles in calm air arrive under the influence of gravity. The physical scenario for a sampler of arbitrary shape with a single orifice at arbitrary orientation is shown in Figure 9.10 (see also Figure 4.18). Some particles falling towards the sampler may come close enough to come under the influence of its aspirating action and so may be drawn towards the sampling orifice. Close to the orifice, inertial effects start to influence particle motion. So, analogous to the moving air case, we have a limiting particle trajectory. It may be shown that aspiration efficiency for calm air sampling is given by Vs a

A =

(9.23)

Q where Q is the volumetric flow rate of sampling and a is the area above the sampler enclosed within the limiting particle trajectory surface. Only particles falling inside this area have any chance of being aspirated into the sampler. Even so, particles may still fail to be collected due to the 'shadow' that the body of the sampler presents to the falling particles, in particular for samplers where the sampling orifice is downwards-facing. As described in Chapter 4, v~ in Equation (9.23) is the falling speed of particles of given size. For the

257

Aerosol science for industrial hygienists

aspiration efficiency for the system described, we may m a k e the general functional s t a t e m e n t

{

A =f

Stc, Rc,

,B,O D

}

z

(9.24)

where we now have a new sampling ratio for calm air Vs Re

(9.25)

--

Us and a new Stokes' n u m b e r d 2

ae0 9 Us

Stc :

(9.26) 18~5

Most research conditions u n d e r has been shown both St c and R c

into calm air sampling has focused on identifying the which aspiration efficiency is close to unity (or 100%). It that 'acceptable' sampling can be achieved provided that are small enough. The conditions for 'acceptability' at the

!o

~0%

Sic

!0 -2

103 10-3

10-2

10-1 Re

Figure 9.11. Summary of the conditions for 'acceptable' sampling in calm air, showing the Davies (1968) criterion and the less restrictive Agarwal-Liu (1980) criterion (from Vincent, J.H., Aerosol Sampling: Science and Practice, Copyright 1989, adapted by permission of John Wiley and Sons Limited).

258

Aerosol sampling in workplaces + 5 % level, as p r o p o s e d by Davies (1968) and A g a r w a l and Liu (1980), respectively, are shown in Figure 9.11. T h e Davies version derives f r o m largely qualitative c o n s i d e r a t i o n of the c o m b i n e d inertial and gravitational b e h a v i o u r of particles n e a r the sampling inlet, and states that ' a c c e p t a b l e ' sampling is achieved w h e n Stc R~

~< ~<

0.016 for the inertial b e h a v i o u r , a n d 0.04 for the gravitational b e h a v i o u r

T h e A g a r w a l - L i u v e r s i o n d e r i v e s f r o m n u m e r i c a l c a l c u l a t i o n s of aspiration efficiency for u p w a r d s - f a c i n g s a m p l e r s . T h e y f o u n d e m p i r i c a l l y t h a t 1.05 >1 A I> 0.95 p r o v i d e d that St c R e <~ 0.05

(9.27)

T h e Davies criterion is the one which has b e e n the m o s t widely used in practical sampling considerations. But it m a y be r e p l a c e d for a p p r o p r i a t e sampling systems (i.e., with upwards-facing orifices) by the m o r e recent and less restrictive version of A g a r w a l and Liu. Example 9.3. A plastic cassette sampler of the type widely used in industrial hygiene has a closed face with a sampling orifice of diameter 4 mm and is used at the sampling flow rate of 2 l min -I. It is usually used as a personal sampler with its orifice facing downwards (see Figure 9.24). What is the maximum particle size for which 'acceptable' sampling can be achieved? Here the Davies criterion should be used since it is more general (the Agarwal-Liu criterion applies strictly only to upwards-facing samplers) The sampling velocity of the sampler in question is (2x 10-3/60) [m 3 S-l] X 4 G

__

3.142 x (42x 10-6) [m2] = 2.65 m s-1 From the Davies criterion, as given by Equation (9.26), acceptable sampling first requires (St c condition) 0.016 x 18 x (18x10-6)[N s m -2] x 4x10-3[m] dae 2 ~<

10-'2[m 2 ixm-2l • 103[kg m-31 • 2.65 [m s-~] 7.8 ixm2

259

A e r o s o l science f o r industrial hygienists In addition, it requires (R c condition) 0.04 x 18 x 18x 10-6[N s m -2] • 2.65[m S-1] dae 2 10-12[m 2 $xm -2] • 103[kg m -3] x 9.81[m S-2]

~< 3501 Dm 2 From the above it is clear that the first condition is the most restrictive. That is, the inertial behaviour of particles near the small sampling orifice is dominant. So this condition should be applied The maximum size of particle aerodynamic for acceptable sampling is 2.8 Ixm. That is, no particle larger than this will fall within the Davies-criterion hatched area shown in Figure 9.11. Note: This is a surprisingly low value, especially bearing in mind that such a sampler is so widely used in practical industrial hygiene. However, it is likely that such a sampler would be infrequently used under calm air conditions

E x a m p l e 9.4. W h a t is the s m a l l e s t p a r t i c l e size t h a t can be efficiently s a m p l e d by t h e o p e n - f a c e d v e r s i o n of t h e s a m e s a m p l e r , w h e r e t h e s a m p l i n g orifice is n o w 35 m m ? The sampling velocity is now U S = 0 . 0 3 5 m s -~ For the St c condition 0.016 x 18 x (18xlO-6)[N s m -2] x (35x10-3)[m] dae 2 ~< 10-12[m2 ixm -2] x 103[kg m -3] x 0.035[m s-1]

~< 5184 txm 2 For the R c condition 0.04 x 18 x (18 x10--6) [N s m -e] x 0.035[ms -1] dae 2 ~<

10-12[m2 i~m-2] x 103 [kg m -31 x 9.81[m s-2] ~< 46.23 txm e

260

Aerosol sampling in workplaces Now we see that the second condition is the most restrictive. That is, the gravitational behaviour of particles near the small sampling orifice is dominant. So this condition should be applied The maximum size of an aerodynamic particle for acceptable sampling is now 6.8 ~xm. That is, no particle larger than this will fall within the Davies-criterion hatched area shown in Figure 9.11 Note" This is again a surprisingly low value In principle, aspiration efficiency for moving air and calm air sampling should be described by a single physical model. At present, none of the models mentioned above achieve this since (a) in moving air gravity is neglected, and (b) in calm air convection is neglected. Recently, Grinshpun et al. (1993) have proposed a model of the 'impaction model' type for the thin-walled sampling probe which incorporates both moving air and calm air ideas. This provides a useful starting point for future discussions of what happens in the case of practical samplers like those used in industrial hygiene in workplaces where the relative motion between the sampler and the surrounding air may be quite slow.

9.6 P H Y S I C A L F A C T O R S W H I C H C A N C O M P L I C A T E S A M P L E R PERFORMANCE There are a number of factors which are not built into the models described above for estimating aspiration efficiency, but which, under certain workplace conditions, can ~ singly or in combination have a significant effect on performance. They are reviewed briefly here, mainly to serve as warnings to the unwary.

Effects of freestream turbulence on aspiration efficiency In Chapter 2, the phenomenon of air turbulence was described briefly. It is clear that, in any relatively large-scale moving air system, the characteristic Reynolds' number (Re) will be such that the motion is turbulent. Certainly as far as aspiration by samplers is concerned, it is inevitable that the workplace air from which the aerosol is being aspirated will almost always be turbulent. What then is the effect on particle motion and, in turn, on the aspiration efficiencies of samplers? It is well known that a major effect of turbulence is to introduce enhanced exchange of airborne material by an effective increase in particle diffusion. The effect of such diffusion is, by Fick's

261

Aerosol science for industrial hygienists law, to cause a net flow of particles from regions of high to regions of lower concentration. However, during the normal process of aspiration, one effect of the inertial forces on particles moving in the distorted flow pattern is to develop spatial concentration gradients. It follows, therefore, that the superposition of turbulence will act so as to try to restore the uniform particle concentration distribution that originally existed in the undistorted flow upstream of the sampler; that is, to try to restore A --. 1. In turn this leads to the useful (and reassuring) conclusion that, for isokinetic sampling with a thin-walled probe of the type described above, A should remain at unity and so should not be affected by the turbulence. Little research has been conducted into this complex phenomenon. But from the work that has been conducted so far (e.g., Vincent et al., 1985), the evidence is that the effects of turbulence on aspiration efficiency are likely in practice to be significant only for large particles moving in a turbulent flow field of high intensity. But theoretical consideration of the problem is further complicated by the fact that, at large enough turbulence length scales, the relatively small sampler 'sees' the turbulence (and the aerosol it transports) not so much as a continuous diffusing system but rather as a slowly-varying mean flow.

Effects of external wall interactions on aspiration efficiency

In sampler theory like that outlined earlier, it is implicit that the aerosol particles that are not aspirated into the sampling orifice either pass by the outside of the sampler or impact onto the sampler body. For the latter, it is usually assumed in aspiration theory that the particles, upon arrival at the external sampler surface, stick and so are permanently removed from the aerosol. However, this is an assumption which has been found to be frequently invalid (e.g., Vincent and Gibson, 1981; Lipatov et al., 1988). Although it may be true for particles that adhere strongly on impact (e.g., fine, smooth particles where Van der Waals forces are strong, or sticky liquid droplets), it may not be the case for coarse, gritty particles which are less strongly retained on impact. What might happen in reality is illustrated in Figure 9.12. Firstly, for example, if the normal component of the velocity on impact of such particles is large enough, the particles may rebound directly (i.e., bounce). This effect depends not only on the particle approach velocity but also on the coefficient of restitution for the particles impacting onto the surface (see Figure 9.12a). This is a dimensionless quantity describing the particle's ability to bounce, being unity for particles which rebound perfectly (leaving with the same velocity with which they arrived) and zero for particles which do not bounce at all. Otherwise, such particles, having been captured by the surface, may be blown off (or re-entrained). This can occur if the drag forces acting on the

262

Aerosol sampling in workplaces (a) I \

High-velocity tmpact

) ~,

(b)

/ o~ c i " t~ ~yJ ~ Low-vel impact, particle ~ ~ is deposited

Boundary layer

Drag force exceeds rolling or sticking friction ------~BLOWOFF

Drag force associated with the boundary layer flow Figure 9.12. Illustration of the physical mechanisms by which particles may rebound from the sampler surface and so contribute to oversampling: (a) bounce, and (b) blow-off.

particles sitting momentarily on the sampler surface in the boundary layer of the air flow close to the surface are large enough to break the adhesion forces holding them there (see Figure 9.12b). What happens to the particles next depends on the direction of the flow into which they are re-entrained. If they find themselves in the flow which is convergent towards the sampling orifice, then they will be drawn towards and may enter the orifice and be sampled. This phenomenon has been referered to in some quarters as 'secondary aspiration' (Lipatov et al., 1988). When it occurs, then the entry efficiency of the sampler cannot be described just in terms of the ideal aspiration efficiency but rather by the apparent aspiration efficiency, as defined in Equation (9.3). Now there will be oversampling compared to the ideal performance. Studies have shown that, for some dusty aerosols (e.g., coal mine dusts), oversampling of total dust for 'typical' samplers can occur by up to 50% (Mark et al., 1982). Effects of electrostatic interference on aspiration efficiency

As mentioned in Chapter 3, workplace aerosols are invariably charged to levels substantially above Boltzmann equilibrium (Johnston et al., 1985). This 263

Aerosol science for industrial hygienists

means that the possibility exists for electrostatic forces to influence sampler performance. Over the years, erratic sampler performance has frequently been explained in terms of such 'electrostatic interference'. This relates to the action of electrical forces both outside the sampler which might influence aspiration efficiency and inside the sampler (en route to the filter) to influence wall deposition. Both are possible if particle charge is large enough and/or if the sampler itself can become charged to sufficient potential. They can occur, for example, by frictional charging during handling of the sampler or in the process of being worn by a worker. In recent years, understanding of these effects has improved. Now it seems likely that, as far as events outside the sampler are concerned, there are few practical situations where particle and/or sampler charge can become great enough for electrostatic interference to significantly influence aspiration efficiency. However, the possibility of some measurable effect should not be entirely ruled out in extreme cases ~ for example, for a personal sampler placed on the body of a worker who is wearing highly-insulating footwear in conditions of very low workplace environmental humidity.

Transport losses of particles after aspiration After aspiration, aerosol is usually transported to a filter (e.g., where particles may be collected and subsequently weighed gravimetrically or otherwise analysed) or to a sensing zone where real-time sensing may take place (e.g., by light scattering). Unless the filter (or sensing zone) is placed immediately adjacent to the sampler entry, then the aerosol must be conveyed through some sort of duct or conduit. During such transport, deposition on the internal walls of the sampler may take place by a variety of

Diffusion

z~-/~-///////

Figure 9.13.

action

Gravity Sampler Summary of some of the mechanisms by which particles may be lost to the sampler inner wall after aspiration.

264

Aerosol sampling in workplaces

mechanisms, including inertial impaction, interception, gravitational settling, diffusion and electrostatic precipitation (some of which are illustrated in Figure 9.13). It is interesting to note that these processes are also those which may influence deposition of inhaled particles in the respiratory tract. Depending on the orientation and geometry of the sampler, the length of the connecting transition, sampling flow rate and particle properties, such deposition can be substantial. It represents a loss from the aspirated aerosol such that the concentration of that which reaches the filter is lower than for that which is aspirated. The overall effect of internal wall losses, therefore, is to cause undersampling (in contrast to the oversampling associated with external wall interactions described above). The following review of some of the more specific mechanisms associated with internal wall loss is intended largely for guidance in assessing transport losses and in designing sampling systems to keep such losses to a minimum. Internal losses near the entrance. Willeke and co-workers (e.g., Okazaki et al., 1987) have identified that, at least for thin-walled sampling probes, significant losses of particles can occur close to the entrance of the sampler. This is the result of the coupling that exists between the external flow and the internal flow. This may be different, depending on whether sampling is super- or sub-isokinetic. Two effects are identified. The first occurs under sub-isokinetic conditions and is associated with the growth of the boundary layer just inside the tube. The second is associated with the vena contracta which characterises the flow in this region under super-isokinetic conditions (when the external flow moving towards the sampler is convergent) or when sampling at large angles with respect to the air flow. These flow phenomena are illustrated in Figure 9.14 for thin-walled sampling tubes. Hangal and Willeke (1990) have considered the mechanisms of particle deposition in this region close to the leading edge of the tube, and have taken into

~

Particle deposition in the separated flow of th? vena contracta v///////////////////////////////~

vl///////////////////////////.

~v / ~

Particle deposition as it decelerates in the boundary layer flow Figure 9.14. Picture illustrating the nature of particle losses to the sampler inner wall in the region close to the entry orifice (after the work of Willeke and his colleagues).

265

Aerosol science for industrial hygienists account the effect of both gravity and inertia (including direct impaction onto the inside tube wall when sampling at an angle) in developing a set of semi-empirical equations for the efficiency of penetration for the inlet region. These equations are only useful if similar equations are available for the other sources of significant particle loss inside the sampler. For these, good solutions are available for simple tube-like transitions (see below). But for many practical samplers like those used in industrial hygiene, the situation is often too complicated for a usable general set of equations to be prescribed. Losses along the length of a sampling tube. First we consider a straight horizontal tube where, once the internal flow has been established (i.e., beyond the inlet transition region), particle deposition can occur by several processes. The first is by molecular diffusion, as described by Gormley and Kennedy (1949). Deposition is greater for smaller particles where the Peclet number (the governing parameter for diffusion defined in Chapter 4) is smallest. Next we have gravitational settling, described by Thomas (1958) and others. This is most effective for larger particles where the gravitational parameter (G, also defined in Chapter 4) is greatest. Then there is turbulent deposition which can occur when the Re value for flow inside the tube is large enough for the flow to become turbulent (see Chapter 2). This problem has been studied by several workers (including Friedlander and Johnstone, 1957; Davies, 1965; Beal, 1970; Liu and Ilori, 1973; Liu and Agarwal, 1974 and others). Deposition is governed by the turbulent diffusion coefficient for particles, increasing with the particle relaxation time in a manner dependent on the nature of the turbulent flow in the tube (dependent in turn on Re). Finally, there is deposition by electrostatic forces. This can be important, especially since it is now known that aerosols in workplaces may be charged to significantly above Boltzmann equilibrium (see Chapter 3). Liu et al. (1985) have shown that the transport of charged particles through long narrow tubes is modified if the particles are charged, and deposition may be considerable if the tubing itself is of insulating material and can become charged. Such electrostatic effects can also occur in some samplers of more direct practical relevance to industrial hygienists, for example, inside the short tube-like plastic cowls which are sometimes placed over the entries of open-filter samplers of the type used in asbestos dust sampling. Here, considerable wall losses by electrostatic deposition have been observed. Losses at bends and transitions. Particle losses may be enhanced by inertial deposition at bends, and has been studied by several workers. It has been shown that penetration increases as the angle of the bend (+) increases and as St decreases (where St here is the Stokes' number for particle transport in the vicinity of the bend, increasing with particle aerodynamic diameter and air velocity and decreasing as the tube diameter increases). Brockmann (1993) has summarised the available results for both laminar and turbulent flows to give

266

Aerosol sampling in workplaces P = 1 - Stbend + for laminar flow (9.28) P - exp(-2.823 Stbend +) for turbulent flow where r is in [radians]. Other losses can occur at other types of transition in conduits like those being discussed. However, in many practical sampling systems, the geometries are so complicated that the flows, and hence particle transport, are difficult to generalise from one to the other.

Overall sampler performance The overall performance of an aerosol sampler depends on the impact of all the processes mentioned, starting with the external processes of aspiration and secondary aspiration, followed by the various internal loss processes. Equation (9.4) may be re-written in the form Aoveral I -- ( 1 - Eloss ) (A + Asec)

(9.29)

where A is the aspiration efficiency as already defined and Ase c is the coefficient of secondary aspiration given by (referring to Figure 9.1)

Nr Ase c --

(9.30)

No and where Eloss is the efficiency of the deposition loss of particles after aspiration ( - l - P ) . If all the processes contributing to this internal deposition were independent and in series, then we could write Eloss = Elos~ 1 Elo~ 2 Elo~s 3 " " "

(9.31)

the product of the individual deposition efficiencies. In reality, the processes in question are not all independent. So Equation (9.29) can only be regarded as an approximation for use in determining Aoveral 1. The ideal sampler is one where E and Ase c are zero, in which case Aoveral 1 becomes equal to the idealised aspiration efficiency A. From the preceding short outline of particle interactions with the sampler wall, some guidelines are suggested to enable these quantities to be minimised in sampling systems. They may be summarised briefly as follows: External design of samplers should be carried out to minimise the ability to collect rebounding particles (e.g., by means of projecting or 'lipped' inlets);

267

Aerosol science for industrial hygienists

Transition sections between the inlet and the sensing zone (or filter) should be kept as short as possible; Bends, contractions, expansions, etc. should be kept to a minimum and designed to minimise particle deposition; and The use of non-insulating materials in sampler construction is strongly recommended to minimise losses due to electrostatic forces.

9.7 S A M P L I N G IN STACKS A N D DUCTS Sampling in stacks and ducts is usually carried out for the purpose of determining the emission of aerosols from industrial processes to the ambient atmosphere or, if exhaust air is to be re-used, to the workplace atmosphere. It therefore forms one basis for assessing the performances of aerosol control systems such as will be described in Chapter 11. Because the air flows in such situations are usually well defined in terms of velocity and direction, isokinetic sampling using thin-walled probes provides the main basis for practical methodology. From the earlier section on thin-walled probes, it is clear that this is one area of aerosol sampling where we do have a clear idea of the criterion for sampling (i.e., true total aerosol) and of how to go about making practical measurements. For isokinetic sampling in stacks and ducts, the aerosol science is generally straightforward and well described in the literature. There are, however, a host of engineering and practical guidelines which have to be adhered to, and these are covered in detail in the various reference and standard procedure documents which have been published. Here it suffices to mention only some of those which relate directly to the actual sampling itself. The first consideration is to choose a sampling location in the duct where possible interfering factors associated with the flow are minimised. The velocity and direction of flow near the sampler need to be well defined so that the sampling flow rate can be readily adjusted to give isokinetic conditions (i.e., Us = U). Therefore, locations close to bends, sudden changes in duct cross-section (where there may be flow separation and reversal, vortex shedding, intense turbulence, etc.) should be avoided. Having chosen a suitable duct location, sampling then involves an appropriate sampling strategy. To obtain a sample representative of the aerosol across the whole duct, it is required to sample at a number of points in the duct cross-section, usually by making traverses with the same probe across the duct. The individual sampling points are chosen so that, by such integration over the duct cross-section, an appropriate measure of the overall mass flow of aerosol in the duct can be obtained. The method of choice of individual locations is illustrated for ducts of both rectangular and circular cross-section in Figure 9.15. The underlying principle is that the duct

268

Aerosol sampling in workplaces

Figure 9.15. Recommended location of sampling points in ducts of (a) circular cross-section, and (b) rectangular cross-section (from Vincent, J.H., Aerosol Sampling: Science and Practice, Copyright 1989, reproduced by permission of John Wiley and Sons Limited).

cross-section is divided into partial areas of equal cross-section, and sampling takes place for equal duration at the centroid of each. The same principle may be applied to ducts of other cross-section. In order to maintain representative measurement, the required number of such partial areas is determined by the overall quality of the flow. For very uniform flow, that number can be kept small. On the other hand, if the flow is non-uniform, or if there are substantial flow disturbances taking place upstream, the number should be increased.

9.8 S A M P L I N G F O R C O A R S E A E R O S O L IN W O R K P L A C E S Static (or area) samplers for 'total' or inhalable aerosol Static (or area or fixed-position) samplers have been used for many years in the sampling of coarse aerosol in workplace atmospheres. The simplest

269

Aerosol science for industrial hygienists

Figure 9.16. Typical static (or area) sampler for 'total' aerosol (of the type widely used by British industrial hygienists) (from Mark, D. et al., Investigation of the entry characteristics of dust samplers of the type used in the British nuclear industry. Atmospheric Environment, 20, 2389-2396, Copyright 1986, with kind permission from Elsevier Science Ltd, The Boulevard, Langford Lane, Kidlington OX5 1GB, UK).

are open-filter arrangements mounted on the box which contains the pump (see Figure 9.16) or systems in which the open-filter holder is mounted independently. The sampler shown is widely used in Britain. Similar devices have been used elsewhere, both in workplace and in ambient air sampling. The performances of such intruments, originally intended as samplers for 'total' aerosol, should now be assessed in the light of the latest health-related particle size-selective criteria described earlier, in particular in relation to the inhalability curve. Results for the sampling system shown in Figure 9.16 are given in Figure 9.17, for a range of orientations with respect to the wind direction from 0 to 180 ~ windspeed up to 1 m s-1, flow rates in the recommended range 30-100 1 min -1, and particle aerodynamic diameter (dae) up to 30 ixm. These data show that, whilst the instrument provides quite a fair measure of true total aerosol, it oversamples significantly with respect to inhalability. Data for several other static 'total' aerosol samplers are also available (as reviewed by Vincent, 1989). The performances of these are characterised variously by tendencies to (a) undersample or oversample with respect to the inhalable fraction, and (b) be strongly dependent on

270

Aerosol sampling in workplaces U Q (m s "t) (1 m i n "l)

~9

30

zx ! 0 5 9 110

0=0 ~

30 100 100

0 = 90 ~

o

A

1.0

IO ,

I 10

21o

0=45 ~

I 20

30

0 = 180 ~ o z~a

1.0

I 10

I 20

3

o

o

oA

I I0

I 20

3

P a r t i c l e a e r o d y n a m i c d i a m e t e r , dae(l~m)

Figure 9.17. Aspiration efficiency (A) as a function of particle aerodynamic diameter (dae) for the static sampler shown in Figure 9.16 (from Mark, D. et al., Investigation of the entry characteristics of dust samplers of the type used in the British nuclear industry. Atmospheric Environment, 20, 2389-2396, Copyright 1986, with kind permission from Elsevier Science Ltd, The Boulevard, Langford Lane, Kidlington OX5 1GB, UK).

windspeed. In s h o r t , none of these examined comes close to a satisfactory match with inhalability. In the light of such data, new generations of aerosol sampler are beginning to emerge, this time designed from the outset to match the inhalability criterion. One designed for use in workplaces is the 3 1 min -1 Institute of Occupational Medicine (IOM) static (or area) inhalable aerosol sampler (shown in Figure 9.18) (Mark et al., 1985). It incorporates a number of novel features. Firstly, the sampler contains a single sampling orifice located in a head which, mounted on top of the housing containing the pump, drive and battery pack, rotates slowly about a vertical axis. The entry orifice forms an integral part of an aerosol-collecting capsule which is located mainly inside the head. This capsule also houses the filter. In the use of the instrument, the whole capsule assembly (tare weight of the order of a few grams) is weighed before and after sampling to provide the full mass of aspirated aerosol. If desired the filter and the internal wall deposit (after washing

271

Aerosol science for industrial hygienists

Figure 9.18. The 3 1 min -1 IOM static inhalable aerosol sampler (photograph courtesy of Perng-Jy Tsai, from his Ph.D. thesis, University of Minnesota, 1994).

(mU-I) 1.0

9

1

/x

3

Inhalability A A

0.5

-

-

A A

9 A

~

A 9

I

0

I

50 ioo Particle aerodynamic diameter dae (lzm)

Figure 9.19. Measured aspiration efficiency (A) of the 3 1 min-1 static inhalable aerosol sampler as a function of particle aerodynamic diameter (dae) and for two windspeeds (from Mark, D. et al., American Industrial Hygiene Association Journal, Copyright 1985, reproduced by permission of the American Industrial Hygiene Association).

272

Aerosol sampling in workplaces

or wiping out) can be recovered and combined so that the whole inhalable 'catch' can be quantitatively analysed for relevant chemical sub-fractions. The system described eliminates the possibility of errors associated with internal wall losses of the type described earlier. Mark (1990) has demonstrated the feasibility of using such capsules, and showed that the high tare weight is not a problem when used with an appropriate modern electronic balance and provided that appropriate care is taken to achieve sample stabilisation in a dessicator before weighing. When the capsule is mounted in the sampling head, the entry itself projects about 2 mm out from the surface of the head, creating a 'lip' around the orifice itself. This has the effect of preventing the secondary aspiration of any aerosol particles which strike the outside surface

Figure 9.20. A prototype 30 1 min -1 static (or area) inhalable aerosol sampler developed for aerosol measurement in the ambient atmosphere (from Vincent, J.H., Aerosol Sampling: Science and Practice, Copyright 1989, reproduced by permission of John Wiley and Sons Limited).

273

Aerosol science for industrial hygienists

of the head and fail to be retained. The performance of this sampler, shown in Figure 9.19, is seen to be in good agreement with the inhalability curve for particles with d ~ up to about 100 Ixm for windspeed up to 3 m s-1. Several prototype higher flow rate versions were built at the Institute of Occupational Medicine during the late 1980s and are still being considered for applications in sampling inhalable aerosol in the ambient atmospheric environment (one is shown in Figure 9.20).

Static (or area) samplers intended primarily for finer aerosol fractions As noted in Chapter 8 in relation to the latest aerosol sampling criteria, if a sampler is to be used for collecting a fine aerosol fraction corresponding to deposition in a particular region of the respiratory tract, it should (ideally) first aspirate the inhalable fraction. Otherwise, if a sampler has a poorly defined aspiration efficiency, or one which varies in an uncontrolled way with (say) windspeed or orientation, then bias can result in the determination of the fine fraction of interest. Therefore, in samplers intended primarily for determination of finer fractions, it is relevant to consider their performances (in terms of their aspiration efficiency) with respect to the inhalability criterion.

Figure 9.21. The British 2.5 1 min -1 MRE Type l13A static sampler, intended primarily to collect respirable dust in coal mines according to the BMRC criterion (from Vincent, J.H., Aerosol Sampling: Science and Practice, Copyright 1989, reproduced by permission of John Wiley and Sons Limited). 274

Aerosol sampling in workplaces One static sampler commonly used for sampling the fine respirable fraction is the British 2.5 1 min -1 MRE Type l13A shown in Figure 9.21. It was developed for use in coal mines. Data for the aspiration efficiency of this and other such instruments have been compared with the inhalability curve, and it has been found that none matches the inhalability criterion particularly well for particles in the coarse size range exceeding about 10 ~xm (see Vincent, 1989). But for finer particles in the respirable range, agreement is much better. This suggests that, for these samplers in this range, reasonable consistency with the inlet efficiency requirement embodied in the latest sampling recommendations is maintained under most conditions. However, there does remain some concern about windspeed dependency. Results, for example, for the performance of the MRE instrument in relation to the collection of respirable dust have shown a strong dependence on windspeed for windspeed greater than about 5 m s-1 (Ogden et al., 1977). These strongly suggest the onset of significant changes in aspiration efficiency for respirable-sized particles at higher windspeeds. So caution is recommended in interpreting respirable dust data obtained using such instruments under conditions of such high windspeed (as might be found, for example, in some mining environments).

Personal samplers for 'total' or inhalable aerosol

For reasons outlined earlier, personal sampling is generally the preferred approach for workplace aerosols. Here, for coarse aerosol, a large number of different devices have been used, again originating ~ historically for the purpose of sampling for 'total' aerosol. Again, the simplest is the open-filter arrangement, the one shown in Figure 9.22 being the 25 mm open filter used by British industrial hygienists in some applications. Other personal samplers for total aerosol currently in use in Britain are the single (4 mm) hole sampler recommended by the Health and Safety Executive for lead-containing aerosol and the modified seven-hole version recommended for general coarse aerosol sampling. These too are shown in Figure 9.22. Both of these closed-face samplers also employ 25 mm filters. All three samplers are intended for use at the sampling flow rate of 2 1 min -~. Wind tunnel experiments have been conducted to compare their performances with the inhalability curve. It is particularly important to note here ~ as well as for all the other personal samplers discussed below that only data obtained with each sampler tested whilst mounted realistically on a life-size torso (e.g., of a mannequin) are considered useful in this context. It should not be assumed that, if such samplers were to be tested independently in isolation, they would necessarily provide the same results. The aerodynamic conditions governing the airflow around the sampler would be quite different. So it follows that devices designed as personal samplers

275

Aerosol science for industrial hygienists

Figure 9.22. Three personal samplers of the type widely used in Britain for the sampling of 'total' aerosol (from Vincent, J.H., Aerosol Sampling: Science and Practice, Copyright 1989, reproduced by permission of John Wiley and Sons Limited). should not be used in the static mode. Unfortunately this commonsense guideline is widely ignored. The results for the three samplers in Figure 9.22 are shown in Figure 9.23. It should be noted that these represent the efficiency with which particles are collected on the filter, the mode of routine use of these samplers. So the performances are portrayed in terms of Aoveral 1. The results indicate that all match the inhalability criterion quite well for particles with dae up to about 15 txm and for windspeeds of 1 m s-1 and below. But for conditions outside these ranges, yet typical of those found in many workplaces, the performances are less satisfactory, with strong windspeed dependency (especially for the single-hole and seven-hole samplers) and with a marked tendency towards undersampling. The physical and design features of these three samplers are, in one way or another, representative of those exhibited by most of the many others which have been designed and used over the years in many countries. One is the 37 mm plastic cassette which is employed widely, either open-faced or closed-faced, by industrial hygienists in the United States and many other countries (see Figure 9.24). Until recently, test results for this sampler were limited in number and range of dae covered (Buchan et al., 1986). But recently, more comprehensive new data have been presented for the closed-face version by Mark et al. (1994). As shown in Figure 9.25, they reveal a striking resemblence to those for the 4 mm orifice sampler shown in

276

Aerosol sampling in workplaces

Open-faced

+

1.0

sampler

+ ++ +

/ vX

9~v+~,.,,_

/

_

A A O

q~

0.5

V

~7

~7 X7X7 V

V

0

I

I

O AV /X

I

I

I

Single-hole

zx

sampler U ( m s -I)

1.0

,+A

I=

0.5

--

+

o zx v +

I

0.5 1.0 2.6 0.6-3.5

+ + V V V

E

A A V

i

1.0

I

OO

o

O

o

~

w

i%l Seven-hole

zx A

o

~7X7

I

~7X7

I

Iw

sampler

!

0.5

V

A

V

V

v

O

v

O

~A

z~

w 0

I I0

I 20

z~

o

o3 A ZX

w

~

W

v I 30

89

I 40

I 50

I 60

iw 70

80

Particle a e r o d y n a m i c d i a m e t e r dae ( ~ m )

Figure 9.23. Measured performance of each of the three samplers shown in Figure 9.22, sampling efficiency plotted as a function of particle aerodynamic diameter (dae) and shown in comparison to the inhalability curve. Data were obtained in a large wind tunnel for the samplers while mounted on the torso of a life-sized mannequin, orientation with respect to the wind averaged uniformly over all directions, based on Mark and Vincent (1986). (Figure from Vincent, J.H., Aerosol Sampling: Science and Practice, Copyright 1989, reproduced by permission of John Wiley and Sons Limited.)

277

Aerosol science for industrial hygienists

Figure 9.24. The 37 mm cassette personal sampler of the type widely used by industrial hygienists in the United States and elsewhere (usually at a flowrate of 21min -1) for the sampling of 'total' aerosol (shown in the closed-face version most commonly used). F i g u r e 9.23. M o s t p a r t i c u l a r l y , t h e y s h o w t h a t the c l o s e d - f a c e 37 m m s a m p l e r s t r o n g l y u n d e r s a m p l e s with r e s p e c t to i n h a l a b l e a e r o s o l for p a r t i c l e s a b o v e a b o u t 20 txm. In the light of the g e n e r a l l y p o o r p e r f o r m a n c e s of m a n y existing ' t o t a l ' a e r o s o l s a m p l e r s with r e s p e c t to the i n h a l a b i l i t y c r i t e r i o n , a n e w p e r s o n a l s a m p l e r was p r o p o s e d ( M a r k a n d V i n c e n t , 1986). This is the 2 1 m i n -1 I O M

1.01

m

r

. ,r. ~

.,..,

I

0

50

I

100

Particle aerodynamic diameter, d~ (l~m)

Figure 9.25. Sampling efficiency for the 37 mm sampler in Figure 9.24, plotted as a function of particle aerodynamic diameter (dae) for a windspeed of 1 m s -1, and shown in comparison to the inhability curve. Data were obtained in a large wind tunnel for the samplers while mounted on the torso of a life-sized mannequin, orientation with respect to the wind averaged uniformly over all directions (from Mark, D. et al., Wind tunnel testing of the sampling efficiency of personal inhalable aerosol samplers, Journal of Aerosol Science, 25, Supplement 1, $339-$340, Copyright 1994, with kind permission from Elsevier Science Ltd, The Boulevard, Langford Lane, Kidlington OX5 1GB, UK).

278

Aerosol sampling in workplaces

Figure 9.26.

The 2 1 min -1 I O M personal inhalable aerosol sampler.

!.0

+i•

U (m s"1) o 0.5

o

0.5

_

9 1.0

+g+

o,

~

++

0

I

I0

I

20

~

++

I

30

I

40

~

+o +

I

50

+

I

60

I

70

I

80

Particle aerodynamic diameter, d~ (l~m) Figure 9.27. Measured performance of the 2 1 m i n -1 I O M personal inhalable aerosol sampler, plotted as a function of particle aerodynamic diameter (dae) and shown in comparison to the inhalability curve. Data were obtained in a large wind tunnel for the samplers while m o u n t e d on the torso of a life-sized mannequin, orientation with respect to the wind averagely uniformly over all directions (based on Mark and Vincent, 1986). (From Vincent, J.H., Aerosol Sampling: Science and Practice, Copyright 1989, reproduced by permission of John Wiley and Sons Limited.)

279

Aerosol science for industrial hygienists

personal inhalable aerosol sampler (see Figure 9.26). It features a 15 mm diameter circular entry which faces directly outwards when the sampler is worn on the torso. Like the IOM static inhalable aerosol sampler in Figure 9.18, the entry is incorporated into an aerosol-collecting capsule which, during sampling, is located behind the face-plate. Use of this capsule ensures that the overall aspirated aerosol is always assessed. Also, as for the static sampler, the lips of the entry protrude outwards slightly from the face-plate in order to prevent oversampling associated with particle blow-off from the external sampler surfaces. Experimental data for this instrument are shown in Figure 9.27, and they show a good match with the inhalability curve for particles with dae up to 80 ~m and for windspeeds up to 2.6 m s-1. Again, the whole inhalable 'catch' can be recovered easily if it is desired to carry out chemical or mineralogical analysis.

Personal samplers intended primarily for finer aerosol fractions Similar to the discussion on static (or area) samplers, a personal sampler for collecting a fine aerosol fraction corresponding to deposition in a particular region of the respiratory tract should (ideally) first aspirate the inhalable fraction. So again it is relevant to consider the aspiration efficiencies of such samplers with respect to the inhalability criterion. The most common personal samplers for finer aerosol fractions are those based on pre-selectors employing the cyclone principle. An example of one of these is shown in Figure 9.32. Unfortunately, it appears that there are no aspiration efficiency data available for such samplers. But based on the experience gained for other types of sampler, it is reasonable to assume that aspiration efficiency will follow the inhalability curve quite well for fine particles in the respirable size range of primary interest in the use of these devices, at least at low ambient windspeeds. There are two other devices which have emerged relatively recently as personal samplers primarily for finer aerosol fractions. These have some interesting and unusual features and so deserve special mention. The first is the French CIP10 (see Figure 9.28) (Courbon et al., 1988). Although this instrument is aimed mainly at collecting a finer (respirable) aerosol fraction, it is also capable of providing the concentration of a coarser fraction. So its performance over particle size ranges extending beyond respirable is relevant. The CIP10 is particularly interesting because it incorporates its own built-in pumping unit, consisting of a battery-driven, rapidly-rotating polyester foam plug. Aerosol is aspirated through a downwards-facing annular entry and is collected efficiently by filtration in two stationary, coarse-grade plastic foam plugs located inside the entry as well as on the finer-grade rotating one. As a result of the low pressure drop characteristics of such foam filtration media, a flow rate of up to 10 1 min -1 can be achieved. By personal sampler

280

Aerosol sampling in workplaces

Figure 9.28. The French CIP10 personal sampler, intended originally for the respirable fraction but also capable of providing information about coarser fractions (as described by Courbon et al., 1988). (From Vincent, J.H., Aerosol Sampling: Science and Practice, Copyright 1989, reproduced by permission of John Wiley and Sons Limited.)

standards, this is very large indeed. Experimental results for the aspiration efficiency of this sampler, again tested mounted on the torso, indicate a fair match with the inhalability curve for particles with dae up to 60 Ixm (see Vincent, 1989). The second interesting novel personal sampler is the Italian P E R S P E C , a device aimed at collecting not only total aerosol but also the finer thoracic and respirable fractions (see Figure 9.29) (Prodi et al., 1986). Aerosol enters at 2 1 min -1 through a pair of crescent-shaped orifices, and is separated by inertial forces into the finer sub-fractions of interest (which are deposited onto different, well-defined parts of the same filter). Again, results for the aspiration efficiency of this sampler indicate a fair match with the inhalability curve for dae up to 60 Ixm (see Vincent, 1989). Despite these encouraging measured performances, however, the use of either of these samplers specifically for the inhalable fraction may be somewhat complicated by the difficulty ~ or degree of skill required ~ in recovering and quantitating the inhalable fraction.

281

Aerosol science for industrial hygienists

Figure 9.29. The Italian PERSPEC personal sampler (as described by Prodi et al., 1986). (From Vincent, J.H., Aerosol Sampling: Science and Practice,

Copyright 1989, reproduced by permission of John Wiley and Sons Limited.) 9.9 S A M P L I N G F O R R E S P I R A B L E A E R O S O L IN W O R K P L A C E S The history of sampling for fine aerosols in workplaces began with the

respirable fraction, in particular with the emergence in the 1950s of the B M R C respirable aerosol criterion (see Chapter 8). A number of types of sampling device have since been developed. Most have in common the fact that they first aspirate a particle fraction which is assumed to be representative of the total workplace aerosol. The desired fine fraction is then aerodynamically separated inside the instrument using physical options such as elutriation (discussed in Chapter 4) and the cyclone, with particle size-dependent penetration characteristics matching the desired criterion. It is the aerosol which remains uncollected inside the selector and passes through to the filter which is the respirable fraction of interest. Static samplers for the respirable fraction

A variety of static samplers for respirable aerosol have been built and successfully used in practical industrial hygiene. One example is the British 2.5 1 min -a M R E Type l13A sampler already mentioned (see Figure 9.21);

282

Aerosol sampling in workplaces

another is the similar, but higher flow rate (and now obsolete), 100 1 min -1 'Hexhlet'. Particle selection in such instruments is based on the principle of elutriation, as shown for the M R E in Figure 9.30a. Here, penetration as a function of particle aerodynamic diameter can easily be tailored to the B M R C respirable aerosol curve using elutriator theory (see Chapter 4) since the B M R C curve was itself originally derived from elutriator theory. Some experimental results for the selection properties of the M R E are shown in (a)

Parallel plate elutriator

k~ Sampled air

-I

/

Glass fibre membrane filter

-I

f

Filter clip

Motor Smoothing reservoir

Rotameter

/ Adjustable crank

\/

Terylene backing filter

Valves

(b) t------t Experiment 1.0

curve 0.5

,

\,

I I I 5 I0 Particle aerodynamic diameter dae (Ixm) I

Figure 9.30. (a) Schematic of the elutriator-based MRE Type l13A static sampler for respirable aerosol (see also Figure 9.21), and (b) results for the penetration characteristics of the MRE as a function of particle aerodynamic diameter (dae) (based on Dunmore et al., 1964). (From Vincent, J.H., Aerosol Sampling: Science and Practice, Copyright 1989, adapted by permission of John Wiley and Sons Limited.)

283

Aerosol science for industrial hygienists

Figure 9.30b, and quite good agreement with the BMRC curve is evident (Dunmore et al., 1964). Other static instruments have been designed operating on the principles of cylcone selection, including such devices as the German 50 1 min -1 TBF50 sampler and the French 50 1 min -1 CPM3 (both developed for use in coal mines). Cyclone and centrifuge-based samplers have also been designed having well-defined penetration characteristics for the respirable fraction, although prediction of performance directly from theory is more complicated than for horizontal elutriators.

Personal samplers for the respirable fraction

Horizontal elutriators have been found to be very satisfactory for static respirable aerosol sampling. But they are inevitably rather bulky and not (a)

Sampled air

(b)

Figure 9.31. Schematic to illustrate the principle of operation of the cyclone (from Vincent, J.H., Aerosol Sampling: Science and Practice, Copyright 1989, reproduced by permission of John Wiley and Sons Limited).

284

Aerosol sampling in workplaces

conducive to miniaturisation. T h e r e f o r e , horizontal elutriation is not promising for the selectors of personal respirable aerosol samplers. On the other hand, cyclones are ideally suited for such purposes, and so have found wide application. The principle of cyclone selection is shown in Figure 9.31. Sampled particles are entrained into the vortex located inside the body of the device, and the larger ones move to the wall under the action of centrifugal forces and so are removed to the pot at the bottom. The finer particles of interest flow back up the centre of the vortex and may be collected there by a filter placed at the cyclone exit. The process has been widely employed in personal samplers for the respirable fraction. There are many well-known examples. The device shown in Figure 9.32 is the British 1.9 1 min -1 Casella cyclone. Typical experimental data for its selection characteristics are shown in Figure 9.33, and they are seen to be in good agreement with the B M R C curve. The French 10 1 min -1 CIP10 sampler mentioned earlier (see Figure 9.28) was intended primarily as a sampler for the respirable fraction. For this device, the selector operates on the basis of foam filtration, where

Figure 9.32. A typical personal sampler for the respirable fraction. The one shown is the British 1.9 1 min -1 Casella cyclone sampler, intended for collecting respirable aerosol according to the BMRC curve (from Vincent, J.H., Aerosol Sampling: Science and Practice, Copyright 1989, reproduced by permission of John Wiley and Sons Limited). 285

Aerosol science f o r industrial hygienists 1.0 -.._

0 0

0 0 00~0 0 0 "JO~, I,j

P

0.5-

1~

o O~Ao

, 0

o

O~ 0 0 o~176 ~176

5

, 10

P a r t i c l e a e r o d y n a m i c d i a m e t e r , dae (p.m)

Figure 9.33. Measured penetration characteristics of the 1.91min-~ Casella cyclone sampler as a function of particle aerodynamic diameter (dae), shown in comparison to the BMRC respirable aerosol curve (results summarised from Higgins and Dewell 1968, Maguire et al., 1973 and Ogden et al., 1983). (From Vincent, J.H., Aerosol Sampling: Science and Practice, Copyright 1989, adapted by permission of John Wiley and Sons Limited). aerosol entering porous polyester foam media is collected by a combination of gravitational settling and inertial forces in a manner rather similar to those taking place in the lung itself. Experimental data for the fine aerosol selection characteristics of the CIP10 are shown in Figure 9.34 (from Courbon et al., 1988), and it is seen that they lie close to the curve for the respirable fraction as defined in the harmonised I S O / A C G I H / C E N criteria (see Chapter 8). Here it is noted that selection also embodies the aspiration efficiency of the sampler. So it is the version of the respirable aerosol curve with the 50% dae value at 4.0 t~m which is plotted in Figure 9.34 (see Chapter 8). One interesting feature of these results is the fall in penetration for small particles with dae less than about 3 I~m. This is the result of the penetration of fine particles through the final foam collector (and so their escape from the instrument). It has been argued that this feature brings the performance of the CIP10 more closely into line with true human exposure than any of the other respirable aerosol samplers described. Impactors (see Chapter 4 and Figure 4.15) also provide some interesting options for particle size-selective aerosol sampling. For example, for bi-modal aerosol like that illustrated in Figure 3.10, it may be considered important to collect the two modes separately (especially if the modes derive from entirely different sources and the aerosols associated with them can give rise to different health effects). It has been found in coal mines in the United States that the modes arising from airborne coal dust on the one

286

Aerosol sampling in workplaces

~ N~o Q o.,,

I.) I.)

o%

ISO/ACGIH/ CEN respirable aerosol curve

0~0 0

,..., !._

9

5

10

Particle

aerodynamic diameter, d~ (l~m)

Figure 9.34. Measured penetration characteristics of the CIP10 personal sampler as a function of particle aerodynamic diameter (dae), obtained as a proportion of total aerosol (from Courbon, P. et al., Annals of Occupational Hygiene, Copyright 1988, reproduced by permission of the British Occupational Hygiene Society). These results are compared with the new ISO/ACGIH/CEN curve, adjusted so that it is a proportion of total aerosol.

hand and diesel particle on the other are clearly divided at d ~ equal to about 0.8 Ixm. In order to collect these two modes separately, a modified personal cyclone-based sampler has been built in which the aerosol passing through the cyclone encounters an impactor with a sharp 'cut' at the desired dae of 0.8 Ixm (Cantrell et al., 1993). The coarser coal dust fraction is therefore collected on the impaction surface of the impactor and the finer diesel particulate passes through to be collected on the backing filter.

Sampling for 'respirable' fibres As already indicated in Chapter 8, the definition of a 'respirable' fibre is based on purely geometric criteria. So selection is best carried out not aerodynamically but visually under the microscope. This means that, in practical sampling, the main priority is to achieve deposition onto a suitable surface (e.g., a m e m b r a n e filter) which can then be 'cleared' to m a k e the collection surface transparent and m o u n t e d for subsequent visual analysis by optical microscopy under phase-contrast conditions. It follows that actual physical sampling can be very simple, usually involving the collection of particles directly onto an open filter (sometimes with the use of a cowl or some other baffle to protect the filter from large airborne material as well as curious fingers!). Such sampling is carried out routinely in both the static and personal modes.

287

Aerosol science for industrial hygienists

5

. ,. , . , . , .

.

",

3~

'

,

.-r.

3

.

,

:

,

;

,

:

,

:

'5

~,~~'b

Figure 9.35. The Walton-Beckett graticule used for the assessment of asbestos fibres in the membrane filter method (from Walton, W.H. and Beckett, S.T., Annals of Occupational Hygiene, Copyright 1977, reproduced by permission of the British Occupational Hygiene Society).

In asbestos measurement, great emphasis is placed on the visual assessment of the sampled fibres. For routine assessment of workplace asbestos, this is usually carried out using an optical microscope under phase-contrast conditions at a magnification of x450. An appropriate graticule, like that described by Walton and Beckett (1977) (see Figure 9.35) is used to provide ease of classification of fibres matching the criteria referred to earlier. Sets of 'counting rules' have been recommended to aid the microscopist in what and what not ~ to count. These guide, for example, the assessment of fibrous aggregates, fibres in the presence of other, non-fibrous particles, fibres not fully contained within the microscope field of view, etc. The technical methods for sample preparation and microscopy, and the processes of selecting and counting fibres, have been extensively researched and fully documented in the various reference methods that have been published (e.g., AIA, 1979; NIOSH, 1979) and will not be enlarged upon here. One important practical aspect is the setting of the sampling flow rate since sufficient flow is required to achieve, over a sampling shift, a sample which is dense enough to provide good counting statistics and reliable visual counting (e.g., Cherrie et al., 1986), yet not so dense as to cause problems with fibre overlap (e.g., Iles and Johnston, 1983). As far as the effects of sampling flow rate on aspiration efficiency for fibres are concerned, it has been demonstrated that, over a very wide range of flow rates, fibrous particles of asbestos are so aerodynamically fine that aspiration efficiency is nearly always close to unity (Johnston et al., 1982). Therefore, for practical purposes, sampling bias due to aspiration effects can be neglected. This provides considerable flexibility in the choice of flow rate in a given situation.

288

Aerosol sampling in workplaces 9.10 P R A C T I C A L S A M P L I N G F O R T H O R A C I C A E R O S O L Methodology for the sampling of thoracic aerosol in the occupational context has not been widely considered prior to the emergence of the first ISO and A C G I H criteria. For workplaces, the nearest we have come to a thoracic aerosol standard is in the United States cotton industry where a criterion was established in 1975 by the US National Institute of Occupational Safety and Health (see Chapter 8) based on a selection curve which falls to 50% at 15 i~m (compared with the 10 ~m 'cut-point' in the ISO and A C G I H thoracic fractions). This implies recognition of the role of particle deposition in the large airways of the upper respiratory tract in cotton workers' byssinosis. The recommended static sampling method employs the concept of vertical elutriation (Walton, 1954). In the future, as the new I S O / A C G I H / C E N criteria begin to be translated into new standards, more energetic consideration will be given to the development of samplers for the thoracic fraction. It is expected that first attempts will be based on the modification of existing respirable aerosol samplers.

9.11 S A M P L I N G F O R M O R E T H A N O N E F R A C T I O N

SIMULTANEOUSLY The important concept that not only is thoracic aerosol a sub-fraction of the inhalable fraction but that respirable aerosol is a further sub-fraction of the thoracic sub-fraction provides a framework by which all three health-related aerosol fractions can be obtained simultaneously. Such aerosol measurements could be important for assessing aerosol-related risk in certain situations, and appropriate practical sampling devices are just beginning to emerge. One such personal sampler is the Italian P E R S P E C already mentioned (see Figure 9.29). The aerosol entering at 2 1 min -1 through the two crescentshaped orifices is winnowed by a clean air sheath in a strongly divergent flow. Particles are separated by inertial forces into the finer sub-fractions which are deposited onto different, well-defined parts of the same 47 mm filter which can later be divided up for mass analysis using an appropriately-designed 'cookie-cutter'. This sampler can provide the inhalable fraction (the whole aspirated catch of aerosol, including that deposited on the internal walls) and the thoracic and respirable sub-fractions (based on measurements of the mass deposited on the different parts of the filter). A second instrument (see Figure 9.36) (Vincent et al., 1993) is derived directly from the IOM personal inhalable aerosol sampler shown in Figure 9.26. Here, aerosol is again aspirated through a 15 mm circular entry and, as before, the entry forms an integral part of an aerosol-collecting capsule

289

Aerosol science for industrial hygienists

Figure 9.36.

The 21min -1 IOM personal sampler for the simultaneous measurement of exposure to inhalable, thoracic and respirable aerosol: (a) sampler shown assembled, (b) sampler shown dismantled to reveal the porous foam selectors. (From Vincent, J.H., Aerosol Sampling: Science and Practice, Copyright 1989, reproduced by permission of John Wiley and Sons Limited).

290

Aerosol sampling in workplaces which acts as a receptacle for the whole inhalable fraction. Now, however, the capsule is extended in length in order to house two porous polyester foam selectors, each using different grades of foam. The first is chosen (i.e., grade of foam, dimensions) to provide penetration characteristics matching the thoracic fraction. The second selector, placed immediately behind the first, is chosen to provide penetration characteristics matching the respirable aerosol curve. In the practical use of this instrument, the whole capsule is weighed before and after sampling to provide the inhalable mass fraction. Then the second foam plug and the backing filter are removed and weighed separately. The sum of the resultant two masses provides the thoracic mass fraction. The mass on the backing filter is the respirable mass sampled. Again if desired, the respective aerosol catches can be recovered for quantitative analysis (e.g., for the assessment of chemical subfractions). This instrument is still at the prototype stage at the time of writing, so is not yet commercially available.

9.12 A E R O S O L S P E C T R O M E T E R S In principle, if we know the particle size distribution and the mass of the sampled aerosol, then we can determine the particle size distribution and mass contained in any sub-fraction. The process is illustrated fully by Example (8.1) (Chapter 8). Aerosol spectrometers that can provide such information are more versatile than the dedicated samplers described above since they can provide data about any number of sub-fractions from just one sample. This can have important implications, in particular for epidemiological research. For example, the general approach outlined above has been used to examine the effects on the dust uptakes in mineworkers of changing breathing patterns and hence different lung deposition characteristics ~ associated with different work rates (Vincent and Mark, 1984). A wide range of physical possibilities exists upon which to base a family of aerosol spectrometer devices. A number of them, including horizontal elutriators, centrifuges and inertial devices, have been reviewed by Mark et al. (1984). The class of spectrometer which has achieved the greatest popularity since it first emerged in the 1940s is the cascade impactor. In this device, sampled aerosol passes through a succession of impactor stages (as shown schematically in Figure 9.37), where each stage takes the form of a jet directed onto a solid surface (see also Figure 4.15). Particle deposition takes place by impaction onto the surface, strongly dependent on particle aerodynamic size and jet width. Decreasing jet width at each successive stage ensures that smaller and smaller particles are deposited as the aerosol penetrates from stage to stage. From the masses of aerosol collected at each

291

Aerosol science for industrial hygienists Sampled air

I I I An impactor

Impaction

I

stage, comprising a jet-nozzle and

impaction plate

Impaction Impaction

?'~-

t

U//r U////A

Impaction

'

/////'//////',/i

Filter

71

~

To pump

Axis of cylindrical symmetry

I I Figure 9.37.

Schematic to illustrate the principle of o p e r a t i o n of the cascade impactor.

stage, together with knowledge of the deposition (or 'cut') characteristics of the impactor stages, the cumulative ~ and, in turn, the frequency size distribution of the sampled aerosol can be obtained. More detailed information of the principles, performances and types of cascade impactors and on data reduction appears widely elsewhere in the literature (see especially Lodge and Chan, 1986). Whilst such instruments might not be the first choice for routine aerosol exposure assessment, they can play an important role in the non-routine investigations which continue to be a significant part of professional industrial hygiene. Here it is widely accepted that personal samplers provide the most reliable estimates of the true nature and magnitude of the aerosol exposures of individual workers. It follows therefore that, for aerosol spectrometry, personal cascade impactors are the most appropriate.

292

Aerosol sampling in workplaces

Figure 9.38. The 2 1 min-~ Marple personal cascade impactor as described by Rubow et al. (1987) (Grazeby Anderson, Model 298, photograph courtesy of Grazeby Andersen, Smyrna, GA).

Here just two specific instruments are mentioned. The first is the sampler proposed by Marple and his colleagues, shown in Figure 9.38 (Rubow et al., 1986). It is an eight-stage device, with radial slot-shaped jets at each stage where aerosol is collected onto polycarbonate membrane films. By weighing the films before and after sampling, the mass of aerosol collected on each is assessed gravimetrically. Alternatively, the aerosol collected at each stage can, if desired, be recovered and prepared for chemical analysis. The second device is the IOM personal inhalable dust spectrometer (PIDS) shown in Figure 9.39 (Gibson et al., 1987). The general configuration is similar to that for the 'Marple' device, except that the slot jets are replaced by circular ones. 293

Aerosol science for industrial hygienists

Figure 9.39.

The 21min -1 IOM personal inhalable dust spectrometer (PIDS) as described by Gibson et al. (1987).

The aerosol is collected directly onto the back of each disc-shaped aluminum impactor surface which also incorporates the jets for the next stage. All the collection surfaces are greased prior to sampling and the masses of collected aerosol are obtained by weighing each disc before and after. Despite the relatively large tare weight of each stage, modern electronic balances are available that can provide accurate measurement of the collected aerosol. The key feature of the PIDS that distinguishes it from the previous instrument is that it incorporates a 15 mm circular entry like that for the I O M inhalable aerosol sampler. So it begins by aspirating the inhalable fraction. This entry is incorporated into a 'cassette' which, by also weighing before and after sampling, provides the mass of aerosol which is collected between the entry and the first impactor stage. Using this mass together with knowledge of the

294

Aerosol sampling in workplaces 2 -'E

I

alable fraction

E

. _

,

o

90

[Resp'irable[/ lsubfraction[

[ Thoracic '1 Isubfraction]

E 0

1 10

0

10

20

Particle aerodynamic diameter dae(l~m)

Figure 9.40. Illustration of how results for the particle size distribution of the inhalable fraction may be used to provide information about health-related aerosol subfractions (from Gibson, H. et al., Annals of Occupational Hygiene, Copyright 1987, reproduced by permission of the British Occupational Hygiene Society). penetration characteristics of the entry stage, the particle size distribution obtained from the cascade impactor part of the instrument may be corrected to allow for deposition (of both coarse and fine particles) in the entry, thus providing the particle size distribution of the inhalable fraction. From this, the size distribution and mass contained in any subfraction may be determined, as illustrated in Figure 9.40. Thus, the performance of this instrument closely matches the health-related sampling rationale outlined in Chapter 8.

9.13 SAMPLING OF B I O A E R O S O L S As mentioned in Chapter 3, bioaerosols represent a category of particle which is receiving increasing interest from the industrial hygiene standpoint. As for other aerosol types discussed in this book, sampling by aspiration is the most appropriate means by which to conduct occupational exposure assessment.

295

Aerosol science for industrial hygienists

Criteria for bioaerosol sampling It is clear from the review of Lacey and Dutkiewicz (1994) that a very wide range of health effects may occur as a result of exposure to the diverse bioaerosols that might be encountered in occupational environments. These include local effects in the regions of the respiratory tract, including mucous membrane irritation, rhinitis, bronchitis and chronic obstructive airways disease, asthma and extrinsic alveolitis. They also include systemic effects, including organic dust toxic syndrome and infection. From such considerations it follows that the particle size-selective criteria for aerosol sampling which were discussed in detail in Chapter 8 are just as appropriate to biological ~ as well as to inert ~ aerosols. But it appears that, either in the past or currently, this rationale has not been widely recognised or applied in aerobiology. However, the need to make the assessment of exposures of people to bioaerosols in the manner described in the harmonised recommendations of ISO (1992), CEN (1992) and A C G I H (1993-1994) is urged in the recent works of Griffiths and DeCosemo (1994) and Upton et al. (1994).

Sampling Assessment of exposure to bioaerosols involves: the physical aspiration and collection of the particles of interest; survival of the collected particles in their original airborne state; and subsequent microbiological assay. It is reasonable to expect that the first stage of the process is dictated by considerations which are the same as for inert aerosols. That is, aerodynamic diameter is the sole particle property which is relevant. For the second stage, however, the biological properties of the particles become important. More specifically, this involves considerations of viability, survivability and culturability (in subsequent microbiological assay). These refer to the particle's ability to remain in a form which can form a colony on a substrate made up of a suitable nutrient. Factors which can adversely influence such survival include such stresses as dessication (i.e., drying) and exposure of the organisms to oxygen, ozone and air pollutants. The first of these is the most problematical in most practical industrial hygiene situations, and relates to the continued exposure of the collected organisms to the air flow associated with the action of the sampler (Griffiths and DeCosemo, 1994; Henningson and Ahlberg, 1994).

296

Aerosol sampling in workplaces In their review, Griffiths and DeCosemo critically examined the performances of a range of candidate samplers in relation to both the latest particle size-selective criteria and the issues of survivability. Of those mentioned earlier in this chapter in relation to their ability to collect relevant health-related particle size fractions, the 3 1 min -~ IOM static inhalable aerosol sampler (see Figure 9.18), the 2 1 min -1 seven-hole sampler (see Figure 9.22), the 2 1 min -1 IOM personal inhalable aerosol sampler (see Figure 9.26) and the 2 1 min -1 IOM personal inhalable dust spectrometer (PIDS) (see Figure 9.39) were all identified as being potentially useful for bioaerosol sampling. However, no versions exist at present which enable the collected bioaerosol sampler to be retained in a suitable, surviving form. On the other hand, of the many samplers which have been designed specifically as bioaerosol samplers (and, indeed, many of which have been widely used in aerobiology for many years), none has been sufficiently characterised in terms of the new health-related, particle size-selective criteria. Clearly, therefore, this is an important area for future work.

9.14 SAMPLING SYSTEM C O M P O N E N T S In this chapter, attention has been focused largely on the sampling heads which deliver to the collection or sensing region an aerosol fraction which has undergone particle size selection in a manner appropriate to the healthrelated criterion of interest. However, no discussion of aerosol sampling can take place without some reference to the auxiliary components that go to make up the sampling 'system'. Fuller details can be found elsewhere (e.g., Hering, 1989).

Pumps Most samplers require a source of air movement so that particulate-laden air can be aspirated into the instrument. We have discussed both personal and static (or area) sampling, the main difference in terms of pump requirements being the flow rate, which tends to be low for personal sampling (usually 1-4 1 min -1) and larger (up to 100 1 min -1 and even higher) for static sampling. The main limiting factor for a personal sampling pump is its weight, since it must be light enough to be worn on the body (usually on a belt) without inconvenience to the wearer. A wide range of lightweight, battery-powered pumps is available for personal sampling (and also static sampling, if desired). These instruments are based on diaphragm, piston and rotary pumping principles, and most of those in practical use are equipped with damping devices to reduce the

297

Aerosol science for industrial hygienists

effects of flow pulsations. The actual volumetric flow rate will depend first on sampling considerations (e.g., entry conditions to provide the desired performance), and then the amount of material to be collected for accurate assessment, analytical requirements, etc. Internal flowmeters, usually of the rotameter type or digital counters, are incorporated into most pumps, but these must always be calibrated against a primary flow rate standard such as a bubble spirometer. It should also be noted that the flow rate may vary with the resistance imposed by the filter and its collected aerosol mass. For this reason, flow rates should be checked periodically during sampling and adjusted if necessary. However, flow-controlled pumps are now available which eliminate the need for such regular attention during sampling. Finally, for sampling in potentially-exposive atmospheres (e.g., coal mines, chemical plants), intrinsically-safe or flame-proof pumps should be used.

Filters

A filter is the most common means of collecting the aerosol sample in a form suitable for assessment. That assessment might include gravimetric weighing on an analytical balance before and after sampling to obtain the sampled mass. It might also include visual assessment using an optical or electron microscope, and a whole range of analytical and chemical techniques. The choice of filter type for a given application depends greatly on how it is proposed to analyse the collected sample. Many different filter materials, with markedly different physical and chemicial properties, are now available, including fibrous (e.g., glass), membrane (e.g., cellulose nitrate) and sintered (e.g., silver). Membrane filters have the advantage that they can retain particles effectively on their surface (good for microscopy), whereas fibrous filters have the advantage of providing in-depth particle collection and hence a high load-carrying capacity (good for gravimetric assessment). Such filters are available in a range of dimensions (e.g., 25-100 mm in diameter) and pore sizes (e.g., about 0.1-10 p~m). Collection efficiency is usually close to 100% for particles in most size ranges of interest, although sometimes some reduction in efficiency might be traded against the lower pressure drop requirements of a filter with greater pore size. For some types of filter, electrostatic charge can present aerosol collection and handling problems ~ in which case, the use of a static eliminator is recommended. For other types, weight variations due to moisture adsorption can cause difficulty, especially during the gravimetric assessment of low masses. It is therefore recommended that the stabilisation of filters overnight in the laboratory should be carried out before each weighing, together with the use of blank 'control' filters to establish the level of variability. In certain cases, humidity control in the balance room might be recommended. The chemical requirements of filters depends on the nature of the analysis

298

Aerosol sampling in workplaces

which is proposed. As already mentioned, weight stability is important for gravimetric assessment. If particle counting by optical microscopy is required, then the filters used must be capable of being rendered transparent (i.e., cleared). Direct on-filter measurements of mineralogical composition (e.g., by infrared spectrophotometry, X-ray diffraction, scanning electron microscopy, energy-dispersive X-ray analyses, X-ray fluorescence, etc.) are often required. For these, filters must allow good transmission of the radiation in question, with low background scatter or interference. Collected samples may also be extracted from the filter prior to analysis, using a range of wet chemical methods, ultrasonication, ashing, etc., each of which imposes a range of specific filter requirements.

9.15 Q U A N T I T A T I O N OF C O L L E C T E D SAMPLES Determination of aerosol exposure by sampling is not considered complete until the particles which have been collected have been quantified to provide an appropriate concentration. Depending on the particular practical sampling situation, the options include: weighing the collection filter ~ or other collecting surface ~ before and after sampling (with appropriate conditioning of the filter to reduce variability or bias associated with moisture uptake) to provide the total mass of all the collected aerosol; determining the collected content within selected chemical species by means of appropriate analytical instrumentation (after applying appropriate chemical assays to prepare the sample for such analysis); or determining the collected content within selected biological species by means of appropriate microbiological assay. The details of the available analytical techniques and assays are diverse and highly specialised. So they lie outside the scope of this book. But, for the interested reader, they are well documented elsewhere in the literature.

REFERENCES Agarwal, J.K. and Liu, B.Y.H. (1980). A criterion for accurate aerosol sampling in calm air. American Industrial Hygiene Association Journal, 41, 191-197. American Conference of Governmental Industrial Hygienists (ACGIH). (1993-1994). Threshold limit values for chemical substances and physical agents and biological exposure indices. ACGIH, Cincinnati, OH, pp. 42-45.

299

Aerosol science for industrial hygienists Asbestos International Association (AIA) (1979). Airborne asbetsos fibre concentrations at workplaces by light microscopy (membrane filter method). AIA Health and Safety Publication RTM1, AIA, Paris. Beal, S.K. (1970). Deposition of particles from turbulent flow on pipe or channel walls. Nuclear Science and Engineering, 40, 1-11. Belyaev, S.P. and Levin, L.M. (1974). Techniques for collection of representative aerosol samples. Journal of Aerosol Science, 5, 325-338. Brockmann, J.E. (1993). Sampling and transport of aerosols. In: Aerosol Measurement (Eds. K. Willeke and P.A. Baron), Van Nostrand Reinhold, New York, pp. 77-111. Buchan, R.M., Soderholm, S.C. and Tillery, M.I. (1986). Aerosol sampling efficiency of 37 mm cassettes. American Industrial Hygiene Association Journal, 47, 825-831. Cantrell, B.K., Williams, K.L., Watts, W.F. and Jankowski, R.A. (1993). Mine aerosol measurement. In: Aerosol Measurement (Eds. K. Willeke and P.A. Baron), Van Nostrand Reinhold, New York, pp. 591-611. Cherrie, J.W., Jones, A.D. and Johnston, A.M. (1986). The influence of fibre density on the assessment of fibre concentration using the membrane filter method. American Industrial Hygiene Association Journal, 47, 465-474. Comit6 Europ6en de Normalisation (CEN) (1992). Workplace atmospheres: size fraction definitions for measurement of airborne particles in the workplace, CEN Standard EN 481, Brussels. Courbon, P., Wrobel, R. and Fabries, J.F. (1988). A new individual respirable dust sampler: the CIP10. Annals of Occupational Hygiene, 32, 129-143. Chung, I.P. and Dunn-Rankin, D. (1992). Numerical simulation of two-dimensional blunt body sampling in viscous flow. Journal of Aerosol Science, 23, 217-233. Davies, C.N. (1965). The rate of deposition of aerosol particles from turbulent flow through ducts. Annals of Occupational Hygiene, 8, 239-245. Davies, C.N. (1968). The entry of aerosols into sampling tubes and heads. British Journal of Applied Physics, 25, 921-932. Davies, C.N. and Subari, M. (1982). Aspiration above wind velocity of aerosols with thin-walled nozzles facing and at right angles to the wind direction. Journal of Aerosol Science, 13, 59-71. Dunmore, J.H., Hamilton, R.J. and Smith, D.S.G. (1964). An instrument for the sampling of respirable dust for subsequent gravimetric assessment. Journal of Scientific Instruments, 41,669-672. Dunnett, S.J. and Ingham, D.B. (1986). A mathematical theory to two-dimensional blunt body sampling. Journal of Aerosol Science, 17, 839-853. Dunnett, S.J. and Ingham, D.B. (1988). An empirical model for the aspiration efficiencies of blunt aerosol samplers oriented at an angle to the oncoming flow. Aerosol Science and Technology, 8, 245-264. Durham, M.D. and Lundgren, D.A. (1980). Evaluation of aerosol aspiration efficiency as a function of Stokes' number, velocity ratio and nozzle angle. Journal of Aerosol Science, 11, 179-188. Erdal, S. and Esmen, N.A. (1995). Human head model as an aerosol sampler: calculations of aspiration efficiencies for coarse particles using an idealised human head model. Journal of Aerosol Science, 26, 253-272. Friedlander, S.K. and Johnstone, H.F. (1957). Deposition of suspended particles from turbulent gas streams. Industrial and Engineering Chemistry, 49, 1151-1156. Gibson, H., Vincent, J.H. and D. Mark, D. (1987). A personal inspirable dust spectrometer for applications in occupational hygiene research. Annals of Occupational Hygiene, 31, 463-479. Gormley, P.G. and Kennedy, M. (1949). Diffusion from a stream flowing through a cylindrical tube. Proceedings of the Royal Irish Academy, 52A, 163-169.

300

Aerosol sampling in workplaces Griffiths, W.D. and DeCosemo, G.A.L. (1994). The assessment of bioaerosols: a critical review. Journal of Aerosol Science, 25, 1425-1456. Grinshpun, S.A., Willeke, K. and Kalatoor, S. (1993). A general equation for aspiration efficiency by thin-walled sampling probes in calm and moving air. Atmospheric Environment, 28, 375. Hangal, S. and Willeke, K. (1990). Overall efficiency of tubular inlets sampling at 0--90 degrees from horizontal aerosol flows. Atmospheric Environment, 24A, 2379-2386. Henningson, E.W. and Ahlberg, M.S. (1994). Evaluation of microbiological aerosol samplers: a review. Journal of Aerosol Science, 25, 1459-1492. Hering, S.V. (Ed.) (1989). Air Sampling Instruments for Evaluation of Atmospheric Contaminants. American Conference of Governmental Industrial Hygienists, Cincinnati, OH. Higgins, R.I. and Dewell, P. (1968). A gravimetric size-selecting personal dust sampler. British Cast Iron Research Association Report, 908, 112-119. Iles, P.J. and Johnston, A.M. (1983). Problems of asbestos fibre counting in the presence of fibre-fibre and particle-fibre overlap. Annals of Occupational Hygiene, 27, 389--403. Ingham, D.B. (1981). The entrance of airborne particles into a blunt sampling head. Journal of Aerosol Science, 12, 541-549. Ingham, D.B. and Hildyard, M.L. (1991). The fluid flow into a blunt aerosol sampler oriented at an angle to the oncoming flow. Journal of Aerosol Science, 22,235-252. Ingham, D.B. and Wen, X. (1993). Disklike body sampling in a turbulent wind. Journal of Aerosol Science, 24, 629-642. International Standards Organisation (ISO) (1992). Air q u a l i t y - particle size fraction definitions for health-related sampling, ISO CD7708, ISO, Geneva. Johnston, A.M., Jones, A.D. and Vincent, J.H. (1982). The influence of external aerodynamic factors on the measurement of the airborne concentration of asbestos fibres by the membrane filter method. Annals of Occupational Hygiene, 25, 309-316. Johnston, A.M., Vincent, J.H. and Jones, A.D. (1985). Measurements of electric charge for workplace aerosols. Annals of Occupational Hygiene, 29, 271-284. Lacey, J. and Dutkiewicz, J. (1994). Bioaerosols and occupational lung disease. Journal of Aerosol Science, 25, 1371-1404. Lipatov, G.N., Grinshpun, S.A., Shingarov, G.L. and Sutugin, A.G. (1986). Aspiration of coarse aerosol by a thin-walled sampler. Journal of Aerosol Science, 17,763-769. Lipatov, G.N., Grinshpun, S.A., Semenyuk, T.I. and Sutugin, A.G. (1988). Secondary aspiration of aerosol particles into thin-walled nozzles facing the wind. Atmospheric Environment, 22, 1721-1727. Liu, B.Y.H. and Agarwal, J.K. (1974). Experimental observation of aerosol deposition in turbulent flow. Journal of Aerosol Science, 5, 145-155. Liu, B.Y.H. and Ilori, T.A. (1973). Inertial deposition of aerosol particles in turbulent pipe flow. Presented at the ASME Symposium on Flow Studies in Air and Water Pollution, Atlanta, GA (June). Liu, B.Y.H., Pui, D.Y.H., Rubow, K.L. and Szymanski, W.W. (1985). Electrostatic effects in aerosol sampling and filtration. Annals of Occupational Hygiene, 29, 251-269. Liu, B.Y.H., Zhang, Z.Q. and Kuehn, T.H. (1989). A numerical study of inertial errors in anisokinetic sampling. Journal of Aerosol Science, 20, 367-380. Lodge, J.P. and Chan, T.L. (1986). Cascade Impactors: Sampling and Data Analysis. American Industrial Hygiene Association, Akron, OH. Maguire, B.A., Barker, D. and Wake, D. (1973). Size selection characteristics of the cyclone used in the SIMPEDS 7000 Mk 2 gravimetric dust sampler. Staub-Reinhalt Luft, 33, 95-99. Mark, D. (1990). The use of dust collecting cassettes in dust samplers. Annals of Occupational Hygiene, 34, 281-291.

301

Aerosol science for industrial hygienists Mark, D. and Vincent, J.H. (1986). A new personal sampler for airborne total dust in workplaces. Annals of Occupational Hygiene, 30, 89-102. Mark, D., Lyons, C.P., Upton, S.L. and Kenny, L.C. (1994). Wind tunnel testing of the sampling efficiency of personal inhalable aerosol samplers. Journal of Aerosol Science, 25 (Supplement 1), $339-$340. Mark, D., Vincent, J.H., Aitken, R.J., Botham, R.A., Lynch, G., Van Elzakker, B., Van der Meulen, A. amd Zierock, K.-H. (1990). Measurement of suspended particulate matter in the ambient atmosphere. Institute of Occupational Medicine Report TM/90/14, Edinburgh, Scotland, U.K. Mark, D., Vincent, J.H., Gibson, H., Aitken, R.J. and Lynch, G. (1984). The development of an inhalable dust spectrometer. Annals of Occupational Hygiene, 28, 125-143. Mark, D., Vincent, J.H., Gibson, H. and Lynch, G. (1985). A new static sampler for airborne total dust in workplaces. American Industrial Hygiene Association Journal, 46, 127-133. Mark, D., Vincent, J.H. and Witherspoon, W.A. (1982). Particle blow-off: a source of error in blunt dust samplers. Aerosol Science and Technology, 1,463-469. National Institute for Occupational Safety and Health (NIOSH). (1975). Criteria for a recommended standard occupational exposure to cotton dust. DHEW (NIOSH) Publication No. 75-118, USGO, Washington, DC. National Institute for Occupational Safety and Health (NIOSH). (1979). USPHS/NIOSH membrane filter method for evaluating airborne asbestos fibers. Criteria for a recommended s t a n d a r d - occupational exposure to cotton dust. NIOSH Technical Report. Ogden, T.L., Barker, D. and Clayton, M.P. (1983). Flow dependence of the Casella respirable dust cyclone. Annals of Occupational Hygiene, 27, 261-271. Ogden, T.L., Birkett, J.L. and Gibson, H. (1977). Improvements to dust measurement techniques. Institute of Occupational Medicine (Edinburgh, Scotland, U.K.), Techncial Report No. TM/77/11. Okazaki, K., Wiener, R.W. and Willeke, K. (1987). The combined effect of aspiration and transmission on aerosol sampling accuracy for horizontal isoaxial sampling. Atmospheric Environment, 21, 181-185. Prodi, V., Belosi, F. and Mularoni, A. (1986). A personal sampler following ISO recommendations on particle size definitions. Journal of Aerosol Science, 17, 576-581. Rader, D.J. and Marple, V.A. (1988). A study of the effects of anisokinetic sampling. Aerosol Science and Technology, 8, 283-299. Rubow, K.L., Marple, V.A., Loin, J. and McCawley, M.A. (1987). A personal cascade impactor: design, evaluation and calibration. American Industrial Hygiene Assosication Journal, 48, 532-538. Stevens, D.C. (1986). Review of aspiration coefficients of thin-walled sampling nozzles. Journal of Aerosol Science, 17, 729-743. Thomas, J.W. (1967). Particle loss in sampling conduits. In: Proceedings of the Symposium on Assessment of Airborne Radioactivity (Vienna), pp. 727-735. Tsai, P.J. and Vincent, J.H. (1993). Impaction model for the aspiration efficiencies of aerosol samplers at large angles with respect to the wind. Journal of Aerosol Science, 24, 919-928. Tsai, P.J., Vincent, J.H. and Mark, D. (1995). Semi-empirical model for the aspiration efficiencies of personal aerosol samplers of the type widely used in occupational hygiene. Annals of Occupational Hygiene, in press. Tsai, P.J., Vincent, J.H., Mark, D. and Maldonado, G. (1994). Impaction model for the aspiration efficiencies of aerosol samplers in moving air under orientation-averaged conditions. Aerosol Science and Technology, 22, 271-286. Tufto, P.A. and Willeke, K. (1982). Dependence of particulate sampling efficiency on

302

Aerosol sampling in workplaces inlet orientation and flow velocities. American Industrial Hygiene Association Journal, 43, 436-443. Upton, S.L., Mark, D., Douglass, E.J., Hall, D.J. and Griffiths, W.D. (1994). A wind tunnel evaluation of the physical sampling efficiencies of three bioaerosol samplers. Journal of Aerosol Science, 25, 1493-1501. Vincent, J.H. (1987). Recent advances in aspiration theory for thin-walled and blunt aerosol sampling probes. Journal of Aerosol Science, 18, 487-498. Vincent, J.H. (1989). Aerosol Sampling: Science and Practice. John Wiley and Sons, Chichester, UK. Vincent, J.H. and Mark, D. (1984). Inhalable dust spectrometers as versatile samplers for studying dust-related health effects. Annals of Occupational Hygiene, 28, 117-124. Vincent, J.H, Aitken, R.J. and Mark, D. (1993). Porous plastic foam media: penetration characteristics and applications in particle size-selective sampling. Journal of Aerosol Science, 24, 929-944. Vincent, J.H., Emmett, P.C. and Mark, D. (1985). The effects of turbulence on the entry of airborne particles into a blunt dust sampler. Aerosol Science and Technology, 4, 17-29. Vincent, J.H. and Gibson, H. (1981). Sampling errors in blunt dust samplers arising from external wall loss effects. Atmospheric Environment, 15,703-712. Vincent, J.H., Mark, D., Miller, B.G., Armbruster, L. and Ogden, T.L. (1990). Aerosol inhalability at higher windspeeds. Journal of Aerosol Science, 21,577-586. Vincent, J.H., Mark, D., Miller, B.G., Armbruster, L. and Ogden, T.L. (1990). Aerosol inhalability at higher windspeeds. Journal of Aerosol Science, 21,577-586. Vincent, J.H., Stevens, D.C., Mark, D., Marshall, M. and Smith, T.A. (1986). On the aspiration characteristics of large-diameter, thin-walled aerosol sampling probes at yaw orientations with respect to the wind. Journal of Aerosol Science, 17, 211-224. Walton, W.H. (1954). Theory and size classification of airborne dust clouds by elutriation. British Journal of Applied Physics, 5 (Supplement 3), $29-$40. Walton, W.H. and Beckett, S.T. (1977). A microscope eyepiece graticule for the evaluation of fibrous dust. Annals of Occupational Hygiene, 20, 19-23.

303

C H A P T E R 10

Direct-reading monitoring of workplace aerosols 10.1 I N T R O D U C T I O N In most of the sampling instruments described in Chapter 9, the sampled aerosol is collected in a form (e.g., on a filter) after which it may be assessed gravimetrically by weighing on an analytical balance or may be recovered for chemical analysis. Such instrumentation is suitable when time-averaged measurement can be justified. However, there are occasions where shortterm ~ approaching ' r e a l - t i m e ' ~ measurement is required; for example, where: the aerosol in question is thought to be particularly hazardous and where an immediate alert to high concentrations is required; or monitoring is desired in order to examine the effects of adjustments in process or dust control. So whereas time-weighted averaged aerosol concentrations (like those obtained using samplers like those described in Chapter 9) provide information like that shown in Figure 10.1a, there is sometimes a need for information like that shown in Figure 10.lb. There is a range of physical options for direct-reading aerosol detection and measurement, all based on sensing the particles either individually or in ensembles using some sort of transducer which in turn can provide a representative electric signal which can be read out or recorded and referred to some appropriate calibration. The majority of such instruments fall into one of five categories ~ optical, electrical, molecular, mechanical and nuclear. The physical basis for optical measurement was given in Chapter 5. This chapter sets out to extend that discussion and to widen it to include the other types of instrumentation. As before, the emphasis is placed on instrumentation of the type that finds application in industrial hygiene, some as routine instruments for workplace air monitoring and some of the more specialised types which find use in industrial hygiene research.

304

Direct-reading monitoring of workplace aerosols (a)

Time-weighted average exposure concentration mg m -3

(b)

Real-time exposure concentration

mg m -3

0

2

4

6

8

Exposure time (hours)

Figure 10.1. Graph to illustrate a hypothetical working shift-long exposure, and the comparison between (a) time-averaged exposure, and (b) real-time exposure.

10.2 G E N E R A L C H A R A C T E R I S T I C S OF O P T I C A L M O N I T O R I N G Optical techniques provide a means by which aerosol can be assessed in, or close to, real time. They have the great advantage over many types of instrumentation that they provide the opportunity that the measurement can be made remotely and without disturbing the aerosol. Available optical techniques fall into two basic categories. The first approach is based on measurement of the interaction between a light beam and the aerosol as a whole. This is useful for providing aerosol concentration, although there are some practical difficulties since optical instruments do not generally respond directly to aerosol mass concentration which, of course, is usually the index of interest to the industrial hygienist. However, light scattering theory suggests regimes of particle size, light wavelength and refractive index where responses can be achieved which in turn can be related consistently to aerosol mass in specific health-related fractions (see Chapter 5). This approach is well suited to workplaces which are consistently characterised by well-defined aerosol types (e.g., respirable dust in coal mines). It has been achieved for some respirable aerosol monitors by generating a calibration curve from field trials conducted side-by-side with gravimetric respirable aerosol sampling instruments.

305

Aerosol science for industrial hygienists The second approach is based on measurement of the optical response of individual particles. This is useful if what is wanted is aerosol concentration in terms of the numbers of particles falling within defined particle size bands.

Extinction monitoring Optical instruments based on the principles of light extinction are the simplest. For these the basic response is given, from Equations (5.3) and (5.5), by

- exp ( -

cpQX)

(10.1)

/o where, as before in Chapter 5, I and I o are the incident and transmitted beams, respectively, X is the optical path length through the aerosol, Q is the particle extinction coefficient and Cp is the projected area concentration. Inspection of Equation (10.1) reveals that the change in light intensity, and hence the ability to distinguish between I and I o, increases with the product cpX. So, for low concentration, a long path length is desired. In modern workplaces, aerosol concentrations are generally not high enough to provide sensitive extinction measurement for a practicable optical path length (except, possibly, for aerosol tomography as described in Chapter 5). Such difficulties in using light extinction are compounded by the problems in obtaining a readout which reflects the aerosol mass concentration, especially when coarse particles are present. So it is not surprising to find that instruments based on the principles of extinction are not widely used in the monitoring of general workplace aerosols. The few industrial applications that have been reported have been for situations where aerosol concentrations are very high; for example, in the monitoring of aerosol emissions from industrial processes flowing through exhaust chimneys and ducts. Over the years, a number of instruments for this purpose have been commercially available and applied for this purpose. A typical such instrument takes the simple form of a light source placed at the focus of a converging lens, thus producing a parallel beam which is directed onto a photocell located at the other side of the aerosol of interest (see Figure 5.5). The electrical output from the photocell is taken as the indicator of I and hence (knowing Io) the average aerosol concentration along the path of the light beam. Most such commercially-available instruments have not met the important physical requirements identified in Chapter 5, in particular the need to limit the amount of forward-scattered light entering the detector. As a result, they have produced results which are often difficult to interpret. However, the use of such equipment might be acceptable if sound calibration

306

Direct-reading monitoring of workplace aerosols methods are employed (e.g., by comparing the time-averaged outputs with gravimetric samples taken isokinetically with thin-walled probes) and checked regularly. Apart from monitoring in ducts, very little has been reported on the use of extinction-based monitors in the actual workplace environment itself. In the 1950s, one such instrument was reported which had been used in coal mines in the Soviet Union for the monitoring of dust concentrations high enough to constitute an explosion risk (see Hodkinson, 1966), but there appears to have been little else.

Light scattering photometry As indicated in Chapter 5, optical instruments operating on the basis of the detection of the scattered ~ rather than transmitted ~ light are more sensitive at lower concentrations. This is because it is easier to detect a change in a small light intensity against a dark background than to detect a small change in an intensity which is already bright. For this reason, light scattering photometry has been a more popular option for workplace aerosol monitoring. Furthermore, because so much information about the aerosol is contained in the angular distribution of the scattered light, there are many more options on which to base an instrument. Hodkinson (1966) classified light scattering instrumentation according to the angle at which the scattered light was detected, and identified some possibilities for optical aerosol measurement. For example, for particles with geometric size greater than about 2 p~m irradiated by white light, the flux scattered through between 40 ~ and 45 ~ is roughly proportional to the aerosol surface (or projected) area concentration. On the other hand, scattering in the forwards direction for particles in the same size range offers the possibility of obtaining an output which can be related to aerosol volume concentration (and hence, for known substances, mass concentration).

10.3 L I G H T S C A T T E R I N G P H O T O M E T E R S A wide range of instruments has appeared on the market, and a representative selection of those which have found applications in the monitoring of workplace aerosols is described here to illustrate how knowledge of the optical properties of aerosols has been applied in practice. The industrial hygienist is usually primarily interested in the airborne mass concentration of an aerosol. But, as has been stated, the effects of particle size and refractive index on scattering properties are considerable over most particle size ranges. Thus, accurate and consistent measurements of mass concentration from optical instruments can be difficult. This is especially

307

Aerosol science for industrial hygienists

so when large particles are involved which can account for a relatively high proportion of the aerosol mass. For this reason, the most successful practical light scattering photometers have been those designed for the monitoring of aerosol fractions within restricted particle size ranges, in particular, the fine respirable fraction. In Europe, two contrasting approaches have been reported. In Britain, the Safety in Mines Light Scattering Instrument (SIMSLIN) is based on the horizontal elutriator system like that of the MRE Type l13A gravimetric sampler described in Chapter 9. That is, the aerosol is aspirated and passed through the plates of a horizontal elutriator designed to have penetration characteristics matching the B MRC respirable aerosol curve. In the M R E itself, the penetrating aerosol is deposited onto a filter so that, at the end of a sampling shift, a time-weighted average of the respirable aerosol mass concentration can be obtained by direct weighing of the filter. In the corresponding SIMSLIN, the aerodynamically-selected aerosol enters an optical sensing zone in which scattered infrared light from a diode laser ()~ = 0.85 ~m) is collected at an angle 12-20 ~ from the forwards direction and focused onto a photodiode detector. For these conditions, experiments in the laboratory have shown that the scattered light flux (given by the output from the photodiode) is reasonably proportional to particle mass concentration over the respirable range. It was also found that the signal from the photodiode is only weakly dependent on refactive index for the range of mineral dust aerosols expected to be encountered in coal mines. The performance of SIMSLIN in relation to the respirable aerosol mass (as 10o 9

i

A iDifferent mine sites 0 A 0

o -

5

-

/oo

0

Z r~ r~

0

I 5 BMRC respirable (mg m-3)

I I0

F i g u r e 10.2. P e r f o r m a n c e of the S I M S L I N p h o t o m e t e r for respirable dust, shown in relation to respirable aerosol mass as m e a s u r e d according to the B M R C respirable aerosol curve. T h e data are s u m m a r i s e d from results of e x p e r i m e n t s carried out in u n d e r g r o u n d coal mines (from F o r d et al., 1983).

308

Direct-reading monitoring of workplace aerosols defined by the BMRC respirable aerosol selection curve) has been examined in experiments in the workplaces of underground coalmines, and the results summarised in Figure 10.2 show an approximately linear 1"1 relationship (from data reported by Ford et al., 1983). In practical industrial hygiene, this instrument has proved to be very useful for investigating and diagnosing certain underground dust control problems. But it has not found widespread routine use in workers' exposure assessment. Part of the reason for this is that the dust control regulations for British coal mines are written in terms of time-weighted average mass concentrations, requiring full-shift sampling and gravimetric assessment of collected dust. Extension of the SIMSLIN concept led to the development of the Optical Scattering Instantaneous Respirable Dust Indication System (OSIRIS) (Leck, 1983). This instrument was intended to provide the respirable dust concentration data in a form that can be telemetered to a central monitoring station. Again, although substantial research and development resources have been devoted to this instrument, it has not yet been widely deployed in the field. In the German coal mining industry, a somewhat different approach is adopted in the TM-Digital (Armbruster and Breuer, 1983). This instrument is shown schematically in Figure 10.3. Here, the aerosol enters the sensing

Figure 10.3. The TM-Digital respirable mine dust photometer.

309

Aerosol science for industrial hygienists

region as a result of the 'natural' convection in the workplace atmosphere, and so is not aspirated p e r se (i.e., it has no pump and flow control system). So the instrument is, in effect, a 'passive', non-aspirating dust monitor. It is therefore reasonable to assume that the respirable-sized particles entering the selection zone will undergo no significant aerodynamic particle size selection. In the sensing region, scattered light from a parallel beam of monochromatic infrared light (k = 0.94 ~m) is detected at an angle of 70 ~ to the forward-facing direction. The optical response of the TM-Digital is shown in Figure 10.4 as a function of particle aerodynamic diameter for an aerosol typical of that found in mining applications. It is seen to be broadly similar in shape to the respirable dust curves shown in Chapter 8, albeit shifted somewhat towards smaller particle sizes. Performance of the instrument in relation to respirable aerosol mass is summarised in Figure 10.5 (from data reported by Ford et al., 1983). As for the SIMSLIN, the relationship is seen to be quite 1:1 linear and relatively independent of dust type. One drawback with the practical use of the TM-Digital, however, arises from the fact that the instrument's sensing region is ~ unlike SIMSLIN ~ fully exposed to the workplace total aerosol. This means that the optical surfaces tend to become contaminated by the deposition of particles, with the result that performance tends to drift when the instrument is used unattended over long periods. In the USA, the corresponding instrument that has emerged is the Respirable Aerosol Monitor ( R A M ) (MIE Inc., Billerica, MA). This instrument has appeared in a number of forms over the years and has been widely characterised. The instrument shown in Figure 10.6 is a recent fully-automated digital version with built-in data-logging capability (the DataRAM). When used with a cyclone pre-collector (not shown), the

1.0

8 o.5

,

l,

l

0 2 4 6 Particle aerodynamic diameter, d (p.m)

Figure 10.4.

Particle size-selective response of the TM-Digital dust p h o t o m e t e r (based on A r m b r u s t e r and Breuer, 1983).

310

Direct-reading monitoring of workplace aerosols 10 n~ Different mine sites

~9t

ta0

: /o- ~176176 / o - ~~176176

v

O

5

//oo ,,

I

5 B M R C respirable (rag m -3)

.

I 10

Figure 10.5. Performance of the TM-Digital respirable dust photometer, shown in relation to respirable aerosol mass as measured according to the BMRC curve. The data are summarised from results of experiments carried out in underground coal mines (from Ford et al., 1983).

Figure 10.6. The Respirable Aerosol Monitor (photometer). The version shown is the DataRAM (photograph reproduced courtesy of P. Lilienfeld, MIE Inc, Billerica, MA).

311

Aerosol science for industrial hygienists instrument provides direct measurement of the respirable fraction. In this version, aerosol is aspirated with the aid of a pump and then passed through the cyclone which allows the respirable fraction to penetrate to the optical sensing zone. Pulsed infrared light (h = 0.88 I-m) scattered in the angular range 45-90 ~ from the forward direction is collected and detected by a photodiode detector. It is predicted from light scattering theory that the photodiode signal should be approximately proportional to the respirable aerosol mass concentration in the sensing zone of the instrument. The measured performance of an earlier version (the R A M - l ) is given in Figure 10.7, based on the results of laboratory experiments reported by Rubow and Marple (1983). It is shown in relation to respirable dust as defined by the earlier 1968 A C G I H curve, for which concentrations were obtained using a cascade impactor. As for the other respirable dust monitors described above, reasonable 1:I linearity is achieved, relatively independently of aerosol type. An important extension of the R A M is the so-called Mini-RAM shown in Figure 10.8. Like the TM-Digital, it is an essentially 'passive' device, having no pump. Also like the TM-Digital it responds to the respirable aerosol fraction. For practical industrial hygiene applications, this instrument is important in that it has been miniaturised to the point where it can be used as a personal aerosol monitor. Whilst the instruments described above provide a good illustration of how the principles of light scattering can provide useful real-time information about the concentrations of workplace aerosols, the list is far from exhaustive.

10-

~ 9

o

Different mine sites

Arizona road dust

tao

5 O

O0

0

I

5 A C G I H respirable (mg m -3)

I

I0

Figure 10.7. Performance of the RAM-1 photometer (an earlier version of the instrument shown in Figure 10.6), shown in relation to respirable aerosol mass as measured according to the ACGIH curve (1968 version). The data are summarised from Rubow and Marple (1983), based on experiments carried out in a laboratory test chamber.

312

Direct-reading monitoring of workplace aerosols

Figure 10.8. The Mini-RAM portable photometer for respirable aerosol (photograph reproduced courtesy of P. Lilienfeld, MIE Inc, Billerica, MA).

Many such aerosol photometers based on the principles of light scattering photometry have been developed, and a more complete listing is given in the A C G I H handbook on air sampling instruments (Hering, 1989).

10.4 O P T I C A L P A R T I C L E C O U N T E R S The light scattering photometers described in the previous section all base their performances on the angular scattering of light from an ensemble of particles, the signals from which can be related ~ directly or indirectly ~ to a relevant index of concentration. However, with careful design of the sensing zone and choice of sensitive detection instrumentation, the same principles can be extended to enable the detection and assessment of individual particles. If the light scattered from an individual particle can be detected and registered electronically, it can not only be counted but also sized that is, placed into a given size band or 'channel' based on the magnitude of the impulsive electrical signal arising from the detection of the scattered light by the particles. This general principle is illustrated in Figure 10.9. By such means, instruments can be designed capable of providing an overall particle size distribution. Practical optical particle counters fall into two categories; those for the diffraction-dominated light scattered at low angles (low-angle devices), those for the light scattered in the generally forwards direction below 90 ~ (forward-scattering devices), and those for the light scattered at angles close to or beyond 90 ~ (large-angle devices). A large number of such devices are described in the literature. A relatively small number have found applications in industrial hygiene.

313

Aerosol science for industrial hygienists Optical pulses produced by individualparticles Count only particles producing signals bigger than this

A Figure 10.9.

General principle of particle counting and sizing using an optical photometer.

The optical configuration for a typical forward-scattering instrument is shown in Figure 10.10 (not to scale). This has formed the basis of the PC series of particle counting instruments manufactured by Hiac/Royco (Menlo Park, CA), commonly referred to as the 'Royco'. For these instruments, scattered light is collected in the angular range 10-30 ~ A much wider effective scattering angle may be achieved by the use of an elliptical collecting mirror in the manner shown in Figure 10.11 (not to scale). This has formed the basis of another popular series of particle counting instruments, the 'Climet' (Climet Instruments Company, Redlands, CA), with scattered light collected in the angular range 15-105 ~. The measured responses of

Figure 10.10. Optical configuration for a typical forwards-scattering ('Royco'-

type) optical particle counter. 314

Direct-reading monitoring of workplace aerosols two versions of these instruments, the Hiac/Royco 245 and the Climet 208, are shown in Figure 10.12, with the output (typically in millivolts of signal for a single particle passing through the sensing zone) plotted as a function of particle size (based on results published by Mfikynen et al., 1982 and Chen et al., 1984). Here it is seen that, for both instruments, response is a monotonically increasing function of particle size. Since such responses are dependent on the type of particle, it is common practice to calibrate instruments of this type using polystyrene latex (PSL) test spheres of well-defined diameter and density. In this case, therefore, it is important to realise that, in practical use, the instrument provides particle size information which is 'PSL-equivalent'. Fibrous aerosols are of great importance to industrial hygienists because of the severe health risks that can be associated with inhaling such particles. Here, as discussed in Chapters 8 and 9, the criteria and methods for sampling and assessment are distinctly different to those for non-fibrous aerosols. It has long been recognised as a significant challenge to develop a directreading instrument which can distinguish and count fibres in the appropriate 'respirable' size range (NIOSH, 1979) and in the presence of other, nonfibrous particles. The Fibrous Aerosol Monitor (FAM) (MIE Inc., Billerica, MA) is the first such instrument (Lilienfeld et al., 1979). The basic principle

Figure 10.11. Optical configuration for the 'Climet'-type optical particle counter. 315

Aerosol science for industrial hygienists Calibration for non-absorbing particles 10-

10Royco

i

cI

m

|

> E

00.1 -

0.01 0.1

0.1-

l 1

I 10

0.01 0.1

[ 1

I 10

Particle diameter (p.m)

Figure 10.12. Summary of the measured optical responses of the Royco and Climet optical particle counters (based on results reported by Mfikynen, J. et al., Optical particle counters: resolution and counting efficiency. Journal of Aerosol Science, 13, 529-535, Copyright 1982, and by Chen, Y.S., et al., Experimental responses of two optical particle counters. Journal of Aerosol Science, 15, 457-464, Copyright 1984, with kind permission from Elsevier Science Ltd, The Boulevard, Langford Lane, Kidlington OX5 1GB, U.K.).

of operation is that airborne fibrous particles are subjected to a combined unidirectional and oscillating electric field. The larger unidirectional field aligns each fibre and the smaller oscillating component introduces a 'rocking' motion about its axes of alignment. The alignment and the rocking motion are dependent on the applied fields but are independent of fibre dimensions (Lilienfeld, 1985). In the physical arrangement of the instrument, the electric fields are arranged so that the rocking motion takes place in the plane at right angles to an incident light beam (see the optical arrangement shown schematically in Figure 10.13a). Light scattering associated with the rocking motion of the fibre is detected in this plane (Lilienfeld, 1987). It is greater for long fibres and becomes non-existent for the limiting case of spherical particles. Thus the opportunity exists for the selective detection of long fibres (e.g., those meeting the measurement criteria for 'respirable' fibres). The packaged instrument is shown in Figure 10.13b. Its performance capability is illustrated in Figure 10.14, where the scattered light intensity is plotted as a function of the cylinder roll angle + (see Figure 10.13a) for particles of fixed diameter and varying length. From this it is clear that particles can be discriminated according to their length and counted electronically. In practice, however, it has been found that the response of the instrument cannot be predicted theoretically. So it is necessary to calibrate it side-byside with one of the more conventional m e m b r a n e filter-based reference methods.

316

Direct-reading monitoring of workplace aerosols

Figure 10.13. The Fibrous Aerosol Monitor (FAM): (a) optical configuration (from Lilienfeld, P., Light scattering from oscillating fibres at normal incidence. Journal of Aerosol Science, 18, 389-400, Copyright 1987, with kind permission from Elsevier Science Ltd, The Boulevard, Langford Lane, Kidlington OX5 1GB, UK), and (b) photograph of practical instrument (reproduced courtesy of P. Lilienfeld, MIE Inc, Billerica, MA).

317

Aerosol science for industrial hygienists

.,,-4

r~

O

L = 20 ~m

U

o,..

/

-IO

f

~-

10 i.Lm 5 ~m

o Fibre roll angle, 4) (degrees)

I

IO

Figure 10.14. Graph indicating the performance of the FAM in terms of its ability to discriminate between fibres of different length (from Lilienfeld, P., Light scattering from oscillating fibres at normal incidence. Journal of Aerosol Science, 18, 389-400, Copyright 1987, with kind permission from Elsevier Science Ltd, The Boulevard, Langford Lane, Kidlington OX5 1GB, U.K.).

Apart from optical particle counters like those described where particle size is determined directly from the intensity of the scattered light pulse, there are also optically-based particle counting instruments which work on different principles. These include, for example, instruments where operation depends on the dynamic properties of particles as detected using phase-Doppler and imaging techniques (as reviewed by Rader and O'Hern, 1993). One instrument which is beginning to find increasing use in industrial hygiene, especially industrial hygiene research, is the 'Aerodynamic Particle Sizer R' (APS) (TSI Inc., St. Paul, MN). This was first proposed by Wilson and Liu (1980). The principle of operation of this instrument is shown schematically in Figure 10.15 and the instrument itself in Figure 10.16. The particles are introduced into the sensing zone through an acceleration nozzle, and each one is detected optically and its velocity relative to the surrounding air jet determined by laser Doppler velocimetry. The difference in velocity between the particle and the air is a direct measure of the particle's ability to respond to changes in the motion of the surrounding air, and so may be related directly to particle aerodynamic diameter. Since this index of particle size is highly relevant to particle inhalation and deposition in the human respiratory tract (see Chapter 6) as well as to aerosol sampling (see Chapter 9), this therefore links up strongly with the needs of industrial hygiene aerosol measurement. Baron et al. (1993) have provided a concise review of the

318

Direct-reading monitoring of workplace aerosols Aerosol in I Outer nozzle ~, ~ ~ . J 5 Imin-' ~'::.'*'.] Filter Inner nozzle ___~2, ~.~ ~ ["'] Flowmeter 1 1 min-I '. ":'"-. -..I'2Fr:T'Ng-"~ I x ~i":':~--~ ~

n

cus" g optics Laser

"" [7

"

'

1 ~

Shear4r ,valve nozzle min-I

Photomuitiplier

1

Pressure transducer

"Nx

If ~Filter I V Flowmeter [ I

I

II

I

Internal vacuum .... L.___ pump I

Figure 10.15. Schematic of the Aerodynamic Particle Sizer a (APS) (TSI Model 3310, schematic diagram courtesy of TSI Inc, St. Paul, MN).

Figure 10.16. Photograph of the Aerodynamic Particle Sizer R (APS) (TSI Model 3310A, photograph courtesy of TSI Inc, St. Paul, MN).

319

Aerosol science for industrial hygienists

APS, as well as other devices with related principles of operation such as the 'Aerosizer' (Amherst Process Instruments, Amherst, MA) (e.g., Dahneke, 1973) and the 'Electric Single Particle Aerodynamic Relaxation Time Analyser' (E-SPART) (e.g., Renninger et al., 1981). Although the optical response, reflecting the light-particle interaction, is a primary performance feature of optical particle counting instruments like those described, there are other important performance parameters that need to be taken into account. Firstly there is the question of defining the sensing zone into which the particles of interest may enter and from which the scattered light is received by the detector. This is determined by the design of the collector/detector optics. Here the problem lies in making the sensing volume large enough so that there is always a good chance of finding a particle there, but not so large that there will be more than one particle at any given time. This may be achieved firstly by constraining the aerosol flow as a thin particle beam in a sheath of clean air and then ensuring that the light-collecting optics receives light from a region which is small enough to contain only one particle. One difficulty is that, if more than one particle is present in the sensing zone at the ~ame time, they will be counted and sized together and recorded as if they were a single, larger particle. Thus, in some practical situations, aerosol concentrations may be great enough that, even at low sampling flow rates, such 'coincidences' may preclude the use of a particular instrument. In some instruments (e.g., for the APS), dilution of the sampled aerosol has been employed in order to alleviate this problem.

10.5 E L E C T R I C A L P A R T I C L E M E A S U R E M E N T The diverse physical properties of aerosols enable a range of further options for particle detection and characterisation. The electrostatic charge carried by aerosol particles, and which can be placed on particles in a controlled way through corona discharge techniques, provides a powerful alternative to the optical approaches discussed above. There are two primary objectives of electrical measurement, the electric charge distribution in an aerosol and the particle size distribution. The particle charge distribution is important in the industrial hygiene context since the charges carried by workplace aerosols may have a significant influence on the lung deposition of inhaled particles and on the performances of sampling and control devices (as reviewed by Vincent, 1986), it having been shown that relatively freshly-generated aerosols in workplace atmospheres are charged to levels significantly above that corresponding to Boltzmann equilibrium (Johnston et al., 1985). Those workplace studies were carried out using a 'split-flow electrostatic elutriator' of the type shown in Figure 10.17a (Johnston, 1983). In this device, particles are sampled into a rectangular duct between a pair of conducting plates, and are transported towards the upper or 320

Direct-reading monitoring of workplace aerosols lower plate d e p e n d i n g on the m a g n i t u d e and polarity of the charge carried and on the applied voltage between the plates. Particles of p r e - d e t e r m i n e d size are detected (using an optical particle counter set to the a p p r o p r i a t e particle size channel) at the exits of the upper and lower halves of the duct, respectively, for a range of applied voltages. By analysis of the two ' p e n e t r a t i o n ' curves using the ' m e t h o d of tangents' which has been described earlier for a simple electrostatic elutriator by H u r d and Mullins (1962) and Vincent et al. (1981), the distribution and m a g n i t u d e of the charge on the sampled aerosol m a y be determined. The p r o c e d u r e with the p r o t o t y p e apparatus used was found to

Figure 10.17. The 'split-flow' electrostatic elutriator: (a) schematic of the basic electrical arrangement, and (b) drawing of a fully-aut0mated version (where A is the parallel plate device, B is an Aerodynamic Particle Sizer (APS), C is a high voltage supply, D is a microcomputer, and E is a printer) (from Wake, D. et al. 1991, Crown Copyright is reproduced by permission of the Controller of HMSO).

321

Aerosol science for industrial hygienists be effective and reproducible. But it is rather laborious. This prompted Wake et al. (1991) to develop the fully-automated version shown in Figure 10.17b. However, at present, no commercial versions of this instrument have been developed. Hochrainer (1985) suggested that it is better to use a device where the aerosol is introduced into the inter-electrode space through an entrance slit whose location is well defined, thus rendering the 'method of tangents' redundant and so providing a direct measure of the electrical mobility of the particles (and hence their charge). Such instruments are referred to as electrical mobility spectrometers. Two versions have been developed, both based on cylindrical geometry. The first is the electrical aerosol analyser ( E A A ) shown schematically in Figure 10.18a. Here all the aerosol entering through the narrow annular slit at the top of the inter-electrode space and which is not deposited by electrical forces on either electrode passes through to a particle counter. The second is the differential mobility analyser ( D M A ) , shown schematically in Figure 10.18b, in which all those particles are detected which arrive at the central electrode exactly at the location of the exit slit

Clean air Clean air ,ll

~,

. ~

.

er

Charger

.:::.:...-:.:.:, 9

9.

-.:: ....:..:.. :

sampled

:..lliH..-i ..1l ..-. High voltage supply

~ [

{:.:.:

aerosol

/.~::{.

~ii-"~i'!"

Exit slit

supply

Excess

I:- 9 .'7-

Clean air

...

air

~..;:.:::: counter

To particle counter

(a)

(b)

10.18. Schematics to show the operation of two electrical mobility spectrometers: (a) the Electrical Aerosol Analyser (EAA), and (b) the Differential Mobility Analyser (DMA).

Figure

322

Direct-reading monitoring of workplace aerosols shown in the diagram. For both the E A A and the D MA, the electrical mobility distribution of the aerosol entering both instruments is provided directly from the curve of particle counts versus applied voltage. As for the electrostatic elutriator, the determination of the distribution of particle charge requires independent measurement of particle size. Conversely, however, if the charge on particles of given size is predetermined (for example by contact with air ions under controlled charging conditions), then either of the mobility spectrometers described can be used to determine the particle size distribution. Devices currently available commercially (e.g., from TSI Inc., St. Paul, MN) can provide this information only for very fine particles in the size range 0.01-1 Ixm.

10.6 C O N D E N S A T I O N NUCLEI P A R T I C L E C O U N T E R S (CNC or CPC) Instruments like those described in Section 10.5 require the detection of ultrafine particles. This has frequently involved another class of particle counters whose operation depends on the condensation of liquid vapour onto the particles acting as nuclei. This follows the theory of heterogeneous nucleation as outlined in Chapter 2 and particle growth as outlined in Chapter 3. The principle of operation of the condensation nuclei counter (CNC) ~ sometimes

Filter

~=~]P

'~ f~$[ Ill m

r

i

|

.

Filter

~

..-

Flowmeter 300 cm3/min 1

Pump exhaust 1500 cm3/min "n

/ o -tios I !

I I I I Cooled c o n d e n s e r 9 o

Vacuum pump

....... ]

Optional make-up air 0 or 1200 cm3/min "1

Heated saturator Optional bypass flow

Sample inlet 300 or 1500 cm3/min "n

Liquid-soaked felt

Figure 10.19. Schematic to show the operation of the Condensation Nuclei Counter (CNC) (TSI Model 3022A, schematic diagram courtesy of TSI Inc, St. Paul, MN). 323

Aerosol science for industrial hygienists

called the condensation particle counter (CPC) ~ involves firstly the sampling of aerosols and their introduction into a region which is saturated with water or some other appropriate substance (e.g., alcohol). The atmosphere is then made to supersaturate, preferably by convective cooling since supersaturation by expansion will require a non-uniform flow, usually considered undesirable from the sampling point of view. This causes molecules from the vapour phase to condense onto the small particles. The particles then grow to become large enough for detection by conventional optical particle counting techniques. A typical such device shown schematically in Figure 10.19 has been described by Agarwal and Sem (1980) and is now available commercially (TSI Inc, St. Paul, MN). It is widely used in fine-particle aerosol research. One particularly interesting version of this device, the TSI P O R T A C O U N T R, is frequently employed by industrial hygienists. As shown in Figure 10.20, this instrument is just about small enough to be used as a personal sampler for very fine particles. But perhaps its most common use is as a detector for personal respirator fit testing.

Figure 10.20. The PORTACOUNTR, a portable version of a CNC of the type shown in Figure 10.19 (photograph courtesy of TSI Inc, St. Paul, MN). 324

Direct-reading monitoring of workplace aerosols 10.7 M E C H A N I C A L A E R O S O L MASS M O N I T O R S For most of the direct-reading instruments described so far, the physics of the particle detection process is such that mass concentration of the aerosol of interest is usually not provided directly and unambiguously. This means that they may have somewhat limited application for routine industrial hygiene use. There are, however, some further direct-reading options which, in principle at least, can directly provide the mass concentration of the aerosol. These are generally referred to as 'mass balances', and share the common feature that the aerosol of interest must first be sampled and collected efficiently onto an appropriate surface. One such device is the piezoelectric mass balance, shown schematically in Figure 10.21. The main sensor is the piezoelectric crystalline material (e.g., so-called 'AT-cut' quartz). When an oscillating (or ac) potential is applied between the coated conducting surfaces placed on each face of the crystal, the crystal vibrates mechanically in its transverse mode. The resonant frequency for such mechanical oscillations is a strong function of the mass of the crystal (as, by analogy, for any physical spring-mass system). So any change in effective mass of the crystal will produce a change in resonant frequency of the form

Af = kf2Am

(10.2)

R

~/' "" 9 ..... U / / / / / / / ~ / / ~

Sampled" m . ' -

aroo, "

9

9

"

"

~

9

". "H "'V.//~

",

9

9

]

N:Z 9

I

9

High voltage

supply

llll~

Piezoelectric crystal

/

/I /

/

Mode of transverse vibration Figure 10.21.

A' [ I

" .... t~lectronlcs

i

t

] / i

iI

L n ---

I

Schematic of the piezoelectric mass b a l a n c e system.

325

Aerosol science for industrial hygienists where Af is the change in frequency from its value at f corresponding to a change in mass Am, and where k is a coefficient which describes the sensitivity of mass detection and depends on the type of crystal, its geometry and size, etc. A summary of the piezoelectric mass balance concept and its applications has been given by Ward and Buttry (1990). From Equation (10.2), it can be seen that if Af and f can be detected and measured in the external driving circuitry, then Am may be obtained, provided that k is determined by means of appropriate calibration. Such a change in mass can occur as the result of aerosol particles which are deposited on the vibrating surface of the crystal. Therefore it follows that monitoring Af over a short time interval will provide information about corresponding changes in the mass present on the surface, and hence on the sampled aerosol concentration. So an instrument based on the piezoelectric mass balance concept can be seen as useful for providing direct-reading information (in close-to-real-time) about aerosol concentration. With all this in mind, an important part of a piezoelectric mass balance system is the efficient sampling and deposition of the particles. Electrostatic precipitation (see Chapter 11) has been shown to be effective for this purpose. In some versions of the device, impaction has also been employed. There are some basic practical limitations associated with this type of instrument. Firstly, the accurate detection of mass requires that the deposited particles are rigidly attached to the active crystal surface and are able to remain so during the rapid accelerations that are experienced during the mechanical oscillation of the crystal. Although this is usually the case for very fine particles for which the short-range adhesion forces are large, it may not be so for large particles (e.g., greater than a few micrometres). For the latter, particles may be less well detected, in which case significant undersampling of mass can occur. Secondly, if the crystal becomes heavily loaded, the response of the crystal as given by Equation (10.2) may become non-linear and unpredictable. These and other difficulties such as particle size effects, sampling losses, etc. have been widely reported (see Lundgren et al., 1976, and others). A related device is the Tapered Element Oscillating Microbalance (TEOM R) (first reported by Patashnick and Hemenway, 1969). Now the oscillating element is a tapered hollow glass tube (see Figure 10.22). One end of the tube is anchored, while the other end supports a filter through which the sampled air is drawn and on which the particulate material is deposited. Again, as the mass on the filter ~ and hence on the tapered element ~ increases, the frequency of electromechanically-generated oscillation decreases, and the rate of change can be related to the mass rate at which the aerosol is being sampled (and hence to aerosol concentration). This device has been considered for close-to-real-time sampling of respirable dust in mines. It is a promising technique provided that attention is given to limitations similar to those reported for the piezoelectric mass balance.

326

Direct-reading monitoring of workplace aerosols

Figure 10.22.

Schematic of the Tapered Element Oscillating Microbalance (TEOMR).

10.8 N U C L E A R MASS D E T E C T O R S Beta attenuation provides yet another alternative for the determination of deposited mass. Negatively-charged beta particles are emitted during the radioactive decay of isotopes such as ~4C or 147pm, and may interact with and so be attenuated by matter. The physical nature of the interaction between a beta particle and an individual atom in the attenuating medium is such that the efficiency (i.e., the 'cross-section') of the interaction is proportional to the ratio between the atomic number (i.e., the number of protons in the nucleus) and the atomic weight for the substance in question. This ratio does not vary much between the elements. So, to a good approximation, this means that the interaction relates uniquely to mass. For a beam of beta particles passing through an attenuating medium, the intensity falls according to the familiar exponential law, this time in the form I = Ioexp(-ixX)

327

(10.3)

Aerosol science for industrial hygienists

Detector ................ ~.I .............. j ~',////////////////////////A ~I0

Deposited particles "*"'~ Substrate/coilecting

surface

~-source

Figure 10.23. Basis of the operation of an aerosol monitor working on the beta attenuation principle.

where Ix here is the mass absorption coefficient (e.g., in cm 2 g-l, the calibration coefficient for an instrument based on this principle) and X is the mass thickness (g cm -2, directly proportional to the mass present). In practice, the particles need to be deposited onto a collecting surface such as a filter or an impaction surface. Therefore, it is necessary to take into account the attenuation of the beta radiation not just by the collected particulate material itself but also by the collection substrate. This means that, in a direct-reading instrument, means have to be developed to remove the contribution of the latter, for example, by alternately scanning areas of clean and mass-deposited substrate (e.g., as described by Macias and Husar, 1970 and Vincent et al., 1982). Although from the preceding it might appear feasible to determine the mass directly from measurements of I and Io, the coefficient Ix contains a number of complicating factors, including the geometry of the beta source and detection 'optics'. So careful calibration against known gravimetrically-assessed reference samples needs to be carried out for each practical instrument based on the beta attenuation principle. The basis of a typical practical measurement system is shown schematically in Figure 10.23, and several practical instruments operating on this principle have been built and used in the industrial hygiene setting.

10.9 O V E R V I E W Direct-reading aerosol instrumentation very definitely has a place in the industrial hygienist's repertoire of tools and methods. However, here perhaps more so than for other measurement methods, where the scientific bases

328

Direct-reading monitoring o f workplace aerosols

for the m e t h o d s available are usually so complex, considerable caution is r e c o m m e n d e d in choosing an instrument to p e r f o r m a particular task. T h e r e are m a n y potential traps and pit-falls. Most of the direct-reading instruments described here detect the presence and quantity of airborne particles within a p p r o p r i a t e size ranges. H o w e v e r , none appears capable of also identifying particle species in the aerosol of interest. This appears to be a very i m p o r t a n t area for future work, and some basic research leading to the d e v e l o p m e n t of direct-reading instruments for aerosol chemistry (e.g., Carson et al., 1995) and for detecting the presence of certain bioaerosols (e.g., Evans et al., 1994) is currently u n d e r way and looking very promising. In addition, b e y o n d such particle characterisation by species, direct-reading i n s t r u m e n t a t i o n which can provide quantitative information a b o u t particle shape is now b e c o m i n g available (e.g., Clarke et al., 1994).

REFERENCES Agarwal, J.K. and Sem, G.J. (1980). Continuous flow, single particle-counting condensation nucleus counter. Journal of Aerosol Science, 11,343-357. Armbruster, L. and Breuer, H. (1983). Dust monitoring and the principle of on-line dust control. In: Aerosols in the Mining and Industrial Work Environments (Eds. V.A. Marple and B.Y.H. Liu). Ann Arbor Science Publishers, Ann Arbor, MI, pp. 689-699. Baron, P.A., Mazumder, M.K. and Cheng, Y.S. (1993). Direct-reading techniques using optical particle detection. In: Aerosol Measurement (Eds. K. Willeke and P.A. Baron), Van Nostrand Reinhold, New York, pp. 381--409. Carson, P.G., Neubauer, K.R., Johnston, M.V. and Wexler, A.S. (1995). On-line chemical analysis of single aerosol particles by rapid single-particle mass spectrometry. Journal of Aerosol Science, 26, 535-545. Chen, B.T., Cheng, Y.S. and Yeh, H.C. (1984). Experimental responses of two optical partical counters. Journal of Aerosol Science, 15, 457-464. Clark, J.M., Reid, K., Burke, J.S. and Shakeshaft, D. (1994). A performance assessment of a real-time instrument for the size and shape analysis of airborne particles. In: Proceedings of the Fourth International Aerosol Conference (Ed. R.C. Flagan) held in Los Angeles in August 1994, American Association for Aerosol Research, Cincinnati, OH, pp. 1059-1060. Dahneke, B. (1973). Aerosol beam spectrometry. Nature Physical Science, 244, 54-55. Evans, B.M.T., Yee, E., Roy. G and Ho, J. (1994). Remote detection and mapping of bioaerosols. Journal of Aerosol Science, 25, 1549-1566. Ford, V.H.W., Minton, R. and Mark, D. (1983). Comparative trials in coal mines of the TM-Digital and SIMSLIN dust monitors. In: Aerosols in the Mining and Industrial Work Environments (Eds. V.A,. Marple and B.Y.H. Liu). Ann Arbor Science Publishers, Ann Arbor, MI, pp. 759-775. Hering, S.V. (ed.) (1989). Air Sampling Instruments, 7th Edn. American Conference of Governmental Industrial Hygienists (ACGIH), Cincinnati, OH. Hochrainer, D. (1985). Measurement methods for electric charges on aerosols. Annals of Occupational Hygiene, 29, 241-249.

329

Aerosol science for industrial hygienists Hodkinson, J.R. (1966). The optical measurement of aerosols. In: Aerosol Science (Ed. C.N. Davies), Academic Press, London, pp. 287-357. Hurd, F.K. and Mullins, J.C. (1962). Aerosol size distributions from ion mobility. Journal of Colloid Science, 17, 91-100. Johnston, A.M. (1983). A semi-automatic method for the assessment of electric charge carried by airborne dust. Journal of Aerosol Science, 14, 643-655. Johnston, A.M., Vincent, J.H. and Jones, A.D. (1985). Measurements of electric charge for workplace aerosols. Annals of Occupational Hygiene, 29, 271-284. Leck, M.J. (1983). Optical scattering instantaneous respirable dust indication system. In: Aerosols in the Mining and Industrial Work Environments (Eds. V.A. Marple and B.Y.H. Liu). Ann Arbor Science Publishers, Ann Arbor, MI, pp. 701-717. Lilienfeld, P. (1985). Rotational electrodynamics of airborne fibers. Journal of Aerosol Science, 16, 315-322. Lilienfeld, P. (1987). Light scattering from oscillating fibres at normal incidence. Journal of Aerosol Science, 18, 389-400. Lilienfeld, P., Elterman, P. and Baron, P. (1979). The development of a prototype fibrous aerosol monitor. American Industrial Hygiene Association Journal, 40, 270-282. Lundgren, D.A., Cater, L.D. and Daley, P.S. (1976). Aerosol mass measurement using piezoelectric crystal sensors. In: Fine Particles (Ed., B.Y.H. Liu). Academic Press, New York, pp. 485-510. Macias, E.S. and Husar, R.B. (1970). High resolution on-line aerosol mass measurement by the beta attenuation technique. In: Proceedings of the 2nd International Conference on Nuclear Methods in Environmental Research (Eds. J.R. Vogt and W. Meyer). CONF-740701, p. 413. Mfikynen, J., Hakulinen, J., Kivisto, T. and I~ehtim~iki, M. (1982). Optical particle counters: resolution and counting efficiency. Journal of Aerosol Science, 13, 529-535. National Institute for Occupational Safety and Health (NIOSH). (1979). USPHS/NIOSH membrane filter method for evaluating airborne asbestos fibers. Criteria for a recommended s t a n d a r d - occupational exposure to cotton dust. NIOSH Technical Report. Patashnick, H. and Hemenway, C.L. (1969). Oscillating fibre microbalance. Review of Scientific Instruments, 400, 1008-1011. Rader, D.J. and O'Hern, T.J. (1993). Optical direct-reading techniques: in situ sensing. In: Aerosol Measurement (Eds. K. Willeke and P.A. Baron). Van Nostrand Reinhold, New York, pp. 345-380. Renninger, R.G., Mazumder, M.K. and Testerman, M.K. (1981). Particle sizing by electrical single particle aerodynamic relaxation time analyser. Review of Scientific Instruments, 52, 242. Rubow, K.L. and Marple, V.A. (1983). Instrument evaluation chamber: calibration of commercial photometers. In: Aerosols in the Mining and Industrial Work Environments (Eds. V.A,. Marple and B.Y.H. Liu). Ann Arbor Science Publishers, Ann Arbor, MI, pp. 777-795. Vincent, J.H. (1986). Industrial hygiene implications of the static electrification of workplace aerosols. Journal of Electrostatics, 18, 113-145. Vincent, J.H., Johnston, W.B., Jones, A.D. and Johnston, A.M. (1981). Static electrification of airborne asbestos: a study of its causes, assessment and effects on deposition in the lungs of rats. American Industrial Hygiene Association Journal, 42,711-721. Vincent, J.H., Mark, D., Gibson, H., Aitken, R.J., Bothyam, R.A. and Lynch, G. (1982). Development of a portable gravimetric dust spectrometer. Technical Memorandum No. TM/82/15, Institute of Occupational Medicine, Edinbrugh, Scotland, U.K. Wake, D., Thorpe, A., Bostock, G.J., Davies, J.K.W. and Brown, R.C. (1991). Apparatus for measurement of the electrical mobility of aerosol particles: computer control and data analysis. Journal of Aerosol Science, 22, 901-916.

330

Direct-reading monitoring of workplace aerosols Ward, M.D. and Buttry, D.A. (1990). In situ interracial mass detection with pizeoelectric transducers. Science, 249, 1000-1007. Wilson, J.C. and Liu, B.Y.H. (1980). Aerodynamic particle size measurement by laserDoppler velocimetry. Journal of Aerosol Science, 11, 139-150.

331

C H A P T E R 11

Control of workplace aerosols 11.1 I N T R O D U C T I O N Ultimately, the reduction of risk to workers associated with exposure to aerosol involves technical control measures. Inevitably, it is one of the jobs of the industrial hygienist to become involved in this process, liaising with the engineers who install and commission the necessary equipment and, afterwards, monitoring the workplace environment in order to check that the equipment is operating effectively. A number of technical approaches are available, including: Reducing the 'aerosolisability' (or, where appropriate, the dustiness) of the working material used in the industrial process in question; Modifying the industrial process such that the aerosol generation process is rendered less effective; General exhaust ventilation (GEV) to achieve sufficiently rapid turnover of the workplace air and so achieve effective dilution of airborne contaminants; Local exhaust ventilation (LEV) to remove the airborne contaminant directly at the source so that it does not become dispersed into the general workplace atmosphere; Effective transport of particulate material following its exhaust from the working environment; Separation of the particles from the air extracted by ventilation; Containment of the aerosol generating process; Localised aerosol suppression; As a last resort, the occupational hygienist might recommend: personal protective measures in the form of respiratory filtration equipment. 332

Control of workplace aerosols It is clear that ventilation is a major component, and this is applicable not only to aerosols but also to the control of contaminant gases and vapours. The general problem of ventilation is a broad field in its own right. Much of it falls outside the immediate scope of this book and is fully covered elsewhere (e.g., Burgess et al., 1989). But for aerosols, after air has been removed from the worksite by general or local exhaust ventilation, it is usually necessary to separate the particles from the air before the air can be discharged to outdoors or re-circulated to the workplace atmosphere. For this there are a wide range of physical air cleaning options which can be applied, and much of this chapter will be devoted to reviewing their underlying principles.

11.2 ADJUSTMENTS TO I N D U S T R I A L PROCESSES The first consideration in relation to the control of worker exposure to aerosols relates to the processes by which aerosols are generated in the first place. This involves both the industrial process as well as the materials which are used in that process. In Chapter 3, we discussed the question of 'dustiness', and how quantitative indices can be obtained by which materials may be ranked according to their dustiness. For industrial processes involving the use of dusty materials, the first control option is, where possible, to change to an appropriate alternative material which has a lower dustiness index. But for industrial processes where the material which is producing the aerosol is itself the primary industrial medium, this is not a real option. Then other options need to be explored, and there are a range of possibilities; for example: Modification of the mechanical process by which the material is being worked (e.g., in minerals extraction by the appropriate design of picks, speed of rotation of cutting heads, etc.); Reduction of the degree of agitation of bulk material (e.g., in powder handling by avoiding vibration and unnecessary drafts, reducing the 'roughness' of conveying, avoiding or minimising drops, etc.); Reduction of the number of friction points where abrasion of working materials can take place (e.g., in textiles manufacturing by modifying the paths of carded, spun, wound and woven material, etc.); The use of additives to reduce dustiness (e.g., water or other liquid media as used in mineral extraction, including both spraying onto working surfaces and infusion into rock strata, etc.); and Enclosure (or partial enclosure) of the industrial process (e.g., in processing particularly hazardous substances such a nuclear materials, asbestos, etc.);

333

Aerosol science for industrial hygienists and so on. Most of these adjustments are largely empirical, based on experience and on trial and error. In the mineral extraction industries, for example, there has been a large amount of engineering research to reduce the aerosolising capability at various stages of the winning, conveying and preparation processes by combinations of the above (e.g., Commission of European Communities (CEC), 1982).

11.3 B E H A V I O U R OF A E R O S O L S IN T H E W O R K P L A C E ATMOSPHERE After the aerosol has become airborne and dispersed, the behaviour of the particles in the macroscopic aerodynamic environment of the workplace atmosphere becomes important. Here, transport of particles is governed by many of the aerosol mechanical processes already referred to, including gravitational settling, inertia, Brownian and turbulent diffusion, thermophoresis and electrostatic forces. Particles are removed from the room air by such mechanisms and, in the absence of ventilation, the aerosol concentration that can build up depends on the balance between them and the rate of generation. For workplaces, the 'perfectly-mixed' scenario is an idealised but nonetheless appropriate ~ simple model to use as a starting point towards gaining some useful insight. For this, first consider aerosol in an unventilated room of horizontal cross-sectional area A and volume V (see Figure 11.1). Particles with aerodynamic diameter dae fall under the influence of gravity with a terminal velocity, v~. This motion is superimposed on the random motions associated with the mixing. The build-up of the concentration of particles of size dae as a function of time is governed by the expression /

Aerosol source, strength 13

I

iiiil.i-iiiiii!i'!ii Vo,m. ili!iiiil i! !iiiiii'iiiii!!ii'!i!ii:it iii!ii.liii.!i:!i!ii!:i} ! :i!i!ii.i!]il i!.iii :." i::i-i :..i ;.: " .:"~'~":." !:: ~:..i

t

t

.-'... 9 "'..-.':'" . ." . ' : ' . . "~...-.i~-'.'.- .':'.' 9. " . : ~Jl,~i.: ~ _ - .

.... 9 ...........;,.. :,,/...-'.--..:.. ~iii)i!-~~"~~... ": : :: '"":: :.~<'..* ."i."..;.-; ?".1/" "

/

Volume A

9

Figure 11.1. Schematic on which to base development of a model for particle loss b y g r a v i t a t i o n a l s e t t l i n g f r o m a n u n v e n t i l a t e d , p e r f e c t l y - m i x e d r o o m .

334

Control of workplace aerosols Vdc = f5 d t - vscAdt

(11.1)

where 13 is the number of particles released into the room per unit time and c is the uniform particle concentration at time t. Here it is assumed that, because of the hypothetical perfect mixing, the rate of transport of particles to the floor by gravitational settling is proportional to the settling velocity, to the concentration in the room as a whole, and to the area of the floor. Starting with a zero particles in the room at time t = 0, solution of this linear first-order differential equation shows that the particle concentration rises exponentially according to

c =

(0){

1 - exp

(vsm)}

vsA

(11.2)

V

where the time constant for an equilibrium aerosol concentration to be established is V/v~A and the equilibrium concentration is

(11.3)

Cequi I ~-

vsA In the absence of any externally-induced air movement, the mixing required for this model may be assumed to be derived from the thermal currents, drafts and diffusion which might be present. A more likely situation, however, will be one where air is being removed from the room by the action of general exhaust ventilation (GEV). With this mode of ventilation, the engineering objective is to provide dilution of the airborne contaminant in the workplace environment. This dilution can be included in the above model by the inclusion of an additional term. If the ventilation volumetric flow rate is Qvent, the modified version of Equation (11.1) is

Vdc - f5 d t - vscAdt - Qventcdt (11.4)

fSdt- (vsA + Qvent)cdt for which the solution shows that the equilibrium particle concentration is now

(11.5)

Cequi 1 = ( v s A -t--O v e n t )

335

Aerosol science for industrial hygienists Example 11.1. For a room with floor area 100 m 2 and for airborne particles with to 10 I~m, estimate the ventilation flow rate for which aerosol removal by the dilution ventilation is more dominant than gravitational settling.

dae equal

In Equation (4.17) for the falling speed of a particle, assume that Stokes' law applies and that the slip correction may be neglected. Then for the particles in this example, (10-5) 2 [m 2] X 103 [kg m -3] • 9.81 [m s-2] VS -'-

18 x 18x10 - 6 [ N s m -z] = 0.003 m s-1 From Equation (11.5), dilution ventilation dominates if Qvent > > vsA

or, say, if Qvent > 10 vsA

It is estimated that dilution ventilation dominates if Qvent > 3 m3 s-1 (or about 6400 cubic feet per minute in more traditional ventilation engineering units) F r o m the highly simplified example given, it becomes clear that in most practical industrial workplace control situations, particle losses by gravity may usually be neglected in relation to the direct effects of the ventilation.

11.4 E X T R A C T I O N O F W O R K P L A C E A E R O S O L S BY E X H A U S T SYSTEMS Local exhaust ventilation (LEV) involves the capture of airborne contaminants released from industrial processes before they can become widely dispersed into the workplace environment. During L E V , aerosols are extracted from the workplace air and drawn into inlets or hoods as the first stage of the elimination process. Ventilation engineers select inlet configurations that can provided m a x i m u m 'reach' for the capture of contaminants, providing sufficient capture characteristics at appropriate distances away from the inlet itself. For aerosols, the physical problem of bringing an airborne particle into a ventilation inlet from a r e m o t e location is essentially the same as that for the aspiration of a particle into a sampling inlet. This has already been discussed in some detail in Chapters 4 and 9.

336

Control of workplace aerosols In LEV systems, familiar inlet geometries include unflanged and flanged inlets of both large and small aspect ratio. The advantage of the flanged type is that it provides greater inwards 'capture' velocities than the unflanged type of (otherwise) similar configuration. For L E V in general, 'reach' is defined as the ability for the exhaust flow to capture contaminants and bring them into the inlet. For aerosols uniformly distributed throughout the space in the vicinity of the inlet, the ventilation engineer's concept of 'capture' or 'reach' is equivalent to the aspiration efficiency which is used by aerosol scientists in relation to aerosol sampling. However, in most considerations about sampler aspiration efficiency, it is usually assumed that the spatial distribution of the aerosol concentration outside the sampler is uniform over the region of influence of the sampling device. In many practical sampling situations, this is a fair working assumption. But for a ventilation inlet, the dimensional scale and the flow rates are usually much greater, and so the spatial volume of influence is much greater. As a result, consideration of the effects of spatial non-uniformity, including location of the aerosol source, becomes much more important. So it is seen that the concept of 'reach' embodies more than just aspiration efficiency but involves considerations of the extent to which the aerosol source is contained within the 'reach' volume. This is illustrated for a simple flanged hood in an external crosswind in Figure 11.2. In discussions of the effectivess of a practical LEV inlet for capturing aerosols, the starting point is the aspiration efficiency of the inlet for the given sets of conditions which pertain. This can be determined from considerations of inertial and gravitational effects like those outlined in Chapters 4 and 9. An appropriate parameter for defining the range of

Figure

11.2.

Concept of 'capture' of particles for a local exhaust ventilation (LEV) hood in a crosswind.

337

Aerosol science for industrial hygienists i n e r t i a l effects likely to i n f l u e n c e a s p i r a t i o n e f f i c i e n c y u n d e r b o t h c a l m air a n d m o v i n g air ( c r o s s w i n d ) c o n d i t i o n s is t h e h o o d S t o k e s ' n u m b e r (Sthood) d2ep, Uhood Sthood =

(11.6) 18ta,~ihood

w h e r e t h e ' h o o d ' s u b s c r i p t r e f e r s to c o n d i t i o n s at t h e p l a n e of t h e L E V h o o d i n t a k e ( a n d so c a n be r e l a t e d to e q u i v a l e n t q u a n t i t i e s for a e r o s o l s a m p l i n g as d i s c u s s e d in e a r l i e r c h a p t e r s ) , a n d w h e r e o t h e r t e r m s a r e as d e s c r i b e d previously.

Example 11.2. Consider a circular slot of d i a m e t e r (~ihood) 0.2 m through which air is extracted at 1 m 3 s-1 (representing a situation which is not atypical of what might be e n c o u n t e r e d in practice). Discuss the effectiveness of this system for capturing airborne particles with dae equal to 10 Ixm. Hood inlet velocity is

1 [m 3 s-'] Uhood :

(3.142 x 0.2 x 0.2/4) [m 2] = 31.8 m s-1 From Equation (11.6), we have (10-5) 2 [m 2] x 103 [kg m -31 x 31.8 [m s-' l Sthood --

18 x 18• 10 -6 [N s m -2] • 0.2 [m] 0.05 Based on considerations like those outlined in Chapter 9, this relatively low value of Sthood suggests that aspiration efficiency for the inlet in question will be close to 100%. However, it should be noted that, even if aspiration efficiency is 100% for particles throughout the size range of interest, the effectiveness of the ventilation for particles will be zero if the source of the aerosol is not contained within the exhausted air volume. This shows that, for hood performance, it is its 'capture' or 'reach' characteristics- rather than the inertial transport of particles in the hood v i c i n i t y - which usually dominate control effectiveness

338

Control of workplace aerosols The particle capture problem as discussed so far assumes that the movement of the air near t h e inlet is unrestricted. However, in most workplaces, the presence of objects near the inlet will be the rule rather than the exception. In particular, for many types of ventilation hood, the worker himself will impose constraints on the air movement and so modify not only the air streamlines (e.g., as described by Flynn and Miller, 1991) but also, undoubtedly, the transport of airborne particles. The nature of the problem is illustrated in Figure 11.3. It is usually very difficult to predict quantitatively the effect of such flow modifications on the effectiveness of aerosol capture under such conditions in practice. But, at the very least, an awareness of the existence and nature of the resultant complications will enable the designer to make qualitative assessments and ~ where appropriate ~ adjustments.

Figure 11.3.

Illustration of the nature of air flow problems near a hood associated with the presence of a worker.

11.5 T R A N S P O R T OF A E R O S O L S IN V E N T I L A T I O N D U C T S Once an aerosol has been captured by the inlet of a ventilation system, the aim is then to transport the particles with maximum effectiveness to a part of the system where they can be either separated from the flow or discharged safely. During transport through the ventilation ducts, particles may be deposited by a combination of gravity, inertia and turbulent diffusion (with Brownian diffusion, thermophoresis and electrostatic forces expected to be negligible in the relatively fast-moving air which usually prevails). Such deposition can cause build-up of particulate material in certain parts of the

339

Aerosol science for industrial hygienists ductwork (particularly in horizontal sections, and near bends, contractions, expansions, baffles and fittings) which, at some point, would need to be removed if significant blockage (and hence increased pressure loss and reduced ventilation performance) is to be avoided. It is clear that such deposition is undesirable. One aim of the design of ventilation ducting is therefore to minimise the conditions under which such unwanted particle deposition can occur. The nature of the flow in the ductwork obviously has an important bearing on the fate of particles which are being transported. It is a reasonable assumption that, for all situations of practical interest, the Reynolds number for the duct flow will be high enough (Reduct > 2000) that the flow will be fully turbulent.

(

i

duct

_ l:t_---_-t_- 7) ~9

Lduct

Figure 11.4. Schematic on which to base development of a model for the loss of particles from flow in a ventilation duct by gravitational settling.

Consider the transport of an aerosol in horizontal, turbulent, perfectlymixed flow through a straight rectangular duct of height Hduct, width Wduct and length L d u c t , a s shown schematically in Figure 11.4. As before, a very simple 'perfectly-mixed' model can provide useful initial insight. Here, particles may be considered to be removed by gravitational settling uniformly along the length of the duct. For any duct element of length dx at a distance x from the duct intlet, the particle loss by gravitation deposition is therefore proportional to the amount of aerosol contained within the element and the gravitational settling velocity. Let the concentration of particles of given size entering the element per unit time be c. The change of particle concentration across the element is therefore given by Qventdc = -- VsCWduct dx

(11.7)

where the volumetric flow rate through the duct is Qvent -

VductHductWduct

340

(11.8)

Control of workplace aerosols I n t e g r a t i n g o v e r the length of the duct, the p e n e t r a t i o n of particles t h r o u g h the duct is

p

eL : exp { ( VsLduct

.__

Co

)}

gductHduct

(11.9)

=exp { H e r e the goal in a practical v e n t i l a t i o n s y s t e m is to m a k e P as large (i.e., as close to unity) as possible, and this is seen f r o m E q u a t i o n (11.9) to be associated with m a k i n g Uduct a n d Hduct as large as possible a n d Lduct as s h o r t as possible. Example 11.3. Consider particles with dae equal to 10 ~m in air flowing along a 10 m long square horizontal duct of side (Hduct = Wduct) equal to 0.3 m. The volumetric flow rate is 0.5 m 3 s-1 (i.e., about 1000 cubic feet per minute). Estimate the penetration efficiency. .

.

.

.

.

.

.

.

.

.

.

As shown in Example (11.1), particle settling velocity is vs = 0.003 m s-1 From Equation (11.9), P is therefore P = f

-(0.003[m s-11

(

X

10[m] X 0.3[m]

J

0.5[m 3 S-1]

= 0.982 *

P = 98.2%

Note that, if the flow rate is raised to 1 m 3 s-1 in the same calculation, we see that P rises to 99.1%. This is a direct result of the shorter residence time of the aerosol as it passes through the duct ,

,

,

F r o m the simple ' b a c k - o f - t h e - e n v e l o p e ' e s t i m a t e s h o w n , t h e r e w o u l d a p p e a r to be very little particle loss by g r a v i t a t i o n a l settling. H o w e v e r , we n e e d to be a w a r e that o t h e r m e c h a n i s m s can also c o n t r i b u t e significantly to particle d e p o s i t i o n in ducts. T u r b u l e n t d e p o s i t i o n is o n e such m e c h a n i s m w h i c h can be i m p o r t a n t u n d e r certain c o n d i t i o n s . T h e p r o b l e m of t u r b u l e n t d e p o s i t i o n

341

Aerosol science for industrial hygienists Quiescent boundary layer

Turbulent core

~ ' / / ~ & Uduct Lduct Figure 11.5.

Schematic to illustrate the nature of turbulent particles from flow in a ventilation duct.

deposition

of

from air flow in idealised smooth cylindrical pipes has been studied by Friedlander and Johnstone (1957), Davies (1965), Sehmel (1967), Beal (1970), Liu and Ilori (1973), Liu and Agarwal (1974), and many others. These studies have been reviewed in relation to sampling tubes by Vincent (1989). Physically, such deposition is thought to involve the turbulent mixing of particles in the core flow of the tube and the inertial motion (in the form of 'free flight' projection) towards the tube wall of particles which come close enough to cross into the relatively quiescent boundary layer near the wall. This phenomenon is illustrated in Figure 11.5. If it is assumed that, to a fair working approximation, the results of these studies can be related to a corresponding square duct, then the effect of turbulent deposition on penetration of particles through an idealised smooth square duct of side/-/duct may be estimated from the following set of equations: number of particles deposited per unit area of duct per unit time W --

number of particles per unit volume of the bulk flow 0.316

f

4 D " 1/4 x,,tZduct W

W* ~--

(11.10) {(f/2) v2 Vduct } ,rPair (f]2) U d2u c t

342

Control of workplace aerosols W

p

exp{ 4 t uct )( uct )/

where it is seen that the equation for penetration, P, is similar to Equation (11.9). Here, however, the gravitational settling velocity is replaced by the quantity w which represents the effective velocity at which particles are deposited by the influence of turbulence ~ the deposition velocity. The factor 4 inside the exponential term (cf. Equation (11.9)) arises from the fact that deposition takes place at all four walls of the duct, and not just on the floor. In the above set of equations, w* is the dimensionless version of the deposition velocity while "r is the particle relaxation time and -r* its dimensionless counterpart. In addition, f is the friction factor which is expressed as a function of Re. Strictly it applies only to flow in a smooth cylindrical pipe and only for Reduct up to about 105 (Schlichting, 1968). A key piece of information which is needed in order that we may estimate P for a given situation is the relationship between the dimensionless quantities w* and -r*. This has been determined experimentally ~ again for flow in smooth cylindrical tubes ~ and the curve shown in Figure 11.6 summarises the data published by Liu and Agarwal (1974).

I -

10-t

10 -2

10-3

10 4

10-5 0.1

I 1

I 10

I 100

1 1000

T*

Figure 11.6. Relationship between the dimensionless deposition velocity (w*) versus the dimensionless particle relaxation time ('r*), for turbulent deposition of particles in a duct (summarised from Liu, B.Y.H. and Agarwal, J.K., Experimental observation of aerosol deposition in turbulent flow. Journal of Aerosol Science, 5, 145-155, Copyright 1974, with kind permission from Elsevier Science Ltd, The Boulevard, Langford Lane, Kidlington OX5 1GB, U.K.).

343

Aerosol science for industrial hygienists Example 11.4. R e - e s t i m a t e t h e p e n e t r a t i o n efficiency for t h e s y s t e m d e s c r i b e d in E x a m p l e (11.3). Assume that (a) the duct is smooth, and (b) the equations which have been derived for the idealised smooth cylinder provide a reasonable basis for estimating the penetration of the corresponding smooth square duct. For the particle with dae equal to 10 Ixm, the relaxation time is given by Equation (4.13), so that (10-5) 2 [m 2] • 103 [kg m -3] I" --

18 x 18x 10-6 [N s m -2] = 0.00031 s From Equation (11.8) the duct velocity is 0 . 5 [ m 3 s -1 ] Uduct

=

(0.3 x 0 . 3 ) [ m 21 = 5.6 m s-l So Reduct = 7 X 104 X 0.3 X 5.6 118,000 Note that this lies somewhat outside the range (<105) indicated for the use of Schlichting's expression for the friction factor (see Equation (11.10)). But assume that it is close enough for an order of magnitude approximation. From the Equations (11.10), we get the following: f = 0.0043 a'* = 1.5 s, where Pair 1.293 kg m -3 and Ix = 18x 10--6 N s m -2 for workplace air (standard temperature and pressure). =

344

Control of workplace aerosols w* = 0.0015, estimated from Figure 11.6. so that w = 0.00035 ms-1 This leads to *

P = 0.992 (or 99.2%)

Note again that, if the flow rate is increased to 1 m 3 S- 2 and if it is assumed that the equation for friction factor still applies even at the resultant higher value of Re (~240,000), we find that P falls to 92%. This is a direct result of the more intense turbulent motions at the higher airflow Again, closer inspection of the t u r b u l e n t deposition model reveals that as for the gravitational model ~ the loss is least for the shortest and widest ducts and for the greatest duct velocity. As already m e n t i o n e d , application of the model contains significant simplifying assumptions. In addition to those already m e n t i o n e d , it is i m p o r t a n t to note that practical ventilation ducting is far from smooth. Thus it is expected that f for a rough duct will be appreciably larger than that calculated. This in turn will result in deposition g r e a t e r than that calculated, and hence p e n e t r a t i o n even less. H o w e v e r , there m a y be a c o m p e n s a t i n g effect in some situations due to the fact that, at high duct velocities and for dry aerosol particles, there is likely to be considerable r e - e n t r a i n m e n t and r e b o u n d of depositing particles. The situation is m o r e complicated still in practice since, in addition to straight duct sections, most ventilation systems contain 'distorting' flow features such as bends, contractions, expansions, fittings and baffles. S o m e

x ~

/ Dust

deposit

Illustrations of situations where particle losses by deposition can occur in distorted flows in ducts (e.g., expansions, bends, constrictions, etc).

Figure 11.7.

345

Aerosol science for industrial hygienists

examples are given in Figure 11.7. The mechanisms governing the inertial behaviour of particles discussed in earlier chapters have a significant bearing on the transport of particles in ducts when such distortions are present. As usual, the governing inertial parameter is a Stokes' number for the particle motion in the flow distortion in question (Stdist) where d ae 2 P 9Uduct

(11.11)

Stdist = 181XDdist

where gduct is now the duct velocity just upstream of the distortion and Ddist is the characteristic dimension associated with the distortion (e.g., the radius of curvature of the bend). The aim in practice is to try to make Stdist as small as possible for the aerosol which is being transported. This can be achieved by designing the ductwork so that changes in flow direction and velocity are made as gentle as possible (i.e., Ddist should be as large as possible and Uduct should be as small as possible). However, there is a conflict between the desire to make Uduct as small as possible in order to minimise inertial losses and the alternative wish to make Uduct as large as possible to minimise losses by gravitational settling in other parts of the ventilation system (see Section 11.4). Ultimately, the problem is best dealt with by keeping the duct layout as simple and undistorted as possible. From models like those outlined, with their many simplifying assumptions, it becomes clear that they are useful only in providing rough guidelines for particle deposition in the ducts of practical ventilation systems. In reality it is extremely difficult to quantify all the factors that can influence particle losses. So industrial hygiene engineering practice takes a more pragmatic approach, based largely on extensive previous practical experience. Taking all the various factors into account, ventilation engineering manuals (e.g., A C G I H , 1988) suggest the working guideline that, for the minimisation of aerosol deposition, the duct velocity should be maintained well above about 15 m s-1 (about 3000 feet per minute).

11.6 P A R T I C L E R E M O V A L SYSTEMS Ultimately the particle-laden gas exhausted from the workplace atmosphere must be disposed of, either by discharging it outdoors into the atmosphere or by returning it to the workplace. In either case, it is desirable to remove the particulate material. Therefore, an important part of a control system for workplace aerosols is that concerned with separating the particles from the exhaust air. It is inevitable that this aspect is strongly linked with air pollution control as is practised for the protection of the ambient atmospheric

346

Control of workplace aerosols environment. So this represents an extension of the industrial hygienist's traditional role. Industrial gas cleaning has been the subject of many papers and books. The excellent book by Strauss (1975) remains an outstanding reference work for all those interested in this subject. The work of Theodore and Buonicore (1976) and the more recent texts by Cooper and Alley (1986) and Licht (1988) are also recommended.

11.7 G R A V I T A T I O N A L S E P A R A T I O N Gravitational settling provides the simplest way of separating suspended particles from a moving airstream. For example, it is well known that this occurs in any case when a particle-laden gas moves through a horizontal duct. Starting from ideas like those outlined earlier in this chapter, gravitational settling chambers have been designed and constructed specifically for the purpose of removing particles from gas streams.

,.!!iiiii.iilli!iii iil'ii Figure 11.8.

Schematic on which to base consideration of particle collection in a gravitational separator.

The basic principle of operation of a practical device may be based on the system shown in Figure 11.8. It consists of a chamber of arbitrary general shape, with a 'working' region having characteristic vertical and horizontal dimensions of H and L, respectively. The mean velocity of gas throughput is U. The controlling dimensionless quantity for collection efficiency is the

settling parameter HU Gs =

(11.12)

Lvs where, again, v~ is the particle gravitational settling velocity. Note the relationship between this quantity and the gravitational parameter G

347

Aerosol science for industrial hygienists

described by Equation (4.29) in Chapter 4. It is intuitive from Equation (11.12) that collection efficiency of the system will increase as v~ and L increase and as H and U decrease ~ hence as G s decreases. We could be sure that all particles will be retained within the chamber when G s < < 1. For practical purposes, we could begin by saying that particles will be collected with reasonable efficiency if G s ~< 1. This is equivalent to saying HU Vs,mi n =

(11.13)

L In turn this means that, for efficient collection, the aerodynamic diameter (dae) must be greater than the particle 'cut' size, dae,cut. From Equations (4.13) and (4.15) this gives 18IxHU

) 1/2

dae,c m -

(11.14) Lgp*

or

18p.Q dae,cut

) 1/2 (11.15)

--

W L gp *

for particles obeying Stokes' law. In Equation (11.15), Q is the air volumetric flow rate and W is the width of the settling chamber. As before, tx is the viscosity of the air, p* is the density of water (10 3 kg m -3) and g is the acceleration due to gravity.

Example 11.5. Estimate the minimum falling speed for a particle which will be efficiently collected by a gravitational settling system where the duct width is 1 m and the duct length is 5 m, and where the gas flow rate is 2 m3s-1. What does this mean in terms of particle size and actual collection efficiency? From Equation (11.13), we have for efficient collection

*

HU

Q

2 [m 3 s - l ]

L

WL

1 [m] x 5 [m]

Vs,mi n = 0.4 m s -1

348

Control of workplace aerosols For a particle obeying Stokes' law, this result yields 18 • 18x10--6[N s m -2] • 2[m 3 S-1]

dae,cut 2 -- ( l[m] • 5[m] x 9.81[m S"2] X 103[kg m-3] = 1.32 x 10-8 m2

*

dae,cut = 115 vLm

Note, however, that this calculation is favourable since particles of the size range indicated will certainly not obey Stokes' law during their settling under the influence of gravity. The actual falling speed will be substantially less. From a calculation like that shown in Example (4.2), the actual aerodynamic diameter m corrected for the non-Stokesian conditions should be 135 p.m. Finally, we may estimate the actual collection efficiency for particles of this size. From Equation (11.9), this is given by

{ ( 4Lms1,xSLm, xl,m, ) } Cgrav = 1 - P = 1 - exp

-

= 1 - 0.37 2[m 3 s-i]

Collection efficiency for particles of falling speed 0.4 ms-1 diameter 135 ~m w is greater than 60%

-

-

or aerodynamic

Expressions like those given a b o v e should be used only as a guideline or starting point for actual design b e c a u s e , in a c h a m b e r large e n o u g h to h a n d l e gas volumes of the m a g n i t u d e r e q u i r e d in practice, particle t r a n s p o r t will be influenced significantly by airflow complications such as t u r b u l e n c e , flow separation, recirculation, etc. In addition, r e - e n t r a i n m e n t of o n c e - d e p o s i t e d material is likely to f u r t h e r complicate m a t t e r s , especially if t h e r e are local regions of high velocity n e a r the collecting surfaces. T h e r e f o r e , a design b a s e d on E q u a t i o n (11.15) w o u l d r e p r e s e n t a 'best case' situation. No m a t t e r h o w the n u m b e r s in the E x a m p l e (11.5) are juggled, and a c k n o w l e d g i n g the fact that the situation will be less f a v o u r a b l e in m o s t practical cases, it b e c o m e s i m m e d i a t e l y obvious that gravitational settling c h a m b e r s are effective control devices only for very large particles. D e s p i t e such limitations, h o w e v e r , practical devices have b e e n built b a s e d on the principles of gravitational settling and have b e e n used in certain applications w h e r e they are suited (e.g., as pre-filters in very gritty o p e r a t i o n s , as mist e l i m i n a t o r s b e h i n d w e t collectors, etc). A typical practical device is illustrated in F i g u r e 11.9. Such air cleaning systems have the a d v a n t a g e s that they are c h e a p a n d easy to install and m a i n t a i n , p r e s s u r e drops are small (so that e n e r g y costs associated

349

Aerosol science for industrial hygienists Settling chamber

Dirty air

9

.

.

.

:~..-':'~ ..'~i'~.~.'~. ~'~.-~.-"..-~."'..~-~-.'~.'-:~1~/~,/~ . ~ 'oF~II'Out' large particles

~~

~

~!ii!~ ~

~

.

.

.

.

.

cleaned air

~r l

l Hoppers

Figure 11.9. Configurationof a typical gravitational separator.

with moving the air through them are correspondingly small), and disposal of collected material is simple. In addition, because the separation process is relatively 'gentle', there is little problem in the collection of highly abrasive material which, in high gas velocity systems, could give rise to significant wear and damage. The main disadvantage is, as already stated, that efficiency is low except for very coarse aerosol.

11.8 INERTIAL SEPARATION The process of inertial (or momentum) separation has a number of features in common with gravitational settling chambers, most notably in that it is an essentially passive process and no external force on the particles is actually applied (or assumed). Here, therefore, the energy required to separate the particles from the fluid derives entirely from the airflow itself. Particles are separated as the air flow is made to undergo sharp changes in direction, so that the separation mechanism involves impaction (described in Chapter 4). As in the case of gravitational separation, unwanted inertial separation may occur inadvertently in this case during aerosol flow along any duct where there are bends, expansions or contractions, or where there are bluff obstacles placed in the flow (e.g., in the form of baffles, vanes, ribs, etc.). By exploitation of these ideas, a basis exists for practical aerosol collection devices based on inertial separation, involving attempts to optimise the effectiveness of the impaction process. The starting point is the inertial 350

Control of workplace aerosols impaction parameter, already identified in Chapter 4 and earlier in this chapter as the Stokes' number, St = d2e p* U/181xD, where as usual U is the characteristic air velocity and D is the characteristic dimension of the flow distortion which is responsible for the impaction. The aim here is to devise a system where, for a given aerosol, the values of St covering the particle size range of interest are as large as possible. That is, gas velocities should be high and changes in flow direction should be as sharp as possible. Note that this is directly opposite to what is desired for the ventilation ductwork itself where, as discussed earlier, minimum deposition is desired. From impaction theory, it is reasonable to estimate that most particles are separated inertially if St I> 1. This means that particles with dae~>dae,cut will be efficiently collected where 181xD

) 1/2

dae.cut =

(11.16) Up*

Quantitative estimates for dae,cut depend in the first instance on the flow configuration which is used. Overall collection efficiency depends on the

9

9

9

eo

9

9

9

9

9

9

9

.

.

9

9

9

C

.

c

9

.

Dirty air

o 9

.

.

9

.

9

9

C ! /,l~,~

"

"

C

..

~C 9

~

9 9

9

9

~ e

9.

0"

".

s

9

-.

9

"

9

s

~

e

~

""

Partially

9

9

. . .

9

"

9 9 " "

.

o

9

cleaned air

J "-'...: o

.

"

.

" "

C

" ""

9

"

.

" "

. "

""

"

9

"

"

9

9

9

9

: . .

--"-'---

Deposited particulate material

Figure 11.10. An example of a 'curtain-type' inertial separator (based on Strauss, 1975).

351

Aerosol science for industrial hygienists impaction efficiencies of the individual collecting elements as well as their number and geometrical configuration. Some examples of inertial separators of the types that have been used in practical control systems are shown in Figures 11.10 (an example of a 'curtain-type' collector) and 11.11 (an example of a 'louvre-type' collector). Under the most favourable conditions ~ with tight changes in flow direction and with high local gas velocity ~ and juggling with ranges of plausible values for D and U in Equation (11.16) yields dae,cut no better than about 50 ~m. Although this represents a substantial improvement over the gravitational separators described in Section 11.7, it is achieved at somewhat higher pressure drop and hence higher energy costs. Another disadvantage is that re-entrainment losses are high and, for abrasive, gritty materials, wear and tear is substantially worse. Nevertheless, such devices have found useful applications in some areas of controlling aerosol emissions. Outstanding examples are their deployment as coarse pre-filters in front of more efficient air cleaning systems and as secondary collectors for eliminating liquid droplet mists and sprays exiting from wet collectors and scrubbers. Here the use of aerodynamically-designed shielded collecting surfaces in devices like the

Dirty air

.-; 9

9

9

9 9

9

.

9 9

'Louvre' plates

9 9

Airflow streamlines i01 ".~

9:,.,

"

rT/ .'i}"

.opp r

ora2ia;'Yr Deposited particulate material

Figure 11.11. An example of a 'louvre-type' inertial separator (based on Strauss, 1975).

352

Control of workplace aerosols curtain-type collector (see Figure 11.10) can be arranged to enable the collected liquid to be irrigated away without significant reentrainment loss.

11.9 CYCLONE S E P A R A T I O N The next class of aerosol collector, and one which has been much more widely used than any of the preceding collectors, utilises the concept of particle separation under the influence of centrifugal forces. The basic operating principle is similar to that of the gravitational elutriator. That is, as the aerosol flows through the separation device, an external force acts so as to drive particles towards one of the walls of the device. Whereas in the gravitational elutriator the collection mechanism depends on the force due to gravity, in the cyclone it results from the centrifugal acceleration of particles derived from the rotation of the body of the particle-laden air as it is passed through the device. In the cyclone, angular momentum is given to the air as it enters the device, either by the deployment of appropriately-designed vanes at the entrance or by some appropriate asymmetric design of the entry configuration. This general principle of operation was shown schematically in Figure 9.31. In Chapter 9 it was applied to the very small-scale particle collecting system of a personal aerosol sampler. Here the concept is scaled up massively so that it can be applied to the cleaning of large volumes of industrial gas particulate emissions. In the case of gravitational elutriation, the force is fixed (at least terrestially) and cannot be made either larger or smaller. For cyclones, however, the magnitude of the centrifugal force is determined by parameters (such as geometry, dimensions and air flow rate) over which the designer has some control. As a result, forces much greater than gravitational are possible, and hence a more efficient removal of particles can be achieved. The first approach is to consider the cyclone as an elutriator of cylindrical geometry, with the outward centrifugal force on a particle of mass m located at radius R given by

m4 Fcent -"

(11.17)

R where u x is the effective tangential velocity of the rotating body of air at radius R. There are a wide variety of design options by which particles can be separated from flowing gases along the lines suggested. Unfortunately, because the flow is so complicated and usually accompanied with strong turbulence, it is very difficult to proceed directly from Equation (11.17) to the determination of realistic quantitative estimates of performance. However, a 353

Aerosol science for industrial hygienists general trend can be identified; namely that finer particles can be collected if u T is made greater and R smaller. Some progress can be made by focusing attention on a particular form of centrifugal separator which has found widespread use in controlling workplace aerosols, the reverse-flow cyclone. This is the configuration first shown in Figure 9.31. It is re-drawn in Figure 11.12 to show the parameters which govern collection efficiency. In this apparatus the air enters the device tangentially at the top, spirals downwards through the cylindrical/conical section and then returns upwards ~ in a tighter spiral (inside the first one) to exit from the axial exit duct. During the downwards spiral, particles accelerated outwards by the centrifugal force reach the outer wall where, in the relatively quiescent flow in the layer close to the wall, they fall under the influence of gravity into the hopper at the bottom. In the system shown in Figure 11.12, the force described in Equation (11.17) can be applied in the equation of particle motion as given in Chapter 4. Although the air flow here is very complicated, indeed, various

Partially cleaned air

air _../i__~

I./ ~---D !

~

B To hopper Figure 11.12. Schematic to illustrate the principle of operation of the reverseflow cyclone, showing the important dimensions needed in the development of an empirical model for the performance of the device.

354

Control of workplace aerosols assumptions can be made in order to arrive at an approximate result for the minimum size of particle collected. Several empirical or semi-empirical models have been proposed. Based on inspection of experimental data, Iozia and Leith (1989) suggested the following set of equations (referring to Figure 11.12 for nomenclature):

dae ,cut

_{

UT,ma x =

} ( )061(Oe) 74 ( H ) 33

'rrzp* UT,ma x

6.1 U

~ D2

~ D

D

025(Oe)153 __o Acore-

0.52D

(11.18)

D 2

{ (dc/B)- 1 } for dc> B

Zcore

[(D/B)-I] or Zcore

- ( H - S ) for d c < B

where U is the average inlet air velocity, UT,ma x is the maximum tangential velocity in the body of the cyclone, and Aco~e and Zcore are the effective diameter and length of the core flow. As before, Q is the air flow rate. The other quantities contained in the above set of equations derive from the geometrical dimensions of the cyclone as identified in Figure 11.12. Iozia and Leith extended this model to obtain an empirical equation for the collection efficiency as a function of particle aerodynamic diameter. Thus

Ccyclon e --

1 + (dae,cut/d)B (11.19) with In [3 = 8 . 6 3 - 0.87 ln{dae,cut } + 5.21 In (ab/D 2) + 1.05 {ln(ab/d 2)}2 When ranges of plausible values for the various system parameters are inserted into the above equations, it is found that practical systems can be 355

Aerosol science for industrial hygienists designed which efficiently collect particles with dae down to about 5 Ixm. This is a substantial improvement over the gravitational and inertial separation devices described in the preceding sections. Such performance, as for those other devices, can be achieved with a system which is relatively simple (and cheap) to engineer. In addition, it is very compact so is ideal when space is at a premium. Once again, the device is essentially 'passive' in that the centrifugal force derives from the airflow itself. As a direct result, the price for this improved performance is the energy requirement, where pressure drop (and hence energy cost) is significantly larger. So too are wear and tear, especially if the cyclone is used for collecting highly abrasive dusts. Finally it is noted that the performance of cyclones still lies substantially below that which would be required for the effective cleaning of gases carrying significant amounts of fine particles. Therefore, for air cleaning, cyclones are usually employed as pre-separators to be used in conjunction with more efficient systems such as high-energy wet scrubbers, filters or electrostatic precipitators.

11.10 WET SCRUBBERS We now move on to separation systems which are potentially capable of more effective particle collection over a wider range of particle sizes. Therefore, whereas for the less-efficient classes of device we have generally discussed performance in terms of the 'cut-off' particle size, it is now more appropriate to talk in terms of collection efficiency. Wet collectors (or 'scrubbers') are the first family of devices in this high-efficiency category. In such collectors, the separation process involves the deposition of particles onto surfaces by combinations of impaction, interception and diffusion, all basic deposition mechanisms which have been described earlier. But for scrubbers, the main feature they all have in common is that the collection surfaces in question are wet, taking the form either of liquidirrigated solid surfaces or of actual liquid droplets moving in the same airstream as the aerosol which is to be collected. Within this framework are a large number of possibilities, usually categorised by pollution control engineers as 'low-energy' and 'high-energy'. Here, by way of illustration, just one example of each is discussed. Greater detail may be found in the more specialised texts already cited.

Spray towers In spray tower scrubbers, relatively large droplets of water (or, where appropriate, some other liquid) are generated by the use of spray nozzles

356

Control of workplace aerosols near the top of the tower and allowed to fall downwards under the influence of gravity. These droplets pass through the aerosol which is rising vertically upwards. The process of particle collection is shown schematically in Figure 11.13. It is basically the same as that by which atmospheric aerosols are 'scavenged' by falling rain drops, it being a well-known phenomenon that atmospheric particulate levels are lower following a heavy rain shower. For a typical droplet of diameter as large as 1 mm, it is recalled that the gravitational settling velocity cannot be calculated under the simplifying Stokes' law assumption. From Table 4.2 it is seen that, whereas we might estimate the falling speed of a 1 mm droplet of diameter to be about 30 m s-1 from the straight application of Stokes' law, this is in fact a considerable overestimation. When the appropriate correction for particle Reynolds' number (Rep) is applied, the correct falling speed of about 4 m s-1 is obtained. For an upwards-moving airstream, this falling speed becomes the relative velocity between the droplets and the aerosol particles. Note that this is true even for air velocities exceeding 4 m s-1. But at such high velocities the droplets, as well as the particles, will all end up exiting the stack. Collection of particles by the droplets is primarily by impaction and so

Figure 11.13.

Schematic to illustrate the principle of operation of a spray-tower wet scrubber.

357

Aerosol science for industrial hygienists

is governed by Stokes' n u m b e r (Stdrop), in the form described by Equation (4.26). Therefore, to achieve maximum efficiency for the impaction of aerosol particles onto droplets (Earop) requires a combination of high relative velocity and small droplet diameter. Thus, it is seen that we have competing demands on droplet size. It must be large enough to provide a rapid falling speed yet small enough to present a small impaction target. From such considerations, it may be shown that optimum performance is achieved for droplet diameter about 0.8 mm. Here, for an airstream with mean velocity about 1-2 m s-1, Edrop is about 90% for particles of dae -- l0 Ixm, falling to about 20% for dae = 3 I~m. A typical spray-tower configuration is shown in Figure 11.14. For efficient overall collection efficiency ( C t o w e r ) , a sufficiently large total water flow rate is required to ensure that the cross-section for the falling curtain of droplets 'covers' the whole aerosol stream flowing in the opposite direction. Performance is given by

C t o w e r --

1 - exp

{ (Qwater)( )) - Kspray

Edrop

Qair

ddrop

Figure 11.14. A typical spray-tower configuration.

358

(11.20)

Control of workplace aerosols

where Kspray is a system constant which depends on the geometry of the spray-tower system. Also, H is the height of the tower and ddrop is the geometrical size of the droplets which, for dimensional consistency, should be in the same units. When plausible practical values for the various system parameters are inserted into the above equations, it is seen that the performance of this device is not significantly better than for the cyclones described in Section 11.9. But the apparatus is very simple and has relatively low energy requirements. In practical versions, the aerosol-laden droplets fall to the bottom of the tower where, if desired, the particulate material can be separated (e.g., in a settling tank) and the water re-cycled. One disadvantage of this type of system, however, is that, if the water re-cycled in this way is not sufficiently clean, blockage of the spray nozzles can occur. On the other hand, if the water is not re-cycled, then a waste water-disposal problem exists. Finally, for the smaller droplets that do get carried upwards in the airstream, they may be removed by means of a de-mister (e.g., in the form of an inertial collector). Venturi scrubbers

At the other end of the performance spectrum, we have venturi scrubbers. The basic idea is shown schematically in Figure 11.15, where water is injected into the region of increasing air velocity in the converging section just upstream of the throat of the venturi. Here, the high fluid mechanical shear forces cause disintegration of the water stream and the production of droplets. Depending on the relaxation time of the water droplets, they cannot instantaneously acquire the velocity of the airstream in the venturi throat. So there is momentarily a high relative velocity between the droplets and the aerosol particles during the time that the droplets are 'catching up' with the flow. The converse occurs in the region where the flow expands downstream of the throat. Here the aerosol particles will decelerate faster than the more massive water droplets. The overall principle of particle collection, therefore, is basically the same as for the spray tower collector discussed in the previous section. Therefore, the impaction of aerosol particles onto droplets is the predominant separation mechanism. It is therefore easy to see that collection efficiency (Cventuri) increases as the droplet concentration increases and as the efficiency of impaction of particles onto droplets (Edrop) increases. The working equation for Cventur i is difficult to derive from first principles for a two-phase (droplet-aerosol) system which is so complicated. But empirically we have

Cventuri 1 exp( vent(

Qwater Qair

359

Stdr~ 1/2 }

(11.21)

A e r o s o l science f o r industrial hygienists Dirty air

:.:... )...".":. ...: ":~"..'~:...... ~:'.. ...:" .:.--:.:....:.-:.:.-:....:.

iii!!iili!iilii!i : :' ': 9

~

9

.

~ :::".!- ;-:-".!.:i. .

i:!i:ii!iiii:ill !!iil

.Watern

:i."".."..:~:.:i.:".!.~.! ; ..

9 9 .'." . . : . . .

. 9 " . . . .

o.

.~

" : " . " . ' " . " ' . . . . "

Venturi throat

I

I

o~ o

oc~ o,oO~1 o0O Igoo. ~

/

-----------

,~//~Particle-laden drops, to de-mister

Figure 11.15.

Schematic to

illustrate the principle of operation of a venturi scrubber.

where the coefficient Kvent is a system constant which depends on the system geometry and the operating conditions. In this expression, Stdrop increases with increasing particle size, increasing throat velocity, and decreasing droplet size. In the venturi scrubber ~ as for the spray-tower ~ the relationships between the air flow rate, the water flow rate and the throat geometry have an important bearing on droplet size and velocity, and hence on the efficiency of the impaction process. Here, however, the droplet size is much smaller and the relative droplet-particle velocity is much higher. So the collection efficiency for each individual droplet is much higher. Other physical mechanisms also come significantly into play, including coagulation, thermophoresis (if the water droplets and the airstream have different temperatures) and diffusion. In addition, in practical devices, pre-conditioning of the aerosol entering the venturi can be usefully applied; for example, air humidification so that aerosol particles can grow by condensation.

360

Control of workplace aerosols As is clear from Figure 11.15, the interaction between the droplets and the particles results in a cloud of particle-laden airborne small droplets leaving the venturi. These themselves must be eliminated, and this is achieved using a de-mister which usually takes the form of an inertial or cyclone-type separator. As usual, overall performance of the system as a whole depends on many factors, including the particle size distribution of the original aerosol. Collection efficiencies for well-designed practical systems range from about 95% for fume collection to greater than 99% for coarser dusts.

Energy considerations in wet scrubbers

As already indicated, there are many versions of the basic wet scrubber concept, and just two contrasting types have been described. Because of the diversity of types, there is no universal single working equation for collection efficiency based on the physics of the collection processes. However, it is particularly interesting that, for all wet scrubbers and for each given aerosol, there is a tendency towards a general relationship between collection efficiency (Cscrubber) and the energy consumption associated with both the air movement and the water injection. Empirically this may be expressed in the form Cscrubber = 1 - exp(- aP~)

(11.22)

where P~ is the power consumed per unit of gas volume handled, and a and b are fitted coefficients which depend only on the aerosol type and particle size distribution and are independent of scrubber type (Semrau, 1960). This shows that Cscrubber rises progressively as P~ increases. If P~ is expressed in units of [kW/1000 m 3 hr--l], then values of a and b range from about 0.4 to 3 and 0.5 to 1.4, respectively, for a wide range of types of industrial particulate emission.

11.11 F I L T R A T I O N All the above aerosol control approaches may be applied to the reduction of airborne particles in the workplace atmosphere by separating them from the air extracted by ventilation. Filtration is different in that, in addition to its application for such purposes, it also has important applications in individual personal respiratory protection systems. The term 'filtration' is frequently applied to any means by which air enters a system at one end and emerges 'clean' at the other. More specifically, however, it refers to a particular class of air cleaning system. In this context,

361

Aerosol science for industrial hygienbsts

Figure 11.16. Electron micrograph of the structure of a typical fibrous filter media (photograph courtesy of Richard C. Brown, Health and Safety Executive, Sheffield, England, UK).

a filter may usually be described as a mat of porous media through which air can pass but by which airborne particles are eliminated by deposition onto the solid elements of the mat. The most common type of filter media found in aerosol control systems, for both general cleaning and for personal filtration equipment, are made up of fine fibres (e.g., of wool, glass, etc.) packed or woven together. A picture of such a mat of fibrous material is shown in Figure 11.16. In such a filter, particles may be separated from the air by a combination of physical mechanisms, including inertial impaction and interception, gravitational sedimentation, diffusion and electrostatic forces. Because of their wide ranges of application and extensive usage, filters have been researched perhaps more than any other air cleaning system. The resultant scientific background is reviewed in detail in the excellent texts of Davies (1973) and Brown (1993).

Macroscopic picture of a filter To begin with, consider the filter from the macroscopic viewpoint, based on the picture shown schematically in Figure 11.17. If c is the number of

362

Control o f workplace aerosols Element of filter, thickness dx

Co---~

~

Ce

~//~/~ Clean air

Dirty air

//

.

x=O Figure 11.17.

x

x+dx

x=X

Macroscopic picture of a filter on which to base a model for collection efficiency.

particles of given size per unit volume of air (i.e., concentration) passing into an element of thickness dx at a distance x from the face of the filter, then the number of particles deposited from that unit volume in the filter element is dc -

- cBdx

(11.23)

This says that the number of particles collected within the element is proportional to the number entering and to the length of the element. So B here is the proportional rate per unit filter thickness at which particles are removed as the air passes through the element, and must have units of [length-~]. It is strongly dependent on particle size, mean air velocity through the filter, filter porosity and bulk solidity, and individual fibre dimensions. Implicit in this is also the fact that B is strongly dependent on the microscopic nature of the flow around the individual elements that go to make up the filter media. If it is assumed that the macroscopic flow is uniformly distributed across the width of the filter and that the properties of the filter itself are homogeneous throughout its bulk, then integrating over the whole thickness (X) of the filter gives Ce =

Coexp(-BX )

(11.24)

where c~ is the number of particles exiting from the filter per unit air volume and c o is that at the inlet. Filter penetration (P) is therefore Ce

P =

= exp(-BX) Co

363

(11.25)

A e r o s o l science f o r industrial hygienists

and, conversely, overall collection efficiency is Cfilte r =

1- P

(11.26)

In order to move from this generalised picture towards incorporating the physical mechanisms by which particles are separated inside the body of the filter, we now define the single fibre collection efficiency, Efibre, as No. of particles deposited per unit length of fibre (11.27)

Efibr e =

No. of particles geometrically incident per unit length of fibre which is seen to be directly analogous to the impaction efficiency defined in Chapter 4 (except that, now, E f i b r e embodies all contributions to particle collection). From Equation (11.27), for fibres of diameter D f i b r e and total overall length Lnbre per unit filter volume, the number of particles collected when unit volume of aerosol passes through the element shown in Figure 11.17 is dc - - c E f i b r e D f i b r e L f i b r e d x

(11.28)

so that, by comparison with Equation (11.23), we have B ~-

EfibreDfibreLfibre

(11.29)

For a filter which has a bulk solidity, o'filter, where fibre volume (11.30)

O'filter --

total filter volume it can easily be shown from geometrical considerations that the total length of fibre per unit volume of filter is 40"filte r Lfibre

(11.31)

-, r r D 2fibre

Equations (11.29) and (11.31) together yield o'filtergfibre

(11.32)

B'rrDfibr e

364

Control of workplace aerosols so that for the filter as a whole

- 4o'filterEfibreX 1 Cfilte r =

1 - exp

(11.33)

xrOfibre This is the basic working expression for all filters. The important ingredient, and one that needs to be evaluated before Equation (11.33) can be applied in practice, is the single fibre collection efficiency, Efibre. This requires detailed consideration of the physics of what happens at the microscopic particle-fibre level.

Single fibre collection efficiency Contributions to the single fibre deposition efficiency (Efibre) can come from a number of physical collection mechanisms, including impaction interception gravitational sedimentation diffusion and, if either particles and/or the fibres are electrically charged, also electrostatic deposition. For each of these mechanisms operating singly, we may define the individual deposition efficiencies Eimp, Eint, EG, Ediff and E E, respectively. Each can in principle be determined from a model based on a mathematical simulation of the flow field in the vicinity of the fibre and the determination of particle motion based on physical ideas about particle motion like those outlined in Chapter 4. The calculations are complicated but have been performed for many situations of practical interest (as reviewed fully by Pich, 1966; Davies, 1973; Brown, 1993). Hinds (1982) has summarised very concisely a set of equations by which the single fibre efficiency for each mechanism can be calculated to a fair approximation sufficient for most practical purposes. For present purposes, a short summary of the main features will suffice. These are as follows: For impaction, Eimp increases steadily with the inertial parameter (St, referred to the fibre diameter, Df, and the air face velocity to the

365

Aerosol science for industrial hygienists filter), and is modified by the effect of o'filter on the flow about individual fibres. For interception, Ein t increases with the ratio of particle geometrical size to fibre size (d/Dfibre) and, again, is modified by o'filter. For diffusional deposition, Ediff increases as the Peclet number (Pe, see Chapter 4) decreases, and it too is modified by o'filter. For gravitational deposition, E G increases with the gravitational parameter (G, see Chapter 4). For electrostatic deposition, we cannot identify a single governing dimensionless parameter because there can be more than one version of the mechanism, depending on whether the fibre or the particles are charged ~ or both ~ and on their dielectric properties. We can be

Efibre, overall Co

Co

J

El

E2

En

v

ClVl

C2

c~

CI = (1 - El); C2

To

~ =~l- El)...

so that Ce C o = (1 - e l ) ( l

- e2)...(1

- en)

so overall single fibre efficiency is Efibre, overall = 1 -- (1 -- El) (1 - E2)...(1 - En)

Figure 11,18,

Schematic

to i l l u s t r a t e h o w t h e i n d i v i d u a l s i n g l e f i b r e e f f i c i e n c i e s in a filter c a n b e c o m b i n e d .

366

Control of workplace aerosols sure that the magnitude of E E is strongly d e p e n d e n t on the magnitude of the charge carried by either the particles or the fibres. H o w e v e r , whether or n o t E E increases or decreases with charge level depends on the relative polarities of the charges involved. In reality, the individual collection mechanisms described do not act singly, but rather in combination. The resultant interactions b e t w e e n them can be very complicated. H o w e v e r , for most practical purposes, a simple approach may be taken in which it is assumed that each m e c h a n i s m is assumed to operate quite independently of the others. In this case, the overall single fibre efficiency, Efibre,overall c a n be estimated from the scheme shown in Figure 11.18. Then Efibre,overall-- l-(l-Eimp)(l-Eint)(l-Ediff)(1-EG)(1-EE )

(11.34)

which, if all the E's are very small ( < < 1 ) , reduces to Efibre,overall =

E i m p -+- E i n t 4-

EG

+

Ediff +

(11.35)

E E

Some trends for these various collection mechanisms are shown in Figure 11.19 for a typical example, based on results of calculations by Hinds (1982).

(a) I0-

(b)

df = 2 l~m

Particle size of maximum penetration

lmpaction

U = l O c m s -I

=o cep 0.n

~

d~

o,,,~

r~

0.01

i cr

005

I

I

1

10

0.001 0.01

0.1

1

10

0.01

0.1

Particle diameter (lam)

Particle diameter (l~m)

Figure 11.19. Some trends in collection efficiency as a function of particle diameter for filter media where the fibre diameter (dr) is 2/xm and the air face velocity is 10 cm s-l: (a) single fibre efficiency for several mechanisms, and (b) overall filter efficiency for media thi,,kness 1 mm and a bulk solidity of 0.05 (plotted from data calculated by Hinds, W.C., Aerosol Technology, Copyright 1982, adapted by permission of John Wiley and Sons Inc).

367

Aerosol science for industrial hygienists These calculations were performed for particles of density 10 3 kg m -3 and for an uncharged aerosol and filter. In Figure l l.19a are plotted the single fibre collection efficiencies for each of the mechanisms referred to above (except electrostatic). We see the predominance of diffusion at very small particle sizes, and of impaction and interception at larger particle sizes. Gravity plays a generally smaller role. It is noted that single fibre efficiencies for impaction and interception can exceed unity. This results from the geometry of the filter where each fibre has neighbouring fibres in close proximity. It can occur only to the point w h e r e Eimp+Ein t <~ l+(d/Ofibre ). The net result of all the contributions to collection efficiency are shown for the filter media as a whole in Figure l l.19b. Here we see again clearly the respective regions where diffusion and impaction/interception dominate. There is an intermediate region where neither are particularly effective, hence the dip in collection efficiency at a particle diameter of about 0.2 Ixm. This is referred to as the 'particle size of maximum penetration'. It is important to recognise its existence in a practical filter because it represents the particle size at which the filter is least effective and so provides the least protection to the workplace, the worker or the outside environment (depending on the application).

Practical filters

Practical filters intended for applications in ventilation systems are usually made from fibrous material packed together in the form of a felt or woven into a cloth-like structure. Among the materials used have been wool, cotton, man-made mineral fibre and metal. From the above discussion, it is clear that the finer the fibres woven into the mat the greater the value of the single fibre collection efficiency (Efibre); and so too, for a given bulk solidity and face velocity, the greater the overall filter efficiency (Cfilter). For this reason, fine fibrous asbestiform materials at one time ~ many years ago ~ found applications in some types of filter, although this has fallen out of favour in recent years as we have become increasingly aware of the health risks associated with the use of such material. High overall efficiency (Calter) is obtained for a filter where the overall collection rate (B) is also high. To achieve this for low bulk solidity (filter), and hence low pressure drop across the bulk filter media, requires high Efibre. Alternatively, a lower value of Efibre can be tolerated if o-filter is high although at the price of higher pressure drop and, hence, higher energy costs. However, this pressure drop limitation can be reduced by designing the filter such that the velocity of the airflow through the filter is reduced for example by increasing the cross-sectional area of the flow through the filter. One way to achieve this increased cross-section is by pleating the filter media or by forming the filter media into a 'sock'. The highest

368

Control of workplace aerosols efficiency is achieved when E f i b r e and o'filter are both high. For filters which o p e r a t e on the mechanical deposition mechanisms (impaction, interception, gravitational deposition and diffusion), this balance b e t w e e n Cnlt~ ~ and energy r e q u i r e m e n t s becomes a direct trade-off. The constraints associated with such a trade-off may be b r o k e n by the use of electrostatic forces (i.e., by ensuring that E E is high). To achieve this, the filter media itself may be artifically electrified. O n e version that has been widely used for m a n y years is the resin wool filter, where the wool fibre is i m p r e g n a t e d with small ( ~ 1 Ixm diameter) insulating particles of resin which, during the p r e p a r a t i o n (in particular the 'carding' process) of the wool material in the m a n u f a c t u r e of the felt media, b e c o m e electrified by triboelectric (or friction) charging. High electrical forces inside a filter m a d e of such material can be achieved until, over a period of time, the resin particles b e c o m e neutralised (e.g., due to contact with the naturally-occurring charge

Bag supporting structure and shaking system

I

I ~.

t~

'

l

c;;a.

9 . ~'...

i..."

~---- Air-flow ' B ag-house'--------~

9 .

-

-U::! ,

.

. . . . . .

.~

9

9

' 'Sock' (or 'bag' )-type filter media

. . .

..

i.i.i.i...-.-,. "i/:: r

.

....

.

-

:...f

::.....: Z

.

"-

9-

' " :

..'.

."

..-

:

..-...~t . . " . : :.-:::::}.: ... _

:'

~i~.~.!.[~:.:.~.{{.~~.~.:~:~../.!.~.~.!.[~.~:. " ":." L """: L

Di.rty air

..........

9

. ~...:.."..?~...-!".."::~":..">::..::.i:...".

!

I

Figure 11.20. Configuration of a typical 'bag-house'-type fabric filter system. Note here the 'bag' or 'sock' arrangement of the filter media. The pressure associated with the airflow inflates the bag. The particulate material collected inside each bag is caused to fall into the hoppers by the action of the shaker system which is activated periodically.

369

Aerosol science for industrial hygienists

Figure 11.21. Some typical filter systems of the type used for personal respiratory protection: (a) disposable half-mask 3M T M 8715 Dust/Mist Respirator, of the type recommended for dusts, mists and fumes where the permissable exposure limits (PELs) are not less than 0.05 mg m-3; and (b) disposable half-mark 3M T M 9970 High Efficiency Respirator with exhalation valve, of the type recommended for dusts, mists and fumes where the PELs are less than 0.05 mg m -3. (Photographs courtesy of 3M Occupational Health and Environmental Safety Division, St Paul, MN.)

370

Control of workplace aerosols in the ambient atmosphere such as air ions and cosmic rays, accumulation of charged aerosol, etc.). Other versions employ filters made of highlycharged, insulating electret material, also pre-charged during the production process, this time by corona discharge. Others employ externally-applied electric fields. In many practical air cleaning devices, especially those which are intended for the separation of particles at high concentrations over extended periods of time, deposited material builds up quite rapidly and the resultant particulate layer (the 'cake') itself becomes an active participant in subsequent particle collection. This interesting phenemenon has been described in detail by Brown (1993). The build-up of particles on the fibres of the filter media is portrayed as 'dendritic', having 'tree-like' branching structures whose features need to be taken into account in any attempt to describe the physics of filtration by loaded filters. During such filter loading, interception becomes a more predominant mode of collection, and the overall efficiency of collection increases. In fact, in some such filters, optimum performance characterisation is based on the achievement of the 'cake' of particulate material in this way. So, for a clean filter, collection efficiency might be quite low at first, then rise steadily during the initial stages of particle collection, reaching its specified performance level only after a certain amount of time has elapsed. In practical devices, all the considerations discussed above can be brought into play in varying combinations. There are many physical options. A typical common practical arrangement is shown schematically in Figure 11.20. This is the 'bag-house' filter which finds applications both in medium-scale air cleaning operations (e.g., in filtering extracted workplace air before it is returned to the workplace) and in larger-scale operations (e.g., in filtering exhaust gases directly from industrial processes into the atmosphere). A much smaller-scale filtration system is that which is used in personal respiratory protection. Again, there are many physical options. Two are shown in Figure 11.21.

11.12 ELECTROSTATIC PRECIPITATION Electrostatic precipitation of aerosols utilises the principles of particle charging and the forces which charged particles experience when subjected to an externally applied electric field. This leads to the family of air cleaning devices known as electrostatic precipitators (ESPs) which, although they find their widest range of applications in controlling particulate emissions to the atmosphere from large-scale industrial processes (e.g., fossil-fuelled electricity generating stations), do find occasional use in ventilation systems for controlling workplace aerosols. They are also commonly used in residential air cleaning systems.

371

Aerosol science for industrial hygienists In the conventional precipitation process, suspended particles are electrically charged by contact with gaseous ions in a corona electric discharge, and are then caused to migrate under the influence of an electric field to a collecting boundary, whereupon they adhere and so are removed from the air flow. Charging and precipitation may take place in the same electric field or in separate regions with separate fields. The layer of particulate matter that builds up over a period of time can subsequently be removed mechanically. These processes are described in detail in the classic text by White (1963). The simplest useful configuration is the concentric pipe-wire arrangement shown schematically in Figure 11.22a. The gas to be cleaned is passed through the pipe. The central wire placed along the axis of the pipe is maintained at a high electrical potential with respect to the electrically grounded outer conducting cylinder. This potential is high enough to maintain a corona discharge but not so high as to allow actual sparkover. Wire polarity may be positive or negative. Although the latter allows a more stable electrical potential for the corona discharge (and hence greater particle charging and in turn more effective collection), it has the disadvantage that a greater amount of ozone is generated. Ozone is a toxic gas, and so it is undesirable to recirculate ozone-containing air back to the workplace environment. Therefore, there are many situations where negative corona are undesirable, and we have to settle for the less efficient positive corona. Particles entering the cylindrical duct are charged and precipitated in

DirtYair

:~:.9 i]:)i';.:i::~:.4.:i::(:.:i.:.?:]i;::].;f.; -'i;i "

o

"

"

"

High voltage supply

o ~..--o

t ..

J

tLL)

I

......'~......-.-o.-:. 9149 9".'" " I "" r. ."-.'['..'.-: 9 .t-"-"t

"

""'"'" "'"""'"'""'""lI"T

... "::.~ ~.:.e.!.6.-.-(...... 9~i,":~.o-:~/

.t"..

~

Clean air

(a,

t'

l, t .I'*l ...

stiffeners/ribs t

.9" - : - . ) 9 9 9 ".'- .'. .~- .- .','. ~-.',-'~ ;.'1 ,. . t I "-~.'-o.'. "o :-.~'.:-.o I.o 9Y / / h . " . . ;

,

---

~Plate ....

I .... )

o

9 9 I I 9 9 '

~

"I

'

I

...

9

t

I

9 I

9

t

9 t

oft

t

~*- - . . :)-.. :-.-

..":1~".~"/"7: ~.o .:o...~...:. :::...-..: .:..:...-.:..

(b)

9

EHT

(c)

'' EHT

Figure 11.22. Schematics of some typical electrostatic precipitator configurations: (a) pipe-wire, (b) single-stage plate-wire, and (c) two-stage plate-wire.

372

Control of workplace aerosols the same region. For handling large gas volumes like those commonly encountered in large-scale industrial applications, the alternative plate-wire configuration is more widely used, the simplest version being as shown in Figure 11.22b. A version which is more common in practice is the one where the two-dimensional fence-like ribs sticking out from the collecting plates are designed not only to provide plate stiffening but also to prevent erosion (by re-entrainment) of the collected layer of particulate material. Figure 11.22c shows a further configuration where particle charging takes place in the entrance region of the precipitator and most of the collection takes place in the subsequent corona-free region. This is the type which is found in most residential 'electronic' air cleaners, primarily because it generates less ozone.

Corona discharge The term 'gas discharge' refers to the situation where a gas is made conducting, as manifested by the production of free ions and electrons which can appear as an electrical current. A corona discharge occurs in a non-uniform electric field like that found in a pipe-wire or plate-wire electrode configuration where the wire is held at a high potential with respect to the grounded plate. The resultant electric field is higher near the wire where the lines of electric flux converge sharply. In this region, if the potential is high enough, any free electrons present may receive kinetic energy from the field such that, when they collide with gas molecules, they can release one or more molecular electrons. This is the phenomenon known as ionisation, by which further free electrons and ions are created. Ionisation is usually accompanied by excitation, whereby some molecular electrons which are not released in this way may be elevated to higher molecular states and, in the process of decaying back down again to their original 'ground' states, cause the release of photons at wavelengths characteristic of the molecular transitions that have taken place. This process accounts for the fact that, in an electrical discharge, visible light is usually emitted. For a corona discharge in air, the faint blue glow in the high-field region around the wire is characteristic of the excitation of nitrogen and oxygen molecules. The new electrons created by ionisation are themselves accelerated in the high field, giving rise to further ionisation, leading to 'amplification' of the charge associated with the initial free electrons. This amplification has been referred to by gas discharge physicists as an 'avalanche' process. When the wire is positive, the electrons are attracted to the wire and so are lost. The positive ions migrate outwards and enter the lower field region further away from the wire. As a result, for a positive corona, that outer region becomes filled with positive ions which are then available for the charging of aerosol particles that pass through the region. Alternatively, when the wire is

373

Aerosol science for industrial hygienists

negative, it is the positive ions that migrate towards the wire and are removed and the electrons which flow outwards. For a negative corona discharge in air, these electrons combine readily with the oxygen molecules in the air by virtue of the strong mutual affinity that exists in oxygen for electron attachment. The region outside the immediate vicinity of the wire or point is now filled with negative ions. A corona is a relatively low-current, low-power electrical discharge and is maintained only when the field which is high enough to support ionisation is confined to the region near the wire. Otherwise, for higher potential on the wire, ionisation takes place further out from the wire and can eventually lead to amplification in the avalanche process which is so great that the gas becomes, in effect, short circuited. At this point, the corona undergoes a transition to a spark and, in turn if the current is not restricted in the external circuit, an arc. The electrical currents flowing in such discharges can be many orders of magnitude higher than in coronas, and are accompanied by the generation of much greater light, heat and sound (e.g., as in lightning discharges in the atmosphere during thunderstorms). Such high current discharges are entirely unsatisfactory for electrostatic precipitation.

Particle charging Aerosol particles receive electric charge by being bombarded by the free ions of the corona discharge. In most electrical air cleaning devices, one of the principle mechanisms is field charging. The presence of a particle in the electric field of the corona represents a dielectric discontinuity, so that there will be lines of electric flux that terminate at the particle surface. Ions moving along flux lines can therefore arrive at the surface of the particle, and will continue to do so until the total charge that has built up on the particle modifies the field distribution outside the particle sufficiently to prevent the influx of any further ions. At this point, the particle is regarded as having been charged to saturation. During the initial stages of charging, the charge (qf) on a spherical particle of diameter d varies with time t according to the expression

qf

(~r%Ecd2) (%+2)

(11.36) (l+~rZiNit)

where N i is the ion density and Z i their mobility, E~ is the electrical field in which charging is taking place, and Er is the dielectric constant ~ or relative permittivity ~ for the particulate material in question. The quantity % is the absolute permittivity of a vacuum, and appears in Equation (11.36) for dimensional consistency in the SI system of units. Equation (11.36) reveals

374

Control o f workplace aerosols

that N i is the quantity that governs the rate at which saturation is achieved. It relates directly to the current flowing in the corona discharge. For ion densities typical of most corona discharges in electrostatic precipitators, N i will exceed 1013 ions m -3, so that saturation will be complete within the order of one second. When that is achieved, the saturation particle charge is given by

-

(-rr%Ec d2)

qfs

(11.37)

(e-r+ 2)

Note here that E!r - - 1 for pure vacuum, 1-10 for many intermediate dielectric materials, and ~ for conducting materials. Also note that, for most small aerosol particles in atmospheric air, the presence of monolayers of water molecules on the surfaces of the particles tends to make them behave as if they were conducting. In addition, for pure vacuum note that the absolute permittivity (%) is a universal constant with the value 8.85 x 10-12 As V-1 m-1. A second mode of charging needs also to be considered ~ namely, d i f f u s i o n charging. Here charging takes place by virtue of the random collisions that take place between ions and particles, derived from the thermal motions of both. Here, charging needs to take into account the statistical distributions of the particle and ion random velocities, respectively, and of the repulsive electrical forces that can occur once a particle has started to accumulate charge. From such considerations, the charge (qd) o n a particle of diameter d at time t is given by

qo =

In

1+

2e

(11.38) 2kT

where T is the gas temperature, c i is the mean random velocity of the ions and k B is Boltzmann's constant. Note that ~ 2 x 102 m s - 1 for air ions at STP, and k B = 1.38 x 10 -23 J ~ -~, Boltzmann's constant.

Ci

It is interesting to note from Equation (11.38) that, unlike for field charging, there is no prediction of saturation charge. This results from the theoretical distribution of ion random velocities by which, even for very high particle charge, there remains a finite probability of finding an ion that has enough random velocity to overcome the electrical repulsion force. In a corona discharge, particle charging will be taking place simultaneously by both the mechanisms outlined. But Equations (11.37) and (11.38) show

375

Aerosol science for industrial hygienists that diffusion charging p r e d o m i n a t e s for smaller particles and field charging for larger particles. Liu and Kapadia (1978) have analysed the complex intermediate situation where both are taking place simultaneously. But for many practical purposes, it is a reasonable first approximation to assume that one or the other predominates. For example, for particles with d i a m e t e r greater than about 0.5 ~m, field charging may be taken as the primary charging mechanism.

Particle electrical migration velocity W h e n a charged particle now experiences a precipitating f i e l d Ep, the force it experiences is qEp so that, the electric migration (or drift) velocity for a spherical particle of diameter d is given by E q u a t i o n (4.19) as

qE P vE =

(11.39) 3wl~d

where the drag force is assumed to follow Stokes' law. For a particle charged by field charging to saturation as given by E q u a t i o n (11.37), this becomes 3~ r

)

(%E c

Epd)

vE -

(11.40) (~r+2)

31~

For a device where charging and precipitation take place in the same field, E c a n d Ep are approximately the same and can be estimated from geometrical and discharge current-voltage parameters.

Example 11.6. In a typical plate-wire precipitator, the charging and precipitating fields are approximately the same at 5 kV cm -1. Estimate the drift velocity for a spherical particle of diameter 5 ~m. As in most practical situations, we may assume that the particle is conducting. Note that, for a conducting particle, Er ~. So the first bracketted term in Equation (11.40) becomes equal to unity -

Note that 5 kV cm-1

= 5 •

-

105 V m-1.

Note that % is a universal constant, and is given by 8.85• 10-12 As V-1 m-1.

376

Control o f workplace aerosols

From Equation (11.40), we get 8.85X

1 0 -12

[As V -I

m -l]

X

(5X

105) 2 [V 2 m -2]

x

5x10 -6

[m]

VE --

3x 18x 10 -6 [N

m -2]

= 0.205 A V N -l 12E 0.20 m s-t, noting that [A V N -1] in the SI system of units is equivalent to [m s-l]. =

V a l u e s of E c a n d Ep of the o r d e r of 5 k V cm -~ are c o m m o n in e l e c t r o s t a t i c p r e c i p i t a t o r c o r o n a s . C o r r e s p o n d i n g c u r r e n t s are of the o r d e r of a few m i l l i a m p s p e r c e n t i m e t r e of c o r o n a wire.

Simple

model

of precipitator

collection

efficiency

T h e simplest m o d e l for p r e c i p i t a t o r efficiency is t a k e n f r o m the t h e o r y of D e u t s c h (1922). It is b a s e d on the single p l a t e - w i r e d u c t s h o w n in F i g u r e 11.23. H e r e it is a s s u m e d that the flow is t u r b u l e n t a n d t h a t , as a result, the a e r o s o l passing t h r o u g h the duct is u n i f o r m l y a n d c o n t i n o u s l y m i x e d . P a r t i c l e s which, by this m i x i n g p r o c e s s , c o m e close to the p l a t e are r e m o v e d to the plate surface at the p a r t i c l e electrical m i g r a t i o n velocity, v E. In this p r o c e s s , the rate of p a r t i c l e r e m o v a l at any c r o s s - s e c t i o n in the d u c t is p r o p o r t i o n a l to the a v e r a g e c o n c e n t r a t i o n o v e r t h a t c r o s s - s e c t i o n . F r o m this p i c t u r e , the p r o g r e s s t o w a r d s o b t a i n i n g an e x p r e s s i o n for c o l l e c t i o n efficiency is similar to t h a t for filters as d i s c u s s e d earlier. So, if c is the n u m b e r of particles of given size p e r unit v o l u m e of air (i.e., c o n c e n t r a t i o n ) p a s s i n g into an e l e m e n t of thickness dx at a d i s t a n c e x f r o m the inlet to the e n t r a n c e to the p r e c i p i t a t o r

SINGLE PLATE-WIRE DUCT (plate h e i g h t - into plane of page m is H)

... :'.'7.7--'.'..' .'- ". ,.,. 9 .-.~ " .

9. . . . . . .

..

c " " " VZ//A _ 1 . I c - ax lT~J/~'d ''~" 9

"': ".'"-i"-"~--:-"~'.'.".'." ""'" "".'Di.rty...: -".-..." .... . 9 a~r

Figure 11.23.

x=0

X

x

x+dx

x=L

Clean air

Schematic on which to base development of a model for the collection efficiency of an electrostatic precipitator.

377

Aerosol science for industrial hygienists duct shown, then the n u m b e r of particles deposited from that unit volume in the filter e l e m e n t is dc=

- c + dx

(11.41)

and + is a collection p a r a m e t e r (similar to B for the filter) which has dimensions of [length -a] and must contain v E. Integrating over the whole length (X) of the precipitator duct gives CES p = 1 -- exp (- + X )

(11.42)

It can be shown, either from simple dimensional a r g u m e n t s or by physical reasoning, that

2 Hv

E

+ -

(11.43)

Q where Q is the gas volumetric flow rate and H is the height of the plates (into the plane of the page in Figure 11.23). This m e a n s that the term 2 H represents the area of the two plates in contact with the flow per unit length of the duct. If A here is the total area of the collecting plate surfaces exposed to the aerosol, then E q u a t i o n (11.42) takes the form

_ YEA

)

CEs p = 1 - exp

(11.44)

Q This e q u a t i o n also applies for an electrostatic precipitator consisting of m a n y ducts (like those ones shown in Figure 11.22) and so is the general working equation for electrostatic precipitators. Example 11.7. A plate-wire electrostatic precipitator is designed to collect particles of 5 txm in diameter. It comprises 21 plates of height 5 m and length 10 m. The plate spacing is 0.2 m. Air flows through the 20 ducts of this system at a mean velocity of 2 ms -1. Estimate the collection efficiency of this device? Recall from Example (11.6) that vE for these particles is 0.2 m s-~. The volumetric air flow rate through this device is given by Q = 20 x (0.2 x 5)[m 2] x 2 [m S-1] = 40 m3 s-1

378

Control of workplace aerosols Plate area in contact with the gas is given by A = 2 x 20 x (5 x 10)[m 2] = 2000 m2 From Equation (11.44), we get

cEsp=,exp{ ( *

0.2[m s-1] x 2000[m2] 40[m3 s-l]

)}

CEsp = 99.99%

From this example we see the potential for achieving very high collection efficiencies by electrostatic precipitation. But it should be noted that the physical picture represented by the Deutsch model is rather simplistic. A number of much more complex models have been developed, involving, for example, considerations of (a) the detailed lateral and longitudinal turbulent motions of the particles in the turbulent air flow in the ducts (e.g., Cooperman, 1971); (b) the aerodynamic effects of the ribbed collecting plates (e.g., Vincent and MacLennan, 1980); (c) the effects of the 'electric wind' arising from the motion of the ions ~ and the transfer of momentum to neutral gas molecules ~ in the weakly ionised gas of the corona discharge; (d) particle re-entrainment (e.g., Tsai and Mills, 1995); and so on. Although electrostatic precipitators have been around since the turn of the century, such research seems to have had relatively small impact on their practical deployment. So the Deutsch model remains the basis of much of the engineering design of practical ESPs.

Practical precipitator systems A typical practical electrostatic precipitator is shown in Figure 11.24. It can be a very large piece of equipment. The dimensions given in Example (11.7) are fairly typical of devices that are used for controlling emissions from large-scale industrial processes (e.g., coal-fired electricity generation). For such devices, engineers have come to realise that the theory as outlined above is indeed rather simplistic. Practical experience has shown that collection efficiency always tends to be lower than predicted theoretically, but consistently by an amount which correlates strongly with the type of aerosol to be collected and with the conditions under which it is collected. The result is that, although the form of the basic working equation for the precipitator

379

Aerosol science for industrial hygienists

Figure 11.24. A typical practical electrostatic precipitator (from Burgess, W.A. et al., Ventilation for Control of Work Environment, Copyright 1989, adapted by permission of John Wiley and Sons Inc).

performance (Equation (11.44)) has been found to hold reasonably well, the magnitude of the particle electrical migration velocity (rE) to be used in those equations is consistently lower than predicted from Equation (11.40). In practice, therefore, engineers very often replace the physical quantity vF with a quantity which is altogether more empirical, based on previous experience and dependent on what is known about the type of aerosol to be collected, its particle size distribution, gas temperature and humidity, etc. In Example (11.7), for instance, an experienced engineer might decide that a typical value of 'v E = 0.1' might be more realistic for a particular situation. Then the collection efficiency estimate falls to the more conservative 99.3% (that is, the emissions rise by a factor of three times). Despite the reduced collection efficiency arising from the factors indicated, electrostatic precipitators can achieve very high collection efficiency. Although the installation cost is higher than for most other types of air cleaning system, the cost of maintenance is relatively low and, since the

380

Control of workplace aerosols

pressure drop is much less than for other types of air cleaner of comparable efficiency, the power requirements are relatively small. In addition, although high voltages are called for, the electrical currents are low. By the use of appropriate safety interlocks, etc., therefore, safety is no more of a problem than in any other device which utilises mains electrical power. Electrostatic precipitators do, however, carry the disadvantage that, in the event of a high-voltage power failure, performance falls dramatically. So there is no 'fail-safe'.

11.13 C O M P A R I S O N B E T W E E N A I R C L E A N I N G SYSTEMS In the preceding sections, the essential elements of aerosol science underpinning the various particle separation methods are summarised. In the process, many of the relative merits of the various particle separation methods, in terms of their ability to effectively perform the basic tasks required, have become apparent. Typical trends in collection efficiency are summarised as functions of particle size in Figure 11.25. They show that performance improves as particle size increases for all the devices described. Even the least effective air cleaners can find useful applications in specific situations, if not as primary collectors then as pre-separators. But factors

Fabric filters and electrostatic precipitators

1O0~ ~ ' ~ - ~High_energy f [ ~---scrubber~ / ~%" ] / /Low-energy/

/

~- 5o

el//

v

o0.I

Cyclones /

/

nal rs

Inertial /

/ II0 Particlediameter,d (~m) I

1

/

I I00

Figure 11.25. Comparison between collection efficiencies as a function of particle diameter for typical practical air cleaning devices like those described in this chapter (from Burgess, W.A. et al., Ventilation for Control of Work Environment, Copyright 1989, adapted by permission of John Wiley and Sons Inc).

381

Aerosol science for industrial hygienists

other than particle size also come into play when making decisions about which system to install for a given industrial application. These include gas temperature, moisture content and corrosiveness, gas flow conditions, the nature of the aerosol, cost (both installation and running) and degree of complexity. The suitability of the various air cleaner types on the basis of these various considerations are summarised in Tables 11.1-11.4. T a b l e 11.1.

Basic scientific factors for air c l e a n e r selection. Particle size to be collected

Cyclones

Gas temperature at inlet to collector

Above 10 ~m

1-10 ~m

Below 1 ~m

Above 400~

250 ~ to 400~

Dewpoint to 250~

Near dewpoint

Yes

Care

No

Yes

Yes

Yes

Avoid

Yes

Care

No

Care

Yes

Yes

Yes

Yes

Yes

Yes

Care

Yes

Yes

Yes

Yes

Yes

Yes

No

No

Yes

Avoid

Yes

Yes

Yes

No

Care

Yes

Avoid

Yes

Yes

Yes

Care

Yes

Yes

Avoid

Low energy wet scrubbers High energy wet scrubbers Fibrous filters Fabric filters Electrostatic precipitators

T a b l e 11.2.

Practical technical factors for air c l e a n e r selection. Practical process requirements Low capital cost

Low running cost

Low technical complexity

Dry product

Fire or explosion risk

Cyclones

Yes

Yes

Yes

Yes

Care

Low energy wet scrubbers

Yes

Yes

Yes

No

Yes

High energy wet scrubbers

Care

Avoid

Yes

No

Yes

Fibrous filters

Yes

Yes

Yes

Yes

Care

Fabric filters

Care

Care

Care

Yes

Care

Avoid

Yes

Avoid

Yes

Avoid

Dry electrostatic precipitators

382

Control of workplace aerosols Table 11.3.

A e r o s o l p r o p e r t y factors for air cleaner selection. Dust properties High inlet burden

Erosive

Sticky

Light/ Fluffy

Difficult to wet

High resistivity

Yes

Yes

Avoid

Avoid

Yes

Yes

Fabric Filters

Care

Care

Avoid

Yes

Yes

Yes

Fibrous Filters

Avoid

Yes

Care

Yes

Yes

Yes

Wet scrubbers Low Energy

Yes

Yes

Yes

Yes

Care

Yes

Wet Scrubbers High energy

Yes

Care

Yes

Yes

Care

Yes

Electrostatic precipitators

Care

Yes

Care

Care

Yes

Care

Cyclones

Table 11.4.

Gas p r o p e r t y factors for air cleaner selection. Gas conditions

Other factors

Constant pressure drop

Varying flowrate

Corrosive

Cyclones

Yes

Care

Care

Yes

Yes

Yes

Fabric filters

Care

Yes

Care

Care

Care

Care

Fibrous filters

Care

Care

Care

Yes

Yes

Avoid

Wet scrubbers Low energy

Yes

Care

Care

Yes

Care

Yes

Wet scrubbers High energy

Yes

Care

Care

Yes

Care

Yes

Electrostatic precipitators

Yes

Care

Care

Care

Care

Yes

High pressure

Minimum ancill. equip,

On-line cleaning

11.14 C O N T A I N M E N T OF A E R O S O L S The preceding sections have dealt with the means by which suspended particles behave in the workplace air and how they might be removed from the air which is extracted during general or local exhaust ventilation. A complementary option is to contain the aerosol at or close to the course of its release. One of the problems associated with overall systems for controlling workplace aerosol exposures is that the effectiveness of the control system as

383

Aerosol science for industrial hygienists a whole is strongly d e p e n d e n t on the efficiencies of a n u m b e r of c o m p o n e n t s . In some cases, especially where highly toxic substances are c o n c e r n e d , such measures are sometimes insufficient ~ even with the best available designs to maintain exposure levels to within prescribed limit values. A l t h o u g h this is not usually the first a p p r o a c h t a k e n (because it is often not c o n v e n i e n t in relation to the industrial process), enclosure of the source of aerosol m a y be applied in such cases. In any such o p e r a t i o n , a system of interlocks can ensure that the e q u i p m e n t in question cannot be o p e r a t e d while the doors are open. W h e n m a i n t e n a n c e or any attention to the process is r e q u i r e d , the w o r k e r must shut d o w n the o p e r a t i o n before the doors can be o p e n e d for access. W o r k e r s in such industrial o p e r a t i o n s have l e a r n e d to accept the necessity to work within such constraints. Quite recently, consideration has been given to the application of 'localised' electrostatic precipitation. H e r e the idea is to r e d u c e the c o n c e n t r a t i o n of aerosol at source by using c o r o n a discharge to charge particles and cause t h e m

A laboratory-based pilot system designed to investigate electrostatic containment of aerosol generated from a textile winding process: (a) un-modified winding rig, and (b) modified with the addition of the enclosure shown (where the high-voltage discharge brushes are located inside the front door) (from Johnston, A.M. et al. 1988, from Applied Occupational and Industrial Hygiene, A laboratory pilot study to investigate the feasibility of localized electrostatic precipitation for dust control in industrial workplaces, reprinted with permission of the American Conference of Governmental Industrial Hygienists). Figure 11.26.

384

Control of workplace aerosols

to be deposited close to their points of release (Johnston et al., 1988). By way of illustration, a pilot arrangement which was designed for application on a textile winding rig is shown in Figure 11.26. By partial enclosure of the parts of the process where most of the dust is generated (i.e., at friction points), and by placing corona discharge electrodes ~ in the form of fine carbon fibre brushes ~ inside the enclosure and maintaining them at a high potential, the level of airborne dust near the process is much reduced. Just as important, perhaps, is that, when the need arises to open the door to the enclosure (e.g., for machine maintenance or inspection), the worker is not exposed to the sudden 'puff' of aerosol which would otherwise occur (as illustrated in Figure 11.27). This could be a serious problem if the aerosol is known to be particularly hazardous. At present, however, such systems are in the prototype stage and have not yet found their way into widespread practical application.

(a) Voltage off

I I I I m r

o

I I I I I I I I I

t~

0

Door

(b) Voltage on

o~pcned

4

Time (minutes)

Figure 11.27. Some typical results using the pilot rig shown in Figure 11.26: (a) the concentration of dust released when the door is opened after running without the discharge brushes energised, and (b) the concentration of dust released when the door is opened after running with the discharge brushes energised (from Johnston, A.M. et al. 1988, From Applied Occupational and Industrial Hygiene, A laboratory pilot study to investigate the feasibility of localized electrostatic precipitation for dust control in industrial workplaces, reprinted with permission of the American Conference of Governmental Industrial Hygienists).

385

Aerosol science for industrial hygienists

11.15 P E R S O N A L R E S P I R A T O R Y P R O T E C T I O N The various options for controlling worker exposure were listed in order of priority at the beginning of this chapter. It is part of industrial hygiene philosophy that, as far as possible, such control is primarily the responsibility of the workplace ownership and management rather than the worker himself. However, situations frequently occur where reasonable practicable measures along the lines indicated do not, in themselves, reduce exposure levels sufficient to acceptably minimise risk. In such situations, personal respiratory protection is applied as a last resort. This might take the form of full breathing apparatus (with its own independent air supply); powered respirator/helmet devices in which clean, filtered workplace air is provided inside the visor; and 'self-powered' face mask filters. The latter is the most common approach nowadays, increasingly of the disposable variety. Such masks are formed from fibrous filter material like that described above, moulded into a form that can fit comfortably and effectively onto the face of the worker (see examples in Figure 11.21). Of course, basic filtration efficiency is a primary index of performance for such devices, and the underlying aerosol science has already been outlined earlier in this chapter. However, other considerations also come into play. First there is the question of comfort (and in turn worker acceptability). Breathing resistance too is especially important for prolonged usage, and so much of the research by the manufacturers of filter media and devices has concentrated on the use of high-performance, low-pressure drop, electrified material (e.g., resin wool, electret). Such materials are widely found nowadays in commercially-available personal respiratory protection equipment. To further reduce breathing resistance, one-way mechanical valves are used in some types of face mask to ease pressure-drop during exhalation. But such valves are a potential source of leakage during inhalation, so need to be carefully designed and maintained. A particularly important consideration involved in the practical use of personal face-mask respirators includes the quality of face-fit. Face-seal leakage can ruin the overall performance of a face mask, even one where the intrinsic filter performance is very good. This in turn is highly dependent on the facial features of individual wearers, and so use of such equipment must be accompanied by adequate fit-testing procedures and worker education.

386

Control o f workplace aerosols

REFERENCES American Conference of Governmental Industrial Hygienists (ACGIH) (1988). Industrial Ventilation: A Manual of Recommended Practice, 20th Edn. ACGIH, Cincinnati, OH. Beal, S.K. (1970). Deposition of particles in turbulent flow on pipe or channel walls. Nuclear Science and Engineering, 40, 1-11. Brown, R.C. (1993). Air Filtration: An Integrated Approach to the Theory and Applications of Fibrous Filters. Pergamon Press, Oxford. Burgess, W.A., Ellenbecker, M.J. and Treitman, R.D. (1989). Ventilation for Control of the Work Environment. John Wiley and Sons, New York. Commission of European Communities (CEC) (1982). Synthesis Report of the Dust Measurement and Control Working Groups. CEC, Directorate-General V, Luxembourg. Cooper, C.D. and Alley, F.C. (1986). Air Pollution Control: A Design Approach. PWS Publishers, Boston, MA. Cooperman, P. (1971). A new theory of electrostatic precipitation. Atmospheric Environment, 5, 541-551. Davies, C.N. (1965). The rate of deposition of aerosol particles from turbulent flow through ducts. Annals of Occupational Hygiene, 8, 239-245. Davies, C.N. (1973). Air Filtration. Academic Press, London. Deutsch, W. (1922). Bewegung und Ladung der Elecktrizitatstrager im Zylinderkondensator. Annals der Physik, 68, 335-344. Flynn, M.R. and Miller, C.T. (1991). Discrete vortex methods for the simulation of boundary layer separation effects on worker exposure. Annals of Occupational Hygiene, 35, 35-50. Friedlander, S.K. and Johnstone, H.F. (1957). Deposition of suspended particles from turbulent gas streams. Industrial and Engineering Chemistry, 49, 1151-1156. Hinds, W.C. (1982). Aerosol Technology. John Wiley and Sons, New York. Iozia, D.L. and Leith, D. (1989). Effect of cyclone dimensions on gas flow pattern and collection efficiency. Aerosol Science and Technology, 100, 491. Johnston, A.M., Hughson, G.W., Jones, A.D. and Vincent, J.H. (1988). A laboratory pilot study to investigate the feasibility of localized electrostatic precipitation for dust control in industrial workplaces. Applied Industrial Hygiene, 3, 321-325. Licht, W. (1988). Air Pollution Control Engineering: Basic Calculations for Particulate Collection, 2nd Edn. Marcel Dekker, New York. Liu, B.Y.H. and Agarwal, J.K. (1974). Experimental observation of aerosol deposition in turbulent flow. Journal of Aerosol Science, 5, 145-155. Liu, B.Y.H. and Ilori, T.A. (1973). Inertial deposition of aerosol particles in turbulent pipe flow. Presented at the ASME Symposium on Flow Studies in Air and Water Pollution, Atlanta, GA, June. Liu, B.Y.H. and Kapadia, A. (1978). Combined field and diffusion charging of aerosol particles in the continuum regime. Journal of Aerosol Science, 9, 227-242. Pich, J. (1966). The theory of air filtration. In: Aerosol Science (Ed. C.N. Davies). Academic Press, London, pp. 223-285. Schlichting, H. (1968). Boundary Layer Theory, 6th Edn. McGraw-Hill, New York. Sehmel, G.A. (1967). Validity of air samples as affected by anisokinetic sampling and deposition within sampling lines. In: Proceedings of the Symposium on Assessment of Airborne Radioactivity, Vienna, pp. 727-735. Semrau, K.T. (1960). Journal of the Air Pollution Control Association, 10, 200. Strauss, W. (1975). Industrial Gas Cleaning. Pergamon Press, Oxford. Theodore, L. and Buonicore, A.J. (1976). Industrial Air Pollution Control Equipment for Particulates. CRC Press Inc., Cleveland, OH.

387

Aerosol science for industrial hygienists Tsai, R. and Mills, A.F. (1995). A model of particle re-entrainment in electrostatic precipitators. Journal of Aerosol Science, 26, 227-239. Vincent, J.H. and MacLennan, A.S.M. (1980). Aerodynamic considerations in electrostatic precipitation. Journal of Electrostatics, 8, 325-342. Vincent, J.H. (1989). Aerosol Sampling: Science and Practice. John Wiley and Sons, Chichester, U.K. White, H.J. (1963). Industrial Electrostatic Precipitation. Addison-Wesley, Reading, MA.

388

C H A P T E R 12

Aerosols and vapours 12.1 I N T R O D U C T I O N At the outset of this book, it was declared that the subject would be 'Aerosol Science'. However, it is clear from the properties of m a t t e r as summarised in Chapter 2 that a molecule is the smallest unit of matter which, in turn, might therefore be considered to represent the lowest possible limit in particle size. So the range from particle to molecule is continuous, and the distinction between particles and vapours may not always be clear cut. Indeed, in some industrial hygiene cases, some volatile substances may ~ depending on their vapour pressure ~ exist simultaneously as airborne contaminants in the forms of both aerosol and vapour. There are many similarities between what we regard as aerosol and what we regard as gases, and how we deal with them in the industrial hygiene context. But there are also many important differences. This chapter therefore sets out to show these similarities and differences. The t r e a t m e n t is largely qualitative.

12.2 T H E T R A N S P O R T O F G A S E S For clean air flowing through a tube, we may see from kinetic theory that there are interactions between the gas molecules and the wall of the tube. In fluid mechanics this accounts for the concepts of pressure and friction. The nature of the interactions is illustrated in Figure 12.1a which shows that the air molecules m a k e elastic collisions with the wall of the tube. That is the colliding molecules r e b o u n d and do not stick to the wall, so that the mass of air which exits the tube is the same as that which enters. Physically, this is t a n t a m o u n t to saying that the concentration of the molecules at the wall is the same as in the body of the tube. If, as shown in Figure 12.1b, the air also contains suspended particles then, as discussed in preceding chapters, those which ~ by the variety of mechanisms available ~ collide with the wall will usually stick and so be removed from the flow. So the mass concentration of

389

Aerosol science for industrial hygienists Molecule

~

i) I

(a)

Particle

i)

(b)

Figure 12.1. Nature of interactions for air and particles flowing through a tube" (a) elastic collisions between air molecules and the wall, and (b) 'sticking' collisions of aerosol particles with the wall.

particles exiting the tube is less than that for particles entering. Physically, this is equivalent to saying that the particle concentration at the wall falls to zero. If the air now contains a contaminant gas, then molecules of that gas may also collide with the tube wall. This is illustrated in Figure 12.2. Such collisions are the same as for the molecules of the air itself, and so are determined by kinetic theory (and hence by thermal considerations). Unlike macroscopic aerosol particles, there are no contributions to deposition from gravity, inertia, thermophoresis or electrostatic forces. However, some such molecules may adhere to the tube wall on impact (i.e., be adsorbed). If all colliding molecules are adsorbed in this way,

O, 9

I

(a)

:

(b)

,

Non-adsorbent

""

Adsorbent

Figure 12.2. Nature of interactions between contaminant gas molecules and t h e tube wall for air flowing through the tube: (a) elastic collisions for non-adsorbent molecules, and (b) 'sticking' collisions for molecules that are adsorbed.

390

Aerosols and vapours

then the concentration of those molecules falls to zero at the wall. If only a proportion of the colliding molecules are adsorbed (depending on the nature and effectiveness of the adsorption process), then the effective concentration is greater than zero, but less than that in the body of the tube. Either way, there is a concentration gradient for the contaminant molecules between the wall and the main tube flow. The net result is a diffusive flux of these molecules towards the wall. So the concentration of the contaminant molecules exiting the tube is greater than those entering. Thus we have a situation which is somewhat analagous to that for the more macroscopic aerosol particles. The same physical situation exists for a large solid surface in contact with air which is not moving. Again, for air itself there is no adsorption of molecules. For aerosols, those arriving at the surface by gravity, diffusion, thermophoresis or electrostatic forces may adhere. For contaminant gas molecules, again those which arrive at the surface by diffusion may be removed by adsorption (see Figure 12.3). The resultant concentration gradient between the surface a n d the air outside results in a flux of the molecules in question to the surface. In all the scenarios given, the collecting surface is portrayed as a solid medium onto which molecules are adsorbed. The same issues of molecular transport occur when the surface is a liquid and where the colliding molecules enter the surface and are absorbed into the body of the liquid.

Contaminant Air molecule

Adsorbent (for contaminant)

Figure 12.3.

Interactions between a flat surface and airborne contaminant gas molecules where the wall adsorbs those molecules.

391

Aerosol science for industrial hygienists

12.3 I N H A L A T I O N OF GASES The preceding may be applied directly to a discussion of the fate of inhaled air and contaminant gas molecules. For aerosols, it remains reasonable to assume that particles which arrive at the wall of the respiratory tract are removed with 100% efficiency, regardless of the part of the respiratory tract. This forms the basis of what is referred to as the 'regional deposition' of inhaled particles and which leads to the definition and quantification of criteria for the assessment of deposition and dose in humans (as discussed in Chapter 8). For contaminant gas molecules, the mechanism of absorption at the moist wall of the respiratory tract leads to the similar regional deposition of those molecules. Here, whereas for aerosol particles, calculation of regional deposition is complicated by the nature of the air flow and of the particle mechanics, the problems for gas molecules are somewhat different. A number of models have been developed, including that proposed by Davies (1985). The Davies model takes account of the air flow in the nasopharyngeal (extrathoracic), tracheobronchial and alveolar regions of the human respiratory tract, and embodies the roles of molecular diffusion and solubility of contaminant molecules in transport towards and adsorption at the respiratory tract wall. For sulphur dioxide in air inhaled at 12 1 min -~, for example, Davies calculated that 14% of the inhaled SO 2 would be deposited in the nasopharyngeal region, and 86% in the tracheobronchial region. None would reach the alveolar region. For other gases the distribution of deposition is different. Here we see, therefore, the importance of regional deposition of inhaled contaminant gases. At first sight this appears somewhat analogous to the particle size-selective criteria which have been identified for aerosols. However, Davies points out the important difference (cf. aerosols) that the parameters that govern deposition of gas molecules ~ molecular diffusivity and solubility at the wall of the respiratory tract ~ are intrinsic to the molecular species in question. Thus, identification of the contaminant gas, in itself, provides a full description of the factors influencing regional deposition. Hence it is a sufficient criterion for health-related sampling.

12.4 T H E SAMPLING OF GASES The conventional sampling of aerosols requires the aspiration of a known volume of particle-laden air, and the subsequent collection or sensing of the particles inside the body of the sampling instrument. Such instruments require air moving equipment in the form of pumps, and are therefore often referred to as 'active'. There is another class of sampling device which does not require such pumping apparatus, and such devices are referred to as 'passive'. These include deposit gauges of the type used in Britain for 392

Aerosols and vapours

collecting large particles in the ambient atmosphere and which operate on the principles of gravitational settling or inertial behaviour of particles in the ambient atmospheric wind (reviewed by Vincent, 1989). More recently a passive aerosol sampler, based on the application of electret media, has been proposed as a possible pump-free personal sampler (Brown et al., 1992). However, with all these passive aerosol samplers, a major problem lies in matching their particle size-dependent performances with health-related criteria like those described in Chapter 8. Active samplers for gaseous contaminants are widely used by industrial hygienists. The simplest are referred to as 'denuders', and involve the scenario described above in which the air and the gaseous contaminant are drawn through a duct or a channel where the walls are coated with some adsorbing substance. For deposition on the walls of a straight circular tube, the equation for collection efficiency of the contaminant gas by diffusion (for a perfectly adsorbing wall) is given to a fair approximation by (Cheng, 1989) Cdenude r --

1 - 0.819 exp(-3.66 13)

(12.1)

where rrDL

13 =

> 4

(12.2)

Q in which Q is the sampling flow rate, D and L are the tube diameter and length, respectively, and the flow in the tube is laminar (i.e. Re < 2300). In this device, the choice of coating depends on the gas which is to be collected. Some are listed in Table 12.1. One very simple denuder system which has found application for personal sampling by some industrial hygienists is shown in Figure 12.4. This consists of an appropriately coated denuder tube followed by a filter holder. Here, and in all other gas sampling systems, it is obvious from the preceding

Table 12.1. List of adsorbent media suitable for use in denuders for the gases indicated. Sampled gas

Coating

Sulphur trioxide Sulphur dioxide Ammonia Formaldehyde Chlorinated organics

Oleic acid Nylon sheet Copper sulphate Bisulphite-triethanolamine Tenax powder

393

Aerosol science for industrial hygienists discussion that there are no entry effects comparable with those experienced for aerosols, and so it can always safely be assumed that aspiration efficiency for contaminant gas molecules is unity (or 100%). The particular system shown in Figure 12.4 can be used for the simultaneous collection of both gaseous and aerosol contaminants. A more common approach to the active sampling of gases in the industrial hygiene setting involves drawing the sampled air through a tube which is packed with the appropriate sorbent material (see Figure 12.5). For organic solvent vapours ~ a very common class of gaseous contaminant in workplace atmospheres ~ activated charcoal provides an excellent adsorbing medium. For personal samplers of the type most frequently encountered, the sampling flow rates are much lower than for aerosol samplers (of the order of 0.1 1 min -1 cf. 2 1 min -1 for aerosols) and so sampled air volumes are correspondingly low. This ensures that the residence time of the gas in contact with the sorbent is as high as possible to ensure efficient collection by diffusion yet the total sampled volume is not so high that the sorbent becomes overloaded leading to 'break-through'. For the example given, the sorbent material is a solid. It is also an option to sample gas molecules by absorption into a liquid medium (as, for example, in the 'bubbler'-type impingers which are used in industrial hygiene for some gaseous contaminants). However, this is a less-convenient approach for the purpose of personal sampling. It is clear that the preceding gas or vapour collection methods relate to the concept of molecular diffusion from moving air outlined above. For gases, the 'un-pumped' passive mode of sampling is a more feasible option than for aerosols. It is also more attractive for personal sampling for gases because

Figure 12.4. Typical simple gas denuder and aerosol sampling system of the type that has sometimes been used for personal exposure assessment in industrial hygiene.

394

Aerosols and vapours

Figure 12.5.

Typical 'active' sampling system for contaminant gases and vapours, of the type widely used in industrial hygiene.

the worker is not asked to wear a pump. The principle of operation is based largely on the concept of diffusion of gas molecules from (essentially) still air to an adsorbent surface (see Figure 12.3). As first proposed for gas sampling by Palmes and Gunnison (1973), diffusive sampling occurs by virtue of Fick's law as originally expressed in the form of Equation (2.11) but now written as

Dgas A ( c - Ca) Mass collected per unit time =

(12.3) L

where Dgas is the molecular diffusivity of the gas of interest, A is the collecting area, L is the length of the 'diffusion layer', and c and c a are the concentration of interest and at the sorbent surface, respectively. Note here, in relation to Equation (2.11), that it is c, c a and L which define the concentration gradient adjacent to the sorbent surface. Note also that the term DgasA/L has dimensions of volumetric flow rate (e.g., [ml min-1]) which can be used as a basis for working backwards from the collected mass to the time-weighted average sample concentration. A wide range of types of passive diffusive sampler is available, and some are shown in Figure 12.6. There are some limitations which derive from the basic properties of molecular motion near the sorbent surface. The first is that Fick's law, on which the operation of the device is based, assumes that the concentration outside the sampler is constant. In practice, of course, this is unlikely to be the case. The effects of fluctuating gas concentrations has been studied (e.g., Bartley et al., 1983) and it has been shown that such fluctuations do not present a significant problem so long as the timescale of sampling is much greater than the time constant associated with the sampling process (typically

395

Aerosol science for industrial hygienists

Figure 12.6. 'Passive' or 'diffusive' samplers of the type widely used for assessing personal exposures to contaminant gases by industrial hygienists: (a) 3M T M 3500 Organic Vapor Monitor, in which organic vapor molecules pass through a diffusion membrane and are collected on a charcoal adsorbent pad, after which they may be recovered and analysed by gas chromatography; and (b) 3M T M 3600 Mercury Vapor Badge, in which mercury atoms are collected on a gold film which enables the formation of mercury/gold amalgam which can subsequently be analysed by the change in electrical conductivity. (Photographs courtesy of 3M Occupational Health and Environmental Safety Division, St. Paul, MN.)

396

Aerosols and vapours

a few seconds, representing the time taken for a molecule to diffuse across the 'diffusion layer' adjacent to the sorbent surface). The second limitation is based on the requirement that the concentration just outside the sorbent surface remains representative of the atmospheric concentration of interest. For this to be met, there needs to be sufficient external convective air movement to replenish the molecules removed from the region just adjacent to the 'diffusion layer'. It has been found that a minimum external wind of about 0.2 m s-1 is required for this purpose, and this requirement is satisfied in most workplaces. In all such gas collection systems like those described, appropriate recovery (e.g., thermal desorption, carbon disulphide extraction) and analytical methods (e.g., gas chromatography) need to be employed. These are usually standard procedures in a well-equipped industrial hygiene analytical chemistry laboratory.

12.5 D I R E C T - R E A D I N G INSTRUMENTS F O R GASES As for aerosols, a very wide range of options is available for the detection and measurement of contaminant gases and vapours. Most derive from the physics of how the gas molecules (a) interact with radiative energy, or (b) behave electrically, thermally or chemically. In the first category are infrared, visible or ultraviolet photometers which operate on the basis of the changes in the molecular states of the gases of interest due to the absorption of radiation at characteristic wavelengths. The general principle of such instruments is similar to the extinction of light by aerosols. That is, the attenuation of photons of light by the absorption of energy by molecules (i.e., by changing their vibrational molecular state) is described by the exponential law, similar to that shown in Equation (5.3) for aerosols. From this the concentration of the molecular species of interest can be obtained if a suitable radiation wavelength is chosen which is characteristic of absorption by the gas of interest without significant absorption by molecules of other species. Some photometric devices operate on principles associated with other molecular changes of state (e.g., emission, scattering). In the second category, electrical monitoring devices are based on the electrical properties of the gas or vapour of interest, either in the air or in solution. These are based, for example, on detecting changes in electrical conductivity, changes in pH, ionisation, etc. Thermal options include detecting changes in thermal conductivity or in the heat of combustion. Chemical options include detecting changes in colour or the emission of characteristic radiation following mixing the gas of interest with a suitable reagent.

397

Aerosol science for industrial hygien&ts Options for direct-reading instrumentation for contaminant gases have been reviewed by Nader et al. (1989). Also provided in this article is an extremely useful listing and critique of the instrumentation which is available commercially.

12.6 A I R C L E A N I N G F O R T H E C O N T R O L OF GASES The physical basis of most of the options for controlling contaminant gases and vapours is as described earlier in this chapter. Unlike for aerosols, where there is a wide range of physical mechanisms that can be brought to bear in separating the particles from the air, the sole mechanism available for gas molecules is diffusion. The problem therefore becomes an engineering one which, for a given gaseous contaminant, involves (a) a choice of a suitable adsorbent or absorbent medium, and (b) choice of a flow rate and configuration which maximises both the residence time of the gas in contact with the adsorbent/absorbent medium and the specific surface area of contact (i.e., (surface area in contact}/{volumetric flow rate)). Whereas for personal sampling purposes it is usually convenient to employ a dry adsorbent material, it is usually more common to use liquid absorbents for the relatively large-scale applications involved in controlling exhaust gases from industrial process. Options include irrigation of solid surfaces (the so-called 'disperse liquid-continuous gas' approach) or the bubbling of the gas through bulk liquid (the so-called 'continuous liquid-disperse gas' approach).

12.7 I N D U S T R I A L H Y G I E N E I M P O R T A N C E OF A E R O S O L S A N D VAPOURS Some of the above principles can be applied to thinking about materials which, while normally existing in the liquid phase ~ and so can be present as aerosol ~ can also appear as vapours in air. This is a situation commonly of interest to industrial hygienists, since not all such materials are harmless. As discussed in Chapter 2, saturation vapour pressure (Psv) represents the pressure which would be exerted by vapour molecules in equilibrium with the same material in liquid form inside a closed container. For a material starting out as 100% liquid in such a closed system, some of the molecules will evaporate into the vapour phase. For some materials, the attractive molecular forces between liquid molecules are relatively weak, so that the pressure exerted by that liquid in the closed container will be relatively high since a high proportion of the material will be present in the vapour phase. Conversely, for materials with stronger intermolecular forces, relatively fewer molecules will be present in the vapour phase. So the vapour

398

Aerosols and vapours

pressure will be correspondingly lower. Thus it follows that materials with high vapour pressures are more likely to evaporate into the air than those with relatively lower vapour pressure. For example, hydrazine (NzH4, a colourless liquid) has a vapour pressure at STP (VPsTP) of 10 mm Hg, while hexane (CH3(CHz)4CH3, another colourless liquid) has a a vapour pressure of 124 mm Hg. Thus the absolute magnitude of vapour exposure is likely to be greater for hexane than for hydrazine. However, this in itself can be misleading because it says nothing about the hazard associated with exposure. This discussion therefore leads to a concept useful to industrial hygienists the vapour hazard ratio (VHR). For a given material, this is defined as SC VHR =

(12.4) OEL

where O E L is the relevant occupational exposure limit for the material in question (in parts per million by volume, ppm) and where SC is the saturation concentration (also in ppm) given by Pv,STP X 10 6

SC =

(12.5) BP

In this expression, barometric pressure (BP) is 760 mm Hg. Applying the above to the examples of hydrazine and hexane, we obtain the following: Hydrazine: SC = 13,158 ppm and O E L = 1 ppm ~ V H R = 13,158 Hexane: SC = 163,158 ppm and O E L = 500 ppm ~ V H R = 326 This indicates that hydrazine is potentially much more hazardous to health, despite its lower vapour pressure and, hence, lower magnitude of exposure. In some cases it is also important to consider the extent to which a material, when it is airborne, can exist as a vapour or an aerosol. To quantify this, SC (as defined in Equation (12.5)) is first converted into a mass concentration (mg m-3). This is then compared to the O E L (also expressed in mg m-3). Thus we have the following possible scenarios (Perez and Soderholm, 1991): (a) If SC/OEL < 1, the hazardousness of the airborne material will be mostly associated with it in the aerosol form. (b) If 1 < SC/OEL < 100, the hazardousness of the airborne material will be associated with both aerosol and vapour forms. (c) If SC/OEL > 100, the hazardousness of the airborne material will be mostly associated with it in the vapour form.

399

Aerosol science for industrial hygienists Table 12.2.

Examples of vapour versus aerosol hazard potential

Substance

TLV (mg m -3)

SC/TLV

Most likely

Mercury Gluteraldehyde

0.05 0.70

400 100

Vapour Vapour and

Trimellitic anhydride

0.04

1

concern

aerosol Aerosol

P e r e z a n d S o d e r h o l m c a l c u l a t e d the ratio S C / T L V ( w h e r e T L V a n d O E L are e q u i v a l e n t ) for s u b s t a n c e s listed in the A C G I H - T L V b o o k l e t ( A C G I H , 1993-1994). S o m e i n t e r e s t i n g e x a m p l e s are s h o w n in T a b l e 12.2. T h e v a l u e of S C / T L V for m e r c u r y v a p o u r , for e x a m p l e , confirms t h a t the s e t t i n g of an O E L for m e r c u r y b a s e d on the a s s u m p t i o n of its a p p e a r a n c e p r e d o m i n a n t l y as a v a p o u r is correct.

REFERENCES American Conference of Governmental Industrial Hygienists (ACGIH) (1993-1994). Threshold limit values for chemical substances and physical agents and biological exposure indices. ACGIH, Cincinnati, OH. Bartley, D.L., Doemeny, L.J. and Taylor, D. (1983). Diffusive monitoring of fluctuating concentrations. American Industrial Hygiene Association Journal, 44, 241. Brown, R.C., Wake, D, Thorpe, A., Hemingway, M.A. and Roff, M.W. (1992). A passive sampler for airborne dust using an electret. Journal of Aerosol Science, 23 (Supplement 1), $623-626. Cheng, Y.S. (1989). Diffusion batteries and denuders. In: Air Sampling Instruments (Ed. S.V. Hering), American Conference of Governmental Industrial Hygienists (ACGIH), Cincinnati, OH, pp. 405-419. Davies, C.N. (1985). The absorption of gases in the respiratory tract. Annals of Occupational Hygiene, 29, 13-25. Nader, J.S., Lauderdale, J.F. and McCammon, C.S. (1989). Direct-reading instruments for analysing airborne gases and vapours. In: Air Sampling Instruments (Ed. S.V. Hering), American Conference of Governmental Industrial Hygienists (ACGIH), Cincinnati, OH, pp. 507-581. Palmes, E.D. and Gunnison, A.F. (1973). Personal monitoring device for gaseous contaminants. American Industrial Hygiene Association Journal, 34, 78. Perez, C. and Soderholm, S.C. (1991). Some chemicals requiring special consideration when deciding whether to sample the particle, vapor or both phases of an atmosphere. Applied Occupational and Environmental Hygiene, 6, 859-864. Vincent, J.H. (1989). Aerosol Sampling: Science and Practice. John Wiley and Sons, Chichester, U.K.

400

POSTSCRIPT

This book was aimed from the outset to be a comprehensive view of aerosols in industrial hygiene. The term 'comprehensive' here does not mean to imply that all of the subjects are covered in their full depth. That would require a text far larger than this one. Rather, it is intended that the subject is covered broadly in an integrated way with appropriate depth in key areas. No such book is ever complete. New knowledge and perspectives emerge by the day, and so any text may be out of date long before it eventually appears in print. Again, therefore, as suggested in Chapter 1, motivated readers should continue to maintain vigilant eyes on the relevant literature contained in the field's most prominent peer-reviewed journals. It is hoped that the material here will provide the framework on which the new developments which will inevitably emerge can be assessed critically. The book is therefore seen as a starting point for in-depth study ~ and not just an end in itself. What might be expected to be the priorities for new developments in this branch of industrial hygiene during the next few years? Several areas emerge from the review contained in the preceding pages. It is clear, for example, that bioaerosols will feature increasingly, partly because of our greater awareness of the potential importance of such aerosols in all industrial settings and partly because of the increase in the scope of industrial hygiene to include such 'industrial' workplaces as offices and hospitals. At present we remain uncertain about the extent to which the behaviour of particles of biological origin is similar to that for inert particles, both before and after deposition on surfaces, in filters, or in samplers. Study and resolution of many of the scientific issues will require a close interaction between microbiologists, aerosol scientists and industrial hygienists. Elsewhere, the relationship between exposure, dose and response for inhaled aerosols requires further illumination. Although we now know how to define exposure in relation to most aerosol-related ill-health, we still do not yet know how to extrapolate what has been learned from animal inhalation studies to humans in relation to the subsequent fate of the deposited material. Similarly, our understanding of the dynamics of the biological response remains inadequate. Progress in this area will require the cooperative efforts of aerosol scientists and inhalation and

401

Aerosol science for industrial hygienists

molecular toxicologists. Aerosol sampling will continue to require further understanding and development, driven by the needs identified by the emergence of the new particle size-selective criteria. It is perhaps surprising but nonetheless true ~ that we still do not fully understand the nature of particle motion near a body into which air is being aspirated, especially for the complex ranges of conditions like those which are met during actual industrial hygiene aerosol sampling. Until these gaps in our knowledge are filled, the design and development of new samplers ~ and the interpretation of results obtained using those of the previous or current generations will remain highly empirical. There have been considerable strides in the development of direct-reading instruments for assessing workplace aerosols. However, many such instruments remain costly and not always convenient for use in 'robust' industrial settings. So further progress towards reducing unit cost and in the miniaturisation of instrumentation is expected. Finally, much current practice for ventilation and air cleaning is based on engineering and science that is decades old. So the search for improved technical control of workplace aerosols will continue, and it is expected that this will include practical implementation of many of the new and advanced aerosol science and technology concepts that have emerged in recent years. From the above it is clear that aerosol science and its applications will continue to be a major priority for industrial hygienists for the foreseeable future.

402

Index Absorption light 117 molecules 392 Adsorption, molecules 390-397 Aerobiology 67-70 Aerodynamic diameter 55, 91-95 Aerodynamic Particle Sizer R (APS) 318-319, 320, 318-319, 320 Aerosizer 320 Aerosol photometry 128, 131-132 Aerosolisability, reduction 332, 334 Agarwal-Liu criterion 258-259 Agglomeration 43 Air 11 Airborne particles, motion 72-115 Alveolar region 138, 139 see also respiratory tract macrophage cells 139, 167-168, 194 Alveolitis 227 American Conference of Governmental Industrial Hygienists (ACGIH) 209-217,230-231 Analytical instrumentation 205 Animal experiments 170-174 APS see Aerodynamic Particle Sizer Asbestos 51 amosite aerodynamic diameter 95 lung accumulation i77 lung clearance 175 chrysotile 51 inhalation 181 cumulative size distribution 61 electric charge 65 electrostatic forces 162 kinetic models 195 lung penetration 169-170 mixed dust 181 particle shape 47

sampling 223-225,288 Asbestos International Association (AIA) 224 Asbestosis 2, 225 Aspiration 103-107 see also inhalability; inhalation efficiency 104, 144--145, 146-147,240-268 apparent 241,263 electrostatic interference effects 263-264 external wall effects 262-263 particle loss 264-267 turbulence effects 261-262 ventilation inlet 337 sampling 238-241 secondary 263,267 Atomisation 41 Atoms 21-22 Bacteria 67-68 Bends, deposition 266-267 Bernouilli theorem 32 Beta attenuation 327-328 Bias map 222 Bioaerosols 3, 67-70, 329 sampling 295-297 Blow-off 246-247,262-263 Bluffness 106 Bluntness 106 Boltzmann equilibrium 63-64, 263,266, 320 Bolus 149 Bounce 262-263 Boundary layers 27-28 Boyle's law 14 Breathing rate 140, 154 Breathing zone 205,229 British Medical Research Council (BMRC) 213 Brownian motion 107 Byssinosis 226,289

403

Aerosol science for industrial hygienists Cadmium 66 kidneys 170 Cadmium oxide fume 230 Calm air, sampling 107, 256-261 Cancer lung 2, 225,227 nasal 2, 226 Carbon disulphide extraction 397 Carcinogens 225,227 Cascade impactors 214, 291-295 Casella cyclone 1.9 1 min-~ 285-286 Centrifugal force 353 Chemical formation of aerosols 42 Chemical industries, dry aerosol generation 37 Chemotaxis 168 Chrysotile see asbestos Cilia 137, 139, 167, 168 Classification 2-4 Clean technology 5 Clearance mechanical 166-170 chemical 170 Coagulation 43, 111-113 coefficient 112 thermal 111-112 Coal dust, lung burden 183-185 Coal mines optical monitoring 307-310 samplers 275,284, 296-297 Coalescence 43 Collection efficiency 97-99, 356, 361 comparison 381-383 electrostatic precipitators 377-379 filters 363-368, 371 Combustion aerosols, particle size 55-56 Comit6 Europe6n Normalisation (CEN) 21 4-217,222-223 Compartmental models 187-197 Concentration 51-52 measurement 120-122 Condensation 22, 41-42, 43-46 Condensation nuclei particle counters 323-324 PORTACOUNTR 324 Containment 332, 383-385 Continuous liquid-disperse gas 398 Control, aerosols 332-388 Convective diffusion 110-111 Corona discharge 373-374 Cotton dust 226 sampling 289

Coulomb force 159, 160 Cumulative distribution 58, 60-61 Cumulative exposure 199-200 Cunningham slip correction factor 74 Cyclone samplers 284-287 Cyclone separation 353-356 Dalton's law 20 Davies criterion 258-259 Davies model 392 Definition 1-4 Deposition, respiratory tract 147-163 Deutsch model 377-379 Dew point 23 Dielectric properties 116-117 Diesel exhaust lung accumulation 179 lung clearance 175 Differential Mobility Analyser (DMA) 322-323 Diffusion 16-18, 107-115 convective 110-111 efficiency 151 gases 395-397, 398 molecular 107-111 turbulent 113-115 Dimensions xvi-xvii Direct reading instruments, aerosols 304-331 gases 397-398 Disintegration 43 Disperse liquid-continuous gas 398 DMA see Differential Mobility Analyser Dose cumulative 198-199 definition 197 Dosimetry 197-201,227 Drag force 72-77 Droplets equilibrium saturation ratio 44-46 falling speed 82-83 light scattering 131 mechanical generation 40-41 molecular formation 41-42 Dry aerosols, mechanical generation 37-40 Ducts sampling 268-269 ventilation 339-346 Dust 2-3 nuisance 181,225 Dustiness 333 estimation methods 38-40 Dynamic shape factor 76

404

Index EAA see Electrical Aerosol Analyser Electric Single Particle Aerodynamic Relaxation Time Analyser (E-SPART) 320 Electrical Aerosol Analyser (EAA) 322-323 Electrical drift velocity 84-85 Electrical forces, motion 83-85 Electrical migration velocity 376-377 Electrical mobility 84 spectrometers 322-323 Electrical monitoring, gases 397 Electrical properties 62-65 Electrostatic deposition 266 Electrostatic elutriator, split-flow 320-322 Electrostatic forces, respiratory tract 151, 159-162 Electrostatic interference 263-264 Electrostatic precipitation 83-84 Electrostatic precipitators 371-381 collection efficiency 377-379 corona discharge 373-374 diffusion charging 375 localised 384-385 migration velocity 376 particle charging 374-375 permittivity 374 pipe-wire 372 plate-wire 372-373,376-379 practical 379-381 saturation 374-375 Elutriation 99-103,282-283,320-323 Emphysema 225,227 Entry efficiency 241 Equivalent projected area diameter 53-54 Equivalent surface area diameter 53-54 Equivalent volume diameter 53-55 Evaporation 22, 44-46 Evolution of aerosols 42-46 Experiments animals 170-174 deposition 147-149, 150 extinction measurements 125-7 inhalability 143-145,207-209 inhalation 170-174 lung build-up 176-182 lung clearance 174-182 nose-only 171, 173 whole-body 171-172 Exposure assessment 227-229 cumulative 199-200

data 206 history 197,200-201 limit 399 limit values 6-7, 230-236 measurement 7, 8, 204-206 real-time 304-305 Extinction 118-127 Face masks 370, 386 Falling speed 79 FAM see Fibrous Aerosol Monitor Feret diameter 52-53 Fibres see also asbestos aerodynamic diameter 95 concentration 51-52 durable 51 electric charge 65 morphology 51 optical particle counters 315-318 particle size 55 respirable 223-225 respiratory tract deposition 158, 162 sampling 223-225,287-288 Fibrosis 66 Fibrotic response 167 Fibrous Aerosol Monitor (FAM) 315-318 Fick's law 16-17, 107, 261-262,395 Filters 298-299, 362-371 bag-house 369, 371 collection efficiency 363-368, 371 electrostatic forces 369 macroscopic view 362-365 materials 368 particle size of maximum penetration 368 particulate layer 371 personal respiratory protection 332, 370-371,386 pressure drop 368,386 resin wool 369 Filtration 361-371 Fluid mechanics 24-36 Flux 104 Foundries, dry aerosol generation 37 Fractal geometry 50, 55-56 Frequency distribution 57-58 Friability 38 Friction 18, 29 Froude number 91 Fume 3 Fungal spores 67-70

405

Aerosol science for industrial hygienists collection efficiency 99 Industrial processes, adjustments 332, 333-334 Inertia 90 Inertial separation 350-353 Inhalability 143-147,207-211 see also aspiration, efficiency curves 208-211 definition 207 experiments 143-145 sampler results 270-272,274, 275, 276-280, 281 Inhalable fraction 215-219, 274-275, 280-282, 294-295 Inhalation 103, 136-203 aerosols 140-147 efficiency 140 experiments 170-174 gases 392 Inspirability, see Inhalability Institute of Occupational Medicine samplers 2 1 min -1 IOM 278-280, 289-291,297 3 1 min -~ IOM 271-274, 297 PIDS 293-295,297 International standards 214-217 International Standards Organisation (ISO) 208-217 IOM see Institute of Occupational Medicine Ionisation 373

Fused alumina 66 impaction 97 Future prospects 401-402 Gas chromatography 397 Gases 9-10, 389-400 adsorbtion 390-391 air cleaning 398 diffusion 395-397, 398 direct reading 397-398 inhalation 392 kinetic theory 11-15 mixtures 20-21 properties 11-36 sampling 392-397 transport 389-391 Generation 37-42 Gluteraldehyde 400 Granite dust, thermophoretic drift velocity 87 Gravitational elutriation 100 Gravitational parameter 91,266 Gravitational separation 347-350 Gravitational settling 334-336, 340-341 respiratory tract 151, 154 Gravity, effects 77-83,257 Gyration, radius of 55 Harmfulness 198,231 Hatch-Choate equations 60 Hay-fever 226 Hazardousness 399 Health and Safety Executive (HSE) 230 Heavy metals 210 Hexane 399 Hexhlet sampler 283 HSE see Health and Safety Executive Human head aspiration efficiency 144-145, 146 sampler 143-144, 207,251,255 Humidity 23 Hydrazine 399

Kelvin effect 44 Kinematic coagulation 43 Kinetic models, lung clearance 187-197 Kinetic theory of gases 11-15,389-390 Laminar flow 27, 34 elutriation 100-103 Lead 66, 225 dissolution in blood 170 kinetic model 195-196 Light dispersion 134 extinction 118-127 measurement 125-127 monitoring 306-307 polarisation 129-131 scattering 117-118, 127-131, 133 monitoring 307-313 Limit values 66, 230-236 comparison 233-235 definition 205,230

Ideal gas law 14 Image forces 159-160 Impaction 95-99, 242-243 efficiency 96 model 146, 243-247, 254-255,261 parameter 244-245 respiratory tract 149, 150, 152 Impactors 286-287 cascade 214, 291-295

406

Index Monodisperse aerosols 56 Monomorphic exposure distribution 228 Motion airborne particles 72-115 electrical force effects 83-85 gravity effects 77-83 thermal gradients 85-88 without external forces 88 Mouth, breathing 150-151, 153, 155-156 Mucociliary clearance 166-168

Liquid droplet aerosols, mechanical generation 40-41 Liquids 22 Lung 138-139, 174-182 see also respiratory tract burden 178-180, 183-185, 188-191 Lymph nodes 169, 192 dust accumulation 182-185 Macrophage cells see alveolar region Marple cascade impactor 293 Martin diameter 53 Mass balances 325-326 piezoelectric mass balance 325-326 TEOM R 326-327 Mass concentration 120, 131-132, 217-221 Mass frequency distribution 57-58 Mass median particle diameter 58 Mathematical models 162-163 Mean free path (mfp) 15-16 Measurement criteria 207-214 health-related 204-237 Mechanical mobility 79 Mercury Vapor Monitor 396 Mercury vapour 400 Mesothelioma 2, 170, 225 Metals 66--67 mfp see mean free path Microscopy 134-135,224 asbestos sampling 288 Mie intensity parameter 130 Mie theory 118, 124-125 Mineral extraction aerosolising reduction 334 dry aerosol generation 37 particle shape 47 Mining see also coal mines bimodal aerosol 60-61 Mist 3 Molecular diffusion 16-18, 107-111 Molecules 22 gases 389-391 Monitoring see also sampling condensation nuclei 323-324 direct-reading 304-331 electrical 320-323 mechanical 325-327 nuclear 327-328 optical 305-320

Nasopharynx 137 see also respiratory tract National Institute for Occupational Safety and Health (NIOSH) 224, 226, 230 Navier-Stokes equations 24-25 Nebulisation 41 Neutrophil cells 199 Nickel 67, 225 exposure 228 Nose 137, 149-151 see also respiratory tract Nuclear mass detectors 327-328 Nucleation 41-42 Nuisance dust 181,225 Occupational exposure, limit 399 Occupational health 5-9 Occupational Safety and Health Administration (OSHA) 230 Optical microscopy 134-135 Optical monitoring 305-320 light extinction 306-307 light scattering 307-313 particle counters 313-320 Aerodynamic Particle Sizer R (APS) 318-319, 320 Aerosizer 320 Climet 314-316 E-SPART 320 Fibrous Aerosol Monitor (FAM) 315-318 PSL equivalent 315 Royco 314-316 photometers gases 397 light scattering 307-313 OSIRIS 309 Respirable Aerosol Monitor (RAM) 310-313 SIMSLIN 308-310 TM-Digital 309-311

407

Aerosol science for industrial hygienists sensing zone 320 Optical properties 116-135 extinction coefficient 122-7 extinction concept 118-22 microscopy 134-5 physical basis 116-118 scattered light distribution 127-32 usual appearance 132-4 Optical Scattering Instantaneous Respirable Dust Indication System (OSIRIS) 309 Organic Vapor Monitor 396 Overload, lung 178, 185-187 Ozone 372

Particles aerodynamic diameter 91-95,318 beta 327-328 biological 226 chemistry 66-67 concentration 51-52 counting condensation 323-324 optical 313-320 deposition, ducts 339-346 diameter aerodynamic 91-95, 155, 162 equivalent projected area 53-54 equivalent surface area 53-54 equivalent volume 53-55 mass median 58 dielectric properties 116-117 drag force 72-77 electric charge 62-65, 83-85 electrical charge 374-375 electrical measurement 320-323 electrical migration velocity 376-377 extinction coefficient 119, 122-125, 126 inertia 90 inhaled 166-203 insoluble 66, 166-167, 174-181 labelling 149 large 207-211 motion 72-115 multi-modal distribution 60--61 optical counting 313-320 perfectly-mixed 334, 340 relaxation time 80 removal systems 346-383 comparison 381-383 sampling criteria 225-227 scattering coefficient 127, 129-131

separation 332, 346-383 cyclone 353-356 electrostatic 371-381 filtration 361-371 gravitational 347-350 inertial 350--353 wet scrubbers 356-361 shape 46-50 size 3, 91-95, 154, 368 estimate 134 measurement 52-56 parameter 122 statistics 56--61 size-selective criteria 215-217, 226, 232 soluble 66, 181-182, 195-196 spherical 46-47, 52 Stokesian 73 trajectories 96-98, 246, 257 ventilation ducts 339-346 workplace behaviour 334--336 Particulates not otherwise classified (PNOC) 225 PCM see phase contrast microsopy Peclet number 111,266 Penetration 103, 142 Personal inhalable dust spectrometer (PIDS) 293-295,297 Personal samplers see samplers Phagocytosis 168 Phase contrast microscopy (PCM) 224 Phase transitions 21-23 Photometers see optical monitoring Photometry 128, 131-132, 307-313 see also optical monitoring Photoreproductive toner lung accumulation 180 lung clearance 176 PIDS see Personal inhalable dust spectrometer Piezoelectric mass balance 325-326 Pneumoconiosis 2, 213,225,227 POCK model 194-195 Polarisation, of light 129-131 Pollutants 1, 5 Polycyclic aromatic hydrocarbons 210 Polydisperse aerosols 46, 56 extinction coefficient 122 Potential flow 31 Pressure 32 Properties 37-71 Protection, personal 332, 370-371,386 Pumps 297-298

408

Index inhalation 103, 136-165 lung burden 177-180, 183-185, 188-191 lymph nodes 169, 182-185, 192 penetration 142-143, 155-156, 213 Restitution coefficient 262 Retention function 198 Reynolds' number 29-30, 139, 340 Rock dust 181

Quartz 66-67 aerodynamic diameter 93-94 falling speed 81-82 harmfulness 199 lung accumulation 176 lung clearance 174 lung sequestration 192 lymph node accumulation 182-183 mixed dust 181

Safety in Mines Light Scattering Instrument (SIMSLIN) 308-310 Samplers 205,269-299 see also impactors; optical monitoring; sampling active 392, 393-395 aspiration efficiency 240-268 bioaerosol 297 blunt samplers 251-256 components 297-299 cyclone 284-287 denuders 393-394 design 267-268 efficiency 270-272, 276-277 electrostatic interference 263-264 filters 298-299 gases 392-397 human head 143-144, 207, 251,255 inhalable fraction 274-275,280-282 measurement criteria 239 open filter 270-271,275-276 orientation effects 247-251,254, 255 oversampling 263 particle loss 264-267 passive 392-393,394-397 performance 267-268 performance assessment 221-223 performance indices 239-241 personal samplers 229,275-282,284-287 3M T M 3500 Organic Vapor Monitor 396 3M T M 3600 Mercury Vapor Badge 396 37mm cassette 276-278 Casella cyclone 1.9 1 min -1 285-286 CIP10 280-281,285-287 gas 393-397 IOM 2 1 min -1 278-280, 289-291,297 IOM PIDS 293-295,297 Marple cascade impactor 293 PERSPEC 281-282, 289 P O R T A C O U N T R 324 seven-hole 275-277, 297 single-hole 275-277 pumps 297-298

Radioactive material 210, 226 Rainbow 134 RAM see Respirable Aerosol Monitor Rats, experimental use 170-174 Rayleigh theory 118, 124, 129 Real-time sampling 304-305 Refractive index 117, 135 Resin wool filter 369 Respirable Aerosol Monitor (RAM) 310-313 Respirable fraction 213, 216-219, 225 monitoring 308, 311,312 samplers 282-288 Respiration, efficiency 142 Respiratory tract 136-140 aerosol inhalation 140-147 air volume 139 bioaerosol effects 296 build-up studies 173, 176-182 clearance 166-170, 174-182 alveolar 167-169 dissolution 170 kinetic models 187-197 mucociliary 166-168 sequestration model 190-194 studies 173, 174-182 tracheobronchial 166-167 defence mechanisms 137-140 deposition 140-142, 147-163 alveolar 142, 151, 153-155, 159-162, 212-214, 226-227 electrostatic 159-162 experiments 147-149, 150 extrathoracic 140, 149-151 fibres 158 mathematical models 162-163 thoracic 142, 151-156, 211-212 total 142, 156-157 tracheobronchial 141, 151, 152-153, 212-214, 226 flow rate 140 gas inhalation 392

409

Aerosol science for industrial hygienists respirable fraction 282-287 static 229,269-275,282-284 CPM3 284 Hexhlet 283 IOM 3 1 min-l 271-274, 297 MRE Type l13A 274-275,282-283,308 TBF50 284 thin-walled probe 241-251,265 undersampling 265 Sampling 238-303 see also monitoring; samplers acceptable 258-259 aspiration 238-241 bioaerosols 295-297 calm air 256-261 ducts 268-269 efficiency 241-268 fibres 287-288 fractions 211-214, 215-217 frequency 227-228 gases 392-397 isokinetic 243,268 quantitation 299 real-time 304-305 respirable fraction 282-287 size-selective 225-226 stacks 268-269 strategy 205 sub-fractions 289-291 thoracic fraction 289 total aerosol 207-208,225,232, 269-274, 275-280 Saturation ratio 23, 41 Saturation vapour pressure 23,398 Sedimentation speed 79 Self-similarity 50 Separation 32-33 Sequestration model 190-194 Shear 18 Silica, see Quartz SI System of units xvi Similarity 28-31, 89-91 Slip 73-74 Smoke 3 colour 133 particle size 55-56 Smokers, tracheobronchial deposition 152 Solids 22 Solvents, organic 394 Soot, colour 133 Space charge 159

Spectrometers 291-295 DMA 322-323 EAA 322-323 Spray 3 Spray towers 356-359 Spraying 41 Stacks, sampling 268-269 Stagnation 32 Standards 225-236 criteria 204-205, 214-225 definition 204-205 Static pressure 32 Stokes' law 73-77, 357 Stokes' number 89-90, 245,258, 338, 346, 351 Stop distance 88, 146 Streamlines 25-27, 96 limiting 240 thin-walled sampling probe 242-243 Sulphur dioxide, inhaled 392 Surface tension 22 Tangents, method 321-322 Tapered Element Oscillating Microbalance (TEOMR) 326-327 TEOM see Tapered Element Oscillating Microbalance Textile industries dry aerosol generation 37 particle shape 47-50 Thermal coagulation 111-112 Thermal desorption 397 Thermal gradients 85-88 Thermophoresis 85 Thermophoretic velocity 86-87 Thin-walled probe samplers 241-251,265 Thoracic fraction 211-212, 215-219 sampling 289 Thoracic region 138 see also respiratory tract Titanium dioxide 66 lung accumulation 178 lung sequestration 192 lymph node accumulation 182-183 Tomographic measurement 126 Total aerosol, sampling 207-208, 225,232, 269-274, 275-280 Toxicity 66 Trajectories 96-98, 246, 257 Trimellitic anhydride 400 Tubes, deposition 266

410

Index Turbidity 119 Turbulence 34-36, 113-115,261-262, 266 Turbulent deposition 341-343 Turbulent diffusion 113-115 Tyndall spectra 134

Visual appearance 132-134 Volume scattering function 131-132 Walton-Beckett graticule 288 Water droplets falling speed 82-83 light scattering 131 Welding fume colour 133 particle shape 50 Wet scrubbers 356-361 energy efficiency 361 spray towers 356-359 venturi scrubbers 359-361 Wind tunnel 143-144 Windspeed 208,210-211,275 Wood dust 226 thermophoretic drift velocity 87 Woodworking industries, dry aerosol generation 37 Workers personal respiratory protection 332, 370-371,386 ventilation hoods 339 Workplaces aerosol behaviour 334-336 aerosol control 332-388 aerosol generation 37-42 aerosols 5-9 monitoring 304-331 sampling 238-303

Ulceration 226 Units and dimensions xvi-xvii

Vapour hazard ratio 399-400 Vapour pressure 398-399 Vapours 22-23,389-400 Velocity pressure 32 Vena contracta 265 Ventilation 333-346 ducts 339-346 penetration efficiency 341,344-345 turbulent deposition 341-343 velocity 346 general exhaust ventilation (GEV) 332, 335 inlet reach 336-337 local exhaust ventilation (LEV) 332, 336-339 worker effects 339 Venturi scrubbers 359-361 Viscosity 18-19 Visibility 119 index 135

411

This Page Intentionally Left Blank