EPA-600/1-76-027a
August 1976 Environmental Health Effects Research Series
OZONE AND OTHER PHOTOCHEMICAL OXIDANTS
Volume 1 (Chapters 1-7)
T TRRAKY
11 q ENVIRONMENTAL PROTECTION
Health Effects Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
Ozone and Other
Photochemical
Oxidants
Volume 1 of 2 Volumes
Subcommittee on Ozone and Other Photochemical
Oxidants
Committee on Medical and Biologic Effects of Environmental Pollutants
Assembly of Life Sciences
-------
Ozone
and Other
Photochemical
Oxidants
Volume 1 of 2 Volumes
Subcommittee on Ozone and Other Photochemical Oxidants
Committee on Medical and Biologic
Effects of Environmental Pollutants
National Research Council
LIBRAPY
NATIONAL ACADEMY OF SCIENCES
Washington, D. C. 1976
-------
NOTICE
The project reported on here was approved by the
Governing Board of the National Research Council, whose
members are drawn from the Councils of the National
Academy of Sciences, the National Academy of Engineering,
and the Institute of Medicine. The members of the commit-
tee responsible for the report were chosen for their
special competences arid with regard for appropriate repre-
sentation of experience and disciplines. The findings and
conclusions presented are entirely those of that committee.
This report has been critically reviewed according to
procedures approved by a Report Review Committee consisting
of members of the National Academy of Sciences, the National
Academy of Engineering, and the Institute of Medicine. Only
after completion of the review process has it been released
for publication.
The work on which this publication is based was performed
pursuant to Contract No. 68-02-1226 with the Environmental
Protection Agency.
-------
SUBCOMMITTEE ON OZONE AND OTHER PHOTOCHEMICAL OXIDANTS
SHELDON K. FRIEDLANDER, California Institute of Technology, Pasadena, Chairman
BERNARD ALTSHULER, Institute of Environmental Medicine, New York University
Medical Center, New York
KYLE D. BAYES, Department of Chemistry, University of California, Los Angeles
ALAN Q. ESCHENROEDER, Environmental Research and Technology, Inc., Santa
Barbara, California
JACK D. HACKNEY, University of Southern California, Downey
WALTER W. HECK, Agricultural Research Service, U.S. Department of Agriculture, North
Carolina State University, Raleigh
JAMES R. MC CARROLL, Medical Services Division, City of Los Angeles, California
JAMES R. MC NESBY, Office of Air and Water Measurement, National Bureau of
Standards, Washington, D.C.
PAUL R. MILLER, Pacific Southwest Forest and Range Experiment Station, U.S. Forest
Service, Berkeley, California
PETER K. MUELLER, Environmental Research and Technology, Inc.. Westlake Village,
California
SHELDON D. MURPHY, Department of Physiology, Harvard School of Public Health,
Boston, Massachusetts
Consultants
KARL A. BELL, Environmental Health Department, Rancho Los Atnigos Hospital,
Downey, California, and Departments of Environmental and Chemical Engineering,
School of Engineering, University of Southern California, Los Angeles
BERNARD D. GOLDSTEIN, Institute of Environmental Medicine, New York University
Medical Center, New York
DANIEL GROSJEAN, Statewide Air Pollution Research Center, University of
California, Riverside
MARGARET HITCHCOCK, Yale University School of Medicine, New Haven, Connecticut
JOHN B. MUDD, Department of Biochemistry, University of California, Riverside
MARSHALL WHITE, University of California, Berkeley
JAMES A. FRAZIER, Division of Medical Sciences, National Research Council,
Washington, D.C., Staff Officer
-------
COMMITTEE ON MEDICAL AND BIOLOGIC EFFECTS OF ENVIRONMENTAL POLLUTANTS
HERSCHEL E. GRIFFIN, Graduate School of Public Health, University of Pittsburgh,
Pennsylvania, Chairman
RONALD F. COBURN, Department of Physiology, University of Pennsylvania School
of Medicine, Philadelphia
T. TIMOTHY CROCKER, Department of Community and Environmental Medicine,
University of California College of Medicine, Irvine
CLEMENT A. FINCH, Department: of Hematology, University of Washington, Seattle
SHELDON K. FRIEDLANDER, W. M. Keck Laboratories, California Institute of
Technology, Pasadena
ROBERT I. HENKIN, Department of Pediatrics, Georgetown University Hospital,
Washington, D.C.
IAN T. T. HIGGINS, School of Public Health, University of Michigan, Ann Arbor
JOE W. HIGHTOWER, Department of Chemical Engineering, Rice University, Houston,
Texas
HENRY KAMIN, Department of Biochemistry, Duke University Medical Center,
Durham, North Carolina
ORVILLE A. LEVANDER, Nutrition Institute, Agricultural Research Center,
Beltsville, Maryland
DWIGHT F. METZLER, Kansas State Department of Health and Environment, Topeka
I. HERBERT SCHEINBERG, Department of Medicine, Albert Einstein College of
Medicine, Bronx, New York
RALPH G. SMITH, Department of Environmental and Industrial Health, School
of Public Health, University of Michigan, Ann Arbor
ROGER P. SMITH, Department of Pharmacology and Toxicology, Dartmouth Medical
School, Hanover, New Hampshire
T. D. BOAZ, JR., Division of Medical Sciences, National Research Council,
Washington, D.C., Executive Director
-------
ACKNOWLEDGMENTS
This document was written by the Subcommittee on Ozone and Other
Photochemical Oxidants under the chairmanship of Dr. Sheldon K. Friedlander.
The members of the Subcommittee and its consultants were chosen for their
competence to prepare sections of the report. The entire document was
critically reviewed by the Subcommittee, and it represents the combined
effort and cooperation of all its members and consultants.
The authors of the individual sections were as follows: Dr. Kyle D.
Bayes, the material on chemical origin; Dr. Daniel Grosjean, on aerosols;
Dr. Alan Q. Eschenroeder, on atmospheric concentrations of photochemical
oxidants and models for predicting air quality; Dr. Peter K. Mueller, on
measurement and methods; Drs. Karl A. Bell and Bernard Altshuler, on respira-
tory transport and absorption; Dr. Sheldon D. Murphy in collaboration with
Drs. Bernard D. Goldstein and Margaret Hitchcock, on toxicology; Dr. Jack D.
Hackney, on controlled studies on humans; Dr. James R. McCarroll, on
epidemiologic studies; Dr. Walter W. Heck in collaboration with Drs. John B.
Mudd and Paul R. Miller, on plants and microorganisms; Dr. Miller in
collaboration with Dr. Marshall White, on ecosystems; and Dr. James R. McNesby,
on effects of photochemical oxidants on materials. Dr. Friedlander prepared
the executive summary.
The document was reviewed by the Report Review Committee of the National
Academy of Sciences; by the parent Committee on Medical and Biologic Effects
of Environmental Pollutants (MBEEP); by the Associate Editor, Dr. Ronald F.
Coburn, and several anonymous reviewers; by the Advisory Center on Toxicology
of the Assembly of Life Sciences (ALS); by the Committee on Atmospheric
Sciences of the Commission on Natural Resources' Environmental Studies Board;
and by the Committee on National Statistics of the Assembly of Mathematical
and Physical Sciences.
-------
The Subcommittee is indebted to Mr. James A. Frazier, staff officer
in the ALS Division of Medical Sciences, for his special efforts and
assistance. The report was edited by Mr. Norman Grossblatt, Editor for
the Assembly of Life Sciences. This is the largest of the MBEEP reports
yet produced, and we wish to acknowledge completion of the task under
difficult time constraints., We also acknowledge the editorial assistance
of Mrs. Renee Ford on one of the chapters.
-------
COMMITTEE ON MEDICAL AND BIOLOGIC EFFECTS OF ENVIRONMENTAL POLLUTANTS
Division of Medical Sciences, National Research Council
Assembly of Life Sciences
SUBCOMMITTEE ON OZONE AND OTHER PHOTOCHEMICAL OXIDANTS
OUTLINE OF THE REPORT
1. EXECUTIVE SUMMARY
2. CHEMICAL ORIGIN
3. AEROSOLS
4. ATMOSPHERIC CONCENTRATIONS OF PHOTOCHEMICAL OXIDANTS
5. MODELS FOR PREDICTING AIR QUALITY
6. MEASUREMENT METHODS
7. RESPIRATORY TRANSPORT AND ABSORPTION
8. TOXICOLOGY
9. CONTROLLED STUDIES ON HUMANS
10. EPIDEMIOLOGIC STUDIES
11. PLANTS AND MICROORGANISMS
12. ECOSYSTEMS
13. EFFECTS OF PHOTOCHEMICAL OXIDANTS ON MATERIALS
14. GENERAL SUMMARY AND CONCLUSIONS
15. RECOMMENDATIONS FOR FUTURE RESEARCH
-------
Chapter 1
EXECUTIVE SUMMARY
1
In the early 1950's, it was reported by Haagen-Smit that many of
the characteristics of photochemical smog could be explained by the presence
of ozone and other photochemical oxidants. These substances, he believed,
were formed in the atmosphere as a result of chemical reactions involving
nitrogen oxides and hydrocarbons present in automobile exhaust. Signif-
icant quantities of nitrogen oxides were also emitted by power plants.
Considerable time elapsed before there was general acceptance of
Haagen-Smit!s important discovery, in part because of its subtle nature.
For the first time, a major air pollution problem was demonstrated to be
caused by a pollutant generated in the atmosphere. Its effect often did
not become apparent until many miles downwind from the source. (The same
suspicion has been attached to sulfate-containing aerosols for many years,
but the proof that the sulfate is damaging is not as well established.)
In addition, a new pollution source, automobile exhaust, had been shown
to be of prime importance.
After the pioneering studies of Haagen-Smit, an extensive scientific
literature developed on the properties, measurement, and effects of photo-
chemical smog. The attempt to control engine emission has had a profound
effect on the automobile and petroleum industries. Estimated costs and
associated benefits of automobile emission control each run into the
2
billions per year.
By the Clean Air Act Amendment of 1970, Congress set automobile
emission standards and instructed the Environmental Protection Agency (EPA)
to set ambient air quality standards. Included in the Act was a requirement
-------
to review the standards periodically. It is hoped that this report will
be of value to both Congress and the EPA in discharging their responsibilities
for the review of these standards.
However, the Subcommittee on Ozone and Other Photochemical Oxidants
did not discuss—and does not necessarily endorse—the adoption of fixed
federal standards as the prime approach to pollution control. The Sub-
committee also did not attempt to determine the concentration at which the
standard should be set, exce.pt to recognize the difficulty of arriving at
such a number. There was, however, general skepticism concerning the
applicability of the concept: of the threshold concentration - the concentration
below which there are no biologic effects.
This report deals primarily with the origins and effects of ozone
and other photochemical oxiclants. It is limited, more or less, to the
problem of urban pollution and to such closely related topics as natural
background in the earth's boundary layer. No consideration is given to the
stratospheric ozone layer and the effects produced by SST emissions or
halocarbons.
The reference method recommended by the federal government for the
determination of oxidant measures ozone, which serves as an indicator of
photochemical smog. Other agents formed in the photochemical system include
a variety of free radicals in the gas phase and sulfates, nitrates, and
oxygenated organic compounds in the particulate phase. A measurement of
ozone, alone, provides only limited information on the concentrations of
the other agents, because of the complex chemical and mixing processes
involved. How these other agents form and what their effects are remain
poorly understood. Recommendations on studies involving such agents will
be found throughout the report.
*Supersonic Transport
1-2
-------
We still lack an adequate dose-response relationship for humans
exposed to ozone, particularly at concentrations less than about 0.2 ppm.
The data base for the development of such a relationship for both short-
and long-term exposures is inadequate. Although some data from controlled
studies are available for concentrations above 0.3 ppm, methods for extra-
polating to lower concentrations are needed. Moreover, it is not clear
how to weight the results of pulmonary function tests on humans, animal
studies, and epidemiologic studies in a general dose-response relationship.
Despite uncertainties concerning the causative agents and their
effects, we must proceed with the regulation of emissions that lead to the
formation of photochemical smog. At the same time, research should continue
on identifying the individual harmful agents in photochemical smog and
their effects. Otherwise, there is danger of focusing on an indicator
(ozone) while the formation and behavior of associated pollutants, which
create a major part of the problem, are not adequately understood.
Roughly speaking, the first third of this report is concerned with
the origins and measurement of ozone and other photochemical oxidants and
the relationship of atmospheric concentrations to emissions. The middle
third deals with toxicologic studies and effects on humans, and the last
with effects on plants, ecosystems, and materials.
Each chapter is accompanied by a summary and/or a set of recommen-
dations prepared by the individual author. Some of the most important points
have been identified in the Executive Summary, particularly those which cut across
several fields. In some cases, the recommendations have been abstracted
directly from the other chapters. For detailed recommendations, however,
the reader should turn to the separate chapters.
1-3
-------
ORIGINS AND MEASUREMENT
The extensive scientific literature covering the chemistry of smog
reactions is reviewed in Chapter 2. Even in the case of a single hydro-
carbon with typical concentrations of the oxides of nitrogen, carbon
monoxide, water vapor, and other trace components, several hundred chemical
reactions take place. The urban atmosphere contains not just one but hundreds
of different hydrocarbons, each with its own reactivity and oxidation pro-
ducts. Most of the reaction mechanisms and rate constants needed to con-
struct realistic models of polluted atmospheres have been determined in
laboratory studies under carefully controlled conditions. Serious gaps
remain in the present models, and further fundamental research on kinetics
and mechanisms is necessary. For example, rate constants are needed for
almost all the reactions of hydroperoxy and alkylperoxy radicals. The
homogeneous and heterogeneous reactions of the oxides of nitrogen with
water also need study.
Smog-chamber studies are needed for validating both detailed chemical
models and lumped models. Measurements of more products and the reactive
intermediates, including such free radicals as hydroxyl and hydroperoxy,
will provide more stringent tests for models. There are useful inter-
actions among modeling studies, smog-chamber experiments, atmospheric measurements,
and fundamental chemical kinetics; it is not possible to ignore one without
hindering progress in the others.
The possibility that free radicals, particularly hydroperoxy, have
significant effects on biologic surfaces should be investigated.
The available information on aerosol formation in photochemical
smog is reviewed in Chapter 3. The story told there is still not complete,
1-4
-------
but there is evidence that reactions involving ozone contribute signifi-
cantly to the formation of both the organic and sulfate-containing compon-
ents of the aerosol. Laboratory studies show that both cyclic olefins and
C diolefins are efficient aerosol precursors that lead to the formation
DT
of difunctional oxygenated organic compounds (such as dicarboxylic acids)
of low vapor pressure. These compounds have also been found in the smog
aerosol. Cyclic olefins have been identified in both gasoline and auto
exhaust and might be an important source of secondary aerosol organics;
there are no known sources of diolefins. The role of aromatics as aerosol
precursors is not understood.
Aerosol organics in the atmosphere could be reduced by control
of emission of nitrogen oxides and total hydrocarbons. However, the iden-
tification and control of a few specific aerosol precursors in gasoline
and other sources might prove a more efficient approach.
Our knowledge of the chemical and physical processes that govern
aerosol formation in the atmosphere is limited, and further research in
the field is badly needed. Attention should be focused on laboratory
studies of aerosol formation from aromatic hydrocarbons. The concentrations
of aerosol precursors in the atmosphere should be determined; more data
on organic compounds in ambient aerosols are needed to estimate the rela-
tive importance of olefinic and aromatic hydrocarbons as aerosol precursors.
The health effects of difunctional oxygenated organic compounds
should be investigated in both animal and human studies.
A critical question concerning atmospheric concentrations of ozone
and other photochemical oxidants is: "What fraction of the observed values
in each locale can be controlled by reduction of emissions?" Some; contend
1-5
-------
that natural background concentrations exceed the federal ambient air
quality standard (0.08). Another point of view is that background ozone
concentrations rarely exceed about 0.05 - 0.06 ppm at the surface and that
higher concentrations are caused by man-made sources.
The data reviewed in Chapter 4 support the second point of view.
Measurements in remote areas of the Northern Hemisphere, when compared with
those in the lower 48 states of the United States, support the contention
that man-made sources are involved in cases where the standard is exceeded.
Further measurements are needed to establish this contention with more
certainty. Some of the difficulties involved in such studies become
apparent when it is noted that the effect of pollution—particularly nitric
oxide emissions—-is to reduce ozone concentrations locally.
Theoretical interpretation of the experimental observations will
help in determining the relative roles played by stratospheric injection,
plant emissions, background methane, and transport to surfaces in the
natural portion of the tropespheric ozone cycle.
The most complete data on ozone and other oxidant concentrations
have been obtained for the Los Angeles air basin, because of the severity
of the problem there. Further measurements are needed in the central and
eastern areas of the United States, to broaden the foundations of a national
control strategy. Such studies should be designed with specific goals in
mind, and not carried out as routine monitoring exercises.
Rational air pollution control strategies require the establishment
of reliable relationships beitween air quality and emissions.(Chapter 5).
Diffusion models for inert (nonreacting) agents have long been used in air
pollution control and in the study of air pollution effects. Major advances
1-6
-------
have been made in incorporating the complex chemical reaction schemes of
photochemical smog in diffusion models for air basins. In addition to these
deterministic models, statistical relationships that are based on aero-
metric data and that relate oxidant concentrations to emission measurements
have been determined.
Improvements in deterministic (photochemical/diffusion) methods are
based largely on accounting for more physicochemical effects in the structure
of the model. Specific research subjects for improved models include
photochemical aerosol formation and the effects of turbulence on chemical
reaction rates. The challenge to the researcher is to incorporate the
study of these subjects without needlessly complicating already complex
models. How accurate a mathematical simulation is required? What, roughly,
will be the effect of omitting some particular chemical or physical compo-
nent? What is the sensitivity of model outputs to inaccuracies in the inputs?
One of the most important contributions of research in this field
will be the development of criteria to define the limits of applicability
of existing models, rather than a single supermodel that will incorporate
all effects.
Specific goals are essential in model development and in data
collection for model-testing. Examples of goals are the determination of
oxidant isopleths and the relating of visibility degradation to emission
sources. Monitoring programs should be designed with specific goals of data
analysis or modeling. It should not be expected that from the data alone
useful information will emerge directly or that someone else will spontan-
eously dig out the important results. Two important steps that can be
undertaken by those who produce models to encourage application and aid
the user are the compilation of a catalog of air quality models that
1-7
-------
describes their capabilities in terms of a common set of performance stan-
dards and the improvement of model output to permit easy access by the user.
Methods of measuring the components of photochemical smog are
reviewed in Chapter 6. There have been significant advances in the
calibration of instruments for monitoring ozone in ambient air. A method
based on the absorption of ultraviolet radiation at 254 nm has been
adopted by California for the calibration of air monitoring instruments.
The method is based on the use of a commerically available instrument that
measures ultraviolet absorption as a transfer standard in the calibration
process.
It is important to separate, conceptually and in practice, the
calibration process from the monitoring process. Photochemical oxidants
consisting primarily of ozone were continuously monitored first in southern
California by measuring the color change of potassium iodide solutions
brought into contact with ambient air. This measurement continues to
yield valid photochemical oxidant data in California. However, it has
yielded questionable data at ambient air monitoring sites elsewhere in
the United States. For this reason, at the end of 1971, the EPA adopted
a continuous monitoring process that involves the measurement of the
chemiluminescence produced when ozone in air is brought into contact with
ethylene. When it is calibrated with the ultraviolet-absorption method,
this reference procedure for monitoring ozone in ambient air is widely
accepted. The evaluation of nationally applicable primary calibration
procedures for ozone measurement should continue.
Instruments based on differential ultraviolet absorption still
need to be evaluated, and possibly modified, before their acceptance for
1-8
-------
monitoring ozone in polluted atmospheres on a nationwide scale. The
California Air Resources Board and other air pollution control agencies
are evaluating ultraviolet absorption side-by-side with chemiluminescence
and potassium iodide instruments.
There is no commercially available instrument for the continuous
monitoring of any of the chemical species present in the particulate
component of photochemical smog. Methods should be developed for the
direct and continuous measurement of such species. Species of interest
include sulfates, nitrates, some oxygenated organic compounds, and lead.
HEALTH EFFECTS
A great deal is known about the deposition of aerosol particles in
the lung and their later clearance. Less is known about the uptake of
gases such as ozone and other oxidants that can react with biopolymers
in the mucous and tissue layers. Such information is important in under-
standing the site and mechanism of pollutant gas action in humans and the
effects of copollutants like nitrogen dioxide and ozone, and in the extra-
polation of dose-response data from animals to humans. What has been done
in this field is reviewed in Chapter 7, which also discusses the information
necessary for improved understanding of the transport process.
The solubility of the gas is important. For example, experimental
data from studies carried out with dogs show that nearly 100% of highly
soluble sulfur dioxide inhaled through the nose is removed before reaching
the first bifurcation in the lung, whereas 27-70% (depending on initial
concentration) of ozone, which is less soluble in water, is removed in the
same region. In addition to solubility, chemical reactions in the surface
layers are of great importance.
1-9
-------
A concerted effort is needed to increase our understanding of the
transfer and uptake of reactive gases in the lung. A program in this
field should involve in vitro model studies, animal experiments, and
clinical studies. More information is required on the chemical, physical,
and morphologic properties of the mucous layer and the kinetics of the
reactions of ozone in the mucous and tissue layers. Experimental data on
uptake and dosage for ozone and other oxidants are difficult to obtain for
the tracheobronchial and pulmonary regions. Such data for animals and
humans will be needed to test the present simple transport models,
before further refinements are made.
Toxicologic research (Chapter 8) on the effects of ozone in
laboratory animals has demonstrated that exposure to airborne ozone at
less than 1 ppm for a few hours produces numerous changes in cell and
organ structure and function. The lowest concentrations that produce
these changes differ somewhat among different species of laboratory animals
and with the effect under observation. However, several functional and
morphologic indexes of response to ozone are altered with exposures to
concentrations of about 0.2 - 0.5 ppm over periods ranging from a few
minutes to several weeks.
Recent studies involving repeated or prolonged exposures of
laboratory animals to ozone have suggested that changes indicative of
chronic lung disease (such as decreased elasticity of the lungs) also
require concentrations of 0.2 - 0.5 ppm.
Exposures to ozone Eor a few hours result in a marked increase in
the susceptibility of animals to controlled doses of infectious organisms
introduced into the lung. This is the most sensitive test of any yet
1-10
-------
reported; significantly increased susceptibility of mice to one micro-
organism occurred after exposure to ozone at a concentration as low as
0.08 ppm. Other reports with different microorganisms or different species
have suggested that somewhat higher concentrations are required. These
findings suggest the need for carefully planned epidemiologic studies on
the incidence of lung infection in human populations exposed to oxidant
air pollution. How do such studies relate to reported cases of human
adaptation to long-term oxidant exposure?
Extrapulmonary effects have also been observed in laboratory
animals at concentrations of about 0.2 ppm. These include reduced
voluntary activity, chromosomal aberrations in circulating lymphocytes of
hamsters, increased neonatal mortality, and greater incidence of jaw
abnormalities in offspring of mice exposed to ozone. The mechanisms of
these effects are largely unknown. Reports of chromosomal aberrations in
hamsters and of mutagenic activity of ozone in microorganisms and tissue
cultures raise the question of a possible genetic or carcinogenic hazard.
This should be tested experimentally and epidemiologically.
There is evidence that nutrition affects animal response to ozone.
Increased susceptibility has been reported in animals deficient in
vitamin E—or the converse (protection conferred by administration of
vitamin E).
Convincing new information on the health effects of oxidant
exposure has emerged from controlled studies on humans, from which tenta-
tive dose-response curves have been constructed. These data are reviewed
in Chapter 9, with the types of experimental facilities now available for
such measurements. The new data show reduced pulmonary function in
1-11
-------
healthy smokers and nonsmokers after exposure to ozone at 0.37 ppm and
higher for 2 h. (The federal standard is 0.08 ppm for a 1-h exposure.)
Other gases and aerosols found in an urban atmosphere were not present in
these experiments.
With various tests of ventilatory function, it has been shown that
healthy male college students experienced no effect of sulfur dioxide at
0.37 ppm, a 10% decline in function with ozone at 0.37 ppm, and a 20-40%
decline in function with a combination of sulfur dioxide at 0.37 ppm and
ozone at 0.37 ppm. Other experiments have suggested an adaptation of
southern Californians to chronic exposure to ambient ozone.
Further studies are needed to give better dose-response informa-
tion, and to provide a frequency distribution of the population response
to oxidants alone and in combination with other pollutants at various
concentrations. Such studies should include the effects of mixed
pollutants over ranges corresponding to the ambient atmosphere. The
mixtures should be carefully characterized to be sure of the effects of
trace pollutants on sulfate aerosol formation. The design of such studies
should permit extrapolation from animals to humans and from small groups
of humans to populations. Further research on the possibility of human
adaptation to chronic exposure to oxidants is desirable.
Safety, ethical, and legal considerations require that the utmost
care be exercised in human experimentation. The risk inherent in this
work can be minimized by the proper design of facilities for human
exposure to reactive gases, such as ozone and sulfur dioxide, and reactive
gas mixtures. Standards for the exposures of humans to such controlled
1-12
-------
atmospheres should be discussed by national groups and agencies, such as the
American Medical Association, and the National Institutes of Health.
Studies of the reactions of population groups to photochemical smog
are reviewed in Chapter 10. Such studies played a major role in the
establishment of the current federal standards. Included were eye irritation
studies, effects on asthmatics, and the responses of groups of high school
athletes. Uncertainties in the design of these experiments and interpretation
of the data make further epidemiologic studies essential.
Two major studies are being conducted by the EPA in Los Angeles on
the effects of photochemical oxidants on health. The first is a survey of
schoolchildren in seven communities representing different degrees of oxidant
exposure. In addition to rather detailed environmental monitoring data,
specific health characteristics will be followed, including chronic respiratory
disease in adults, lower respiratory disease in children, acute respiratory
disease in both children and adults, pulmonary function in children, aggravation
of asthma, irritation of mucous membranes, and tissue residues of trace metals.
Complete data from this study will not be available for another 5 years.
The second study is only beginning and will attempt to correlate the
effects of photochemical oxidants and cigarette-smoking in promoting chronic
respiratory signs and symptoms in cohorts of adolescents and their families.
Pulmonary function tests will be included.
These studies are being carried out by EPA as part of the CHESS program.
The results of these studies should be released as soon as possible for evalua-
tion by the general scientific community. This will permit the design and
initiation of additional studies with modifications where necessary to supplement
what has been done. The continuation of epidemiologic studies, including those
of the CHESS program, is vital to our understanding of the effects of air
pollution on health.
1-13
-------
Other epidemiologic studies should be designed to seek analogues
in human populations of effects observed in toxicologic and clinical
studies, including the results of pulmonary function tests, evidence of
chronic lung disease in animals, and evidence of increased susceptibility
to microbial infection.
EFFECTS ON PLANTS. ECOSYSTEMS, AND MATERIALS
The major phytotoxic components of the photochemical oxidant
system, discussed in Chapter 11, are ozone and peroxyacetylnitrate (PAN),
but there is indirect evidence that other phytotoxicants are present.
Considerable effort has gone into controlled exposures to ozone and into
field studies. Leaf stomata are the principal sites for ozone and PAN
entry into plant tissue. Closed stomata will protect plants from these
oxidants. Both ozone and PAN may interfere with various oxidative
reactions within plant cells. Membrane sulfhydryl groups and unsaturated
lipid components may be primary targets of oxidants. Young leaf tissue is
more sensitive to PAN; newly expanding and maturing tissue is most sensi-
tive to ozone. Light is required before plant tissue will respond to PAN;
that is not the case with ozone.
Oxidants reduce yields of many plants, especially sensitive
cultivars. Chronic exposures to concentrations between 0.05 and 0.15 ppm
will reduce soybean, corn, and radish yields. The threshold appears to be
between 0.05 and 0.1 ppm for some sensitive cultivars—well within values
monitored in the eastern United States. Growth or flowering effects on
carnation, geranium, radish, and pinto bean have been found at chronic
1-14
-------
exposures to ozone at 0.05 - 0.15 ppm. Estimated costs to consumers
of agricultural losses from oxidant damage are several hundred million
dollars a year.
There have been three main approaches to protecting plants from
air pollution. Several researchers are including pollutant stress in
standard breeding programs and thus are breeding for tolerance. Interim
measures involve the use of chemical sprays. Such sprays are not now
economically feasible, but they are being tested, and some are protective.
Cultural and land-use practices may also be used to control pollution
effects, especially on a short-term basis.
A few definitive experiments are needed to complete our knowledge
of acute dose-response relationships for ozone. Research is necessary in
the case of PAN and other oxidants. More important is the need for studies
of crop and native species over growing seasons with chronic oxidant
exposures. At the same time, additional work with field chambers,
filtered or nonfiltered, is needed.
There is a critical need to understand the interaction of multiple
pollutants on the plant systems. These are believed to be important, but
little is known about these interactions with respect to most plants.
Although the plant membrane is considered the primary site of
action for the oxidants, there is no definitive work on this. The mechan-
ism of response and the biochemical systems affected are not understood.
An understanding of these responses would be supportive of breeding and
spray protective programs.
Effects on ecosystems are considered in Chapter 12. The permanent
vegetation constituting natural ecosystems receives much greater chronic
1-15
-------
exposure than the short-lived vegetation that makes up the agroecosystem
subject to intermittent short-term fumigations. Each situation has
measurable economic and aesthetic effects, but on different time scales.
The single agricultural ecologic system (the agroecosystem) has little
resilience to pollutant stress; losses are sometimes immediate and
occasionally catastrophic. The natural ecosystem is initially more resistant
to pollutant stress because of species diversity, but the longer chronic
exposures disrupt the system. Simulation models of ecosystem components are
under development. The study of such models and their interaction offers the
possibility of determining the long-term effects of pollutants on natural
ecosystems and agroecosystems.
There is convincing evidence of large-scale damage to natural
ecosystems in regions downwind from Los Angeles. The injury to the mixed-
conifer stands of the San Bernadino National Forest began in the early
1940's and is well advanced. A similar problem is developing in the
forests of the southern Sierra Nevada. Both direct and indirect effects
have been observed on most components of the forest ecosystem, including
producers, consumers and decomposers. These effects are the results of
reactions in the Los Angeles urban plume, which generate oxidants as
the pollutants are transported downwind. Other cities of the western
United States—namely, Salt Lake City and Denver, where basin-mountain
terrain is contiguous—may show similar injury to forest ecosystems as
their oxidant air pollution problems grow.
The socioeconomic consequences of the continued degradation of
natural ecosystems and agroecosystems should be investigated in more
detail. The indirect effects on man's health and the direct effects
1-16
-------
on his welfare resulting from ecosystem deterioration by oxidant injury
should be taken into account in developing air pollution control strate-
gies. Land-use planning and airshed classification schemes may be useful
in averting further deterioration.
Dose-response relations in key primary-producer species in
service-important ecosystems should be examined under field conditions.
Material damage by oxidants is reviewed in Chapter 13. In test
chambers with external "ozone" generators that operate at or near atmos-
pheric pressure, ozone is the only likely oxidizing species. In ambient
air, however, the presence of other oxidants and sunlight may contribute
to material damage. Laboratory studies of the mechanisms and effects of
oxidants other than ozone—including PAN, atomic oxygen, and some free
radicals—on specific materials are needed. The concentration of these
agents in the atmosphere should be determined.
The most economically important materials with respect to ozone
damage are paint, elastomers (rubbers), and textile fiber-dye systems.
Damage to polyethylene by ozone is considered to be negligible. The ozone
damage in 1970 to materials has been estimated as follows: paint, $540 mil-
lion; elastomers, $569 million; and textile fibers and dyes, $84 million—
for a total of over $1 billion. Thus, the total combined material and
crop damage falls between $1.5 and 2 billion per year. Estimates of
damage to natural ecosystems are not available.
1-17
-------
REFERENCES
1. Haagen-Stnit, A. J. Chemistry and physiology of Los Angeles smog.
Ind. Eng. Chem. 44:1342-1346, 1952.
2. National Academy of Sciences. National Academy of Engineering. Air
Quality and Automobile Emission Control. A Report by the
Coordinating Committee on Air Quality Studies. Vol.4. The
Costs and Benefits of Automobile Emission Control. U.S. Senate
Committee Print Serial No. 93-24. Washington, D.C.: U.S.
Government Printing Office, 1974. 470 pp.
1-18
-------
CHAPTER 2
CHEMICAL ORIGIN
The photochemistry of the polluted atmosphere is exceedingly complex.
Even if one considers only a single hydrocarbon pollutant, with typical
concentrations of nitrogen oxides, carbon monoxide, water vapor, and other
trace components of air, several hundred chemical reactions are involved
in a realistic assessment of the chemical evolution of such a system.
The actual urban atmosphere contains not just one, but hundreds of diffe-
rent hydrocarbons, each with its own reactivity and oxidation products.
The atmosphere is complicated in other ways. The emission of
primary pollutants occurs throughout the day and night (varying with
time and location), adding to some of the previous day's well-aged
pollutants. As the sun rises, the light intensity increases in a non-
linear fashion. The movement of air is important — vertical mixing
and lateral transport from one community to another. It is not practical
to construct a model that includes the detailed chemistry and all these
variables with present computers.
Several approaches have been used to reduce the problem to
manageable proportions. The chemistry of photochemical oxidant formation
can best be understood by considering laboratory experiments with one
hydrocarbon (at most, two) and typical amounts of the nitrogen oxides,
carbon monoxide, and water vapor. A model is developed on the basis of
all the chemical reactions that are thought to be relevant, with their
measured or estimated rate constants. Calculations and observations are
then compared to assess the accuracy and completeness of the chemical
model. Alternatively, one can reduce the chemistry to just a few generalized
2-1
-------
or lumped reactions, and then include the temporal and spatial changes
that occur as a polluted air mass moves through an urban environment. Both
approaches provide valuable insights.
This chapter offers a brief introduction to present understanding
of the chemistry of photochemical oxidant formation (thorough reviews are
available elsewhere 1~->> a) „ some comments on the state of knowledge of
the chemistry of the polluted atmosphere, a review of some recent develop-
ments in instrumentation that promise to increase understanding of atmos-
pheric chemistry.
2-2
-------
BASIC CHEMISTRY OF OXIDANT FORMATION
Brief Review
This section covers some of the more important chemical reactions
that occur in the polluted atmosphere and attempts to show how these reac-
tions result in photochemical oxidant formation. For a more thorough
understanding of the chemistry involved, the reader should consult several
1—4 5a
of the recent reviews ' and computer modeling studies by Demerjian,
Kerr, and Calvert and by Calvert and McQuigg. Unless otherwise noted,
the mechanisms and rate constants of these modeling studies are used in
this discussion.
Three properties of photochemical smog were evident first in Los
Angeles: eye irritation; haze (aerosol) formation; and the degradation
of rubber products. All three are associated with oxidants, although
^
aerosols can also be formed other pollutants, particularly sulfur dioxide.
f.
The photochemical oxidants that are observed in the atmosphere are
ozone, 0^; nitrogen dioxide, N02? and peroxyacetylnitrate (PAN). Several
other substances, such as hydrogen peroxide, Vifi^* may ^e classified as
photochemical oxidants, but their common presence in smog is not well
established. The oxidants are secondary pollutants; i.e., they are formed
as a result of chemical reactions in the atmosphere. Primary pollutants
are those emitted directly by pollution sources.
The classes of major primary pollutants that are important in urban
areas are listed in Table 2-1. The pollutants most responsible for oxidant
formation in the air are the nitrogen oxides, hydrocarbons, aldehydes, and
carbon monoxide. The internal-combustion engine is a major source of
emission of these primary pollutants, although many stationary sources, such
as electric power generating plants, also contribute heavily to emission
of nitrogen oxides.
2-3
-------
TABLE 2-1
Some Primary Pollutants Involved in Photochemical Oxidant Formation
Hydrocarbons:
Alkanes; ii-butane, isopentane, isooctane
Cycloalkanes: cyclohexane, methylcyclopentane
Olef ins (sometimes called "alkenes"): ethylene, propylene, butene
Cycloolef ins; cyclohexene
Alkynes : acetylene
Aromatics; toluene, xylene
Chlorinated hydrocarbons
a
Aldehydes, RCHCT": Formaldehyde, acetaldehyde
Ketones. RCOR— : Acetone, m£thylethylketone
Nitric oxide.
Carbon monoxide, CO
Sulfur dioxide, S02
— R = a hydrocarbon group, such as methyl, CH^, or benzyl, CgHc.
— "NOX" is often used to indicate "oxides of nitrogen." In practice,
this usually means the sum, NO + N02, although it should include
such other forms as N0.(, ^Ocj, and HNO^. Nitrous oxide, N20, is
relatively inert in the lower atmosphere and is not included in
the term "NO ."
x
2-4
-------
The amount of oxidant formed In the atmosphere has a complex
dependence on time of day, meteorologic conditions, and amounts of the
various primary pollutants. A typical time dependence for a smoggy day
in Los Angeles is shown in Figure 2-1. Early in the morning the concentra-
tion of ozone is very low. As the day progresses, the ozone increases.
The complementary behavior of nitric oxide and nitrogen dioxide is of
major importance here. The rapid increase in nitric oxide is the result
of the morning rush-hour traffic. The nitric oxide concentration then
falls, even though automobile emission is still strong, and nitrogen
dioxide begins to increase. As the nitrogen dioxide concentration
increases, so does that of ozone.
It should be noted that the concentrations shown in Figure 2-1
represent averaged hourly values. Recent continuous monitoring of nitrogen
dioxide at one site shows a more complex time dependence, seen in Figure 2-2.
These rapid fluctuations in concentration within a few minutes are probably
the result of the movement of air masses and varying emission from local
Q
sources. Figure 2-2 underlines the importance of air transport and variable
emission, as well as of chemical changes, in the modeling of concentration
at a single monitoring station.
The time dependence of the oxidant concentration shown in Figure 2-1
can be mimicked in laboratory studies. The results of a typical smog-
chamber experiment are shown in Figure 2-3. A sample of air initially
containing propylene at 2.2 ppm and nitric oxide at 1.0 ppm is irradiated
starting at time zero. Although the hydrocarbon begins to disappear almost
immediately, ozone does not develop until almost all the nitric oxide has
been converted to nitrogen dioxide. This relationship between nitric oxide,
nitrogen dioxide, and ozone is of central importance in attempts to under-
2-5
-------
HOUR OF DAY
FIGURE 2-1. Diurnal variation of nitric oxide, nitrogen dioxide, and ozone
concentrations in Los Angeles. July 19, 1965.
(Reprinted from Air Quality Criteria for Photochemical Oxidants.19
2-6
-------
"
t "
r
i::
!••
j-»>
»AOnC MTllOff lAMm TM« lw|
I I I I I I I I I
I I I I I I
FIGURE 2-2.
Atmospheric nitrogen dioxide concentrations, El Segundo, California,
April 4-5, 1974. (Reprinted with permission from Tucker et al.? )
2-7
-------
CHjCHD
SO
ISO
to
ttminl—
FIGURE 2-3. Typical concentration-time profiles for irradiation of a propy-
lene-NO,. mixture-io.a smog chamber. (Reprinted with permission
from NiKi et al.20 )
2-8
-------
stand the chemistry of these systems. Note also in Figure 2-3 the develop-
ment of PAN, CH3C03N02.
Laboratory experiments of this type have the great advantage that the
initial conditions can be well defined (although often they are not-*) , in
contrast with the average sample of urban air, which is a mixture of new and
old pollutants. Also, in laboratory experiments, the same sample of air is
observed over a long period, which is not possible with most air pollution
monitoring networks. For these reasons, most attempts to understand the
chemistry of oxidant formation have concentrated on smog-chamber experi-
ments, rather than the real atmosphere.
The major oxidant in smog is ozone, and early research efforts
concentrated on the mechanism of its formation. Attention was focused on
the nitrogen oxides, specifically nitrogen dioxide, because it was known
that the nitrogen dioxide molecule could absorb blue and near-ultraviolet
sunlight and break apart (undergo photolysis):
N02 + light(A<430 nm) -*- 0 + NO. (1)
In the lower atmosphere, the oxygen atoms react quickly with molecular
oxygen to form ozone:
0 + 02 + M -> 03 + M. (2)
Equation 2 is actually a three-body process, in that another molecule (M),
usually nitrogen or oxygen, is necessary to carry off the energy released
in the newly formed bond. These two reactions then form a mechanism for
ozone formation in the atmosphere. They would not be complete without
the additional reaction,
NO + 03 •+ N02 + 02 (3)
which is known to be fast. Because Eq. 3 consumes the molecule of
ozone that was formed in Eq. 2 and regenerates the molecule of nitrogen
2-9
-------
dioxide that was photolyzed in Eq. 1, the three reactions form a cycle,
which is shown schematically in Figure 2-4. The net result of this cycle
is that the absorbed sunlight is degraded into thermal energy. Several
other minor reactions occur in this system, but they will not be discussed
here (see page 17 of reference 5).
Equations 1-3 show the most common chemical reactions that occur
in the polluted atmosphere. That is because nitrogen dioxide is the
strongest absorber of sunlight. At a latitude of 40°, the typical turnover
lifetime for nitrogen dioxide is about 1.4 min. This means that, on the
average, half the nitrogen dioxide molecules are photodisssociated in Eq. 1
and reformed in Eq. 3 every 1.4 min. No other molecule in smog is so
active.
An important consequence of this rapid turnover is the establish-
ment of a steady-state concentration of ozone. One can express this
dynamic equilibrium as follows:
[03] "
k3 [NO]
(4)
where k-^ is the rate at which sunlight dissociates nitrogen dioxide, k3 is
the rate constant for Eq. 3, and the brackets indicate concentrations. With
a typical value for k^ (8 x 10~^/s) and the known rate constant for Eq. 3,
Eq. 4 becomes:
[NOo]
[03] = (0.021 ppm) Z . (5)
[NO]
Modeling studies show that Eq. 4 should be obeyed quite closely.
Tests of this equation on atmospheric data show good agreement, at least
for ozone concentrations of 0.1 ppm or less.°^>"-' At higher ozone concen-
trations, deviations have been observed, although it was suggested that
2-10
-------
Ozone
0,
<
f j
^02
^ s
Oxygen
Atom
0
^j
o
>
Nitrogen
Dioxide
+ light
J<
Oxide
NO
<
FIGURE 2-4. The NO-NO^-C^ cycle in air contaminated with N0x only (above)
and with NOX and hydrocarbons (below). Above, the dissocia-
tion of nitrogen dioxide by sunlight forms equal numbers of
nitric oxide molecules and oxygen atoms. The latter are
rapidly converted to ozone molecules. The ozone then reacts
with the nitric oxide, again on a 1:1 basis, to reform nitrogen
dioxide. Only a small steady-state concentration of ozone results
from this cycle. Below, when hydrocarbons, aldehydes, or other
reactive contaminants are present, they can form peroxy radicals
that oxidize the nitric oxide, pumping it directly to nitrogen
dioxide. This leaves very little of the nitric oxide to react
with the ozone, so the ozone builds up to large concentrations.
2-11
-------
the method of averaging was responsible, rather than a real failure of
Eq. 4."
Equation 4 does appear to explain, at least qualitatively, the
time dependence of the ozone concentration. For example, as long as the
ratio [NC>2] : [NO] is less than 1:1 in Figure 2-3, the ozone concentration
is very low. However, when most of the nitric oxide has been converted
to nitrogen dioxide, the ozone concentration increases rapidly. Similar
behavior is observed in the Los Angeles atmosphere (Figure 2-1).
Although of central importance in smog chemistry, Eqs. 1-3 cannot
by themselves explain the buildup of ozone. If only these three reactions
were important, the photodissociation of nitrogen dioxide would rapidly
establish a small ozone concentration within a few minutes, after which
no further changes would occur. During this ozone buildup, the nitrogen
dioxide concentration would only decrease. Contrast this behavior with
the observations in Figures 2-1 and 2-3. The actual ozone buildup occurs
over a period of hours and is accompanied by an increase in the nitrogen
dioxide concentration. Thus, the photolysis of nitrogen dioxide alone cannot
explain the ozone buildup, even though it is the mechanism of ozone
formation. The dominant factor in these systems is the ratio of nitrogen
dioxide to nitric oxide. The challenge, then, is to explain the conversion
of nitric oxide to nitrogen dioxide. Once that is done, the ozone concen-
tration will follow the [NQ2] : [NO] ratio.
Another difficulty was apparent in the early chemical studies on
Q /
polluted air.— It was known from laboratory studies that both ozone and
the ground-state oxygen atoms that are formed in Eq. 1 would attack reactive
hydrocarbons. However, the experimentally observed rate of loss of the
hydrocarbons was often greater than could be explained by the attack of ozone
2-12
-------
and oxygen atoms. Figure 2-5 shows this effect for the case of propylene.
Note that the discrepancy is especially large in the earlier parts of the
reaction.
Some new mechanism was required to explain the rapid oxidation of
nitric oxide to nitrogen dioxide; and one or more reactive intermediates,
in addition to ozone and oxygen atoms, were needed to explain the observed
hydrocarbon loss rates. Several people suspected that these two problems
were connected to some free-radical chain mechanism. Many reactive inter-
mediates were suggested, including the hydroxyl radical, OH; the hydroperoxy
radical, H02; the methoxy radical, CHUO; nitrogen trioxide, N03; and "singlet
oxygen" (raeaning 02 in one of its low-lying metastable states, a A or b I.) .
About 1970, two research groups suggested that hydroxyl radicals
9 10
(OH) were the solution to the above problems. ' They suggested the following
reaction cycle:
OH + CO + H + C02, (6)
H + 02 + M -> H02 + M; (7)
H02 + NO ->• OH + N02. (8)
Equation 6 is sufficiently fast to be important in the atmosphere. For a
carbon monoxide concentration of 5 ppm, the average lifetime of a hydroxyl
radical is about 0.01 s, owing to Eq. 6 (other reactions may decrease the
lifetime even further). Equation 7 is a three-body recombination and is known
to be fast at atmospheric pressures. The rate constant for Eq. 8 is not well
11—13 13b
established, although several experimental studies support its occurence. '
On the basis of the most recently reported value for the rate constant of Eq. 8,
13a
which is an indirect determination, the average lifetime of a hydroperoxy
radical is about 2 s for a nitric oxide concentration of 0.05 ppm. Equation 8
is the pivotal reaction for this cycle, and it deserves more direct experimen-
tal study.
2-13
-------
FIGURE 2-5. Comparison of the experimentally observed
rate of propylene loss with that calculated
for its reactions with ozone and oxygen atoms.
(Reprinted with permission from Niki et al. )
2-14
-------
Equations 6-8 form a catalytic cycle, in that the hydroxyl radical that
is used in Eq. 6 is regenerated in Eq. 8. The net results of this cycle are
the oxidations of nitric oxide to nitrogen dioxide and carbon monoxide to
carbon dioxide by the oxygen present in the air.
Other oxidation chains can be constructed. For example, when methyl
radicals are generated by other reactions, as in the ozonolysis of olefins,
then the following reactions can occur:
CH3 + 02 + M •> CH302 + M, (9)
CH302 + NO -> CH30 + N02, (10)
CH30 + 02 -»• CH20 + H02. (11)
The methyl radical rapidly (in 10 s) combines with oxygen to form the
methylperoxy radical, CH302. A recent study has confirmed that nitric oxide is
14
oxidized by methylperoxy, although the rate constant is still unknown.
The methoxy radical, CH.^ should then react predominantly with oxygen to
form formaldehyde, CH20, and hydroperoxy radical. The net result of this
sequence is the oxidation of one molecule of nitric oxide to nitrogen
dioxide and the conversion of an alkyl radical into a hydroperoxy radical,
which can then react as in Eq. 8. Similar sequences can be written for
larger alkyl radicals.
There were two important innovations in the development of these
oxidative cycles: the use of carbon monoxide, which had previously been
considered a relatively inert molecule in the atmosphere, to regenerate the
hydroperoxy radical via Eqs. 6 and 7; and the use of peroxy radicals, HO-
and R02, to oxidize nitric oxide to nitrogen dioxide.
These oxidative cycles have a drastic effect on the concentration
of ozone, as is summarized in Figure 2-4. The peroxy radicals oxidize the
nitric oxide back to nitrogen dioxide, increasing the ratio [N02J:[NO] and, as
2-15
-------
a result of Eq. 5, the concentration of ozone. Expressed in words, the
photolysis of nitrogen dioxide continues to generate ozone, but the balancing
reaction of ozone with nitric oxide becomes less probable, because most of
the nitric oxide is reacting with peroxy radicals. As a result, more ozone
is being formed than is being destroyed, and so its concentration increases.
It is this interaction of the NO-NC^-C^ cycle with the free radicals generated
from hydrocarbons and other reactive pollutants that is the basis of photo-
chemical oxidant formation.
The participation of hydroxyl and hydroperoxy radicals in the oxida-
tion of nitric oxide raises the possibility that these radicals might also
attack hydrocarbons. In the case of hydroxyl, these reactions are known
to be fairly rapid. On the basis of the rate constants that have been
measured and estimates of those which have not, the rates of attack of
hydroxyl and hydroperoxy radicals appear to be large enough to explain the
excess consumption of propylene shown in Figure 2-5.
A detailed chemical model has been constructed with these free-
radical chain reactions and all the other reactions that are thought to be
important in the atmosphere. It is possible to evaluate quantitatively the
various reactions that destroy the olefin and to determine which intermediates
are most important. Figure 2-6 shows the results of such a summation for a
mixture of trans-2-butene, NO , and aldehydes.6 The vertical distances
" X
between lines on Figure 2-6 are proportional to the various rates of attack.
The graph shows that destruction of olefin by hydroxyl radical is a major
process, although attack by ozona, hydroperoxy radical and oxygen atoms is
also significant. Olefin attack by other intermediates, such as nitrogen
trioxide and methoxy radical, are less significant, but not insignificant.
Taking propylene as a typical example, the reactions with hydroxyl
radical would be:
2-16
-------
0(3P)
30 60
IRRADIATION TEffi, minutes
90
120
FIGURE 2-6. Calculated rates of reaction of various species with
trans-2-butene as a function of irradiation time. The
initial conditions are as specified in Table 2-2, except
that the aldehydes are not initially present. (Reprinted
with permission from Calvert and McQuigg." ).
2-17
-------
•OH + C3H6 ->• .CH2CHCH2 + H20, (12)
•OH + C3H6 •> CH3CHCH2OH, (13)
where the dots on the free radicals indicate the dominant positions of the
unpaired electrons. Evidence of both abstraction reactions, such as
Eq. 12, and addition reactions, such as Eq. 13, has been obtained recently.
The relative importance of addition and abstraction will depend on the
structure of the olefin; these numbers are not well established.
Each of the radicals formed in Eqs. 12 and 13 will react with
molecular oxygen to form a peroxy radical,
R- + 02 -»• R0|, (14)
which can then oxidize nitric oxide:
RO + NO -> RO' + NO . (15)
The RO' formed in Eq. 15 is still a free radical, so it will react further.
Most free radicals contain odd numbers of electrons, and most stable
molecules contain even numbers of electrons (nitric oxide and nitrogen dio-
xide are two important exceptions, being stable molecules with odd numbers
of electrons). Therefore, in the reaction, free radical + stable molecule
another free radical is usually generated. This free-radical chain process
is stopped only when one of the following types of processes occurs.
Radical-radical reactions, e.g.,
H&2 + H62 + H202 + 02, (16)
CH362 + H02 + CH302H + 02. (17)
Radical-N0v reactions, e.g.,
2t
H02 + N02 •> HONO + 02 (or HOON02 ?) , (18)
CH3C002 + N02 •+ CH3C002N02 (PAN). (19)
Radical-surface reaction, e.g.,
H02 + surface ->• absorbed radical. (20)
2-18
-------
The rate constants for these chain-terminating steps are not well established.
However, present estimates are probably not greatly in error because radical-
radical reactions tend to be fast.
What is the initial source of the free radicals that are so important
for oxidant development? Calvert and McQuigg attempted to answer this question
by evaluating the many proposed reactions with their detailed chemical model.
Although the actual importance of any particular source will depend on the
concentration of pollutants assumed and the time of irradiation, they found
for a typical mixture (nitric oxide, nitrogen dioxide, trans-2-butene,
formaldehyde, acetaldehyde, carbon monoxide, water, and methane) that the
following reactions were the most important radical sources:
HONO + sunlight -»- OH + NO, (21)
CH 0 4- sunlight -v CHO + H, (22)
CH CHO + sunlight -+ CH + CHO, (23)
°3 + C4il8 "* various radicals. (24)
Of special interest here is the radical generation during the early part of
the irradiation, before the oxidant concentration has developed; photodissocia-
tion of nitrous acid, HONO, and aldehydes is very important. The concentration
of nitrous acid in the atmosphere due to the reaction,
NO + N02 + H20 -*• 2HONO, (25)
is controversial, but the presence of aldehydes in the urban atmosphere
is not. It is established that aldehydes will absorb sunlight and dissociate
into free radicals. The mechanism of the reaction of ozone with olefins in the
gas phase appears to be very complicated, but recent experimental evidence
shows that some free radicals are formed.
Although the above reactions generate a few free radicals, most of
the oxidation of nitric oxide to nitrogen dioxide is carried out by the alkyl-
peroxy, R02, and hydroperoxy radicals that are formed in later reactions
2-19
-------
involving reactive hydrocarbons, aldehydes, or even carbon monoxide. One such
example is shown in Figure 2-7. There is still considerable uncertainty as
to the mechanism of these secondary reactions. The modeling studies should
5,6
be consulted for details.
The importance of the secondary reactions can be expressed as a radical
chain length, which is the total rate of all reactions involving a particular
radical divided by the primary rate of formation of that radical. For a
particular set of conditions chosen in the modeling studies, chain lengths
of about 4 for the hydroperoxy radical and about 8 for the alkylperoxy and
hydroxyl radicals were calculated early in the irradiation period. After
further irradiation, the chains became shorter.
The oxidizable pollutants — such as hydrocarbons, aldehydes, and
carbon monoxide — serve the function of regenerating free radicals that will
react with the oxygen in the air to form alkylperoxy and hydroperoxy. Thus,
these oxidizable pollutants can be thought of as pumping the nitric oxide
to nitrogen dioxide. In the process, they become degraded to other compounds,
some of which are still reactive (e.g., formaldehyde, CH^O, in Figure 2-7).
The amount of pumping that can be done, and thus the amount of photochemical
oxidant formed, depends on both the reactivity of the oxidizable pollutant
and its concentration in a nonlinear way. As the oxidant concentration builds
up, the probability of ozone's reacting with hydrocarbons and various free
radicals increases, and the rate of ozone accumulation decreases.
Early attention focused on the most reactive of the hydrocarbons,
the olefins, because it was expected and was observed by atmospheric sampling
18
that they were preferentially consumed during smog formation. Laboratory
studies confirm that olefin-NO mixtures are very prolific sources of ozone.
X
However, these olefins are not essential to oxidant formation. Both the
2-20
-------
CHCH-CHCHj
(a)
• ThrW proaucU unotrgo ngnrficam protcxMCOnxxXUat •
(b)
as Before
CHJOCH(CHJ)CHCH3
CHJOCHC
IP,
CHjOCHf^CK,
HjCH-CHCHj. CH3OH
|«bc.<«
-------
modeling studies and smog-chamber simulations show significant oxidant formation
with NO + aldehydes, NO + alkanes (except methane), or even NO + carbon
XX X
monoxide in moist air. The development of significant oxidant from NO +
A.
aldehydes is particularly ominous, because aldehyde emission is not now
controlled. As the modelers state:
"It appears from these data that the Oo standard (i.e. 0.08 ppm for
one hour) could not be met if the aldehydes remained high, [CH^O] =
0.10, [CH^CHO] = 0.06 ppm, even if nearly all of the olefinic
hydrocarbon were removed .... We should learn from these data that the
true relationship between non-methane hydrocarbons and maximum 1-hour
oxidant at low hydrocarbon levels could be a critical function of a
variable which is not routinely measured now, namely the concentration
of the impurity aldehyde."
A similar statement could probably be made concerning ketones. These compounds
are commonly used as solvents, and they are known to form free radicals when
photolyzed. Many chlorinated hydrocarbons, which are also widely used as
solvents, can be attacked by hydroxyl radicals and thus contribute to peroxy
radical formation.
No one pollutant can be blamed as the major cause of ozone formation.
Replacing the more reactive hydrocarbons with less reactive ones would delay
the formation of ozone, but would not prevent it. Reducing the NOX concentra-
tion seems to reduce the maximal oxidant concentrations observed, but the effect
is nonlinear. Heavy injections of nitric oxide into the air can temporarily
reduce the local ozone concentration, as often happens in urban centers, but
additional oxidant formation can be expected later downwind. Although these
effects can be understood qualitatively, it is not yet possible to make accurate
predictions of oxidant formation, even in laboratory experiments.
2-22
-------
In summary, the concentration of ozone in the polluted atmosphere is
controlled by the intensity of sunlight and the ratio of nitrogen dioxide to
nitric oxide. Hydrocarbons and other pollutants — such as aldehydes, ketones,
chlorinated hydrocarbons, and carbon monoxide — react to form peroxy radicals.
These, in turn, react with nitric oxide, causing the ratio [NC^J'tNO] to
increase. As a consequence of Eq. 5, the ozone concentration also increases.
This brief description of oxidant formation in polluted air is based
on our current understanding of the chemistry involved. It is evident from
an examination of the detailed mechanism that many of the important reactions
have not been well studied. For example, the sequences of degradation
reactions for the hydrocarbons are only poorly understood. As a result
of these uncertainties, it is not possible to make accurate predictions of
photochemical oxidant concentrations. However, with another 5 years of
progress similar to the last 5, it should be possible to construct chemical
models that will permit ozone predictions accurate to within 30% for labora-
tory studies. Although many of the chemical details in the current models
are certain to be altered as more experimental data become available, it
seems probable that at least the backbone of our present understanding is
correct.
Recent Chemical Modeling Studies — Results and Uncertainties
The state of our current understanding of the chemistry of oxidant
formation can be judged by examining recent modeling studies, in which all
the reactions that are considered important are combined and the resulting
differential equations are integrated numerically. Experimental rate
constants from the literature are accepted (unless they are judged unreason-
able), and others are estimated. In the system NO -propylene-air, 242
2£
different reactions are included; another 100 reactions are neglected as
2-23
-------
probably unimportant. Roughly half the rate constants used in the model
are not firm experimental values.
The success of these computer simulations must be rated as quite
good. Figure 2-8 compares concentration-time measurements from a smog-
chamber study of N0x-propylene-air with computer-calculated results based
on the same initial conditions. The time dependences and absolute concentra-
tions agree fairly well, but not perfectly. Note that the calculated maximal
concentrations of ozone and PAN are less than the measured values by about
a factor of 2. The rapid conversion of nitric oxide to nitrogen dioxide
and the time dependence of ozone and PAN are reproduced quite well. These
results are impressive, considering that the mechanism and rate constants
in the model were not considered variables to be adjusted until a best fit
is obtained. Once the reactions and rate constants were selected, they were
used for a variety of simulations,. with fair to excellent agreement.
However, the number of concentrations being fitted is fairly small,
compared with the number of parameters going into the model. The computer
calculations can be used to predict the concentrations of other trace
compounds that should be present in smog chambers. If these compounds are
later found at approximately the predicted concentrations, the model will be
strengthened. If not, changes will have to be made in the model. A more
stringent test of the model will occur when it becomes possible to measure
the actual concentrations of the free radicals in the atmosphere.
Many of the values of the rate constants that go into the model
are not critical, but some are. As the modelers state:
"Even these systems are very complex and difficult to treat quanti-
tatively since many reactions which appear to be important in theory
have not been studied in detail, and theoretical estimates of rate
constants must be made in desperation...."
2-24
-------
£0
LO
°C
-8-
\
\
1
SP
\
(^
y
\ (
y
\
*or
\
i\.e
•^-,
HE
,C
^==
ii _
=».
^BC:
:
100 200 300 40
RRADtATOW TIME. MIN
100 200 300 400
IRRADIATION TIME. MIN
FIGURE 2-8. Photooxidation of propylene in irradiated C3H,-NO-NC>2
mixtures in moist air. A, experimental rate data from
smog-chamber experiment of A. P. Altshuller et al.
(Environ. Sci. Technol. 1, 899 (1967)). Initial concen-
trations: C3H6, 2.09 ppm; NO, 0.90 ppm; N02, 0.09 ppm.
Relative humidity at 31.5 C, 50%. B, computer simula-
tion of product concentration-time curves for same
initial conditions. (Reprinted with permission from
Demerjian et al. ).
2-25
-------
Some of the most serious current uncertainties are as follows:
• Direct determinations of rate constants are needed for almost
all the reactions of hydroperoxy radical and RC^.
• "The rate constants for both homogeneous and heterogeneous
reactions related to the oxides of nitrogen and water vapor
should be characterized carefully.... Furthermore, the actual
levels of nitrous acid, nitric acid, l^O, NO, and N0£ should
be determined simultaneously in real auto-exhaust-polluted
atmospheres...."
• "The primary quantum yield of radical formation in nitrous
acid, CHJD, and CH-jCHO photolyses should be better established
as a function of the appropriate wavelength range of sunlight."
• "The chemical details of the reaction sequence following OH
radical addition and abstraction from olefins should be explored
fully...."5
• The chemical details of the reactions of representative alkyl
radicals, alkoxy radicals, and biradicals with oxygen should be
established. Both the rate constants and the immediate products
are needed to construct realistic mechanisms for the model.
• The gas-phase stability of a variety of possible products —
such as the olefin ozonide, peroxyformyl nitrate, and peroxy-
nitric acid — are not known.
These assignments represent formidable tasks for the experimentalists,
but recent developments in instrumentation and techniques suggest that
substantial progress should be made soon.
The present smog models are probably most vulnerable to the
possibility that some important mechanism or intermediate has been omitted
altogether. The modelers themselves recognize the uncertainties involved:
2-26
-------
"One must not take too seriously the results from complex simulations
at this stage of our knowledge. In particular the development of sound
reaction schemes and realistic smog models require much more quantitative
kinetic information as to the detailed reaction paths which appear to
be important in photochemical smog. ...There is no question that as
such information becomes available, present models will require sub-
stantial changes."
A realistic and detailed chemical model has great value. The stepwise
addition of various primary pollutants can be made to evaluate the importance
of each. The effects of various emission control strategies on the chemistry
of oxidant formation can be studied easily and quickly. It is possible to
calculate the importance and concentration of various reactive intermediates.
One can estimate the concentrations of various compounds that have not yet
been observed in smog. And it is possible to pinpoint some of the important
gaps, in order to stimulate future experimental studies.
The detailed chemical model has been used to make the following
observations and predictions; some of these statements are common to other
models and are in accord with experimental observations:
• Complete elimination of olefinic hydrocarbons without controlling
aldehyde emission will not ensure low oxidant readings in the
21
atmosphere.
• If no hydrocarbons or aldehydes were present in the atmosphere,
but carbon monoxide and NO were present, significant ozone
X
concentrations would develop. With reactive hydrocarbons present,
the addition of carbon monoxide does not have a strong effect on
oxidant concentration, unless it is added in very large amounts
22
(2,000 ppm).
2-27
-------
• Reactions of the hydroxyl radical dominate the removal of hydro-
carbons. However, several other reactants make significant
contributions, including hydroperoxy radical, ozone, and oxygen
atoms. (This conclusion depends on the hydrocarbon being considered:
OQ
it is claimed that some terpenes in air are attacked mainly by ozone. J)
• The hydroperoxy radical has the highest concentration of all the
free radicals in smog. The concentrations of both hydroperoxy and
hydroxyl radicals are rather insensitive to primary pollutant
concentration.
• The concentration of ozone generated photochemically goes through
24
a maximum as the NO concentration is increased.
x
• The total oxidant dosage varies in a nonlinear manner with dilution
J C *) f.
of the primary pollutant. '
• The participation of singlet oxygen in the development of photo-
chemical oxidant is minor, but its human health consequences should
27
be considered.
A good understanding of the detailed chemistry of oxidant formation
makes it possible to construct more compact chemical models. These generalized
or lumped mechanism models reduce the number of individual chemical reactions
28-30
by combining similar or sequential reactions and ignoring minor ones.
Averaged reactivities and averaged product yields may be used. The simplified
mechanisms can also reproduce smog-chamber data very well, although the construct-
ion of the simplified model is semiempirical, so good agreement is not un-
expected. These generalized mechanisms are very valuable for modeling
the real atmosphere, because the reduction in chemical complexity allows
the inclusion of other important variables, such as transport and pollutant
inputs. (The use of these models is covered in Chapter 5.)
2-28
-------
Which Reactive Intermediates Are Important?
The concentrations of the various reactive intermediates can be
calculated with a detailed chemical model. This Information is not
available elsewhere, because it is not yet possible to measure these reactive
intermediates in the atmosphere. Table 2-2 shows the calculated concentra-
tions of several intermediates and ozone for a simulated polluted atmosphere.
Table 2-3 gives the calculated rates of attack of the same intermediates on
the olefin trans-2-butene for the same irradiation times. By comparing
these two tables, one can judge the importance of various intermediates.
For example, the concentration of singlet oxygen is high, but its rate of
attack on the olefin is insignificant, compared with that of hydroxyl radical.
The steady-state concentration of hydroxyl radical is very low, owing mainly
to its high reactivity. Early in the irradiation, hydroxyl radicals
account for 90% of the trans-2-butene removal and even after 60 min, when the
oxone concentration has developed, they still account for 50% of the removal.
The hydroperoxy radicals show both a relatively high concentration — by far
the highest of the free radicals — and a significant rate of attack on the
olefin.
Average Lifetimes of Reactive Intermediates
It is well established that both ozone and PAN can cause damage to
biologic systems (see Chapters 8 and 11). The possibility that the reactive
intermediates in smog could directly cause biologic damage has been suggested, 1
but experiments are seldom designed to test this possibility.
The lifetimes of the various reactive intermediates shown in Table 2-4
were calculated with the assumption of a sudden termination of sunlight, as would
occur when air is inhaled. The long lifetime of the hydroperoxy radical means
that almost all of it will survive long enough to be transported into the lungs.
2-29
-------
TABLE 2-2
Calculated Concentrations of Reactive Species in a Simulated
o
Smog Mixture at Several Irradiation Times—
Concentration, ppb
Species
Ozone, 0
Hydroperoxy radical, HO,.,
Singlet oxygen, CL (a A)
Nitrogen trioxide, NO.,
Hydroxyl radical, OH
Oxygen atoms, 0
2 min
8.5
0.21
3.9 x 10~3
0.03 x 10~3
1.7 x 10~4
3.8 x 10~6
30 min
84
0.32
5.1 x 10~3
2.2 x 10~3
0.88 x 10~4
8.9 x 10~6
60 min
139
0.37
5.2 x 10~3
7.4 x 10~3
0.72 x 10~4
. 9.5 x 10~6
a Data from Demerj ian et a^. Initial conditions: [NO], 0.075 ppm; [N02],
0.025 ppm; [trans-2-butene], 0.10 ppm; [CO], 10 ppm; [CH20], 0.10 ppm;
[CH3CHO], 0.06 ppm; [CH4], 1.5 ppm; relative humidity, 50%.
2-30
-------
TABLE 2-3
Calculated Rates of Attack on trans-2-Butene by Various Reactive Species
a
in Simulated Smog Mixture at Several Irradiation Times."
Attack Rate, ppb/min
Species
Ozone
Hydroperoxy radical
Singlet oxygen
Nitrogen trioxide
Hydroxyl radical 1.72 0.55 0.27
Oxygen atoms 0.013 0.018 0.011
2
0.
0.
2.
0.
min.
026
16
9 x 10~6
05 x 10~4
30 min.
0.16
0.15
2.1 x 10"6
2.2 x 10
60 min.
0.16
0.09
1.3 x 10~6
2.6 x 10~4
a J
Data from Demerjian et^ ai. Conditions same as in Table 2-2.
2-31
-------
TABLE 2-4
Calculated Average Lifetimes of Several Reactive Intermediates
a
in a Simulated Smog Mixture.
Species Lifetime,s
Ozone 250
Hydroperoxy radical 7
Singlet oxygen 0.05
Nitrogen trioxide 0.1
Hydroxyl radical 0.0036
Oxygen atoms 10
Data from Pitts and Finlayson (p.8).4 Conditions same as
in Table 2-2. Irradiation time, 60 min.
2-32
-------
In contrast, the hydroxyl radicals will decay within a few milliseconds. As
a computer simulation has shown, even though the half-life of singlet oxygen
appears short, compared with the breathing cycle of several seconds, a
31
significant fraction of it should survive into the lungs.
Whether such species as hydroperoxy radicals and singlet oxygen will
survive collisions with the surfaces of the upper breathing tract and reach
the lower lungs, as ozone does, must await further experiments. Clearly, if
all the ozone molecules and all the hydroperoxy radicals are eventually absorbed
in the lungs, the ozone damage will dominate, unless hydroperoxy radicals
are several hundred times more damaging, which seems unlikely.
Free radicals could have a significant effect on biologic surfaces
that are in direct contact with irradiated smog, such as leaves and human
skin. The effective dost, delivered to a surface depends on the bulk concen-
tration of the species, the sticking coefficient (the probability that a molecule
that hits a surface will be absorbed), and the damage once it is absorbed.
Very little is known about sticking coefficients, especially of free radicals.
It is known that singlet oxygen will survive approximately 10,000 collisions
32
with aqueous phosphoric acid and that ozone is not immediately absorbed in the
upper bronchial tubes. There is an indication that hydroperoxy radicals might
be absorbed readily on surfaces. If the sticking coefficient of hydroperoxy
radicals were unity for biologic surfaces (which might be reasonable, inasmuch
as it has the ability to make hydrogen bonds) and the sticking coefficient
for ozone were 10 , the actual influxes of hydroperoxy radicals and ozone
to a surface might be comparable, even though the ozone concentration is
approximately 300 times greater. This calculation is pure speculation, but
it does show that the influx of free radicals to a sunlit surface should not
be neglected.
2-33
-------
It is known that free radicals are formed when ozone reacts with carbon-
carbon double bonds. Recently, it has been suggested that PAN probably forms
free radicals when it reacts with aldehydes.3^ Because hydroperoxy radicals are
free radicals, they may have biologic effects similar to those of ozone and
PAN. Certainly, for experiments in which the observed biologic damage cannot
be attributed to the measured concentrations of ozone and PAN, free radicals
or unstable compounds should be considered.
Summary
The major primary pollutants of importance to oxidant formation are
nitric oxide, hydrocarbons, aldehydes, and carbon monoxide. A few free
radicals are formed by photolysis of aldehydes and nitrous acid by sunlight
or by the reaction of traces of ozone with reactive hydrocarbons. These
free radicals initiate chain reactions involving hydroperoxy and alkylperoxy
radicals. During these chain reactions, the nitric oxide is converted to
nitrogen dioxide, and the hydrocarbons and aldehydes are degrated. The
photolysis of nitrogen dioxide by sunlight forms a free oxygen atom, which
combines with an oxygen molecule to form ozone. Because of the NO-NC>2-03
cycle (Eqs. 1-3), the ozone concentration is determined primarily by the
ratio [N02J : [NO] and so does not become large until most of the nitric
oxide has been converted to nitrogen dioxide. The total amount of oxidant
formed depends, in a nonlinear fashion, on the amount of hydrocarbons
available to continue pumping the nitric oxide to nitrogen dioxide. Al-
dehydes and even carbon monoxide can also serve this pumping function.
When some of the peroxy radicals recombine or react with the nitrogen oxides,
many secondary products, such as hydrogen peroxide and PAN, are formed.
Recent chemical modeling studies have been reasonably successful
in reproducing the concentration — time histories of smog-chamber experi-
2-34
-------
merits. An examination of these models shows a need for much more detailed
chemical knowledge. Modeling studies also point out the necessity of
carefully defining the initial conditions of smog-chamber experiments. Some
observations that have been made with these models are:
• Even if hydrocarbons are completely removed from the air,
aldehydes and NO can generate high concentrations of photo-
X
chemical oxidants.
• If both hydrocarbons and aldehydes are eliminated, carbon
monoxide and NOX alone can generate significant concentrations
of ozone.
• The concentration of ozone generated photochemically goes through
a maximum as the NOX concentration is increased.
• The steady-state concentration of free radicals in smog is
approximately 0.3 ppb and is rather insensitive to primary
pollutant concentration.
The concentrations, average lifetimes, and rates of attack of the reactive
intermediates can be calculated with chemical models.
The effects of free radicals on biologic surfaces cannot be
ignored.
NEW EXPERIMENTAL METHODS OF STUDYING AIR CHEMISTRY
Lasers
The development of reliable lasers with a variety of wavelengths,
both fixed and variable, has generated many ideas of applications to air
pollution monitoring and to chemical kinetics. Most of these techniques
are still in the developmental stage. Whether they will have the reliability
and low cost needed for widespread use is not known. Proposed methods
include the following:
2-35
-------
• Long-path infrared absorption, using a tunable diode laser,
which is claimed to have a sensitivity of 5 ppb for carbon
35
monoxide over a 610-m pathlength.
• Differential laser absorption, with measurement of two or more
wavelengths simultaneously and claimed sensitivity in the parts-
O f
per-billion range.
• Laser backscattering, either Raman or fluorescence, which does not
require a remote detector, can thus be used for detecting atmos-
pheric pollutants at a distance, and has sensitivity less than
that with direct absorption techniques.
• Laser-induced electronic fluorescence; two devices reported
recently look very promising for continuous atmospheric monitoring;
sensitivities of 0.6 ppb for nitrogen dioxide and 50 ppb for
8 37
formaldehyde are claimed; ' careful attention to possible inter-
ference from other species is necessary; detection of the hydroxyl
Q O
radical in air (^10 molecules/cm ) has been claimed for this
38
technique, but it has been pointed out that this concentration
seems much too high, especially because the air had been removed
from the sunlight 6 s before analysis; spurious effects, such as
photolysis of the ozone in the air by the laser beam and two-photon
absorption by water vapor, might have been responsible for the
hydroxyl radical that was observed.
• Photoacoustic or optoacoustic spectroscopy, which detects the
absorption of a pulsed laser in a cell by the pressure pulses
39
generated when the. light energy is degraded to heat, which is
claimed to have sensitivities of 0.4 ppb for nitric oxide and
5 ppb for ethylene, and which can measure the absorption spectra
of solids and dusts.
2-36
-------
• Laser magnetic resonance, which has already been used to detect the
free hydroxyl, methynyl (CH), hydroperoxy, formyl (HCO), and
41
amino radicals in low-pressure gases and could be used to
determine rate constants for the reactions of the smaller free
radicals.
Several recent reviews on lasers and laser spectroscopy should be consulted
42-44
for details of these promising new techniques.
Photoionization Mass Spectrometry
The use of high-intensity resonance lamps in the vacuum ultraviolet
as photoionization sources for mass spectrometers allows many free radicals
to be observed directly in reacting gases. With the proper choice of lamp,
photoionization causes no fragmentation of other molecules to interfere with
the free-radical peaks — a major problem in conventional electron bombard-
ment sources. Steady-state concentrations of free radicals have been observed
17 46-48
when oxygen atoms react with hydrocarbons and in ozone-olefin reactions. '
With this technique, it was possible to resolve an argument of long standing
concerning the immediate products formed by the attack of oxygen atoms on
ethylene (mainly methyl and formyl radicals, with about 5% formaldehyde
and hydrogen).
The photoionization mass spectrometer can detect singlet oxygen
in a large excess of ground-state oxygen, and nitric oxide an excess of
nitrogen dioxide. ' Attempts to detect hydroperoxy radicals have not
been successful, probably because of a low photoionization cross section,
CO
but the methylperoxy radical has been observed with this technique.
Future research using the photoionization mass spectrometer should
result in significant progress in resolving some of the uncertainties in
2-37
-------
current chemical smog models, namely, detection of the immediate products
formed when hydroxyl radical reacts with olefins, determination of rate
constants and mechanisms for oxygen reacting with free radicals, identifica-
tion of the immediate products formed in ozone-olefin reactions, and deter-
mination of the rate constants and products of the reactions of alkylperoxy
radicals with olefins and other hydrocarbons.
Computer-Controlled High-Resolution Mass Spectrometry
The powerful technique of coupling a computer to a high-resolution
53
mass spectrometer has been used to analyze air pollutants. Both particulate
matter and gases can be scanned for up to 300 pollutants. Only stable
compounds will be detected by this method, because the samples are concen-
trated before analysis.
The advantage of this technique is the rapidity of monitoring for
many compounds simultaneously, including some of the liquid and solid inorganic
materials — such as sulfuric acid, ammonium sulfate, and ammonium nitrate
which may be the final products of the primary pollutants nitric oxide
and sulfur dioxide. Also, monitoring the many partially oxidized hydro-
carbons, such as aldehydes and acids, will give useful insight into the
reaction mechanisms involved in the atmosphere. One disadvantage of this
system is that only compounds that the computer is programed for will be
reported; unexpected compounds may be overlooked. A careful study of
sampling efficiency will be needed before quantitative concentrations
in the atmosphere can be reported.
Summary
The development of lasers has opened up several new techniques
for monitoring pollutants in the atmosphere. Sensitivities down to the
parts-per-billion range are claimed, and continuous monitoring is possible.
2-38
-------
The photoionization mass spectrometer has been developed as a sensitive
detector for free radicals in the gas phase. A high-resolution mass
spectrometer coupled to a computer is capable of detecting up to 300
compounds in air, both in particulate form and in the gas phase.
PRODUCTS OF PHOTOCHEMICAL SMOG — OBSERVATIONS AND SPECULATION
A large number of compounds can be formed in the polluted atmosphere.
As a result of the small concentrations involved and the great variety of
possible products, very few compounds have actually been observed. The gaseous
compounds for which quantitative measurements have been reported are listed
in Table 2-5 with typical concentrations. Compounds observed in particulate
matter will be discussed in the next chapter.
Which of the compounds in Table 2-5 are considered oxidants depends
on the reactant being considered. Certainly ozone, PAN, and hydrogen peroxide
are strong oxidants when biologic materials are considered. This list is
certain to grow as more sensitive analytic techniques are used and as modeling
studies suggest other important species that should be present.
Table 2-6 is a list of some compounds that may be present in
photochemical smog, but have not yet been reported. The presence of some
of these compounds seems very probable, in that they have been observed in
smog-chamber studies (such as PBzN and ketene), whereas others are very
speculative. For example, organic peroxy radicals, RC^, are almost certainly
important intermediates in the conversion of nitric oxide to nitrogen dioxide.
When these radicals undergo a chain termination reaction with hydroperoxy
radical, the corresponding organic hydroperoxide, ROOH, or peracid, RCOO?H
(if an acylperoxy radical is involved), will be formed. Although the rate
of this hydroperoxide formation is estimated to be less than that of the
similar formation of hydrogen peroxide by approximately a factor of 100,
2-39
-------
Compound
TABLE 2-5
Compounds Observed in Photochemical Smog
Typical (or Maximal)
Concentration Reported, ppm
Ozone, 0-j
PAN, CH3C002N02
Hydrogen peroxide, H.2°2
Formaldehyde, Cl^O
Higher aldehydes, RCHO
Acrolein, CH2CHCHO
Formic acid, HCOOH
0.1 (0.7)
0.004 (0.01)
(0.18)
0.04
0.04
0.007
(0.05)
Reference
Chapter 4
Chapter 4
58
Chapter 4
Chapter 4
Chapter 4
59
2-40
-------
TABLE 2-6
Compounds That May Be Formed In Photochemical Smog
Compound
Possible Reaction of
Formation
Reference
Peroxybenzoylnitrate,
65 22
Nitric acid, HONO
Organic hydroperoxides,
ROOH
Organic peracids, RCOO?H
Organic peroxynitrates,
Ozonides, Oo-olefin
Ketene, CH2CO
Nitrous acid, HONO
Pernitric acid, HO NO
Pernitrous acid, H02NO
Sulfoxyperoxynitrate,
R02 + HO
RC00
HO
RO + NO + M
2 2
0 + olef in + M
0 + olefin
N02 + H02
N02 + H02 + M
NO + H02 + M
HOS0202 + N02
M
57
60, 59
5 (p. 75)
5 (p. 75)
54
17
61
62
12
6
2-41
-------
the potential concentration must be combined with toxicity and other
information before a substance can be dismissed as unimportant.
Similarly, chain termination by R02 + NC^ could form organic
peroxynitrates. These would probably be less toxic than PAN, but they
could be present in comparable amounts. The formation of organic per-
54
oxynitrites and peroxynitrates has been considered, but they have not
been observed in smog.
If ozone-olefin adducts are stable in the gas phase, as a recent
study hinted, then they are almost certainly present in the urban atmos-
phere. Their concentrations will depend on their stability in the sunlit
atmosphere. If present, they are expected to be very reactive.
Pernitric (or peroxynitric) acid is an example of a compound that
could be present in significant quantities. This proposal is very
speculative, because there is no evidence of this compound in the gas
phase, although there is evidence of some such species in solutions.
The reaction of hydroperoxy radical with nitrogen dioxide is usually
written as,
H02 + NC»2 -> HONO + 02> (18)
which is certainly reasonable. However, at a pressure of 1 atm, the
three-body recombination step,
H<52 + N02 + M -*- HOON02 + M, (26)
also seems probable. The laboratory experiments that support Eq. 18
cannot rule out the formation of some pernitric. acid. The chemical modeling
studies show that, if Eq. 26 is assumed to be half as probable as Eq. 18,
6
the formation of pernitric acid would be comparable with that of PAN.
12
Similarly, recent experiments have been interpreted to mean that
about 10% of the reaction of hydroperoxy radical with nitric oxide gives per-
nitrous acid, HOONO, instead of nitrogen dioxide and hydroxyl radical. Because
2-42
-------
this reaction is of major importance, even 10% of a second channel would be
important, although it has been argued that such compounds would not be
54
sufficiently stable to accumulate in the atmosphere. Whether such peroxy-
nitrogen compounds are stable in the gas phase and whether they can be found
in the atmosphere must await further experiments.
The last entry in Table 2-6, sulfoxyperoxynitrate, is an inorganic
analogue of PAN. It was suggested in a computer study that investigated
6
the addition of sulfur dioxide to the NO -hydrocarbon system. Although
X
this is only speculative, such unpleasant compounds may become more important
if the sulfur content of fuels is allowed to increase in areas where an
oxidizing photochemical smog is common. It is known that sulfur dioxide, when
present in photochemical smog, is rapidly oxidized to sulfuric acid, and
a recent mass-spectrometric study also indicated the presence of organic
63
sulfur dioxide compounds.
Table 2-6 is only a sampling of the compounds that might be found
in photochemical smog in the future. The possible combinations among the
many free radicals and the oxides of sulfur and nitrogen are almost limitless.
Many undiscovered exotic compounds are present in photochemical smog, but
their concentration and importance remain to be established.
RECOMMENDATIONS FOR FUTURE RESEARCH
• Rate constants are needed for almost all the reactions of HO
and R02-
• The homogeneous and heterogeneous reactions of the oxides of
nitrogen with water vapor need study.
• The yields of free radicals from -the photolysis of nitrous acid
and of aldehydes should be established.
2-43
-------
Equation 4, [03] = k-jNC^]/k3[NO], should be tested in the real
atmosphere, as well as in laboratory experiments. Simultaneous
measurements of the concentrations of ozone, nitric oxide, and
nitrogen dioxide and of the intensity of sunlight for a variety of
conditions will provide a much-needed check on this dynamic
equilibrium.
A quantitative measure of the concentration of free radicals in
smog (probably OH or HO ) under well-defined conditions will provide
an important test of present chemical models.
Strong support for fundamental gas-phase kinetics is needed. Most
of the reaction mechanisms and rate constants that are needed to
construct realistic and detailed models of the polluted atmosphere
are determined in laboratory studies under very special conditions,
not in smog simulations at a pressure of 1 atm. Because there are
still very serious gaps in the present models, further research
should be supported.
Smog-chamber studies are needed for validating both the detailed
chemical models and the lumped models. Many of the past chamber
studies have not used sufficiently well-defined initial conditions.
Measurements of more products and of the reactive intermediates
will provide more stringent tests for models.
Modeling studies are very useful in pointing out the important
kinetic data that are lacking, in clarifying some of the past
smog-chamber studies, and generally in making the very complex
chemistry more comprehensible. Accurate models can make unique
predictions about the polluted atmosphere. There are very useful
interactions between the modeling studies, the smog-chamber
2-44
-------
experiments, and fundamental chemical kinetics; it is not
possible to ignore one without hurting progress in the others.
It seems probable that many new and unstable compounds are
present in the polluted atmosphere or in smog chambers. A care-
ful search for some of these compounds may provide some surprises.
Promising new instrumental techniques should be supported, both
for monitoring pollutants and for following reactive intermediates
in kinetic studies. A reliable and accurate method of standardizing
concentrations in the parts-per-billion range is needed.
The possibility that free radicals, particularly H0«, have signifi-
cant effects on biologic surfaces exposed to the irradiated atmosphere
should be investigated. Sticking coefficients are needed. In
experiments in which the observed biologic effects cannot be attributed
to the measured ozone and PAN concentrations, the possibility of
damage by the steady-state concentrations of free radicals in the
atmosphere should be considered.
2-45
-------
REFERENCES
1. Leighton, P. A. Photochemistry of Air Pollution. Hew York: Academic Press,
1961. 300 pp.
2. Altshuller, A." P., and J. H.~ Bufalini. Photochemical aspects of air pollution.
A review. Photochem. Photobiol. 4:97-146, 1965.
3. Altshuller, AT P., and Jf. J." Bufalini. Photochemical aspects of air pollution;
A review. Environ. Sci. Technol. 5:39-64, 1971.
4. Pitts, jf NTi Jr., and B," J.~ Pinlayson. Mechanisms of photochemical air pollu-
tion. Angew. Chem. (Engl.) 14:1-15, 1975.
5. Demerjian, K. L., J. A. Kerr, and J. G. Calvert. The mechanism of photochem-
ical smog formation. Adv. Environ. Sci. Technol. 4:1-262, 1974.
<
5a. Levy, H., II. Photochemistry of the troposphere. Adv. Photochem. 9:369-
524, 1974.
6. Calvert, J. G., and R. D. McQuigg.
Int. J. Chem. Kinet. (in press) (UNVERIFIED)
7a. Stern, A. C., Ed. Air Pollution. Vol. 1. Air Pollution and Its Effects.
(2nd ed.) New York: Academic Press, 1968. 694 pp.
7b. Stern, A. C., Ed. Air Pollution. Vol. ' 2. Analysis, Monitoring, and
Surveying. (2nd ed.) New York: Academic Press, 1968. 684 pp.
7c. Stern, A. C., Ed. Air 'Pollution. Vol. 3. Sources of Air Pollution and
Their Control. (2nd ed.) New York: Academic Press, 1968. 866 pp.
7d. Stem, A. C. , Ed. Air Pollution. Vol. 4. ( ed.) New York: Academic
Press, (in press) (UNVERIFIED)
2-46
-------
8. Tucker, A. W., M. Birnbaum, and C. I. Fincher. Atmospheric N02 determination
by 442-nm laser induced fluore«CMC«. Appl. Optics 14:1418-1422, 1975.
9. Heicklen, J., K. Westberg, and N. Cohen. Conversion of NO to N02 in Polluted
Atmospheres. Center for Air Environment Studies Publication No. 115-69.
University Park: The Pennsylvania State University, 1969. 5 pp.
10. Stedman, D. H. , E. D. Norris, E. E. Daby, H. Niki, and B. Weinstock. The
role of OH radicals in photochemical smog reactions. Paper Presented
at 160th National Meeting of the American Chemical Society, Chicago,
September, 1970.
11. Payne, W." AT, L.~ J." Stief, and D? D7 Davis. A kinetics study of the reaction of
H02 with S02 and NOT J7 Amer. Chem. Soc. 95:7614-7619, 1973.
4 , I _ f
12. Cox, R. A., and R. G. Derwent. Kinetics of the reaction of H02 with nitric
oxide and nitrogen dioxide. J. Photochem. 4:139-153, 1975.
13. Simonaitie, R., and J." Heicklen. Reaction of H02 with NO and N0£. J. Phys.
Chem. 78:653-657, 1974.
13a. Hack, W. K. Hoyermann, and H. G. Wagner, liber einiae Radikalreaktionen im
H-O-N-System. Z. Naturforschung. 29A:1236-1237, 1974.
13b. Hack, W., K. Hoyermann, and H. Gg. Wagner. The reaction NO -4- H02 —} N02
+ OH with OH -f H202 ^ H02 -f H20 as an H02-source. Int. J. Chem.
Kinetics, Symp. No. 1 (Chemical Kinetic Data for the Lower and Upper
Atmosphere):329-339, 1975.
14. Pate, C. T., B. J. Pinlayson, and J. N. Pitts, Jr. A long path infrared
spectroscopic study of the reaction of methylperoxy free radicals with
nitric oxide. J. Amer. Chem. Soc. 96:6554-6558, 1973.
2-47
-------
15. Slagle, I. R., J. R. Gilbert, R. E. Graham, and D. Gutman. Direct identifi-
cation of reactive channels in the reactions of hydroxyl radicals with
allene, propylene arid 2-butene. Int. J. Chem. Kinetics, Symp. No. 1
(Chemical Kinetic Data for the Lower and Upper Atmosphere):317-328, 1975.
16. Demerjian, K. L., J. A. Kerr, and J. G. Calvert. The mechanism of photochem-
ical smog formation. Adv. Environ. Sci. Technol. 4:166, 1974.
17. Atkinson, R., B. J. Finlayson, and J. N. Pitts, Jr. Photoionization mass
spectrometer studies of gas phase ozone-olefin reactions. J. Amer.
Chem. Soc. 95:7592-7599, 1973.
18. Stephens, E. R., and F. R, Burleson. Distribution of light hydrocarbons in
ambient air. J. Air Pollut. Control Assoc. 19:929-936, 1969.
19- U. S. Department of Health, Education, and Welfare. Public Health Service.
Air Quality Criteria for Photochemical Oxidants. National Air Pollution
Control Administration Publ. No. AP-63. Washington, D. C.: U. S.
Government Printing Office, 1970.
* * - *
20. Niki, H., E. E. Daby, and B. Weinstock. Mechanisms of smog reactions. Adv.
Chem. Ser. 113:16-57, 1972.
21. Bufalini, J. J. , and K. L. Brubaker. The photooxidation of formaldehyde at
low partial pressures, pp. 225-238. In C. S. Tuesday, Ed. Chemical
Reactions in Urban Atmospheres. Proceedings of the Symposium held at
General Motors Research Laboratories, Warren, Michigan, 1969. New York:
American Elsevier, 1971.
22. Glasson, W. A. Effect of carbon monoxide on atmospheric photooxidation of
nitric oxide--hydrocarbon mixtures. Environ. Sci. Technol. 9:343-347,
1975.
2-48
-------
23. Grimsrud, E. P., H. H. Westberg, and R. A. Rasmussen. Atmospheric reactivity
of monoterpene hydrocarbons, NO photooxidation and ozonolysis. Int.
X
J. Chem. Kinetics Symp. No. 1, 1975 (Chemical Kinetic Data for the Lower
and Upper Atmosphere):183-195, 1975.
24. Tuesday, C. S. The atmospheric photooxidation of trans_-butene-2 and nitric
oxide, pp. 15-49. In Chemical Reactions in the Lower and Upper
Atmosphere. Proceedings of an International Symposium Arranged by
Stanford Research Institute, San Francisco, California, April 18-20,
1961. New York: Interscience Publishers, 1961.
25. Schuck, ETAf, AT P." Altshullet, D? S? Barth, and C? S? Morgatt. Relationship
of hydrocarbons to oxidants in ambient atmospheres. J7 Air Pollut. Control
Assoc. 20:297-302, 1970.
26. Glasson, W, A., and C. S. Tuesday. Hydrocarbon reactivity and the kinetics
of the atmospheric photooxidation of nitric oxide. J. Air Pollut.
Control Assoc. 20:239-244, 1970.
27. Pitts, J. N., Jr., and B. J. Finlayson. Mechanisms of photochemical air pollu-
tion. Angew. Chem. (Engl.) 14:8, 1975.
28. Eschenroeder, A. Q., and J. R. Martinez. Concept and applications of photo-
chemical smog models. Adv. Chem. Ser. 113:101-168, 1972.
29. Hecht, TTAT, and J." H." Seinfeld. Development add validation of a generalized
mechanism for photochemical smog. Environ. Sci. Technol. 5:47-57, 1972.
30. Hecht, Tf A., J? H? Seinfeld, and M? CT Dodge. Further development of general-
ized kinetic mechanism for photochemical smog. Environ. Sci. Technol. 8:
327-339, 1974.
31. Firestone, R. F., and J. G. Calvert.
(in press)
2-49
-------
32. Wayne, R. P. Singlet molecular oxygen. Adv. Photochem. 7:311-371,
33. Lloyd, A. C. Evaluated and estimated kinetic data for phase reactions of
the hydroperoxyl radical. Int. J. Chem. Kinetics 6:169-228, 1974.
34. Wendschuh, P. H., C. T. 'Pate, and J. N. Pitts, Jr. The reaction o£ peroxyacetyl
nitrate with aldehydes. Tetrahedron Lett. 31:2931-2934, 1973.
35. Ku, R. T., E. D. Hinkley, and J. 0. Sample. Long-path monitoring of atmos-
pheric carbon monoxide with a tunable diode laser system. Appl. Optics
14:854-861, 1975.
36a. O'Shea, 0." Cf, and L." GT Dodge. N02 concentration measurements in an urban
atmosphere using differential absorption techniques. Appl. Optics 13:
1481-1486, 1974.
36b. Menzies, R."T., and M." T.' Chahine. Remote atmospheric sensing with an airborne
laser absorption spectrometer. Appl. Optics 13:2840-2849, 1974.
36c. Patty, R? Rf, G." M." Russwurm, W." A." McClenny, and D." R.'Morgan. Co2 laser
absorption coefficients for determining ambient levels of 0,, Ntt,, and
C2H4. Appl. Optics 13:2850-2854, 1974.
37. Becker, K." H., U." Schurath, and T." Tatatcayk. Fluorescence determination of low
formaldehyde concentrations in air by dye laser excitation. Appl. Optics
14:310-313, 1975.
38. Wang, C. C., and L. I. Davis, Jr. Measurement of hydroxyl concentrations in
air using a tunable uv laser beam. Phys. Rev. Lett. 32:349-352, 1974.
39. Kreuzer, L." Bf, Nf D." Kenyon, and C.~ K." N? Patel. Air pollution: Sensitive
detection of ten pollutant gases by carbon monoxide and carbon dioxide
lasers. Science 177:347-349, 1972.
40. Maugh, T. H., II. Photoacoustic spectroscopy: New uses for an old tech-
nique. Science 188:38-39, 1975.
2-50
-------
41. Davies, P. B., D. K. Russell, B. A. Thrush, and F. D. Wayne. Detection of
amino radical NH~ by laser magnetic resonance spectroscopy. J. Chem.
Phys. 62:3739-3752, 1975.
I
42. Colles, M. J., and C. R. Pidgeon. Tunable lasers. Rep. Prog. Phys. 38:
329-460, 1975.
43, Melngailis, I.
IEEE J. Quant. Electr. QE-10:7- , 1972.
44. Hlnkley, E. D., K. W. Kill, and P. A. Blum.
pp. . In H. Walther, Ed. Laser Spectro^-
scopy of Atoms and Molecules. (in press)
45. Jones, 1." TT K., and K.~ Df Bayes. Detection of steady-state free-radical con-
centrations by photoionization. J7 Amer. Chem. Soc. 94:6869-6871, 1972.
46. Jones, I. T. N., and K. D. Bayes. Free radical formation in the atomic
oxygen plus acetylene reaction, pp. 277-284. In Proceedings of 14th
Symposium (International) on Combustion, 1972. Pittsburg: The
Combustion Institute, 1973.
47a. Washida, N., and K. D. Bayes. The reaction of methyl radicals with atomic
oxygen. Chem. Phys. Lett. 23:373-375, 1973.
47b. Slagle, IT R., P." J^Pruss, Jr., and D. Gutman. Kinetics into the steady state.
I, Study of the reaction of oxygen atoms with methyl radicals. Int. J.'
Chem. Kinetics. 6:111-123, 1974.
48, Washida, N., R. I. Martinez, and K. D. Bayes. The oxidation of formyl radicals.
Z. Naturforsch. 29A:251-255, 1974.
49a. Kanofsky, J. R. , and D. Gutman. Direct observation of the products produced
by the 0-atom reactions with ethylene and propylene studied in high-
intensity molecular beams. Chem. Phys. Lett. 15:236-239, 1972.
2-51
-------
49b. Kanofsky, J. R., D. lucas, and D. Gutman. Direct identification of free-
radical products of oxygen atom reactions with olefins, using high-
intensity molecular beams, pp. 285-294. In Proceedings of 14th
Symposium (International) on Combustion, 1972. Pittsburgh: The
Combustion Institute, 1973.
49C- Pruss, FT J., Jr., I. R. Slagle, and D. Gutman. Determination of branching
ratios for the reaction of oxygen atoms with ethylene. J7 Phys. Chem.
76:663-665, 1974.
50. Jones, I." Tf H., and R? E>T Bayes. Formation of 02 (*!A.) by electronic energy
O "
transfer in mixtures of N02 and Cy J. Chem. Phys. 59:3119-3124, 1973.
,' J t i >
5-^ Jones, I. T. N. , and K. D. Bayes. Photolysis of nitrogen dioxide. J. Chem.
Phys. 59:4836-4833, 1973.
' - . '' . '
52. Washida, N., and K. D. Bayes.
(in press)
53. Schuetzle, D., A7 17 Crifctenden, and R." J." Charlson. Application of computer
controlled high resolution mass spectrometry to the analysis of air
pollutants. J. Air Pollut. Control Assoc. 23:704-709, 1973.
54. Demerjian, K. L., J. A. Kerr, and J. G. Calvert, The mechanism of photochem-
ical smog formation. Adv. Environ. Sci. Technol. 4:77, 1974.
55. Moeller, T. Inorganic Chemistry. New York: John Wiley & Sons, 1952.
.p. 613.
Delete 56--omitted.
1| .. ... .,.. r -•
57. Heuss, JTM., and W7 A." Glasson. Hydrocarbon reactivity and eye irritation.
Environ. Sci. Technol. 2:1109-1116, 1968.
2-52
-------
58. Bufalini, J. J., B. W. Gay, Jr., and K. L. Brubaker. Hydrogen peroxide forma-
tion from formaldehyde photooxidation and its presence in urban atmospheres.
Environ. Sci. Technol. 6:816-821, 1972.
59. Hanst, P. L. , W. E. Wilson, R. K. Patterson, B. W. Gay, and I. W. Chaney.
Paper Presented at 167th National Meeting of the American Chemical
Society, Los Angeles, April 1974. (UNVERIFIED)
60. Miller, D. F., and C. W. Spicer. A Continuous .Analyzer for Detecting Nitric
Acid. Paper 74-17, Presented at 67th Annual Meeting of the Air Pollution
Control Association, Denver, Colorado, June 1974.
61. McAfee, J. M., A. M. Winer, and J. N. Pitts.
In CODATA (Committee on Data for Science and
Technology) Symposium on Chemical Kinetics Data for the Lower and Upper
Atmosphere, Warrenton, Virginia, Sept. 1974. (in press) (UNVERIFIED)
62. McAfee, J. M., J. N. Pitts, and A. M. Winer.
In Pacific Conference on Chemistry and Spectroscopy, San Francisco,
October 1974. (in press) (UNVERIFIED)
63. Schulten, H.-R., and U. Schurath. Analysis of aerosols from the ozonolysis
of 1-butene by high-resolution field desorption mass spectrometry. J.
Phys. Chem. 79:51-57, 1975.
64. Stedman, D. H., and J. 0. Jackson. The photostationary state in photochem-
ical smog. Int. J. Chem. Kinetics Symp. No. 1, 1975 (Chemical Kinetic
Data for the Lower and Upper Atmosphere):493-501, 1975.
65. Eschenroeder, A. Q., and J. R. Martinez. Analysis of Los Angeles Atmospheric
Reaction Data from 1968 to 1969. General Research Corp. CR-1-170, 1970.
(UNVERIFIED)
2-53
-------
CHAPTER 3
AEROSOLS
Organic compounds were recognized long ago as key ingredients
of the polluted atmosphere and constitute a significant fraction of
I
the urban aerosol associated with photochemical smog. Although
there is no air quality standard for organic aerosols, ambient concen-
trations of particulate organic substances are related to the concentra-
tions of total suspended particles, for which there are federal and
state standards,* and to prevailing visibility. Primary organic
compounds are emitted directly into the atmosphere, and their concentra-
3
tions can be reduced through emission control. Secondary organic
compounds result from gas-phase photochemical reactions involving
hydrocarbons, nitrogen oxides, and ozone and thus imply the same type
4,5
of control strategies as for ozone and photochemical oxidants.
However, control of secondary organic compounds would require a more
elaborate approach, with the identification of specific hydrocarbon
precursors as a necessary step (Figure 3-1). Because of their accumu-
lation in the submicrometer range, secondary organic aerosols fVve^ b«-
responsible for adverse health effects and contribute significantly to
visibility degradation. Despite its importance, the gas-to-particle
conversion of organic pollutants has received much less attention than
other aspects of air pollution. This section deals with the identification
of secondary organic aerosols in the atmosphere and the physical and
chemical aspects of their formation. The relative importance of their
gas-phase hydrocarbon precursors, including naturally emitted terpenes,
*The national ambient air quality standards for particulate matter are;2
24-h averages, not to be exceeded more than once a year: primary, 260
yg/mj; secondary, 150 yg/m3. Annual geometric means: primary, 75 yg/m3;
secondary, 60 yg/m3.
-------
O
a:
o
c_>
CL
UJ
i
co
o
or
o
o
LU
CO
o
w
o
!-i
QJ
cfl
O
•H
C
CO
60
J-!
CO
T3
O
O
0)
U)
•H
M
O.
J-i
O
CO
CU
•H
bO
a)
4-1
cd
^
4-1
M
O
J-I
4-1
O
u
I
ro
00
•H
3-2
-------
will be discussed, and the contribution of hydrocarbon photochemical
reactions to the formation of inorganic sulfate and nitrate aerosols
will be briefly reviewed.
Atmospheric Observations
Analytic Techniques. Identifying the chemical composition of urban
aerosols, in which hundreds of components may be present, is a formida-
ble task for the analyst. Although rather simple techniques can be used
to quantify the aerosol composition by elemental analysis (e.g., organic
carbon) and groups of compounds (e.g., nitrates, sulfates, and organics
or aliphatics, aromatics, and oxygenates in the organic fraction), only
more sophisticated techniques can provide individual identification of
compounds present at very low concentrations. The necessary selectivity
and sensitivity are then gained at the expense of a general picture of
the aerosol composition.
Aerosol samples can be analyzed without further preparation for
organic carbon and for carbon, hydrogen, oxygen, and nitrogen (CHON);
with infrared spectroscopy; ' with photoelectron spectroscopy; and
11
with mass spectrometry. However, organic analysis is generally con-
ducted after solvent extraction of samples collected on glass-fiber
12
filters. The organic extract can be further fractionated by thin-layer,
paper, liquid, or column chromatography, by ion-exchange chromatography,
14
or by separation into different types of solvents. The extracts or
6 15-19
fractions are then analyzed for organic carbon or for CHON by
15,16,18-21 22-29 23,30-34
infrared, ultraviolet fluorescence, gas chromatography,
35 36-40
high-pressure liquid chromatography, and mass spectrometry. For
example, more than 70 polynuclear aromatics have been identified with
3-3
-------
41
mass spectrometry. Some of these techniques have been used in
9,10,18,19,29
conjunction with size distribution measurements. Vacuum
34 42
sublimation and, more recently, ultrasonic extraction have been
proposed as alternative techniques for the time-consuming process that
uses the Soxhlet extractor. However, most of the later analyses depend
on the extraction efficiency of the organic solvent. Although nonpolar
solvents were most widely used in the past (cyclohexane, benzene
43 '
for the National Air Surveillance Network ), it has been found
6
recently that solvent extraction efficiencies depend on solvent
polarity characteristics (Figure 3-2) and that binary mixtures of a
polar solvent and a nonpolar one can extract up to 48% more organics
than benzene alone from samples collected in areas of heavy photochemical
pollution (Table 3-1)
Primary Organic Aerosols. Primary organics are emitted to the atmosphere
by industrial sources (oil refineries, chemical plants, producers and users of
solvents and plasticizers), vehicles (as a result of incomplete fuel combus-
tion, oxygenated degradation products of lubricating oil, polymers from
tires), and agricultural activities (use of pesticides). An exhaustive
literature survey is beyond the scope of this section, but can be found
44
in Air Quality Criteria for Particulate Matter; many useful references
6-43
are also available.
Among the identified primary organics are linear and branched alkanes
and alkenes, substituted benzenes and styrenes, quinones, acridines, quino-
lines, phenols, cresols, phthalates, fatty acids, carbonyl compounds, and
some pesticide compounds. Diurnal concentration profiles of primary pollu-
tants emitted in auto exhaust parallel vehicle activity and show two
3-4
-------
h-
LJ
211
Q Q_
lO
000
com
o
o
o
o
ro
O
O
oo
O
o
Ol
T3
01 -H
M 4H
CU i-H
4J 3
CU CD
ttj 'O
o
cfl
CO Pi £! M
ft O « O
,£> 0) 4H
>, M •£ O
4J CO O M
•H O rH O
1-1 >> ^
,0 oT y u
3 T)
i—t -i-i " •
O M M CU
01 o o) c
,
O
M en
ft O
o
CO M
O
*-< lw* ^
O (U TO
cu
_ - O
•H M M
^ « rH
"S^
CD O (-1
s e
cfl 0)
><; ft
T3 4-1
o
d cu
o >
•rl r-l
4-1 O
(3
4-1 •
CO r~r-
COW ..,5
s-^ T3 O
W n)
T)
CU
c" a
CO -H
3 ft
<4-l CU
O P**!
CO M
M 0)
4-1 4-1
0) CO
CO
CU
o co fx 6
QJ CU *^ 0)
•H 4-) JJ M
O CU >> 0)
4-1
0)
. . „
14-1 B
14-! I-l .
0) n) ro
J-i rH
O /^
rH Ol
U -H
X
01 O
C MH
01 rH
H D
t^, CD
01
01
C r-l
O >^
•H 4-1 (3 -
4J -H O N
y s-i cu fi
CO CO M CU
^i i—i pL4 ,n
4-J O
cu cu cu
T3 C (3
4-i pi n) cO
PI CO 4J 4J
CU Pi O
> " 0) O
rH ^-N ft O
,8£i|S
SI
•H ~a
B
CO •
rH rH
r^ O
4-i nj
0) J3
•H 4-1
M 0)
4-1 E
LJ
CNI
I
ro
60
•H
3-5
-------
rH
1
co
01
H
cd
H
PN"
W 1
O CO
rH
03 CX
cu S
•H cd
O CO
(3
CU rH
•H O
U 03
•H 0
4-1 !-4
<4-l CU
w <
f3 CJ
0 iH
*rH J~l
4J CU
0 43
cd cx
!-l CO
4-1 O
X 0
W 4J
a
0 M
43 O
m 14-1
cd
O CO
CU
CJ VJ
•H 3
fi 4J
cd X
60 -H
^•4 S
o
13 S-i
(3 cd
cd C3
•H
s-~. pq
Pfl 13
Sl"— S ri
cd
CO
CU CO
•H 4-1
0 13
C CU
CU >
•H rH
0 0
•H CO
14-1
M-l CO
pq y
O
C3 -H
O f-4
•H cd
4-1 ?*
CJ
cd m
M O
4-1
X
W
P4 ^ i*H -^ r**»
O i — 1 co co *3"
r^ i-l oo O
pm to G** O
U vo r-~
O
pt< CJ\ Q , — | \Q
pq vo r-^ r*^ r*^
CU
13
•H
^i
O
rH CU
43 13
0 -H
4J Cd <4-l
a M H
CU 4J 3
> cu ca
rH 4-1 1-1 CU CO
O 13 £3 rH
CO C3 cd rH
0 C 4J
CJ O CJ (3
•H 43 CU O
rH M pn CU
•H td 1 . M
CO 0 C3 Fn
mvOO^r-ICMO-*
OOCN(NC5rHiTiiH
T— i co *d~ 10 oi -^" 10 oo
CN CN CN Ol Ol CsJ Ol Ol
i j
n"t T~T
0 0
a rH £3 rH
ty (U T~H O cj CD rH O
P, f^ o c3 cx G o fi
oocJcdooscd
J-l 4-1 Cti 42 M -W C^J rC
PL, OJ^^PH aj^^J
i u -u> cu i o 4-i a)
oi i CO
O H
i^ oo oo oo o\ o
1-1
r*^ oo tH CM •'j in o CM
r~r-~oooocoaNOO
rH rH
CU
13 /-,
•H CU
J-t CJ
0 fi
H CU
43 t-i
U CU
Ol cd M 14-1
C M 0) CU CU
4-1 CU 4-1 td CU *— '
£3 CU 4-1 CU X C3
CU C3 rH CU td PJ G)
ex cd (3 t*"i 43 4-i £5 C3
O X O ,*~i O O fT1 CU
rH CU 43 4-J iH O t*J 3
CJtBMCUOOSi-l
>> 1 . cd -H ^ CO W O
0 (30OUMPQH
m
n
m
CM
CM
iH
O
td
ex
o
PH
1
O)
-,rH CdvOr~CX4-ICU
O 4343 cocMroS£3c3
rH 4J cj ca 01 a)
43 CU CU CO [> N
US 43 •• rH 0
4-1 •• H CU O CU
CU O a CO 43
!-! (3 (3 cd
O O Cd 03 « "
14-1 4J 43 iH (3 (3
0) 4-1 •> O O
inOvor^OOO O 13UCU4-1CU4343
rHrHrHvDt-^CTiO CO CUCdaf3£3t-lM
I^JvOOrHrHrHrHi— IrHCMlCM J-( CUCU CdCd
rHrHcMIIIIIII 1 3 >NOO
in ^^ in vo vo \o vo r*^ td o cu o o
CU 1 — CM CO 43 -H -H
ocMinr~-oOrHoo-* SOCM Cc3
vDvOCT\CNOOCMvDCMO\ rHrH^'-tdtd
CM OO CO *>O CM CM CM CO CO *^ in in in fT l r^ 'O 6C 60
iHrHCNCMIItllllll Fx3T3T3CUcU>-iM
rHrHrHrHOCMinOrHininOCO OC3C34-14-J OO
OrHrHr^r~cr.ooo>oo ocdcdou
rHiHrH rH Cdtd13TJ
\0 TJOrHt-lhCUCU
£3 CO CT> 4-1 4J 4-14-1
£3 • CdCMCMXK OO
td cu cu cu cd cd
CU Ol CU 6 fe M t-i
13 13 T->3W"rHrH4-14-l
•H -H COrH rH-'Cdcd.XX
!-i£3 X OO4-iOrH4-i4JCUCU
Ocd O S-i>t3c3OOO
CUHM iw O 0)td!34J4->X
CUC3433 rH |^>CXCd
£3^HO«4-ICU 3 6431-IO43X O
•H00OCiH CO O OM4J O
Bficdcut-itd o Ms-^cocxcuo rH
CdOrH£3'dXcU£3 rHrH M-lOII O
rH IW ^» CU r**! O £3 Cd CU rH O P*> in J-t CM rH ||
F*^» O 43 rH 43 *H "H CX £3 O £3 43 cd •• cd
43 J-i 4J $*°t cd O T3 O O c3 cd 4-1 M 4-1 &^S rH II *H
4-1 O CU 43 ^4 1 *H (-1 4-1 cd 43 CU CU td CO O PH
CUrH-H4-l4-l,l04JCU'HCd , , , Pn CJ
QUHgHrHfXiCM-^pLlgOS «d43 0 Fd O
3-6
-------
characteristic peaks associated with morning and evening traffic
45
periods. Because of their potential carcinogenic activity, poly-
nuclear aromatic hydrocarbons have been extensively studied (see Foster
35
et^ al. ). Recent studies have reported their identification in
40
automobile exhaust, their ambient distribution with respect to parti-
29
cle size, the close relation between their ambient concentrations and
28,46
automobile traffic in the Los Angeles area, and their possible loss
47
during sampling as a function of their volatility. The recent observa-
48
tion of chemiluminescence associated with particles in auto exhaust is
of interest, in view of the fact that the carcinogenic action of electron-
ically excited molecules is much greater than that of the corresponding
49
ground state.
Secondary Organic Aerosols. Organic aerosols formed by gas-phase photo-
chemical reactions of hydrocarbons, ozone, and nitrogen oxides have been
identified recently in both urban and rural atmospheres. Aliphatic organic
19,50,51 50
nitrates, such dicarboxylic acids as adipic and glutaric acids,
carboxylic acids derived from aromatic hydrocarbons (benzoic and phenyl-
acetic acids) and from terpenes emitted by vegetation, such as pinonic
52
acid from a-pinene, have been identified. The most important contribu-
tion in this field has been that of Schuetzle et al., who used computer-
11,53,54
controlled high-resolution mass spectrometry and thermal analysis,
the only available technique that combines the resolving power necessary
to identify individual pollutants at very low concentrations with the
ability to detect the wide range of compounds (metals, inorganics, organics)
present in polluted atmospheres. Organic aerosols of secondary origin
11,53,54
identified by the University of Washington group—Schuetzle et al.
3-7
-------
45
and Knights e_t aJ^. —are listed in Tables 3-2 and 3-3. It can be
seen that most of them are difunctional compounds, most probably formed
in the atmosphere by photochemical oxidation of cyclic alkenes and alka-
dienes. The sensitivity of the technique permits obtaining diurnal
variations of primary and secondary organics from 2-h size-resolved
samples. Diurnal concentration profiles of secondary organic aerosols
follow ozone variations closely. Such typical profiles are illustrated
for some selected compounds in Figure 3-3.
Relative Abundance of Primary and Secondary Organic Aerosols. If primary
organic aerosols are preponderant in urban areas where photochemical
reactions are not significant, secondary organic aerosols are predominant
in photochemically polluted areas, such as the south California basin.
As part of the recently completed California Air Characterization Study
55,56,57
(ACHEX), a close relation between visibility reduction and photo-
chemical smog was observed at eight sampling locations where the extinction
coefficient, b , due to light scattering from airborne particles was
scat
measured when ozone concentration was at its maximum (Figure 3-4). Good
51,58
correlations were obtained between the organic aerosol fractions,
51
their carbonyl infrared absorption band intensities, and ozone concentra-
tions .
59
With the concept of chemical-element balance developed by Friedlander,
secondary organic aerosols have been estimated to account for 82% and 76%
of the aerosol carbon balance in 24-h and 2-h samples (Pasadena, September 29,
60
1973 ). During a severe photochemical episode, secondary organics reached
up to 95% of the total organics (Pasadena, July 25, 1973; maximal ozone
3-8
-------
labla "-:
..,:,^y Organic Aerosols"
Compounds Identified
Al
1.
2.
iphatic multi
X-(CH2)n-Y
X
COOH
COOH
COOH
COOH
or^COH
COH
COH
COOH
or b_coH
COH
COOH
COOH
Others:
functional compounds:
(n-3,4,5):
Y
CH2OH
COH
COOH
CH2ONO
CH2ON02
CH2OH
COH
COONO
COON02
COONO
COON02
CH2ON02
CH2OH-CH=C (COOH) -CHO
CH2OH-CH2-CH=C (COOH) -CHO
CHO-CH=CH-CH (CH 3 ) CHO
CH2OH-CH=CH-CH=C(CH3)CHa
C5H803 isomersk
Nitrocresols b
Aromatic monofunctional compounds;
3. C6H5-(CH2) -COOH (n = 0,1,2,3)
n
4. C6H5-CH2OH
C..H.CHO
b b
Hydroxynitrobenzyl alcohol
Terpene-derived oxygenates:
5. Pinonic acid
Pinic acid
Norpinonic acid
b
6. Tsomers of pinonic acid:
CgHii+02 isomers
ClflHit+'-'s isomers
C10H16°2 isomers
— Data from Knights et al.;
1974.
Possible Gas-Phase Hydrocarbon
Precursors
1. Cyclic olefins
'CH
>C=CH-(CH2) -CH=C<
n
2. Not known; possibly from aromatic
ring cleavage
3. Alkenylbenzenes
C6H5-(CH2) -CH=CHR;
n
also toluene for C6H5COOH
Toluene, styrene, other
monoalkylbenzenes?
a-Pinene
6. Other terpenes?
compounds identified at West Covina, California, July 24,
— Isomers not resolved by mass spectrometry.
3-9
-------
Table 3-3
Relative Importance of Aliphatic and Aromatic Precursors
Gas-Phase Hydrocarbon
Precursors
^-— — — CH
r ||
(CH2) II
C=CH-(CH2) -CH=C<
n
CrHc-CH=CHR
b 0
Secondary Organic
Aerosols
X-(CH ) -Y
2 n
X Y
COOH CH2OH
COOH COH
COOH COOH
or rCOOH CH2ONO
COH CH2ON02
COH CH2OH
COH COH
or rCOOH COONO
COH COON02
COH COONO
COOH COON02
COOH CH2ON02
Total
COOH-CH2-COOH
COOH- (CH2) 2 -COOH
Concentration ,—
n=3 n=4
2.18 3.40
1.39 2.59
1.35 0.78
1.01 0.40
0.31 0.40
0.30 0.24
0.14 0.24
1.01 0.14
0.12 0.15
7.81 8.34
0.15
0.57
yg/m
n=5
0.65
0.82
0.15
0.27
0.13
2.02
Total difunctional compounds: 18.89
CCH -(CH2) -COOH
b b fl
Total from aromatics:
n=0:0.38
n=l:0.41
n=2:0.52
n=3:0.03
1.34
Of aerosols in Pasadena, California, September 22, 1972; sampling period, 7:30 a
t-n 1 9 • ^ r> TTI ArlflDtpd f-rnm '^r-Viiiot-^1 o ot- al -'^
.m.
to 12:35 p.m. Adapted from Schuetzle et al.
— The same response factor (that of adipic acid) was used for all difunctional
compounds.
3-10
-------
9-
8-
7-
6-
5-
4-
3-
2-
1-
DICARBOXYLIC ACIDS
IN PARTICLES <1/um DIA.
(CONC 'fn ug/m3)
•
West Covina, California
(4O km E of Los Angeles) ^=
*
7/24/73
»
/-COOH
1- -j V-COOH
' PENTANEDIOIC ACID
(GLUTARrC ACID) fT*™
•
^mmm^mm
HEXAN
(ADIP
»
MEim
,
1
/-COOH
1 \ /COOH
IEDOC ACID
1C ACID)
| CH3^
'LHEXAls
sss^ssa
C^OOH
COOH
IEDIOIC AC
[—
— — — TT~-
ID
^•••1^
•^^••i
mmmmm
••«••
/
mmmmm
S
ma;
9"9V
^^^m
/
ozo
< 0.5
fmmmm^
•••••
\
ne
>4pp
BSS
MtaMM
»^ ^ «t •
ITN\
STTSTTa
_=
v 1 >•*
21
0
12
15
18 21
24
TIME OF DAY (PDT) 7/24/73
Figure 3-3. Dicarboxylic acid diurnal profiles.
from Knights et al.45)
(Reprinted with permission
3-11
-------
14
12
10
B „ AT PEAK OZONE
5 Ca u
(SUMMER 1973)
O ROUBIDOUX (RIVERSIDE)
A WEST COVINA
D POMONA
O DOMINGUEZ HILLS
(TORRANCE)
(SUMMER 1972)
PASADENA
RIVERSIDE
POMONA
HARBOR FWY.
(DOWNTOWN LA)
BLIMP FLIGHT
(9/6/73)
1315-1500 PST
BLIMP
NO AEROSOL IN THE ATMOSPHERE
.1
.2 .3 .4
03 ppm (MAXIMUM)
Figure 3-4. Correlation between b and maximal ozone concentration. Based on
c* f* 3 1~
2-h averaged data taken in the Los Angeles area. (Reprinted with
permission from Hidy et al.56)
3-12
-------
51
concentration, 0.67 ppm ). In such extreme situations, all secondary
material resulting from gas-to-particle conversion (secondary organics plus
nitrates plus sulfates) accounts for up to 95.5% of the total aerosol
mass (Figure 3-5). Size distribution measurements on short-period sam-
61
pies, first developed by Lundgren, show that difunctional compounds
45,54
listed in Tables 3-2 and 3-3 accumulate in the submicrometer range
and that a significant fraction of the organic carbon is found below
55,56
0.5 ym« These results clearly illustrate the potential health
3
hazards associated with human exposure to high concentration (^300 yg/m )
of respirable secondary particles during acute photochemical episodes.
Gas-Particle Distribution Factors. The extent of gas-to-aerosol conver-
sion of secondary pollutants can be estimated by measuring gas-particle
51
distribution factors for carbon, nitrogen, and sulfur species. For
3
example f = P/(P+G), where P = particulate organic carbon ( pg/m as
C
carbon) and G = gas-phase reactive hydrocarbons (total hydrocarbons -
3
[methane + acetylene]) converted from ppm as methane to yg/m as carbon.
It can be seen from measured values of f , f , and f (Table 3-4) that
N S> C
the extent of conversion of organic gases is much lower than that of
sulfur and nitrogen oxides. Similar f , f , and f values were calculated
56,57 N S C
from ACHEX data. The maximal f , measured over a 1-h period when
d
the highest ozone concentration in several years was recorded in Pasadena,
was only about 6%, and average values are in the range of 1-2%. This,
combined with estimates of air-mass travel times based on air trajectory
62
analysis, provides an upper limit of about 2%/h for the ambient gas-to-
aerosol conversion rate of organic gases in photochemically polluted
51
atmospheres. Although organics always account for an important
3-13
-------
100%
80%
60%
40%
20%
0
S04~~
N03'
Organics
6--30
I
I2--30 16:30
time, PDT
0.6
0.4 1
0.2
Figure 3-5. Hourly variations of secondary aerosol: organics, nitrates, sulfates, and
ammonium as percent of total aerosol. Pasadena, July 25, 1973. (Reprinted
with permission from Grosjean and Friedlander.51)
3-14
-------
sf
1
CO
CD
rH
rH
lv
H
ts
x^
C/l
U-4
CO
Sf CO O CO rH \& |-^
en CO CO «-H I-H LO \O CO -3" i— I >3"
i — 1 i— 1 LO LO VO CN i— 1 i — 1 r— 4
CO
C CN r*"^ ^D LO oo ^H co o^ ^^ ^D r^" oo CN *~H 10 oo F^* ^o LO co LO oo
Cd CN CO CN «
O CM in rH
0 rH CS| -v^
cfl — -^ O
cfl r>. r-^ rH
C
o
0) -H
tO 4J
CO CO
r. • P, 4J
«r-~ CM i c!
O rH O CO CJ
O *'*•- c/3 CO cj
.. O 60 C
Sf rH CO O
rH CO |IO
1 •"
O O CO O
O CO & ' — '
" " "tr, ^
CM ON 60 i — i
rH rH ^1
1
»O « C
O CO CO o
O " -U -H
.. r~- 'H 4J
CM i-H C CO
— 1 3 ,-1
1 01 4-1
O rH •" (3
O A O Cl)
.. B i — i o
oo cO CM c
en q o
« e/i o
CM 4J ' — '
rH en CD
^*^ co T- co
r*^ rH co
•> ' ' Q*
'• 0 II 1
CD CO sf
rH I-- • Sf )-|
CX O O cfl
0 43 O en ex
cfl •• ' — i
en rH r- II
rH II PM
• mo en x
CO CM O «w ' — i
pj *• — •• ^)
a) r^ m
T3 rH
cO ••*
CO O T3
CO O Cl
PH •• CO
sf •
C CM « CM
•H I O O
O O ^
OCO ••
fi •• in W
•H O -H CO
rH CM 1
CX O CO
B t3 o B
cfl C •• ^
i— 1 3.
ft
r-H O * r.
moo en
• •• O 4J
r-l O •• -H
CU CM CO C
*TJ | rH rj
C O 1
CO O O ••>
rH •• O O
TJ 00 •• ' '
0) rH CM X
•H rH O
r-l " 123
fV| f^*> «\ • '
O O
T3 •• O +
C 00 ••
Cfl rH CM PH
1 rH ' '
C 0 1 1
cfl O O co
CU .. O O
•r-)VO •• "Z
t/J rH O 1 '
O « — I ' —
rJ •« pM
rj o •> ' — i
O O 1
e •• o co
O vD •• O
r-l rH 0 f5
M-l | rH 1 1
O 1
CO O O II
4J " O
cfl sf •• £3
O rH 00 IH
« rQ
3-15
-------
fraction of urban aerosols, only a small fraction of the organic gas
is converted to particulate material. This, and the fact that most of
the secondary organics identified in ambient aerosols are difunctional
compounds that can be formed from a few possible precursors (such as
cyclic olefins and diolefins), leads us to review next data obtained in
smog-chamber studies of organic aerosol formation from olefinic and other
hydrocarbons.
Smog-Chamber Studies
Since the pioneering work of Haagen-Smit and co-workers more than
-------
Table 3-5
Compounds Investigated in Smog-Chamber Studies, with References
I. OLEFINS
Alkenes:
Ethylene 65,66,68,69,73,75,79,88,89,91.
Propylene 69,70,73,79,86,90,91,95,96.
1-Butene 68,70,75,80,84,87,88,89,95,99.
cis- and trans-2-Butene 65,66,68,70,72,85,97.
Isobutene 65,66,68,69,70.
2-Methyl-l-butene 70
2-Methyl-2-butene 65,66,67,70,88,89,90.
3-Methyl-l-butene 68,70.
1-Pentene 65,66,68,70,74,88,89.
cis- and trans-2-Pentene 68,70,85,90.
1-Hexene 64,68,70,72,74,77,80,82,85,87,95,98
cis- and trans-2-Hexene 68,70,82.
2-Methyl-l-pentene 85
4-Methyl-l-pentene 70,90.
cis-3-Methyl-2-pentene 68
2,3-Dimethyl-2-butene 65,66,69,70,100.
1-Heptene 52,72,79,84,88,89,94.
cis- and trans-3-heptene 68,72,79,88,89.
5-Methyl-l-hexene 94
1-Octene 76,79,81,94,95.
trans-4-Octene 94
2,4,4-Trimethyl-l-pentene(isooctene) 68,79.
3-17
-------
Table 3-5 (Cont.)
Compounds Investigated in Smog-Chamber Studies, with References
1-Decene 95
1-Dodecene 95,98.
Cyclic Olefins:
Cyclopentene 68,79,100.
Cyclohexene 52,64,65,66,68,72,79,81,82,88,89,92,94,95,98,100.
Diolefins;
1,3-Butadiene 65,66,68,75,79.
Isoprene 79
1,3-Hexadiene 98
1,5-Hexadiene 68,,82,94,98.
1,6-Heptadiene 94
2-Methyl-l,5-hexadiene 94
1,7-Octadiene 94., 100.
2,6-Octadiene 94
Dicyclopentadiene 68
Other Olefins:
Indene 94
Turpentine 81
Styrene 79
a-Pinene 52,78,79,82,92,94,98.
3-18
-------
Table 3-5 (Cont.)
Compounds Investigated in Smog-Chamber Studies, with References
II. PARAFFINS
Methane 65,66,79.
Cyclopropane 68
n-Butane 68,88,89.
n-Pentane 65,66.
2,2-Dimethylpropane 65,66.
Cyclopentane 65,66.
2-Methylbutane 65,66,88,89.
2-Methylpentane 68
3-Methylpentane 75
n-Hexane 75
Cyclohexane 68,75,82,88,89.
2,4,4-Trimethylpentane (isooctane) 75,79,84,88,89.
2,6-Dimethylheptane 94
III. ACETYLENIC
1-Butyne 65,66.
IV. AROMATICS
Benzene 65,66,71,79,81,88,89,93.
Toluene 52,65,66,71,75,80,81,82,83,84,88,89,92,93,98,100,147.
o-, m-, and £-Xylene 65,66,71,75,81,88,89,93,94,98,147.
Ethylbenzene 65,66,71,79,94.
3-19
-------
Table 3-5 (Cont.)
Compounds Investigated in Smog-Chamber Studies, with References
1,3,5-Trlmethylbenzene (mesitylene) 71,72,75,79,80,88,89,94.
Isopropylbenzene 71,81.
1,2,4-Trimethylbenzene 71
1,2,3-Dimethylbenzene 71
3-Ethyltoluene 71
1,2-Diethylbenzene 88,89.
tert-Butylbenzene 88,89.
n-Butylbenzene 88,89.
1,4-Diethylbenzene 71
1,2,3,5-Tetramethylbenzene 71
V. ALDEHYDES
Formaldehyde 75
Propionaldehyde 75
Hexanal 88,89.
Heptanal 147
Benzaldehyde 81
Glutaraldehyde 94
VI. KETONES
Mesityl oxide 83
Isophorone 83
Methylisobutylketone 83
Cyclohexanone 82
3-20
-------
Aerosol Formation from. Different Types of Hydrocarbons. Extensive
discussion of data represented in Table 3-5 is beyond the scope of this
review, and only the most important aspects of aerosol formation are
reported here and in the next few sections. Many of the conclusions
presented thereafter were reached in the early studies reported by the
Stanford Research Institute, Air Pollution Foundation, and Franklin
64-69,101-103
Institute groups. Results obtained in the presence of
sulfur dioxide are discussed toward the end of this chapter.
In the absence of sulfur dioxide, aerosol formation depends strongly
on the type of hydrocarbon precursor studied. The following qualitative
trends are observed:
• Most paraffins do not generate aerosol, even when irradiated at
high concentrations. However, some aerosol can be formed from the more
"reactive" branched paraffins having more than six carbon atoms (such as
isooctane) after long irradiation periods.
• Acetylenics do not form aerosol.
• All unsaturated compounds can form organic aerosol when reacting
70
with ozone at high concentrations, as observed by Cvetanovic. However,
studies conducted at much lower alkene concentration (1-10 ppm) show a
marked effect of alkene chain length on aerosol formation. Alkenes
with fewer than six carbon atoms do not form aerosol; those with six or
more carbon atoms form aerosol when they yield (after rupture of the
double bond) a fragment with at least five carbon atoms. For example,
1-heptene forms much more aerosol than 3-heptene, and 2,4,4-trimethyl-l-
pentene (isooctene) forms more aerosol than its isomer trans-4-octene.
3-21
-------
Amounts of aerosol formed from 1-alkenes increase regularly with the
number of carbon atoms (Figure 3-6).
• Cyclic olefins and diolefins form much more aerosol than 1-alkenes
that have the same number of carbon atoms (for example, cyclohexene »
1-hexene, and 1,7-octadierie » 1-octene) . The same effect of chain
length and double-bond position is observed for diolefins (1,7-octadiene >
1,6-heptadiene > 1,5-hexadiene, and 1,7-octadiene » 2,6-octadiene).
Heavier unsaturated cyclic compounds, such as indene and terpenes, form
even more aerosol.
• Conflicting results have been reported for aromatic compounds.
Aerosol formation has been reported from benzene, toluene, and other
88,92,93
alkylbenzenes by several investigators, whereas no aerosol
66,94,100
formation was observed in other studies. This merits further
investigation, in view of the large fraction of aromatic hydrocarbons
present in polluted atmospheres.
• Carbonyl compounds (ketones, C aldehydes, dialdehydes) do
not generate aerosol.
88,101-105
• Data on aerosol formation from irradiated automobile exhaust
confirm the marked sensitivity to hydrocarbon type observed in individual
103
hydrocarbon studies. Aerosol formation increases with the olefinic
105
and aromatic fuel content. However, changes in mode of engine operation
(acceleration, idle) and inorganic variables (sulfur dioxide, relative
humidity) have a more pronounced effect on aerosol formation than change
88
in fuel composition.
3-22
-------
0)
O
o
6J
Ul
ID
it
E
QL
Q.
CVJ
rO
0
UJ
Z
UJ
_i
>-
0.
o
or
Q.
H
•
• = I-BUTENE
o= I-HEXENE
UJ
z
UJ
X
UJ
X
o
o
o
II
<
| 0= I-OCTENE
3N303Q-I = A
• = I-DODECENE
<*-{ O
a) 4J
i-H t-l
O 3
c n
O 4-1
•H
4-> e
o o
C -H
3 CO
4-1 to
•H
td g
CO 0)
cd p*
C rC
O 4J
•H -H
4-1 15
O
cd -a
0) OJ
t-i
CD
-H
o a<
N a)
O Pi
0) M
^H o
O 4->
o
§ 3
K (-1
c o
O ^H
•H «4-l
4J
g a)
o cu
« 6
O -rl
M 4-1
(U
cd a>
o
4-( 0
o cu
C -H
o to
•rl Q)
4-> VJ
cd
cu
o
C •
O C
o o
•H
CO 4-1
to cd
i
CO
3.
•H
3-23
-------
Photochemical Reactivity and Aerosol Formation Ability. We have listed
in Table 3-6 the relative rates of conversion of nitric oxide to nitro-
106
gen dioxide measured by Glasson and Tuesday and the amounts of aerosol
66
formed, relative to cyclohexene, from data of Renzetti and Doyle,
68 79 88,89
Prager et^ al.., Groblicki and Nebel, Wilson et^ aJk, O'Brien
94 100
et^ al^, and Grosjean for different types of hydrocarbon precursors.
Table 3-7 summarizes the experimental conditions for data presented in
Table 3-6. As mentioned before, data reported by these investigators
are in good qualitative agreement for all precursor classes except
aromatics. It is clear from the data in Table 3-6 that there is no
straightforward relation between the amount of aerosol formed and the
gas-phase photochemical reactivity of the various precursors. The
overall hydrocarbon photochemical reactivity is a complex function of the
rate constants of reaction with various species (ozone, oxygen atom,
hydroxyl and hydroperoxy radicals, other free radicals) during the
photooxidation process. Therefore, relative rates of hydrocarbon decay,
of ozone (or oxidant) formation, or of conversion of nitric oxide to
nitrogen dioxide can be used as empirical photochemical reactivity (PR)
indexes. As we will see in a more quantitative manner later, the ability
of a given hydrocarbon to form aerosol depends on the nature of the
products formed (chemical dependence) and on their volatility (physical
dependence), and an aerosol formation ability (AFA) index can be derived
from physical and chemical aerosol data. Therefore, the amount of aero-
sol formed appears to be the product of two terms:
[aerosol]=(photochemical reactivity)x(aerosol formation ability).
3-24
-------
Table 3-6
Aerosol Formation and Gas-Phase Reactivity
Aerosol Formation Ability'
a
Without Sulfur Dioxide
Precursor bad e f
I. OLEFINS
Alkenes :
Ethylene 0 2.8 -
Propylene - - 12.4
1-Butene - 1.4 -
cis-2-Butene - } - - -
1.4
trans-2-Butene 0 J
Isobutene 0 0 -
2-Methyl-2-
butene 3.6 - -
3-Methyl-l-
butene 2.8
1-Pentene 0.9 2.8 -
cis- and trans-
2-Pentene - 0 - - -
1-Hexene - 1.4 - - -
cis- and trans-
2-Hexene 7 - - -
cis-3-Methyl-
2-pentene - 5.6 - - -
2,3-Dimethyl-
2-butene 2.7 -
1-Heptene - 10.5 =0 1
cis- and trans-
3-Heptene - 12.6 0
5-Methyl-l-
hexene - - - - 1
With Sulfur Dioxide
g b ad eg
12.6 63 44 1
69
- - 81-1
1
1 86
144 J -
- 40 - - -
- 96 - - 2.6 -
- - - - - -
- 96 87 2.7
- 86 - -
- - 96 - -
- - 86 - - -
- - - - - -
0 100 - - - -
- - - - 6.5 -
96 11.7
_ _ _ _ _
Gas-Phase
Reactivity
a,h
48.5
100
83
202
320
100
543
77
60
187
48.5
171 (trans)
10
43
134 (trans)
3-25
-------
Table 3-6 (Cont.)
Aerosol Formation and Gas-Phase Reactivity
Q
Aerosol Formation Ability
Precursor
1-Octene
trans-4-
Octene
Isooctene
Without
b a
-
-
- 94
Sulfur
d
-
-
28
Dioxide
e f
1
1
_ _
With Sulfur Dioxide
ff b a
_
_ _ _
94
d e q
88
_ _
81
Gas-Phase
Reactivity
42.8
94.5
—
Cyclic Olefins:
Cyclopentene - 75 58
Cyclohexene 100 100 100
a-Pinene - - 140
100 100 107 100 91 100
200 - - - 137
657
100
Diolefins:
1,3-Butadiene 3.6 33
Isoprene -
1,5-Hexadiene - 104
1,6-Heptadiene
2-Methyl-l,5-
hexadiene -
1,7-Octadiene
2,6-Octadiene
Dicyclopentadiene
- 124
45
178
122
200
1
150 91 85
- - 75
111
123
106
111
3-26
-------
Table 3-6 (Cont.)
Aerosol Formation and Gas-Phase Reactivity
Aerosol Formation Ability
Without Sulfur Dioxide
Precursor bad e f g
II. PARAFFINS
Methane 0 -
•Cyclopropane - 4.3 - - - -
n-Butane - - - - -
n-Pentane 0 - - - -
2,2-
Dimethylpropane 0 - - -
Cyclopentane 0 - - -
2-Methylbutane 0 - - -
2-Methylpentane - - - -
Cyclohexane - 0 - - - —
Isooctane - - 6.2 - - -
With Sulfur Dioxide
b ad eg
0 - 25 - -
_____
2.8 - .55
0 - - - -
0 - - - -
0 - - - -
0 .65
- 0 - - -
- 1.4
- 12.4 .25
Gas-Phase
Reactivity
2.86
-
21.2
26.5
10.0
26.2
28.0
26.8
27.6
21.1
2,6-
Dimethylheptane -
III. AROMATICS
Benzene
Toluene
o-Xylene
m-Xylene
jj- Xylene
Ethylbenzene
Mesitylene
0 - 10.5 4.5 10.8
0 - - 8.5 - 0 9 8.6
___ _ 9 _ _ __ __
- - - - - - 29.4
0-- ___ o -- --
0 - - - 1 - 12.6 - 69 -
- - - 9 1 - - - 50 9.0 -
9.4
37.2
74.5
106
60
34.3
146
3-27
-------
Table 3-6 (Cont.)
Aerosol Formation and Gas-Phase Reactivity
Aerosol Formation Ability
Without Sulfur Dioxide
Precursor
d
With Sulfur Dioxide
bode
1,2-
Diethylbenzene
tert-
Butylbenzene -
n-Butylbenzene -
23.4
24.4
22.2
Gas-Phase
Reactivity
48.5
16.8
IV. ALDEHYDES
Hexanal -
Heptanal
Glutaraldehyde -
0
0
V. ACETYLENIC
1-Butyne
0.9
aRelative to cyclohexene = 100.
^Data from Renzetti and Doyle, ref. 66
°T>3.ta. from Prager et al., ref. 68
"Data from Groblicki and Nebel, ref. 79
eData from Wilson ^ ^., ref. 88,89
/Data from O'Brien et_ al_., ref. 94
9"Data from Grosjean et^ al., ref. 100
"Data from Glasson and Tuesday, ref. 106; nitric oxide photooxidation rates.
3-28
-------
I
ro
42
cti
H
CO
01
•H
T3
3
4J
CO
t-l
0)
I
O
60
I
CO
CD
C
o
•H
4-J
•H
T3
C
O
o
Ol
01
CX
o
Co
a)1"1-
•I-)rH
CD cfl
0
r4 4-1
O 01
C
o
rH
UH
01
H
J-
c CD
0) .
•H rH
P Cfl
M
» 4-1
O Ol
CD
CD
CO
rH
O
en
CO
«
CO
CO
C •
0 rH
CO CO
rH
•H 4-1
tS 0)
CO
CD
tO
rH
O
CD
t--
•H rH
Ai Ol
0 42
•H O)
rH 3
o nd
^ C
O to
CD
CO
0)
rH
C
•H
CO
4-J
CO
CO
U}
>-l •
0) rH
60 CO
CO
rJ 4-1
PH 01
CD
CO
tfl
rH
O
T3
C
cfl
•H
4JU3
4-1 10
CU CU
N rH
C !*
0) O
PS 0
X
M
r*"l
PM
CO
E
o
CO
ro
4J
<4H
O
O
r.
i-H
Cfl
^
01
4-1
•H
rH
O
0
CN
ro
rH 4-1
0) 4-1
Ol
4-1 O
CO O
rO
CO
p
CU
4-1
•H
rH
O
s
en
|H
CU
4J
•H
rH
O
n
CO
1— 1
o
C/l
y
cj
to
i-H
M
1
co
Ol
^H
O
3
rH
"4-1
M
O
CO
rH
M
4-1
C
01
O
CD
01
o
3
rH
K^»
IH
3
o
)-l
:§
CD
CX
E
cfl
rH
^*»
IH
3
O
IH
01
S
4-1
G
0)
o
co
CU
t-l
o
3
rH
<4H
CO
CX
E
CO
rH
4-1
C
Ol
CJ
co
0)
42
3
4J
CD
CX
E
CO
rH
(-J
•H
tfl
4J
C
01
•H
42
1
M
•H
cfl
en c
ex cfl
E o)
cO rH
rH O
M
•rl
Cfl
C
CO
01
I— 1
C_>
H;-
IH
•H
-l
CO
01
CO
01
4-1
C
01
•H
42
1
CJ
CM
r\i
PH
CM
CO
fl_|
LO
C^
4_)
c
01
•H
42
1
4-1
C
0)
•H
42
•<
ro O O
E
In 43
co
01
4-1 -H
C 4J
CU CO
•H rH
42 CU
E 1-1
13 CO 4J
3 i-H -H
4J 0) T3
CD rJ -H
S
4-1 ^s 3
O O 43
01
•H
4-1 >.
Cfl 4-1
rH -H
Ol T3
rJ -H
^S 3
O 43
o
4-1
rH CO
E
CX
ex
o
CN O
01
•H
O
I
4-1 >,
Cfl 4-1
rH -H
ai T3
rJ lH
E
S^S 3
O 43
tH
CU
42
CJ
6£
0
E
CO
rH
to
•rl
M
CU
4J
CO
s
Ol
E
rH
0
>
C
o
•H
4-1
CO
•rl
"d
cfl
}H
^i
h- 1
CO
CO
60
4-1
C
CU
3
rH
•H
O
CO
01
•H
•H
i-l
3
ex
'H
c
o
42
(H
cfl
CJ
0
T3
P3
CJ
tfl
CU
60
C
rJ
•H
4-J
CO
3
4-J
CO
CU
CX
01
H
CO
c
0
•H
4-1
CO
M
4-1
C
Ol
CJ
C
O
CJ
rH
CO
•H
4-1
•H
C
M
E
ex
ex
M
C
o
42
tfl
0
O
^j
TJ
£ri
H
ex
t\
cu
•a
•H
x
o
o
^4
4-1
•rl
&
E
CX
ex
A
cu
13
•H
X
0
•H
T3
C
CU
60
O
4-1
•H
^
C
01
43
JS
01
T3
•H
X
O
•H
*d
^i
3
<4-J
i-H
3
CO
E
ex
ex
M
/— s
T3
OJ
T3
T3
cfl
J-l
CU
4-1
tfl
s
rQ
rQ
rC!
60
C
cfl
01
(H
Cfl
01
C
CU
O
rH
0)
CX
01
C
60
C
•H
4-1
Cfl
J-l
60
CU
4-1
-S
co
CD
cfl
E
o
4-J
cfl
C
o
o
CX
o
CU
co
CO
o)
CJ
r" °,
Ol -rl
01 O
E -<
o A
4J X
j2 CO
d, CD
cfl
cu E
O II
a
O)
CO
•rl 60
tO C
O M CJ
4-1 cu q
4J -rl
O) 4-1
> tfl 60
•HOC
4-1 CD 'H
CO Id
i—I T3 *
CU !H
C O -H
•rl «H C
M 3
01 X
4J -H rH
4-1 G
Cfl 0) •>
CJ O O
CO 43 -H
PH §
4J 43
43 (H 4J
60 -rl -H
•H cfl M
rH rH Cfl
CJ 60
o C o
O -rH rH
CTi CO ^—'
(3 rQ
3-29
-------
Bearing in mind that product volatility decreases regularly when the
size of the hydrocarbon precursor increases, this provides a rational
approach for a qualitative estimate of aerosol formed from any type of
precursor. Aerosol formation can be "PR-controlled" (paraffins, acetylenics;
PR = 0) or "AFA-controlled" (low-molecular-weight alkenes, carbonyl com-
pounds; high PR, but AFA =0). In the same way, aerosol formation from
1-alkenes having similar PR is controlled by their AFA factor, which
increases with chain length; whereas relative amounts of aerosol formed
from hydrocarbons leading to products of similar volatility (for example,
n-heptane, toluene, and 1-tieptene, with similar AFA factors) are controlled
by their relative gas-phase reactivities. Therefore, hydrocarbons pro-
viding the bulk of secondary organic aerosols in the atmosphere are not
necessarily those prevailing in oxidant formation. An important conse-
quence is that control of atmospheric organic aerosol concentrations can
be achieved by two types of control strategies: specific control of pre-
cursors with high AFA may prove to be as efficient as total hydrocarbon
and nitrogen oxides emission control.
Application of Smog-Chamber Data to the Atmosphere. Many difficulties
arise in the application of laboratory data to the more complex atmospheric
processes. This is illustrated by the following examples:
• Depending on meteorologic conditions, aerosol formation in the
atmosphere is better approximated, but never fully simulated, by smog-
chamber studies under either static (batch-reactor) or dynamic (flow-
reactor) conditions.
3-30
-------
» Small differences in light sources have definite effects on
photochemical processes. Some irradiation systems match the solar
spectrum poorly, thus affecting the relative importance of the various
107
photodissociation processes occurring in the atmosphere. Ambient
variations in the nitrogen dioxide photolysis pseudo-first-order rate
constant, K,, are difficult to reproduce with artificial irradiation
systems. "Transparent" smog-chamber materials, such as fluoroolefinic
polymers, fail to transmit the entire solar radiation: Mylar absorbs
108
strongly in the < 3300 A region, and Teflon shows a pronounced de-
100,109
crease in transmission, owing to aging. Different smog profiles
were obtained when identical experiments were conducted in aluminum,
107
stainless steel, Pyrex, and Teflon smog chambers. Furthermore, a
71,110
so-called inert material, such as Teflon, reacts with ozone.
80
• Mechanical stirring dramatically inhibits aerosol formation,
and data obtained in smog chambers equipped with stirring accessories
are questionable.
• Significant aerosol losses on the walls are observed in smog
chambers with high surface:volume (S :V) ratios. Typical smog chambers
-1
have S:V ratios of around 3:1 to 5:1 m . Although wall losses are
-1
minimized in bigger smog chambers, with S:V ratios of 0.1:1 to 1:1 m
range, this is still much higher than typical ambient S:V values. For
3
example, a typical ambient concentration of 150 yg/m corresponds to an
-4 -1
S'V ratio of about 9x10 :1 m (assuming a density of 1 and a diameter
of 1 ym). Formation of water layers on the smog-chamber walls would
promote heterogeneous reactions and thus affect the chemical composition
3-31
-------
of the aerosol formed. Nitric oxide and, to a lesser extent, nitrogen
dioxide are poorly soluble in water, whereas sulfur dioxide and nitrogen
pentoxide are readily soluble and their liquid-phase oxidation to form
sulfate and nitric acid, respectively, is quite significant.
• Errors may be due to the instrumentation itself. Improper cali-
bration of condensation nuclei counters may lead to poor nuclei concen-
111
tration data. In many experiments, the aerosol mass concentration was
deduced from the light-scattering measurements by using Charlson's
112
relation, which was established from data on ambient, well-aged aerosols.
105,169
Use of this relation is questionable for automobile exhaust data,
94,100
and it does not apply to fresh aerosols generated in smog chambers.
Interference due to ozone and to nitrogen-containing compounds has been
reported when nitrogen oxide concentrations have been measured by the
113 114
colorimetric and chemiluminescence techniques. Significant
differences have been observed in the measurement of ozone by different
115
techniques, and a standard procedure has been recently recommended.
• Use of clean air, although important for obtaining quantitative
gas-phase data, would favor homogeneous nucleation to the detriment of
the heterogeneous process, which is preponderant in the polluted atmos-
phere, thus providing aerosol kinetic and growth data of questionable
significance. Moreover, the background composition of individual hydro-
carbons in clean air is generally not specified, and only an upper limit
of "total hydrocarbon" concentration is provided. Therefore, a significant
fraction of the nuclei formed may well originate from a few C/-, alkenes
or cyclic olefins present as traces in the hydrocarbon background.
3-32
-------
• Because of the abundance of smog-chamber data on homogeneous
gas-phase photochemical reactions, the possible importance of hetero-
geneous reactions in the formation of organic aerosols has been somewhat
overlooked. However, reactions in the liquid phase and reactions catalyzed
by airborne particles containing metals, oxides, and salts—similar to
the reactions occurring in the well-known heterogeneous oxidation of
sulfur dioxide—may be of some importance. As demonstrated by Judeikis
116
and Siegel, heterogeneous reactions can compete with and, under
favorable conditions, outweigh the homogeneous gas-phase reactions that
are usually considered to be important. Free radicals (hydroxyl,
hydroperoxy, and so on) may be efficiently trapped by aerosol particles
117
and react further in the particle-water layer or at the particle sur-
118
face. These reactions have been incorporated in a recent mathematical
117
simulation study.
We have seen that aerosol data obtained in environmental chambers
may be seriously affected by many empirical characteristics, such as
smog-chamber design and materials, stirring, mode of irradiation, and
wall losses. Moreover, only homogeneous systems have been studied, and
the importance of heterogeneous organic reactions is not assessed.
Therefore, caution must be exercised when extrapolating smog-chamber
data to atmospheric processes.
Chemical Composition of Model Aerosols
As mentioned earlier, there is little information on the chemical
composition of "model" organic aerosols generated from a single hydro-
68 69
carbon in a smog chamber. Prager et^ al. and Endow £t al. reported
the presence of absorption bands due to carbonyl, carboxylic acid,
3-33
-------
and nitrate eater groups in the infrared spectra of aerosols produced
from various olefins. In their study of the gas-phase olefin-ozone
119
reaction at high concentrations, Vrbaski and Cvetanovic reported
the formation of aerosols that initially had a peroxidic character and
whose later decomposition produced carbonyl compounds identified by gas
71
chromatography. Kopzcynski reported absorption bands due to carbonyl,
hydroxyl, nitrate ester, and nitro groups in the infrared spectra of
74
aerosol generated from mesitylene at 25 ppm. Barlage and Alley studied,
by mass spectrometry, aerosols formed from 1-pentene and 1-hexene at 10
ppm. They found peaks corresponding to masses in the range 120-160 a.m.u.,
79
indicating the presence of polymeric material. Groblicki and Nebel
82
and Ripperton et al. reported similar infrared spectra for a-pinene
aerosol generated by either dark reaction with ozone or irradiation
with nitrogen oxides. Infrared spectra of a-pinene, 1-hexene, and
98 93
dodecene aerosols were recently reported by Lipeles. Chu and Orr
studied aerosols formed from benzene, toluene, and o-xylene and found
similarities between o-xylene aerosol and diethylphthalate (infrared
spectra) and between benzene aerosol and such aromatic ketones as
94
fluorenone (mass spectra). O'Brien et^ al^ reported carbonyl,
hydroxyl, carboxyl, and nitrate ester absorption bands in the infrared
spectra of 1-octene and 1,7-octadiene aerosols and found that infrared
spectra of indene aerosol resembles that of homophthalic acid, an
expected product of indene photooxidation. Paper chromatography indi-
cated the presence of several carboxylic acids in the 1-octene aerosol
and the presence of acidic difunctional compounds (a, w diacids, a, co
acid nitrate, and other w-substituted acids) in the 1,7-octadiene
3-34
-------
aerosol. Some ammonium nitrate was found, indicating that nitric acid
was formed photochemically. The molecular weight of 1,7-octadiene
256
aerosol was determined, and its value indicates that the difunctional
compounds are present as dimers, or possibly higher polymers, in the
aerosol.
This survey shows that highly oxygenated compounds (carbonyls,
acids, nitrate esters) constitute the bulk of model organic aerosols,
and that smog-chamber data on chemical composition are in qualitative
agreement with atmospheric observations. However, positive identifica-
tion of individual organic compounds present in model aerosols cannot be
achieved from infrared, carbon-hydrogen-nitrogen, and paper-chromatography
analysis or from limited mass-spectrometry data. Only since 1973 have
87
firm product identifications been reported. Lipeles et^ al^. and
99
Schulten and Schurath studied the composition of aerosol from the
reaction of ozone with 1-butene. Aerosol composition from irradiation
of nitrogen oxides with cyclohexene, toluene, and a-pinene was reported
92
by Schwartz, and with cyclopentene, cyclohexene, and 1,7-octadiene,
100
by Grosjean. Mass spectrometry was used in these four studies.
87
Ozone-1-Butene Aerosol. The Rockwell study was conducted in a flow
reactor with a dry nitrogen stream and typical residence times of
20-60 s. Freshly collected aerosol showed a marked peroxidic character,
but would no longer oxidize iodide after several days. Polarographic
analysis confirmed the presence of peroxides. Aerosol composition was
obtained by a combination of gas chromatography and electron-impact
99
mass spectrometry. Schulten and Schurath also used a flow-reactor
system coupled with a high-resolution-field desorption mass spectrometer.
3-35
-------
Hydrogen peroxide was positively identified. Other peroxides may have
been present, but would decompose at the low (- 10 torr) pressure in
the ion source of the mass spectrometer. The complete data are listed
in Table 3-8. Aldehydes, ketenes, and C , monocarboxylic acids were
1-4
identified. Schulten and Schurath observed many unidentified complex
ions in the higher mass region and postulated the presence of aggregates
of aldehydes and acids held together by strong hydrogen bonds. This
assumption is consistent with the release of carbonyl compounds observed
by Cvetanovic when aerosols generated from alkenes decompose during gas-
chromatography analysis.
92
NO -Toluene Aerosol. This system was studied by Schwartz. Toluene at
X L ^
10 ppm, nitric oxide at 1 ppm, and nitrogen dioxide at 1.2 ppm were
3
irradiated with ultraviolet lamps in a 17-m batch reactor for 270 min.
Collected aerosols were successively extracted with methylene chloride
and then methanol. The methylene chloride extract was fractionated into
water-soluble and water-insoluble material, and the latter fraction was
further divided into acidic, neutral, and basic fractions. The acidic
and neutral fractions were analyzed by gas chromatography and chemical-
ionization mass spectrometry; the compounds identified are shown in
Figure 3-7. The two analyzed fractions represented only about 5.5% of
the total aerosol mass. It is noteworthy that "classical" nitration of
an aromatic ring appears to be an important process during aerosol
formation and that both the alkyl group and the aromatic ring undergo
attack by oxidizing species, thus leading to polyfunctional compounds of
very low volatility.
3-36
-------
Table 3-8
Composition of Aerosol from the Reaction of Ozone and 1-Butene
Identified by Lipeles et al.87
Carbon dioxide
Ethane
Water
Formaldehyde
Methanol
Acetaldehyde
Formic acid
Propionaldehyde
Acetic acid
Propionic acid
Other unidentified trace
Compounds
Identified by Schulten and Schurath99
Formaldehyde
Hydrogen peroxide
Ketene
Acetaldehyde
Formic acid
Methylketene
Propionaldehyde
Acetic acid or methylformate
Ethylketene
Alcohol (tentative)
Butyraldehyde or ether, C H 0
4 8
Propionic acid
Formic acid-formaldehyde cluster, C H 0
243
Butanoic acid
Butanoic acid - water
C H 0 : dimer of formic acid or dimer
244 zwitterion
C H 0 : ozonide
483
+ Unidentified peaks from m/e=105 to
m/e=199 (C H 0 with x=3 - 8, y=5 - 13,
x y z
z=2 - 6), presumably from condensates or
aggregates of aldehydes and acids held
together by strong hydrogen bonds
3-37
-------
CH,
COOH
- CHO
COOH
TENTATIVE
TOLUENE
CH,
s
CHO
J:HO
Figure 3-7. Composition of aerosol from the reaction of NO and toluence
(according to Schwartz, ref. 92).
3-38
-------
Aerosols from Cyclic Olefins and Diolefins. Data on cyclohexene and
92
a-pinene aerosols were reported by Schwartz after a preliminary report
52
from the Battelle Institute group. The experimental conditions and
analytic techniques were identical with those just described for the
toluene aerosol study. Here again, only the methylene chloride-soluble,
water-insoluble fractions were studied. They accounted for about 7%
and 65% of the total aerosol mass generated from cyclohexene and a-pinene,
100
respectively. Grosjean reported chemical composition of cyclopentene,
cyclohexene, and 1,7-octadiene aerosols. Experiments were conducted in
3
an 80-m Teflon smog chamber filled with ambient air, with irradiation
by sunlight. Typical initial concentrations were 1 ppm for hydrocarbon,
0.33 ppm for nitric oxide, and 0.16 ppm for nitrogen dioxide. Aerosols
were extracted after collection and analyzed without further fractionation
by combined gas chromatography and electron-impact mass spectrometry, and
by combined gas chromatography and chemical-ionization mass spectrometry.
Data obtained by Schwartz and Grosjean are listed in Figure 3-8 and
Table 3-9. Most of the products are difunctional compounds bearing in
many cases a carboxylic acid group. The ethylenic bond is retained in
some compounds, indicating free-radical attack on the aliphatic chain.
Good agreement is observed in the case of cyclohexene (the only compound
common to the two studies), except for adipic acid, which was not reported
by Schwartz, but was the major aerosol compound in Grosjean's study.
However, adipic acid would not be expected to be present in the methylene
chloride-soluble, water-insoluble fraction analyzed in the former study.
Data for the 1,7-octadiene aerosol are also in good agreement with those
94
obtained by O'Brien by infrared and paper chromatography.
3-39
-------
CM,
X^bc
TENTATIVE
a-PINENE
Figure 3-8. Composition of aerosol from reaction of NQX and a-pinene
(according to Schwartz, ref. 92).
3-40
-------
Table 3-9
Composition of Aerosol from the Reaction of NO and Various Olefins
x
From Cyclohexene Aerosol
Identified by Schwartz (ref. 92) Identified by Grosjean (ref.100)
COOH-CCH ) -CHO COOH-CCH ) -COOH major
24 24
COOH-(CH ) -CHO COOH-CCH ) -CH ONO
23 2422
COOH-CCH ) -CH ONO COOH-CCH ) -CHO
2322 24
COOH-(CH ) -CH OH COOH-(CH ) -CH OH
232 242
CH -CH=CH-C-C-CH COOH-(CH ) -COOH
3 || || 23
0 0
COOH-(CH ) -CH ONO
23 22
CH =CH-CH=CH-CH -CH OH COOH-(CH ) -CHO
2 22 23
cyclopentene-2-aldehyde or COOH-CCH ) -CH OH
23 2
CHO-CH=CH-CH -CH=CH CHO-(CH ) -CHO
22 23
From Cyclopentene Aerosol Cldentified by Grossjean, ref. 100)
CHO-CCH ) -CHO COOH-CCH ) -CH OH
23 222
CHO-CCH ) -COOH COOH-CCH ) -cnoa
23 22
H ) -COOH CHO-CCH ) -CHQa
23 22
H ) -CH ONO CHO-CCH ) -CH oNOa
2322 2222
3-41
-------
Table 3-9 (Cont.)
Composition of Aerosol from the Reaction of NO and Various Olefins
x
From 1,7-Octadiene Aerosol (Identified by Grosjean, ref. 100)
COOH-CCH ) -COOH CH =CH-(CH ) -CH ONO a
24 22422
COOH-CCH ) -CH_ONO. CH =CH-(CH.) -CHO
2 4 z ^ 2 ^4
COOH-(CH ) -CHO CH =CH-(CH ) -
24 2^
Tentative.
3-42
-------
It is very significant that most of these polyfunctional compounds
have also been identified in ambient aerosols (Tables 3-2 and 3-3) and
that the gas-to-particle distribution factor, f , measured in smog-chamber
C 100
studies for aerosol precursors, such as cyclic olefins and diolefins,
51
exceeds by one order of magnitude those measured in ambient atmospheres.
Chemical Mechanisms of Organic Aerosol Formation
The Ozone-Olefin Reaction. Much evidence has been accumulated that the
ozone-olefin reaction has a predominant role in aerosol formation from
alkenes, cyclic olefins, diolefins, and other unsaturated compounds. Free
radicals are formed in the reaction and can react further, along with
nitric oxide and nitrogen dioxide, either with the various intermediates
120
or with the olefin itself (see the recent review by Pitts and Finlayson ).
121
A mechanism has been proposed recently by O'Neal and Blumstein for
the gas-phase ozone-olefin reaction. This mechanism postulates that molozonide-
biradical equilibrium is reached fast and postulates a competition between
a-, 3-, and y-hydrogen abstraction reactions and the classical mechanism
122
proposed by Criegee for the liquid-phase reaction. The main features
of the Criegee mechanism (Figure 3-9) are the formation, from the initial
molozonide, of the major carbonyl products and a second biradical inter-
mediate, the "zwitterion." The decomposition pathways of the zwitterion
comprise unimolecular rearrangements and bimolecular reactions, the latter
including ozonide formation from zwitterion-aldehyde reaction. Other
aspects of the Criegee mechanism have been discussed and reviewed else-
123-127
where.
The Criegee mechanism, widely accepted for the liquid-phase reaction,
does not adequately explain the available gas-phase data. O'Neal and
3-43
-------
LIQUID PHASE OZONE OLEFIN REACTION: THE CRIEGEE MECHANISM
INITIAL REACTION:
H R2 H R2
Os+ C=C -- *• ^C - C MOLOZONIDE (M)
Rl R3 Rl 0-^ ^-0 R3
^•0""^
MOLOZONIDE SPLITS:
R /™\ X^^^R
M — *• ' Q0^° 3 ZWITTERION + CARBONYL
:• ,C Cf CROSS DIPEROXIDE
/ \ / \
R3 0—0 ^R3
Figure 3-9. Liquid-phase ozone-olefin reaction: the Criegee mechanism.
3-44
-------
121
Blumstein suggested a biradical structure for the first gas-phase
intermediate and proposed three types of unimolecular hydrogen abstrac-
tion reactions (Figure 3-10).
The ex-hydrogen abstraction leads to an ct-ketoperoxide, which has
128,129
been tentatively identified by Pitts in the case of cis-2-butene.
Further reactions of the ketoperoxide include the formation of the "normal
products," i.e., the carbonyl products that can also be explained by the
Criegee mechanism. The 3Hiydrogen abstraction accounts for the observed
a-cleavage products. The y-hydrogen abstraction (possible for C alkenes)
4+
should lead to a rather stable five-membered ring, dioxetane, whose
existence has so far not been demonstrated experimentally. On the basis
of calculations of the relative importance of the Criegee and abstraction
pathways for both spontaneous and thermal reactions of the intermediate
species, O'Neal and Blumstein conclude that the Criegee and abstraction
mechanisms are of equal importance for butenes, that the former is pre-
dominant for ethylene and propylene, and that the latter is predominant
for C alkenes.
5+
All aerosol products identified in the smog chamber can be reasonably
explained in terms of the O'Neal and Blumstein and Criegee mechanisms,
as is illustrated in Figure 3-11 for cyclohexene. The major difference
between alkenes and cyclic olefins lies in the fact that, after opening
of the cyclic olefin double bond, the original number of carbon atoms is
conserved and the chain carries both the carbonyl group and the biradical
intermediate, whose further reactions lead to the observed difunctional
compounds.
3-45
-------
GAS PHASE OZONE OLEFIN REACTION:
THE O'NEAL AND BLUMSTEIN MECHANISM
MOLOZONIDE-BIRADICAL EQUILIBRIUM
RIN /R2
H I CH2-CH2-Rj
0 »0
(B2) 0*
I .
HYDROGEN ABSTRACTION
REACTIONS
-^
|XCH2CHaR3 H \
o 9
(B,)
R! * f R2 t
NC -»- C^ HYDROGEN ABSTRACTION
0 0NCH2CH2R3 REACTIONS
0^
H
CRIEGEE MECHANISM
BIRADICAL HYDROGEN ABSTRACTION REACTIONS
(ILLUSTRATED FOR B,)
a-HYDROGEN ABSTRACTION: /^ Q KETO PEROX|DE
0 OOH
a KETO PEROXIDE
/3- HYDROGEN ABSTRACTION:
2
,
I ~C — C*~ wH~
iii
OH 0-0
y-HYDROGEN ABSTRACTION:
Bi —*RrC— C ^O DIOXETANE
I I I
OH CH2—CH
R20H+RrC-C-CH2-CH2-R3
6 0
R,C +R2-C-CH2CH2R3
SOH 0
"NORMAL"PRODUCTS
0 0
a KETO ALCOHOL
Figure 3-10. Gas-phase ozone-olefin reaction: the O'Neal and Blumstein
mechanism.
3-46
-------
POSSIBLE FORMATION PATHWAYS FOR CYCLOHEXENE AEROSOL PRODUCTS
•>o
/QH ABSTRACTION?
s ^ x"
a Ox f^^rO •
y> -—*• i \ ,° ^H
0 k>° x
ABSTRACTION?
laH ABSTRACTION
*
CRIEGEE^ H
r^c^ rr? rr>ooH
^ts- /^r\ Lk>°
/ I V +°2X/ \
^ ^°2 / \ /^\
r^CHO / ,^\CHO r^-CHO
L ^CHO / •
^^ y l^^COOH ^
IDISMUTAT.ON ^ ^ ^^
S^COOH f^^CHO
COOH I I .
L. COOH \00
0 J + NO,H62>...
|N02,HN02
|+02
+ 02
+N02,
H ABSTRACTION
AEROSOL PRODUCTS
IDENTIFIED IN SMOG
CHAMBER
r^^cHO
H ABSTRACTION
CCHO
~. J2OH
Figure 3-11.
Possible formation pathways for cyclohexene aerosol products
Cafter Grosjean and Friedlander.156)
3-47
-------
The Hydroxyl Radical-Aromatic Hydrocarbon. Reaction. Very little is
known about the mechanisms governing ambient reactions of aromatic
92
hydrocarbons. The Battelle study demonstrates that aromatic-ring
opening, aromatic-ring nitration, and alkyl-group oxidation may occur
at atmospheric concentrations. The former possibility is supported by
130
Altschuller et al., who observed a decrease of infrared absorption
bands due to the aromatic ring during the photooxidation of various aro-
matics at 5 ppm with nitric oxide at 3 ppm. Reactions of aromatic hydro-
carbons with ozone, atomic oxygen, and hydroxyl radical might account
for the observed polyfunctional aerosol products. The reaction of ozone
123
with aromatics has been studied in the liquid phase by Wibaut and co-workers.
They reported relative rate constants of 1.9, 10, 40, 250, and 15,000 for
benzene, toluene, xylenes, mesitylene, and hexamethylbenzene, respectively.
Glyoxal, methylglyoxal, and biacetyl were the major products of the ozone-o-
xylene reactions, and their formation was interpreted in terms of the
Criegee mechanism. The gas-phase reaction has received little attention,
and the rate constants of only three aromatics have been measured. It
appears that the reaction is very slow, compared with that of olefinic
131
compounds. Bufalini estimated the rate constant k for mesitylene,
U3
one of the most reactive aromatics, to be only about 0.37 the rate constant
for ozone reactions with the least reactive olefin, ethylene. Recent data
132
from Stedman show that toluene and xylenes react even more slowly with
ozone than acetylene does.
3
The reaction of atomic oxygen, 0( P), has been investigated in
93,120 133,134
greater detail. Cvetanovic and co-workers reported formation
3-48
-------
of nonvolatile polymeric material from benzene and toluene. Formation
of the observed linear products can be explained by cleavage of the
biradical initially formed by addition of atomic oxygen on the aromatic
ring. Absolute rate constants for atomic oxygen addition to a series of
135
aromatics were recently measured in the temperature range 299-392 K.
The aromatic-hydroxyl radical reaction has been studied by Davis
136
et al. They reported rate constants for benzene and toluene and con-
cluded that hydroxyl additions to the aromatic ring compete favorably with
137
the abstraction of hydrogen atom from the alkyl substituent. Doyle et al.
recently published hydroxyl reaction rate constants for a series of alkyl-
benzenes.
Possible ozone, atomic oxygen, and hydroxyl radical reaction mecha-
nisms, leading to linear polyfunctional products from aromatic hydrocarbons,
are shown in Figure 3-12. In an attempt to assess the relative importance
of the various oxidizing species in the photooxidation of olefinic and
aromatic hydrocarbons, we have compared the rate constants for hydroxyl
radical, atomic oxygen, and ozone reactions with the rate constants for
the conversion of nitric oxide to nitrogen dioxide (Figure 3-13). Good
relations are obtained between k~ and kjgQ _,. JTQ f°r alkenes and between
3 2
k and k™ ^ for aromatics, indicating that the structural effects
prevailing in the overall conversion of nitric oxide to nitrogen dioxide
are those controlling the ozone-olefin and hydroxyl radical-aromatic
reactions. This is further substantiated by comparing the absolute
reactivities of olefins and aromatics. The most reactive aromatics,
m-xylene and mesitylene, react faster than ethylene and as fast as pro-
pylene and isobutene with the hydroxyl radical. The same reactivity
3-49
-------
OH
ALKYL CHAIN OXIDATION CH2OH (CHO, COOH)
CH,
°'+ [O
CH3
^v
O
ADDITION
R
^^^ ^^
N02+ O
CH3
C-0-6
CHO
CH3
CH2
O
DIFUNCTIONAL
COMPOUNDS,
"FURTHER
0X1 DATIVE
DEGRADAHON
CH2 OH (CHO, COOH)
CH3 CH,
,H
= CH-CH = CH-CH =
OH
0
CHO
Figure 3-12. Possible initial steps for ozone, atomic oxygen, nitrogen dioxide, and
dioxide and nydroxyl radical reaction with aromatic hydrocarbons.
3-50
-------
10
Figure 3-13a.
Relations between conversion of nitric oxide to nitrogen dioxide
and ozone, atomic oxygen, and hydroxyl radical reaction rate con-
stants for (a) olefinic and (b) aromatic hydrocarbons. (Reprinted
with permission from Grosjean.l57b) (k „_ from Glasson and
Tuesday.106) NO - N02
Alkenes (k^ from Japar et al.;160 from Morris and Niki;243 k from
°3 °
Furayama et a_l. 2 ** **)
3-51
-------
15 -
I 10
•O
o>
o
05
00
AROMATICS
-0.5
0.0
7b
0.5
20
to
O
10
o>
o
00
-10
Figure 3-13b. Relations between conversion of nitric oxide to nitrogen dioxide, atomic
oxygen, and hydroxyl radical reaction rate constants for (a) olefinic and
(b) aromatic hydrocarbons. (Reprinted with permission from Grosjean.15^)
Aromatics (kQ from Atkinson and Pitts;135 Doyle &t_ £LU kQ liquid-phase,
relative to toluene, from Bailey;123a,123b gas-phase ozone rate constants
have not been measured) .
3-52
-------
sequence is observed for the rates of conversion of nitric oxide to
nitrogen dioxide, whereas ozone-aromatic reaction rate constants are
too low to account for their overall photochemical reactivity. Relevant
to the postulated importance of the hydroxyl radical-aromatic reaction is
89
the observation made by Wilson et al.: an aerosol increase from 1-heptene
and an aerosol decrease from toluene when ozone is added to irradiated
NO -hydrocarbon mixtures. In the latter case, the observed inhibiting
effect results from the reaction,
0 + OH -> H02+ 0 , (1)
which competes with the hydroxyl radical-aromatic reaction. Many investi-
gators also reported that aerosol formation from olefins coincides with
ozone appearance, whereas aerosol buildup from aromatics starts before
ozone formation and is presumably associated with the appreciable hydroxyl
radical concentrations in the early stages of the photochemical process.
On the basis of all available evidence, the hydroxyl radical plays
a major role in the photooxidation and aerosol formation processes for
aromatic hydrocarbons. However, much research remains to be done to
improve our knowledge in this field.
Physical Mechanisms of Organic Aerosol Formation
Dynamics of Gas-to-Particle Conversion. The physical processes involved
in the formation of atmospheric aerosol have been thoroughly investigated
55-57,138-146
in the past few years and will be briefly summarized here,
with some emphasis on recent data obtained in smog-chamber studies.
3-53
-------
Regardless of the chemistry, there are some physical constraints
on aerosol-gas interactions. Particles must be close to or at equilibrium
with respect to the surrounding vapor to exist in air for any substantial
period. Thus, the partial pressure of condensed species on particles
must be less than or equal to the saturation vapor pressure at atmos-
pheric temperature for stability. As we will see later, the requirement
of low vapor pressure is particularly important to the stability of
organic aerosols.
Accumulation of condensed material as aerosols in the atmosphere
may take place by two basic processes: by condensation of supersaturated
vapor or chemical reaction that leads to spontaneous formation of new
particles, and by condensation, absorption, or reaction on existing parti-
cles. In the latter case, the chemical reactions may actually take place
on the surface of or within existing particles.
For condensable precursors, particle formation may occur by homo-
geneous or heterogeneous nucleation. It is generally accepted that
heterogeneous processes are most likely in the atmosphere, because of the
large number of nuclei present.
Growth of particles by accumulation on existing particles can be
classed as two broad processes. If the precursor is supersaturated,
growth will occur at a rate limited by vapor diffusion, which depends on
the supersaturation, the temperature, the particle size, and the accommodation
coefficient at the surface. The proportionality of particle size changes
with the ratio of particle diameter to mean free path of the suspending
gas. At one extreme, the growth depends on volume to the 2/3 power; at
the other, growth is proportional to volume to the 1/3 power. When the
3-54
-------
precursor is unsaturated, growth still may take place by irreversible
absorption or by chemical reactions in the particle. In this case, the
rate law should be proportional to the particle volume, if the reaction
is uniform throughout the particle. If the formation of material is
limited by reactions in the particle, the conversion ratio should not
depend on the concentration of the gaseous precursor.
There is insufficient information available to determine the rate
law or physical mechanism most likely to predominate in atmospheric aero-
sol growth. However, there are clues to differences in the processes
from the Los Angeles data. The shape of the particle volume number dis-
tribution of tropospheric aerosol is such that the 1/3 (diameter) and
2/3 (surface) moments are concentrated in the submicrometer fraction,
whereas the first moment (volume) is weighted toward larger particles.
Thus, the observed accumulation of organic carbon on the small particles
in smog (Figure 3-14) suggest a process controlled by surface or vapor
diffusion.
It is of interest that the influence of thermodynamic equilibrium
must enter the growth process of particles. If the radius of the parti-
cles is too small, the partial pressure of the condensable species can
increase significantly by the influence of radius of curvature (Kelvin
effect). Examination of values of surface tension for a range of materials
suggests that the Kelvin effect will constrain growth to particles greater
than about 0.05-0.1 ym in diameter. This appears to be consistent with
available observations of atmospheric growth and the distribution of
secondary chemical components.
3-55
-------
CSI
I
1
J-
I
ro
co
0)
(-1
00
&
0)
CO
O
o
O
CO
O
(-1
01
cO
60
I
CO
mi- UID/ w°
e t
O 0)|
c >.
O T3
•H -H
4J K
rQ B
•H O
H M
4-1 >4-l
CO
•H C
•O O
•H
QJ CO
6 CO
3 -H
i— I E
o n
> cu
ex
Q)
J3 J2
4-1 4->
•H
C
O
•H
^H P,
O CU
> Prf
~tf
i—I
I
3
bO
3-56
-------
Growth of Secondary Organic Particles. The physical mechanisms governing
the formation of organic aerosols in smog-chamber experiments have recently
147,148
been studied in some detail. Two types of profiles, presented in
Figure 3-15, are generally observed. Aromatic hydrocarbons and alkenes
with fewer than seven carbon atoms, when present at initial concentrations
of about 1 ppm, produce copious quantities of nuclei that are not able to
grow in the light-scattering range. On the contrary, aerosols formed from
C alkenes, cyclic olefins, diolefins, and terpenes always grow in the
light-scattering range and produce appreciable visibility reduction
(Figure 3-16; compare with ambient data presented in Figure 3-14). As
a result of aerosol growth, the light-scattering efficiency per aerosol
mass unit increases rapidly in the early stages of the experiment (Figure
3-17; compare with the light-scattering efficiency as a function of
particle size for monodisperse aerosol presented in Figure 3-18).
In the case of organic constituent formation, growth of particles
is governed by physical laws of condensation, provided that the precursors
are formed in the gas phase. For a diffusion-limited condensation process,
the rate of volume change in particles is
(L + a, 2A_I
dpj
dv = A (dp - dp*) (1 + g, 2A_| , (2)
dt
where A=ZA;A = A x S InS; for the th condensable species, X
iiiiiii i i
is the particulate mole fraction and x is the activity coefficient; dp
i
is the particle diameter; dp* is the initial particle size below which
condensation cannot take place because of the curvature effect on equilibrium
vapor pressure (Kelvin effect); S is the supersaturation ratio for the
th ±
1 species; X is the mean free path in air; and H is a parameter
3-57
-------
CO x->
GO
NOI1VSN3QNOO
o
o
o
o
ro
o
o
CO
o
o
/ /
/ /
V
M
X
o
/\
X
o
in
o
*Jo
i
O
o
1 1 1
m
d
o
CO
d
m
d
o
ro
0
1
i
i i
m c
d
o
CVJ
o
ul
( -- )
'3NOZO
n) a
J-i CO
4J 0)
ti •!-)
a) en
o O
ti n
o o
o
^ §
ca s-i
•H M-(
4J
•H ti
c o
H -H
CO
CO
• -H
0> 0)
>H ft
CU -H
4-1 01
0) 4-1
B C
cd -H
o> a)
4J p^
4-1
GO B
•H ft
c o
0)
X •>
0) CN1
^3 O
O S
iH
O •"
>^ e
o p,
4-j n
m en
OJ •
to 2;
0)
,-i .^
•H 6
"4-1 P<
o ex
M
ft "-H
6 ^
cfl t-t
jl nJ
o o
I O
txO S-i
I
ro
(JO
•rt
3-58
-------
1500
1000-
f
500-
Figure 3-16.
Evolution of the volume distribution of secondary
organic aerosol generated in smog chamber with 1-ppm
cyclohexene, 0.33-ppm NO, and 0.17-ppm NO . Time from
bottom to top: 0, 203, 412, 631, 863, 1109, 1364, and
1626 s. Compare with Figures 3-22 and 3-26. ,
(Reprinted with permission from Heisler et al. )
3-59
-------
, to
o co
B -H
z .2 6
O o
-------
ro
e
o
o
0.08
E 0.06
-------
proportional to the ratio dp/A. This linear relation has been verified
experimentally in the case: of cyclic olefin and diolefin aerosols
(Figure 3-19). Critical sizes of 0.13-0.24 ym and 0.26-0.28 pm have
148
been measured for cyclohexene and 1,7-octadiene, respectively. For
such aerosol precursors, the gas-to-particle conversion process consists
of the formation of supersaturated compounds in the gas phase followed by
condensation on preexisting particles.
Condensable-Species Vapor Pressure and Aerosol Formation. As discussed
earlier, the equilibrium constraint of low vapor pressure is particularly
important for stability of the organic aerosols. Organic aerosol formation
requires accumulation of condensable species in excess to their gas-phase
149-155
saturation concentrations,, In turn, examination of vapor-pressure data
for various oxygenated compounds permits estimating the minimal hydrocarbon
precursor concentration required to achieve aerosol formation in ambient
atmosphere. Table 3-10 gives examples of vapor pressures of various oxy-
genated compounds formed in hydrocarbon photooxidation reactions. It can
be seen from the table and from boiling-point data that volatility decreases
regularly when the number of carbon atoms increases and that, for a given
chain length, carboxylic acids have the lowest vapor pressure. Moreover,
the volatility of difunctional oxygenates is several orders of magnitude
156
lower than that of the corresponding monofunctional compound. Grosjean
estimated the minimal ambient alkene concentration required to form the
corresponding carboxylic acid in excess to its saturation concentration
(Table 3-11). Because a 100% gas-to-aerosol conversion was assumed and
because carboxylic acids actually represent only a fraction of the products,
the precursor concentration data in Table 3-11 are, in fact, lowest
3-62
-------
4.0
3.0
2.0
.1.0
0.0
CRITICAL SIZE d£ =0.27yL6m
0.2 0.3 0.4 0.5
0.6 0.7
dp,
0.8 0.9
1.0
Figure 3-19. Growth 'of organic aerosol generated in
smog chamber; same experiment as in
Figure 3-24. (Reprinted with permission
148
from Heisler. )
3-63
-------
Table 3-10
Vapor Pressures of Oxygenated Compounds
b
Vapor Pressure, mm Hg at 25 C
a
Carbon
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Q
Alkanes
-
-
-
-
-
-
-
15.1(27 C)
-
2.28(33.6 C)
0.48
0.117
-
-
- 2 x 10~3
10~3(21.8 C)
5 x 10~4(25.7
10~4(24.8 C)
Q
Aldehydes Alcohols
100(21.2 C)
795 31.5
200 16
5
1.8
1(24.4 C)
6 0.26
0.11
7.9 x 10~2
1.4 x 10~2 4.3 x 10~2
7.2 x 10~2 3.8 x 10~2
io-2
-
-
— —
-
C) -
1.45 x 10~4
Carboxylic
Acidsc
40(24 C)
16
4
1
0.25
0.02
-
4.4 x 10~3
-
10~3(33.6 C)
-
10~5(25.3 C)
8.0 x 10~5
= io-6
—
* 10~
-
,i aldehyde > alcohol - nitrate ester > carboxylic acid - dialdehyde
> diol - dinitrate - alcohol C , - acid aldehyde » dicarboxylic acid.
3-64
-------
Table 3-11
Hydrocarbon Threshold Concentration
(estimated lowest ambient olefin concentration required to form the
corresponding condensable species in excess to its saturation concentration)
a
Olefinic Precursor
Ethylene
Propylene
1-Butene
1-Pentene
1-Hexene
1-Heptene
1-Octene
1-Nonene
1-Decene
1-Tridecene
Cyclopentene
Cyclohexene
Methylcyclohexene
Condensable
Species
Formic acid
Acetic acid
Propionic acid
Butanoic acid
Pentanoic acid
Hexanoic acid
Heptanoic acid
Octanoic acid
Nonanoic acid
Dodecanoic acid
Glutaric acid
Adipic acid
Methyladipic acid
uonaensao-Le apecj.<
Vapor Pressure,
mm Hg
40
16
4
1
0.25
0.02
= 9 x 10~3
- 4 x 10~3
= 6 x 10~4
io-5
2 x 10~7
6 x 10~8
= 2 x 10~8
as
Minimal Precursor
Concent rat ±onb
52,400 ppm
20,960 ppm
5,240 ppm
1,310 ppm
327 ppm
26.2 ppm
11.8 ppm
5.2 ppm
0.78 ppm
13 ppb
= 0.26 ppb
0.08 ppb
= 0.03 ppb
aData from Grosjean.156
^Assuming that there is 100% gas-to-particle conversion, that the formed
condensable species is the one with the lowest vapor pressure (carboxylic
acid), and that there is no vapor-pressure lowering by condensable species
polymerization or other effect.
3-65
-------
estimates. Vapor-pressure considerations suggest, again, cyclic olefins
and diolefins as the most efficient aerosol precursors. For example,
cyclohexene at only about 0.1 ppb is required to form adipic acid (Vp * 6x
—8
10 mm Hg) in excess to its saturation concentration, whereas 1-heptene
at about 50 ppm (5x10 more than cyclohexene) would be necessary to form
-2
hexanoic acid (Vp - 2x10 mm Hg) aerosol. Obviously, aerosol formation
from C alkenes would require unrealistic ambient precursor concentrations,
even if polymerization of the condensable species lowers the required pre-
cursor concentration by several orders of magnitude. Aggregation or poly-
merization (at least dimerization) of the condensable species formed from
C,, alkenes must take place to achieve aerosol formation. On the contrary,
no polymerization (not even dimerization, although it may occur) is necessary
to form aerosols from cyclic olefins, even when present as traces (1 ppb or
below) in ambient air. Therefore, there is a threshold concentration,
below which no organic aerosol is formed for each hydrocarbon precursor.
The concept of precursor threshold concentration is supported by qualitative
98
experimental observations on nuclei formation (Figure 3-20) and occurrence
94
of light scattering (Figure 3-21) as a function of the precursor concen-
tration. The low-volatility constraint for organic condensable species
has major consequences for the ability of the various precursor classes to
form organic aerosol in the atmosphere:
• Cyclic olefins and diolefins form aerosol even when present at very
82,100
low concentrations, as confirmed by smog-chamber studies for cyclohexene
94
and 1,6-heptadiene.
• Alkenes of C do not form organic aerosol in parts-per-million
jL-v
concentrations. Aerosol data obtained at higher concentrations, although
3-66
-------
u
o
1.
«U
0.
a
z
o
A 1-DOOECENE
O 1,3-HEXAOIENE
D 1-HEXENE
O m-XYLEHE •
HYDROCARBON COWCENTRATION ppm
Figure 3-20. Threshold concentration for condensation nuclei formation from
various hydrocarbons. (Reprinted with permission from
98
Lipeles. ) 3-67
-------
Q
d
ID
CM
§00
8 2 Si
•rl
M
I
Pi
a)
§
•H
4J
O.
0)
•
CNJ
c
o
•rl
O
4-1
O
CO
o
j-i
0)
tfl
C
O
rl
m
o
C
o
o
o
CO
0)
u
CM
I
CO
(U
l-l
60
4J
0>|
g
•H
PQ
O
O
t-l
O
•H
CO
CO
•H
(U
p.
3-68
-------
useful for mechanistic studies, cannot be extrapolated to atmospheric
concentrations for these low-molecular-weight alkenes.
• Threshold concentrations for C alkenes are in the parts-per-
hundred-million to parts-per-million range. Determination of their gas-
phase concentration is required for assessing their contribution to
atmospheric organic aerosol.
• Nothing is known about the threshold concentration for aromatic
hydrocarbons. No vapor-pressure data on the polyfunctional products are
available. However, a possible threshold concentration of several parts
per million seems indicated by data of Kopzynski (mesitylene aerosol at
71
25 ppm, no aerosol at lower concentration), O'Brien (mesitylene aerosol
94 92
at 10 ppm, no aerosol at 2 ppm), Schwartz (toluene aerosol at 10 ppm),
100
and Grosjean (no toluene aerosol at 1 ppm). The situation is further
complicated by the low reactivity of aromatic compounds. Monofunctional
compounds formed early would stay in the gas phase, whereas polyfunctional
compounds formed later by further oxidation would accumulate in the
aerosol phase. Therefore, aerosol formation from aromatics would be
favored by meteorologic conditions that allowed long irradiation periods.
Further research is necessary for assessing in a more quantitative manner
the contribution of aromatic hydrocarbons to atmospheric aerosol.
Kinetics of Organic Aerosol Formation
Assuming that some of the physical and chemical mechanisms just
reviewed are predominant in the formation of organic aerosol, various
schemes can be derived that permit a more quantitative description of
the time evolution of atmospheric organic aerosol. For example, a
3-69
-------
15 7a
kinetic scheme has been proposed recently for aerosol formation from
olefinic precursors that may be applied in principle to other hydro-
carbon classes. Starting with this system.
C = C + \
S N
X
X
(3)
the olefin concentration is given by:
_ d[ol] =
dt
(4)
where k is the rate constant of olefin reaction with the th species X
±. 1 i
(0_, 0_, OH, HO , NOo,...). However, not all the rate constants k. have
J £* "* -i_
been measured, and the olefin consumption can be estimated from empirical
data, such as the nitric oxide to nitrogen dioxide conversion rate constants
106
measured by Glasson and Tuesday. Note also that Eq. 4 can be written:
>± k [03][ol]
3
(5)
where k = b. kO , and that [X ] = a.[0_], where a is only a function of
i i J 1 i J i
time and can be calculated from computer simulation data.
The Ozone-Olefin System. With Eq. 5, the general problem can be reduced
to the ozone-olefin system:
ZP1I i - 1,P
C - C + 0, -»- I
-^ 3
slow
(6)
3-7Q
-------
where EP-i.. and I^j are the products (including other intermediates)
that result from unimolecular and bimolecular reactions of the inter-
mediate I, whose formation is the rate-determining step.
Assuming a steady-state concentration for I,
I - k[03][olefin] / [ **it + Ek2j IRJ]]
(7)
where Rj - 0, 03> 02, OH, H02> N(>2, SO,,, etc.
The condensable species are generally only a fraction of the products:
and ZCS21 " a2
Assuming that the rate of gas-to-particle conversion of any condensable
species is greater than its rate of formation in the gas phase (which is
the case for heterogeneous nucleation predominant in the atmosphere, but
may not be valid for homogeneous nucleation in "clean-air" smog-chamber
studies):
d[aerosol] I dCSli Z dCS2i
dt dt dt
[I]
y j. j-j. t. f.j j i -
(8)
with Eq. 7,
d[aerosol] QlZk14 + a.,Ek04 [R
- k[03][olefin]
dt ikll
3-71
-------
The term a = d[aerosol] 1 is not constant,
dt k[03][olefin]
but is a function of [R.]. However, a can be constant in the following
cases:
• a = a I if k_.[R.] << k-. (fast unimolecular decomposition)
or if cu " 0 (no condensable species are formed in the
bimolecular pathways)
if
• [R. ] - constant, if the species R leading to condensable species
by reaction with the intermediate are either in great excess
(pseudo-first-order) or at steady-state concentrations (free
radicals) .
Note that the proposed scheme is general and that Eq. 9 can be applied to
any simpler system. For example, using sulfur dioxide 5 R , a, = 0,
a = 0 for j ^ 1, and a = 1 (all sulfur trioxide formed leads to
ZJ ZI 85
sulfuric acid aerosol), we find the Cox and Penkett relation:
*I«ero.ol] _ d[H2S04] _ k[olefin] [o^ k2i[S02]
dt . dt
which has been experimentally verified for the system:
.V ZPU, no CSU
k
olefin + 03 -" I
IP2j.noCS2. (11)
9,
+ SO -*• H0SO. aerosol
+S02 3 2 4
3-72
-------
A relation similar to Eq. 9 has been found to account satisfactorily
for organic aerosols formed from cyclopentene, cyclohexene, and 1,7-
157a
octadiene in smog chambers. The aerosol organic carbon concentration
as a function of time was measured with an organic-carbon analyzer
(Figure 3-22). The relation obtained was:
d (organic aerosol carbon) « a a k[0,][olefin] , (12)
dt C J
where a is the percent of organic carbon averaged over all condensable
\j
species.
The Ozone-Multihydrocarbon System. The same kinetic scheme can be used
for the most general situation involving multicompetitive olefin-ozone
reactions:
ol, -1 I,
-
01 > I
1
+*s
2
"i
+R
,.n
k
11
, n
•>
"
i
+R.
2j
(13)
3-73
-------
•rl O
4-1 1-1
nj 4-1
M
•P ti
d O
Qi -H
O CD
d U)
O -H
iH 01
ed ft
•H
4J J3
•H 4J
C -H
•a
• a)
CO 4->
(3 C
O -H
•H H
4J CX
I
O r-~
fi •>
cd B "~i
ft r-l
CO ft
cd
ro
0) fl
4-) •
(d O
3_LVd IMOIlVIAIdOJ 10SOd3V
d CM
o o
•H J3
4J
P a
c ft
O ft
14-1
r^
t-l —i
O •
CO O
o
S-i •
cu o
CM
CM
O
0)
M
60
•H
T3
0)
•H
M
T3
n)
cd
a>
•r-)
CO
O
M
O
3-74
-------
In this case, the ozone concentration is such that it satisfies simultane-
ously the m relations:
k» 'oW ' + Ek ' n -
With the only assumption being, as before, that the intermediate formation
step is slower than any further reactions, including transfer of conden-
sable species into the aerosol phase, we obtain the relation:
"(aerosol] - I <*t*"osol]m . [01 E a k [ol]m
dt m-l,n dt 3 m-l,n m m m
where, as before:
vi " j. „" TL-" m 1 / I Tie" -4- -
a « a. Ek,^ + <»„ ifcojlK^J / \ LK-u T
n 1
Note that Eq. 15 expresses in a more quantitative manner the previously
discussed dependence of aerosol formation on the product of a gas-phase
reactivity term, k , and an aerosol formation ability term, a .
m
m
Applications to the Atmosphere. The general relation of Eq. 15 can be
used for estimating amounts of organic aerosol formed from a multihydro-
carbon mixture, provided that four parameters—[0 ], a , k , and [ol] —
are known. Ozone concentration can be readily measured. Aerosol
3-75
-------
formation ability factors, a , can be measured in smog-chamber experi-
m
ments or estimated (a = 0 for paraffins, acetylenics, and C alkenes).
Note also that the measured ambient gas-to-particle distribution factor
51
f is an upper limit for Zv.m> Rate constants have been measured for
70,85,127,131,159-162
some olefins. Unknown rate constants for the ozone-
olefin reaction can be estimated from published linear relations between
163,164
kn and alkene ionization potentials and between k_ and the rate
3 157a,164 3
constants for other electrophilic additions. The last parameter,
olefin concentration, can be measured or estimated from data on the compo-
sition of gasolines and automobile exhaust (see the following section).
When combined with an estimation of air-mass trajectories and residence
time derived from meteorologic data, use of this type of kinetic relation
would permit predicting the amount of secondary organic aerosol present
at a given location of an urban basin by measuring the local ozone concen-
tration. However, it is not assumed so. far that such a relation as Eq. 15
is valid for all types of unsaturated and aromatic precursors (see Grosjean
15 7a
and Friedlander for discussion of other complex systems), and its
application to the atmosphere is limited by the scarcity of data on
ambient precursor concentrations.
Gas-Phase Hydrocarbon Precursors
More than 100 compounds are released in the atmosphere of urban areas
by automobiles, and there is a close relation between the atmospheric
hydrocarbon composition and the composition of gasolines and automobile
exhausts. The full range of compositions of gasolines has been reported
165,166
by Sanders and Maynard They identified 180 of the 240 compounds
separated by capillary-column gas chromatography. Detailed fuel compositions
3-76
-------
167-169
were reported by other investigators, and exhaust hydrocarbon
168,171 170
compositions were reported by Neligan, McEwen, and, more
172 173 174
recently, Papa, Jacobs, and Dishart. Exhaust compositions
170
were found to be very sensitive to vehicle regime and fuel composi-
174
tion. Methane, ethane, ethylene, and acetylene due to fuel cracking
always represent at least 30% of the exhaust. Ambient hydrocarbons with
171,175-180
no more than six carbon atoms are now measured on a routine basis,
but data on C-,, hydrocarbons have been limited to the paraffins and aro-
matics. Figure 3-23 illustrates the wide variety of hydrocarbons
found in gasoline, auto exhaust, and urban air.
Alkenes. Surprisingly, in view of the wealth of publications dealing
with the hydrocarbon composition of polluted atmospheres, fhere is vir-
tually no information on the existence and ambient concentrations of C7,
alkenes. Because of their reactivity and the stringent restriction on
the olefinic content in gasolines sold in California, C-, alkenes have
not been found in California ambient air. Only a few of them have been
181-184
identified elsewhere. On the basis of the difference between
nonmethane hydrocarbon concentrations and the sum of all C<, hydrocarbons
measured individually, the C fraction (paraffins + olefins + aromatics)
1 56
accounts for about half the nonmethane hydrocarbon concentration. In
the absence of any data for C alkenes, their possible concentrations
were estimated by scaling up gasoline and auto-exhaust data to match
the morning ambient concentration of 1-hexene. This crude estimate
indicates that C alkenes may be present at about 100 ppb in the atmos-
' . 185
phere during the morning traffic period. These alkenes are listed in
Table 3-12.
3-77
-------
NUMBER OF CARBON ATOMS. . .
PARAFFINS
N-ALKANES
BRANCHED ALKANES
CYCLIC
ACETYLENICS
AROMATICS
ALKYLBENZENES
POLYNUCLEAR
OLEFINS
ALKENES
DIOLEFINS
CYCLIC OLEFINS
STYRENE, INDENE, TERRENES
Figure 3-23. Hydrocarbons in gasolines (G), automobile exhaust (E),
and ambient urban air (A). A^ A^ A.^, A^= mono-, di-,
tri^, and tetra-substituted alkenes. X = not found.
Hatching = not found in California,
3-78
-------
CO
3
W
o
e
o
4-1
3
O
CO
3
O
CU
PH
CJ
•H
C
cd
too
cd
CJ
•H
I
O
O
4-1
O
TJ CO
CU 4-1
4-1 CJ
O 3
CU T3-
0, O
CM
•—1
1
CO
CU
rH
o
cd
H
r4
o
cd
o
•H
rH
^
X
o
^
^1
cd
o
•H
Q
rH
cd
o
•H
[ 1
O
a
3
UH
•H
rrj
(-4
CU
43
4-1
O
CO
ti
3
O
i1
O
0
+ + + I I
+ + + + I
11+11
o
CO
o
4-1
CU
4-1
O
Pn
01
01 C
S3 0)
0) 4-1
4-1 3
3 43
XI4-I
I
r-l
H
01 o) 01 01 o*
(3 C (3 C C
01 01 01 01 01
4J 4-1 4-1 4-1 4J
C f3 (3 13 C
OIOIOIOICUCUCUCUOIOIOI
01
O)
4J
0)
ft
IIIIIOIOIOIOIOIOI rH|3
I I I I I C C 0) 0) CU 01 43X C
l(»%>>>>>>>.ftftl I I I OI43C4J
NJ343434343I 1^-irHrHrH eiaicj
|4-l4->4-l4J4J<-,>^!>%CUHrHOCN
t>^43 43 43 43 4-1 I >•> I I
>>g g g 0
43 -rl -H -H -H -
4JOOQQO43434J4-I4-I
01 I I I I I 4J 4-1 01 CU CU
I I I
cu
I
co
0) (3
CU C 0)
S3 0) CJ
CO O 01
4-i cu -a
O T3 O
o o o
CO
(3
ai
•H
cd
CO
(3
CO
(3
CU
a) -H
(3 13
0) cd
•H 4J
cd PQ
ft I
O ro
P-i
3 UH
43 O) 0)
I C rH
en o) o
- -H -H
«—i ^3 T3
ijs +
43 01 lO
4J PL, c_)
0) I I
_ S en 3^
CN i—i a
3-79
-------
4->
(3
O
U
4J
to
td
43
X
W
Ol
rH
•H
43
O
§
"g
td
U]
O
O 3
O> T3
a o
X t-i
H PM
HI
•H
43 t-i
to
3
O ed
4-> 43
3£
C
td
to
T3
•H
0
td
u
•H
rH
>i
X
o
43
)-i
td
0
•H
O
I
rH
tfl
£3
O
•H
J_)
O
(3
3
4H
•H
T3
M
0)
43
4-1
O
to
T3
C
3
0
ex
6
o
o
+ I I I I
+ I
to
4-1
J-l
td
o
A|
cs
01
rH
43
td
H
c
a>
o
to
M
3
O
tu
o
•H
§
+ + I
43
4J
•H
U
II
O
O
to
O
tfl
•H
4-1
(3
01
4-1
O
PM
0)
01
4-1
tU
Ol 01
C (3
0) O>
4-1 4J
ti a
Ol Ol
(X CX
o o
(J 0)
(3 C
(0 Ol
!< X
(I) Ol
43 43
O O
O O Ol O O
>> >, (3 !>s >,
O O Ol (J O
o
rH
o
4J O
Ol rH
O
I I
a —i ro o
43
4J
I I
(3
01
tfl
^
4-4
U
O
4-1
Ml
C
•H
T3
td
0)
CO
0)
C
Ol
td
•H
ts
M
o
td
u
id
0)
•H
4H
•H
4-1
(3
0)
TJ
•H
3-80
-------
Cyclic Olefins. Five cyclic olefins have been identified in gasolines
and auto exhaust, cyclopentene being the only one detected in the atmos-
171,176
phere. The five cyclic olefins account for only about 0.5% by
weight of gasolines and 0.6% by volume of exhausts. Their ambient con-
185
centrations were estimated by scaling up gasoline and auto-exhaust
data to match the measured ambient morning cyclopentene concentration
and from the measured ambient concentrations of C,. aerosol difunctional
compounds. The two estimates agree and indicate that cyclic olefins may
be present in the morning atmosphere at 10-50 ppb. On the contrary, C
6+
diolefins are not present in gasolines and exhaust gases and have not
been found in the atmosphere (Table 3-13).
Terpenes. The possibility of widespread haze formation by sunlight irradi-
186
ation of terpenoid compounds from vegetation was first suggested by Went.
187 188 189
Went, Went and Rasmussen, and Rasmussen estimated the annual world-
wide contribution of forest hydrocarbon emissions and concluded that more
reactive hydrocarbons are released by tree foliage than by man's activities.
The major compounds emitted are monoterpenes (C-,n)—like a-pinene, 3-pinene,
189,190 191
limonene, and myrcene —and the hemiterpene (C ) isoprene. The
fate of these gaseous olefins in the atmosphere is still undetermined.
192
Went noticed that irradiated mixtures of nitrogen dioxide and a-pinene
produced ozone.and fine particles and postulated that terpene-nitrogen
oxide reactions, similar to the olefin-NO reaction of polluted urban
193
areas, may take place in forested rural areas. Rasmussen and Holdren
indicated that, on an individual basis, monoterpene hydrocarbon concentra-
tions are in the low parts-per-billion range in rural air.
3-81
-------
1— 1
o
CO
o
yi
CU
<
CJ
•H
fi
ca
60
M
0
42
ft
ft
•V
(-1
•H
<
4-1
C
CU
•H
43
O -H
iJ Pi
CTi , C
O -H
4-1
<4-l CU
0 iH
O
4-1
CO O
XI -H
1 i ]
•1— ' ^n
^-- CJ
K^
t^l
CM CJ
OO
43
1 1 i o
M— 1 l-J-i
O ft
I
4-J O
4^ CO
60
•H -
0) x~-
S o
0
^-1 «~H
CO
I ^
3 4-1
0 0)
01 l-l
1— I
o B
B o
(j
01 4-1
00^--
cfl
^ a
CU 0
> -H
CO CD
l-l
C 0)
—4 ^
CU K*
00 O
C O •
•H CO
B iH B
30-^-
CD CD 60
CD O 3
CO l-i
CU vo
•• Cfl CO
cfl 1
4-1 O 4-1
cfl -u 3
13 1 O
CO 42
4-1 CO CO
CO 60
3 4-1
cfl B--S cfl
•C 0
X CO CD
CU 13
13 C
01 ti 3
i-i ca o
•H ft
42 * B
0 C O
Boo
O -H
4J 4-1 i-l
3 0 CO
cfl cfl fi
01 O
13 S-i -H
C 4-1
cfl CU O
CD C
CU CO 3
C 42 4-1
•H ft -H
iH 1 13
O CD
CD Cfl i-H
Cfl 60 O
60 CD
B~S O
B 0 l-l
O O CU
1-1 .-i ca
Pn
O
3-82
-------
The gas-phase reactivity of various terpenes has been measured.
194
Stephens and Scott were the first to include two terpenes (pinene and
a-phellandrene) with their study of the relative reactivity of various
hydrocarbons. Both monoterpenes showed the high reactivity predicted
by their olefinic structure. Conversion of nitric oxide to nitrogen
dioxide in the presence of isoprene is at a rate intermediate between
106
those for ethylene, and trans-2-butene, and Niki reported rate constants
for the a-pinene and terpinolene-ozone reactions. Grimsrud et al.
measured the rate constants for the reaction of various terpenes
with ozone and with nitric oxide, the latter at the low concen-
trations (10-ppb hydrocarbon and 7-ppb nitric oxide) observed in rural
areas. Their data are listed in Table 3-14. Structural effects have
a major influence on terpene reactivity, and olefinic terpenes react with
ozone at rates comparable with those of the most reactive alkenes, such
as tetramethylethylene. Grimsrud et^ a^. also established that the
atmospheric reactivity of very reactive terpenes is due to their reaction
with ozone, whereas atmospheric reactivity of less reactive terpenes is
controlled by other mechanisms that involve free radicals. Therefore,
photochemical reactions of terpenes and nitric oxide may contribute, in
part, to the high ozone concentrations C>20 ppb, and sometimes exceeding
197,198
80 ppb) observed in rural daytime air.
3-83
-------
C
O
•H
o
•H
4-J
TO
co
LO
CM
U1
00
vO
•
O
O
•
CM
ft —t
CO
C
o
,0
TO
o
o
J-i
o
CO
u
cu
CO
r-1 CM
• •
o o
CM
cu
cu
ft
rl
cu
4-1
o
I
CO
•rl
CO
8
o
N
o
o
cu
CO
I
ft
CO
I
o
CO
CO
I
CO
I
o
co
o
CM
I
O
X
CO
CO
I
CO
•3
H
CO
cu
4-1
CO
•H
CO
O
C
O
N
O
c
o
•H
TO
TJ •>
*rH /*"*s
o ^
o
4-J CU
O 4J
o
cu
en
10
CM
<
o
o
4J
o
PH
o
TO
o
o
33
0)
c
g
cu
g
4-J
,£1
O
CO
M
cu
(3
CU
fi
•rl
P-l
ca
0)
c
01
ft
o
ca
M
cu
cu
•S
PM
8
CU
01
S-l
TO
O
CO
QJ
c
cu
C
TO
ca
01
a
CU
H
01
(3
CU
4->
C3
0)
6
o
TO
U
3-84
-------
o
•rl •• X
4J ro O
O S3
cfl
Pi
Pi
O
•
CM
00
•
CM
OO
CM
O
•
CM
i—I
I
co
I
CO
a
o
£>
rl
cfl
a
o
/—N
•
4J
C
o
u
-3-
rH
I
CO
•o
EC
0>
0
OJ
P.
rl
CU
4J
O
O
S
<4H
o
CO
a)
4J
CO
Pi
cfl
pi
CO
•H
CO
t^l
rH
O
rj
O
N
O
M
>^
O
N
O
d
cfl
C
o
cfl
•a
Photoo:
a
P,
o
CO
OO
CO
in
IT)
oo
o
CO
CJ
0)
CO
a
P
P
CM
1
O
rH
X
VD
•
i— l
CN
1
O
i — i
X
r-~
•
« — i
CM
1
O
rH
X
rH
•
CO
CM
1
O
1— 1
X
m
o
I—I
X
CM
O
1—I
X
CM
CM
C
o
•H
4J
cfl
TJ *
•H s^.
X r*
O v~'
O
4-1 CU
O 4J
PH Pi
rQ
O
i—l
X
o
CO
00
co
CO
m
l
in
cfl
u
o
rl
T3
0)
0)
a
I
CU
CD
O
B
o
rl
T3
CU
CU
CJ
CU
C
CU
O
O
CU
S
cu
rH
O
G
•H
P.
rl
01
H
OJ
C
QJ
M
T3
a
CO
I
S
cu
C
01
C
•H
P.
M
01
H
8
t£3
cn
i — i
•
rH
CO
4J
01
•a
rl
m
S
•H
M
o
a
0
M-J
tfl
4-1
cfl
p
« -
G
0
rJ
cfl
a
o
>>
o
p,
p.
1
O
i — i
T3
C
cfl
O
i-Q
P.
&
i
T3
0)
rj
•H
cfl
4-1
C
O
O
CO
rH
CU
CO
CO
cu
>
C
o
U
cfl
•H
•O
CO
rl
rl
H
Q
•
g
p,
in
1
1 — 1
•
0
M
CO
C
o
•H
4-1
xW
)H
4J
C
CU
o
C
0
a
rH
cd
•H
4J
•H
C
•H
C
0
^Q
rl
Cfl
O
0
M
T3
^i
^rj
13
C
CO
CU
ti
O
N
O
0 'I
•
p,
o
co
MH
O
C
o
•rl
4-1
Cfl
4-1
C
CU
O
C
O
0
cu
o
N
0
rH
cd
M
3
rl
rH
Cfl
U
•H
p.
4J
cd
J_J
0
14-1
0
•rl
4-1
Cfl
Pi
3
3-85
-------
Aerosol formation from terpenes has not been extensively investi-
78,79,82,92,98
gated. Only one olefinic terpene, a-pinene, has been studied.
78
Ripperton and Lillian reported condensation nuclei formation from irra-
diated mixtures of 0.1-ppm nitrogen dioxide and 0.5-ppm a-pinene. Fewer
condensation nuclei were obtained when increasing the water vapor concen-
79,82,98
tration. Infrared spectra of a-pinene aerosol indicate the
presence of organic compounds, such as carbonyls, carboxylic acid, and
nitrate ester. As mentioned earlier, the composition of a-pinene aerosol
was studied in great detail by Schwartz (Figure 3-8). Similar polyfunctional
compounds of low volatility are expected to be formed from other olefinic
terpenes ($-pinene, limonene) and probably constitute a major fraction
of the natural blue haze a.erosols formed over forested areas. It is note-
worthy that a-pinene and its aerosol products are also present at significant
45
concentrations in urban air. Diurnal profiles of the products are indi-
cative of their secondary origin (Figure 3-24). It has not been determined
whether tree foliage, solvent use (turpentine?), or transport from nearby
rural regions is responsible for the presence of terpenes in urban areas.
Relative Importance of Various Hydrocarbon Classes in the Formation of
Secondary Organic Aerosols. The contribution of the various classes of
hydrocarbons to the formation of particulate organic compounds is a complex
function of their relative: ambient concentrations, gas-phase reactivity,
and ability to form products whose physical properties, especially vapor
pressures, are of prime importance in the physical mechanisms controlling
the gas-to-aerosol conversion process. In view of the results discussed
previously, cyclic olefins appear to be the most important class of
organic aerosol precursors. This is due to their high gas-phase reactivity
3-86
-------
10
ro
E
\
o>
8
| | TERRENE PRODUCTS
a-PINENE
21
6 8 10 12 14 16 18
TIME, PST
21
Ft8ure 3-
Figure 3-11. (Adapted from Knight
3-87
s et
-------
and their ability to form difunctional compounds of very low volatility.
Such difunctional compounds constitute the major fraction of ambient
secondary organic aerosols, and most of them have been identified in
smog-chamber experiments.
The observed ambient organic aerosol formation rates are also con-
sistent with those estimated by extrapolation of smog-chamber kinetic
data. Other heavy unsaturates, such as styrene and indene, are present
in the atmosphere and may contribute, in part, to the formation of ben-
zoic acid and homophthalic acid, respectively. Diesel exhaust and in-
dustrial emission are possible sources of such heavy unsaturates.
Diolefins of C are not present in gasolines and exhaust gases and have
o+
not been found in the atmosphere, and their possible role as precursors
of the Cc_7 difunctional acidic compounds is seriously challenged.
Lower diolefins are emitted in automobile exhaust. Examination of vapor-
pressure data indicates that the bulk of their expected photooxidation
products remains in the gas phase, including most of the less volatile
C dicarboxylic acids.
Although the ambient concentrations of alkenes are about 10 times
higher than those of cyclic olefins, their contribution to the formation
of organic aerosols is much lower in both smog-chamber experiments and
the ambient atmosphere. Here again, examination of vapor-pressure data
reveals that vapor pressures of monofunctional oxygenates expected from
alkenes exceed by several orders of magnitude those of the difunctional
compounds that have the same numbers of carbon atoms. Thus, the
saturation concentration for monofunctional compounds and their later
3-88
-------
gas-to-aerosol conversion can be reached only in "favorable" conditions,
i.e., at high precursor (alkene) concentrations. However, ambient con-
centrations of C alkenes have not been measured, and their role as
o+
aerosol precursors cannot be ruled out. The importance of aromatic hydro-
carbons remains to be determined. A "compensation effect" is expected
between their low reactivity and the possible aromatic ring cleavage
that leads to multifunctional compounds of low volatility. On the basis
of all available ambient-air and laboratory data, cyclic olefins can be
reasonably postulated as the most important class of secondary organic
aerosol precursors. However, aromatics and, to a lesser extent, alkenes
may be important after a sufficiently long period, when the more efficient
aerosol-forming precursors have disappeared. Unfortunately, there are no
laboratory data based on periods long enough to simulate the slow transport
of air masses (2. 12 h) encountered in some urban areas (such as the eastern
part of the Los Angeles basin) where aromatics may outweigh cyclic olefins
as major organic aerosol precursors.
Formation of Secondary Inorganic Aerosols in Photochemical Systems
Recent studies suggest that potentially harmful health effects may
199
be associated with moderately high concentrations of sulfates and
200
nitrates and that both sulfate and nitrate aerosols contribute more
51,56,201
than organic particles to visibility degradation in urban areas.
Nitrates and sulfates in atmospheric aerosols can be formed by a wide
variety of homogeneous and heterogeneous reactions. Heterogeneous
reactions have been thoroughly studied and were considered, until recently,
as the major pathways for sulfur dioxide and NO removal and particle
X
formation. However, studies of various homogeneous reactions initiated
3-89
-------
in the last few years suggest that some of these reactions can compete
effectively with heterogeneous processes. The purpose of this section
is to examine the possible relation between ozone and the two major
secondary inorganic aerosol species, sulfates and nitrates. Accordingly,
only gas-phase photochemical reactions relevant to sulfate and nitrate
formation will be reviewed, and their relative importance in the overall
aerosol formation process will be assessed where possible. Information
on the corresponding heterogeneous reactions can be found in several
56,57,202-206
reviews.
Nitrate Aerosols. The two important aerosol nitrate precursors are nitrous
acid and nitric acid, formed mainly in the reactions:
NO + NO + H20 ->- 2HONO , (16)
N205 + H20 -> 2HON02 , (17)
NO™ + OH + M •> HONO + M . (18)
202
With currently accepted rate constants for these reactions, Calvert
estimated that Reaction 18 is of principal importance for nitric acid
formation in smog. Nitric acid has recently been identified in both smog
206 208
chambers and Los Angeles air. However, both nitrous acid and nitric
209
acid have high vapor pressures, and it does not appear possible that
particulate nitrogen species will exist in the atmosphere in pure form as
acids. Infrared, x-ray, and chemical analyses indicate that ammonium
nitrate is the major constituent of the ambient particulate nitrate
56,61,210,211
fraction. Ammonium nitrate is also found in aerosol formed
3-90
-------
from NO -hydrocarbon mixtures irradiated in the presence of ammonia in
x 94,100
smog chambers. Thus, the production of aerosol nitrate requires
neutralization by ambient ammonia. It is not certain whether the reaction
takes place in the gas phase or in the liquid phase after fast nitric
acid diffusion into the aerosol droplets. Urban ammonia concentrations
3
have been estimated at up to 0.2 ppm (140 yg/m ), with averages of one-tenth
212
that value. High aerosol nitrate concentrations are observed in the
56,211
eastern part of the Los Angeles basin, where ammonia emission from
feed lots is thought to be important. Also, diurnal aerosol nitrate pro-
files do not correlate with nitrogen oxide concentrations. Therefore,
aerosol nitrate formation may be limited by ambient ammonia, rather than
by any of the nitrogen oxide species.
The relative importance of the heterogeneous and photochemical
nitrate formation pathways can be assessed from recently measured aerosol
51,56
nitrate profiles. On humid days, a midmorning nitrate peak is
usually observed, which tracks rather poorly the morning rush hour NO
X
peak and indicates that heterogeneous processes are predominant. On
oxidant episode days, however, particulate nitrate concentrations correlate
with ozone concentrations. Diurnal profiles of several nitrogen species
51,213
are presented in Figure 3-25. In this case, photochemical produc-
3
tion of aerosol nitrate at a rate of about 10 pg/m per hour was observed.
Sulfate Aerosols. As mentioned before, heterogeneous sulfur dioxide
57,202-206
oxidation reactions have been extensively studied. Two reactions
of this type are thought to be particularly important: the aqueous oxida-
tion of sulfur dioxide in water droplets and the catalytic oxidation of
3-91
-------
to
O
0.8 -
).6
0.4
0.2
0
O.I
0.05
0.01
0
6:30
12:30 16:30
time, PDT
60
40
I
' ro
O
20
0
20:30
Figure 3-25. Diurnal profiles of nitrogen compounds, Pasadena, California, July 25,
1974. NO, NO o and PAN in ppm; and inorganic aerosol nitrates
in yg/m . (Data from ref. 51 and 213.)
3-92
-------
sulfur dioxide adsorbed on carbon particles. Because ozone is involved
in the former reaction and carbon particles are generated by incomplete
combustion of hydrocarbons, these two reactions will be examined hereafter.
214,215
Novakov et al. have observed that significant amounts of sulfate ion
can be found on carbon particles generated by combustion of hydrocarbons
in air enriched with sulfur dioxide in the parts-per-million range. It
is difficult to assess the significance of carbon or organic particles
for the sulfur dioxide oxidation in the atmosphere. There is little doubt
that absorption of sulfur dioxide on carbon particles freshly generated by
combustion can provide a surface-catalyzed oxidation medium. Indeed,
216
such experiments as those of Yamamoto et al. have shown that sulfur
dioxide oxidation can be as high as 30%/h on activated-charcoal particles
less than 5 mm in diameter. Their work also indicates that this rate is
strongly reduced by sulfuric acid collection in the micropores of the char-
coal. The work of Yamamoto e_t al. further emphasizes that such a hetero-
geneous oxidation mechanism depends on a variety of factors, including
grain size of the carbon, temperature, and concentrations of sulfur dioxide,
water vapor, and oxygen, as well as the micropore structure of the particle
surface. It would seem that oily, gummy, wet particles collected from
the atmosphere would be poorly suited for nonaqueous reactions to form
sulfate, in that their micropore structure would be minimal. Yet such a
mechanism cannot be ruled out.
The class of reactions that have been used most often to explain high
sulfur dioxide rates in the presence of aerosols that contain water is the
system involving sulfur dioxide absorption in water followed by oxidation
by dissolved oxygen and/or ozone to form sulfate. Catalysis of the
3-93
-------
oxidation by heavy metal ssalts, such as Mn ion, can realize rates
of oxidation in excess of 1%/h in clean water solutions (e.g., Johnstone
217
and Coughanowr ). The absorption of sulfur dioxide can be promoted by
the buffering effect of simultaneous absorption of ammonia. Scott and
218
Hobbs have shown that the aqueous sulfur dioxide oxidation process can
be enhanced significantly by ammonium ion. Indeed, the estimates and
219
experiments of Miller and dePena suggest that rates of sulfur dioxide
oxidation can be achieved in fog approaching 10%/h.
It is well known that ozone is quite soluble in water. Therefore,
one expects that ozone absorption with sulfur dioxide would contribute to
220
significant oxidation of sulfur dioxide. Experiments of Penkett have
shown that oxidation of sulfur dioxide in air at 7 ppb when absorbed in
water droplets with ozone, which is present in surrounding air at 0.05 ppm,
can be as large as 13%/h. Thus, foggy or cloudy air mixed with photochemical
smog, such as occurs sometimes along the Pacific coast, could well be an
important medium for sulfate ion formation. Furthermore, such an aqueous
mechanism could be significant at middle altitudes over continents even
at background ozone concentrations.
The reported rates of sulfur dioxide oxidation in clean water droplets
must be considered maximal. It is questionable whether they can ever be
achieved in the atmosphere, inasmuch as such aqueous reactions have been
shown to be suppressed significantly by organic contaminants. The work
221 222
of Fuller and Christ and later of Schroeter has indicated that the
aqueous absorption of sulfur dioxide and its later oxidation are reduced
by as much as an order of magnitude by dissolved organic acids or alcohols
3-94
-------
that are known to be present in atmospheric aerosol. Much experimental
evidence has been accumulated on homogeneous sulfur dioxide oxidation
reactions. All investigators report aerosol sulfate formation when sul-
fur dioxide is added to mixtures of NO and hydrocarbons in smog-chamber
X
experiments (see, for example, Table 3-5). The observed enhancement of
sulfur dioxide oxidation depends strongly on the hydrocarbon structure
(Table 3-6). Alkenes with fewer than seven carbon atoms, which constitute
the bulk of ambient unsaturates and do not form organic aerosol, always
produce sulfate aerosol when sulfur dioxide is added. On the contrary,
there is no significant sulfate formation and neither the nature nor the
yields of organic aerosol formed by cyclic olefins are affected by adding
sulfur dioxide. An "intermediate" class is represented by C? alkenes
84
(for example, 1-heptene ), which produce some organic aerosol in the
absence of sulfur dioxide and both organic and sulfate aerosol when sulfur
dioxide is present. Addition of sulfur dioxide does not seem to affect
89,105
aerosol formation from aromatic hydrocarbons significantly. Such
a striking hydrocarbon effect is not well understood and may -reflect the
differences in the nature or the stability of the intermediates involved
for each hydrocarbon class.
Various homogeneous reactions have been postulated to account for
sulfur dioxide oxidation in irradiated NO -hydrocarbon-sulfur dioxide
X
systems. Gas-phase inorganic reactions of sulfur dioxide with oxygen plus
light, ozone, oxygen atoms, nitrogen dioxide, nitrogen trioxide, and
nitrogen pentoxide have been considered severely rate-limited on the
202,206
basis of available rate data (e.g., Calvert ). Listed in Figure 3-26
are other reactions that appear to be important in the atmosphere. These
3-95
-------
I. PRECURSOR = SULFUR TRIOXIDE
1. Reactions with intermediates of the ozone-olefin reaction:
RI ' R»
\ / • RI
sos + ^ (p>R* >s°3 + ^c=0 + -c=0
molozonide
SO2 f ^C C ^ > same
ewitterion
II. 2. Reactions with organic free radicals:
SO + HO -> OH + SO
22 3
SO + RO -> OR + SO
II. PRECURSOR = QRGANOSULFUR Sf ECIES
SO + OH -> HOSO
SO + OR -> ROSO
o cr i
i
biradical '
Figure 3-26. Atmospheric SO oxidation to aerosol sulfate homogeneous gas-
phase organic reactions.
3-96
-------
reactions are divided into two subcategories whose end products are
sulfur trioxide and organic sulfur species. The first three reactions
85
in Figure 3-26 correspond to the interpretation of Cox and Penkett's
observations that sulfur dioxide is oxidized at appreciable rates in the
dark in ozone-olefin-air mixtures. The higher rate of 3%/h was found for
cis-2-pentene, whereas the lower rate of 0.4%/h was found for propylene.
Cox and Penkett suggested that either the ozonide or the zwitterion
intermediates were involved in sulfur dioxide oxidation. However, Calvert's
estimated lifetimes of these olefin-ozone intermediates do not favor their
importance as oxidizing agents. Other radical species—such as hydroxyl,
alkoxy, hydroperoxy, and alkyperoxy—may well account for sulfur dioxide
oxidation.
The importance of the hydroxyl-sulfur dioxide reaction was postulated
223
by Castelman. With measured rate constants for hydroperoxy-sulfur
224 225 226
dioxide (Davis et al. ) and other reactions (James et al., Gordon ),
Calvert estimated sulfur dioxide oxidation rates by hydroperoxy, methylperoxy,
hydroxyl, and methory radicals to be 0.85, 0.16, 0.23-1.4, and 0.48%/h,
respectively.
It is generally accepted that aerosol sulfate formation requires
formation of sulfur trioxide and later fast sulfur trioxide reaction with
water. However, consideration should be given to other possible sulfur
227
dioxide and sulfur trioxide reactions. For example, Urone and Schroeder
228
and Bricard et al. reported nitrosylbisulfate, NOHSO , formation during
' 4
229
photolysis of sulfur dioxide-NO mixtures, whereas Daubendiek reported the
x
formation of an unidentified white solid in the sulfur trioxide-nitrogen dioxide
reaction. The significance of this compound in urban aerosol formation
3-97
-------
should be evaluated. Moreover, it appears from reactions listed in
Figure 3-26 that various organic sulfur species must also be considered
as potential sulfate aerosol precursors. The radical addition products ,
such as OHSO , should react: rapidly with other species to form sulf uric
acid, peroxysulfuric acid, alkylsulf ates , and mixed anhydrides, such as
202
HOSO ON02. Any of these ultimately should lead to sulfate in the
presence of water.
Although no conclusive evidence has been reported so far, the possible
importance of organic sulfur species as sulfate aerosol precursors is
supported by several observations. Sulf uric acids, sulfonic acids, and
other organic sulfur compounds are formed in sulfur dioxide-hydrocarbon
230-232
reactions at high concentrations. Organosulfur radical species,
such as RSO and R02S02 have been postulated as intermediates for these
f\ Q Q f\ Q £• f\ O £L
reactions. Suzuki observed polymer formation from
S
/ \
0 0
units in the sulfur dioxlde-cis-butene reaction. More recently, Schulten
99
and Schurath reported several organosulfur compounds in aerosol formed
from the sulfur dioxide-ozone-1-butene system. They tentatively identified
these compounds as zwitterion-sulfur dioxide addition products and/or
mixed anhydrides of sulfuric acid and sulfurous acid with several organic
acids. These results are most interesting, inasmuch as the system was
studied at concentrations approaching those of interest in polluted atmos-
phere .
3-98
-------
Summing all the known homogeneous reactions for sulfur dioxide
oxidation, it is possible to rationalize a theoretical rate of sulfate
formation in the range 1.7-5.5%/h for moderate photochemical smog con-
ditions. Sulfur dioxide oxidation rates in nonphotochemically polluted
236 57
areas, such as central Europe and St. Louis, are in the range of
0.5-1.0%/h. Much higher sulfur dioxide oxidation rates, up to 13%/h,
have been measured in the Los Angeles basin for days of low humidity and
237
high oxidant (see also gas-particle distribution factor f in Table 3-4).
O
A strong influence of photochemistry is demonstrated for submicrometer
sulfate aerosol by the systematic increase in f < 0.5 urn with ozone con-
M
centration at various locations in the southern California air basin
(Figure 3-27). It appears, therefore, that conversion of sulfur dioxide
to sulfate is significantly enhanced by homogeneous reactions in photo-
chemically polluted atmosphere. It is possible, however, that the sulfate
problem in the next few years will be dominated by other factors, such as
the use of high-sulfur fuels and the use of oxidizing catalysts that cause
238-242
conversion of fuel-bound sulfur to sulfuric acid aerosol. More
quantitative data on aerosol formation in sulfur dioxide-NO -hydrocarbon
x
systems at atmospheric concentrations are necessary to estimate whether
significant reduction in sulfate concentrations can be achieved through
the control of oxidants.
Summary
Review of the literature provides ample evidence that aerosol formation
is an important part of the atmospheric chemistry linked with photochemical
oxidant production. The important chemical constituents of concern include
sulfate, nitrate, and secondary organic material.
3-99
-------
0.4
0.3
E
a.
LSI
\/
0.2
0.1
0.1
o
0.2
I
T
T
O WEST COVINA 7/23/73
• WEST COVINA 7/25/73
D WEST COVINA 7/26/73
• POMONA 8/17/73
O RUBIDOUX 9/19/73
O
0.3
(ppm)
0.4
0.5
0.6
Figure 3-27. Scatter diagram of the conversion ratio f based on particles smaller
O
than 0.5 ym vs. 2-h averaged ozone concentration. (Data from Hidy
56,57
^ ^' } 3-100
-------
Secondary organic aerosols, formed by gas-phase reaction between
nitrogen oxide, ozone, and hydrocarbons constitute an important fraction
of urban photochemical smog. Data obtained at high ozone concentrations
(0.67 ppm) can be taken as an upper limit of the contribution of secondary
organic aerosols to the organic aerosol fraction and total suspended
particulate material (95% and 65%, respectively). Most of the identified
ambient secondary organic aerosols are difunctional compounds that bear
carboxylic, nitrate, aldehyde, and alcohol groups. The same compounds
have been identified in smog chambers from C cyclic olefins and diolefins,
with gas-to-aerosol conversion factors exceeding by more than an order of
magnitude those measured for the ambient average conversion of all reac-
tive hydrocarbons. The formation of such species in the gas phase in
excess of their saturation concentration followed by condensation on
preexisting particles and further growth in the light-scattering range
is the predominant physical mechanism that controls the gas-to-aerosol
conversion process.
Because of their very low vapor pressures, difunctional compounds
are readily converted to the aerosol phase, whereas more volatile
monofunctional compounds require much higher precursor and ozone concen-
trations to reach their saturation concentration. This explains why
most of the compounds formed from alkenes remain in the gas phase, whereas
C cyclic olefins and diolefins are efficient aerosol precursors. How-
ever, there is no known source of the latter class, so cyclic olefins,
identified in both gasolines and auto exhaust, can be regarded as the
most important source of secondary organic aerosols. The role of aromatics
as aerosol precursors is essentially unknown. Because of their accumulation
3-101
-------
in the submicrometer range, all secondary organics are potentially
dangerous. However, there is almost no information on health effects
associated with the presence of such compounds in the atmosphere.
Because the conclusions presented in this chapter rely heavily on
a few recent studies, it ±s extremely difficult to relate the urban con-
centrations of secondary aerosols to the concentration of their gas-phase
precursors. Simple relations of the type d(secondary aerosol)/dt = a
(precursor) (ozone) have been derived from smog-chamber data for organic
aerosol formation in mixtures of cyclic olefins and NO and sulfate
X
aerosol formation in mixtures of NO , sulfur dioxide, and C alkenes.
X .J ^
Such kinetic data are consistent with the organic (a few micrograms per
cubic meter per hour) and sulfate (up to 13%/h) aerosol formation rates
observed in photochemically polluted urban areas. More complex kinetic
relations reflect certainly all the possible variations between these
extreme and rather simple systems. Although control of ozone, through
control of NO and total hydrocarbon emission, would obviously have a
x
roughly proportional effect on the formation of organic aerosols, present
data suggest the identification and control of a few specific hydrocarbon
precursors as an alternative approach. The contribution of photochemical
reactions involving hydrocarbons to inorganic nitrate and sulfate aerosol
formation remains to be determined. More data on the identification of
hydrocarbon precursors and on the kinetics of formation, physical parameters,
and health effects of their products would ultimately permit quantifying
the complex relations between secondary aerosols and ozone concentrations
in urban atmospheres.
3-102
-------
RECOMMENDATIONS FOR FUTURE RESEARCH
Our present knowledge of the chemical and physical processes that
govern aerosol formation in the atmosphere is rather limited, and further
studies are needed in most of the relevant areas of research. This may
leave the reader—and the decision-maker—with a feeling of endlessness.
However, substantial improvements could be made in a reasonable period
by focusing research efforts in the subjects most directly involved:
• Laboratory (smog-chamber) studies of aerosol formation from
aromatic hydrocarbons; gas-phase reaction mechanism, physical processes
controlling gas-to-aerosol conversion, kinetic data on aerosol formation
and aerosol growth, identification of the aerosol products, and effect
of hydrocarbon concentration on aerosol formation (threshold).
• Careful search, in the atmosphere, for aerosol precursors, such
as cyclic olefins and C alkenes.
6+
• Study of the possible health effects of exposure to difunctional
oxygenated organics (such as dicarboxylic acids) that are present in urban
aerosols.
• Identification of organic components of ambient aerosols, to
permit estimation of the relative importance of olefinic and aromatic
hydrocarbons as aerosol precursors.
• Estimation of the relative contributions of photochemical and
nonphotochemical pathways to the formation of inorganic nitrate and
sulfate aerosols.
Identification of organic components of ambient aerosols and esti-
mation of the contributions of various pathways are of immediate interest
for control strategies and could be achieved by using the existing
monitoring networks so as to provide more information on aerosol chemical
3-103
-------
composition. In view of the adverse effects (e.g., on health and visi-
bility) associated with subiaicrometer aerosols, an air quality standard
for submicrometer particles might be more adequate than the present
standard for total suspended particles.
3-104
-------
REFERENCES
1. Leighton, P. A. Photochemistry o£ Air Pollution. New York: Academic Press,
1961. 300 pp.
2. U. S. Environmental Protection Agency. Part 410--National primary and secon-
dary ambient air quality standards. Fed. Reg. 36:8186-8201, 1971.
3. TRW Systems Group. An Implementation Plan for Suspended Particulate Matter
in the Los Angeles Region. Final Report to EPA, Contract #68-02-1384.
Redondo Beach, Calif.: TRW Systems Group, March 1975. (UNVERIFIED)
4. National Academy of sciences. National Academy of Engineering. Coordinating
Committee on Air Quality Studies. Air Quality and Automobile Emission
Control. Vol. 3. The Relationship of Emissions to Ambient Air Quality.
U. S. Senate Committee Print Serial No. 93-24. Washington, D. C.:
U. S. Government Printing Office, 1974. 137 pp.
5. see Chapter 5 (this report)
6. Grosjean, D. Solvent extraction and organic carbon determination in atmos-
pheric particulate matter: The organic extraction-organic carbon
analyzer (OE-OCA) technique. Anal. Chem. 47:797-805, 1975.
7. Patterson, R. K. Automated pregl-dumas technique for determining total
carbon, hydrogen, and nitrogen in atmospheric aerosols. Anal. Chem.
45:605-609, 1973.
8. Stephens, E. R., and M. A. Price. Smog aerosol: Infrared spectra.
Science 168:1584-1586, 1970.
3-105
-------
9. Cunningham, P. T., S. A. Johnson, and R. T. Yang. Variations in chemistry
of airborne particulate material with particle size and time. Environ.
Sci. Technol. 8:131-135, 1974.
10. ttovakov, T. , P. K. Mueller, A. E. Alcocer, and J. W. Otvos. Chemical
composition of Pasadena aerosol by particle size and time of day.
III. Chemical states of nitrogen and sulfur by photoelectron spectros-
copy. j. Colloid Interface Sci. 39:225-234, 1972.
11. Schuetzle, D., A." L." Crittettden, and R, J. Charlson. Application of computer
controlled high resolution mass spectrometry to the analysis of air
pollutants. J. Air Pollut. Control Assoc. 23:704-709, 1973.
12. Jutze, G. A., and K. E. Foster (TR-2 Air Pollution Measurements Committee)
Recommended standard method for atmospheric sampling of fine particulate
matter by filter media--high-volume sampler. J. Air Pollut. Control
Assoc. 17:17-25, 1967.
13. Jewell, D. M., J. H. Weber, J. W. Bunger, H. Plancher, and D. R. Latham.
Division of Petroleum Chemistry Paper 13 Presented at Meeting of
the American Chemical Society, Washington, D. C., Sept. 12-17, 1971.
(UNVERIFIED)
14. Tabor, E. C,, T. E. Hauser, J." p\' Lodge, and R4" H, Butttschell. Characteristics
of the organic particulate matter in the atmosphere of certain American
cities. A.M.A. Arch. Ind. Health 17:58-63, 1958.
15. Renzetti, N. A., and G." J. Doyle. The chemical nature of the particulate in
irradiated automobile exhaust. J." Air Pollut. Control Assoc. 8:293-296,
1959.
3-106
-------
16. Cukor, P., L. L. Ciaccio, E. W. Lanning, and R. L. Rubino. Some chemical and
physical characteristics of organic fractions in airborne particulate
matter. Environ. Sci. Technol. 6:632-637, 1972.
17. Sawicki, E., T. W. Stanley, T. R. Hauser, H. Johnson, and W. Elbert. Corre-
lation of piperonal test values for aromatic compounds with the atmos-
pheric concentration of benzo(a)pyrene. Int. J. Air Water Pollut. 7:
57-70, 1963.
18. Ciaccio, L. L., R. L. Rubino, and J. Flores. Composition of organic constit-
uents in breathable airborne particulate matter near a highway. Environ.
Sci. Technol. 8:935-942, 1974.
19 Miller, D. F., W. E. Schwartz, P. W. Jones, D. W. Joseph, C. W. Spicer, C. J.
Riggle, and A. Levy. Haze Formation: Its Nature and Origin, 1973. EPA-
650/3-74-002. Columbus, Ohio: Battelle Columbus Laboratories, 1973.
/178 pp._/
20. Frassa, K. P., R. K. Siegfriedt, and C. A. Houston. Modern analytical
techniques to establish realistic crankcase drains. Society of
Automotive Engineers Paper 951D, New York, January 1965.
21. Mader, P. P., R. D. MacPhee, R. T. Lofberg, and G. P. Larson. Composition of
organic portion of atmospheric aerosols in the Los Angeles area. Ind.
Eng. Chem. 44:1352-1355, 1952.
22. Hoffman, D., and E. L. Wynder. Analytical and biological studies on gasoline
engine exhaust. Nat. Cancer Instit. Monograph 9:91-116, 1962.
23. Sawicki, E., S. P. McPherson, T. W. Stanley, J. Meeker, and W. C. Elbert.
Quantitative composition of the urban atmosphere in terms of poly-
nuclear aza heterocyclic compounds and aliphatic and polynuclear
aromatic hydrocarbons. Int. J. Air Water Pollut. 9:515-524, 1965.
3-107
-------
24. Dubois, L., A. Zdrojewski, P. Jennawar, and J. L. Monkman. The identifica-
tion of the organic fraction of air sample. Atmos. Environ. 4:199-
207, 1970.
25. Stanley, T. M., J. E. Meeker, and M. J. Morgan. Effects of various solvents
and conditions on the recovery of benzo(a)pyrene, benz(£)acridine, and
7H-benz(c[e)anthracene-7-one. Environ. Sci. Technol. 1:927-931, 1967.
26. Shabad, L. M., and G. A. Smirnov. Aircraft engines as a source of carcino-
genic pollution of the environment l_ benzo(a)pyrene studies_/. Atmos.
Environ. 6:153-164, 1972.
27. Sawicki, E., T. W. Stanley, S. McPherson, and M. Morgan. Use of gas-liquid
and thin-layer chrcmatography in characterising air pollutants by
fluorometry. Talanta 13:619-629, 1966.
28. Gordon, R. J. , and R. J,, Bryan. Patterns in airborne polynuclear hydrocarbon
concentrations at four Los Angeles sites. Environ. Sci. Technol. 7:
1050-1053, 1973.
29. Pierce, R. C., and M. Katz. Dependency of polynuclear aromatic hydrocarbon
content on size distribution of atmospheric aerosols. Environ. Sci.
Technol. 9:347-353, 1975.
30. Brocco, D., V. di Palo, and M. Possanzini. Improved chromatographic evalua-
tion of alkanes in atmospheric dust samples. J. Chromatogr. 86:234-
238, 1973.
31t Hauser, T. R., and J. N. Pattison. Analysis of aliphatic fraction of air
particulate matter. Environ. Sci. Technol. 6:549-555, 1972.
32. Liberti, A., G. P. Cartoni, and V. Cantuti. Gas chromatographic determina-
tion of polynuclear hydrocarbons in dust. J. Chromatogr. 15:141-148,
1964.
3-108
-------
33. McPherson, S. P., E. Sawicki, and F. T. Fox. Characterization and estima-
tion of n-alkanes in airborne particulates. J. Gas Chromatogr. 4:156-
159, 1966.
34. Arito, H., R. Soda, and H. Matsushita. Gas chromatographic determination
of polynuclear hydrocarbons in particulate air pollutants. Ind.
Health 5:243-259, 1967.
35. Melton, C. W., R. I. Mitchell, D. A. Trayser, and J. F. Foster. Chemical
and Physical Characterization of Automotive Exhaust Particulate Matter
in the Atmosphere (Year Ending June 30, 1972). Final Report. EPA 650/3-
73-001. Columbus, Ohio: Battelle Columbus Laboratories, 1973. 91 pp.
36. Shepherd, M., S. M. Rock, R. Howard, and J. Stormes. Isolation, identifi-
cation, and estimation of gaseous pollutants of air. Anal. Chem. 23:
1431-1440, 1951.
37. Safe, S., and 0. Hutzinger. Mass Spectrometry of Pesticides and Pollutants.
Cleveland, Ohio: CRC Press, 1973. 220 pp.
3g Majer, J. R., R. Perry, and J. M. Reade. The use of thin-layer chromatog-
raphy and mass Spectrometry for the ra*pid estimation of trace quantities
of air pollutants. J. Chromatogr. 48:328-333, 1970.
39. Perry, R. , R. Long, and J. R. Majer. The use of mass Spectrometry in the
analysis of air pollutants, pp. 557-562. In H. M. Englund and W. T.
Beery, Eds. Proceedings of the Second International Clean Air Congress.
Held at Washington, D. C., December 6-11, 1970. New York: Academic
Press, 1971.
40. Boyer, K. W., and H. A. Laitinen. Automobile exhaust particulates. Proper-
ties of environmental significance. Environ. Sci. Technol. 9:457-469,
1975.
3-109
-------
41. Lao, R. C., R. S. Thomas, H. Oja, and L. Dubois. Application of a gas chro-
matograph-mass spectrometer-data processor combination to the analysis
of the polycyclic aromatic hydrocarbon content of airborne pollutants.
Anal. Chem. 45:908-915, 1973.
42. Golden, C., and E. Sawicki. Ultrasonic extraction of total participate aro-
matic hydrocarbons from airborne particles at room temperature. Int.
J. Environ. Anal. Chem. (in press)
43^ U. S. Environmental Protection Agency. Office of Air and Water Programs.
Air Quality Data for Organics 1969 and 1970 from the National Air
Surveillance Networks. Report APTD-1465. Research Triangle Park,
N. C. : U. S. Environmental Protection Agency, 1973. /~34 pp. ~J
44. u. S. Department of Health, Education, and Welfare. Public Health Service.
National Air Pollution Control Administration. Air Quality Criteria
for Particulate Matter. NAPCA Publ. No. AP-49. Washington, D. C.:
U. S. Government Printing Office, 1969. 211 pp.
45. Knights, R. L., D. R. Cronn, and A. L. Crittenden. Diurnal Patterns of
Several Components of Urban Particulate* Air Pollution. Paper No. 3
Presented at 1975 Pittsburg Conference on Analytical Chemistry and
Applied Spectroscopy, Cleveland, Ohio, March 3, 1975.
46. Gordon, R. J. Patterns in airborne particulate pollutants at several Los
Angeles locations. Division of Environmental Chemistry Paper No. 40,
Presented at 167th National Meeting of the American Chemical Society,
Los Angeles, Califomia, April 3, 1974.
47. Pupp, C. , R. C. Lao, J. J. Murray, and R. F. Pottie. Equilibrium vapour
concentrations of some polycyclic aromatic hydrocarbons, As/Og and
SeO~ and the collection efficiencies of these air pollutants. Atmos.
Environ. 8:915-925, 1974.
3-110
-------
48. Stauff, J., and H. Fuhr. Chemiluminescence of the exhaust gases of an
internal combustion engine. Angew. Chem. Int. Ed. Engl. 14:105-106,
1975.
49. Santamaria, L., G. G. Giordano, M. Alfisi, and F. Cascione. Effects of
light on 3,4-benzpyrene carcinogenesis. Nature 210:824-825, 1966.
50. O'Brien, R. J., J. H. Crabtree, J. R. Holmes, M. C. Hoggan, and A. H.
Bockian. Formation of photochemical aerosol from hydrocarbons.
Atmospheric analysis. Environ. Sci. Technol. 9:577-582, 1975.
51. Grosjean, D., and S. K. Friedlander. Gas-particle distribution factors for
organic and other pollutants in the Los Angeles atmosphere. J. Air
Pollut. Control Assoc. 25:1038-1044, 1975.
52. Wilson, W. E., Jr., W. E. Schwartz, and G. W. Kinzer. Haze Formation.
Its. Nature and Origin. Battelle Report #CPA 70-172. Columbus, Ohio:
Battelle Columbus Laboratories, Jan. 28, 1972. 77 pp.
53. Schuetzle, D. Computer Controlled High Resolution Mass Spectrometric
Analysis of Air Pollutants. Ph.D. Thesis. Seattle: University
€
of Washington, 1972. 242 pp.
54. Schuetzle, D., D. Cronn, A. L. Crittenden, and R. L. Charlson. Molecular
composition of secondary aerosol and its possible origin. Environ.
Sci. Technol. 9:838-845, 1975.
55 Hidy, G. M. , et^ a_l. Characterization of Aerosols in California (ACHEX) .
Interim Report, Phase I. Thousand Oaks, Calif.: Rockwell Interna-
tional, 1973.
56a. Hidy, G. M., et._aj.. Characterization of Aerosols in California. (ACHEX)
Final Report. Vol. I-IV. SC524.25FR. Thousand Oaks, Calif.: Rock-
well International, 1974.
3-111
-------
56b. Hidy, G. M. , B. R. Appel, R. J. Charlson, W. E. Clark, S. K. Friedlander,
D. H. Hutchinson, T. B. Smith, J. Suder, J. J. Sesolowski, and K. T.
Whitby. Summary of the California aerosol characterization experiment.
J. Air Pollut. Control Assoc. 25:1106-1114, 1975.
57. Hidy, G. M., and C. S. Burton. Atmospheric aerosol formation by chemical
reactions. Int. J. Chem. Kinetics, Symp. No. 1 (Chemical Kinetic Data
for the Lower and Upper Atmosphere):509-541, 1975.
58. Appel, B. R. , P. Colodny, J. J. Wesolowski, P. K. Mueller, R. Knights, J.
Huntzicker, and D. Grosjean. The analysis of carbonaceous materials
in California atmospheric aerosols. Paper Presented at 1974 Pacific
Conference on Chemistry and Spectroscopy, San Francisco, California,
October 16, 1974.
59. Friedlander, S.' K. Chemical element balances and identification Of ait pollu-
tion sources. Environ. Sci. Technol. 7:235-240, 1972.
60. Gartrell, G., Jr., and S. K. Friedlander. Relating particulate pollution
to sources: The 1972 California aerosol characterization study.
Atmos. Environ. 9:279-299, 1975.
61. Lundgren, D. A. Atmospheric aerosol composition and concentration as a
function of particle size and of time. J. Air Pollut. Control Assoc.
20:603-608, 1970.
62> White, W. H., R. B. Husar, and S. K. Friedlander. A Study of Los Angeles
Smog Aerosol Dynamics by Air Trajectory Analyses. Paper 73-11, Pre-
sented at 66th Annual Meeting of the Air Pollution Control Association,
Chicago, Illinois, June 24-28, 1973.
63a. Haagen-Smit, A. J. Chemistry and physiology of Los Angeles smog. Ind.
Eng. Chem. 44:1342-1346, 1952.
3-112
-------
63b. Haagen-Smit, A. J., C. E. Bradley and M. M. Fox. Ozone formation in photo-
chemical oxidation of organic suhstances. Ind. Eng. Chem. 45:2086-
2089, 1953.
63c. Haagen-Smit, A. J., and M. M. Fox. Ozone formation in photochemical oxida-
tion of organic substances. Ind. Eng. Chem. 48:1484-1487, 1956.
g4 Cadle, R. D. p. 27. In L. H. Rogers, Ed. Proceedings of the
Conference on Chemical Reactions in Urban Atmospheres. Technical
Report 15.. San Marino, Calif.: Air Pollution Foundation, 1956.
65. Schuck, E. A., and G. J. Doyle. Photooxidation o£ Hydrocarbons in Mixtures
Containing Oxides of Nitrogen and Sulfur Dioxide. Report No. 29. San
Marino, Calif.: Air Foliation Foundation, 1959. 126 pp.
66. Renzetti, N. A., and G. J. Doyle. Photochemical aerosol formation in sulfur
dioxide-hydrocarbon systems. Int. J. Air Pollut. 2:327-345, 1960.
67. Schuck, E. A., G. J. Doyle, and N. Endow. A Progress Report on the
Photochemistry of Polluted Atmospheres. Technical Report 31.
San Marino, Calif.: Air Pollution Foundation, 1960.
68. Prager, M. J., E. R. Stephens, and W. E. Scott. Aerosol formation from gaseous
air pollutants. Ind. Eng. Chem. 52:521-524, 1960.
69. Endow, N., G. J. Doyle, and J. L. Jones. The nature of some model photochemical
aerosols. J. Air Pollut. Contr. Assoc. 13:141-147, 1963.
70. Wei, Y. K., and R. J. Cvetanovic. A study of the vapor phase reaction of
ozone with olefins in the presence and absence of molecular oxygen.
Can. J. Chem. 41:913-925, 1963.
71. Kopczynski, S. L. Photo-oxidation of alkylbenzerte-nitrogen dioxide mixtures
in air. Int. J. Air Water Pollut. 8:107-120, 1964.
3-113
-------
72. Stevenson, H. J. R., D. E. Sanderson, and A. P. Altshuller. Formation o£
photochemical aerosols. Int. J. Air Water Pollut. 9:367-375, 1965.
73. Harkins, J., and S. W. Nicksic. Studies on the role of sulfur dioxide in
visibility reduction. J. Air Pollut. Contr. Assoc. 15:218-221, 1965.
74. ftarlage, W. B., Jr., and F. C. Alley. Sampling and mass spectrometer
analysis of reaction products from the photochemical decomposition
of various olefins. J. Air Pollut. Control Assoc. 15:235-238, 1965.
75. Altshuller, A. P., D. L. Klosterman, P. W. Leach, I. J. Hindawi, and J. E.
Sigsby, Jr. Products and biological effects from irradiation of nitro-
gen oxides with hydrocarbons or aldehydes under dynamic conditions.
Int. J. Air Water Pollut. 10:81-98, 1966.
76 Goetz, A., and R. Pueschel. Basic mechanisms of photochemical aerosol
formation. Atmos. Environ. 1:287-306, 1967.
77. Urone, P., H. Lutsep, C. M. Noyes, and J. F. Parcher. Static studies of sul-
fur dioxide reactiors in air. Environ. Sci. Technol. 2:611-618, 1968.
78. Ripperton, L. A., and D. Lillian. The effect of water vapor on ozone
synthesis in the photo-oxidation of alpha-pinene. J. Air Pollut.
Control Assoc. 21:629-635, 1971.
79. Groblicki, P., and G. J. Nebel. The photochemical formation of aerosols
in urban atmospheres, pp. 241-263. In C. S. Tuesday, Ed. Chemical
Reactions in Urban Atmospheres. Proceedings of the Symposium held
at General Motors Research Laboratories, Warren, Michigan, 1969.
New York: American Elsevier Publishing Company, Inc., 1971.
80. Wilson, W.'E., Jr., E." L< Merryman, A. Levy, and H. R. Taliaferro. Aerosol
formation in photochemical smog. 17 Effect of stirring. J. Air Pollut.
Control Assoc. 21:128-132, 1971.
3-114
-------
81. Goetz, A., and 0. J. Klejnot. Formation and degradation of aerocolloids by
ultraviolet radiation. Environ. Sci. Technol. 6:143-151, 1972.
82. Ripperton, L. A., H. E. Jeffries, and 0. white. Formation of aerosols by
reaction of ozone with selected hydrocarbons. Adv. Chem. Ser. 113:
219-231, 1972.
83. Wilson, W. E. , Jr., A. levy, and E. H. McDonald. Role of SCIU and photochem-
ical aerosol in eye irritation from photochemical smog. Environ. Sci.
Technol. 6:423-427, 1972.
84. Wilson, W. E., Jr., A. Levy, and D. B. Wimmer. A study of sulfur dioxide in
photochemical smog. II. Effect of sulfur dioxide on oxidant formation
in photochemical smog. J. Air Pollut. Control Assoc. 22:27-32, 1972.
85 Cox, R. A., and S. A. Penkett. Aerosol formation from sulphur dioxide in the
presence of ozone and olefinic hydrocarbons. J. Chem. Soc. Faraday Trans,
I. 68:1735-1753, 1972.
gg^ Kocmond, W. C., D. B. Kittelson, J. Y. Yang, and K. L. Demerjian. Deter-
mination of the Formation Mechanisms and Composition of Photochemical
Aerosols. Report No. NA 5365-M-l. EPA-650/3-73-002. Buffalo, N.Y. :
Calspan Corporation, 1973. 101 pp.
87. Lipeles, M. , C. S. Burton, H. H. Wang, E. P. Parry, and G. M. Hidy.
Mechanisms of Formation and Composition of Photochemical Aerosols.
Final Report. EPA-R3-73-036. Thousand Oaks, Calif.: Rockwell
International Corporation, July 1973. 104 pp.
88. Wilson, W. E., Jr., D. F. Miller, A. Levy, and R. K. Stone. The effect of
fuel composition on atmospheric aerosol due to auto exhaust. J. Air
Pollut. Control Assoc. 23:949-956, 1973.
3-115
-------
89. Miller, D, F., A. Levy, and W. E. Wilson, Jr. A Study of Motor Fuel Composi-
tion Effects on Aerosol Formation. Part II. Aerosol Reactivity Study
of Hydrocarbons. Battelle Columbus Laboratories Report to the Committee
for Air And Water Conservation, American Petroleum Institute, February
21, 1972.
90. Cox, R. A., and S. A. Pertkett. Oxidation of atmospheric S02 by products of
the ozone-olefin reaction. Nature 230:321-322, 1971.
91. McNelis, D. N. Aerosol Formation from Gas-Phase Reactions of Ozone and
Olefin in the Presence of Sulfur Dioxide. EPA 650-4-74-034, Research
Triangle Park, N. C.: U. S. Environmental Protection Agency, 1974. 219 pp,
92. Schwartz, W. Chemical Characterization of Model Aerosols. EPA-650/3-74-011.
Columbus, Ohio: Battelle Memorial Institute, 1974S 129 pp.
93. Chu, R. R., and C. Orr, Jr.
Paper 74-156, Presented at 67th Annual Meeting of the Air Pollution
Control Association, Denver, Colorado, June 9-13, 1974.
94. O'Brien, J. R. , J. R. Holmes, and A. H. Bockian. Formation of photochemical
aerosol from hydrocarbons. Chemical reactivity and products. Environ.
Sci. Technol. 9:568-576, 1975.
95, Burton, C. S., E. Franzblau, and G. M. Hidy.
Division of Colloid and Surface Chemistry Paper No. 128, Presented at
the 167th National Meeting of the American Chemical Society, Los
Angeles, California, March 31-April 5, 1974.
96. Smith, J. P., and P. Urone, Static studies of sulfur dioxide reactions
effects of N02, C3H6, and H20. Environ. Sci. Technol. 8:742-746, 1974.
97. Penzhorn, R. D., and W. Schock. pp. 37-51.
In Chemische Reaktionen Atmospharischer Schadstoffe. Institut fur
Radiochemie, Preport No. KFK 1975 UF, Karlsruhe, Germany, 1974.
3-116
-------
98. Lipeles, M. The Formation of Aerosols in a Photochemical Fast Flow Reactor.
Report to the Air Pollution Research Advisory Committee, Rockwell
International, Thousand Oaks, California, 1975. (UNVERIFIED)
99. Schulten, H.-R., and U. Schurath. Analysis of aerosols from the ozonolysis
of 1-butene by high-resolution field desorption mass spectrometry. J.
Phys. Chem. 79:51-57, 1975.
100. Grosjean, D. The nature and formation of secondary aerosol organics. Part
1. Aerosol products from cyclic olefins and diolefins, pp.
In G. M. Hidy and P. K. Mueller, Eds. The Character and Origins of
Smog Aerosol. Ann Arbor, Mich.: Ann Arbor Press. (in press) (UNVERIFIED)
101. Doyle, G. J., and N. A. Renzetti. The formation of aerosols by irradiation
of dilute auto exhaust. J. Air Pollut. Control Assoc. 8:23-32, 1958.
102. Doyle, G. J., and J. L. Jones. Automobile exhaust-gas aerosols: A review
of studies conducted at Stanford Research Institute. J. Air Pollut.
Control Assoc. 13:365-367, 387, 1963.
103. Schuck, E. A., H. W. Ford, and E. R. Stephens. Air Pollution Effects of
Irradiated Automobile Exhaust as Related to Fuel Composition. Report
No. 26. San Marino, Calif.: Air Pollution Foundation, 1958. 91 pp.
104. Hamming, W. J., P. P. Mader, S. W. Nicksic, J. C. Romanovsky, and L. G.
Wayne. Gasoline Composition and the Control of Smog. Los Angeles
County Air Pollution District Report, September 1961. (UNVERIFIED)
105. Wilson, W. E., Jr., D. F. Miller, D. A. Trayser, and A. Levy. A Study of
Motor Fuel Composition Effects on Aerosol Formation. Part III. Visi-
bility Reduction from Automobile Exhaust. Battelle Columbus Laboratories
Report to the Committee for Air and Water Conservation, American
Petroleum Institute, February 21, 1972. (API Publ. 4147) (UNVERIFIED)
3-117
-------
106. Glasson, W. A., and C. S. Tuesday. Hydrocarbon reactivities in the atmos-
pheric photooxidation of nitric oxide. Environ. Sci. Technol. 4:916-
924, 1970.
107. Jaffe, R. J. , and F. R. Smith.
Paper 74-246, Presented at 67th Annual Meeting of the Air Pollution
Control Association, Denver, Colorado, June 9-13, 1974. (UNVERIFIED)
108. Altshuller, A. P., and I. R. Cohen. Structural effects on the rate of nitro-
gen dioxide formation in the photo-oxidation of organic compound-nitric
oxide mixtures in air. Int. J. Air Water Pollut. 7:787-797, 1963.
109. Fox, D. I., J. E. Sickles, M. R. Kuhlman, P. C. Reist, and W. E. Wilson.
Design and operation parameters for a large ambient aerosol chamber.
J. Air Pollut. Control Assoc. 25:1049-1053, 1975.
110. Daubendieck, R. I., and J. G. Calvert. The reaction of ozone with per-
fluorinated polyolefins. Environ. Lett. 6:253-272, 1974.
111. Liu, B. Y. H. , and D. Y. H. Pui. A submicron aerosol standard and the pri-
mary, absolute calibration of the condensation nuclei counter. J.
Colloid Interface ScL. 47:155-171, 1974.
112. Charlson, R. J. Atmospheric visibility related to aerosol mass concentra-
tion. A review. Environ. Sci. Technol. 3:913-918, 1969.
113. Baumgardner, R. E., T. A. Clark, J. A. Hodgeson, and R. K. Stevens. Deter-
mination of an ozone interference in the continuous Saltzman nitrogen
dioxide procedure. Anal. Chem. 47:515-521, 1975.
114. Winer, A. M., J. W. Peters, J. P. Smith, and J. N. Pitts, Jr. Response of"
commercial chemiluminescent NO-NO™ analyzers to other nitrogen-contain-
ing compounds. Environ. Sci. Technol. 8:1118-1121, 1974.
3-118
-------
115. California Air Resources Board. Final Report of the ad hoc Oxidant Measure-
ment Committee, Report No. 75-4-4, February 20, 1975. (UNVERIFIED)
116. Judeikis, H. S., and S. Siegel. Particle-catalyzed oxidation of atmospheric
pollutants. Atmos. Environ. 7:619-631, 1973.
117. Graedel, R. E., L. A. Farrow, and T. A. Weber. The influence of aerosols
on the chemistry of the troposphere. Int. J. Chem. Kinetics, Symp.
No. 1 (Chemical Kinetic Data for the Lower and Upper Atmosphere):
581-594, 1975.
118.. Mulcahy, M. F. R., and B. C. Young. Heterogeneous reactions of OH radicals.
Int. J. Chem. Kinetics, Symp. No. 1 (Chemical Kinetic Data for the
Lower and Upper Atmosphere):595-609, 1975.
119. Vrbaski, T., and R. J. Cvetanovic. A study of the products of the reactions
of ozone with olefins in the vapor phase as determined by gas-liquid
chromatography. Can. J. Chem. 38:1063-1069, 1960.
120. Pitts, J.'Nt) Jr., and B. J, Finlayson. Mechanisms of photochemical air pollu-
tion. Angew. Chem. (Engl.) 14:1-15, 1975.
121. O'Neal, H. E., and C. Blumstein. A new mechanism for gas phase ozone-olefin
reactions. Int. J. Chem. Kinetics 5:397-413, 1973.
122a. Criegee, R. , and G. Werner. Die Ozonisierung des 9,10-Oktalins. Justus
Liebigs Ann. Chem. 564:9-15, 1949.
122b. Delete—same as 122a.
122c. Cfiegee, R. The Course of ozOnization of UiiSaturated compounds. Rec. Chem.
Progr. 18:111-120, 1957.
123a. Bailey, P. S. The reactions of ozone with organic compounds. Chem. Rev. 58:
925-1010, 1958.
3-119
-------
123b. Gould, R. F., Ed. Ozone Reactions with Organic Compounds. Advances in Chem-
istry Series 112. Washington, D. C.: .American Chemical Society, 1972.
129 pp.
124. Gillies, C. W., R. P. Lattimer, and R. L. Ruczkowski. Microwave and mass
spectral studies of the ozonlyses of ethylene, propylene, and c is - and
trans-2-butene with added oxygen-18 formaldehyde and acetaldehyde. J.
Amer. Chem. Soc. 96:1536-1542, 1974.
125. Fliszar, S., and J. Renard. Quantitative investigation o£ the ozonolysis
reaction. XIV. A simple carbonium ion stabilization approach to the ozone
cleavage of unsymmetrical olefins. Can. J. Chem. 48:3002-3018, 1970.
126. Hull, L. A., I. C. Hisatsune, and J. Heicklen. Low-temperature infrared
studies of simple alkene-ozone reactions. J. Amer. Chem. Soc. 94:
4856-4864, 1972.
•^27 Hanst, P. 1., E. R. Stephens, W. E. Scott, and R. C. Doerr. Atmospheric
ozone-olefin reactions. Preprints Div. Petrol. Chem. Amer. Chem. Soc.
4(4):A7-A16, 1959.
128. Atkinson, R., B. J. Finlayson, and J. N. Pitts, Jr. Photoionization mass
spectrometer studies of gas phase ozone-olefin reactions. J. Amer.
Chem. Soc. 95:7592-7599, 1973.
129. Pinlayson, B." J.", J,' N." Pitts, Jr., and R.'Atkinson. Low-pressure gas-phase
ozone-olefin reactions. Chemiluminescence, kinetics, and mechanisms. J.
Amer. Chem. Soc. 96:5356-5357, 1974.
130. Altshuller, A. P., I. R. Cohen, S. F. Sleva, and S. L. Kopczynski. Air
pollution: Photooxidation of aromatic hydrocarbons. Science 138:
442-443, 1962.
3-120
-------
131. Bufalini, J. J., and A. P. Altshuller. Kinetics of vapor-phase hydrocarbon-
ozone reactions. Can. J. Chem. 43:2243-2250, 1965.
132. Stedman, D. H., and H. Niki. Ozonolysis rates of some atmospheric gases.
Environ. Lett. 4:303-310, 1973.
133. Boocock, G. , and R. J. Cvetanovic. Reaction of oxygen atoms with benzene.
Can. J. Chem. 39:2436-2443, 1961.
134. Jones, G. R. H., and R. J. Cvetanovic. Reaction of oxygen atoms with
toluene. Can. J. Chem. 39:2444-2451, 1961.
135. Atkinson, R., and J. N. Pitts, Jr. Temperature dependence of the absolute
rate constants for the reaction of 0(^P) atoms with a series of aromatic
hydrocarbons over the range 299-392°K. J. Phys. Chem. 79:295-297, 1975.
136. Davis, D. D., W. Bellinger, and S. Fischer. A kinetics study of the
reaction of the OH free radical with aromatic compounds. I. Absol-
ute rate constants for reactions with benzene and toluene at 300°K.
J. Phys. Chem. 79:293-294, 1975.
137. Doyle, G. J., A. C. Lloyd, K. R. Darnall, A. M. Winer, and J. N. Pitts, Jr.
Gas phase kinetic study of relative rates of reaction of selected aro-
matic compounds with hydroxyl radicals in an environmental chamber.
Environ. Sci. Technol. 9:237-241, 1975.
138. Fuchs, N. A. The Mechanics of Aerosols. New York: The MacMillan Co.,
1964.
139 Hidy, G. M., and J. R. Brock. The Dynamic of Aerocolloidal Systems. Inter-
national Review on Aerosol Physics and Chemistry Vol. 1. New York: Per-
gamon Press, 1970. 379 pp.
Mirabel, P., and J. L. Katz, Binary homogeneous nucleation as a mechanism
for the formation of aerosols. J. Chem. Phys. 60:1138-1144, 1974.
3-121
-------
!4i. Husar, R. B., K. T. Whitby, and 6. Y. H. Liu. Physical mechanisms governing
the dynamics of Los Angeles smog aerosol. J. Colloid Interface Sci.
39:211-224, 1972.
142. Willeke, K., and K. T. Whitby. Atmospheric aerosols: Size distribution
interpretation. J. Air Pollut. Control Assoc. 25:529-534, 1975.
p^ Kiang, C. S. , D. Stauffer, V. A. Mohnen, J. Bricard, and D. Vigla. Hetero-
molecular nucleation theory applied to gas-to-particle conversion.
Atmos. Environ. 7:1279-1283, 1973.
-,/, Cadle, R. D. Formation and chemical reactions of. atmospheric particles.
J. Colloid Interface Sci. 39:25-31, 1972.
145 Katz, J. L., C. J. Scoppa, II., N. G. Kumar and P. Mirabel. Condensation of
a supersatuated vapor. II. The homogeneous nucleation of the n-alkyl
benzenes. J. Chem. Phys. 62:448-465, 1975.
146. Heisler, S., S. K.Friedlander, and R. B. Husar. The relationship of smog
aerosol size and chemical element distributions to source characteristics.
Atmos. Environ. 7:633-649, 1973.
147. O'Brien, J. R., J. R. Holmes, and A. H. Bockian. Formation of photochemical
aerosol from hydrocarbons. Chemical reactivity and products. Environ.
Sci. Technol. 9:568-576, 1975.
148a. Heisler, S. L. Gas-to-Particle Conversion in Photochemical Smog: Growth
Laws and Mechanisms for Organics. Ph.D. Thesis. Pasadena: California
Institute of Technology, 1976. 215 pp.
148b. Heisler, S. L., and S. K. Friedlander. Gas to particle conversion in photo-
chemical smog: Growth laws and mechanisms for organics. Atmos. Environ.
(in press)
3-122
-------
149. Davis, M. , and G. H. Thomas. The lattice energies, infra-red spectra, and
pollisble cyclization of some dicarboxylic acids. Trans. Farady Soc.
56:185-192, 1960.
150. Jordan, T. B. Vapor Pressure of Organic Compounds. New York: Interscienee,
1954. 266 PP.
151> Hughes, E. E., and S. G. Lias.
National Bureau of Standards Report 6435. Washington, D. C. , 1960.
(UNVERIFIED)
152. Aliphatic dibasic acids, Chapter 1, pp. . In J. K. Stille and
T. W. Campbell, Eds. High Polymers. Vol. 27. Condensation Monomers.
New York: John Wiley and Sons, 1972. (UNVERIFIED)
153. Weast, R. C., Ed. CRC Handbook of Chemistry and Physics. (47th ed.)
Cleveland: The Chemical Rubber Company, 1966.
154. Rossini, F. D. , B. J. Mair, and A. J. Streiff. • Hydrocarbons From Petroleum.
ACS Monograph Series. New York: Reinhold Publishing Corporation,
1953. 556 pp.
155. Bachman, G. B., and N. W. Connon. Nitration studies. XIV. Conversion of
nitrite and nitrate esters into nitro alkanes. J. Org. Chem. 34:4121-
4125, 1969.
156t Grosjean, D., and S. K. Friedlander.
Paper No. 75 Presented at Pacific Conference on Chemistry and Spectros-
copy, Los Angeles, California, October 28-30, 1975. (UNVERIFIED)
157a. Grosjean, D., and S. K. Friedlander. The nature and formation of secondary
aerosol organics. Part II. Kinetics of organic aerosol formation.
(in press) (UNVERIFIED)
3-123
-------
157b. Grosjean, D. Secondary organic aerosols and their gas phase hydrocarbon
precursors. To be presented at the American Chemical Society Centennial
Meeting, Division of Environmental Chemistry, New York, April 4-9,
1976. (UNVERIFIED)
158. Demerjian, K. L. , ,T. A. Kerr, and J. G. Calvert. The mechanism of photo-
chemical smog formation. Adv. Environ. Sci. Technol. 4:1-262, 1974.
159. Herron, J. T., and R. E. Huie.
Int. J. Mass Spectrosc. Ion Phys. 16:125- , 1975. (UNVERIFIED)
160. Japar, S. M., C. H. Wu, and H. Niki. Rate constants for the reaction of
ozone with olefins in the gas phase. J. Phys. Chem. 78:2318-2320, 1974.
161. Becker, K. H. , U. Schurat'h, and H. Seitz. Ozone-olefin reactions in the gas
phase. 1. Rate constants and activation energies. Int. J. Chem. Kinet.
6:725-739, 1974.
162. Vbraski, T., and R. J. Cvetanovic. Relative rates of reaction of ozone with
olefins in the vapor phase. Can. J. Chem. 38:1053-1062, 1960.
Masclet, P., D. Grosjean,, G. Mouvier, and J. Dubois. Alkene ionization
potentials. Part I: Quantitative determination of alkyl group
structural effects. J. Electr. Spectrosc. Relat. Phenom. 2:225-
237, 1973.
164. Grosjean, D., P. Masclet, and G. Mouvier. Potentiels d'ionisation des
alcenes. III. Relations entre les PI, les vitesses d'additions
electrophiles et al reactivite des alcenes en pollution atmospherique.
Bull. Soc. Chim. Fr. 1974:573-576.
165. Sanders, W. N., and J. B. Maynard. Capillary gas chromatographic method
for determining the C3-C hydrocarbons in full-range motor gasolines,
Anal. Chem. 40:527-535, 1968.
3-124
-------
166. Maynard, J. B., and W. N. Sanders. Determination o£ the detailed hydrocar-
bon composition and potential atmospheric reactivity of full-range
motor gasolines. J. Air Pollut. Control Assoc. 19:505-510, 1969.
167. Glasson, W. A., and C. S. Tuesday. Hydrocarbon reactivity and the kinetics
of the atmospheric photooxidation of nitric oxide. J. Air Pollut.
Control Assoc. 20:239-243, 1970.
168. Neligan, R. E. , P, P. Mader, and L. A. Chambers. Exhaust composition in
relation to fuel composition. J. Air Pollut. Control Assoc. 11:178-
186, 1961.
169. Melton, C. W. , R. T. Mitchell, D. A. Trayser, and J. F.Foster. Chemical
and Physical Characterization of Automotive Exhaust Particulate Matter
in the Atmosphere. CRC-APRAC Report No. CAPE-19-70-2. Columbus, Ohio:
Battelle Columbus Laboratories, 1973. 90 pp.
170. McEwen, D. J. Automobile exhaust hydrocarbon analysis by gas chromatography.
Anal. Chem. 38:1047-1053, 1966.
171. Neligan, R. E. Hydrocarbons in the Los Angeles atmosphere. A comparison be-
tween the hydrocarbons in automobile exhaust and those found in the Los
Angeles atmosphere. Arch. Environ. Health 5:581-591, 1962.
172. Papa, L. J., D. L. Dinsel, and W. C. Harris. Gas chromatographic determina-
tion of C, to C1„ hydrocarbons in automotive exhaust. J. Gas Chromatogr.
6:270-279, 1968.
173. Jacobs, E, S. Rapid gas chromatographic determination of C-^ to CIQ. Hydro-
carbons in automotive exhaust gas. Anal. Chem. 38:43-48, 1966.
174. Dishart, K. T. Exhaust hydrocarbon composition. Its relation to gasoline
composition. Proc. Amer. Petrol. Inst. 50:514-540, 1970.
3-125
-------
175. U. S. Department of Health, Education, and Welfare. Public Health Service.
National Air Pollution Control Administration. Air Quality Criteria
for Hydrocarbons. NAPCA Publ. No. AP-64. Washington, D. C.: U. S.
Government Printing Office, 1970.
176. Stephens, E. R., and F. R. Burleson. Distribution of light hydrocarbons in
ambient air. J. Air Pollut. Control Assoc. 19:929-936, 1969.
!77. Altshuller, A. P., W. A. Lonneman, P. D. Sutterfield, and S. L. Kopczynski.
Hydrocarbon composition of the atmosphere of the Los Angeles basin -
1967. Environ. Sci. Technol. 5:1009-1016, 1971.
178. Kopczynski, S. L., W. A. Lonneman, F. D. Sutterfield, and P. E. Darley.
Photochemistry of atmospheric samples in Los Angeles. Environ. Sci.
Technol. 6:342-347, 1972.
179. Siddiqui, A. A., and F. L. Worley, Jr. Hydrocarbons in Houston's atmosphere.
Paper Presented at Conference on Ambient Air Qualtiy Measurements,
Air Pollution Control Association, Lakeway, Texas, March 10-11, 1975.
180. Kopczynski, S. L., W. A. Lonneman, T. Winfield, and R. Seila. Gaseous
pollutants in St. Louis and other cities. J. Air Pollut. Control
Assoc. 25:251-255, 1975.
181< Bertsch, W., R. C. Chang, and A. Zlatkis. The determination 6f btganic
volatiles in air pollution studies: Characterization of profiles. J.
Chromatogr. Sci. 12:175-182, 1974.
182. Raymond, A., and G. Guichon. Gas chromatographic analysis of Cg-C18 hydro-
carbons in Paris air. Environ. Sci. Technol. 8:143-148, 1974.
3-126
-------
183. Grob, K., and G. Grob. Gas-liquid chromatographic-mass spectrometric inves-
tigation of C/--C- organic compounds in an urban atmosphere. An appli-
cation of ultra trace analysis on capillary columns. J. Chromatogr. 62:
1-13, 1972.
184. Jones, P. W.
Paper 74-265, Presented at 67th Annual Meeting of the Air Pollution
Control Association, Denver, Colorado, June 9-13, 1974. (UNVERIFIED)
185. Grosjean, D., and S. K. Friedlander. The nature and formation of secondary
aerosol organics. Part III. Cyclic olefins and diolefins as possible
hydrocarbon precursors. (in press) (UNVERIFIED)
186. Went, F. W. Blue hazes in the atmosphere. Nature 157:641-643, I960.
187. Went, F. W. Organic matter in the atmosphere, and its possible relation to
petroleum formation. Proc. Nat. Acad. Sci. 46:212-221, 1960.
188. Rasmussen, R. A., and R. W. Went. Volatile organic material of plant origin
in the atmosphere. Proc. Nat. Acad. Sci. 53:215-220, 1965.
189. Rasmussen, R. A. What do the hydrocarbons from trees contribute to air
pollution? J. Air Pollut. Control Assoc. 22:537-543, 1972.
190. Gerhold, H. D., and G. H. Plank. Monoterpene variations in vapors from
white pine and hybrids. Phytochemistry 9:1393-1398, 1970.
i
191. Rasmussen, R. A. Isoprene: Identified as a forest-type emission to the
atmosphere. Environ. Sci. Technol. 4:667-671, 1970.
192. Went, F. W. The nature of aitken condensation nuclei in the atmosphere.
Proc. Nat. Acad. Sci. 51:1259-1267, 1964.
^ Rasmussen, R. A., and M. W. Holdren. Analyses of Cc to CIQ Hydrocarbons
in Rural Atmospheres. Paper No. 72-19 Presented at the 65th Annual
Meeting of the Air Pollution Control Association, Miami Beach, Florida,
June 18-22, 1972.
3-127
-------
194. Stephens, E. R., and W. E. Scott. Relative reactivity of various hydro-
carbons in polluted atmospheres. Proc. Amer. Petrol. Inst. 42(Section
III Refining):665-670, 1962.
195. Japar, S. M., C. H. Wu, and H. Niki. Rate constants for the gas phase
reaction of ozone with o^-pinene and terpinolene. Environ. Lett.
7:245-249, 1974.
196. Grimsrud, E. P., H. H. Westberg, and R. A. Rasmussen. Atmospheric reactivity
of monoterpene hydrocarbons, NO photooxidation and ozonolysis. Int.
J. Chem. Kinetics Symp. No. 1, 1975 (Chemical Kinetic Data for the Lower
and Upper Atmosphere):183-195, 1975.
197< Stasiuk, W. N., Jr., and P. E. Coffey. Rural and urban ozone relationships
in New York State. J. Air Pollut. Control Assoc. 24:564-568, 1974.
198. Research Triangle Institute. Investigation of High Ozone Concentration in
the Vincinity of Garrett County, Maryland and Preston County, West
Virginia. Institute Final Report. EPA-R4-73-019. Research Triangle
Park, N. C.: Research Triangle Institute, 1973. 105 pp.
199. U. S. Environmental Protection Agency. Office of Research and Development.
National Environmental Research Center. Health Consequences of Sulfur
Oxides: A Report from CHESS, 1970-71. Human Studies Laboratory.
Program Element 1A1005. Research Triangle Park, N. C.: U. S. Environ-
mental Protection Agency, 1974. 454 pp.
'200. French, J. G., V. Hasselblad, R. Johnson, and K. E. McClain. The Association
of Suspended Nitrates in Ambient Air with Aggravation of Asthma. Human
Studies Laboratory, National Environmental Research Center. Research
Triangle Park, N. C.: U. S. Environmental Protection Agency, 1975.
(UNVERIFIED)
3-128
-------
201. Barone, J., T. A. Cahill, R. G. Flocchini, and D. J. Shadoan. Visibility
Reduction: A Characterization of Three Urban Sites in California.
Air Quality Group, Crocker Nuclear Laboratory, University of California,
Davis, 1975. (UNVERIFIED)
202. Calvert, J. G. Modes of formation of the salts of sulfur and nitrogen in an
NOx-SC>2-hydrocarbon polluted atmosphere. Paper Presented at Conference
on Atmospheric Salts and Gases of Sulfur and Nitrogen in Association
with Photochemical Oxidant, University of California, Irvine, January
7-9, 1974. (UNVERIFIED)
203. Urone, P. Photochemical and thermal reactions of sulfur dioxide, pp. 505-
519. In Proceedings of the International Symposium on Air Pollution,
Tokyo, Japan, 1972. (UNVERIFIED)
204. Bufalirti, M. Oxidation 6f sulfur dioxide in polluted atmospheres - A review.
Environ. Sci. Technol. 5:685-700, 1971.
205. Sulfur in the Environment. Missouri Botanical Gardens, St. Louis, Missouri,
April 1975. (UNVERIFIED)
206. Calvert, J. G. Interactions of air pollutants, pp. 19-101. In National
Academy of Sciences — National Research Council. Assembly of Life
Sciences. Proceedings of the Conference on Health Effects of Air
Pollutants. U. S. Senate Committee Print Serial No. 93-15. Washington,
D. C.: U. S. Government Printing Office, 1973.
207. McAfee, J. M., J. N. Pitts, Jr., and A. M. Winer. In situ long path infra-
red spectroscopy of photochemical air pollutants in an environmental
chamber. Paper Presented at Pacific Conference on Chemistry and
Spectroscopy, San Francisco, Calif., October 16-18, 1974. (UNVERIFIED)
3-129
-------
208. Miller, D. F., and C. W. Spicer. A Continuous Analyzer for Detecting Nitric
Acid. Paper 74-17, Presented at 67th Annual Meeting of the Air Pollution
Control Association, Denver, Colorado, June 1974.
209. Kiang, C. S., D. Stauffer, and V. A. Mohnen. Possibilities for atmospheric
aerosol formation involving NH . Nature Phys. Sci. 244:53-54, 1973.
210. Gordon, R. J., and R. J. Bryan. Ammonium nitrate in airborne particles in
Los Angeles. Environ. Sci. Technol. 7:645-647, 1973.
211. Stephens, E. R. Identification of odors from cattle feed lots. Calif.
Agric. 25(1):10-11, 1971.
212. Jacobs, M. B. Table A-8. Concentrations of comon air pollutants in the
average urban area, p. 413. In The Chemical Analysis of Air Pollutants.
New York: Interscience Publishers, Inc., 1960.
2i3. Hanst, P. L., W. E. Wilson, Jr., R. K. Patterson, B. W. Gay, L. W. Chaney,
and C. S. Burton. A spectroscopic study of Pasadena smog. Division
of Environmental Chemistry Paper Presented at 167th National Meeting
of the American Chemical Society, Los Angeles, California, April 1,
1974. (UNVERIFIED)
214. Novakov, T.
Paper 74-200, Presented at 67th Annual Meeting of the Air Pollution
Control Association, Denver, Colorado, June 9-13, 1974. (UNVERIFIED)
215. Novakov, T., S. G. Chang, and A. B. Harker. Ozone concentrations in New Jersey
and New York: Statistical association with related variables. Science
186:259-261, 1974.
216. Yamamoto, K., M. Seki, and K. Kawazoe. Effect of sulfuric acid accumulation
on the rate of sulfur dioxide oxidation on activated carbon surface.
Nippon Kaguku Kaiski 7:1268-2179, 1973. (in Japanese, summary in English)
3-130
-------
217. Johnstone, H. F., and D. R. Coughanowr. Absorption of sulfur dioxide from
air. Oxidation of drops containing dissolved catalysts. Ind. Eng.
Chem. 50:1169-1172, 1958.
218. Scott> W. D., and P. V. Hobbs. The formation of sulfate in water droplets.
J. Atmos. Sci. 24:54-57, 1967.
219. Miller, J. M., and R. G. de Pena. The rate of sulfate ion formation in
water droplets in atmospheres with different partial pressures of SO
pp. 375-378. In H. M. Englund and W. T. Beery, Eds. Proceedings of
the Second International Clean Air Congress. Held at Washington, D. C.,
December 6-11, 1970. New York: Academic Press, 1971.
220. Psrtkett, S. A. Oxidation of S02 and other atmospheric gases by ozone in aqueous
solution. Nature Phys. Sci. 240:105-106, 1972.
221. Fuller, E. C., and R. H. Crist. The rate of oxidation of sulfite ions by
oxygen. J. Amer. Chem. Soc. 63:1644-1650, 1941,
222. Schroeter, L. C. Kinetics of air oxidation of sulfurous acid salts. J.
Pharm. Sci. 52:559-563, 1963.
223. Wood, W. P., A. U. Castelman, Jr., and I. N. Tang.
Paper 74-153, Presented at 67th Annual Meeting of the Air Pollution
Control Association, Denver, Colorado, June 9-13, 1974. (UNVERIFIED)
224. Davis, D. D., W.'A." Payne, and L.'Ji Stief. The hydroperoxyl radical in atmos-
pheric chemical dynamics: Reaction with carbon monoxide. Science 179:
280-282, 1973.
225 James, F. C., J. A. Kerr, and J. P. Simons. Direct measurement of the
rate of reaction of the methyl radical with sulphur dioxide. J.
Chem. Soc. Faraday Trans. I. 69:2124-2129, 1973.
3-131
-------
226. Gordon, S. , and E. Hamilton. Chem. Div., Argonne National laboratories,
Argonne, Illinois. Personal communication to J. G. Calvert, December
11, 1973. (see ref. 202)
227. Urone, P., and W. H. Schroeder. SCL in the atmosphere: A wealth of monitor-
ing data, but few reactions rate studies. Environ. Sci. Technol. 3:
436-445, 1969.
228 Bourbigot, Y., J. Bricard, G. Madeleine, and D. Vigla. Pollution atmospher-
izue.--IdentificatLon des aerosols produits par photolyse en presence
d'anhydride sulfureux. C. R. Acad. Sci. (Paris) C 276:547-550, 1973.
229. Daubendiek, R. L., and J. G. Calvert. A study of the N20 -S02-03 reaction
system. Environ. Lett. 8:103-116, 1975.
230. Dainton, F. S., and K. J. Ivin. The photochemical formation of sulphuric
acids from sulphur dioxide and hydrocarbons. Trans. Faraday Soc.
46:374-381, 1950.
231. Johnston, H. S., and K. Dev Jain. Sulfur dioxide sensitized photochemical
oxidation of hydrocarbons. Science 131:1523-1524, 1960.
232. Timmons, R. B. The photochemically induced reactions of sulfur dioxide with
alkanes and carbon monoxide. Photochem. Photobiol. 12:219-230, 1970.
233. Badcock, C. C., H. W. Sidebottom, J. G. Calvert, G. W. Reinhardt, and E. K.
Damon. Mechanism of the photolysis of sulfur dioxide-paraffin hydro-
carbon mixtures. J. Amer. Chem. Soc. 93:3115-3121, 1971.
234. Sidebottom, H. W., C. C. Badcock, G. E. Jackson, J. G. Calvert, G. W.
Reinhardt, and E. K. Damon. Photooxidation of sulfur dioxide.
Environ. Sci. Technol. 6:72-79, 1972.
3-132
-------
235. Penzhorn, R. D., W. G. Filby, K. Gunther, and L. Stieglitz.
pp. 16-36. In Chemische Reaktionen Atmos-
" 11
pharischer Schadstoffe. Institut fur Radiochemie Report KFK 1975 UP,
Karlsruhe, Germany, May 1974. (UNVERIFIED)
236. Penzhorn, R. D. , W. G. Filby, and H. Glisten.
pp. 3-15. In Chemische Reaktionen Atmospharischer Schadstoffe.
11
Institut fur Radiochemie Report KFK 1975 UF, Karlsruhe, Germany, May
1974. (UNVERIFIED)
237. Roberts, P. T., and S. K. Friedlander. Conversion of SO to sulfur partic-
ulate in the Los Angeles atmosphere. Environ. Health Perspect. 10:103-
108, 1975.
238. California Air Resources Board. Measurement of Sulfur Oxides in Auto
Exhaust. Project M-274, October 1974. (UNVERIFIED)
239. Begeman, C. R., M. L. Jackson, and G. J. Nebel. Sulfate emissions from
catalyst-equipped automobiles. Society of Automotive Engineers Paper
No. 741060, General Motors Corp., October 1974. (UNVERIFIED)
240. pierson, W. R., R. H. Hammerle, and H. W. Kummer. Sulfuric acid aerosol
emissions from catalyst-equipped engines. Society of Automotive
Engineers Paper No. 740287, Ford Motor Co., March 1974. (UNVERIFIED)
241. California Air Resources Board. Sources and Effect of Sulfates in the
Atmosphere. Informational Report March 6, 1975. (UNVERIFIED)
242. U. S. Environmental Protection Agency. Issue Paper: Estimated Public
Health Impact as a Result of Equipping Light-Duty Motor Vehicles
with Oxidation Catalysts. January 30, 1975. (UNVERIFIED)
3-133
-------
243. Morris, E. D., Jr., and H. Niki. Reactivity of hydroxyl radicals with
olefins. J. Phys. Chem. 75:3640-3641, 1970. (letter)
244. Furayama, S., R. Atkinson, A. J. Colussi, and R. J. Cvetanovic. Determina-
tion of the phase shift method of the absolute rate constants of reac-
Q
tions of 0( P) atoms; with olefins at 25°C. Int. J. Chem. Kinetics 6:
741-751, 1974.
3-134
-------
Chapter 4
ATMOSPHERIC CONCENTRATIONS OF PHOTOCHEMICAL OXIDANTS
The main purpose of this chapter is to survey atmospheric concentrations
of photochemical oxidants with emphasis on surface concentrations and the
distribution patterns associated with them. The reason for that emphasis
is that the photochemical oxidants that affect public health and welfare
are largely concentrated in this region. The whole subject of stratospheric
ozone (and its filtering of ultraviolet light and interactions with supersonic-
transport exhaust products), nuclear weapon reaction products, and halogenated
hydrocarbon decomposition products will not be treated here.
As in Air Quality Criteria for Photochemical Oxidants,1 our concern
will be with the broad subject of oxidant pollution in the atmosphere and
the net oxidizing ability of the contaminants in an air sample. The standard
corrections or adjustments to remove interferences from the data have already
been implied in the primary references in most cases. Therefore, we need
not refer to the early distinction between "oxidant" or "total oxidants" and
"corrected or adjusted oxidant."
It should be noted that there are still unresolved discrepancies in
oxidant data owing to differences in primary standards. lodometric cali-
bration techniques for ozone monitors were compared by an ad hoc committee
appointed by the California Air Resources Board (CARB)<>2 The committee set
out to find an accurate method for measuring ozone, to relate the recommended
method to earlier data, and to recommend procedures for calibrating ozone
monitors in the field. It examined calibration techniques used by the
-------
U. S. Environmental Protection Agency (EPA), the GARB, and the Los Angeles
County Air Pollution Control District and concluded that the Los Angeles
County method reads low by about 4% and that the GARB and EPA methods read
high by about 25 - 30%. An ultraviolet procedure was used as an absolute
standard, and a commercially available ultraviolet ozone monitor was cali-
brated to serve as a secondary standard. The committee stated that
multiplication of previous GARB ozone measurements by 0.78 and of previous
Los Angeles County Air Pollution Control District ozone measurements by
1.04 would yield results accurate to within 10%. (A seemingly disproportionate
amount of attention must still be given to measurements available from
California, because that data base continues to be the largest available.)
Ozone measurement as an indicator of the concentration of oxidants in
general is widely accepted. This is not to presuppose that ozone is the
cause of all oxidant-related health effects or damage, but that it serves
as a surrogate compound for a complex mixture of substances that are
characterized by their oxidizing ability and by the effects attributable
to this property. Consequently, oxidized organic compounds like aldehydes
and organic nitrates are also in the category of our subject matter,,
Implicit in the belief that ozone is an indicator compound is the assumption
that the concentrations, or at least the adverse effects of concentrations,
of nonozone oxidants bear some relationship to the ozone concentration.
This assumption is a measure of the inability to analyze for the other
compounds and of the sparse medical knowledge of their effects on human
beings. The use of ozone as a surrogate should be carefully reevaluated.
4-2
-------
Measurements of natural background photochemical oxidants are extremely
difficult to find in the literature. Measurements that once were thought
to represent the truly rural background are now suspected of having been
contaminated by pollutants of man-made origin. This constitutes a critical
question in the interpretation of atmospheric concentration, because some
of the ozone that is present is controllable and some is not. Some of the
literature cited in this chapter contains assertions that it will be
impossible in some urban areas to achieve the national ambient air quality
standard by local-source reductions, because "natural background" concen-
trations already exceed the standard. If these are not in fact natural
background concentrations, they may be reducible by limiting emission of
chiefly man-made precursor pollutants in upwind areas. Because the measure-
ments are incomplete, no specific section in this chapter will be devoted
to natural background concentrations of ozone; however, measurements that
are available will be cited and interpreted.
It is currently advocated that atmospheric concentrations of air pol-
lutants be expressed in micrograms per cubic meter. For ozone, the factor
for converting from parts per million to micrograms per cubic meter is 1,962
3
for 25 C and 760 mm Hg. That is, 1 ppm = 1,962 yg/m . In this chapter,
conversions are not always made, because of uncertainties as to the temper-
ature and pressure at which readings were taken0 These uncertainties become
especially severe under conditions of high altitude,, In this context, parts
per million is on a molar basis.
This review begins with a summary of the sources of monitoring data
operated primarily by public agencies. The spatial and temporal patterns
4-3
-------
of oxidant concentration are then discussed—urban versus rural and indoor
versus outdoor relationships, diurnal and seasonal patterns, and long-term
trendso The chapter includes brief discussions of photochemical oxidants
other than ozone and of data quality and concludes with a set of recommendations
for guidelines in future monitoring of atmospheric concentrations of ozone
and other photochemical oxidants.
SOURCES OF MONITORING DATA
One of the earliest organized efforts to acquire data on photochemical
oxidants was that of the Los Angeles County Air Pollution Control District,
which began in the middle 1950's and has produced the largest data base now
available for these studies. In 1961, the California Department of Public
Health set up a 16-station Statewide Cooperative Air Monitoring Network (SCAN).
The national network of air pollution measurements is keyed to the 247
air quality control regions (AQCR's), which were classified according to the
relative severity of their pollution problems.3 The classification is a
ranking of measured ambient air concentrations or the estimated air quality
in the area of maximal severity. The priorities for air quality problem
severity are as follows:
Priority I: Ambient concentrations significantly above primary
standards.
Priority la: Ambient concentrations significantly above
primary standards and due to emissions
from point sources.
4-4
-------
Priority II: Ambient concentrations significantly above secondary
standards.
Priority III: Ambient concentrations below secondary standards.
For photochemical oxidants, there were 55 AQCR's in the Priority I classifi-
cation and 192 AQCR's in the Priority III classification in October 1973.4
This indicates how many stations showed a severe oxidant air pollution problem.
A 1973 report issued by the EPA summarizes the history of the federal
air quality program by presenting a comprehensive analysis and interpretation
of data collected from federal, state, and local air quality surveillance
activities.5 Because of data-reduction lags, the report carries information
up to 1971, when 183 continuous oxidant monitoring stations were in operation.
It proposed 458 monitoring stations for 1974, compared with the 208 legally
required. Statistics presented indicate that nine AQCR's were exceeding the
primary oxidant standard in 1969; the number grew to 12 in 1970 and to 15 in
1971. Under the federal regulations, the states submit to the EPA on a
quarterly basis all the air quality data obtained from their monitoring
systems. The regional EPA offices edit the data for inconsistencies and
errors and then forward it for inclusion in the National Aerometric Data
Bank (NADB).
Of the six federal monitoring programs in operation when the report was
written, two (the National Air Surveillance Networks, NASN, and CAMP) were
analyzed for trends. The NASN monitors total suspended particulate matter
and sulfur dioxide at over 200 stations. Nitrogen dioxide is monitored at
some sites. Oxidant data obtained on a national basis are available from
the CAMP, which has been operating in six major urban areas for over 10 years.
4-5
-------
Data from the EPA's air monitoring networks described above and from networks
operated by state and local governments are submitted for storage in EPA's
Storage and Retrieval of Aerometric Data (SAROAD) format. Significant
quantities of the information from the states were not expected to be trans-
mitted until the summer of 1973, so the report had only limited information
in this category,, Extensive tabulations of air quality data and of the
monitoring installations are provided.
Other data in this chapter are from special monitoring programs and
from scattered reports from networks operated by public agencies. The special
programs generally are experimental and are designed to elucidate specific
features of the oxidant problem. The scattered data from public monitoring
networks have been obtained from a variety of international sources.
A DAMAGE INDEX FOR OXIDANT IMPACT
The health effects of photochemical oxidant pollution should be measured
by some kind of aggregate index that involves a weighting of pollutant con-
centrations according to spatial and temporal distribution. The following
sections there fore examine typical values of oxidant concentration as they depend
on location and time. The national ambient air quality standard concerns
itself simply with the worst place at the worst time. However, in the future,
damage associated with each specific effect should be incorporated in the
form of a summation or in an integral, such as
D = JJp (x,t)4> (c )dadt, (1)
ij j ij i
where D = index of damage due to the action of the ±th species on the j_th
ij
type of receptor,
4-6
-------
pj = population density of jth receptor,
$ = impact function for jLth species on jth receptor,
ij
c = concentration of jLth species,
i
x = location,
t = time, and
a = area.
This objective index goes beyond the simpler requirements now imposed by
ambient air standards; however, it can be related to ambient air standards
by normalizing the impact function. For example, suppose that the impact
of pollutant i on receptor j increases as the n±h power of c , so that the
i
expression
$ - (c /c ) (2)
ij i is
permits a comparative evaluation of all species if c is the ambient air
is
standard. Contributions can be aggregated to state the total damage on
receptor j by summing all the D over all i. These interactions of time
and space patterns with long-term health effects are discussed further in
Chapter 10. The application of this index6 to the analysis of entire systems
requires the spatial and temporal distribution of pollutants through the c ,
i
which implicitly depends on these variables. The sections that follow consider
some of the distributions that might enter this type of damage formula.
4-7
-------
SPATIAL PATTERNS OF OXIDANT CONCENTRATION
The early siting of monitoring stations in central business districts
overlooked the fact that nitric oxide emitted by primary polluters reacts
very rapidly with ozone to cause localized decreases in ozone concentration,,
Consequently, the ozone concentration is often higher in surburban and even
rural areas than in urban areas. Before examining detailed patterns within
a given urban area, it is important to look at the degrees of pollution
typically found in cities throughout the United States and elsewhere.
The original criteria document for photochemical oxidants1 included
tabulations of maximal oxidant concentrations and cumulative frequency
distributions of hourly average oxidant concentrations. Tables 4-1 and 4-2
show the results. The ranking of stations by yearly average concentration
differs from the ranking that might come about from using peak concentrations
only. This is born out by the cumulative frequency distribution: at the
high-frequency end, there is very little differentiation among the cities in
Table 4-2; but for the rare event, the cities are widely separated in con-
centration values. This must reflect the differences in mechanisms for
forming the ozone that is largely responsible for the oxidant readings.
This distinction may be drawn between the nonurban ozone formation processes
leading to low concentrations and the urban photooxidation processes leading
to the high concentrations of low frequency.
An international expert panel7 has issued an air quality criteria
document for photochemical oxidants and related hydrocarbons that builds on
the U.S. Department of Health, Education, and Welfare air quality criteria
document for photochemical oxidants. It discusses oxidant concentration
patterns in the context of the same tabular material presented earlier0
4-8
-------
vO
O
vO
O
vO
-*
O
rH
CO
O
m
oo
o
CM
0
m
o
CM
0
oo
CM
o
CM
o
CM
CM
0
rH
O
3 >> 60
8 rH Cd
•HUM
t J I-H H\ r~ '
X r3 CU p3
S o > ft
S rC Cd ft
vO
o
oo
in
o
00
CO
o
m
CM
o
m
CO
o
rH
CM
O
vO
CM
O
VO
CM
O
m
CM
o
CM
0
oo
i — i
o
CO
rH
O
n)
H
a) cu
rC
4J 'rj
•H CU
S rl -H
O <4H
CO -H
>-, O O
cd 4-1 CU
T^ O-
rH CO
rH Cd
cd 3 C
4-1 tr o
O CU -H
cu cd
MH 60 M
o cd 4J
rl C
4J 0) CU
C > 0
cu cd rJ
0 O
0) rH
ft M (3
3 cd
*"O O »C
PJ rC 4J
cd
£3 M
rl 3 CU
Q) H 4J
rQ -H cd
3 cti M
s d £jf
I
ft
m
o
•
o
s
ft
ft
o
rH
•
o
1
ft
m
rH
•
o
4J CO
PJ >,
(ll rrt
vy \\i
0 T3
0) <4H
P-l O
CO
>,
O
4-1 CO
CS >i
n) rrl
\U IU
0 Q
M
cu MH
PM o
&
cd
/^
4-1 CO
C K*^
cu cd
O T3
!H
CU 14H
PM o
fc
P?
^3 ^3 \^
iO -^ ^D
r~*» r**- r**-
vO O O
*^" "si" *^~
m in st
t— i m o>
m oo o
m o in
CM
1^ vO
VO CM
CM vO
CM CO
0"V rH
f^ CO
rH rH
rH ON
m m
ON -vj"
• •
•. n)
Cd rH
T3 -rl
cd
rH >
cd cd cd
•U 4J
o >w td
HOT)
oo
CM
o
CO
co
CM
in CM
oo oo
CM m
m
m
CO
rH
VO
CO
CM
m
vo
o
CO
in
M
cd
M
•s
oc
•H
00
CU
cd
•U
CO
CU
CO
cd
PM
CO
cu
cu
60
O
60
CU
C
cd
CO
rl
cu
c
cu
Q
co
•H
3
O
W
cd
•H
ft
rH
0)
-d
cd
0
cu
|
o
cd
cd
C
pi
•H
a
•S
cd
J-J
cd
rQ
cd
cd
4J
c
cd
Q
A
P!
O
60
a
•H
A
CO
cd
o
o
CO
•H
o
PJ
cd
pTj
fi
cd
CO
o
60
cd
CJ
•H
rC
o
cd
4J
cd
CO
4J
C
i
4-9
-------
CM
td
H
a) m
oo vo
cu I
«3j VO
r^^H
3 co-i.
O 0) J3f
33 i-l •
] t [V]
14H -H •
O O ill
a TS "4-i
O CU O
•H 4-1
4-> O 4J
3 0) ti
J3 rH 0)
7H 0) g
r-l CO 4J
4J r-l
co C cd
•H -H O,
O 0)
CO Q
?-> C
!H
0)
a I
0) }-l
U 14H
O
•H O
4J
Cd 4-1
rH C
3 Cd
E *"U
3 -H
O X
>-. cd a
CM
CO
C
•H
r-l CU
4-1 4-1
C3 cd
01 4-1
a co
o c!
o cd
H**
'is cu a
4-1
ca cd
)H 0)
3 r-i co
O 60 C
rC O
r-l -H
<4H O 4->
O cd
O V-i
4-1 4J 4J
C C
CU rH CU
o cd
o
CO
o
m
*
0
CM
r-H
o
o
i-H
o
oo
o
0
o
o
CO
o
o
CM
o
o
o
i-H
0
0
vO
CO
o
o
CN
0
0
i-H
O
oo
o
o
vo
o
o
-3-
O
o
CO
o
o
CM
O
O
i-H
O
0
0
i-H
CO
0
O
i-H
O
(Ti
O
O
O
O
O
O
o
o
CO
o
o
CM
O
O
i-H
O
O
vO
CM
O
O
1
4J
•H
CJ
td
rt
0)
*t3
cd
ca
cd
[0
01
rH
01
H
"^
CO
,3
o
60
01
•H
P
£
cd
CO
«
o>
a
cu
Q
CO
•H
3
O
•
4-1
CO
P.
P-i
O
4-1
G
0)
1-1
o
cd
CO
cd
C
u
cd
pq
cd
4J
$
CO
o
4J
60
|
co
JJ
o
o
ca
•H
o
C
cd
cd
co
o
60
cd
o
•H
U
cd
4J
cd
MH
o
ca
C
i
4-10
-------
New information is added for the city of Delft:* The monthly means of daily
maximums of hourly ozone concentrations are shown in Table 4—3, and the
monthly average ozone concentrations are shown in Table 4-4. As in other
cities, the worst month seems to be August, with a mean daily maximum (of
3
hourly concentrations) of 0.071 ppm (140 yg/m ). Table 4-5 compares the
number of days in May through July of 1971 when the ozone concentration at
3
one or more sites reached or exceeded the hourly average of 200 yg/m or of
3
100 yg/m in ref. 7. A comparison is made between Delft and five other
monitoring sites in the Netherlands. Amsterdam had a peak value of 0.18 ppm
in March 1971.
Measurements from a downtown site in Frankfurt-am-Main and another on
a nearby 800-m mountain, Kleiner Feldberg, both in the Federal Republic of
Germany, are plotted in Figure 4-1. As opposed to the pattern in New York
State,9'10 the ozone concentration at the urban sampling site was far greater
than that observed at the mountain site,. The peak in Frankfurt lasts from
10 a.m. to 6 p.m. Readings from two stations in Berlin are compared in
Figure 4-2. One station, Steglitz, is in an area with high traffic; the
other, Dahlem, is in the suburbs. It was concluded that the lower oxidant
concentrations at the urban station were due to the high concentrations of
hydrocarbon there. In reality, they were probably caused by the high con-
centration of nitric oxide that reacted with the ozone. For both cases, the
concentrations of ozone were extremely low, with a maximum of 0..0132 ppm at
Dahlem.
* The Delft data were provided to the NATO/CCMS Panel on Air Quality Criteria
by L. J. Brasser of the Research Institute for Public Health Engineering,
Delft, the Netherlands, in April 1973.
4-11
-------
Table 4-3
Monthly Means of Daily Maximums of
Hourly Ozone Concentrations in Delft, 1970a
Month
January
February
March
April
May
June
July
August
September
October
November
December
Ozone concentration
3
4
23
90
104
116
121
88
140
113
58
46
29
Ppm
0.002
0.012
0.046
0.053
0.059
00062
0.045
0.071
0.058
0.030
0.023
0.015
a Reprinted with permission from ref. 7.
4-12
-------
Table 4-4
Monthly Averages of Ozone Concentrations in Delft, 197Oa
Ozone concentration
Month
January
February
March
April
May
June
July
August
September
October
November
December
3
yg/m
2
13
50
58
48
45
46
61
50
32
25
10
ppm
0 = 001
0.007
Oo026
0.030
0.024
0.023
0.023
0.031
Oo026
00016
0.013
0.005
a Reprinted with permission from ref. 7.
4-13
-------
r-
c^
.—i
I
•
> (3
cd cu<
o
H4
O be
O 3.
O
N
O
l"1
I
9
TJ
C
O
CU -H
bO JJ
cd cd
CU 4-1
> c
cd CUro
o _e
rH O bt
rJ O 3-
O CU
O
N
O
-I
o T
d
TI cd
-4 Tj
O
1—I
o\o
i-ICMCSJCO rHOICOCOt-H^-IC-J
c
o
CU iH
60 4J
cd cd
M r4
CU 4J
> a
cd CU ro
0 B
>> e-^.
rH O M
r^ O 3-
3
O CU
43 C
O
N
O
rQ
S-+1
O
IS & &
en co co co
0) OJ CU CU
M toO 60 M
cd cd ctj cd
M t-l n ^
CO OJ CU CU
cd cd cd cd
i-l rH H
O O O O
CM O 00 CT\
CM CO > > >
ctf cd n3 ttJ
CN
4-14
-------
•x.
o
Ul
g
g
OFRANKFURT. JULY 27.1967
"T3KLEINER FELDBERG. JULY 27,1967
aFRANKFURT, JULY 28, 1967
C-KLEINER FELDBERG, JULY 28,1967
1800 2000 2200 2400
Figure 4-1. Diurnal variation in ozone concentration in Frankfurt-am-Main
and at Kleiner Feldberg mountain station on July 27 and 28,
1967. (Reprinted with permission from Weber.8)
4-15
-------
w-
0.6-
E
O.
CL '0,6-
«> *
y 0.4-
8 0.2-
» 12 Vi 16
20 22 24
HOUR
Figure A-2. Diurnal variation in ozone concentration at two sites in
Berlin, 1966-1967. (Reprinted with permission from
Lahmann et al.11)
4-16
-------
The NATO report also supplied ozone information for Amsterdam, Univer-
sity City in Rome, and Ankara, Turkey. These data are shown in Figures 4-3,
4-4, and 4-5.
An air pollution episode in Windsor, Ontario, was recorded on August
17-19, 1971.15 Figure 4-6 shows total-oxidant peaks ranging from 0.12 to
0.20 ppm and comparable peaks for the oxides of nitrogen. An interesting
feature of these measurements is the relationship between oxidants and
hydrocarbon. This interpretation must not be taken too seriously, because
of the large quantity of relatively unreactive material in the total-
hydrocarbon reading. Figure 4-7 refers to a pollution episode at Tunney's
Pasture in Ottawa in 1973 in which oxidant concentrations approached 0.14
ppm. This episode took place on a July weekend, and the nitrogen dioxide
and NO concentrations appear to be very small. Nonmethane hydrocarbons,
x
however, were relatively high, at about 0.6 ppm.
Bilger has documented ozone and other oxidant measurements in Australia16
and compared them with those in other cities. Table 4-6 shows the fraction
of hours during which threshold concentrations were exceeded in Sydney in
1971, 1972, and 1973. Low concentrations were recorded before 1970,
probably because of the proximity to nitric oxide emission sources in the
central portion of the city. It is also noteworthy that, despite the calm
conditions and strong inversion in the winter months (in the southern
hemisphere), high-oxidant days were relatively infrequent0 Figure 4-8
outlines this seasonal variation in oxidant concentration. In Melbourne,
Australia, however, a high-ozone episode was observed during December 1973»
The diurnal concentration variations shown in Figure 4-9 illustrate the episode
between December 10 and December 13, with a peak of 0.28 ppm on December 11,
the second day of the episode. Relatively low levels were observed in other cities,
4-17
-------
0.20
0.15
IU
1
0.10
O.Q5
0200 0400 0800 0600 1000 1200 1400 1600 1800 2000 2200 2400
TIME, hr
Figure 4-3. Ozone concentrations in Amsterdam, March 10, 1971. (Reprinted
with permission from NATO.7)
4-18
-------
0.06
i
0.05
0.04
e
5
o
§ 0.03
«t
=j 0.02
_l
0.01
0
i i i i r
N02
NO
03
ALDEHYDES (HCHO)
HYDROCARBONS-CH4
o
0600 0700 0809 09CO 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900
TIME, hr
Figure 4-4. Nitrogen dioxide, nitric oxide, ozone, aldehyde, and hydro-
carbon concentrations in University City in Rome. (Reprinted
with permission from NATO.7)
4-19
-------
2400
Figure 4-5. Diurnal variation in ozone concentration in Ankara, Turkey,
on selected days of 1972. (Reprinted with permission from
NATO.7)
4-20
-------
Total oxidanfcs and nitrogen oxides ppb
O
O
tn
O
O
O
era
cH
o
*••
CO-
•-* ^
CO |0
to.
o
oo
CO
CO
KJ
O
I
CO
Total hydrocarbons ppm
Figure 4-6. Oxidant, nitrogen oxide, and hydrocarbon concentrations in pollution
episode in Windsor. (Reprinted from Ref. 15.)
4-21
-------
C3,N02 and NOX
CQ so'
S^>
1
>i
0
o
I
t*
o
o
" • —
I
o«
o
o
— : — -«•
i
C3
0
o
I
_4
o
o
0
Non-methane hydrocarbons ppb
Figure 4-7. Ozone, nitrogen dioxide, NOX, and nonmethane hydrocarbon con-
centrations in pollution episode at Tunney's Pasture 1973.
(Reprinted from Ref. 15.)
4-22
-------
Table 4-6
Oxidant Concentrations in Sydney*2
Year
Portion of Hours with Concen-
tration, %
>0.05 ppm >0.10 ppm >0.15 ppm
Annual
Maximal 1-hr
Average,
ppm
1971
A.I
0.2
0.01
0.17
1972
5.6
0.9
0.2
0.24
1973
6.1
0.9
0.3
0.28
a Derived from Bilger.
16
4-23
-------
10
1
I10
GL
in
UJ
i
£ 5
g '
tr
UJ
Q.
J
P x 1971
o 1972
D 1973
o
I •
X
a x
/ \ q
/ \> o.o
^,^D \ * ^^.. — L
\ [y
1 " « *\° L
\ /O
\ / x
A / °
ft m^
' fcJ% ^^
X
o
X x
- F M A M J J ASONO
Figure 4-8. Seasonal variations in oxidant concentration in Sydney,
(Reprinted with permission from Bilger. )
4-24
-------
0.30
0.20 -
a.
a.
I
o
o
o
I
TUES 11 DEC
WED 12 DEC
xTHURS 13 DEC
HCXJR OF DAY - EASTERN SUMMER TIME
Figure 4-9. Diurnal variation in ozone concentration during December 1973
at Parliament Place, Melbourne—a 3-min average on the hour.
(Reprinted with permission from Bilger.16)
4-25
-------
In addition to the analysis of oxidant problems in the United States
and Australia, the Organization for Economic Cooperation and Development's
Air Management Sector Group17 has compiled oxidant concentrations measured
in Japan during 1970 - 1972. Table 4-7 summarizes maximal hourly oxidant
concentrations and the frequencies of concentrations of 0.15 ppm or more.
The figures show that, despite the high latitude, rather large maximal
hourly concentrations are observed in several of the cities, with the highest
being 0.38 ppm in Tokyo in 1972, during which year the oxidant concentration
reached or exceeded 0.15 ppm on 25 days.
Another high-latitude area where substantial ozone concentrations have
been observed is London. Derwent and Stewart17a reported ozone concentrations
in excess of 0.1 ppm. These were recorded at Harwell, Brookshire, on July
3, and July 7, 1971, by Atkins et^ al.17^ The main subject of Derwent and
Stewart's paper, however, deals with measurements in central London. Figure
4-9a (from Derwent and Stewart173) shows the time variation of photochemically
active pollutants for 3 days in central London. On the second day, the late
afternoon peak exceeded 0.1 ppm. It is apparent that some conversion of
nitric oxide to nitrogen dioxide was responsible for the ozone buildups,
but it is not as clear a chemical pattern in these results as it is for cities
in the western United States. The paper did not mention the method of ozone
measurement; however, qualitiitive descriptions of the weather patterns suggest
that the days of high ozone were characterized by light winds and considerable
sunlight.
Coupled closely with the effect causing horizontal distributions are the
vertical distributions of ozone concentrations. These distributions have an
1 Q
intimate influence on the urban-rural interchange of ozone. Miller and Ahrens
4-26
-------
Table 4-7
a
Oxidant Concentrations in Japan, 1970-197217
Maximal Hourly Oxidant Concentration, ppm
City 1969 1970
Tokyo 0.27 0.34
Osaka 0.26 0.24
Kanagawa 0.21
Chiba
Saitama
Aichi
Hyogo
No. Days in 1972 with Oxidant Concentration a. 0.15 ppm
1971
0.29
0.24
0.36
0.25
0.32
0.165
0.21
1972
0.38
0.29
0.33
0.32
0.28
0.20
0.24
Tokyo district: an average of 25
Osaka district: an average of 11
Ise district: an average of 5
Ibaragi: 16
Okayama: 3
Ehime: 3
a Derived from OECD.17
4-27
-------
"08—ft" 16"" 20"" 00 0408 12 16 20 00 04 08 12 16 20 00
July 12,1972 I July 13,1972 ! July 14,1972
Time of day
Figure 4-9a.
Diurnal variations of air pollutants measured in London from
July 12 to July 14, 1972. • , Ozone, p.p.b.; •, nitric
oxide, p.p.b.; CD, nitrogen dioxide, p.p.b.; O, hydrocarbons,
p.p.m. (Reprinted with permission from Derwent and Stewart.17a)
4-28
-------
presented detailed vertical time and space cross sections of ozone concen-
trations at altitudes up to 2,500 m. A low-altitude temperature inversion
may actually lead to lower concentrations of oxidant, because the destruction
rate can be increased by the injection of nitric oxide into a shallow mixing
layer. Very high oxidant concentrations were observed at the edge of the
marine inversion, where the mixing layer is rather deep. Possible high-
altitude transport of ozone within the inversion layer and absence of de-
struction mechanisms could lead to a buildup alofto Later fumigation caused
by surface heating could cause increased ozone concentrations at downwind
locations. Horizontal spatial patterns were mapped out with aircraft
measurements to indicate the connection between transport from areas where
precursor emission occurs and the downwind buildup of ozone after the photo-
chemical reaction.
The compensating factors leading to increased daily maximal oxidant
concentration over the years include both decreased destruction rates and
increased precursor concentration. The decreased destruction rates theoretically
result from urbanization, which tends to destabilize the air and increase
vertical mixing. Thus, the very mechanism that leads to greater dispersion
also leads to higher ozone concentration. The ozone aloft that comes from
polluted air that is transported upward into the inversion has a relatively
long lifetime, compared with that at low altitudes.
To determine the relationship of ozone concentrations in the Sierra
Nevada mountains with those in the central valley of California, airborne
and ground-station measurements were taken of total oxidant, temperatures,
and wind from August 17 to August 27, 1970. Miller et_ al.19 report the
results in a paper that coordinates the measurements on the basis of ozone
4-29
-------
transport,, Figure 4-10 shows concentrations equaling or approaching the
national ambient air quality standard for oxidants at stations at 351 and
2,287 m. Comparison with concentrations in an urban center in the central
valley, with an altitude of 99 m, shows that the peak concentrations at these
very different elevations roughly approximate one another during the after-
noon and early evening. At the urban location, however, nighttime values
decrease dramatically, probably because of reaction with nitric oxide emitted
locallyo These results suggest that the ozone-laden layer moves along the
terrain, rather than touching the surface at some altitude near the mixing
depth. A highly plausible transport mechanism is found in the afternoon
up-slope flows induced by surface heating. Aircraft profile measurements
at Mineral King (2,287 m) indicated that both temperature and oxidant increased
slightly near the surface. The evidence suggests that photochemical oxidant
is transported from population centers in the central valley into the higher-
altitude mountain valleys. The primary pollutants (oxides of nitrogen and
hydrocarbons) may have been the specific pollutants advected (i.e., moved
horizontally), and the abundant light in the ultraviolet range at the higher
altitudes could have been significant in increasing ozone production during
the early morning.
Blumenthal et_ al^.20 obtained aircraft data by taking vertical soundings
over various airports in the Los Angeles basin area. Figure 4-11 shows the
results for a vertical sounding over Hawthorne Airport at 9:15 a.m., on
September 20, 1972. The vertical distributions of condensation nuclei and
temperature suggest a mixing layer approximately 800 ft deep. At the same
altitude, there is a decided increase in ozone concentration that persists
4-30
-------
Hammond CDF Station 351 m.
Mineral King
Guard Station 2287 m.
Fresno Court House 99m
08 10 IZ
Hour* PST
24 •
Figure 4-10.
Comparison of 10-day means of hourly total oxidant
concentrations at Fresno, Hammond, and Mineral King,
California. (Reprinted with permission from Miller
et al.19)
4-31
-------
3-
|2.
|
I1'
0
(
C
_
C
{
c
Fig
M * * T VT
K I /
i\ \ {
•< M V "
X N. >> T
Vh >
f * * 1 H
c «' ^i^/ j i
f X^^Tx
) 0.1 0.2 „ 0.3
Ozone, ppm
) 10 20 30 40 50 60
Relative humidity. %
SO 5 10 15 20 25
Temperature, °C
1 1 2 3 4 5 _4 ., 6
) 5 10 15 20 25 30
Carbon monoxide, ppm
Legend
= Ozone
= Humidity
- Temperature
= Carbon monoxide
'Condensation nuclei
H K^
0.4 0.5
70 80 90 100
30 35 40 45
7 8 9 10
35 40 45 50
1 0.2 0.4 O.'e 0.8 1.0
Condensation nuclei/cm3 ilO5
m
ure 4-11. Data from vertical sounding over Hawthorne Airpc
at 9:15 a.m., September 20, 1972. (Reprinted \n
permission from Blumenthal et al. )
-------
up to a few thousand feet in altitude. This morning profile of ozone up
into the inversion layer corroborates other findings of high ozone concen-
trations in regions that are not well mixed with nitric oxide from the ground.
The profiles of carbon monoxide concentration and b (a measurement of
scat
light scattering in air) also indicate the location of the mixing layer.
Blumenthal and co-workers21 used the earlier data and a case study of
Denver to develop arguments as to the source of high concentrations of ozone
or ozone precursors that are found in some nonurban areas. They believe that
downwind areas as long as 260 km can exceed the standard because of precursor
emission from an urban source. Although they do not present any ozone pro-
duction estimates related to photochemistry, their data analyses confirm
hypotheses of transport from urban to nonurban areas. They point out that
rural areas have had concentrations as high as 0.3 ppm with no local source
of reactant.
The Los Angeles Reactive Pollutant Program (LARPP) obtained three-
dimensional distributions of ozone by using helicopters and ground-based
instruments. The instrument platforms were directed to follow an air parcel
through the Los Angeles basin by means of radio commands based on the radar
tracking of tetroon* arrays that were released below the inversion base and
allowed to float along with the air. Preliminary plots of these data22 show
that each vertical distribution of ozone is very nonuniform during the morning,
but becomes very uniform toward midday as the values increase. This behavior
is shown in Figure 4-12. Vertical profiles of temperature, dew point, and
ozone, shown in Figure 4-13, illustrate a case in which the ozone concentration
* A neutral buoyancy balloon that follows winds along constant density
contours.
4-33
-------
4-34
-------
4
1-1
E
«-«
ct
n
o
6
•3
I
-75
O
O
t
. !
\l
U
\
\
I
80
85"
10
\
.2
20
.3
30
90
.4
40 °F
i
Figure 4-13,
Vertical temperature, due point, and ozone sounding for
operation 14, Glendale, October 4, 1973, 1:01-1;25 p.m.
C Reprinted with, permission from Edinger.°°"
22>
4-35
-------
in the inversion layer 1,000-2,000 ft above mean sea level was approximately
3 times the concentration near the ground. This finding lends further support
to the hypotheses advanced by others regarding the horizontal transport of
ozone at high altitudes.
In rural areas of western Maryland and West Virginia, hourly ozone con-
centrations as high as 0.12 ppm were recorded23 in August and September 1972.
The concentration exceeded the national ambient air quality standard of 0.08
ppm 11% of the time. During one episode, this standard was exceeded for 26
consecutive hours. Precursors were always near background concentrations,
and the horizontal spread of ozone was extensive„ After mentioning strato-
spheric sources and ground sinks, the report hypothesizes that clouds of
precursors emitted from distant urban areas undergo transformation to ozone
during their long journey. Mechanical turbulence is suggested as a transport
mechanism that brings the resulting ozone to the surface.
Airborne measurements made for transects across the Los Angeles basin
by Edinger2Lf confirm the earlier hypotheses that ozone concentrations go
through a spatial maximum above the base of the inversion in many situations.
Figures 4—14 and 4-15 show space sections of the potential temperature and
the oxidant concentrations in a vertical cross section running approximately
west to east from the Pacific cost to Rialto, California. These results are
for June 20, 1970, at 4:30 p.m. The temperature field exhibits a distinct
inversion base that roughly follows the terrain, but increases slightly
moving inland—which is typical of the heating effect that occurs during
onshore flows. The oxidant field corresponding to it shows peaks in excess
of 0.35 ppm at altitudes approaching 1,000 m. At the same time, ground con-
centrations gradually increase to a maximum of 0.25 ppm 60 miles onshore„
4-36
-------
MILES INLAND
Figure 4-14.
Field of potential temperature (°C) in the vertical
cross section from Santa Monica to Rialto-Miro, 4:30
p.m., June 20, 1970. (Reprinted with permission from
Edinger.21*)
4-37
-------
MILES INLAND
JO 40 JO
MILES INLAND
6O 70
Figure 4-15.
Field of oxidant concentrations (ppm) in the vertical
cross section from Santa Monica to Rialto-Miro, 4:30
p.m., June 20, 1970. (Reprinted with permission from
Edinger.24)
4-38
-------
These data suggest that the oxldant within the inversion layer has been
advected into that layer from horizontal injections in the vicinity of heated
mountain slopes. This explanation is preferred to vertical mixing, because
of the extreme stability within the inversion layer. In fact, that stability
insulates the high-ozone air parcel from nitric oxide, which could react
rapidly with the ozone and thereby decrease it. Entrapment in the inversion
layer also suggests an explanation for long-range horizontal transport that
can ultimately lead to the large nonurban ozone concentrations that have been
reported.
Ozone and ozone precursor concentrations at nonurban locations in the
eastern United States were studied extensively.25 The three parts of the
study were field measurements, a quality assurance program, and an airborne
monitoring program. The main objective of the study was to establish a
data base for nonurban ozone and precursor concentrations0 Simultaneous
statistical summaries of the concentrations of nitrogen dioxide and nonmethane
hydrocarbons were also provided. Another objective was to search for rela-
tionships between ozone concentrations and nitrogen dioxide and nonmethane
hydrocarbon concentrations.
Monitoring sites were selected in northwestern Pennsylvania, central
Ohio, western Maryland, and southern West Virginia. The field program was
conducted between June and September of 1973. Ozone was measured at each
of the four sites, and precursor concentrations were measured at three of
them. The diurnal patterns of ozone concentration were not similar to the
rural site measurements reported for New York State.9'10 Instead of a gentle
ozone peak just after midnight, there was a late afternoon peak that was
4-39
-------
several times the minimal concentration of ozone, which generally occurred
early in the morning. On selected dates, however, nighttime ozone concen-
tration maximums were observed at the Maryland and Ohio sites. It was noted
that the National Ambient Air Quality Standard for photochemical oxidants
was exceeded during 37, 30, 2:0, and 15% of the hours at the sites in Maryland,
Pennsylvania, Ohio, and West Virginia, respectively, during the sampling
interval. Correlation coefficients among hourly ozone concentrations measured
at the four stations at the same time were in the range of 0.468-0.678.
Evidently, high ozone concentrations at the nonurban locations were spread
over extensive areas of the four-state region.
The quality assurance phase of the program was conducted to maximize
the comparability of the data generated at the four monitoring sites, as well
as data from the mobile laboratory. Real-time data processing permitted
quick analysis of the performance of the analyzers. Because the nitrogen
dioxide values were very low and the hydrocarbon data from the fixed sites
were declared invalid, only ozone was subjected to statistical analysis»
The bias of measurements between fixed- and mobile-site analyzers was estimated.
The greatest bias observed during the program, 23%, was attributed to span
drift of one of the analyzers.
The airborne monitoring program concentrated on the measurement of ozone
to provide supplementary air quality data for various altitudes over the fixed
sites. The airborne measurements were conducted in a C-45 aircraft that
carried a solid-face chemiluminescent ozone monitor. The ozone meter was
cycled every 2 min to provide calibration, purge, measurement, and purge at
equal intervals. The sparseness of the airborne data precluded detailed
analysis, but the comparisons between ozone aloft and ozone at the ground
4-40
-------
stations showed a high degree of regularity. The airborne values were
comparable with those read on the ground 100 - 600 ft below. A general con-
clusion on the vertical data is that the ozone generated from ground-source
precursors is predominant over ozone that might be transported from the
stratosphere.
Gloria et a.L.26 studied photochemical air pollution with an instrumented
aircraft in various air bases in California. They estimated background ozone
concentrations for the southern coast air basin to be about 0.03 ppm. Ozone
was measured by ultraviolet absorption with a Dasibi ozone photometer. Total
oxidant, carbon monoxide, oxides of nitrogen, hydrocarbons, dew point, and
temperature were measured simultaneously. East-west ozone sections were
made from series of flights following survey patterns mapped on Figure 4-16.
Figure 4-17 shows the vertical cross section of ozone isopleths for August
11, 1971, indicating higher ozone concentrations next to the surface than
up near the inversion base. Layering of ozone at higher altitudes is shown
in the profiles measured at Riverside, California, on August 10-12, 1971.
These profiles, illustrated in Figure 4-18, show the entrapment of high ozone
concentrations far above the surface in strata. Similar ozone layering was
observed in the San Francisco Bay area during the same study.
Atmospheric ozone concentration was measured simultaneously at two rural
and four urban sites in New York State.9 It was noted that during the period
August 1-17, 1973, the average hourly ozone concentrations at the rural sites
stayed well above the urban concentrations throughout each day and that the
urban values peaked at approximately the rural values in the early afternoon
(see Figure 4-19). The average rural ozone concentration for the period of
observations was around 0.05-0.07 ppm. The authors concluded that violations
4-41
-------
SAM KRNARDINO
COUNTY
SOUTH COAST AIR BASIN
1.500 ft contour
SAN oteso
COUNTY
Figure 4-16.
Outline map of the Los Angeles basin area traversed
during survey flights, showing general flight paths,
(Reprinted with permission from Gloria e_t al.26)
4-42
-------
10 x 10s
<0.03 ppm
7/ff/ff/M
0.03-0.06/////\
-------
10X103
—o—Aug. 10. 1971; 5:00 pm PST
"0-- Aug. 11. 1971; 4.15 pm PST
-• - Aug. 12, 1971; 5:00 pm PST
.04 .08 .12 .16 .20 .24 60 80 100
t>3 concentration, ppm Temperature, °F
Figure 4-18. Ozone profiles, Riverside, Calif., August 10-12, 1971.
(Reprinted with permission from Gloria et al.26)
4-44
-------
6
5
ij
2
1
• Kingston
- Rensselaer
• Glens Falls
• New York City
• Utsayantha
• Whiteface
AVERAGE HOURLY OZONE VALUES
8/1/73 to 8/17/73
2 4 6 8 10 12 14 16 18 20 22 24
Hours
Figure 4-19.
Average hourly ozone concentrations during
August 1-17, 1973, at selected sites in
New York State. (Reprinted with permission
from Stasiuk and Coffey.9)
4-45
-------
of the present ambient air quality standard for photochemical oxidants may
not be prevented completely by reducing anthropogenic hydrocarbon emission
in the state. They compared the weight equivalent of 0.24-ppm hydrocarbon
and 0.08-ppm ozone. These numbers were derived from the air quality standards
and suggested that one weight of nonmethane hydrocarbon is capable of generating
an equivalent weight of ozone photochemically. The questionable nature of
these assumptions is brought out by a comparison of the ozone potential for
all anthropogenic emission of hydrocarbon in the nation with the influx of
ozone by advection into New York State. The two ozone values are approximately
equivalent. Clearly, the photochemistry of nitrogen oxides must be included
in such estimates. Explanations of the relatively high rural ozone concen-
trations center on speculation based on the subsidence of stratospheric ozone
or on ozone generation by the photooxidation of natural precursors. It is
argued that transport of ozone from other urban areas and later reactions
with further precursors do not appear to constitute the dominant mechanism.
These arguments apparently were based on the assumed weight equivalency
between hydrocarbon precursors and ozone in the atmosphere,,
The authors later continued the disucssion with the same New York State
data.10 There were close correlations between rural ozone concentrations
and some of the urban peak values for the August 1-17, 1973, period, as shown
in Figure 4-200 Time correspondences between maximal wind and maximal ozone
concentration are cited as evidence to support the hypothesis that high urban
ozone concentrations in New York State result from the transport of background
ozone into these areas. Correlation coefficients of 0.83 for wind and tem-
perature, 0.81 for wind and ozone, and 0.87 for ozone and temperature were
obtained for readings in Kingston, New York. The correlations for the Welfare
4-46
-------
Figure 4-20.
Close correlations between the whiteface ozone concentrations
from August 1-17, 1973, and the smooth curve drawn between the
Glens Falls and New York City daily ozone maximums. (Reprinted
with permission from Coffey and Stasiuk.10)
4-47
-------
Island monitoring site in New York City were not nearly as clear as the ones
for Kingston. It was concluded that the correlation of photochemical ozone
in one area with that in another does not necessarily indicate transport.
An alternative mechanism was based on a belief that the observed correlations
imply a common ozone source. It was suggested that ozone concentrations
exceeding the federal ambient air quality standard exist in rural areas and
are transported into urban areas, but no direct evidence is offered beyond
the concentration graphs.
Air breakdown due to coronal discharges around high-voltage transmission
lines has been considered as a possible rural source of ozone. Several
investigations25>27~3Q suggest that this cannot be a significant source.
In a paper presented at the EPA scientific seminar on automotive pol-
lutants,31 Pitts pointed out that, until the controversy on calibration
methods was resolved, most people believed that the urban plume moving eastward
from Los Angeles was accompanied by increasing ozone concentrations. This
is illustrated in Figure 4-21 for downtown Los Angeles, Pasadena, Pomona,
Azusa, Riverside, and San Bernardino. Figures 4-21 through 4-23 all show the
numbers of days on which the maximal hourly average oxidant concentration
equaled or exceeded 0020 ppm in 1973. Pitts assumed that all data were
adjusted to the Los Angeles Air Pollution Control District (APCD) calibration
method by multiplying the nori-Los Angeles (i.e., Riverside and San Bernardino)
data by 5/7„ Figure 4-22 shows that this procedure lowers the numbers for
the two easternmost stations relative to those for the eastern portion of
Los Angeles County (stations 2, 3, and 4—Pasadena, Pomona, and Azusa).
Proceeding to another alternative, Pitts multiplied the numbers for stations
1-4 by 7/5 as a way of converting Los Angeles APCD data to the ARE calibration
4-48
-------
140
120
or
2 100
an
LU 80
Q_
60
40
20
LOS ANGELES
PASADENA
POMONA
AZUSA
RIVERSIDE
SAN BERNARDINO
PRESENT METHOD
( DATA AS REPORTED )
Figure 4-21. Number of days in 1973 on which the maximal hourly average
oxidant concentration equaled or exceeded 0.20 ppm at six air
monitoring stations in the southern coast air basin. Data as
reported: Los Angeles, Pasadena, Pomona, and Azusa reported by
Los Angeles County APCD; Riverside reported by Riverside County
APCD; and San Bernardino reported by San Bernardino County APCD0
(Reprinted with permission from Pitts.31)
4-49
-------
OC
UJ
Q_
140
120
DC
2 100
>: 60
Q
40
20
L
A
A
P
C
D
v»
t'
LOS ANGELES
PASADENA
POMONA
AZUSA
RIVERSIDE
SAN BERNARDINO
A
R
B
LA APCD METHOD
(|"VERSIDE. SAN BERNARDINO\
\DATA x 5/7 j
Figure 4-22. Number of days in 1973 on which the maximal hourly average
oxidant concentration equaled or exceeded 0.20 ppm at six air
monitoring stations in the southern coast air basin. Non-Los
Angeles APCD data multiplied by 5/7 to convert to Los Angeles
APCD scale. (Reprinted with permission from Pitts.31)
4-50
-------
140
120
o:
2i 100
o:
LJ 80
Q.
>? 60
Q
40
20
O
w
~ p
1
" 1
}• © LOS ANGELES
$ ® PASADENA
|; ® POMONA
4): ® AZUSA
1 © RIVERSIDE
® SAN BERNARDIN
C
D
0
©
@
®
V
ARB METHOD
/LOS ANGELES DOWNTOWN, PASA-\
\ DENA, AZUSA, POMONA DATA * 7/5 I
Figure 4-23<
Number of days in 1973 on which the maximal hourly average
oxidant concentration equaled or exceeded 0.20 ppm at six air
monitoring stations in the southern coast air basin. Los Angeles
APCD data multiplied by 7/5 to convert to the ARB-EPA scale.
(Reprinted with permission from Pitts.31)
4-51
-------
method. The results, illustrated in Figure 4-23, show the numbers for the
two easternmost stations to be lower than those for the stations in Los
Angeles County. Without saying which agency is correct in the controversy,
Pitts points out that a consistent approach with either method produces a
conclusion different from the previous one. Previously, it was believed
that the eastern counties experienced higher ozone concentrations. Now it
appears that the ozone concentration goes through a peak in the eastern
portions of Los Angeles County and decreases as the air parcel moves into
the counties east of Los Angeles. These results point out that extreme care
must be taken in comparing oz:one or oxidant measurements taken at different
places with instruments that are calibrated by different methods,,
The production of ozone in power plant plumes has been suggested to
explain ozone spatial distributions in nonurban areas.32*33 Comparison of
oxidation mechanisms competing for sulfur dioxide suggests that three reactions—
SO + HO •* SO + OH (3)
223
SO + OH + M -> HSO + M (4)
2 3
SO + CH 0 -> SO + CH 0 (5)
2 32 3 3
—are principally responsible for sulfur dioxide removal. Equation 4 is thought
to be the first step in a chain that converts nitric oxide to nitrogen dioxide,
causing an ozone buildup via the photostationary state mechanism involving
the nitrogen dioxide-nitric oxide-ozone cycle. Oxidized species up to HSO
6
are postulated as being responsible for the conversion.
4-52
-------
Descriptions of field studies of power plant stack plumes were given
by Davis ^_t aJL.32 The ozone concentration appears to be lower in regions of
high sulfur dioxide content. At 32 km downwind from the stacks, it was
claimed that the ozone concentration in the plume (now 11 km wide) is higher
than the ambient concentration ("0.08 ppm) by approximately 0.02 ppm.
Simultaneous measurements of nitric oxide and nitrogen dioxide were integrated
across the plume. Values of the ratio of nitric oxide to NO concentration
x
decreased monotonically from 0.9 at 2 km to 0.2 at 15 km and beyond. Other,
similar observations were cited.
It has been suggested by the analysis of organic components in rainwater
that naturally emitted terpene compounds may be the source of some photo-
chemical oxidant.31* Gas chromatography was used to separate individual
components of rainwater that had been purged by helium gas as a carrier.
The individual components thus separated were identified by their mass spectra.
Rainwater samples were taken at the end of a mid-August 1973 smog episode
in the Washington, D.C., area. The episode was characterized by stagnant
weather over a large area east of the Appalachian Mountains between New Jersey
and North Carolina. It ended on August 13, when a gentle rain occurred and
samples were taken. The dominant compound in the rainwater samples was 3-
methylfuran. Plausibiluty arguments based on chemical reaction mechanisms
were advanced for the existence of 3-methylfuran as a product of the photo-
oxidation of isoprenoid compounds. Some aromatics in the sample were attributed
to vehicular hydrocarbon emission, but the relative concentrations suggested
that terpenes could have been the most important compound in the smog. No
definite conclusions were drawn, but it was inferred that naturally occurring
4-53
-------
hydrocarbons could cause summer photooxidant smog. This is highly unlikely,
in view of measurements made on rain that originated from an air mass different
from that in which the smog episode occurred„
In addition to observations in Los Angeles, Blumenthal and White35 have
reported measurements of a power-plant plume and an urban plume 35 and 46 km
downwind from St. Louis, Missouri. Figure 4-24 shows the evidence of extensive
ozone buildup in the urban plume. Simultaneous measurements of scattering
coefficient, b , trace the spread and dilution of suspended particulate
scat
material. It is interesting that in the urban plume, which spreads to 20 km
in width, the ozone increases while the particulate matter decreases; this
suggests considerable photochemical production at an altitude of 750 m.
Contrary to the statements of Davis and co-workers reported above,32»33 the
power-plant plume causes a decrease, rather than an increase, in ozone.
Nitric oxide in the plume reacts with the ozone as it mixes. This is clearly
indicated by the distribution of particulate matter, which acts as a tracer.
Rough plots of ozone isopleths from measurements over rural areas of
Connecticut have been constructed by Rubino e£ al •36 Increased concentrations
(~0.31 ppm) are traced across the state during the June 10, 1974, episode.
The concentrations that built up throughout the day are associated with the
trajectory of an air mass that was over the metropolitan New York area during
morning peak traffic. The conclusion was that the city is an origin of
precursors that cause high ozone concentrations downwind. This is an interesting
contrast with (but perhaps not a contradiction of) the suggestions of Coffey
and Stasiuk,10 who believed that ozone of stratospheric origin is transported
into the metropolitan New York area. One difference between the two studies
4-54
-------
I'J
o,
cc.
UJ
B£
o
a.
o
i I
V -k
ro
CVJ
_
< Ul
o
fO
O
cc
ud
O
Q.
a.
i.l
x'1»DOSq
o
o
o
fO
e
tr>
CVJ
0
•
o
CVJ
O
•
in
— ~
mdd '{
i
0
-o°
^•^ t
lO
O
o
I
o
o
o
I
o
CJ
O
o
o
10
O
O
ro
O
CM
O
to
CVJ
0
1 -
IO
0
CVJ
O
«
^
10
~~
o
uj'dd .
o
o
o
i
o
•H g
CO iH O
6 co H
•^ e 4-1
^H 6 G
cu o
> O -H
CU LO CO
rH 1-^ CO
•H
CT3 -U 6
a) nj s-i
CO 0)
cu eu
(3 co
cfl M ^2
cu cu -u
CU M
> 4J t3
o cu
,n •> -u
cd +J cJ
^3 -H
B bO M
•rl CX
O P4 Q)
4-1 B
cfl • •
m B
CU o
co r^ ft
^ ro
cu •• o>
> ^-i •• n
00
CO CO
-rl -rt
3 3
O O
0) 4J 4-1
fl CO CO
B I I
4J U U
CX M !-J
cu<£!<;
coo
o
cu
s
CU
3
CO
CU
B
T^ *"O
3 S
•H -i-l
O O
3
in vo ,-H
oo
-------
is that the Connecticut work traced geographic movement of increased ozone,
whereas the New York paper showed temporal variations in urban and rural areas.
Sticksel37 discussed vertical profile measurements of ozone in the
stratosphere and the troposphere over the last several years. Transient
ozone maximums in the troposphere are illustrated and explained by three
possible mechanisms: a channel-like region conducted ozone from the strato-
sphere into the troposphere; ozone-laden air descended from the stratosphere
and was compressed as it subsided; and ozone-rich layers leaked through the
break between the polar and middle tropopauses by differential advection.
Surface variations of ozone soundings were mostly attributed to anthropogenic
pollution; however, relatively thick high-altitude layers are held out as a
potential natural source of high surface ozone concentrations. Sticksel
concluded that further investigation was needed to ascertain whether the
stratosphere can play a significant role in raising nonurban ozone concen-
trations.
Johnston38 has pointed out possibilities of significant transport and
transformation of oxidant precursors from the Los Angeles or San Francisco
metropolitan air quality control regions to the central valley of California.
Emission rates per unit area from sources in Los Angeles County and the Los
Angeles metropolitan AQCR are approximately 10-20 times the national average
for photochemically active pollutants. The emission rates per unit area in
the California central valley are about the national average values for
hydrocarbon and carbon monoxide, but only half the national average for NO »
x
Despite the much lower emission rates in the central valley, compared with
Los Angeles, the valley's ambient concentrations of carbon monoxide, NO , and
x
oxidant are half those in Los Angeles. Johnston suggested that a combination
4-56
-------
of urban-rural transport and confinement without substantial ventilation is
responsible for the ambient pollutant concentrations, which are relatively
high, compared with emission rates per unit area.
In considering the more general case of ozone in nonurban locations,
Johnston considered the problem of downward transport of ozone from the
stratosphere, which contains ozone at a peak value of about 10 ppm. .He cited
Fabian and Pruchniewicz's^ measurements of ozone at 19 nonurban sites in
Europe and Africa—from Tromso, Norway, to Hermanus, South Africa—in 1970-
1972. The 19-station network in Europe and Africa averaged approximately
0.022 ppm in the summer,, However, some of the central European areas
occasionally had higher values. ^ In another survey, a 12—site network was
operated from 1963 to 1964 between Thule, Greenland, and Balboa, Canal Zone0uo
Profiles were measured at intervals from ground level to about 30 km. Table
4-8 shows the number of ozone profiles that were made at 0-1 km and the number
of ground-level ozone readings that equaled or exceeded 0.08 ppm. Note that
the average of the 835 values corresponds closely to the summer values
reported from the European-African network. Table 4-9 details the seven
readings that equaled or exceeded 0.08 ppm. Five of them occurred at the
Bedford, Massachusetts, station.
Johnston argued that, if ozone were formed in the stratosphere and downward
transport occurred, there would be a positive concentration gradient extending
from 0.5 km to 5 km in altitude. Most of the cases in Table 4-9 obey this
situation; however, for the seven high-ozone cases in Table 4-9, the reverse
is true. This seems to indicate that the occurrence of ground ozone concen-
tration exceeding 0.08 ppm is characterized by formation reactions near the
ground, which suggests that stratospheric sources are not responsible for
these cases.
4-57
-------
•s
O ^H
o|
oo
CO
£3
O
C/S
0>
o
cfl
ft
c/j
d)
•H
,-1 vo
Oi
cfl I
CO
i-H vo
CU
cfl
01
CO
o
cS
a
o
4-J
bO
C
•H
XI
en
cfl
^
A
01
t»4
4J
4J
cfl
(3
•rl
CO
13
O
O
en
•H
^
•\
13
O
CO
•r-l
*C3
i
0
en
en
CO
^
^
•a
M
O
^H
•a
Ol
s
o
i-H
O
U
A
CO
(3
•H
,—1
i-H
0
O
4-1
O
cr
H
a)
a)
5
fa
0)
CU
en
CO
cfl
CO
3
4-1
CO
I
5
•s
cfl
O
OJ
C
O
[-J
cfl
S
O
cfl
§
fl.
o
J-
•
<1J
T3
S-)
O
PQ
60
•S
§
l-i
0)
Q
4-58
-------
Table 4-9
Ozone Concentrations (1963-1964) Equal to or
Greater than 0.080 ppm at Ground Level (0-1 km)a
Station Location
Bedford, Mass.
Ozone Mole Fraction, ppm
Date
Tallahassee, Fla.
Seattle, Wash.
June 26, 1963
July 3, 1963
Jan. 20, 1964
July 15, 1964
Aug. 26, 1964
Sept. 11, 1963
April 17, 1964
0 km
0.100
0.080
0.100
00090
0.110
0.080
0.100
5 km
0.075
0.075
0.065
0.060
0.040
Oo045
0.090
a Derived from Hering and Borden.40
4-59
-------
Evidently, the high nonurban ozone concentrations observed in 1971-
1973 in the continental United States were not of natural origin. This
conclusion is supported by comparisons of the data from the 1963-1964 North
American study and those from the 1970-1972 European-African study. Because
increased concentrations were not traceable to natural processes, anthropogenic
sources may be necessary to explain the concentrations of ozone currently
observed in the United States. Inhibition of the photochemical smog reaction
in the atmosphere by increased NO emission could delay the oxidant formation
x
until air parcels move out of urban regions. (The early portion of the
vehicle emission control pla^n had increased NO emission, but decreased hydro-
x
carbon and carbon monoxide emission.) Similarly, hydrocarbon substances in
the lower paraffin series that are relatively abundant may be reactive over
long periods and behave synergistically with other hydrocarbons. The high
ozone values in nonurban locations may well result from these processes<,
Nonurban oxidant measurements in Ohio were reported by Neligan and Angus.42
Concentrations of 0.18 and 0.12 ppm were reported for rural sites in Wilmington
and McConnelsville, respectively. At the same time, urban sites had similar
concentrations. However, the nonurban sites violated the ambient air quality
standard more frequently than the urban sites. Trajectory analysis showed
that ozone concentrations of 0.04 - 0.06 ppm were found in air masses that
had not passed over anthropogenic hydrocarbon sources0 These may have been
examples of naturally occurring oxidant. Airborne hydrocarbon bag samples
were obtained over 6-min periods, and ozone was also measured. At 4,000 ft
(1,200 m) above mean sea level (msl), ozone concentrations exceeding 0020 ppm
were observed over northern Ohio and Pennsylvania. Wind was predominantly
out of the Southeast. Forty-eight-hour reverse trajectories (apparently
4-60
-------
based on surface winds) showed that a developing high-pressure system carried
air from central New York State and seaboard metropolitan areas into the test
region. The correspondence of locations with high ozone concentration and
locations with high acetylene, carbon monoxide, and chlorofluoromethanes
("freons") concentrations suggests anthropogenic emission that leads to
increased ozone concentrations. Early in the morning, the rural ozone con-
centration is at a maximum. Profiles at 7:00-9:00 a.m. show increased ozone
in nonurban areas at altitudes of several thousand feet and relatively low
concentrations at the surface. Daytime profiles show increases in ozone at
the surface, with the late afternoon concentrations exceeding the midday
concentrations.
Deep mixing of stratospheric ozone can occur during episodes of frontal
passage and jet stream interaction. The mechanism for this transfer is the
downward advection of air parcels rich in ozone and unaffected by removal
processes until they reach the ground. The meteorologic conditions during
these events are totally different from those attending high-oxidant episodes
over extensive nonurban areas. Therefore, the invocation of stratospheric
transport for high nonurban ozone concentrations cannot explain the long-term
increases in ozone observed in the studies described above.
4-61
-------
INDOOR VERSUS OUTDOOR OXIDANT CONCENTRATIONS
An important aspect of the relationship of indoor to outdoor oxidant
concentration is the scavenging of the oxidant by surfaces, e.g., in ventilating
systems and in indoor areas.
Ozone decay was measured in an office, a home, and several metal test
facilities.43 Measurements were carried out with a Mast ozone meter and an
MEG chemiluminescence ozone detector. The latter was calibrated with a stable
ozone source and the EPA neutral buffered potassium iodide procedure,, (It
was noted over a wide range of concentrations that the MEC meter measurements
were consistently higher than those of the Mast meter by a factor of 1.3.
That this is essentially identical with the findings of the DeMore committee2
is interesting.) Ozone generated by a positive corona ionizer was introduced
into the test facilities. Ozone decay in a metal-walled room was found to
be first-order, with the rate constant highly dependent on the preconditioning
of the metal walls with respect to ozone exposure. An activated-carbon filter
in the air cleaner for the room markedly increased the rate of decay of ozone„
Increases in ambient humidity or temperature also markedly increased the rate
of ozone decay. Decay rates in the bedroom and office environment were much
greater than those in rooms with aluminum or stainles-s-steel walls 0
During a series of smoggy days in Los Angeles, indoor concentrations of
ozone were measured to study the phase relationship of indoor to outdoor
ozone buildup.44 The genersil mechanism of ozone decay within buildings was
investigated with a Dasibi ozone instrument. Various air filters were evaluated
with respect to capacity for removing ozone. A computational model was sug-
gested for correlating indoor and outdoor concentrations of ozone as a
function of time. Buildups of ozone within buildings lagged behind ambient
4-62
-------
outdoor buildups by 3 or 4 h, and the maximums indoors were not as high as
those outdoors. For example, a small number of measurements suggest that
concentrations in private residences reach only about 70% of the peaks
attained outside. Sabersky et^ _al_. ^ constructed a table of ozone decomposition
rates on several common materials. Activated-charcoal and Purafill filters
performed better in their tests than fiberglass filters. Approximately 95%
of the ozone was removed by the activated charcoal under a wide range of flow
conditions. A differential equation model was proposed to calculate concen-
trations-time relationships for indoor ozone concentration . First-order
reactions with empirical constants were assumed, and the predicted ozone
concentration patterns were similar to those based on measurement,,
A ventilation model for relating indoor to outdoor pollutant concentration
was proposed on the basis of an extension of earlier work by these authors.45
When the outdoor pollutant concentrations change slowly, compared with the
indoor changes, a rather simple equation can be used to relate indoor to
outdoor concentrations. Further findings of the model study suggest that
reductions in indoor concentrations down to 20% of the outdoor concentrations
is feasible. Such a program would limit the number of days on which a
threshold value of 0.1 ppm is exceeded to 4 or 5 per year, instead of the
typical 200 or more days/year for outdoor concentrations,, The model was
tested with the assumption that the air flow in a building can be approximated
by a well-stirred chemical reactor, which is influenced by constant sources
and first-order reactions.^ Monitoring measurements were taken for a
laboratory building at the California Institute of Technology to test the
model. For a series of locations within the building, the model predicted
a good upper-bound curve for the observed indoor pollutant concentrations.
4-63
-------
It appeared that ozone concentrations were lower in rooms with excess furniture,
books, and papers than in rooms that had concrete floors and contained few
books or papers.
The use of ozonizers for deodorizing indoor air has been discussed and
3
evaluated with respect to potential health hazards.^7 In a normal 40-m
room, an ozone concentration of 0.1 ppm is established after 3% h of oper-
ation of one of these devices. Evidence on health effects was cited to
support the conclusion that inhalation of the quantities of ozone produced
by these air conditioners should be avoided and that certainly no beneficial
effects should be attributed to ozone inhalation.
PERIODIC TEMPORAL PATTERNS OF OXIDANT CONCENTRATION
Although the previous discussion on spatial differences in ozone con-
tained some references to time dependence, this discussion will explore
periodic time dependence further. Specifically, diurnal and seasonal
variations will be explored here, with data on various cities in the United
States.
The diurnal variations in mean hourly average oxidant concentration are
illustrated in Figures 4-25 and 4-26. Several factors influence the shapes
of these curves. The primary influence is that of sunlight intensity, inasmuch
as photons in the ultraviolet are responsible for the primary photochemical
process that leads to ozone formation. Note that the St. Louis curve for
June 1966 is broader than that for Los Angeles in August. This is explained
by differences in sunlight intensity distribution throughout the day.
Another factor in the curve shapes is the relative proximity of sources
of nitric oxide, which reacts with ozone locally and suppresses it. Figure
4-19 shows how the diurnal variations are larger in a city where nitric oxide
4-64
-------
0.16
I I I I I I I I I I I
LOS ANGELES
AUGUST 1964 AND"l965
HOUR OF DAY
Figure 4-25. Diurnal variation in mean hourly average oxidant
concentration in Los Angeles and St. Louis.
(Reprinted from U.S. DREW.1)
4-65
-------
0.30
0.2S
0.20
2 0.15
UJ
U
z
S o.io
0.05
I I
I I I I I I I
01 I I I I I I I I I I I
12 2 4 6 8 10 12 2 4 6 8 10 12
W O.IH. •!• p.m. 4
HOUR OF DAY
Figure 4-26.
Diurnal variation in mean hourly average oxidant
concentration in Philadelphia, August 6-8, 1966.
(Reprinted from U.S. DREW.1)
4-66
-------
emission dominates than in a rural area, where the ozone is relatively
unaffected. This brings us to a third factor influencing the shape of these
curves: advective transport. If movement of ozone from another area is
dominant in the local photochemical oxidant concentration, then the wind
direction and speed have a great influence on the curve shape„ Figure 4-27
shows that, in a central urban area like downtown Los Angeles, the ozone
pulse shape is relatively symmetrical about the middle of the day, because
of the dominance of local production. In outlying areas, however, such as
Azusa and Riverside, the pulse is skewed to the later hours of the afternoon,
because of the greater role of transport to the area. Riverside shows this
most decidedly, with its own locally produced ozone rising around 8 a.m.
and a large secondary peak at 4 p.m., presumably the result of advection
from areas to the west. Palm Springs peaks after 8 p.m.
Monthly patterns that show the seasonal variation in oxidant concen-
tration for three cities are presented in Figures 4-28 and 4-29° Again,
the combination of availability of sunlight and degree of ventilation governs
the shape of these curves. Los Angeles is subject to high late-season oxidant
production, because the late spring and early summer months are heavily
affected by stratus, which obscures the sunlight. Denver, however, has a
more symmetrical seasonal distribution that follows the sunlight pattern.
Phoenix has relatively lower oxidant peaks in July, August, and September,
because of the dominance of convectively driven mixing, which counteracts
the affect of the higher solar intensity.
4-67
-------
4O
I-
•* A
40
ft a
8 4
8»o
I-
* to
\ .
* ^
!"
B
1 I 1 1 • 1 1
A
"~T_/V_
/\
• Kl$*BtH* / \
V_ -
_ —
- »OVO« '^>A
L ' ^— •
-^ x%
J-^x^' \
- livt*S
-------
0.06
LOS ANGELES
1964-1965
DENVER
1965
PHOENIX
JAN. 1967-JUNE 1969
JAN. FEB. MAR. APR. MAY JUN, JUL. AUG. SEP. OCT. NOV. DEC.
Figure 4-28. Monthly variation in mean hourly oxidant concentration
in three selected cities. (Reprinted from U.S. DREW. )
4-69
-------
0.20
0.16
cc
IU
u
§
u
IU
s
0.12
0.04
LOS ANGELES
1964.1965
PHOENIX
JAN. 1967-JUNE 1969
JAN. FEB. MAR. APR. MAY JUN. JUL. AUG. SEP. OCT. NOV. DEC
0.08,
Figure 4-29. Monthly variation in mean daily maximal 1-h average
oxidant concentration in three selected cities.
(Reprinted from U.S. DREW.1)
4-70
-------
LONG-TERM TRENDS OF AMBIENT CONCENTRATIONS
The limited sample of photochemical oxidant and ozone data precludes
extensive trend analysis for all but a few regions in the United States.
Some such analysis has been done by the California Air Resources Board.49
Its analysis can be applied on a temporal basis to any given location;
however, spatial intercomparisons are subject to the problems pointed out
by Pitts.31
Both annual average and 3-year moving average values of oxidant concen-
tration were plotted in a GARB report by Kinosian and Duckworth49 for several
stations in the southern coastal air basin during 1963-1972. These are
shown in Figures 4-30, 4-31, and 4-32. The 3-year moving average is used
to smooth the data. Comparing the three cities—Los Angeles, Azusa, and
Riverside—they found a distinct downtrend in the 3-year moving average for
Los Angeles, a nearly level trend for Azusa, and a slight uptrend for
Riverside. It should be pointed out that the variability due to weather is
not completely removed by the 3-year averaging process. Thus, the report
applies an adjustment factor for temperature aloft and concludes that low
oxidant concentrations in 1968 were due to the weather, whereas the lower
ones observed in 1970 and 1971 were not.
A report by the National Academy of Sciences Coordinating Committee
on Air Quality Studies to the U.S. Senate50 concluded that "available air
monitoring data do not allow conclusions to be drawn about photochemical
oxidant trends on a nation-wide basis." This report relied heavily on
California data to illustrate trends, because so much information was
available for that region. Maximal 1-h concentrations in the New Jersey
cities of Bayonne and Newark were compared for 4-year periods between 1966
4-71
-------
• o
V
§
-rl
CD
W
CO -H
T3 6
M
14-1 0)
O &
O
r-
o»
CO
01 -
00
cd
VI
0)
f>
Cd
C M
•H a
o PS
-O 4-1
C &
cd o)
to
\o
\o
£
cn <
rH
i
CO t^
VO rH
co OJ-
01 M-l •
rH JS
0) CO 4J
60 C H
COO
to cd o
o n s
nJ 4-1 n
C ST»
•HOC
C cd
w o
T3 O C
fj cd
<1) X! -rl
i-t I CO
4-) ^1 O
4J iH -H
cd E3
^H 'X O
!»! cd !-i
06>w
O
CO
0)
t-l
3
60
4-72
-------
CO X
CU 4J
bO-H
> Q)
cB 4J
bO -H
C M
•H ft
> CU
O Pi
B ^
ca •
cu M
!>> cu
I £>
ro B
cu
T3 4J
C ft
ca cu
co
3 C
C c8
c
CN 00
r» 3
en
VO
CTi
"MM
c8 O O
•H IH 5
c ^
M CO O
O C 3
o
T3
C
" C fS
ca cu ca
M o -H
3 a co
N O O
co
•Sf «
§
M
c ca iw
cu B
M -3 c
•u x o
4-1
9
3
CO
to
ca a)
T3 P.
en
I
Ml
•qdd
4-73
-------
eo
0) .£
00 4->
rt -H
M &
T3
n) 0)
4J
00 C
C -H
•rl M
> P.
O OJ
J-l
flj
CO
a)
en
1-1 H>
.g *
• co
CN 3
co •>
vO >-,
(^ r-t
^H 3 /-.
1-3 en
H J-
cfl H •
•H O JS
o w o
>H S &
•H O A!
iH -H 0
cd 4-> 3
U flj Q
Vl
«. 4J T3
a) 0 C
is co
•H fi O
ft I C
•H iH
a @
en B o
T3 -H »-(
C H t-i
(U TO
^ e c
4J O
^•H
4J rH CO
PS -H CO
R) td -rl
X m a)
OOP.
ro
00
•H
4-74
-------
and 1973. These two cities exhibit 24% and 46% decreases, respectively„
Emphasis was placed on differences in calibration procedure in various
jurisdictions of air pollution control agencies. It is indicated that trend
analysis for each station is still valid, despite the differences in cali-
bration procedure. But values from different places must be compared
cautiously,,
CAMP data presented in Table 4-10 show a generally decreasing trend in
total oxidant concentration as measured by neutral buffered potassium iodide.
The southern coastal air basin in California is also known as the metropolitan
Los Angeles AQCR. The map in Figure 4-33 shows the location of air pollution
monitoring stations. However, the boundaries shown in that figure include
more than the AQCR referred to. Of Santa Barbara County, only the southern
strip bounded on the North by the coastal mountain range is included in that
basin; and of San Bernardino and Riverside Counties, only the partially
urbanized areas in the western portions are included. Figure 4-34 shows
the oxidant trends in the southern coastal air basin. Two selection techniques
were used for stratifying the air monitoring data.51 One is based on the
"rule 57" day, which is defined as a day having an inversion base at 4:00
a.m. lower than 1,500ft, a maximal mixing height below 3,500 ft, and an
average wind speed between 6:00 a.m. and noon below 5 mph. Another selection
technique stratifies the data for an inversion base less than 3,500ft. The
number of days in each year that exceed an oxidant concentration of 0025 ppm
as defined by these two systems shows a distinct downward trend in the graph.
As a result of recent control activities, the downtown Los Angeles
station has experienced successively lower oxidant concentrations,^9»52 as
shown in Figure 4-30. It should be noted that Riverside and San Bernardino
4-75
-------
<".
a*
cd
o
OO
I
o
i-H
1
Cfl
)H
4J
CU
CJ
C
O
u
4->
c
cd
T3
•H
k
cS
i-H
cfl
4->
o
H
3
13
O
,C
4J
CU
S
CU
T3
•H
T3
O
M
g
3
•H
CO
CO
cfl
4->
O
CU
T3
CU
!H
cu
M-l
U-i
3
M
i-l
cd
rl
4-1
3
cu
2
>•>
,£>
CO
C!
O
•H
J-J
cd
4-J
CO
r-l
Cfl
3
<
>4H
O
CU
60
cd
rl
' cu
-
vO
CT\
I— 1
r~-
CTi 00
CN CN
vO
^.o
vU
cr.
i-H
1
CN
vO
-------
Figure 4-33. Air pollution monitoring
4-77
-------
100
50_
IWI pr*|«el4«l *l MIMlll InM!.
JUtMil in«*ltOfiiif 4mtm c«rr««M
»,«.!• 17 ««,!.+
X AclMl nwniurliif «•!• e«r*cM
ky »••»««• MlMtlMn MOO lit
1965
1970
1975
|A rule 57 day Is one on which the Inversion base at 1 *M (PST) Is lower than
1500 feet, the maxlnun nixing height Is not above 3500 feet and the average
surface wind speed between 600 AM and 12 noon (PST) does not exceed 5 MfH.
tibia correction has been adopted by the California Air Resources Board In an
•ttoivt to move this meteorological variability froD the analyzing oxldant
trends.
Figure 4-34.,
Southern coastal air basin oxidant trends,
permission from Heitner and Krier.51)
(Reprinted with
4-78
-------
Counties show great fluctuations in annual average oxidant concentrations,
with Riverside actually having an increase. Again, it should be emphasized
that differences in calibration methods from place to place will not affect
the temporal trend analysis presented here.
It has been argued that the smog reaction has been inhibited by the
higher NO :hydrocarbon ratios brought about by early emission control systems,
x
thereby increasing concentrations at downwind locations,, It has also been
argued that increased NO concentrations could contribute to the high non-
x
urban values by interacting over long periods with natural methane or terpene
to produce ozone«53»51t It is probable that nitric oxide emission reacts
locally to decrease preexisting atmospheric ozone; however, there is little
doubt that increased NO over a long-time (long-range) trajectory will
x
ultimately make more oxidant, if more hydrocarbon is introduced. In summary,
added nitric oxide inhibits ozone locally, but can enhance it regionally,,
Altshuller49 evaluated oxidant results from throughout the United States.
CAMP concentration data were analyzed for 8 or more years of measurements
available from the period 1964 - 1973. Tables 4-11 and 4-12 summarize the
measurementso Data from 2 years were rejected, because the method of
eliminating sulfur dioxide interference was not in operation. Interference
from nitrogen oxides was removed from the data by subtracting 20% of the
combined nitric oxide and nitrogen dioxide concentrations from the total
oxidant reading, because all the data used were from the colorimetric
potassium iodide method of measurement. Where comparisons were available,
colorimetry oxidant data averaged 0.015 ppm higher than the chemiluminescence
measurements. It was observed that this difference should be expected,
because of possible additional interferences from organic peroxides or
peroxyacylnitrates.
4-79
-------
P*> C *
in to a
M 'd O
3 M-i -H
O ^d W
M O CO
i—I QJ JJ
cfl 60 C
a ctf Q)
•H l-i O
X 0) C
ifl > O
S 60
cn *d
Q ^H (U
M
-------
rH Cd fS
J-i T3 O
3 -H -H
OX4J
prj Q (^
rH CU 4-1
CO 60 Pi
g tO o ft
a <: u ft
OCOCMI— r-H i—I PO -J"
CNrOLOr-HCsl!—(r-li—1
OOOOOOOOO
OOOOOOOOO
CN fvj i—) C^J CN »—(r—( r—I F-H
OOOOOOOOO
•
4J
d
o
o
i — i
.-H
1
cu
H
cd
H
«
CM
r^
o>
^H
I
O-
vD
cn
rH
«\
CO
01
•M
•H
c/1
Qj
§
u
4-1
cd
T3
a)
T3
^j
0
o
CU
pi
CO
d
0
•H
4J
n)
M
4-1
£
cu
o
(3
O
U
4-1
c
cd
T3
•H
X
0
rH
cd
4J
O
H
MH
0
>.
rJ
CU
£IJ
O
4J
CO
cd
h4
4J
cd
J3
4-J
, J
Fl
CO
^
cd
a
*
0
&
^
o
0
4-1
rH
3
CT
W
CU
60
cd
cu
^
<^
r*^
rH
V-4
3
Q
|T}
CO
^»
cd
Q
•
0
"Z
ft]
C^
m
i— t
•
0
0
ft
ft
0
i— i
o
o
60
C g
•H ft
13 ft
CU
cu to
0 0
X •
w o
rrj
•H
rH
cd
> cd
4-1
<4H Cd
O Q
rJ
cd
cu
^H
ro CM m r-j
vo ro ro oo
-------
CN
i— 1
1
0)
r-H
r\
CO
H
4J
d
cfl G
Tj CO
•H r^N
X CJN
O -H
1
T3 *^J"
N, 4-1
cfl cfl
Q rJ
<4-i d
O O)
o
M C
O) O
rO CJ
B
3
S3
rQ
co
r*"l
cfl
Q
4_l
o
0>
W
tO
4J
C
0>
a
0)
o
o
S3
i— i
cfl
4-1
O
H
ft
C
0
4-1
00 •
C! CJ
•H •
rC Q
CO
cfl
JZ
CO
•H
3
O
rJ
.
4J
CO
•
o
S3
•
si
•
o
^
CM in o o X-N X-N X-N m mo x-\ X-N
r-.incMrHCMOOi— i incNooo
vOCNvtr^-J-CMrHvO 6-2 OCMONCM
oor^-covC' — i r^« VOCM
CO CO rH O\
<3" CM X-N X-N X-N CN VO 00 X-N , ,
•— i i— i in vo i — i
OO ON •— 1 i— 1 rH in rH
CN
X~\ X-N X™\ X—S
OO X-S X-N X-N X-N VO O in X-N X"N
rHONOOrHO CO UOCNCOO
CM vo ON vo co o co vo oo vo in CN
ON in rH r^ Vf rH
i — i
cfl
•H
rH
CI)
13
cfl
rH
•rl
^3
PH
o
S3
•— !•— ICMCOCNOOCO vOCMrHO
cor^vor-cOrHi-Hoo Or^-cMco
inr^cNi — i r^- &-? r^cM< — '
i— i
Q)
01
Q
.
O
^
*" X x™s ^ ^ ^ N
rHrHrHrHO CN VtrN-J-O
4J
o a
0) 01
M O
O*-.
U
o cj
e
& ft E P
a a, a. o.
P. a. P.
m
o CM m o
•— 1 rH rH CN
o o o o
in in CD ai ai a)
ON o-ininminrH >>>>
ocNi-iON,
> CO
•H -a
n c
0> 3
n w
4-82
-------
The fractions of days with valid oxidant data for May - October and
for June - August were tabulated for the CAMP sites for 1964 - 1973. The
average percentage over all years and all sites was 60%. Seasonal effects
were shown by distributing according to month the days with oxidant concen-
trations exceeding 0.08 ppm. Sunday oxidant concentrations and comparisons
of results by location were also discussed. Annual trends in oxidant con-
centration for the various sites were represented by plotting the number of
days with oxidant concentrations over 0.08 ppm (or the California state
standard of 0.10 ppm) for each year at each site (see Figure 4-35 and Table
4-13). It was suggested that the relatively infrequent high oxidant con-
centrations observed in Chicago were due to a site effect involving the lake
breeze ventilation and the high concentration of automobile-generated nitric
oxide. The higher frequency of high ozone concentrations in Denver was
attributed to Denver's high altitude and attendant increases in ultraviolet
intensity. No distinct Sunday oxidant effect was noted with respect to
exceeding the 0.08-ppm concentration. Neither precursor concentrations nor
favorable reactant ratios should lead one to expect more favorable conditions
for oxidant formation on Sundays, according to the paper. However, it was
noted that in the Los Angeles basin alert concentrations (0.6 ppm) or near-
alert concentrations (0.4 - 0.6 ppm) are never observed on Sundays. Extensive
discussion was devoted to the apparent downward shifts in oxidant measurements
for central urban areas, in contrast with the nearly constant or rising
oxidant concentrations in suburban and rural areas.
The latest Los Angeles County APCD profile gives further insights into
the measurement of trends and their interpretation. It uses California's
ambient air quality standard of 0.10 ppm oxidant (hourly), instead of the
4-83
-------
r_
O Denver
OSt Louis
* Washington, D.C.
v Philadelphia _
Cincinnati
« Chicago
"1963 19M 1965 1966 1967 1968 1969 1970 1971 1972 1373
Figure 4-35.
Trend in oxidant concentration by year at CAMP sites.
(Reprinted with permission from Harrison and Lodge.57)
4-84
-------
en
r-H
|
o-
0)
,1
r_j
cfl
EH
SfCl
cx c
O, -rH
CO
O cd
!-< CO
•
o to
CU
MH rH
o cu
00
CU Jj
m
(H CO
CU O
> I-J
^
d
r-l -H
3
O "°O
W CU
1 t3
-HCU rH
cu cd
T3 O CU
MX r^"*
ct) W
Td
C rH
cfl o
4J
T3
CU CU
•U I—I
nj cd
4-J 3
c/} O*
W
JU
O CO
•H cd
^y
C 4J
0 C
cd
CO 73
>-. -rH
Q ^
vO
2
^O
^O
ON
i-H
m
ON
r-H
^3-
VO
ON
r-H
•JS
ON
i-H
CM
**C
ON
i-H
vO
ON
I— 1
o
ON
rH
ON
in
ON
rH
00
m
ON
i-H
In
ON
rH
£
4J
g
O i—i £f) c\] *^D u".
r-l CN CN
CN r^ *«O CNJ CN vO
r-l CSJ CN CsJ CNJ
0 CM r- ^ - vo
in I-H ON ON vo r*^
rH rH rH CM CM
oo r^ r^ CM ^o CM
i-H CM CM CM
GO CO 00 <3" ^O **O
rH rH CM CM CM
O CM 00 % C
cd cu td D. cd 3
!-) fu S ->
t — ( ON r^- -*^ P** O in
en CM CM CM oo
O ON <}• ejo ~^ *~~* *~H
en CM CM -H
CM
rH rH VD 00 Cn 0 00
CM
I — 1 rH ON -^" r^- CM rH
en en CNI CM rH -d-
CM
O IH O r^ -vt" CM **O
en en en CNI rH rH ~a-
CN|
rH ON in VO O VO CM
en CM CM CM CM in
CM
rH rH CO rH O 00 ON
en en CM en CM in
rH rH OO O ^ VD i-H
en en CM en I-H r^
CM
rH i — i -^ oo m oo ^o
en en CM CM rH en
CM
rH o O t^ o en CNI
en en en CM rH en
CM
ON rH VD r-» \o o oo
CM en CM CNI I-H CM m
CM
rH rH ON *sD 1 • -rJ • • • Cd
rH 60 fX 4J ,> O W
3 3 CU 0 O CU O
•
13
O
4J UD
CO ID
e •
CO
0) O
T3 A!
•H cfl
T3 rH
O -H
•H CQ
•H M
CO 14H
CO
cd T3
j_> a)
0 >
a, T)
rH
>> 01
m n
«-Q
4-85
-------
federal standard of 0.08 ppm. Table 4-13 shows the numbers of days on which
the state standard for total oxidant (by the potassium iodide method) was
equaled or exceeded in the Los Angeles basin (confined here to the county of
Los Angeles). From 1957 to 1973, a distinct downtrend in the number of days
was observed. On an average over this 17-year period there were 258 days of
violation per year. During the last three years tabulated—1971, 1972, and
1973—the numbers were 218, 211, and 185, respectively. This trend is
qualified by the observation that climatologic conditions during 1973 were
favorable for smog formation, with the lower frequency of inversion and weak
winds. It is perhaps more realistic to consider the final 3 years as a
group, rather than using 1973 to interpret the trends. Another index that
the Los Angeles County APCD uses is the total number of days on which eye
irritation was recorded in the Los Angeles basin. This is summarized in
Table 4-14. For the same 17-year span, the average number of days charac-
terized by eye irritation is 163. However, during the final 3 years, the
numbers were 125, 125, and 124, which seem to correspond to the ozone
measurement trend. It is of interest to note the similar statistics (Table
4-15) for violation of the state visibility standard—namely, for one
observation, visibility should not be less than 10 miles when the relative
humidity is less than 70%. The data in Table 4-15 are reported on the basis
of the 1970 state standard through May 1972 and later on the basis of the
June 1972 state standard. The average number of days on which the standard
was violated over the 12-year period is 328. For the last 3 years of the
reporting period, the numbers are 333, 311, and 298. This confirms the
improvement trend exhibited in the other two tables.
4-86
-------
•3-
1
01
43
cd
H
a
c
•t-t
en
(fl
M
en
0)
00
(H
H
C
0
•H
(fl
4J
•H
VJ
M
ai
^-,
w
45
0
•H
43
2
G
0
CD
r*~l
(0
Q
MH
O
SH
01
43
e
3
Z
CO
r^ O O O ON
ON
i-H
CM
i — I
ON
0
ON
^
ON
NO CO fH ON in
ON
fH
00
NO CM r~" in r^
ON
r-H
NO N4O \O O ' — '
ON fH
t-H
NO
NO CM CO 00 00
ON »-H r-H
i-H
in
NO CM r-H ^ in
ON fH fH
i-H
m
r-l T-H
-H ro
r-H r-l
\C «"HI
fH CM
CO fH
CM
m ^j-
CM
CO
CM
CO
CM
00
CM
CM
CM
NO
CM
co
CM
p-H
CM
r-.
CM
CO
CM
J>^
i-l
3
^
CO
CM
O
fv)
CO
CM
NO
CM
fH
CO
ON
i-H
NO
CM
CO
CM
ON
CM
CM
CM
CM
CM
ON
CM
00
CM
NO
CM
CO
CM
r-H
CO
j — .
CM
CO
CM
•
00
<3
f^.
1 — 1
u-l
i — i
0
CM
vj
OJ
vO
t-H
in
i-H
O
CM
O
CM
CO
CM
ON
i-H
in
CM
in
i-H
i — i
CM
CM
i — 1
NO
CM
in
CM
00
CM
.
4-1
a
0)
C/3
p^
fH
CM
t-H
fH
0
rH
CO
fH
CO
CM
ON
i-H
CM
CM
p-H
CM
CM
CM
co
CM
ON
t-H
00
t-H
CM
CM
t-H
CM
t-H
. — 1
f>^
i — 1
4-1
O
0
CO
-d-
-3-
(0
•H
m
e
o
(H
>4H
13
(U
•H
SH
CD
Q
«
4-87
-------
m
i
~, ON CM CM CM CM
S 4J rH
•H
4J -H
•H g ON
rH 3 VO •* CM -* OO
•H X ON CM CM CM CM
O rH
•H 0)
•H -H
> 4J OO
cd vo oo vo oo ON
S-l rH ON CM CM CM CM
O 0> rH
<4H f£| rJ
cd
tJ C 01
rJ 0) >H 1^
cd rC vo oo oo ON -o-
T3 3 ON CM CM CM CM
C rH
cd 03
4-1 0)
OO rH
•H vO
01 S vO vO vO vO CO
4-1 ON CM CM CM CO
Cd O rH
4J rH
CO
C
rC cd m
O r£2 vO vO *^" 00 CM
•H 4J ON CM CM CM CM
-C rH
tS 03
03
C 01
O rJ O-
vO i*"** -*^ *-, ON CM CM CM CM
>-> 4J rH
cd -H
•H
<4-t rO CO
O -H VO CTl CM VO rH
03 ON CM CM CM CM
to -H rH
0) >
D X_X
3 t3 CM
js 01 vo in oo vo ON
t3 ON CM rH CM CM
01 rH
01
o
X
W
rG x: rH
4J • • U iH
C C rQ rH r4
r§ * fe r§ •§*
ON
CM
OO
CM
00
CM
O
CO
rH
CO
O
CO
o
CO
oo
CM
1 — 1
CO
vO
CM
00
CM
tn
CM
r*N
£
OO
CM
00
CM
O
CO
o
CO
ON
CM
O
CO
oo
CM
ON
CM
ON
CM
ON
CM
00
CM
ON
CM
CO
C
3
rH
CO
oo
CM
i— 1
CO
i— i
CO
rH
CO
rH
CO
rH
CO
rH
CO
rH
CO
rH
CO
rH
CO
1 — 1
CO
>>
rH
1-3
0
CO
rH
CO
I— 1
CO
*rH
CO
i — 1
CO
o
CO
rH
CO
rH
CO
rH
CO
1 — 1
CO
rH
CO
— 1
CO
•
60
0
CO
vO
CM
ON
CM
ON
CM
O
CO
o
CO
o
CO
ON
CM
ON
CM
O
CO
ON
CM
0
CO
•
4J
0)
C/3
r-^
CM
^3-
CM
C^
CM
O
CO
oo
CM
O
CO
rH
CO
rH
CO
00
CM
O
CO
o
CO
o
CO
•
4->
U
O
O
53
vO
CM
m
• — i
ON
i — i
vO
CM
VO
CM
vO
CM
vO
CM
m
CM
vO
CM
i — l
CM
ON
CM
i— 1
CO
•
a
01
a
00
ON
CM
1 — 1
1 — 1
CO
CO
CO
CO
CM
CO
CO
VO
CM
CO
in
^~
CO
CO
«^"
CO
rH
to
0)
Q
4-88
-------
It is also of interest in interpreting trends to look at the rare event
and at the frequency distribution, as well as average values. This can be
done by examining the number of ozone alerts called for the series of years
under analysis. For a first alert for ozone, the Los Angeles County APCD
has set the value at 0.50 ppm. Ninety alerts were declared from the inception
of the alert program in 1955 through December 1973. The 1955 - 1972 average
is five alerts/year. However, in 1972 and 1973, only one alert was declared
each year. The highest number of alerts posted during a single year was 15,
in 1955. The highest frequency of alerts occurs during September. January
and February have never experienced ozone alerts. Alerts are most frequent
on Friday and have never occurred on Sunday.
Frequency distributions of ozone concentrations for Azusa and Goleta,
California, are shown in Figures 4-36 and 4-37, respectively. (These curves
were prepared by J. R. Martinez of Environmental Research & Technology, Inc.)
The Azuse distribution consists of points for 8 years (each taken as an
individual sample). The Goleta data are for 1 year only. A year-to-year
scatter of approximately ±15% about a mean curve is observed for Azusa; both
curvature and slope difference distinguish it from the Goleta distribution.
The primary reference standard for the latter followed the GARB procedure;
that for Azusa followed the Los Angeles APCD procedure. Although the DeMore
et^ al. report2 states that zero offsets are suppressed and that direct
multiplicative corrections should be applied, one must wonder whether nonlin-
earities may still emerge from the calibration procedure, in view of the
differences in curvature. Another explanation may be that, because of the
higher concentrations in Azusa, actual atmospheric nonlinearities occur.
4-89
-------
*
"j
;§
"K
;j
4-t—» '
£'._-rp:
iff
H
;inrn
P—-t- :—i , ; —~
._,_^=r;q3^g
---i *
o
CT\
£3EdE3£E3f
nrp
l£fc
~f~ r r~" t i ! ,
-•&--—:
—rj
— I >.
HI
3
D1
01
lj
<4-<
mn:mrcrr-rr
_—,— _V S»—:_,
— a)
^r: c
•— o .
— N (N
o r-.
— o\
— X >-H
— T1 '
ii M u-l
r' a vo
— O ON
1 i fC i—<
vo
01
i£3E
^^
7f-t
^F
^li-^ii
l_ <"
ti
^H^j—i~
i^tP
iwgi
—-t—j—i— t-
— I ( 1 —r
a
•H
fe
^StEHi
>l!i
I '.';• / / / •'
J^jJU<$;j
.LuauiiiL:
^==•{33^==;
mi
H
-i-MU;!:)
XMS
Effl
SB?
_(.^LlL-
•tii'
n
^
I I I
^
tilt'
11
TT
I I l
c. o.
* *
o o
5 J^
c
-------
EschenroecVr , ' "f . i V. , p . > 5 ,
tte
H?T
J«
V "
ih
i Mi'i'iiiiniiiip.mimjj
4iLilJ±rj1.11j_l
ro
ov
id
u
- 8
o
•H
4->
3
X3
•H
-J_- M
H -H
~J^ 'O
0)
3
o1
ilt
h i
Tin
•-t q;
~r o
N
- o
0
pa
ttf
ON
I
m
•I'll !
_U._J
4-U-
nn_
:1-J
:rr.::
TIT
.,4—- -. -i --;----)
—T—•i- -H 1—-1
::q—irr: : -::-"-4":~-
— L , ^ __1 1 ^J
:rrT^-r~ Pclc'iii—
-f— h-
:rt:
Illl
im
j i •! i i i i i ,
i i i ' i ' i _>
j^:lJLL"" - '-'-
=r-f
TTEr , r , _^
=;-
^rp-i:—i—i~r--'f- ^t
±rtS±i4rh±
~*" ' ' rTr.7_-_-_-z
3"
•' /• i
•n-rb^-i-rr
7'
4S
-J-T-I-••'.->" I ••!-!•••:• ;.|. '+-r-
ijI^'u!!llI
,t."ri'|,,|:li'"l!li1
i^V
.„ -j __.
114
irr
5tiiC
r^rd —
li I
iU
1
nr
1 1
_._:——,' a
z=r^]
Us
J
c5 6 o 6 6
v> w r*>
o
p o
~^H\
4 «>
O
* J J
NO:> 3 NOZO hl^JOOH
^i.j^-/j.i -, _^.^
J
O
o
-------
But this will not account for the large slope discrepancy at low concentrations.
It should be noted that all Los Angeles APCD distributions qualitatively
resemble Azusa's, and that other GARB distributions are like Goleta's.
PHOTOCHEMICAL OXIDANTS OTHER THAN OZONE
As mentioned previously, the scope of photochemical oxidants extends
to organic nitrates and other carbonyl compounds. Among the organic nitrates,
the one most often cited is peroxyacetylnitrate (PAN). Electron-capture
detector techniques applied to the gas chromatograph were used to measure
PAN concentrations in Los Angeles late in 1965.57
On each of 16 weekdays in September and 19 in October, seven measurements
were made. The daily means are plotted in Figure 4—38. Note that the PAN
concentrations are considerably below the total oxidant concentrations.
At the University of California at Riverside, the same technique was
used beginning in June 1966 to measure PAN. Hourly samples were collected
between 7:00 a.m. and 4:00 p.m. Figure 4-39 shows the PAN distribution to
exhibit double peaks, which sometimes occur with oxidant concentrations.1
Note that the relationship between the PAN curve and the oxidant curve in
Figure 4-39 is very similar to that in Figure 4-38. Figure 4-40 shows the
seasonal variations in PAN concentration and oxidant concentration for portions
of 1966 and 1967 at Riverside.
Aldehydes may also be thought of as photochemical oxidants. The
definition here becomes a bit hazy, because aldehydes in themselves are
photooxidative reactants as well as secondary pollutants that have adverse
health effects,, Referring to Figure 4-4, we note that aldehyde concen-
tration throughout the day in Rome seems to decay at roughly the same rate
as the nitric oxide concentration. It would be expected to track the
4-92
-------
AVERAGES:
19 WEEKDAYS, OCTOBER
16 WEEKDAYS. SEPTEMBER
HOUR OF DAY, PST
Figure 4-38. Variation in mean 1-h average oxidant and PAN concentrations,
by hour of day, in downtown Los Angeles, 1965. (Reprinted from
U.S. DREW.1)
4-93
-------
0.1*
HOUR OF DAY. PST
Figure 4-39.
Variation in mean 1-h average oxidant and PAN concentrations,
by hour of day, at the Air Pollution Research Center, Riverside,
California, September 1966. (Reprinted from U.S. DHEW.1)
4-94
-------
MONTHLY MEANS OF DAILY MAXIMUM
1-hour AVERAGE CONCENTRATIONS
MONTHLY MEANS OF 1-hour AVERAGE
CONCENTRATIONS
*OXIDANT BY MAST, CONTINUOUS
24 hours; PAN BY PANALYZER,
SEQUENTIAL, 6 a.m. TO
4 OR 5 p.m. ONLY.
OX/DANT*
o-
JUL. AUG. SEP. OCT. NOV. DEC. JAN. FEB. MAR. APR.
MONTH
1966 94* 1967
Figure 4-40.
Monthly variation in oxidant and PAN concentrations at the Air
Pollution Research Center, Riverside, California, June 1966-
June 1967. (Reprinted from U.S. DHEW.!)
4-95
-------
reactive fraction of the hydrocarbons, and this is also borne out approximately
3
by the Rome data. A maximum formaldehyde concentration of 39 yg/m (0.032 ppm)
was measured in Berlin, Germany, in 1967 on a street with high traffic
density.58
Measurements were conducted in Rotterdam during the period January-March
1973, to determine the ambient concentrations of aliphatic aldehydes. The
3
24-h average concentrations were around 5 yg/m , and the 8:00 a.m. - 4:00 p.m.
3
average concentration was 8.4 yg/m .59
Dickinson60 measured total aldehydes by the bisulfite procedure and
formaldehyde by thechromotrophic acid procedure.
Renzetti and Bryan61 measured total aldehydes, formaldehyde, and
3
acrolein. The maximal acrolein concentration observed was 25.2 yg/m
(0.011 ppm), and the maximal total aldehyde concentration was 0.36 ppm for
3
a 10-min sample. The formaldehyde concentration never exceeded 130 yg/m
(0.10 ppm).
Altshuller and McPhersori62 used spectrophotometry to analyze aldehydes
in the Los Angeles atmosphere in the fall of 1961. Table 4-16 shows the
diurnal variation in both formaldehyde and acrolein concentrations. Both
rise early, remain relatively constant throughout the day, and decrease in
the later part of the day. Acrolein apparently accounts for only about 10%
of the total olefinic aldehyde in the atmosphere, with most of the latter
concentration being accounted for by formaldehyde.
Another nonozone photochemical oxidant observed in urban atmospheres is
hydrogen peroxide. Bufalini et al.63»61+ found this compound to be present
at concentrations up to 0.04 ppm in the air in Hoboken, New Jersey, and up
to 0.18 ppm on a smoggy day in Riverside, California. Figure 4-41 shows
4-96
-------
Table 4-16
Average Aldehyde Concentrations by Hour in
Los Angeles, September 25 - November 15, 1961
a
Formaldehyde
Acrolein
Sampling
Time
7 a.m.
8 a.m.
9 a.m.
10 a.m.
11 a.m.
12 noon
1 p.m.
2 p.m.
3 p.m.
4 p.m.
No . Days
7
18
21
28
27
23
25
27
25
15
Average
Concentration, ppm
0.041
0.043
0.045
0.044
0.051
0.044
0.041
0.034
0.026
0.019
No. Days
2
3
3
5
5
3
7
5
4
5
Average
Concentration, ppm
0.007
0.009
0.009
0.008
0.008
0.005
0.008
0.007
0.004
0.004
aDerived from Altshuller and McPherson.62
4-97
-------
THICTMT
Figure 4-41. Measured oxidant at Riverside, Calif., August 6, 1970.
(Reprinted with permission from Bufalini et al.°3)
4-98
-------
that the diurnal hydrogen peroxide variation in Riverside on August 6, 1970,
nearly parallels that of total oxidant. Figure 4-42 indicates, however,
that on at least one occasion (August 11, 1970) it peaked as early as 10:30
a.m.
A dramatic departure of ozone measurements from total oxidant measure-
ments has been reported65 for the Houston, Texas, area. Side-by-side measure-
ments suggested that either method was a poor predictor of the other.
Consideration was given to known interferences due to oxides of nitrogen,
sulfur dioxide, or hydrogen sulfide, and the deviations still could not be
accounted for. In the worst case, the ozone measurements exceeded the
national ambient air quality standard for 3 h, and the potassium iodide
instrument read less than 15 ppb for the 24-h period. Sulfur dioxide was
measured at 0.01 - 0.04 ppm throughout the day. Even for a 1:1 molar
influence of sulfur dioxide this could not explain the low oxidant values.
Regression analysis was carried out to support the conclusion that the ozone
concentration is often much higher than the nonozone oxidant concentration.
DATA QUALITY
Chapter 6 covers most of the questions influencing data quality. At
least three factors enter into the selection of data: the instrumental
technique used for measurement, the exposure of the station and the location
of the sampling inlet, and the choice of standard calibration method.
Because of interaction with nitric oxide, local decreases in ozone con-
centrations can occur near large sources, such as power plants or freeways.
Consequently, ozone monitoring should not be undertaken near any of these
sources if a representative regional ambient concentration is desired. The
early portions of the Continuous Air Monitoring Program were evidently focused
4-99
-------
f.
I
1
I i r
HM.1
T«t OF DAT
Figure 4-42. Hydrogen peroxide concentrations at Riverside, Calif., in
August 1970. (Reprinted with permission from Bufalini et al.63)
4-100
-------
on the measurement of community exposure to pollutants, inasmuch as monitoring
sites were set up in the centers of urban areas. Because of the source
interaction referred to above, this often resulted in an underestimate of
the ozone problem. It is now known that stations around the perimeter of
a central business district or in surburban areas give higher ozone readings
because of the absence of local sources and because of the time required
for photochemical reactions to result in ozone buildups in the air. Dark
reactions in the inlet manifolding of the sampling train can also distort
an instrument's ozone measurements. Large-diameter high-velocity flow
systems made of low-reactivity materials should be used to avoid this
source of data distortion.
The problems with various primary calibration standards are still being
resolved. The earlier discussion of DeMore et_ ai_. indicates the work
undertaken by the GARB in comparing oxidant calibration procedures. It was
recommended that all oxidant analyzers in the California network be calibrated
by a secondary standard consisting of an ultraviolet ozone analyzer. The
primary standard recommended is ultraviolet photometry. These recommendations
have been adopted by the GARB. Potassium iodide, indicated as a second
choice, requires the application of a correction factor of 0.78.
The fraction of valid data days has been summarized by Altshuller55 for
the six CAMP sites. He noted that, in Cincinnati in the summers of 1969 and
1970 and in Denver in 1971, no results were reported for oxidant. At best,
for Washington, D.C., and Chicago, 70 - 75% valid data were obtained; at
worst, 40 - 50% valid data were obtained for summer months in Denver and
Cincinnati. The findings are summarized in Table 4-17. Altshuller called
attention to the large corrections that must be applied to colorimetric
4-101
-------
r^.
i — i
T-l
ca
H
CO
r-l
rQ
r-l
•H
ca
*
J-J
<
cd
4J
ca
o
4-1
c
nd
•H
£
T3
r-l
cd
Which \
r-l
O
<4-l
CO
>!
Percentage of De
CO
4-1
•H
O
cd
4J
60
-,
1
rJ
O
>4-l
ca
p
14-1
o
co
60
ca
c
0)
PM
co
cd
SH
>
OO
CN
i— 1
r-l
r-l
0
g-
<— 1
s
•-*
oo
r-l
IS
1— 1
vO
i— I
m
r-H
f-H
£
0
r--
00
2
m
oo
m
00
s
§
m
o
o
m
r-H
m
r--.
m
m
m
o
oo
o
m
oo
m
00
m
o
VD
m
m
0
oo
0
oo
Ln
o
m
oo
m
00
o1
m
/— N
O
m
i— i
m
ro
m
co
m
m
0
m
o
m
o
00
0
o
CO
o
oo
o
0
m
0
m
m
m
m
m
r-H
m
CN
m
m
o
VD
o
CN
in
i— i
o
o
in
m
i— t
0
ro
in
m
o
0
a\
Q
0
m
m
m
0
m
m
c
01
p
cd
•H
rC
a
rH
O)
cd
rH
•H
JS
PL,
CO
•H
jj
O
rJ
•
4-1
cn
•
•
P
ti
O
4-1
00
a
•H
r^
CO
^2
CO
6C
cd
M
CO
-------
oxidant analyzers for peaks that occur at night late in the fall and in
winter. He pointed out that the NO response correction becomes a major
x
part of the reading, because the instruments are functioning mostly as
analyzers of NO . What might be considered a valid data .base extends much
x
further back in time for the Los Angeles basin. A high fraction of the data
for nearly 20 years is available for that area. Before a final judgment on
the validity of the data, however, one must carefully assess the problems
of calibration consistency and monitoring station siting. It will be found
in most local monitoring systems outside California that the bulk of the
valid oxidant data has been obtained since 1970. Thus, in 1975, valid data
are available for only a few years, because of lags in data processing and
editing.
SUMMARY
In comparison with previously available material1 on atmospheric con-
centrations of photochemicaloxidants, we now have a far richer data base and
a deeper understanding of how to interpret the reported concentrations. The
recent information on hydrogen peroxide and the broader geographic coverage
of measurements abroad are examples of new data that have come to light.
Subtleties in future standard-setting must consider receptor damage in
terms of exposure location and time and receptor distributions and response
functions. The formula for damage function points up the need for improved
knowledge of spatial and temporal distribution. The use of second-to-worst
hourly readings for an ambient air quality standard must give way to a
specification stated in terms of a statistically defensible higher-frequency
event. This will reduce substantially the uncertainty inherent in confining
one's attention to the "worst case."
4-103
-------
Long-term trends in oxidant concentration cannot be identified with
nearly the degree of certainty that we might like. The data suggest a
decrease in oxidant concentrations in central-city areas and an increase in
downwind areas. Measurements of nonurban oxidant are exhibiting a higher
frequency of violations of the ambient standard than was once believed to
occur. Figure 4-43 shows typical ranges of concentrations in various regions
in and out of urban complexes,,
Probably the most critical question today regarding atmospheric concen-
trations of ozone and other photochemical oxidants is, "What fraction of the
observed values in each locale is susceptible to control by anthropogenic
emission reduction?" As brought out in this chapter, there is one school
of thought embracing the idea that nature frequently presents us with con-
centrations that exceed the U.S. national ambient air quality standards.
The other point of view is that global background ozone concentrations do not
exceed 0.05 - 0.06 ppm at the surface and that higher concentrations than
this have anthropogenic sources.
The data presented in the literature reviewed above support the second
point of view. The observation of ozone concentrations exceeding the ambient
standard in nonurban areas does not demonstrate that this is of natural
origin. However, the measurements in remote areas of the northern hemisphere,
compared with those in the continental United States, do support the thesis
that anthropogenic sources are involved in cases where the standard is
exceeded.
This is a very broad conclusion, and additional measurements must be
made. Some of this effort (which is goingon) should address the problem of
other pollutants and condensation nuclei that accompany the nonurban oxidant.
4-104
-------
-Rural (upwind) -
-Urban-
- Rural (downwind)-
Stratospheric
-transport effect
Natural
photochemical input
30-50
ppb
Loss to ground
and aerosols
100-500
PPb,
Urban
photochemistry/
Reaction
Net decay, bi-t effects
from long *erm
^photochemistry and
regional input
Loss to ground
and aerosols
-Rural (upwind) -
-Urban-
-Rural (downwind)-
Figure 4-43.
The tropospheric ozone cycle.
from Corn et al.66)
(Reprinted with permission
4-105
-------
Interpretation of these measurements will increase the specificity of
separating anthropogenic sources from natural background sources. Theoretical
assessments of the existing observations will shed light on the relative
roles played by stratospheric injection, plant emission, background methane,
and dry deposition on surfaces in the natural portion of the tropospheric
ozone cycle.
Geographically, our best measurements have focused on the Los Angeles,
California, region because of the severity of the problem there. The
Regional Air Pollution Study and its extensions will, it is hoped, supply
an additional rich data base Eor the St. Louis, Missouri, region. Airborne-
pollutant measurements aimed at specific experimental objectives are needed
in the central and eastern areas of the United States to broaden the foundations
of a national control strategy. Existing ground-based continuous monitoring
networks will not provide an adequate basis for the regional control of
oxidant. Ad hoc, one—shot aircraft measurements have led mainly to specu-
lation that can establish incorrect attitudes on the origin and fate of nonurban
ozone. In the vigor of environmental control efforts, incomplete data sets
have stimulated hasty targeting on specific sources (for example, rural
vehicular emission, power stations, trees, and frontal passages).
Another subject of recent interest has been the question of indoor-
outdoor oxidant concentrations. Available measurements and models suggest
that indoor exposures may be substantially reduced by appropriate choices
of ventilation systems, air filters, and interior surface materials. The
cost-benefit relationships may well have a great impact on future decisions
based on atmospheric concentrations of oxidant pollutants.
4-106
-------
Much care must be exercised in comparing atmospheric concentrations
between one place and another because of differences in primary
calibration techniques or in instrumentation. Chapter 6 summarizes
these problems in detail.
RECOMMENDATIONS
• Nonmethane hydrocarbons and both oxides of nitrogen should
be monitored concurrently whenever photochemical oxidant or ozone
is monitored.
• Photochemical oxidant monitoring stations should be sited
upwind and downwind from urban areas, as well as within those urban
areas, wherever possible.
• A common primary calibration standard should be established
for all monitoring networks.
• Documentation should be provided in each case to outline
the rationale for location and design of monitoring stations and
the rationale for data validation for photochemical oxidants.
• A clear indication of what constitutes background
concentrations of photochemical oxidants and ozone must be made,
in order to form the basis of emission control programs.
• The results of monitoring data must be generalized, in
order to relate air quality to emission in a stochastic fashion.
4-107
-------
REFERENCES
1. U. S. Department of Health, Education, and Welfare. Public Health Service
Environmental Health Service. Atmospheric photochemical oxidant con-
centrations, pp. 3-1--3-18. In Air Quality Criteria for Photochemical
Oxidants. National Air Polution Control Administration Publ. AP-63.
Washington, D. C.: U. S. Government Printing Office, 1970.
2. DeMore, W. B., J.C. Romanovsky, M. Feldstein, W. J. Hamming, and P. K.
Mueller. Comparison of Oxidant Calibration Procedures. California
Air Resources Board Ad Hoc Oxidant Measurement Committee Report,
February 3, 1975. (UNVERIFIED)
3. U. S. Environmental Protection Agency. Title 42--Public Health. Part 420.
Requirements for preparation, adoption, and submittal of implementation
plans. Federal Register 36:15486-15506, 1971.
4. U. S. Environmental Protection Agency. Progress in the Prevention and
Control of Air Pollution in 1973. Report to the Congress of the
United States, January 1974. pp. 41, 64. (UNVERIFIED)
5. U. S. Environmental Protection Agency. The National Air Monitoring Program:
Air Quality and Emissions Trends. Annual Report. Vol. 1. Monitoring
and Data Analysis Division, EPA-450/l-73-001-a. Research Triangle
Park: U. S. Envirormental Protection Agency, 1973.
6. Eschenroeder, A., J. Martinez, and R. Nordsieck. A view of future problems
in air pollution modeling, pp. 1013-1027. In Proceedings of the 1972
Summer Computer Simulation Conference. La Jolla, Calif.: Simulation
Councils, Inc., 1972. (UNVERIFIED)
4-108
-------
7. North Atlantic Treaty Organization. Committee on the Challenges of Modern
Society. Atmospheric concentrations, pp. 2-1--2-52. In Air Quality
Criteria for Photochemical Oxidants and Related Hydrocarbons. N.29.
1974.
8. Weber, E. Ministry o£ the Interior, Bonn, Federal Republic of Germany,
April, 1973. (UNVERIFIED)
9. Stasiuk, W. N., Jr., and P. E. Coffey. Rural and urban ozone relationships
in New York State. J. Air Pollut. Control Assoc. 24:564-568, 1974.
10. Coffey, P. E., and W. M. Stasiuk, Jr. Evidence of atmospheric transport of
ozone into urban areas. Environ. Sci. Technol. 9:59-62, 1975.
11
11. Lahmann, E., J. Westphal, K. Damschke, and M. Lubke. Kontinuierliche Ozon-
Messungen in einer verkehrsreichen Strasse. Gesund. Ing. 89:144-147,
1968.
12. Brasser, L. J. of the Research Institute for Public Health Engineering,
Delft, The Netherlands, April 1973. (UNVERIFIED)
13. Laboratoria di Inquinamento Atmosferico del CNR, Universita di Roma,
Rome, Italy, April 1973. (UNVERIFIED)
14. Muezinoulu, A. Scientific and Technical Research Council of Turkey, Ankara,
Turkey, May 1973. (UNVERIFIED)
15. National Research Council. Associate Committee on Scientific Criteria for
Environmental Quality. The Effects of Photochemical Smog. July 1974,
Draft. (UNVERIFIED)
ig Bilger, R. W. Oxidants and Their Precursors in the Atmosphere. National
Summary Report. Department of the Environment and Conservation,
Canberra City, Australia, February 19, 1974. (UNVERIFIED)
17. Report on the Problem of Photochemical Oxidants and Their Precursors in the
Atmosphere. Issued by the Organization for Econimic Cooperation and
Development, Air Management Sector Group, September 9, 1974. (UNVERIFIED)
4-109
-------
17a. Derwent, R. G., and H. N. M. Steward. Elevated ozone levels in the air o£
central London. Nature 241:342-343, 1973.
17b. Atkins, D. H. F., R. A. Cox, and A. E. J. Eggleton. Photochemical ozone
and sulphuric acid aerosol formation in the atmosphere over southern
England. Nature 235':372-376, 1972.
18. Miller, A., and D. Ahrens. Ozone within and below the west coast temperature
inversion. Tellus 22:328-339, 1970. (UNVERIFIED)
19. Miller, P/R,, M. H. McCutchan, dnd H.'P. Milligan. Oxidant air pollution ill
the Central Valley, Sierra Nevada foothills, and Mineral King Valley of
California. Atmos. Environ. 6:623-633, 1972.
20. Blumenthal, D. L., T. B. Smith, D. S. Ensor, S. 0. Marsh, R. B. Husar, W.
White, S. L. Heisler, and P. Owens. Three Dimensional Pollutant
Gradient Study--1972 Program. Report to the California Air Resources
Board. Altadena, Calif.: Meteorology Research, Inc., 1972. (UNVERIFIED)
21. Blumenthal, D. I., W. H. White, R. 1. Peace and T. B. Smith. Determination
of the Feasibility oE the Long-Range Transport of Ozone or Ozone
Precursors. EPA-450/3-74-061. Altadena, Calif.: Meteorology Research,
Inc., 1974.
22. Edinger, J. G. Preliminary Analysis of LARPP Data. Informal Report to
Coordinating Research Council, Inc., February 1, 1975. (UNVERIFIED)
23. Research Triangle Institute. Investigation of High Ozone Concentration in
the Vincinity of Garrett County, Maryland and Preston County, West
Virginia. Institute Final Report. EPA-R4-73-019. Research Triangle
Park, N. C.: Research Triangle Institute, 1973. 105 pp.
24. Edinger, J. G. Vertical distribution of photochemical smog in Los Angeles
basin. Environ. Sci. Technol. 7:247-252, 1973.
4-110
-------
25. Research Triangle Institute. Investigation of Ozone and Ozone Precursor
Concentrations at Nonurhan Locations in the Eastern United States.
EPA-450/3-74-034. Research Triangle Park, N. C.: Research Triangle
Institute, 1974. /~i4i pp.J7
26. Gloria, H. R. , G. Bradburn, R. F. Reinisch, II. N. Pitts, Jr., J. V. Bahar,
and L. Zafonte. Airborne survey of major air bases in California. J.
Air Pollut. Control Assoc. 24:645-652, 1974.
27 • Fern, W. J., and R. I. Brabets. Field Investigation of Ozone Adjacent to
High Voltage Transmission Lines. IEEE (Institute of Electrical and
Electronics Engineers) Transactions Paper T 74 057-6, Presented at the
PES Winter Meeting, New York, New York, Jan. 27-Feb. 1, 1974.
28. Roach, J. F., V. L. Chartier, and F. M. Dietrich. Experimental oxidant pro-
duction rates for EHV transmission lines and theoretical estimates of
ozone concentrations near operating lines, pp. 647-657. In IEEE (Insti-
tute of Electrical and Electronics Engineers) Transactions on Power
Apparatus and Systems, Mar-April, 1974.
29. Frydman, M., A. Levy, and S. E. Miller. Oxidant measurements in the vicinity
of energized 765 KV lines, pp. 1141-1147. In IEEE (Institute of Electri-
cal and Electronics Engineers) Transactions on Power Apparatus and
Systems, May-June, 1973.
30. Sherer, H. N., Jr., B. J. Ware, and C. H. Shih. Gaseous effluents due to
EHV transmission line corona, pp. 1043-1049. In IEEE (Institute of
Electrical and Electronics Engineers) Transactions on Power Apparatus
Systems, May-June, 1973.
31. Pitts, J. N., Jr. Air Pollutants and Public Health: Old Problems and New
Horizons for NO Control. Statewide Air Pollution Research Center
JC
Report No. 2. Riverside: University of California, 1975.
4-111
-------
32. Davis, D. D., G. Smith, arid C. Klauber. Trace gas analysis of power plant
plumes via aircraft measurement; CL , NO and S0_ Chemistry. Science
J X Z
186:733-736, 1974.
33. Davis, D. D. Atmospheric Gas Phase Oxidation Mechanisms for the Molecule
S02- Chemistry Department. College Park: University of Maryland,
1974. (in press)
34. Satinders, R. A., J. R. Griffith, and P. E. Saalfeld. Identification of
some organic smog components based on rain water analysis. Biomed.
Mass Spectrosc. 1:192-194, 1974.
35. Blumenthal, D. I., and W. H. White. The Stability and Long Range Transport
of Ozone or Ozone Precursors. Paper 75-07 Presented at 68th Annual
Meeting of the Air Pollution Control Association, Boston, Massachusetts,
June 15-20, 1975.
36. Rubino, R. A., L. Bruckman, and J. Magyar. Ozone Transport. Paper 75-07
Presented at 68th Annual Meeting of the Air Pollution Control Associa-
tion, Boston, Massachusetts, June 15-20, 1975.
37. Sticksel, P, R. The Stratosphere as a Source of Surface Ozone. Paper 75-07.6
Presented at 68th Annual Meeting of the Air Pollution Control Association,
Boston, Massachusetts, June 15-20, 1975.
38. National Academy of Sciences. National Academy of Engineering. Coordinating
Committee on Air Quality Studies. Air Quality and Automobile Emission
Control. Vol. 3. The Relationship of Emissions to Ambient Air Quality.
U. S. Senate Committee Print Serial No. 93-24. Washington, D. C.:
U. S. Government Printing Office, 1974. pp. 76-82.
39. Fabian, P., and P. G. Pruchniewicz. Miriodonal distribution of tropospheric
ozone from ground based registrations between Norway and South Africa.
Pure Appl. Geophys. 106-108:1025-1035, 1973.
4-112
-------
40. Hering, W. S., and T. R. Borden, Jr. Ozonosonde Observations over North
America. Volumes I, IT, and III. Air Force Cambridge Research
Laboratory, Report AFCRL-64-30, January, 1964.
41. Fabian, T., and P. C. "Pruchniowicz. Ozonesondo Observations over North
America. Air Force Cambridge Research LabornLory, Report AFCRL-64,
January 1964. p. 1027.
42. Neligan, R. E., and R. M. Angus. The Validity of the Strategy of Linear
Rollback of Hydrocarbons to Achieve Oxidant Air Quality Standards.
Presented at the UC-ARB Conference Technical Bases for Control Strata-
gies of Photochemical Oxidant: Curent Status and Priorities in
Research, December 16-17, 1974, Riverside California.
43^ Mueller, F. X. , L. Loeb, and W. H. Mapes. Decomposition rates of ozone in
living areas. Environ. Sci. Technol. 7:342-346, 1973.
44. Sabersky, R. H., D. A. Sinema, and F. H. Shair. Concentrations, decay rates,
and removal of ozone and their relation to establishing clean indoor
air. Environ. Sci. Technol. 7:347-353, 1973.
45. Shair, F. H., and K. L. Heitner. Theoretical model for relating indoor
pollutant concentrations to those outside. Environ. Sci., Technol.
8:444-451, 1974.
46. Hales, C. H., A. M. Rollinson, and F. H. Shair. Experimental verification
of linear combination model for relating indoor-outdoor pollutant con-
centrations. Environ. Sci. Technol. 8:452-453, 1974.
47. King, C. S. Ozone and air conditioning. Royal Soc. Health J. 93:84-66, 1973,
48. Pitts, J. N., Jr., A. C. Lloyd, and J. L. Sprung. Chemical reactions in
urban atmospheres and their applications to air pollution control stra-
tegies, pp. 27-61. In Proceedings of the International Symposium on
Environmental Measurements, Geneva, Oct. 2-4, 1973.
4-113
-------
49. Kinosian, J. R., and S. Duckworth. Oxidant Trends in the South Coast Air
Basin 1963-1972. California Air Resources Board Report, April 1973.
5Q^ National Academy of Sciences. National Academy of Engineering. Coordinating
Committee on Air Quality Studies. Air Quality and Automobile Emission
Control. Vol. 3. The Relationship of Emissions to Ambient Air Quality.
U. S. Senate Committee Print Serial No. 93-24. Washington, D. C.:
U. S. Government Printing Office, 1974. pp. 65-88.
51. Heitner, K., and J. E. Krier. Outline of an Approach to Management Stand-
ards. California Institute of Technology, Environmental Quality Lab-
oratory, Memorandum Mo. 13, January 1974. p. 11.
52. Tiao, G. C., G. E. P. Box, M. Grupe, S. T. Liu, S. Hillmer, W. S. Wei, and
W. J. Hamming. Los Angeles Aerometric Ozone Data 1955-1972. University
of Wisconsin, Technical Report No. 346, Oct. 1973.
53 Chameides, W., and J. C. G. Walker. A photochemical theory of tropospheric
ozone. J. Geophys. Res. 78:8751-8760, 1973.
54. Ripperton, L. A., H. Jeffries, and J. J. B. Worth. Natural synthesis of"
ozone in the troposphere. Environ. Sci. Technol. 5:246-248, 1971.
55. Aitshuller, A. P. Evaluation of oxidant results at CAMP sites in the
United States. J. Air Pollut. Control Assoc. 25:19-24, 1975.
56. Birakos, J. N., Ed. 1074 Profile of Air Pollution Control. County of Los
Angeles Air Pollution Control District, 1975.
57. Bartel, A. W., and J. W. Temple. Ozone in Los Angeles and surrounding
areas. Ind. Eng. Chem. 44:857-861, 1952.
4-114
-------
58. Lahmann, E. Luftverunreinigungen durch den Kraftverkehr. Bundesgosund-
heitsblatt 12:284-286, 1969.
59. Brasser, I. J. Research Institute for Public Health Engineering, TNO
Delft, The Netherlands, April, 1973. Private communication to the
NATO/CCMS Panel on Air Quality Criteria.
60. Dickinson, J. E. Air Quality of Los Angeles County. Los Angeles County
Air Pollution Control District Technical Progress Report, Vol. 2,
February, 1961.
6i. Renzetti, N. A., and R. J. Bryan. Atmospheric sampling for aldehydes and
eye irritation in Los Angeles smog. J. Air Pollut. Control Assoc. 11:
62. Altshuller, A. P., and S. P. McPherson. Spectrophotometric analysis of
aldehydes in the Los Angeles atmosphere. J.' Air Pollut. Contr. Assoc.
13:109-111, 1963.
63. Bufalini, J. J., B. W. Gay, Jr., and K. L. Brubaker. Hydrogen peroxide
formation from formaldehyde photooxidation and its presence in urban
atmospheres. Environ. Sci. Technol. 6:816-821, 1972.
64. Gay, B. W., Jr., and J. J. Bufaiini. Hydrogen peroxide in the urban atmos-
phere. Adv. Chem. Ser. 113:255-263, 1972.
65. Severs, R. K. Simultaneous total oxidant and chemiluminescent ozone meas-
urements in ambient air. J. Air Pollut. Control As ROC. 25:394-396, 1975.
66. Corn, M., R. W. Dunlap, L. A. Goldmuntz, L. H. Rogers. Photochemical
oxidants: Sources, sinks, and strategies. J. Air Pollut. Control
Assoc. 25:16-18, 1975.
4-115
-------
CHAPTER 5.
MODELS FOR PREDICTING AIR QUALITY
Trends in air pollutant concentrations can be predicted with simple
empirical models based on atmospheric and laboratory data. Concentrations
of nonreactive pollutants from point sources can be predicted with accuracy
to within a factor of 2; predictions are more likely to be too high than
too low, especially predictions of concentration peaks. Concentrations
of reactive pollutants, such as ozone and other photochemical oxidants,
can be predicted reasonably well with photochemical-diffusion models when
detailed emission, air quality, and meteorologic measurements are available;
most such predictions of air pollution in Los Angeles have been accurate to
within 50%.
Statistical models based on data correlations and on Markov chains
are being actively developed and their fidelity evaluated by several research
groups. Photochemical-diffusion models based on deterministic equations are
also being developed, but because of their complexity will probably be used only
as research tools for some time.
-------
As expressed in Eqc 1 of Chapter 4, the spatial and temporal distributions
of pollutants must be known to assess their damage to receptors. Traditionally,
monitoring station measurements are used to estimate the concentrations for
entire regions. Statutory mandates, however, require the prediction of
future concentrations under differing emission conditions0 More and more
refined predictions will be required as the standards become more specific.
An air quality model is a method of relating air quality to emission
under specific environmental conditions. There are many types of air quality
models, and the purposes of this chapter are to describe models that are avail-
able for the analysis of photochemical oxidants and to give some information
on how well the models perform,,
We must first gain a perspective on present modeling capabilities,, The
photochemical-diffusion computer programs are aimed at prediction of ozone and
nitrogen dioxide concentrations. It must be understood that ozone is used as
an indicator or as a surrogate for other oxygenated organic compounds or
radicals that may actively cause adverse effects,. Most models do not treat
aldehydes and oxygenated organic aerosols specifically, although they do
address the problems of peroxy-radical and hydrogen peroxide concentrations
in some cases.
Because of the dominance of distributed sources over local single sources
in the production of photochemical oxidants, point-source models will not be
discussed here. Related research regarding the measurement of diffusion or
the development of atmospheric chemical submodels will not be emphasized.
5-la
-------
Chapter 2 is devoted to the chemical processes that govern atmospheric trans-
formation and removal, and this aspect of the models will not be repeated here0
Numerous reviews of air quality models have appeared in the literature
in the last several yearsc Wanta1 considered the meteorologic role in air
pollution, including the effects of vertical temperature structure, topography,
windfields, and dispersion mechanisms. He tabulated in detail numerous
characteristics of the mathematical air pollution models that existed in 1967„
The relationship of meteorology to air pollution was also reviewed in papers
by Pack2 and Neiburger03 Theories of diffusion in the lower atmosphere were
outlined by Gifford,^ who stressed the empirical aspects of both differential
equation diffusion theory and Gaussian diffusion formulas, traced the
historical connections between the two, clarified their relationship to
atmospheric turbulence, and reviewed cases of plume formulas to cover special
phenomena, such as plume fluctuation, looping, coning, fumigation, and lofting.
Various meetings on urban diffusion models have been held in recent years.
One was a symposium^ sponsored by the Environmental Protection Agency0 Its
proceedings include studies of Gaussian plume and puff modeling techniques
available in 1969. Each paper on a specific model gave some detail as to the
mathematical assumptions and the types of measurements that were used to test
it. Several participants noted a need to develop finite-difference
numerical techniques to handle the nonlinearities of atmospheric
reactions. Another review resulted from a series of meetings and working
groups related to Project Clean Air, which was carried out by the University
of California in 1969 and 1970. One of the task force reports6 is especially
helpful in its review of mathematical simulation modeling up to 1970. The
5-2
-------
stated purpose of the review was to identify future research needs for
California's Project Clean Air. A tabulation summarized fifteen simulation
schemes by commenting on each of nine points of information, including
statements of working equations and quantitative aspects of cell size, time
resolution, and verification tests. It is gratifying to note that sub-
stantial progress has been made on each of the 10 recommendations given for
future research.
Another useful collection of papers is the proceedings of the second
meeting of the Expert Panel on Air Pollution Modeling, sponsored by NAT007
The volume contains 15 papers from both the first and second meetings of the
panel. Three of the main topics are: air quality modeling projects that were
going on in Ankara, Turkey, and Frankfurt, Germany; applicability of physical
models that use hydraulic or aerodynamic replications of flow fields; and
problems in atmospheric chemistry of pollutants, with particular emphasis on
photochemical transformation. In his review of atmospheric transport processes,
Reiter8 stressed the behavior of the diffusion equations under the combined
effects of frictional, buoyant, and Coriolis forces. The distortion of the
velocity profile as it approaches the geostrophic wind was discussed<> The
main approach was theoretical and was based on K theory, but Gaussian plume
results were also cited. The fluid dynamic limitations of the Gaussian plume
formulas were reviewed critically, and extensive references for various special
aspects of the formulas were given.
In another review, Hoffert9 discussed the social motivations for modeling
air quality for predictive purposes and elucidated the components of a modelo
Meteorologic factors were summarized in terms of windfields and atmospheric
stability as they are traditionally represented mathematically0 The species-
5-3
-------
balance equation was discussed, and several solutions of the equation
for constant-diffusion coefficient and concentrated sources were suggested.
Gaussian plume and puff results were related to the problems of developing
multiple-source urban-dispersion models. Numerical solutions and box models
were then considered. The review concluded with a brief outline of the
atmospheric chemical effects that influence the concentration of pollutants
by transformation.
Motivated by statutory mandates for environmental evaluation of trans-
portation systems, Darling10 solicited information from each originator of an
air pollution model by a questionnaire advertised in the Commerce Business
Daily, a widely circulated publication in the United States„ Of the 78 ques-
tionnaires on models sent out, 44 were completed and returned„ The questionnaire
dealt in some detail with computer programs involved with each model, in addition
to the analytic foundations of each approach. Principles, implementation,
applications, and validation were discussed in the report. Whenever infor-
mation was available, there was comparative analysis of the models,, An
important conclusion of the work was that in 1972 there had been very little
performance evaluation of modeils related to transportation-generated air
pollution. Johnson11 reviewed EPA programs in air quality simulation
modeling in 1972, covering the various policy questions that models can help
to answer, summarizing modeling approaches, and outlining advantages and
disadvantages of the various techniques. The mission of UNAMAP (User
Network for Applied Modeling of Mr Pollution) was described as a system
that would provide easy user aiccess to the models for practical applications.
5-4
-------
The third meeting of NATO's Expert Panel on Air Pollution Modeling12
updated reports of modeling efforts for Cologne, St. Louis, Milan, Ankara,
Frankfurt, Stockholm, Oslo, and Manchester. It also included research papers
on model-related meteorologic topics and reviews of various national programs
and facilities represented by the participants. The fourth meeting of the
panel13 departed from the format of the others, in that it consisted of five
workshop sessions, which covered applications of modeling and user's needs,
validations of air quality simulation models, regional air pollution studies,
empirical-statistical modeling of air quality, and the question of simplicity
versus sophistication in air quality modeling. Each workshop discussion was
summarized in narrative form, and several prepared papers were included after
the workshop sessions.
Lamb and co-workers14 reviewed techniques of diffusion modeling for air
quality with relation to transportation-generated pollution. They discussed
the theory and structure of models, presented a series of tabulated comparisons,
analyzed the function and design of each model, and offered simple diagrams to
illustrate the functions and problems of the various techniques. The report
was intended to survey a great deal of unpublished material and therefore was
important in bringing the earlier surveys up to date (to about mid-1973)o
Air quality simulation models for photochemical pollutants were reviewed
by Sklarew1 for a new edition of Air Pollution. Some of the models developed
for simulating photochemical smog were reviewed from the viewpoints of module
logic and evaluation results. The Los Angeles-based developments were outlined,
including the format and preprocessing of emission inventory data and meteo-
rologic data. Lumped-parameter chemical approaches were described, and smog
chamber kinetics validations were outlined.
5-5
-------
One of the purposes of this chapter is to add recent material to that
collected in the reviews just described. In contrast with previous reviews,
however, this chapter will emphasize the critical evaluation of performance,.
The sections that follow deal with objectives of models (from research to
applied control systems), the elements of schemes for predicting air quality,
specific methods of modeling, and the evaluation of prediction techniques.
OBJECTIVES OF MODELS
At the heart of the problem of relating improvements in air quality to
reductions in pollutant emission is a reliable method of prediction. Only
with such a method can there be rational planning for air pollution control
through regulation of transportation, indirect sources, and stationary sourceso
Decision-makers need it as a tool and must specify it in their regulations.
Otherwise, their administration of an air quality plan would be based on sheer
guesswork tempered by political negotiation.
Pollutant concentrations are related to sources under specified
meteorologic conditions by using what is called an air quality model.
Models vary from simple arithmetic exercises to complex computer simulations.
There are many paths to the needed answers. Which of the available methods is
appropriate depends on the specific problem. Most agencies charged with en-
forcement of air quality rules have used only the most rudimentary models.
In some cases, short deadlines have forced the situation. But the recognition
of the social costs of air pollution control has led to deeper interest in
scrutinizing the results of air quality models, and it is not likely that
casual calculation will be accepted by regulatory agencies—especially if the
results (in the form of abatement strategies) involve severe socioeconomic
dislocation and large financial outlays.
5-6
-------
In examining how well a model meets its objectives, one must ask several
questions:
• Who will use the model?
• What questions must the model be able to answer?
• How well is the model adapted to the intended application?
• Has the model been thoroughly tested against an adequate data base
for validations?
• How readily accessible is the model with respect to practical calcu-
lations and computer implementation?
This section examines various purposes of developing and using air quality
models.
Scientific Purposes
In any environmental management plan, it is essential to understand the
pertinent quality criteria and their relationship to the variables that can be
manipulated directly. The construction of mathematical models imposes the
requirement of a logical framework that connects causes and effects, identifies
all pertinent variables, and defines their interrelationships. Even if a
particular technique never becomes practical, the discipline imposed by the
logical structure is valuable in highlighting important relationships„
Methods for predicting air quality were first applied in conjunction with
field measurement programs and routine monitoring programs. Such applications has
emphasized both the strengths and the weaknesses of various modeling schemes
and eluciadated the main points of technical understanding that must underlie
the establishment of control systems. Considering the responses of models to
various inputs, "parametric sensitivity evaluations" carried out in parallel with
5-7
-------
supporting field programs will continue to be helpful in the development of the
understanding on which pollution abatement plans can be based.
Experience has followed an iterative pattern in playing the model exercises
against field measurements. Usually, the first indication of the relative
importance of variables is seen in bodies of observational data. The next
step is to build a model on the basis of either intuition or a deterministic
physical equation that reflects the trends seen in the data. The model is then
used for the range of conditions in the data base, and uncertainties as to the
correctness or completeness of the model become evident. The questions that
arise can usually be answered only through further field experimentation
Thus, the models themselves are used in the design of both laboratory and field
experiments that will ultimately provide a basis for the improvement of the
modeling art.
Regulatory Purposes
Long-term air quality forecasts are implicit in any scheme that is des-
igned to improve the atmospheric environment through specific sanctions on
primary pollutant emission. The writing of legislation and regulations is an
obvious application of air quality modeling. For example, vehicular emission
controls are specified in the U. S. Clean Air Amendments of 1970. Originally,
a modeling scheme was used to specify the control that would be required to
achieve improvements in air quality. One of the difficulties with such legis-
lation is that the ambient air quality standards are set by an administrative
process, whereas the vehicular emission standards are set by the original Iaw0
In this case, the extent of control was specified before the establishment of
the target reduction of pollution for the future.
5-8
-------
A broader regulatory application of air quality modeling is the examination
of regional plans for abatement of air pollution. Ideally, each plan should
be tested by a reliable model that will tell what the pollutant concentrations
will be in the future if emission is reduced as specified by the plan,,
Therefore, alternative approaches must be evaluated both with respect to their
effectiveness in cleaning the air and with respect to their social costs, and
cost models and possibly even damage models must be used concurrently with the
air quality model. Only in this way can a least-cost abatement strategy be
implemented.
The successful application of quantitative predictions to the design of
strategies requires full cooperation of the scientific community with the
decision-makers. Regulatory and legislative bodies need the most reliable
tools available to assess the impact of their decisions. The assessment of
environmental impact is an integral part of the process of engineering design,
whether the object of the process is a steam generator, a highway, or an
airport,, As a result of the new requirements, many branches of government—
not only the EPA—must be able to predict air quality.
Urban Planning
Human land-use patterns and meteorologic conditions both directly deter-
mine the degree of air pollution. Land-use planning includes the design of
transportation systems. Air quality models must be able to predict the
pollution that will be caused by various patterns of land use. In this appli-
cation, the predictive scheme is used inversely; i0e0, it has to answer the
question: Given air quality standards, what patterns of land use are acceptable?
5-9
-------
Experience has shown that conventional wisdom can be dangerously misleading
in regard to the impact of various land—use and transportation plans on air
pollution. In the models to evaluate proposed changes, one gains an under-
standing of the sensitivity of environmental variables to different types of
facilities and to where they are placed.
Emerging from analyses like these is a new concept of population holding
capacity,, Earlier notions of maximal population density were based on avail-
ability of space for transportation systems, subsoil conditions related to the
support of buildings, availability of off-street parking space, and many other
considerations. With the new set of requirements imposed by environmental
quality standards, the earlier criteria will be superseded, in effect, by the
output of a model that specifies how great a population can occupy a given
regionjand where and when. In many urban areas, air quality may already be a
controlling factor in the determination of this holding capacity. Hence,
accurate forecasts are important.
Episode Control
Systems are now being devised that require a real-time system that uses
model logic to control emission sources to maintain acceptable air quality for
all meteorologic conditions. In the United States, such control systems may
be needed for large combustion sources of air pollution. Such systems use
meteorologic and power-demand data to produce load schedules by fuel type.
Indeed, the concept of computer-aided controls could be used for the control
of sources throughout an urban area, as has been proposed in some prefectures
in Japan. For real-time warning systems or control systems, the air quality
model provides the logic in the feedback loop that links air quality improvement
5-10
-------
to emission reduction under specific predicted meteorologic conditions. With
the coming need for flexibility in fuel logistics, these systems may become
more and more prevalent.
The operational logic for an episode control system must have much finer
resolution and higher reliability than that for long-term applications„ For
forecasts covering a decade or two, progress in a cleanup campaign can be
closely monitored, and emission controls can be continually adjusted, Episode
management, however, is far more difficult. One the one hand, if too many
errors are made by the system, unduly high public exposure to harmful pollutants
will result. On the other hand, if the episode control system generates too
many false alarms, the social costs in industrial operational modification
will become so onerous that public support for the system may fade.
Implicit in the success of any episode control system is the ability to
predict weather accurately. Conventional applications of air quality models
that use a wide variety of meteorologic information require that such details
as wind speed and direction and mixing depth be reliably predicted. Some of
the statistical-empirical models that are discussed below obviate such a large
mass of prediction by selecting a few sensitive measurements and bypassing the
detail of deterministic logic. The episode-control application of models
clearly illustrates the need to develop a multiplicity of air quality models.
For example, a model that finds a natural application in the regulatory or
planning process may be extremely poorly suited to the control of discrete
episodes.
5-11
-------
ELEMENTS OF AIR QUALITY PREDICTION SCHEMES
Structure of Deterministic Models
All deterministic schemes have some elements in common. The completeness
or detail of any of these elements varies greatly from model to model, but a
diagrammatic representation of the basic structure will clarify the relation-
ships among the various techniques to be presented. Figure 5-1 shows that
three streams of input information enter the preprocessing module, and the
background and initial pollutant concentrations may become unimportant if the
simulation is run for a long enough period and covers a large enough area.
The preprocessing module renders this information useful for the main compu-
tational part of the air quality model, transmitting to the main program the
various kinds of data shown in Figure 5-1. Depending on the complexity of
the logic, one or more of these kinds of data may be unnecessary. Pollutant
concentrations constitute the output of any of the models. The form of this
output varies in detail, ranging from a single concentration averaged over a
long period for an entire region to hourly maps of concentration in three
dimensions over the region.
Classification of Types
There are many ways of categorizing air quality models. One differentiation
is between statistical and deterministic models. The structure of statistical
models is based on the patterns that appear in the extensive measured data. The
structure of deterministic models is based on mechanistic principles wherever possible.
Most deterministic models contain some degree of empiricism. For example,
few models, if any, use turbulent-diffusion formulations that are based on
first principles, but rather use measured values of dispersion,, The same is
5-12
-------
Source Emission
Data
Background and __
Initial Pollutant
Concentrations
Meteorologic Data
Preprocessing
Module
Data Transmission;
Initial and Boundary
Conditions
Diffusion and Windfield
Input Data .
Photochemical Rate
Input Data .
Emission Intensity
Input Data ,
Model
Logic
Concentration
Field Output
Figure 5-1. Diagram of elements of deterministic model.
5-13
-------
true in regard to the atmospheric chemistry of photochemical formulations. A
common caution is that, if too many disposable quantities exist in a deter-
ministic model, it becomes a statistical model in disguise, because all it
accomplishes is a version of curve-fitting.
Another classification of model is related to the time and space scales
of interest. Ambient air quality standards are stated for measurement averaging
periods varying from an hour to a year. However, for computational purposes,
it is often necessary to use periods of less than an hour for a typical
resolution-cell size in a model. Spatial scales of interest vary from a few
tenths of a meter (e.g., for the area immediately adjacent to a roadway) up
to hundreds of kilometers (e.g., in simulations that will elucidate urban—rural
interactions). Large spatial scales are also warranted when multiday simula-
tions are necessary for even a moderate-sized urban area. Under some climato-
logic conditions, recirculations can cause interaction of today's pollution
with tomorrow's. Typical resolution specifications couple spatial scales with
temporal scales. Therefore, the full matrix of time scales and space scales
is not needed, because of the dependence of time scales on space scales.
Some typical categories by scale are as follows:
• A roadway impact model in the microscale ("0.1 km; "10 min-1 h)0
• Large-point-source or indirect-source model (~10 km; ~1 h) .
• Urban regional scale model (~50 km; ~1 h).
• Urban-rural regional scale model (~300 km; 1 h to 1 week).
The reviews by Johnson11 and by Seinfeld16 give helpful guidelines in the
classification of models by space and time scale.
5-14
-------
Perhaps the most fundamental method of classifying models is by methodology.
Examples of specific methods will be discussed later, but a brief summary is
appropriate here0 Great emphasis on historical data as embodied in empirical
1 7 ^O 7
formulas is found in the methodology of statistical and rollback models.1
Rollback models embody the principle that reductions in emission are reflected
by improvements in air quality, as may be shown by a straight line, a curved
line, or a complex surface that expresses some proportionality relationship,,
These models work best if the geographic and temporal distribution of the
emitters is not changed. The straight-line versions can apply only to pollutants
that do not undergo chemical transformations in the atmosphere,. An example
of a proportional linear rollback is a rapidly instituted retrofit control of
carbon monoxide emission imposed on an entire vehicle population. Corrections
for irregularities in distribution caused by nonuniform growth, however, have
been suggested.22'23 The great advantages of the linear rollback model are its
use of aggregated emission statistics and its mathematical simplicity. Sug-
gested improvements of the rollback to account for chemical change have been
based on the use of air quality data or smog-chamber data.
Dispersion models may take the form of a simple box model28"32 or a Gaussian
formula. Dispersion models that use superposition of Gaussian plumes improve
the rollback approach for nonreacting species by accounting for geographic
distribution of emission sources. Emission generally consists of an array of
area elements or effective point sources, each characterized by an output
intensity. The plumes from the several sources contribute in an additive
manner to the pollution at any downwind field point. Summation of the con-
tributions over all sources constitutes the superposition aspect of the
5-15
-------
approach. Computationally, this is convenient, because it allows sequential
consideration of each source element, but the same feature renders the method
inapplicable to multireacting systems undergoing chemical transformations.
The Gaussian puff approach removes the limitation of steady-state assumptions
from the calculation and treats discrete emissions as puffs that spread
according to the Gaussian law.
Special balance equation models combine the effects of diffusion, advection
and some chemistry and normally may use finite difference techniques for the
33-44
solution. Currently, these are the most elaborate simulations applied
to air quality analysis. Both time and space are subdivided into cells in
these models, so that the cumulative effects of emission, transport, and (when
chemistry is included) reactions are simultaneously accounted for. The potential
danger in using this kind of model is that the available data base will be
outstripped by mathematical detail. When that happens, the large volumes of
data output can delude the user into a high degree of confidence when, in
reality, only very sparse data bases are available for verifying the model.
The advantage of finite-difference models is the potentially greater fidelity
where greater detail exists in the input data base and in the validation testing.
Emission Description
The quantitative expression of the introduction of primary pollutants into
the atmosphere is basic to any air quality model. Emission is most generally
described as a geographic, temporal, and chemical distribution that requires
a rather massive array of numbers. Some simple models need only the aggregated
numbers found in ordinary tabulations of emission inventory (e.g., kilograms
per day of carbon monoxide, NO , etc., for a large air quality control region).
X
5-16
-------
A geographic breakdown of emission is needed, however, if we are to determine
key "hot spots" of high pollution to which ambient air quality standards are
directed. This spatial distribution of emission is often given on a grid or
on a traffic-flow network system. Similarly, the hourly variation in emission
must be known if we are to predict the peak hourly averages that are often
mentioned in the standards.
For purposes of characterization, emission sources are generally divided
broadly into stationary and mobile or transportation sources. Stationary
sources are further divided into point and area emitters. Typical point
sources must include petroleum refineries and electric power plants„ Commercial
solvent emission and gasoline marketing emission may generally be represented
as area sources. A third category has been defined recently—indirect
sources —which takes into account hybrid sources like sports arenas and
shopping centers. These have fixed locations, but the traffic that is gen-
erated by or attracted to such a facility constitutes the source of emission
that is combined with the emission of the facility itselfo
Gathering emission data and putting them in condition for use in air
quality models are often among the most tedious and time-consuming parts of
their handling. For this reason, the preprocessing module is identified as a
separate automatic operation in the procedure outlined in Figure 5-1.
Transport Formulation
The movement of pollution from one place to another and its dilution by
atmospheric mixing are both based on the meteorologic conditions of the airshed
in question. Air flow patterns are in turn based on the interaction of the
large-scale flow with the topographic details of the region, with regard to
5-17
-------
altitude variation, roughness of surface, and heating characteristics. The
part of the flow-surface interaction that influences the degree of pollution
must be taken into account in a model. Rather than computing the local
weather as part of the prediction, most models use meteorologic measurements
to construct atmospheric flow fields that represent, on a local scale, the
driving factor of the transport mechanisms.
With the velocity field and the atmospheric dispersion mechanisms given,
the basic equation is that for mass balance for an individual species, which
can be expressed in the following forms
3c 3 8(c v )
j + I j jfc = w + S (j = 1, 2, 3, ..co,s), (1)
3t H=l 3x j j
a
where
c = mass concentration of jth species,
j
t = time,
H = an index referring to each coordinate direction,
v
j£ = velocity of jth species in £th direction,
x
distance in &th direction,
w
j = net molar production rate of jth species per unit volume by chemical
reaction,
S
j = source strength for emitters of jth species at some location above
the ground, and
s = number of species 0
5-18
-------
The derivation of the mixture-balance laws has been given for a binary
mixture by Chapman and Cowling.52 Its generalization to multicomponent
mixtures, as in Eq. 1 , uses a determination of the invariance of the Boltzmann
equation. This development has been detailed by Hirschfelder et a.U 53 These
derivations were summarized in the notes of Theodore von Ka'rma'n's Sorbonne
lectures given in 1951-1952, and the results of his summaries are stated in
Penner's monograph.54 For turbulent flow, the species-balance equation can be
represented in the Boussinesq approximation as:
35-
_JL
9t
£=1
3c
£ 3x
£
3
9x
£
92 '
K j_
££ 9x
£
+ w + S
j J
(j = 1,2,3,..,o,s), (2)
where overbars denote time-averaged values and K is the eddy diffusivity
££
that relates the flux of a diffusing species to the species concentration
gradient, both acting in the same direction. The Boussinesq approximation is best
applied to large scale turbulent flow or motions typifying the problems in urban
and regional areas.
The development of this form of the equation is given by Bird et_ al.
The species-mass-conservation models use numerical integrations of various forms
of these equations.
The Gaussian plume formulations, however, use closed-form solutions of the
turbulent version of the species-mass-balance equation subject to simplifying
assumptions. Although these will not be treated further here, their description
is included for comparative purposes. The assumptions are: reflection of
species off the ground (that is, zero flux at the ground), constant value of
vertical diffusion coefficient, and large distance from the source compared
with lateral dimensions. This Gaussian solution to the species-mass-balance
5-19
-------
equation is obtained under the assumption that chemical transformation source
and sink terms are all zero. In some cases, an_ exponential decay factor is
applied for reactions that obey first-order kinetics. A typical solution
(with the time-decay factor) is:
c (x ,x ,x ) =
j 123
Q /P
j
2ira a v
2 3 1
exp
X
- 1 - 1
V T 2
1
X
2
a
2
- 1
2
x -h
3
a
I 3 ,
+ exp
- 1
2
x +h
3
a
I 3
-
•
(3)
where
Q = emission rate of jth species,
j
p = air density,
h = source height,
T = chemical decay time for first-order reaction, and
1/2
a - (2K x /v)
ft, i!i I I
The three coordinate directions—1, 2, and 3—are taken to be downwind, cross-
wind, and vertical. The origin is fixed to the ground at the location of the
source.
5-20
-------
Physicochemical Transformation Simulation
The atmospheric chemical processes undergone by most pollutants are not
readily describable by first-order kinetics. Hence, the simple Gaussian plume
solution in Eq. 3 is inapplicable in most cases where physicochemical trans-
formations significantly alter concentrations on a time scale or space scale
appropriate to an urban airshed,,
The general case must be solved by numerical integration with finite-
difference schemes or other approaches to the solution of Eq0 2 for the species
of interest. As written, this equation requires that the partial-differential
equation be solved for each species in the reactive mixture. In reality,
however, the number of partial-differential equations that must be solved can
be reduced by imposing stationary-state assumptions. That is, some species
are so reactive that their rate of production nearly equals their rate of
depletion, and these rates may be effectively equated. This being the case,
algebraic expressions are used to relate the stationary-state specie's concen-
tration to all the reactive compounds and radicals that are responsible for
simultaneous
its production and removal. For multicomponent mixtures, these are/nonlinear
algebraic equations that must be solved by numerical techniques, such as
Newton's method for finding roots of equations»
A generalized notation for representing chemical reactions is:
k
s is
I v A -> £ v' A (i = 1,2,3 ,r), (4)
j=l ij j j=l ij j
5-21
-------
where, for the jth species participating in the ith reaction,
v = reactant stoichiometric coefficient,
ij
v* = product stoichiometric coefficient,
ij
A = chemical formula of jth species,
j
r •» number of reactions, and
k = rate coefficient of ith reaction
i
The generality of the index notation permits any specification of a
chemical mechanism that the us;er desires. With the notation, the source term
becomes:
r s v
w = (y /y) I 6 k n (yc /y ) iX (j - 1,2,3, ,s), (5)
j j i=l ij i X=l XX
where
y = molecular weight of jth species,
j
y = molecular weight of mixture (air),
3 = v1 , and
ij ij
X = dummy species index.
Because of fluctuations in turbulent flows, Eq. 5 is only an approximation,,
Research is underway to correct this deficiency. The use of Equation 5 instead
of specific terms retains a degree of generality in the computer program that
greatly simplifies alterations of the mechanism.
5-22
-------
For stationary-state species, the partial-differential equation repre-
sented by Eq. 2 is replaced by the source term in Eq. 5 set equal to zero»
This provides the algebraic means for solving for individual species concen-
trations. These algebraic formulas are carried along in the calculation as
constraints on the remaining partial-differential equations,,
Discussions of specific chemical mechanisms are found in Chapter 2, and
examples of working mechanisms that have been used in models can be found in
Friedlander and Seinfeld,35 Eschenroeder and Martinez,36'37 Wayne et_ all,,,140
and Reynolds ^t al.46'48
EXAMPLES OF SPECIFIC METHODS
Statistical, Rollback, and Box Models
Both linear rollback and modified rollback models were used by Earth17
to examine federal motor-vehicle emission goals for standards governing carbon
monoxide, hydrocarbon, and oxides of nitrogen. The linear rollback principle
was suggested and applied to these primary pollutants,,
R = (GF) (PAQ) - (DAQ). (6)
(GF) (PAQ) - B
where R is the fractional reduction required. GF is the growth factor, PAQ
is present air quality, DAQ is desired air quality, and B is background con-
centration. Linear rollback and related models have been heavily emphasized
in the regulatory approaches taken to date.
Linear rollback involved direct application of Eq. 6 to carbon monoxide,
hydrocarbon, and oxides of nitrogen emission. Modified rollback implies
reading from a graph of peak oxidant versus hydrocarbon concentration the
needed hydrocarbon reduction and then using Eq. 6 to relate hydrocarbon
5-23
-------
emission to atmospheric concentrations. The reduction in nitric oxide emission
was chosen to achieve the nitrogen dioxide ambient air quality standard, and
the reduction in hydrocarbon was tailored to achieve the oxidant standard via
the Schuck >e_t al. diagram18 following the modified rollback scheme. This
diagram gives envelope curves of the maximal 1-h oxidant observed versus the
6:00-9:00 a.m. average hydrocarbon concentration in five cities. To avoid
the inaccuracies of the linear rollback scheme as applied directly to oxidant,
Schuck and Papetti1^ have refined the nonlinear or modified rollback scheme.
First, they modified the nonlinear rollback method to include 1968-1971 air
quality data from the Los Angeles basin. The new approach to modified rollback
used an eight-station concentration average in response to the objection that
the hydrocarbon concentrations that caused the oxidant buildup occur somewhere
other than the oxidant station location. That is, winds will generally sweep
the air with the morning hydrocarbon contamination to a location some distance
away. Therefore, averaging eight stations tends to distribute this inaccuracy
and, it is hoped, cancel it out. The upper-limit curve is stated to be in
agreement with regression analyses done at Chevron Research.20 Because of
NO inhibition effects, decreasing hydrocarbon emission faster than nitric
x
oxide emission will reduce ozone even more than is predicted by the curve0
Working against this bonus, however, is the effect of growth that must be
used to correct the percentage hydrocarbon reductions read directly from the
curve. The original rollback formula, of course, included gross growth factors.
The nonlinear rollback approach has serious deficiencies and should not be used
for planning purposes.
Another approach that uses a linear rollback relating primary pollutants
to ambient air quality is described by Hamming et al_0,21 who presents a series
of graphs to show photochemical smog effects in terms of primary pollutant
5-24
-------
concentrations. Contours of constant oxidant, eye Irritation, peak nitrogen
dioxide concentration, and time to nitrogen dioxide peak are plotted on NO —
x
versus-hydrocarbon planes. The data for the graphs were based on interpretation
of the Los Angeles County Air Pollution Control District's experimental-
chamber measurements, as well as those of Korth e^t al_.56 and of Dimitriades.57
It is cautioned that the chamber results cannot be transferred to the atmosphere
on a one-for-one basis, because of wall effects, dilution effects, and other
experimental artifacts. The 6:00-9:00 a.m. concentrations of hydrocarbon and
NO are traced out on graphs. Projections of hydrocarbon and NO concentrations
x x
are made for future years, but the method of carrying out these projects is
not described. Presumably, linear rollback was used to relate emission to
precursor concentrations,, Hence, in theory, this procedure is one step more
nearly complete than the modified rollback used in the oxidant envelope curves.
It reflects the influence of the hydrocarbonrnitric oxide ratio on the four
photochemical smog effects contoured on the plots. This would have been a
direct extension of the modified rollback method, if atmospheric data had been
used; however, the authors cite atmospheric trends that tend to support the
conclusions that they draw from chamber-data plots. The lack of any explicit
means of considering mixture-ratio effects is a major drawback in the use of
modified rollback for the present oxidant control strategies, which impose
more rapid reduction in hydrocarbon than in NO emission0 The technique of
x
Hamming et al. holds out the possibility of eliminating this deficiency.
Chang and Weinstock22 represent the general rollback formula as integrals
over time and space of a population of emitters weighted with influence
functions. The influence functions (or source-receptor interaction functions)
5-25
-------
contain the elements of meteorologic and geometric effects,, The analysis is
restricted to carbon monoxide or any other pollutant that may be nonreactive.
The role of an inhomogeneously distributed growth factor is discussed, and a
method of correcting rollback for this effect with a diffusion model to obtain
the source-receptor interaction functions is proposed. This follows very
closely a methodology introduced earlier by the same authors.23 The rollback
formula was corrected for the nommiformity of growth with the diffusion model
and contrasted with the usual straight tonnage approach^ The result is a
lessening of the requirement for percentage reductions in emission to achieve
a specific air quality goal. For cases with regularity in the forms of the
source-receptor interaction function, there appears to be a good chance to make
general use of this modified rollback approach.
Statistical relationships between air quality measurements and meteorologic
variables are also used for calculating future air quality,, It is important to
note, however, that the statistical relationships are usually related to a
fixed-emission-source distribution pattern or one that changes with time in
some regular manner. Peterson24 analyzed 24-h averaged sulfur dioxide concen-
trations at 40 sites in St. Louis for the winter of 1964-1965. Assuming some
linear relationships, he isolated three basic patterns that accounted for
most of the variance in the observations. The method of empirical orthogonal
functions was used to get regression equations that related pollutant patterns
to meteorologic variables. Because this approach performed better than a
diffusion model, it appeared to be useful for forecasting, although it
contains no explicit relationship to source emission,, It might be possible
to add an adaptive algorithm to a model like this one and use continuously
5-26
-------
updated data to reflect changing emission patterns0 With meteorologic
forecasts, it could then be useful for short-term pollution forecasts, but
it would not be applicable to the long-range study of control strategies0
Forecasting formulas for pollutant concentrations at the downtown Los
Angeles monitoring station were developed by Merz et al.,20 who used a time
series analysis of the monitoring data. The logarithm of oxidant was fit to
the logarithms of hydrocarbon and nitric oxide concentrations up to second-
order terms in the logarithm for each of the dependent variables. The formula
obtained was based on 624 data points for eight 3-month periods and yielded a
correlation coefficient of 0.55 with a standard error of estimate for the
oxidant concentration logarithm (base 10) of 0.5. The diagram was constructed
to provide a graphic working basis for the formula derived. This model relates
pollutants with one another, but does not relate emission to air quality and,
therefore cannot be used for planning emission control strategies.
Trijonis25 determined statistical empirical relationships for primary and
secondary pollutants., The primary-pollutant formula assumed linear rollback
for each concentration along the frequency distribution. This was based on
the usual supposition that the pollutant is inert and that changes in emission
are proportional in space and time. The method was applied to nitrogen
dioxide for central Los Angeles on the basis of NO emission projections»
x
Another relationship was concerned with photochemical secondary pollutants,
specifically oxidant. It followed the assumption that the probability of
exceeding a prescribed concentration depends on the extent of emission of
primary pollutants that react to form oxidant. The specific oxidant-precursor
relationship was implicit in a series of probability curves that depended on
morning NO concentration in central Los Angeles. Curves were developed for
x
5-27
-------
various constant concentrations of morning hydrocarbon (adjusted for methane)„
The relationships were derived from 5 years of monitoring data. This approach
constitutes another method of accounting for the influence of hydrocarbon:
NO ratios on smog predictions.
x
Correlations of ozone concentrations with solar radiation, wind speed,
and temperature data were prepared by Bruntz et al.26 for a monitoring station
at Welfare Island in New York City. Insights into these correlations were gained
by means of a "weathervane plot," which displays ozone versus solar radiation
with circular symbols whose diameters are proportional to temperature,, Lines
on each circle were applied to represent wind speed and direction vectorially.
6 1 9 2
An equation of ozone proportional to (solar radiation) x (wind speed) x
93
(temperature) is suggested by an examination of the graph. While interesting
correlations emerge from this work, no relationship to emissions is provided.
McCollister and Wilson27 proposed two time series models based on fore-
casting some future event by applying a linear formula to a past event. One
model uses today's maximal value of peak carbon monoxide or oxidant to predict
tomorrow's peak value. The other model uses data averaged over each hour today
to predict concentrations for corresponding hours tomorrow. Each model is
evaluated by dividing the mean absolute error by the mean value for days in
1972 using coefficients derived from years before 1972. Evaluation results
are described later,, Again, nothing in this model explicitly shows how emissions
affect air quality.
The box model is one of the simplest forms of solutions of Eq. 2 that
appears in air quality simulation techniques. It assumes that the air bounded
by the ground and the mixing height is uniform. It further considers that the
source intensity of pollution emanating from the ground is constant and that
the wind speed is constant. The consequence of these assumptions is a simple
5-28
-------
formula that states that the ambient concentration of some pollutant is
directly proportional to the source emission rates and inversely proportional
to the wind speed. The constant of proportionality is determined from mixing
height or is derived empirically. One form of this type of model was suggested
by Hanna.28 He assumed the constant of proportionality to be taken as the
width of the region divided by the average mixing depth appropriate to the
scale of the region. This constant is approximately 225 for many cities.
Gifford and Hanna29 conducted tests of the simple model for particulate
matter and sulfur dioxide predictions for annual or seasonal averages versus
diffusion-model predictions. Hanna30 went on to apply the simple dispersion
model to the analysis of chemically reactive pollutants. This required that
each reaction achieve a steady state within the space and time scale of the
airshed of interest. It was concluded that chemical concentrations indicate
the lack of steady state for nearly calm conditions of low mixing depth, but
that the chemistry does not significantly interfere with the use of a simple
model for sunny and windy conditions. For calculational purposes in the
example taken for Los Angeles, the reactive hydrocarbon emission was assumed
equal to the "published propylene emission." Gifford and Hanna state that
"detailed urban diffusion models developed so far have the property that they
generate more pollution variability than is actually observed to occur. This
seems to us to be a strong argument for the use of simpler models." "Vari-
ability" is undefined, and the degree of variability that is generated by the
simpler models is not stated0
After the application of the simple model to chemically reactive pollu-
utants,30 Lamb and Seinfeld31 disagreed sharply with the contention that the
5-29
-------
simple dispersion model proposed by Hanna could be applied to the photochemical
smog problem. They argued that the rate of change depends on more variables
than concentration alone, tha.t spatial variations cannot be neglected, that
steady-state conditions are highly unlikely, that the height of the pollution
layer does indeed extend to the edge of the mixing layer, and that more than
local sources influence ground-level pollution concentrations.
In reply to these criticisms, Hanna32 addressed the objections point by
point. He stated that the use of regionally averaged variables is a necessary
first step and has no special limitations. He asserted that the simple model
formulation does not assume steady-state conditions, but that such conditions
often occur. He restated the belief that local sources are the chief influence
on ground concentrations. The number of chemical parameters available with the
simple model depended on the complexity of the mechanism and was therefore con-
sidered to be arbitrarily complex. It was further contended that the reason
that there are few restrictive assumptions listed for the simple model is
that it has few limitations. Further comparisons with more complex models
were drawn on the grounds that the correlation coefficients for observed versus
predicted data were comparable or better with the simple model.
On the basis of these discussions it does not appear that the simple model
is applicable to chemically reactive pollutants.
Finite-Difference Simulations
Correct modeling of variable diffusivity, time-dependent emission sources,
nonlinear chemical reactions, and removal processes necessitates numerical
integrations of the species-mass-balance equations0 Because of limitations of
dispersion data, emission data, or chemical rate data, this approach to the
5-30
-------
modeling of air pollution may not necessarily ensure higher fidelity, but it
does hold out the possibility of the incorporation of more of these details
as they become known,,
An early analog-computer study of the solution to the species-mass-
balance equations was done by Karplus et al.3 3 This work consisted on a one-
dimensional time-dependent diffusion equation with chemical source terms
representing a multicomponent atmospheric kinetic system. An electronic
analog computer was used, with one integrator at each node between space cells
to handle the combined effects of mass transfer and chemical reaction. Results
were obtained for a simple mechanism, but no tests of validity were made.
Ulbrich3^ also used a series of boxes that were coupled in a model of the
species-mass-balance equations. He integrated a data-management system with
a model system for real-time prediction and controlo A long-term cost-minimizing
strategic command and control system was formulatedo Adaptive features were
built into this control system. The system model conceived for Los Angeles
was a series of seven boxes, each covering 288 square miles. Each box was
assumed to be well mixed and was advectively coupled to its neighbors. A smog
delay time of 1.5 h was set into the system. The only emission that was con-
sidered was that of nitric oxide.
Similarity solutions of the species-mass-balance equations were assumed
by Friedlander and Seinfeld35 for a simple photochemical-smog reaction scheme0
(This scheme assumed a steady-state condition for ozone.) Demonstration runs
were shown for parametric variations in the system of ordinary differential
equations that emerged from the partial-differential equations. Analytic
solutions yielded atmospheric reaction criteria that were usable for correlating
smog variables.
5-31
-------
Pollutants emitted by various sources entered an air parcel moving with
the wind in the model proposed by Eschenroeder and Martinez.35'37 Finite-
difference solutions to the ispecies-mass-balance equations described the pol-
lutant chemical kinetics and the upward spread through a series of vertical
cells. The initial chemical mechanism consisted of seven species participating
in 13 reactions based on smog-chamber observations. Atmospheric dispersion
data from the literature were introduced to provide vertical-diffusion co-
efficients. Initial validity tests were conducted for a static air mass over
central Los Angeles on October 23, 1968, and during an episode late in 1968
while a special mobile laboratory was set up by Scott Research Laboratories.58
Curves were plotted to illustrate sensitivity to rate and emission values, and
the feasibility of this prediction technique was demonstrated. Some problems
of the future were ultimately identified by this work,38 and the methodology
developed has been applied to several environmental impact studies (see, for
example, Wayne et al.4°).
•5-32
-------
Concern is sometimes expressed regarding the role of surface reactions
on aerosol as an interference in the buildup of photochemically produced
ozone. An upper-limit analysis was carried out by Eschenroeder, Martinez,
3
and Nordsieck38 to assess the effect of 200-yg/m loading of 0.5-ym-diameter
particleso Ratios of surface-reaction rates to gas-phase rates for 100%
surface reaction efficiencies were computed. Early in the day, nitric oxide
and nitrogen dioxide surface reaction rates can be a few hundred times the gas-phase
reaction rate, under the assumption that each molecular collision with the
particle surface constitutes a successful reactionevent. In the middle of
the day, nitrogen dioxide and ozone surface-reaction rates in this limit are
again
10-50 times the corresponding gas-phase reaction rates,/assuming 100% efficiency.
-6 -4
The efficiencies of surface reactions are more nearly in the 10 -10 region.
Therefore, the high upper-limit values must be reduced considerably to reflect
realistic conditions. It is therefore likely that the ozone buildup is affected
little by aerosol surface reactions; however, the formation of nitrate in the
aerosol, which is observed to take place in the morning, may be marginally
important, compared with the gas-phase reaction.
Another Lagrangian photochemical model was developed by Wayne et al.40
This photochemical model uses a moving "cell" or "air mass" that follows an
air trajectory either from a specified source or to a specified receptor.
The emission input from the ground-based sources is expressed as a time-
dependent influx of primary pollutants. Diffusive spread need not be computed
in this model, because the air in the cell is assumed to be homogeneously
mixed at all times. This brings about a tremendous advantage in reduced
computing time. The chemical mechanism represents reactive hydrocarbons
as propylene and lumps others into a generic dummy hydrocarbon,. Reactivity
is handled by specifying the mix of propylene and generic hydrocarbon.
Performance evaluation results are reported for September 30, 1969, on the
5-32a
-------
basis of arbitrarily assigned initial values for hydrocarbon concentrations
and neglecting stationary emission of NO .
x
Mahoney and Egan1*1»^developed a two-dimensional time-dependent diffusive
and advective model that neglects vertical velocity0 Chemical reaction is
also neglected. The source term is an effective volume source in the bottom
grid mesh cell of the calculation. The authors discussed the pseudodiffusion
errors that arise with the large grid spacing that is appropriate to urban-
scale calculations. They pointed out that this error is orders of magnitude
larger than natural diffusion. A puff example from a volume source was
presented in the paper. The scheme proposed avoids pseudodiffusion by using
moments of the concentration distribution in the governing equations» The
concentration profile was reconstructed by using the computed first and second
moments. Egan and Mahoney43 applied the model to estimate ground concentrations
under different meteorologic conditions. Velocity profiles and vertical-
diffusivity profiles were introduced on the basis of various stability con-
ditions. Two-dimensional time-dependent solutions with variable advective
velocity were obtained. Test cases were presented, including the effects of
wind-velocity vector changes with height and the effects of this velocity
field distortion on the dispersion of air pollutants0 Elevated inversions
and time-dependent mixing heights were also investigated. The height
variation of the velocity field was shown to be important under stable con-
ditions „ Although this model does not treat air chemistry, it can resolve
subgrid scale elements because of the moment method used.
A solution to the species-balance equation was generalized by
Lamb and Neiburger44 to allow for the space and time distribution of pollutant
5-33
-------
emission, diffusion coefficients, and wind field. Pollutant removal at the
ground and leakage through the upper boundary were also allowed for in this
model. A transformation to s. Lagrangian coordinate system was made, and the
model was adapted to various source emission configurations. It was tested
for carbon monoxide concentrations in the Los Angeles basin for September
23, 1966. It considered vehicular sources based on freeway and surface-
street traffic counts. The authors plotted observed versus predicted concen-
trations for four stations and found that the morning "peak" in concentration
was exaggerated by the model,, but that the midday "valley" was underestimated.
Seinfield et al.^5 and Reynolds et al.1*6 discussed theoretical aspects
of urban air pollution modeling in terms of the species-mass-balance equations
cast into a problem requiring specification of initial and boundary values of
field quantities. Restrictions of K theory in turbulence approximations were
reviewed. The vertical coordinate system was mapped between the ground and
the inversion base by a line^^.r stretching transformation0 The methods were
detailed for interpolating discrete data on winds and vertical stability to
obtain field values needed for the calculations. The kinetic mechanism that
uses lumped parameters for hydrocarbon change was outlined0 Eulerian
difference equations were integrated numerically, and a method of fractional
steps was described. Explicit differencing was used for horizonal coordinates,
and implicit differencing for the vertical terms of the equations. In a
later paper in the series, Roth et al.^7 gave a detailed description of an
emission model and an inventory for the Los Angeles basin. Automotive emission
was discussed with a breakdown into surface-street and freeway categories0
Average-trip-speed correction factors were allowed for, and cold-start corrections
5-34
-------
for automotive emission factors were introduced. Spatial distributions of
daily vehicle miles traveled were laid out on a 2-mile grid for the Los Angeles
basin. An overall temporal distribution was adopted, and deviations from this
at various locations were shown to be small. Ground and flight operations of
aircraft were discussed, but only ground emission was included in the actual
application of the model. Reactivity of hydrocarbon from aircraft was con-
sidered to be the same as that from automobiles. Stationary sources were
categorized as power plants, refineries, and distributed sources0 This paper
serves as a prototype for future source—emission inventory processes that are
intended to supply modelers with input data.
Reynolds et al. ° described the process of evaluation of the airshed
photochemical model. Kinetic mechanisms were checked against smog-chamber
tests, yielding branching factors and rate parameters for the simplified
lumped-parameter scheme. A microscale model was established to correct for
local effects around monitoring station sites. An evaluation procedure
involved preparing data, preparing initial and boundary conditions, checking
for agreement with carbon monoxide, and testing computed versus observed
values for reactive pollutants. The results agreed rather well with observa-
tions, but no statistical performance measure was used at this stage of the
work.
Sklarew et al.*49 used a particle-in-cell K. theory (PICK) approach to
calculate atmospheric diffusion and reaction. The method was applied to
tcarbon monoxide pollution in Los Angeles and predicted daily averaged concen-
trations reportedly within 20% of measured averages for 12 monitoring stations.
Photochemical smog simulation was demonstrated, but the results were hampered
5-35
-------
by the use of the early prototype of lumped-parameter chemical mechanisms,36
which did not adequately represent the influence of differing hydrocarbon:
NO ratios. Also, the carbon monoxide results required inordinately high
x
carbon monoxide emission by the Pacific Ocean.
Knox50 summarized recent developments at the Lawrence Livermore Laboratory.
JL
The computation using the Lagrangian large-cloud dispersion model was shown
to agree moderately well with the measurement of gross beta-particle activity
in the cloud of a nuclear test burst. A three—dimensional atmospheric—diffusion
particle-in-cell code was described in which marker particles are traced
through an Eulerian grid that was distorted to fit topography. A grid followed
the cloud center to minimize the number of cells and to minimize pseudodiffusion.
Mass-consistent wind-field modeling was also discussed. This form of modeling
considers the confining effects of an elevated inversion and the variable
topography on the surface wind measurements. A multibox regional air pollution
model was run for the San Francisco Bay area for carbon monoxide, arid the
calculated surface concentrations were approximately 20-50% above the observed
concentrations. The frequency distribution predicted for the concentrations,
however, paralleled closely the observed frequency distributions.
Gradient diffusion was assumed in the species-mass-conservation model of
Shir and Shieh. •"• Integration was carried out in the space between the ground
and the mixing height with zero fluxes assumed at each boundary. A first-order
decay of sulfur dioxide was the only chemical reaction, and it was suggested
that this reaction is important only under low wind speed. Finite-difference
numerical solutions for sulfur dioxide in the St. Louis area were obtained with
a second-order central finite-difference scheme for horizontal terms and the
*A model for calculating the time dependent airborne concentration,
surface air concentration, cind dry and/or wet deposition of pollutants
from large clouds.
5-36
-------
Crank-Nicolson technique for the vertical-diffusion terms. The three-dimensional
grid had 16,800 points on a 30 x 40 x 14 mesh.
PERFORMANCE OF PREDICTION TECHNIQUES
Before we examine the performance of various models for predicting ambient
air quality, it is important to review the criteria for selecting a particular
model. In this selection process, it is essential to consider the following
factors:
• The precision required for forecasting air quality impact.
• The conditions of high concentrations of air pollutants in the selected
region.
• The severity of vehicle-generated pollution relative to that from
other sources.
• The availability of meteorologic and air quality data for the selected
region.
• The availability of geographic and temporal distributions of emission
sources.
• The effect of stack heights associated with large point sources on
regional air quality.
• The relative influences of spatial emission distribution, time depen-
dence, and chemical reactions on regional pollution patterns.
• The time and money available for modeling.
Many standard statistical tests are available to evaluate the performance
of models against observations. It should be pointed out, however, that
graphic checks on the performance of models are also necessary. Anscombe59
has brought this out strikingly with specific numerical examples. He showed
5-37
-------
that pure regression analysis can be extremely misleading and demonstrated
what can happen in purely numerical analysis of data. In his paper, four
artificial data sets of x and y values were given; they are reproduced in
Table 5-1. Suppose that x is an observed pollutant concentration, and y is a
concentration computed from a model. If standard statistical analytic tech-
niques are performed on these four sets, it can be shown that they posess
identical values of the following calculations:
• Number of observations: 11
• Mean of the x's: 9.0
• Mean of the y's: 7.5
• Equation of the regression line: y = 3 + 0.5x
• Sum of squares of x - x: 110.0
• Estimated standard error of slope of regression line: 0.118
• Multiple correlation coefficient: 0.667
5-38
-------
Table 5-1
Sample Data Sets Given By F. J. Anscombe59
(x = observed pollutant concentration; y = model-computed concentration)
Set 1
Set 2
Set 3
Set 4
X
10.0
8.0
13.0
9 = 0
11.0
14.0
6.0
4.0
12.0
7.0
5.0
y
8.04
6.95
7.58
8.81
8.33
9.96
7.24
4.26
10.84
4.82
5.68
X
10.0
8.0
13.0
9oO
11.0
14.0
6.0
4.0
12.0
7.0
5.0
y
9.14
8.14
8.74
8.77
9.26
8.10
6.13
3.10
9.13
7.26
4.74
X
10.0
8.0
13.0
9.0
11.0
14.0
6.0
4.0
1200
7.0
5.0
y
7.46
6.77
12.74
7.11
7.81
8.84
6.08
5.39
8.15
6.42
5.73
X
8.0
8.0
800
8.0
8.0
8.0
8.0
19.0
8.0
8.0
8.0
y
6.58
5.76
7,71
8.84
8.47
7,04
5.25
12.50
5o56
7.91
6.89
5-39
-------
Figures 5-2 through 5-5 show rhe great contrasts in the actual data that are
revealed by a graphic presentation, but are totally concealed in the above results of
standard analytic procedures. Figure 5-2 shows a typical scatter plot, but
the other three reveal regularities that may be missed by linear regression
analysis—a smooth curve, a different straight line with an outlier, and x
values all the same except for an outlier.
The meaning of this example for model performance evaluation is signif-
icant: if measures are confined to the. usual statistical tests, a great deal
of information can be lost. The loss of information tends to perpetuate
biases in the calculations, which might otherwise be eliminated. The message
is clear. One should beware of correlation analysis alone. Questions of data
regularity that can be resolved by graphic presentations must be pursued, as
well as those which pertain to regression analysis alone.
Rollback, Statistical, and Box Model Performance
Eschenroeder compared linear rollback results with photochemical-
diffusion model results in assessing the effects of various stages of the
original transportation control plan formulated by a federal regulatory agency„
Technologic controls without a reduction in the number of vehicle miles
traveled (VMT) were considered as a first stage, and successive percentages of
VMT reduction were considerec. as further stagesc. For the first stage of con-
trol, the photochemical-diffusion models suggested more rapid reduction in peak
photochemical oxidant concentration than linear rollback, because the latter
fails to account for nitric oxide inhibition effects. Reactive hydrocarbon
(RHC) is the only precursor considered by the rollback method. In reality,
the ratio of RHC to NO also has a decided influence on the oxidant production.
x
5-40
-------
to
10
y
I I I 1 I
I I I I I I ! I I
"lO 15
X
ZO
10
15 20
Figure 5-2. Figure 5-3„
Graph of data in Anscombe's sample set I59 Graph of data in Anscombe's sample set 259
(see Table 5-1) (see Table 5-1)
10
10
i t I i i i i I i i i t I i i i i I Oi ' i i i I i i t
10
15 20
10
15
•x,
Figure 5-4.
Graph of data in Anscombe's sample set 359
(see Table 5-1)
2O
Figure 5-5.
Graph of data in Anscombe's sample set 459
(see Table 5-1)
5-41
-------
Excess NO slows the production of oxidant. Successive stages of emission
x
control brought about by reductions in VMT show less than proportional reduc-
tions in photochemical oxidant according to the photochemical diffusion model.
This occurs because RHC and NO are being reduced in the same ratio, which
x
has a weaker influence on oxidant than in the case of more rapid reductions in
•RHC. This comparison illustrates the need to consider nonlinearities in the
case of reactive pollutants when using rollback models.
In another study, Martinez and Nordsieck61 found that comparing linear
rollback with a photochemical model gave very nearly the same results for peak
oxidant for the year in the distant future when extensive controls would be
imposed. The agreement between the two methods, however, was poor in a case
in which NO controls were frozen at 1974 values, instead of being progressively
x
applied as rapidly as RHC controls. This case is designated 1980d in Table 5-2,
which summarizes the results for forecast ozone concentrations on the basis of
both linear rollback and the photochemical-diffusion model. The frozen-controls
case (1980d) shows the large disparity between the methods because of the
failure of rollback to consider the effects of mixture ratios of precursor
pollutants on oxidant formation.
Reynolds and Seinfeld62 compared the statistical model of Trijonis25 with
their dynamic model145""1*8 and with linear rollback. Considering 1977 emission
with and without one of the EPA abatement plans in effect, they obtained a
reduction from 80 to 20 in the number of days on which the oxidant standard is
exceeded and from less than 15 days to less than 10 days for exceeding the
nitrogen dioxide standard. (The oxidant standard is defined on the basis of
hourly average, and the nitrogen dioxide standard is defined as an annual
5-42
-------
Table 5-2.
Baseline Values of Maximal Hourly Average
Ozone Concentration in Vicinity of Riverside^
Maximal Hourly Average Ozone Concentration, ppm
Year Diffusion Kinetics Model Linear Rollback
1980 0.16 0.16
1990 0.15 0.13
2000 0.15 0.13
1980d 0.12 0.17
a. Derived from Martinez and Nordsieck.61 See text for explanation
5-43
-------
average0) These findings, which were obtained with the Trijonis model, were
limited to one monitoring station because of the method by which that model
was calibrated. The dynamic model was run with two ways of specifying future
initial conditions and grid boundary conditions for the calculations0 It was
not stated whether linear or nonlinear rollback was usedo Comparisons were
then made between 1969 (the reference year) and 1977 both with and without the
EPA abatement plan in effect in 1977. The dynamic-model calculation showed a
76% reduction in carbon monoxide from 1969 to 1977 with the EPA abatement plan
in effect and only a 65% reduction without it. Rollback agrees with the 65%
reduction, because the conditions for its validity with an inert pollutant are
fulfilled. Nitrogen dioxide peak concentrations were reduced by 5-8% without
the plan and by 60-76% with the plan, according to the dynamic model. Rollback
predicted only a 34% reduction with the plan. For ozone, the dynamic-model
results suggested that the abatement plan is more effective than the rollback
model would predict—namely, that up to a 94% reduction in maximal ozone was
predicted by the dynamic model, and a 74% reduction was predicted by rollback.
The 1969-1977 ozone reductions without the abatement plan in effect are 39%
according to rollback and 65-75% according to the dynamic model. As was the
case with the Hamming et al.21- and Eschenroeder60 calculations, this illustrates
for ozone that it is necessary to consider both RHC and NO in the rollback
x
calculation, not only RHC. If the RHC emission is reduced faster than the NO
x
emission, the ozone concentrations at a given station may decrease in greater
proportion than the RHC concentrations.
Peterson2^ used the skill score to evaluate the performance of his empirical
statistical model based on orthogonal functions. The skill score equals 1.0
5-44
-------
when all calculated and observed concentrations agree, but 0 when the number
of correctly predicted results equals that expected by chance occurrences.
The statistical technique had a skill score of 0.304. An 89-day, 40-station
set of the data was used to check a Gaussian diffusion model, and this tech-
*
nique gave the diffusion model a skill score of only 0015. It will be recalled
that the statistical empirical model was used for 24-h averaged sulfur dioxide
concentrations at 40 sites in St. Louis for the winter of 1964-1965.
The time-series analysis results of Merz et al.20 were expressed in first-
order empirical formulas for the most part. Forecasting expressions were
developed for total oxidant, carbon monoxide, nitric oxide, and hydrocarbon,,
Fitting correlation coefficients varied from 0.547 to 006590 As might be
expected, the best results were obtained for the primary pollutants carbon
monoxide and nitric oxide, and the lowest correlation was for oxidant0 This
model relates one pollutant to another, but does not relate emission to air
quality,, For primary pollutants, the model expresses the concentrations as a
function of time.
Bruntz et al.26 applied multiple regression analysis and found that the
method of least squares yielded a set of coefficients that produced an 0084
correlation of ozone concentration with the data. Adding mixing height to the
correlation yielded no statistically significant improvement in agreement with
the assertions of Hanna.30
The first time-series model of McCollister and Wilson27 yields results
around 0.4 for the mean absolute error divided by the mean value for the days
in 1972, on the basis of parameters derived before 1972. By comparison, a
persistence model yields about 0.5 for the same parameter, and Los Angeles APCD
5-45
-------
forecasts lie between the time-series results and the persistence-model
results. The hourly oxidant time-series model yields errors from 0.3 to 0<>45,
whereas persistence yields errors of 0.4-0.5. For carbon monoxide, the hourly-
model results lie between 0.15 and 0.5—consistently below the persistence
*
resultSo
Gifford and Hanna29 tested their simple model for particulate matter and
sulfur dioxide predictions for annual or seasonal averages against diffusion-
model predictions. Their conclusions are summarized in Table 5-3. The cor-
relation coefficient of observed concentrations versus calculated concentrations
is generally higher for the simple model than for the detailed model. Hanna30
calculated reactions over a 6-h period on September 30, 1969, with his chemically
reactive adaptation of the simple dispersion model. He obtained correlation
coefficients of observed and calculated concentrations as follows: nitric
oxide, 0.97; nitrogen dioxide, 0.05; and RHC, 0.55. He found a correlation
coefficient of 0.48 of observed ozone concentration with an ozone predictor
derived from a simple model, but he pointed out that the local inverse wind
speed had a correlation of 0.66 with ozone concentration. He derived a
"critical wind speed" formula to define a speed below which ozone prediction
will be a problem with the simple model. Further performance comparison of
the simple box model with more detailed models will be discussed later.
Species-Mass-Balance Model Performance and Comparative Evaluations
Sklarew et al^9 evaluated their particle-in-cell K-theory approach for
atmospheric diffusion of carbon monoxide and for photochemical smog. All-day
averages of carbon monoxide concentration were predicted to be within 20% of the
5-46
-------
Table 5-3.
Test of Simple Model for Particulate
Matter and Sulfur Dioxide Predictions?
Correlation Coefficient, observed
vs. calculated concentrations
City
Memphis
Nashville
Ankara
Bremen
"Test City"
No. Sampling
Sites
9
16
10
4
8
Source Area
Size, km2
25
25
9
16
25
Simple Model
0.68
0.91
0.63
0.65
0.98
Detailed Model
0.73
Oo80
0.59
0.05
0.96
a
Derived from Gifford and Hanna»29
5-47
-------
measured averages at 12 monitoring stations, and the correlation coefficient
of measured with observed concentrations was 0.73.
In his survey,15 Sklarew covered evaluation studies and summarized the
results of correlation coefficients and root-mean-square (rms) errors from a
linear regression between observed and calculated values for three photochemical
models,,63'64'65 The statistical comparisons are shown in Tables 5-4 and 5-5
for carbon monoxide and ozone, respectively. It is notable that the correlation
coefficients are considerably higher than those reported for many Gaussian
models. Similarly, the regression lines have slopes closer to unity than those
from the Gaussian models. Extreme caution must be exercised in comparing
performance measures directly, because of the intrinsic differences between
trajectory and grid models and between evaluation test designs. Nappo66
evaluated eight mathematical models and 24-h persistence with six different
measures of performance evaluation for carbon monoxide predictions. The
usual measure of correlating the time-averaged concentrations was applied to
the computed versus observed values. Table 5-6 summarizes these results, with
computer time and computer costs estimated for an IBM 360/65 system. This is
one of the few attempts to supply information that could enter a cost-
effectiveness analysis. The data base used by Nappo consisted exclusively of
computed and observed concentrations that are reported in the literature (see
table for references). The averaging time for each data set varied according
to the time interval for which each model was run. For example, if a model
were designed to predict hourly averages and was run for an 8-h interval, the
averaging time for the data set was chosen to be 8 h.
5-48
-------
Table 5-4
Statistical Comparison of Model Calculations
and Observations of Carbon Monoxide Concentration,13
Model
SAI
PES
GRC
b
No.
Ref. Data Points
99 514
100 71
101 149
Correlation
Coefficient
0.79
0.82
0.80
RMS
Error, ppm
3.4
3.0
3.7
Regression Line
(y = ax + b)
1.09 0.68
0.91 -0.70
1.01 -0.37
a Derived from Sklarew015
b SAI = Systems Applications, Inc., PES = Pacific Environmental Services,
Inc., GRC = General Research Corporation.
a y = observed„ x = calculated.
5-49
-------
Table 5-5
Statistical Comparison of Model Calculations
and Observations of Ozone Concentration12
No.
Correlation
Regression
RMS (y = ax + b)
Model0
SAI
PES
GRC
Ref.
99
100
101
Data Points
574
63
151
Coefficient
0.69
0.49
0.92
Error, ppm
Oo063
0.069
0.021
a
0.76
0.46
0.84
b
3.9
-7.6
2.3
a
b
15
Derived trom bklarew.
SAI = Systems Applications., Inc., PES = Pacific Environmental Services,
Inc., GRC = General Research Corporation.
y = observed, x = calculated.
5-50
-------
Table 5-6.
Model Evaluation Based on Temporal Characteristics'2
Model
MacCracken et al.67 multibox
24-h persistence
Roth et al.68 primitive equation
Hanna69 ATDL simple model
Sklarew et al.70 particle-in-cell
Pandolfo and Jacobs71 primitive
equation
Reynolds et al.63 primitive
equation
Eschenroeder et al.65 trajectory
Lamb and Neiburger'*'* trajectory
Average
Temporal
Correlation
Coefficient
0.37
0.47
0.52
0.60
0.65
0.66
0.73
0.73
0.90
Computer Time Computer Cost
for 24-h for 24-h
Prediction, Prediction, $
min
106
0
60
0
49
20
30
15
35
350
0
200
0
160
70
100
50
115
a Derived from Nappo.66
5-51
-------
In addition to the temporal correlation coefficient, the spatial cor-
relation coefficient was calculated approximately for fixed values of time.
Except for one of the mathematical models,69 all techniques showed a better
temporal correlation than spatial correlation. The two correlation coefficients
are cross plotted in Figure 5-6. Nappo stressed that correlation coefficients
express fidelity in predicting; trends, rather than accuracy in absolute con-
centration predictions. Another measure is used for assessing accuracy in
predicting concentrations: the ratio of predicted to observed concentration.
Nappo averages this ratio over space and over time and extracts the standard
deviation of the data sample for each. The standard deviation expresses
consistency of accuracy for each model. For example, a model might have a
predicted: observed ratio near unity, but with a wide variability of the
ratio about its mean value. All the models tested have ratios that average
within ±40% of unity (including 24-h persistence), with two exceptions—the
Hanna simple model69 and the Lamb and Neiburger trajectory model.^ It should
be noted that this measure of model performance produced a rank ordering very
different from that of the usual temporal correlation test. Figure 5-7 shows
the mean ratios of predicted to observed concentration with uncertainty bars
characterized by the standard deviation of the ratios about their mean.
Figure 5-8 shows the space-averaged temporal standard deviation plotted against
the time-averaged spatial standard deviation of these ratios. The models have
both standard deviations less than 0.5, except those of Roth et al.,68 Pandolfo
and Jacobs,71 and Hanna69 and 24-h persistence.
5-52
-------
10
0.8
0.6
0.4
0.2
0
• R.OTH et ol. (68)
A REYNOLDS et of. (63)
• HANNA (69)
T PANDOLFO AND JACOBS ( 71)
0 SKLAREW et of. ( 70)
A LAMB AND NEIBURGER (44)
* MocCRACKEN et a/. ( 67)
D ESCHENROEDER et of. ( 65 )
o 24 hr PERSISTENCE
0.2
0.4
0.6
0.8
msr
1.0
Figure 5-6. R(t) versus R(s) —space-averaged temporal correlation coefficient
versus time-averaged spatial correlation coefficient. Points are
averages for each model tested. (Reprinted with permission from
Nappo;66 reference numbers changed to conform with this volume.)
5-53
-------
4.0
3.5
3.0
CO
2.0
1.5
1.0
0.5
• ROTH el of. ( 68 )
A REYNOLDS el a!. (63)
o HANNA ( 69 )
v PANDOLFO AND JACOBS ( 71)
ft SKLAREW el ol. ( 70 )
A LAMB AND NEIBURGER (44)
* MocCRACKEN el ol. ( 67)
o ESCHENROEDER el ol. ( 6s )
o 24 hr PERSISTENCE
0.5
1.0
1.5
2.0
2.5
3.0
3.5
Figure 5-7. r(t) ± a(t) versus r(s) ± a(s) (see text for definitions),
Points are averages for each model tested. (Reprinted with
permission from Nappo;66 reference numbers changed to conform
with this volume.)
5-54
-------
e ROTH et ol. ( 68 )
A REYNOLDS et ol. (63)
D HANNA ( 69 )
V PANDOLfO AND JACOBS ( 71)
0 SKLAREW et al. ( 70)
A LAMB AND NEIBURGER (44)
* MocCRACKEN et ol. ( 67)
o ESCHENROEDF.R et ol. (65)
o 24 hr PERSISTENCE
s t
Figure 5-8. o(t) versus a(s) . Points are averages for each model tested.
(Reprinted with permission from Nappo;66 reference numbers changed
to conform with this volume.)
5-55
-------
Whitney/- evaluated the evaluation studies of predicted vaJues
derived by pnotochemical models with actual measurements. To assess the
63-65
validation tests sponsored by the EPA for three different approaches,
Whitney proposed several statistical evaluation measures of model
performance. In addition to the correlation coefficient and the standard
deviation, the chi-squared test is imposed. Table 5-7 shows the comparison
for four pollutants modeled. The symbol SAIC denotes station values selected
from the Systems Applications, Inc., validation runs to be in correspondence
as nearly as possible with the station values computed in the Pacific Envi-
64
ronmental Services, Inc., validation runs (PESC). This was done in an
effort to compare models on a common-denominator basis. Likewise, trajectories
are traced through the SAI hourly-grid results, to form the SAIT data set to
compare with the General Research Corporation validation runs (GRCT). Clear
trends are difficult to identify, but, as the author points out, the SA1 model
probably excelled in carbon monoxide prediction because it has more detail in
advective diffusive coupling, and the GRC model did best with ozone prediction
probably because of careful calibration. The PES model had the minimal
standard deviation with nitric oxide prediction because the data sample centered
on midday concentrations away from the morning traffic peak, which is difficult
to predict. The chi-squared test suggested that most of the time none of the
models adequately represented the data on the basis of a 10% chance that the
results are randomly distributed. Exceptions to this finding are noted with
GRCT and SAIT, both passing the chi-squared test to a 90% confidence level for
carbon monoxide results.
5-56
-------
Table 5-7
Statistical Analyses of Photochemical Air Quality
Model Performance for the Los Angeles Basin a
Correlation Standard Degrees of Measured Standard
Model^ Coefficient^ Deviation" Freedom Chi-Squarede Chi-Squared/
Carbon monoxide
PESC 0.68 4.27 143 233.75 171.62
SAIC 0.84 3.52 290 287.35 330.44
GRCT 0.82 3.39 172 75.12 203.32
SAIT 0079 3.18 175 84.42 206.58
Nitric oxide
PESC 0.77
SAIC 0.87
GRCT 0.87
SAIT 0.86
Nitrogen dioxide
PESC 0.68 8.36 49 157.15 66.05
SAIC 0.65 6.82 276 310.56 315.47
GRCT 0.43 10.05 172 316.91 203032
SAIT 0.52 12.62 175 159.81 206.58
Ozone
PESC 0.50
SAIC 0.60
GRCT 0.91
SAIT 0.69
3.18
9.13
8.70
8.13
49
277
171
173
227.67
1,081.01
1,053.03
1,042.44
66.05
316.54
202023
204.41
8.26
8.56
3.82
8.75
53
255
127
131
380.85
3,480.79
237.74
387.30
70.71
292.97
154.02
158.43
a Data from Whitney.,72
b See text for abbreviations,,
o Correlation of predicted and actual measurements.
d Difference of predicted and actual measurements.
e Normalized by a where a is assumed to be 0.7 + 0.1 x (observed value).
/ A 90% confidence level was assigned for the purposes of the chi-squared
test.
5-57
-------
Other performance measures were carefully chosen to reflect accuracy of
prediction. They consisted of scatter plots of predicted versus observed
concentrations and four other types of plots involving analyses of residuals.
The residual study took the (predicted - observed) values and displayed them
in histograms in plots of time against predicted concentrations and against
observed (or interpolated) concentrations. The results of the residual
analysis are too extensive to review here, but the user who is interested in
modeling a particular pollutant or who is considering one of the specific
models tested should examine the computer printouts to answer specific per-
formance questions. Whitney has devised a useful set of model performance
measures for the task of evaluating the earlier evaluations. The results of
similar tests on log statistics would have been interesting. It was not
stated whether the residuals; were normally distributed.
5-58
-------
The sensitivity of diffusion-model output to variations in input has
been assessed by workers at Systems Application Inc. and at the California
Department of Transportation. In each case, reports are in preparation and
are therefore not yet available. It is important to distinguish between
sensitivity and model performance. True physical or chemical sensitivity
that is reflected by the simulation-model equations is a valid reflection of
reality. But spurious error propagation through improper numerical integration
techniques may be regarded as an artificial sensitivity. Such a distinction
must be drawn carefully, lest great sensitivity come to be considered
synonymous with unacceptable performance.
-------
SUMMARY
The literature contains reviews of air quality modeling that stress
special purposes. Some concentrate on meteorologic aspects, and others combine
this with air chemistry. Proceedings of several conferences are another
information resource,, Recent surveys have been addressed specifically to
photochemical modeling problems. It may be concluded that, although they are
relatively complex, the photochemical-diffusion models perform as well as, if
not better than, available inert-species models.
A variety of goals and objectives may be met with air quality modeling.
They are summarized as:
• Scientific understanding of atmospheric phenomenology,,
• Rational application of the regulatory process.
• Land-use planning within environmental constraints,
• Real-time control of episodes.
The fundamental elements of deterministic models involve a combination of
chemical and meteorologic input, preprocessing with data transmission, logic
that describes atmospheric processes, and concentration-field output tables or
displayso In addition to deterministic models, there are statistical schemes
that relate precursors (or emission) to photochemical oxidant concentrations.
Models may be classified according to time and space scales, depending on the
purposes for which they are designed.
Specific model applications to the oxidant problem include both the
simple rollback (with modifications) and the photochemical-diffusion tech-
niques. Very little modeling of intermediate complexity seems to have been
attempted for the oxidant system.
5-61
-------
Model performance is now receiving critical attention because of the need
for cost-effective control measures. Standard statistical performance de-
scriptors can sometimes mislead a prospective user; therefore, more, specialized
tests are being devised. Various model types are being compared for a specified
set of initial and boundary conditions. It is apparent from these studies
that added fidelity is purchased at the expense of added complexity of a logical
structure that must represent the controlling phenomenology.
SUGGESTED RESEARCH DIRECTIONS
Internal improvements in deterministic methods will be based on accounting
for more physicochemical effects in the logical structure. One challenge to
the researcher is to do this without making something that is already complex
still more difficult to understand, and another challenge is to avoid needless
elaboration of detail. Both pitfalls will be avoided, first, by asking how
accurate a modeling job is demanded and, second, by carrying out order-of-
magnitude analytic appraisals of the otaitted phenonfenology.
Perhaps the most important thing that research will contribute is a set
of criteria delineating the fidelity of existing models, rather than a single
supermodel that will consider all effects. Much remains to be done in statis-
tical modeling. The scientific community is on the threshold of potentially
great strides with these methods, because of the veritable explosion of data
from measurement programs. It is absolutely essential for all agencies
interested in environmental management to begin mounting analysis programs that
are carefully designed to capitalize on the data base. Traditionally, support
has been more readily obtained for making additional measurements in hope
5-62
-------
that useful information would emerge directly or that someone would sponta-
neously dig out the useful information. Seldom has either been the case0
Specific research subjects have emerged with respect to improved des-
criptions of specific phenomena. Some time ago,38 it was speculated that gas-
solid interactions and turbulence effects on reaction kinetics would be
important areas of advance in the modeling art. Gas-solid interactions include
both chemical formation of aerosols and reactions on surfaces of preexisting
suspended particulate matter. Because of differing effects of a material in
the gas phase and in some condensed phase, it will be important to characterize
transformation processes. The ACHEX (Aerosol Characterization Experiment)
program89 recently carried out under the direction of Hidy will provide an
extensive data base with which to test new ways of treating the gas-solid
interaction problem.
The turbulent mixing of emitted reactant gas (such as nitric oxide) with
atmospherically formed reactant gas (such as ozone) results in macroscopic
heterogeneities, which under some circumstances can significantly change the
reaction rate from the value that the mean concentrations used in a rate
equation would predict. Airborne measurement from some 40 operational days
from the LARPP (Los Angeles ^Reactive JPollutant Frogram) study gives 6-s-interval
gas-phase data for six gas-phase species simultaneously. This program (under
the field management of W. Perkins and under the direction of Coordinating
Research Council's CAPA-12 committee, chaired by J. Black) has produced archives
of these data that can serve as a test bed for theories of turbulent inter-
actions with kinetics.
In a broader sense, the data obtained from the Regional Air Pollution
Study (RAPS) and the California Three-Dimensional Pollutant Gradient Study
5-63
-------
Program should also serve as bases of further model development. It is
incumbent on the agencies responsible for air quality control to identify
resources specifically aimed at using these data for improving techniques for
designing pollution abatement strategies.
SUGGESTED APPLICATIONS TO POLLUTION ABATEMENT
Without doubt, the top-priority application of air quality models is the
determination of emission controls needed to achieve ambient air quality
standards. With the reexamination of transportation control strategies and
with the pressures of fuel substitutions, refinements well beyond the
traditional proportional models are imperative. Where validated diffusion
models are available, they should be used to recalculate the emission require-
ments that came from initial hasty efforts to implement the Clean Air Act
Amendments of 1970. This is the greatest national service that could be
performed by the air quality modelers at present. Before this can be achieved,
however, the institutional apparatus must provide the impetus and resources
called for in a recent National Academy of Sciences report to the U.S0 Senate0^°
Much of the research work will add content to the model structures, but
future applications demand simplifications that are oriented toward the non-
specialist user. One of the largest obstacles to the effective use of air
quality prediction schemes is the resolution of this apparent conflict„ At
least two steps can be taken by those who produce models to encourage appli-
cations and to aid the user: *
• Compile a catalog of air quality models that describes their capa-
bilities in terms of a common set of performance standards.
• Clarify data communication in the input-output interfaces between
user and model.
5-64
-------
To accomplish the first step, model standards will be evolved on the
basis of legislative mandates and regulatory needs. Each of the various
types of model has undergone performance evaluation through the application
of a set of tests peculiar to its own structure or output. For example,
Gaussian models that predict long-term averages are often evaluated by
computing only the correlation coefficient between measured and computed con-
centrations. Early evaluations of species-mass-balance models stressed hour-
by-hour comparison of the predicted and observed concentrations. Recently,
a broader range of descriptors has evolved, as evidenced by the work of
Nappo66 and Whitney.72
The performance indexes must be designed with the model applications in
mind. Will the model be used to predict local effects around a highway or a
smelter where short-term high doses are as important as long-term averages?
Will the model be called on to compare trends in air quality between two
different scenarios of urban population growth? Will the model be used to
select a control plan that will result in a given hourly air concentration's
being exceeded only once a year? A properly designed set of performance
standards will allow a potential user to compare models with respect to
suitability for any specific application. The particular performance char-
acteristic of interest influences strongly the rank-ordering of models on a
scale of goodness.
Fundamental to the definition of an optimal set of performance measures
will be the relationship of risk (of health, property, or aesthetic attributes)
to exposure (average pollutant concentration, time-integrated pollutant con-
centration, synergistic combination of pollutant dosages, or dosages integrated
5-65
-------
with respect to space, time, and populations^ Derived from the risk factor
will be, not a single number, but a distribution of effects for each degree of
exposure.For example, a range of pulmonary effects can be expected in a sample
population in which each individual has been exposed to ozone at an average
concentration of 100 yg/m^ for 5 years. The expectation value of the effect
will be the risk factor that is the function of exposure described above.
The model performance index will utilize these relationships to connect a
probability density distribution output from the model (associated with imperfect
knowledge) to a probability density distribution of the threat to public health
and welfare. Stated in a different way, each model will be assessed on the
basis of the uncertainty of damage estimate that arises from it imperfections.
This must be done in an unambiguous way for the user, who may not be a specialist.
The second step that will be needed to ensure ready application of air
quality models is largely a question of packaging and presentation. User-
oriented documentation will be needed for personnel at data processing centers,
who may not be specialists in chemistry, mathematics, or meteorology. Expe-
rience has shown that the user desires to operate the model in his own data
center and wishes to understand enough about the model structure to explain it
to others in his field. Models that cannot be adapted to these requirements
have not been widely applied. In some cases, an operating manual intended for
persons with some knowledge of programing will need to be rewritten to allow
the user to supply completed data forms to a computer center and routinely
receive output in return. Other adaptations may require a user to punch data
in on a teletypewriter and receive output on the same machine in an interactive
mode. This involves a network of remote terminals served by the computer
center, such as that under development in UNAMAP.11
5-66
-------
Output displays will be required to bring the abstract aspects of
voluminous output data into some form that appeals to the experience of the
user. Isopleth maps are useful, as are three-dimensional isometric plots
like SYMVU, produced by Harvard University. Printer plots of concentration
maps will undoubtedly enjoy an even greater application, because of the common
availability of line printers or teletypewriters as output devices. Examples
of these techniques are SYMAP and GRID, both produced by Harvard University.
Another aspect of matching output to user needs involves presentation of
results in a statistical framework—namely, as frequency distributions of con-
centrations. The output of deterministic models is not directly suited to this
task, because it provides a single sample "point" for each run. Analytic
linkages can be made between observed frequency distributions and computed
model results. The model output for a particular set of meteorologic conditions
can be on the frequency distribution of each station for which observations are
available in sufficient sample size. If the model is validated for several
different points on the frequency distribution based on today's estimated
emission, it can be used to fit a distribution for cases of forecast
emission. The fit can be made by relating characteristics of the distribution
with a specific set of model predictions. For example, the distribution could
be assumed to be log-normal, with a mean and standard deviation each determined
by its own function of output concentrations computed for a standardized set
of meteorologic conditions. This, in turn, can be linked to some effect on
people or property that is defined in terms of the predicted concentration
statistics. The diagram below illustrates this process:
5-67
-------
Concentration field
predictions
Statistical Predicted Injury or Expected
^ module frequency distributions damage —. harm to health
Historical air^^ ^- module-^"- and weifare
quality data
We have seen the wide variety of methods now available to calculate air
quality. The priority for adapting these methods to current needs is clearly
established. Only through clear expositions of model performance and simple
implementation procedures will the present techniques have a favorable impact
on air quality management. A growing appreciation by the specialist community
of the policy requirements will be essential for the successful fulfillment
of these goals.
5-68
-------
REFERENCES
1. Wanta, R." cT Meteorology and &it pollution, pp. 187-226. In A. C. Stern, Ed.
Air Pollution. Vol. 1 (2nd ed.) New York: Academic Press, 1968.
2. Pack, DTK." Meteorology of air pollution. Science 146:1115-1128, 1964.
3. Neiburger, M. The role of meteorology in the study and control of air
pollution. Bull. Araer. Meteorol. Soc. 50:957-965, 1969.
4. Gifford, F. A., Jr. An outline of theories of diffusion in the lower layers
of the atmosphere, pp. 65-116. In D. H. Slade, Ed. Meteorology and
Atomic Energy. Oak Ridge, Tenn. : U. S.. Atomic Energy Commission, 1968.
5. Stern, A. C., Ed. Proceedings of Symposium on Multiple-Source Urban Diffu-
sion Models. Air Pollution Control Office Publ. AP-86. Research
Triangle Park, N. C. : U. S. Environmental Protection Agency, 1970.
6. Keiburger, M. , J. G. Edinger, and H. C. Chin. Meteorological Aspects of Air
Pollution and Simulation Models of Diffusion Transport and Reactions of
Air Pollution. Project Clean Air Report of Task Force No. 4. Riverside:
University of California, 1970.
7. North Atlantic Treaty Organization. Committee on the Challenges of Modern
Society. Proceedings of the Second Meeting of the Expert Panel on Air
Pollution Modeling, Paris, France, July 26-27, 1971.
8. Reiter, E. R. Chapter 7, pp. . In Atmospheric Transport Processes.
Part 3: Hydrodynamic Traces. AEC Critical Review Series. Oak Ridge,
Tenn.: U. S. Atomic Energy Commission, 1972.
9. Hoffert, M. I. Atmospheric transport, dispersion, and chemical reactions
in air pollution: A review. Amer. Instit. Aeronaut. Astronaut. J. 10:
377-387, 1972.
5-69
-------
10. Darling, E. M., Jr. Computer Modeling of Transportation-Generated Air Pollu-
tion, A State of the Art Survey. U. S. Department of Transportation
Report DOT-TSC-OST-72-20, June 1972.
11. Johnson, W. B. The status of air quality simulation modeling, pp. 114-127.
In Proceedings of the Interagency Conference on the Environment.
Washington, D. C.: U. S. Atomic Energy Commission, 1972.
12. North Atlantic Treaty Organization. Committee on the Challenges of Modern
Society. Proceedings of the Third Meeting of the Expert Panel on Air
Pollution Modeling, paris, France, October 2-3, 1972.
13. North Atlantic Treaty Organization. Committee on the Challenges of Modem
Society. Proceedings of the Fourth Meeting of the Expert Panel on
Air Pollution Modeling, Oberursel, Federal Republic of Germany, May
28-30, 1973.
14. Lamb, D. V., F. I. Badgley, and A. T. Rossano, Jr. A Critical Review of
Mathematical Diffusion Modeling Techniques for Air Quality with Rela-
tion to Motor Vehical Transportation. Washington State Highway Depart-
ment Research Report 12.1, Research Project &-1540, University of
Washington, June 1973.
15. Sklarew, R. C. Air quality simulation models for photochemical pollutants,
pp. . In A. C. Stern, Ed. Air Pollution. Vol. . ( ed.)
New York: John Wiley and Sons, (in press)
16. Seinfeld, J. H. Mathematical models of air quality control regions, pp. 169-
196. In A. Atkisson and R. S. Gaines, Eds. Development of Air Quality
Standards. Columbus, Ohio: Charles E. Merrill Publishing Co., 1970.
17. Earth, D. S. Federal motor vehicle emission goals for CO, HC, and N0x based
on drsired air quality levels. J. Air Pollut. Control Assoc. 20:519-
523, 1970.
5-70
-------
18. Schuck, E. A., A. P. Altshuller, D,, S. Earth, and G. B. Morgan. Relation-
ship of hydrocarbons to oxidarits in ambient atmospheres. J. Air
Pollut. Control Assoc. 20:297-302, 1970.
19. Schuck, E. A., and R. Papetti. Examination of the photochemical air pollu-
tion problem in southern California. Appendix D, pp. . In
Technical Support Document for the Metropolitan Los Angeles Intrastate
Air Quality Control Region Transportation Control Plan Final Promulga-
tion, Environmental Protection Agency, Oct. 30, 1973. (UNVERIFIED)
20. Merz, Ph. H., L. J. Painter, and P. R. Ryason. Aerometric data analysis--
time series analysis and forecast and an atmospheric smog diagram.
Atmos. Environ. 6:319-342, 1972.
21. Hamming, W. J., R. L. Chass, J. E. Dickenson, and W. G. MacBeth. Motor
Vehicle Control and Air Quality. The Path to Clean Air for Los Angeles.
Paper No. 73-73 Presented at 66th Annual Meeting of the Air Pollution
Control Association, Chicago, Illinois, June 24-28, 1973. (UNVERIFIED)
--i «*.*- _ -
22. Chang, T. Y., and B. Weinstock. Rollback modeling for urban air pollution
control, pp. 184-189. In Proceedings of the Symposium on Atmospheric
Diffusion and Air Pollution, American Meteorological Society, Santa
Barbara, California, September 9-13, 1974. (UNVERIFIED)
23. Chang, T. Y., and B. Weinstock. Urban CO concentrations and vehicle emissions.
J. Air Pollut. Control Assoc. 23:691-696, 1973.
24. Peterson, J. T. The calculation of sulfur dioxide concentrations over a
metropolitan area by using empirical orthogonal functions. American
Institute of Aeronautics and Astronautics, Paper 70-113, January 1970.
(UNVERIFIED)
5-71
-------
25. Trijonis, J. C. Economic air pollution control model for Los Angeles County
in 1975. General least cost-air quality model. Environ. Sci. Technol.
8:811-826, 1974.
26. Bruntz, S. M., W. S. Cleveland, B. Kleiner, and J. I. Warner. The dependence
of ambient ozone on solar radiation, wind, temperature, and mixing height
pp. 127-128. In Proceedings of the Symposium on Atmospheric Diffusion
and Air Pollution, American Meteorological Society, Santa Barbara,
California, 1974. (UNVERIFIED)
27. McCollister, G. M., and K. R. Wilson, linear stochastic models for forecast-
ing daily maxima and hourly concentrations of air pollutants. Atmos.
Environ. (in press) (UNVERIFIED)
28. Hanna, ST RT A simple method of calculating dispersion from urban area sources
J. Air Pollut. Control Assoc. 21:774-777, 1971.
29. Gifford, F. A., and S. R. Hanna. Modelling urban air pollution. Atmos.
Environ. 7:131-136, 1973.
30.Hanna, ST RT A simple dispersion model for the analysis of chemically reactive
pollutants. Atmos. Environ. 7:803-817, 1973.
31. Lamb, R? G., and J." Hf Seinfeld. A simple dispersion model for the analysis oi
chemically reactive pollutants. Atmos. Environ. 8:527-529, 1974.
(discussion)
5-72
-------
32. Hanna, S.~ R." Author's reply. /To discussion o£ "A simple dispersion model f6r
the analysis of chemically reactive pollutants.^7 Atmos. Environ. 8:
529-530, 1974.
33. Karplus, W. J., G. A. Bekey, and P. J. Pekrul. Atmospheric diffusion of air
pollutants: Analog computer study. Ind. Eng. Chem. 50:1657- 1660,
1958.
34. Ulbrich, E. A. Adapredictive air pollution control for the Los Angeles
basin. Socio-Econ. Plan. Sci. 1:423-440, 1968.
35. Friedlander, S?R., and J? HT Seinfeld. A dynamic model of photochemical smog.
Environ. Sci. Technol. 3:1175-1182, 1969.
•» - - -,
36. Eschenroeder, A, Z., and J. R. Martinez. Mathematical Modeling of Photo-
chemical Smog. General Research Corporation IMR-1210. Santa Barbara,
Calif.: General Research Corporation, 1969.
37. Eschenroeder, A. Q., and J. R. Martinez. Concept and applications of photo-
chemical smog models. Adv. Chem. Ser. 113:101-168, 1972.
38. Eschenroeder, A. Q., J. R. Martinez, and R. A. Nordsieck. A view of future
problems in air pollution modeling, pp. 1013-1027. In Proceedings of
the 1972 Summer Computer Simulation Conference. Lajolla, Calif.:
Simulation Councils, Inc., 1972.
5-73
-------
39. Martinez, J: R., R.~* A." Nordsieck, and A7 Q." Eschenroeder. Morning vehicle-start
effects on photochemical smog. Environ. Sci. Technol. 7:917-923, 1973.
. - . - - ' " • ' • 9f- , ,
40. Wayne, LT, RT Danchick, M? Weiaburd, AT Rokin, and A." Stein. Modeling photo-
chemical smog on a computer for decision-making. J. Air Pollut. Control
Assoc. 2l!334-340, 1971.
41. Mahoney, J. R., and B. A,, Egan. A mesoscale numerical model o£ atmospheric
transport phenomena in urban areas, pp. 1152-1157. In H. M. Englund
and W. T. Beery, Eds. Proceedings of the Second International Clean
Air Congress. New York: Academic Press, 1971.
42> Egan, B. A., and J. R. Mahoney. Numerical modeling of advection and diffusion
of urban area source pollutants. J. Appl. Meteorol. 11:312-322, 1972.
43^ Egan, B. A., and J. R. Mahoney. Applications o£ a numerical air pollution
transport model to dispersion in the atmospheric boundary layer. J.
Appl. Meteorol. 11:1023-1039, 1972.
44^ LambVR."* G., and M.~ Neiburger. An interim version of a generalized urban air
pollution model. Atmos. Environ. 5:239-264, 1971.
45 Seinfeld, J. H., S. D. Reynolds, and P. M. Roth. Simulation of urban air
pollution. Adv. Chem. Ser. 113:58-100, 1972.
46. Reynolds, S. D., P. M. Roth, and J. H. Seinfeld. Mathematical modeling of
photochemical air pollution--!. Formulation of the model. Atmos.
__ Environ. 7:1033-1061, 1973. _.._ _ . . _ „ _ -
47. Roth, PT M.~7 P." Jr W7 Roberts, M-K. Liu, 57 D? Reynolds, and J7 H? Seinfeld.
Mathematical modeling of photochemical air pollution— II» A model and
inventory of pollutant emissions. Atmos. Environ. 8:97-130, 1974.
48. Reynolds, S. D., M-K. lieu, T. A. Hecht, P. M. Roth, and J. H. Seinfeld.
Mathematical modeling of photochemical air pollution. III. Evaluation
of the model. Atmos. Environ. 8:563-596, 1974.
5-74
-------
49. Sklarew, R. C., A. J. Fabrick, and J. E. Prager. Mathematical modeling of
photochemical smog using the PICK method. J. Air Pollut. Control Assoc.
22:865-869, 1972.
50. Knox, J. B. Numerical modeling of the transport diffusion and deposition
of pollutants for regions and extended scales. J. Air Pollut. Control
Assoc. 24:660-664, 1974.
51. Shir, C. C., and L. J. Shieh. A generalized urban air pollution model and
its application to the study of SCL distributions in the St. Louis
metropolitan area. J. Appl. Meteorol. 13:185-204, 1974.
52. Chapman, S., and T. G. Cowling. Chapter -8, pp. . and Chapter 14, pp.
In The Mathematical Theory of Nonuniform Gases. Cambridge:
Cambridge University Press, 1939.
53a. flirschfelder, J. 0., C. F. Curtiss, and R. B. Bird. The equation of state
of gases at low and moderate densities, pp. 131-233. In Molecular
Theory of Gases and Liquids. New York: John Wiley and Sons, Inc., 1954.
53^ Hirschfelder, J. 0., C. F. Curtiss, R. B. Bird, and E. L. Spotz. The equation
of state of dense gases and liquids, pp. 234-335. In Molecular Theory
of Gases and Liquids. New York: John Wiley and Sons, Inc., 1954
53c. Hirschfelder, J. 0., C. F. Curtiss, and R. B. Bird. The transport properties
of dense gases and liquids, pp. 611-667. In Molecular Theory of Gases
and Liquids. New York: John Wiley and Sons, Inc., 1954.
53d. Hirschfelder, J. 0., C. F, Curtiss, R. B. Bird, and E. 1. Spotz. The trans-
port properties of dense gases and liquids, pp. 611-667. In Molecular
Theroy of Gases and Liquids. New York: John Wiley and Sons, Inc., 1954.
54- Penner, S. S. Chapter 2, pp.
In Introduction to the Study of Chemical Reactions in Flow Systems.
London: Butterworths Scientific Publications, 1955.
5-75
-------
55. Bird, R. B., W. E. Stewart, and E. M. Lightfoot. Chapter 20
pp. . In Transport Phenomena. New York: Wiley, 1960.
56. Korth, M. W., A. H. Rose, Jr., and R. C. Stahman. Effects of hydrocarbon to
oxides of nitrogen ratios on irradiated auto exhaust. Part I. J. Air
Pollut. Control Assoc. 14:168-175, 1964.
c-, Dimitriades, B. Effects of hydrocarbon and nitrogen oxides on photochemical
j / •
smog formation. Environ. Sci. Technol. 6:253-260, 1972.
58a. Scott Research Laboratories, Inc. 1969 Atmospheric Reaction Studies in the
Los Angeles Basin. Vol. 1. Program Design and Methodology. Data
Summary and Discusision. Final Report. Plumsteadville, Pa.: Scott
Research Laboratories, Inc., 1970. l_ 113 pp._/
58b. Scott Research Laboratories, Inc. 1969 Atmospheric Reaction Studies in
the Los Angeles Basin. Vol. 2. Commerce Ground Data. Final Report.
Plumsteadville, Pa.: Scott Research Laboratories, Inc., 1970. 535 pp.
59 Anscombe, F. J. Graphs in statistical analysis. Amer. Statist. 27:17-21,
1973.
60 Eschenroeder, A. Q. Comments on the California Transportation Control
Stratege. Invited Testimony before the Environmental Protection
Agency Hearing at Santa Barbara, California, March 19, 1973.
(Available as General Research Corporation IM-1741, April 1973.)
61 Martinez, J. R., and R. A. Nordsieck. Air Quality Impacts of Electric Cars
in Los Angeles. General Research Corporation RM-1905. Santa Barbara:
General Research Corporation, 1974.
g2 Reynolds, S. D., and J. H. Seinfeld. Interim evaluation of strategies for
meeting ambient air quality standard for photochemical oxidant. Environ.
Sci. Technol. 9:433-447, 1975.
5-76
-------
63. Reynolds, S. D., M. K. Liu, T. A. Hecht, P. M. Roth, and J. H. Seinfeld.
Urban Air Shed Photochemical Simulation Model Study. Vol. 1. Devel-
opment and Evaluation. EPA-R4-73-30a. Systems Applications Incorpor-
ated, July 1975.
64. Wayne, L. G., A. Kokin, and M. I. Weisburd. Controlled Evaluation of the
Reactive Environmental Simulation Model (REM) Vol. 1. Final Report
Santa Monica, Calif.: Pacific Environmental Services, Inc., 1973. _/176 pp.J7
65. E.schenroeder, A. Q., J. R. Martinez, and R. A. Nordsieck. Evaluation of
a Diffusion Model for Photochemical Smog Simulation. EPA-R4-73-012a.
Santa Barbara, Calif.: General Research Corporation, 1972. 226 pp.
66. Nappo, J., Jr. A method for evaluating the accuracy of air pollution models,
pp. 325-329, In Proceedings of the Symposium on Atmospheric Diffusion
and Air Pollution. American Meteorological Society, Santa Barbara,
California, September 9-13, 1974.
67. MacCracken, M. C., T. V. Crawford, K. R. Peterson, and J. B. Knox. Develop-
ment of a Multibox Air Pollution Model and Initial Verification for the
San Francisco Bay Area. Lawrence Radiation Lab. No. UCRL-73348, 1971.
68. Roth, P. M., S. D. Reynolds, P. J. W. Roberts, and J. H. Seinfeld. Devel-
opment of a Simulation Model for Estimationg Ground Level Concentrations
of Photochemical Pollutants. Systems Applications Inc., Final Report
71-SAI-21, 1971.
69. Banna, S. R. Urban Air Pollution Models—Why? ATDL Contribution File No.
83. Atmospheric Turbulence and Diffusion Laboratory, Oak Ridge, Tenn.
1973.
5-77
-------
70. Sklarew, R. D., A. J. Fabrick,- and J. E. Prager. A Particle-in-Cell Method
for Numerical Solution of the Atmospheric Diffusion Equation, and
Applications to Air Pollution Problems. Systems, Science and Software
Final Report 3SR-844, 1971.
71. Pandolfo, J. P., and C. A. Jacobs. Tests of an Urban Meteorological-Pollut-
ant Model Using CO Validation Data in the Los Angeles Metropolitan
Area. Vol. 1. EPA R4-730-025A. Hartford, Conn.: The Center for
the Environment and Man, Inc., 1973. 176 pp.
72. Whitney, D. C. Analysis of Results of Past Model Verification Studies.
Systems Application, Inc. Draft Report, August 1974.
73. Lissaman, P. B. S. A Simple Unsteady Concentration Model Explicitly Incor-
porating Ground Roughness and Heat Flux. Aerovironment Report No. 73-
129, 1973.
74> Wang, I. T., and D. M. Rote. A Finite line Source Dispersion Model for
Mobile Source Air Pollution. Argonne National Laboratory Paper 75-
135, 1973.
75. Beaton, J. L., A. J. Ranzieri, and J. B. Skog. Air Quality Manual: Vol. 1.
Meteorology and Its Influence on the Dispersion of Pollutants from
Highway Line Sources. Department of Transportation Report FHWA-RD-72-
33, 1972.
76 Beaton, J. L., A. J. Ranzieri, and J. B. Skog. Air Quality Manual: Vol. 2.
Motor Vehicle Emission Factors for Estimates of Highway Impact on Air
Quality. Department of Transportation Report FHWA-RD-72-34, 1972.
77. Beaton, J. L., E. C. Shirley, and J. B. Skog. Air Quality Manual. Vol. 3.
Trtffic Information Requirements for Estimates of Highway Impact on Air
Quality. Department of Transporation Report FHWA-RD-72-35, 1972.
5-78
-------
78. Beaton, J.. L., A. J. Ranzieri, E. C. Shirley, and J. B. Skog. Air Quality
Manual. Vol. 4. Mathematical Approach to Estimating Highway Impact
on Air Quality. Department of Transportation Report FHWA-RD-72-36,
1972.
79. Beaton, J. L., A. J. Ranzieri, E. C. Shirley, and J. B. Skog. Air Quality
Manual. Vol. 5 (Appendix to Vol. 4) Department of Transportation
Report FHWA-RD-72-37, 1972.
80> Beaton, J. L., A. J. Ranzieri, E. C. Shirley, and J. B. Skog. Air Quality
Manual. Vol. 6. Analysis of Ambient Air Quality for Highway Projects.
Department of Transportation Report FHWA-RD-72-38, 1972.
81. Beaton, J. L., E. C. Shirley, and J. B. Skog. Air Quality Manual. Vol. 7.
A Method of Analyzing and Reporting Highway Impact on Air Quality.
Department of Transportation Report FHWA-RD-72-39, 1972.
82. Beaton, J. L., and J. B. Skog. Air Quality Manual. Vol. 8. Synthesis of
Information on Highway Transportation and Air Quality. Department of
Transportation Report FHWA-RD-72-40, 1972.
83. Chen, T. C., and F. March. Effect of highway configurations on environmental
problems—dynamics of highway associated air pollution, pp. 35-40. In
H. M. Englund and W. T. Beery, Eds. Proceedings of the Second Interna-
tional Clean Air Congress. New York: Academic Press, 1971.
84. Danard, M. B., R. S. "Koneru, and P. R. Slawson. A numerical model for carbon
monoxide concentrations near a highway, pp. 152-157. In Proceedings of
the Symposium on Air Pollution, Turbulence and Diffusion, December 7-10,
__ 1971' - _
85, Calder, K. L. Air Pollution Concentrations from a Highway in an Oblique
Wind. Environmental Protection Agency, Division of Meteorology,
Research Triangle Park, N. C., August 1972.
5-79 :
-------
86. Sakuraba, s. Sensitivities of air quality prediction to input erros and
uncertainties, pp. 8-11. In A. C. Stern, Ed. Proceedings of Symposium
on Multiple-Source Urban Diffusion Models. Air Pollution Control
Office Publ. AP-86. Research Triangle Park, N. C.: U. S. Environmental
Protection Agency, 1970.
87. Knox, J. B. Numerical Modeling of the Transport, Diffusion and Deposition
of Pollutants for Regions and Extended Scales. Lawrence Livermore
Laboratory Report UCRL-74666, 1973.
88. Shieh, L. J., P. K. Halpem, B. A. Clemens, H. H. Wange, and F. F. Abraham.
A multiple-source diffusion model for "New York City, pp. 108-115. In
Proceedings of the Symposium on Air Pollution, Turbulence and Diffusion.
December 7-10, 1971.
89< Hidy, G. Final Report on the Aerosol Characterization Experiments. State
of California Air Resources Board. (in press)
90. National Academy of Sciences. National Academy of Engineering. Coordinating
Committee on Air Quality Studies. Conclusion and recommendations ]_ for
/
measuring oxidants_/,p pp. 6-8. In Air Quality and Automobile Emission
Control. Vol. 3. The Relationship of Emissions to Ambient Air Quality.
Committee Print Serial No. 93-24. (Committee on Public Works, U. S.
Senate) "Washington, D. C. : U. S. Government Printing Office, 1974.
5-80
-------
Chapter 6
MEASUREMENT METHODS
Photochemical oxidants are atmospheric pollutants produced by
a series of reactions between hydrocarbons and oxides of nitrogen in
the presence of sunlight. The recognized photochemical oxidants that
have been measured in ambient air are ozone, the peroxyacyl nitrates
(mostly as peroxyacetyl nitrate, abbreviated as PAN), and hydrogen
1,2
peroxide.
Chemical radicals such as hydroxyl, peroxyhydroxyl and various
alkyl and aryl species have either been observed in laboratory studies
or have been postulated as photochemical reaction intermediates.
Atmospheric photochemical reactions also result in the formation of finely
divided suspended particles, (secondary aerosols), which create
atmospheric haze. Their chemical content is enriched with sulfates (from
sulfur dioxide), nitrates (from nitrogen dioxide, nitric oxide and peroxyacyl
nitrates), ammonium (from ammonia), chloride (from sea salt), water and
oxygenated, sulfurated and nitrated organic compounds (from chemical
combination of ozone and oxygen with hydrocarbon, sulfur oxide and nitrogen
3
oxide fragments).
The chemistry of the photochemical reaction milieu including the
formation of secondary aerosols is covered in Chapter 3. Table 6-1 lists
the substances which have been identified and associated with the impact
of photochemical oxidants on breathing, eye irritation, plant damage and
visibility reduction. It is important to recognize this chemical complexity
when formulating strategies for controlling the emissions of the primary
-------
pollutants. It is also important that a review of the issues relates
air pollutants to their sources. This requires not only the measure-
ment of end products but also of their precursors. Table 6-1 shows
in general terms the present state of monitoring practice for each
of these substances.
The effects of photochemical smog initially observed in Los
Angeles were severe haze formation, eye irritation, plant damage and
4
rubber tire cracking. Haa.gen-Smit was the first researcher to
recognize that the severity of photochemical smog could be quantified
by measuring oxidants. This oxidizing property of smog was subsequently
more efficiently monitored by measuring the increase in color intensity
produced by the iodine released from potassium iodide solutions in
5,6
contact with air. As a consequence, photochemical oxidants have
been defined by air pollution agencies as atmospheric substances that
oxidize certain reagents.
The reagent most frequently used is a neutral phosphate -
buffered potassium iodide solution, calibrated -with a known source of
ozone. This reagent, which is particularly sensitive to ozone, is also
somewhat responsive to other atmospheric oxidants such as nitrogen
dioxide, the peroxyacyl nitrates, and to a lesser extent, to hydrogen
1,2
peroxide. Reducing agents present in smog (e. g. , sulfur dioxide,
SO ) have an effect on the reagent solution opposite to that of oxidants.
2
Various terms are used to describe photochemical oxidant con-
centrations. Two of these, "oxidant" and "total oxidant," are used
because the measurement method (usually the potassium iodide method)
cannot distinguish between oxidizing and reducing agents, which in
6-2
-------
Table 6-1
Substances and Monitoring Practices for Species Present in
Photochemical Smog as Precursors or Products
Substance
Gaseous
ozone
peroxyacetyl
nitrate (PAN)
hydrogen peroxide
nitrogen dioxide,
nitric oxide
hydroxyl
peroxy hydroxyl
alkyls
aryls
aldehydes, formaldehyde
acetaldehyde
acrolein
benzaldehyde
hydrocarbons, total
methane
total non-
methane
alkanes
olef ins
diolef ins
acetylene
aroma tics
Formula or Symbol
0
3
fi
CH COONO
3 2
H 0
2 2
NO
2
NO
HO'
HOO "
CH (CH ) * ,CH CO ", etc.
3 2 n 3
C H CH ° etc.
652
CH 0
2
CH CHO
3
CH=CHCHO
2
C H CHO
6 5
THC
CH
4
NMHC
C H to C H
26 n 2n+2
C H to C H
24 n 2n+l
C H to C H
46 n 2n-2
C H
2 2
C H to C H
66 n 2n-6
Monitoring
Practice
Routine
Occasional
Research
Routine
Research
None
None
None
Occasional
Occasional
Occasional
Occasional
Routine
Routine
Routine
Occasional and
research
Occasional and
research
Occasional and
research
Occasional and
research
Occasional and
research
6-3
-------
Table 6-1 (Cont.)
Substances and Monitoring Practice for Species Present in
Photochemical Smog as Precursors or Products
Substance
Particulate
sulfate
Formula or Symbol
SO
Monitoring
Practice
Routine
nitrate
ammonium
NO
3
+
NH
Routine
Occasional
chloride
Cl"
Occasional
water
oxygenated organics
H 0
2
HOOC(CH ) CH OH
2 n 2
HOOC(CH ) COOH
2 n
HOOC(CH ) CHO
2 n
HOOC(CH ) CONO
2 n 2
Research
Research
Research
Research
Research
Many others
6-3a
-------
combination, produce the "net" oxidizing ability of the air. The terms
"corrected oxidant" or "adjusted oxidant" are used to indicate that a
measurement has been corrected for responses to certain components
known to be present (usually sulfur dioxide and nitrogen dioxide),
3a
other than ozone, peroxyacetyl nitrate or hydrogen peroxide. None
of these terms: "photochemical oxidant," "total oxidant," "oxidant,"
"corrected oxidant, " and "adjusted oxidant" has an exact meaning since
the reaction measured in the potassium iodide method is produced by
the presence of a number of substances, each responding differently.
Ideally, each of the major oxidants: ozone, nitrogen dioxide, PAN,
and hydrogen peroxide, and each of the major reductants: sulfur
dioxide and hydrogen sulfide, should be measured separately.
The potassium iodide method has been used extensively in
California to measure the net oxidizing properties of atmospheric
pollutants -without identifying the particular species of oxidizing or
reducing agents present. Studies have shown, however, that in
2,7
California, ozone is the major oxidant. At most sampling
locations in California the negative interference from sulfur dioxide,
and the positive interferences from other oxidants have not substantially
1,2
altered the perception of the photochemical smog problem. This
is both because the levels of sulfur dioxide are in general small
compared to the relatively high observed ozone levels and because
the positive interference from nitrogen dioxide tends to offset the
negative interference from sulfur dioxide. In other parts of the United
States where the peak ozone concentrations have generally been much
lower, the potassium iodide measurement method has produced data
6-4
-------
which has been useful in establishing that an oxidant pollution problem
8,9
exists. But because it lacks sufficient specificity, it is unreliable
for measuring the ozone occurrence in these areas. In contrast,
there are specific measurement methods that provide reliable
relationships between ozone concentrations and other aerometric
10
variables.
In addition to the specificity of the monitoring method, another
important requirement for the measurement of atmospheric pollutants
is the accuracy of the calibration technique. The calibration procedure
for the measurement of oxidants or ozone utilizes a stable and reproducible
sample of dilute ozone in air. The ozone concentration of this sample is
established, using a reference method that is not necessarily suitable for
monitoring ambient air. This reference method must agree with the
scientifically accurate measurement of ozone in the calibration
sample.
Whether ozone or oxidant measurements can be correlated with
other variables is highly dependent on the choice of sampling site and
the manner of sample transfer. The importance of these criteria is
discussed in Chapter 5 and the desirable specifications are described
in this chapter.
The emphasis here is on the various routinely used measurement
processes. These create huge banks of data which become influential
in deciding public policy. Measurements made only occasionally or
during research (Table 6-1) are more easily scrutinized at the time
of data application and therefore are not discussed here. Specifications
6-5
-------
for most types of available; instrumentation have been compiled and
discussed by the Environmental Instrumentation Group, Lawrence
11
Berkeley Laboratory.
UNITS OF EXPRESSION
Ozone or oxidant concentrations are commonly reported in the
units; volumes of ozone per million volumes of air (ppm), or -weight of
ozone per cubic meter of air (yg/rn-^) . The scale for the units in volume
per volume or weight per volume may be varied to avoid small decimals or
large whole numbers.
The commonly used units of expression are explained in
3
Table 6-2. Expressing concentration data in yg/m facilitates relating
ambient concentrations to emissions. This practice is generally
accepted as standard by the Environmental Protection Agency (EPA) in
the United States, and similar agencies in other countries.
Expressions of volume per volume units (ppm, pphm or ppb)
simplify measurements because their value is independent of atmospheric
temperature and barometric pressure. The volume units are equivalent
to the ratio of the number of molecules of ozone to the number of molecules
of air. This facilitates quantification of the atmospheric chemical re-
actions which lead to the formation of ozone. These units are also
preferable when the molecular weight of a substance is uncertain, as
is the case when reporting; total nitrogen oxides or total aldehydes.
Particulate matter components cannot be expressed on a volume
to volume basis, but thev can be expressed on a mole per unit volume
O
basis, u moles/m , If the molecular weight is known. This would
be a convenient unit to use when investigating the relationship between
particulates and gaseous species.
6-6
-------
Table 6-2
Units of Ozone Concentration and their Interconversion Factors
Units
Explanation
Units
ppm
yi/A
pphm
ppb
3
mg/m
3
yg/m
3
160 yg/m
Volume of ozone per million volumes of
air (or molecules of ozone per million
molecules of air)
Microliters of ozone per liter of air
Volume of ozone per hundred million
volumes
Volume of ozone per billion volumes
of air
Milligrams of ozone per cubic meter of
air at 25 C and 1 atmosphere pressure
Microgram of ozone per cubic meter of
air at 25 Cand 1 atmosphere pressure
Ozone air quality standard
= yl 0 /A air
3
= 100 pphm
= 1000 ppb
3
= 1.96 mg 0 /m air*
3 3
= 1960 yg 0 /m air*
3
= ppm
= 0.01 ppm
= 0.001 ppm
= 0.51 ppm
= 0.00051 ppm
= 0.051 pphm
= 0.51 ppb
= 0.08 ppm
* When the temperature is 25 C and the barometric pressure is
1 atmosphere.
Temperature and pressure do not influence the value of the volume of
volume units.
6-7
-------
3
The conversion of ppm to yg/m is calculated with equations
(1) and (2):
3
pg/m = (ppm) 1.219 x 1Q4 PM (1)
T
and
3
ppm = ( ug/m ) T , (2)
4
(1. 219 x 10 ) P M
•where M is the molecular weight of the gaseous substance being
measured (for ozone, 0 , M = 48 g/mole),
3
P is the total gas pressure in atmospheres (atm)
T is the temperature in degrees Kelvin, (degrees Celsius
4
+ 273), and the number, 1. 219 x 10 is the reciprocal of
the gas law constant, R, in units moles T/atm m-5,
consistent with the units used in the equations.
As these conversion equations show, in addition to knowing the
molecular weight of a. measured gaseous pollutant, the temperature and
pressure at the time of the measurement must also be known. Since
this information is frequently not given in the literature, no attempt
has been made in this report to convert to a common unit.
6-8
-------
SAMPLE GATHERING
Several sample gathering techniques are available for oxidant
and ozone measurements. Typically, outdoor air, taken with a sampling
probe is ducted to a sampling manifold located in a temperature regulated
space. Air is withdrawn from the manifold either continuously into
automated ozone monitoring instruments or intermittently using cumulative
ozone absorbers. Because ozone reacts quickly with some substances,
reliable measurement requires careful attention to the details of
sampling site selection, sampling frequency, and sampling train materials.
Ozone measurements are made indoors, outdoors near ground level, and
from aircraft, for different though related purposes. Nationwide uni-
formity in sampling techniques is being achieved through cooperation
among the various control agencies in specifying sampling conditions.
The location of the probe necessitates careful design and
documentation to ensure that surrounding environmental conditions do not
interfere with the subsequent interpretations of the measurement data.
Examples of such minimum site descriptions are given in Table 6-3.
Some of this information is now included with all the data logged into
the Environmental Protection Agency's National Aerometric Data Bank
and some was provided by the control agencies. An example of a site
description in Southern California based on the requirements of the
California Air Resources Board is shown in Table 6-4 (2 pages). In
addition to the site description, there is a statement of all the con-
ditions being monitored, the methods used, and the numerical specifi-
cations for the sampling probe both for the sampling manifold and for
the connections from the manifold to the instruments. A third page
6-9
-------
I
\D
43
CO
H
B
0 i-J
•H W
n) 01
> >
cu o
rH 43
W <
*
*y?
£_|
D
B
O
•H
4-1
£ B
IH O
CJ -H
CO 4-1
CU P.
O -H
M
0) O
4-> CO
•rl CU
CO Q
00 CU
B 4-1
•ri -H
IH CO
0
3 *a
B co
O
a .
o
B
CU
00
*"O
B
CO
B
O
•H
4-1
CO
O
O
I-H
s-*.
co
cu
4-1
CU
£
CO
CU
CJ
IH
3
O
CO
CU
^4
•H
MH
O
•
>
•H
Q
A
B
O
•H
4J
5
IH
CU
CO
B
O
o
rH
CO
4-1
B
CU
O
IH
•H
W
MH
O
•
4-1
P.
CU
O
cu
4-1
CO
CO
s-^
en
o
fy]
u
0"
^^
>^
^i
o
CU
a
•
T-H
w a
CO
B
O
•H
4J
O
3
rl .
4-1 IH
CO CU
43 >
0 -rl
Pi
e
O 4-1
IH CO
MH cO
W
CU
cu B
M iH
MH
T3
CO B
CU CO
M rH
CO CO
M
B
CU CU
ft, ^4
O CO
MH
XrH
rH CU
M IS
•H
CO MH
MH O
• « p.
rH -rl
CO 4-1
•H
cu cu
•a 43
•H 4-1
CO IH
01 O
erf B
CO
o
1
en
0
•£
•d CU
B >
CO iH
rH A
CO
H 4J
co
CU CO
IH W
CO
MH O
rH iH
CU Z
IS ^
,-H
^D
O
4-<
vD
~*
w a
CO CN
ON rH
rH rH
NO in
o
•t
cO
CU
IH
cO
B -d
cu B
P. 3
O O
• « CO
CO rH
EC rH
CU CO
o
p.
4-1 CO
i — 1 M
01 00
> o
01 P,
CO O
O 4-1
SrH
0)
IH >
cO CU
CU rH
B
•> CO
rH B
CO O
•H -H
4-1 4-1
B 0
cu 3
13 IH
•H 4-1
CO CO
CU 43
pcj o
O
i-H
O
C7N
CM
t
H
^
•
^
• PH
CU
> rl
•5) cu
fi*i o
0 43
•H B
IH CU
M CO
S W
W 55
ON 00
m ^H
o on
NO in
•*
• 43
rH O
0) -H
> 43
CU ft
rH
B
X -H
rH
IH B
•H -H
CO CO
MH CO
43 •
4-1 CO
3 rH 3
43 rH -H
4-1 B CO
O CO M
P.
10 B -
•H O
Es o
O T3 OO
rH CU
... CO CO
rH CJ
cO O B
•H ,-3 -H
0 43
H 4->
0) « -H
B CU
O CD -
03O
o m
T3 43
B co
cO X cu
43 CO
rH -rl
CO -d IH
•H CU
4J T3 T3
B B B
CU 3 co
•O O rH
•H IH
CO H CU
CU 3 43
prf CO 4-1
t— i
O
In
CTN
m
•
^
•
55
O CO
CU 4-1
CO • -rl
4) CU X
B *
J "5 tt1
^ m &H
rH
O
IH
4-1
B
o
o
B
O
•H
4-1
3
rH
rH
O
IH
•H
•^3
MH
o
3
CO
01
3
*
B
O
•H
4J
0
CU
4-1
O
PH
rH
CO
4-1
B
0)
cj
p3
0
IH
•H
1
MH
O
a
4-1
P.
CU
O
CU
4J
CO
4J
CO
s—^
en
o
PM
u
Of
*~*
^
0)
to
O)
cu
a
f
CN
VD
"*
W 55
-* m
& 6
s*^
^
f_l
ca
O-(
^*>
4-1
B
0
u
cu B
B O
B CO
o -a
prj v_^
CN
CN
rH
W E5
^o in
ON O\
VO ^3
m m
o-
T3
r, CU
B >
O O
•H B
4-1
o •••
3 cu
IH M
4-1 CO
CD 3
43 0-
O CO
^*> r^J
43 0
O
T3 rH
CU 43
T3
B rH
3
O 4J
IH O
IH rH
3
to oo
B
• -H
rH .^
CO IH
•H CO
O P,
IH
CU CO
E MH
0 0
O
IH .
* cu m
f*. B r^
4-1 IH ON
•H O rH
0 o
IH cu eg
a) B S
4-1 O
C rO
CU B -rl
0 o B
CU
CJ
CO
rH
C«
T3
IH
O
MH
cO cO
B M
CU FQ
i — i
[jq ^.
o ^
^O u-i
^O QN
U~>
^j
cu
B
•H
MH
0)
^i
0
4-1
4-1
B
CU
O
CO
•I—)
T3
cO
i-H
CO
•H
4-1
CO
3
T3
•rl
•t
B
CO
43
IH
3
43
3
C/3
,— -^
4J
B
CO
rH
PH
0)
00
CO
B
01
CO
4J
B
•rl
O
PH
£**•>
cu
B rH
cu B
T3 CU
fi IH
•H H
, ~] ^^
•^
W tt
*~l "^
m m
OO CT*
CN (N
en
M
(B
0)
B
(S
B
o
•rl
4-1 •
cj !>,
3 43
IH B,
4J CO
CO IH
43 00
0 0
B 0
O 4J
MH rH
CU
01 >
cu cu
IH rH
MH
*
CO O
CU rH
co m
H
B
01 T3
P. B
O cO
A •
rH T3
CO Prf
•H
IH rH
4-1 CO
•O CO IH
IH 30)
cd *"0 *"O
o B oi
M M tu
i-H
0
IH
4J
B
O s~-
U -H
IH O CU
•H ^O -rl
< 1 B
in B
CD i-H O
co in u
i*5 ^-3" ^~^
tu
H
/•— X
^-1
CN
(-^
8-
""— ' ,^l /^-x
IH •
B CO 4-1
O PH CO
4-1
CO IH 01
3 cu cO
o o) S
D3 Q *"-^
t
en
o^
r^
W fs
o 00
-H B
•H
• •» i — I
B -rl
O CO
•H >
4.1 0)
CJ IH
3 PH
IH
4-1
to •
43 4J
O M
O
e p.
0 IH
IH -rl
MH cO
0) MH
CU O
IH
MH CD
B
cO J2
CU
IH CN
cO •
CO
B
0) •"
P.ON
O CM
•\ •
cO B
•rl 33
4-1
B MH
CU O
•H 4-1
CD CD
0) 0)
/—s
oo
cu
•H
B
B
O
u_
X— N
B
0
4-1
to
3
(U O
B 33
•rl
•d •
1-1 5
•
s~^
•^
1
vTJ
01
rH
43
CO
H
CU
CU
CD
^-s
4-1
O
•H
IH
4J
CO
•H
Q
rH
o
IH
4-1
B
O
o
B
0
•rH
4-1
3
rH
rH
O CO
PH •*-'
•H
^ B
•H 3
IH
>> CU
4-1 4-1
B eu
3 B
o o
U rH
•H
cn A!
01
rH 0
CU rH
00 —
B rH
^
n
CO IH
3 S
CO
s-~, CJ
sr M
CN CU
o s
o to
0- rj
<; cu
^~^ ^
CO
to B
cu co
rH IH
CU H
00
B rH
•52 co
to
to M
O V
hJ >
•H
B
• | i
-------
Table 6-4
State Air Resources Board Sampling Site Report
Date July 1, 1974
60
City
Azusa
Prepared by_
WDH
Air Monitoring Station No.
Street Address 803 North Loren Avenue, Azusa, California
Longitude: W 117 Peg. 55 Min. 22 Sec., Latitude: N 34 Peg 8 Min 9 Sec
Nearest Street Intersection Loren Avenue and Foothill Boulevard
Date Operation Initiated at this Location January 30, 1957
Name of Operating Agency Los Angeles County Air Pollution Control District
Address of Operating Agency_
District
Los Angeles
434 South San Pedro Street, Los Angeles, California 90013
Basin South Coast
SITE DESCRIPTION
CATEGORY
f CENTER CITY [TjSUBURBAN
RURAL
I 1 REMOTE
Type of Street Traffic Arterial
SUBCATEGORY (Dominating Influence)
Dp Industrial Q Residential £3 Commercial £7Mobile
QNear Urban £J Agri f"7 Comtn /~7 Ind CJ None of these
(Residential, Expressway, Commercial, Downtown, Arterial)
Outstanding Landmarks and Relation to Site:
Station location at foothills of San Gabriel Mountains.
Foothill Freeway a mile south.
Aerojet plant S.E. 2 miles;
Surrounding terrain and community characteristics (hills, valleys, flat, bodies of water>
residential, industrial, commercial, rural, open, forested, row crops, grassland, orchard,
etc.).
San Gabriel Mountains four miles to the north; small manufacturing plants around station
site. Some residence near site - single family next door to south, trailer park across
Loren to the east. Commercial area of Azusa 2 miles to the east. Flat, gently sloping terrain.
Possible nearby sources (refineries, stacks, chimneys, gasoline stations, Dower plants,
parking lots, traffic, agricultural operations, etc.)
Source
Direction
Small Iron Foundry
(usual equ ipment)
Large Gravel Plant
Large Gravel Plant
Freeway
Limited Parking
Metal Melting Furnace
(no permits on file so
apparently a minor source)
Distance
Pollutants
southeast
northwest
southwest
south
east and north
southwest
1000 feet
1% miles
3 miles
1 mile
Adjacent
100 feet
Normal
Particulate
Particulate
Auto Exhaust
Auto Exhaust
Normal - Operated
intermittently,
regular schedule
Attach 8-V1 X 11" portion of local street map indicating station location.
6-11
-------
Table 6-4 (Conto)
Air Monitoring Station No. 60
Gaseous Measurements
Gases Method
0
3
0 X Potassium Iodide - Colorimetric
X
NO X Modified Saltzman - Colorimetric
NO x Modified Saltzman •- Colorimetric
Air Flow
L/Min.
3.7
0.225
0.225
2
NO X Addition of NO and NO from above
x 2
THC X Flame lonization
THP TP Y VI amo Trim* *r al-n nt-i
4
CO X Non-dispersive Infrared
SO X Conductivity - Hydrogen Peroxide
0.05
0.05
Oo5
20 CFH
2
OTHER
Instruments located on 1st floor of 1
PROBE DESCRIPTION
Size 1 inch I.D. Vertical
story building.
AIR INTAKE
(up, down)
Material Pyrex pipe Angled 45° down ; Direction
Flow rate 30 L/Min. (degrees, up,
Length 23 Ft. Horizontal X
Sample Residence Time in Probe 7.1 Sec. Height above:
Number of Changes in direction 1 sea level
down) (N,S,E,W)
Direction S
(N,S,E,W)
Ground level 71/3 ft.
607 ft0; roof - ft.
Number of Instruments connected 5
Horizontal distance to nearest street curb 43 ft.
If the probe does not extend above the roof, indicate the intake distance down from the
roof 4 ft.; out from the building wall \\ ft.
Note: If more than one probe utilized, provide the above information for each. Attach a
single line schematic diagram of probe arrangement(s) indicating inside diameters and
cumulative air flows of all segments»
Three to four feet of Js-inch Teflon tubing connects each instrument to sampling man.
Are wind instruments on same mast as station probe(s) inlet(s)? XXX No. If yes how much
higher? - ft. If no, distance and direction from probe(s) 25 ft.; direction W .
Height above: Ground level 24 l/3fi., roof 13 ft., sea level 624 ft.
Nearby obstruction(s) to wind flow: None
Type(s), distance(s), direction(s), size(s)
6-12
-------
(not included here) shows a schematic drawing with the dimensions and
locations of the bends in the ducting. The sampling probe specifications
currently in effect at four major air pollution control agencies are
summarized in Table 6-5.
When selecting a monitoring site, it is important to take account
of environmental features. For example, ozone measured in or near
automotive traffic can drop to 50% of the area-wide value owing to
reaction with the nitric oxide from exhaust emissions. While ozone
measured 7. 5m from a large tree in green leaf can drop to 70% of the
area-wide value, it may also be reduced within 1m of shrubs and grass.
Paint, asphalt, concrete, dry soil, and dead vegetation are not as
reactive and so have a lesser effect. Peak ozone values observed in
sunlit-^windscreened-concrete and-asphalt courtyards tend to be higher than
those recorded on an adjacent rooftop.
A preferred sampling location is open to the free movement of the
ambient air, at least 3m above grass, 8m from shrubs, 40m from large trees,
and 120m from any heavy automotive traffic. The flat roof of a one-or two-story
1
building is ideal. A probe which projects through a wall and extends
about 2m beyond, is an acceptable alternate approximating the roof.
To minimize undesirable gas phase reactions in the probe, the
gas flow must be regulated to keep the transit time in the sampling ducts
as short as possible. When a probe is longer than 30cm, passage of the
ozone-air sample should take less than 20 or 30s. Losses can also occur
from reaction -with the probe itself or with the accumulated dust which
coats the inside.
6-13
-------
g,
•rl
CO
CU
t~\
Table 6-5
Sampling Probe I
Houston
cu
CO
I-l
CU
>->
§
3
^
IH
O
!*
I
CD
CU
rH
0)
OC
3
CD
3
0
T3
M
T-H
O
M
4J
P
0
CJ
rl
•H
<:
CD
cd
s
E-i
IH
•H
.
o
i g
J 00
^ <
vO
-*
o
4J
CO
vo
o
4-1
-a-
VD
o
4-1
CM
CM CO
0 >.
4J rH
4-1
CM CO
• 0
CM g
S~\
\™ ,/
i^
g
3
Q
^.j
00
QJ
o
•3
4-1
00
•H
CU
m
O
T3
D
13
CU
4J
CO I-l
CD CU
Cd >
rH P CO
60 -rl m rH
O
•o
CO
CO
rd
rH
00
en
CO
cd
rH
00
CO
CO
cd
rH
00
rH
cd
Materi
p
13 v£>
CU •
4-1 r» CT\
IH
CU O o
> -U 4-1
•rl in r^-
p
o
T3
£>
•o
CU rH
4-> rH
rl
CU O
li ^ u
•H r~- t^-
p
o
T3
P
T3
CU VO
4J rH
M
cu o
> LI 4J
P "
•H c~J i—
f—\
0
0
*^>
^^
• B
rl -2-
cu
4-1 A
9J A
H 4^
cd £0
•H P
•a cu
t-H
cu
M O
4J CO CU
cd SH
IH 14H
O W
U CU
o B 4-1
rH 3 Cd
Pn cn 2
Heat obviates
condensation.
none
none
none
CO
rH
O
t4
4-1
P
0
O
>-l
4J
•rl
TJ
•H
3
43
•
rH
cu
Pd
U
CM
+1
CJ
CO
-H
/— s
CJ
CM
CM
1
O
CM
s_>
a
^1
CJ
m
-H
CD
rH
O
IH
4J
P
O
o
cu
rl
3
4-J
td
M
CU
I-
CU
H
no screen none
r>
M
cu
rH
•H
M-l
g
rl
CU
4J
screen, no fil
none
co
M
CU
•y
iH
•rl
MH
Inlet
-------
The materials of which probes are constructed must be sufficently
inert to trace amounts of ozone to prevent adsorption, absorption, and
reaction between the ozone-air sample and the sample line during the
transit period. Teflon and glass are both relatively inert, but Teflon
has the advantage of being unbreakable and thus more durable. Nalgon
vinyl tubing is satisfactory for connectors and even for a short probe, if
nothing else is available. Ozone reacts with Tygon vinyl, stainless steel,
and aluminum, all of which should not be used.
To prevent the formation of reaction products from the interaction
of the ozone-air sample with filters, they are intentionally not used at
probe inlets (see Table 6-5). Some of the newer instruments however,
require filters at the inlet of their sampling ports to prevent the
particulate matter in the ambient air from fouling reaction chamber cells
or from clogging the gas-flow controllers. When the same type of filter
also precedes the calibration and zero gas sampling ports (-which has not
always been the practice), the problem is minimized to the extent that similar
events occur during the calibration and sampling.
CALIBRATION
Most currently used oxidant and ozone monitors need to be cali-
brated -with a predetermined concentration of ozone in air. Regardless
of the principle used to measure ambient ozone or oxidant concentrations,
the primary reference standard for calibrating each monitoring device
or system should be identical everywhere. This requirement remains to
be achieved in practice. Up to June 1975, at least seven calibration
6-15
-------
procedures were practiced in the U.S. These are listed in Table 6-6
along -with the agency, primary use and current status of each method.
12-16
Comparisons of these methods were recently completed.
Ozone Generation
For reliable calibrations, it was necessary to develop a stable
and reproducible ozone source which could produce ozone in air at con-
19
centrations smaller than ppm. After this was accomplished several
different versions were engineered which are now available commercially
20-24
from most ozone-monitoringvendors of instruments. The factors
affecting the production and survival of oxygen species other than ozone
are discussed in Chapter 12. Care must be taken to prevent these
species from creating interference when generating ozone for instrument
ca.lib ration.
A typical source (see Figure 6-1) consists of an ultraviolet
mercury lamp which irradiates a quartz tube through which clean air
flows at 5-10 1/min. A small amount of the oxygen in air is converted
to ozone by photolysis. It is important that the incoming air is free of moisture
nitrogen oxides, sulfur oxides, hydrocarbons and particles to avoid
producing inadvertent interferences with subsequent sensors. Ozone
concentrations from 0 to 1 ppm can be generated by varying the ultra-
violet radiation intensity by means of an adjustable shield around the lamp
envelope. The air stream flow rate is controlled by a needle valve and measured
by a rotameter or a mass-flow meter. The ozonized air passes to a manifold
from which the testing monitor draws its sample. The ozone concentration produced
in this way is solely an empirical function of the settings of the lamp
6-16
-------
co
00 00
'•O
1
'•O
01
rH
Xi
cd
H
CO
•H
CO
[
cd
c
<
H
c
CO
T3
•H
><
"
^
o
OJ
c
o
N
o
MH
o
c
o
•H
4-1
cd
XI
•H
rH
cd
CJ
IH
o
MH
CO
T3
r-l
td
TJ
G
cd
4-J
en
cu
CJ
CU
IH
cu
CO
3
4J
cd
4-1
CO
P-I
w
CO
&
rH
(0
•H
CJ
•H
MH
MH
0
a
o
•rH
4-1
CO
U
•H
rH
p.
o.
^^
K^
o
g
60
60
c
•H
l-l
o
4J
•H
rj
o
e
J^
•H
CO
<2
P-I
w
en
CD
c
o
•rH
4-1
•rl •
4J T3
cd o
4-1 XI
El 4-1
0) CU
H g
c
o
•H
4-1
cd
o
•rl
MH
•H
^
cu
•*^
^4
&
CO
H
P-I
o
mental
CX
o
rH
Ol
^
Ol
o
60
pi
•rl
(-1
o
4J
•H
o
e
J_l
•H
CO
rj
•rH
^\
1 — 1
^s
XI
4-1
•H
^
^^
rH
CO
01
0)
* CO
o
1
td
CJ
a
•H
rH
cd
•rl
o
•H
MH
MH
o
• n
60
c
•H
^J
o
4J
•H
c
o
E
p
•H
CO
pq
P-H
^
•
MH
•rl
rH
cd
CJ
^.
m
CM
i
rH
cu
h— )
01
o
rj
•rl
CO
•
CO
XI
cd
rH
XI
o
^1
cd
CU
CO
cu
X!
o
M
cd
01
CO
01
•
CO
Xi
cd
rH
X!
0
y^
CO
CU
CO
cu
cd
en
m
o
rH
H
a
0
•H
4-1
CO
o
•rl
rH
ex
a
CO
cu
r-H
a
•H
o
c
rH
cd
IH
4J
3
01
c
gsg
t-H
• *>
o
•H
J_l
•p
(U
•1
IH
o
rH
o
CJ
cu
•a
•H
•a
o
•H
g
3
•H
CO
CO
cd
4-1
o
ex
r^
IH
01
MH
MH
3
Xi
cu
rH T3
cd -H
IH -O
4-1 O
3 -H
0)
c e
3
r>« -H
CM CO
CO
• » cd
a 4-i
•H 0
>-i a
4-J
CU T3
e cu
•H IH
VJ CU
O MH
rH MH
O 3
CJ ^
-d
0)
H
01
MH
MH
3
Xl
c
3
r>S
CM
• f.
0
•H
^
4J
cu
3
H
4J
•ri
EH
01
•rH
•a
0
•H
c
3
•H
CO
CO
cd
4-1
o
ex
0)
>H (3
o o
N
C O
o
•H X!
4-1 4J
cd -H
^1 &
•H Ol
4J T3
•H
cu X
co o
cd
XI 0
l_l
CO 4-1
Cd -rl
O CJ
MH
o
3
cd
cu
IH
3
rH
td
o
•H
4J
CO
a
M
o
CO
0)
c
cu
60
01
a
o
N
o
CO
•o
cd
T)
cd
4J
CO
MH
o
4-1
pj
cu
•rH
o
•H
MH
IH
cu
o
U
Hi
o
•H
4J
o
a
•H
4J
^
w
4J
CU
rH
o
•H
^
CO
S-i
4J _
rH e
3 a
cu o-
XI m
4J CM
C 4-1
•rl Cd
cu cu
C 60
o a
N Cd
O |J
MH
o
4-1
c
01
•rl
o
•rH
MH
MH
01
o
o
c
o
•rH
4-1
o
•H
4-1
^
w
60|
fj £i
ctf o
M
m
T3 m
-------
PLOW-
METER
NEEDLE
YALYE
CLEAS AIR
ULTRAVIOLET
LAMP
ADJUSTABLE &RADUATED1
' SHIELD
OZOtfE SOURCE
TO
INSTRUMENTS ^
TO SAMPLING
FOR
CALIBRATION
QUARTZ TUBE
Figure 6-1. Ozone source and manifold system.
r
VElfT
6-18
-------
shield and air flow rate, the air temperature, and the humidity.
Therefore, the output concentration must be measured using a scien-
tifically acceptable reference standard.
Ozone Analysis
All of the iodometric reference methods (items 1, 2 and 3,
Table 6-6) are similar in principle. They assume that ozone (O ) •when
3
in contact with iodide ions (I ) in aqueous (H O) solution releases a
2
stoichiometric amount of iodine (I ) according to the following chemical
2
equation:
O +21 + H O = I +O + 2OH (3)
3 222
where O stands for oxygen and OH for hydroxyl ions.
2
In California, the Los Angeles County Air Pollution Control
District (LAAPCD) and the California Air Resources Board (ARB) both
began continuous air monitoring for oxidants more than 20 years ago.
Oxidant monitoring at urban sites was begun over 5 years ago by the
National Air Pollution Control Administration and continued by the
Environmental Protection Agency (EPA). All these groups initially used
instruments containing the same absorbing solutions but their iodometric
calibration methods differed in detail.
California Air Resources Board (ARB) Procedure (prior to June
26
1975. The ARB reference procedure for ozone uses a 2% neutral
buffered potassium iodide reagent. Ozone is generated in air previously
humidified to about 50% relative humidity. The air sample to be analyzed
is drawn through the reagent in a midget impinger, and the ozone present
in the sample air liberates iodine from the iodide reagent. The quantity
6-19
-------
of iodine liberated is determined using a spectrophotometer which has
been calibrated with an iodine solution standardized with sodium thio-
sulfate solution which, in turn, has been standardized against primary
grade potassium biiodate.
25
U.S. Environmental Protection Agency (EPA) Procedure.
The EPA reference procedure for ozone is similar to the ARB procedure
except that the ozone is produced from dry air, the reagent is 1% neutral
buffered potassium iodide, and the iodine solution is standardized with
primary standard grade arsenious oxide. Oxidant minitors outside
California have generally been calibrated with this procedure.
Los Angeles County Air Pollution Control District (LAAPCD)
27
Procedure. The LAAPCD reference procedure for ozone uses a 2%
unbuffered potassium iodide reagent. Ozone is generated in air previously
humidified to about 50%. The air sample to be analyzed is drawn through
the reagent in an impinger of LAAPCD design and the ozone present in the
sample air liberates iodine from the iodide reagent. The quantity of
iodine liberated is determined by titrating the reagent-with sodium thio-
sulfate solution. The sodium thiosulfate used in the titration is standardized
with potassium dichromate solution. The procedure utilized by the
LAAPCD was modified in January 1974 to include the use of dry ice
in the standardization of the dilute sodium thiosulfate in order to improve
the precision of the titrations. This does not affect the mean values
28
obtained with their procedure.
These iodometric calibration methods are based on the assumption
that there is a stoichiometric reaction between ozone and the iodine in the various
6-20
-------
potassium iodide procedures. Three essentially independent
methods have been used to test the accuracy of this assumption. These
are: measuring the absorption of ultraviolet radiation at 254 nm by ozone
in air; measuring the absorption of infrared radiation at 9480 nm by ozone
in air; and determining the ozone in air concentration by titration -with
nitric oxide.
The first two determinations by radiation absorption require
accurate measurements of the extinction coefficients of ozone (a measure-
ment of the absorption efficiency of the incoming radiation at a maximum
absorption wavelength) in the ultraviolet and the infrared. Three different
principles have been used over the past 20 years to measure the extinction
coefficient of ozone in the ultraviolet at 254 nm: manometric, decom-
position stoichiometry and gas-phase titration. The manometric method
which is based on pressure measurements of gaseous ozone requires, in
34
at least one case, a substantial ar.d somewhat uncertain correction
for decomposition; and the method of decomposition stoichiometry depends
on the pressure change that accompanies the decomposition of ozone
31
to oxygen, 2 O -> 3 O . Clyne and Coxon determined ozone concen-
3 2
trations in a flow tube by titration with nitric oxide, a method essentially
equivalent to gas phase titration. These methods with their results are
summarized in Table 6-7. The best value for the ultraviolet extinction
-1 -1
coefficient at 254 nm is considered to be 134 cm atm at O C and
31a
1 atm (STP).
6-21
-------
Authors
Table 6-7
Ozone Extinction Coefficient Measurements
Reference
Extinction Coefficient
-1 -1
cm atm STP.base 10
Inn and Tanaka
Hearn
34
33
133
134
DeMore and Raper 30
Griggs 32
Clyne and Coxon 31
Becker, Schurath, and 29
Seitz
135
132
136 (250 nm)
135
Method"
Manometric
Decomposition
stoichiometry
Decomposition
stoichiometry
Manometric
Gas-phase titration
Manometric
*Methocl used to establish the ozone concentration.
6-22
-------
Measuring the absorption by ozone in air of ultraviolet radiation
at a wave length of 254 nm. with this value for the extinction coefficient,
a laboratory photometer is used to measure the primary reference con-
13
centration of the ozone in the gas stream used for calibration. A schematic
for this arrangement is in Figure 6-2. Highly accurate measurements are
attainable -with this type of photometer because the pressure, temperature
and path length of the light beam in the cell can all be controlled and
measured precisely.
The ozone concentration in parts per million is given by:
6
O (ppm) = 10 T log lo (4)
3 273 Pk£ I
t
where T = temperature, K (kelvin)
P = total pressure, atm.
k = extinction coefficient of ozone is 134 cm~^ atm at STP
£ = path length, cm
1 - intensity with carrier gas only
o
I = intensity with O present.
t 3
Because at concentrations smaller than parts per million, the
radiation intensities, I and I , have to be measured with very great
o t
precision, the determination using absolute photometry requires a
physical chemistry laboratory with an experienced staff.
Calibrations at monitoring sites, therefore, require a trans-
6-23
-------
CK
w
n
i-J
0,
M
a
p
«i
o
EH
O
X.
PL.
T_T
(X,
O
E-«
W
J_
-
W
O
K.
EH
C3
w
n
N> ?=
IHO
iK Q
•< s
§
M
01
4J
O
1U
O
N
O
4-1
rH
!=>
a)
4J
rH
O
in
CM
W)
•H
o
K
>.B
K. CO
^1
O A,
cK S
tJ
-------
ferable standard. The use of a commercially available photometer, which
displays the difference in signals between the ultraviolet light absorbed
at 254 nm with and without ozone in the sample cell, is one approach.
It has proven to be stable in performance even after transport between
12, 15
different laboratories. A schematic diagram of the Dasibi Model
1003-AH ozone monitor is shown in Figure 6-3. The operating principle
36
of the Dasibi instrument has been described by Behl, Bowman and
37 15
Horak, and by Hodge son je_t al_. An evaluation of this instrument has
35
been carried out by deVera, Jeung and Imada.
As it enters the instrument, the sample is diverted by the valve
into a chemical scrubber which changes any ozone present
to oxygen. This ozone-free sample then passes through an absorption
chamber 71 cm in length, where a detector measures the amount of
ultraviolet light at 254 nm transmitted through it. The integrated
intensity of the transmitted light is digitized and stored electronically
as the reference signal. On completion of this measurement the valve
is opened and the absorption chamber is flushed with the ozone-containing
sample. The digital counter now records the reduction of ultraviolet
light due to the ozone in the sample. This value is subtracted from the
stored reference measurement and the difference is displayed in concen-
tration units equivalent to parts per million ozone.
The ozone concentration (c) as measured by the Dasibi ozone
meter is in principle given by equation (4), which can be expressed
in terms of the instrument itself as:
6-25
-------
a)
S-j
BO
a)
•H
.M
O
O
•H
i
a)
g
tn
n)
O
3
SO
•rl
X
6-26
-------
O (ppm) = 22.4 T £n S (5)
3 kP S-R
where T = temperature, K, degrees kelvin.
S = span setting of instrument, « I
o
R= instrument reading °c (I - I )
° t _! -1
k= extinction coefficient of ozone is 134 cm atm at STP
P = pressure, atm.
At low ozone concentrations (R small compared to S), the above expression
can be simplified as follows:
O (ppm) = 22.4 T [n] . (6)
3 k P [S
The quantity 22. 4 T/k P is the theoretical absolute span setting
of the instrument. In practice, calibration against the ozone
standard is required for maximum accuracy, both because the length
of the light path is altered by possible multiple reflections off the inner
wall, and because the pressure, temperature and absorption inside the
15
instrument cannot be precisely determined. Hodgeson et al. compared
a Dasibi instrument spanned according to equation (4) with the absolute
photometer. They obtained the following relationship when repeated (61)
measurements were made on different days at ozone concentrations of
0. 05 to 2. 5 ppm:
O (phot) = (1.05± 0.01) O (Dasibi) + (0. 025± 0.013). (7)
3 3
6-27
-------
When the span setting of a Dasibi instrument was adjusted to
agree with an absolute laboratory photometer and repeated measurements
were made on different days, the folio-wing relationship with the absolute
photometer was obtained:
O (phot) = (0.999± 0. 003)O (Dasibi) - (0.007 ±0.002). (8)
3 3
The uncertainties refer to 9-'% confidence limits. These small error
limits are indicative of the high precision of the readings obtainable from
both instruments.
Measuring the absorption by ozone in air of infrared radiation at
a wavelength of 9480 nm. This method which is identical in principle to
measurement of the absorption in the ultraviolet requires an accurate
measurement of the extinction coefficient of ozone at a wavelength of 9480 nm.
Making this measurement in the infrared energy region is more complicated
because the determination of the extinction coefficient is more strongly
influenced by temperature, pressure and instrumental characteristics such
as spectral slit width. Also, the extinction coefficient is much smaller
than in the ultraviolet. This means that for measurements of concentrations
smaller than parts per million, the absorption path length in the infrared
infrared has to be about 100 times longer than the 1 meter ultraviolet cell
(see Figure 6-2). All these factors create problems which detract from
the practicality of this method as a primary reference. Nevertheless, at
laboratories experienced in this technique, accurate measurements
have recently been made that are in close agreement with the values
14
obtained by absolute ultraviolet photometry. Using the Dasibi ozone
meter as a transfer standard, the following agreement has been obtained:
6-28
-------
O (Dasibi) = (0. 99± 0. 02) O (IR) + (0. 016 ± 0. Oil). (9)
3 3
The uncertainties are within one standard deviation. Because ultraviolet
photometry carried out with the Dasibi instrument is a relatively
simple procedure, it may be advisable to calibrate even infrared
analyzers, by means of the ultraviolet transfer standard.
Determination of ozone in air by titration with nitric oxide.
This calibration technique is based on the application of the rapid gas-
phase reaction between nitric oxide (NO) and ozone to produce a
stoichiometric quantity of nitrogen dioxide (NO ) according to the
2
following equation:
NO + O ^ NO + O . (10)
3 22
The reaction rate is extremely fast so that even at concentrations
smaller than parts per million, the reaction is virtually complete after
a few seconds when there is an excess of nitric oxide present. Under
these conditions the amount of ozone added during the titration is
equivalent to the amount of nitric oxide consumed, and to the amount of
nitrogen dioxide formed. The accuracy of this calibration method is
critically dependent on the accurate measurements of the nitric oxide
concentration, and on the nitric oxide and ozone flow rates.
6-29
-------
Nitric oxide at about 50 ppm compressed with very pure nitrogen
in gas cylinders is provided for this purpose, and the true concentration
is established by comparison with that of a compressed gas tank which can
be obtained from the National Bureau of Standards, as a standard reference
material. The nitric oxide meter is calibrated repeatedly at several con-
centrations of nitric oxide and the mass flow meters are recalibrated
frequently with absolute bubble meters.
The titration is shown schematically in Figure 6-4.
Nitric oxide, compressed with nitrogen in a gas cylinder, is then
metered into the apparatus along with the clean air stream but without
the ozone. In this way, the initial nitric oxide concentration is measured
with the nitric oxide meter. When the initial nitric oxide concentration
is established, the lamp shield in the ozone generator is withdrawn
to various settings and the final nitric oxide concentrations are measured.
The initial nitric oxide concentration, the reaction residence times,
and the maximum ozone concentration added are established such that
the unreacted ozone is negligible (i. e. , <1%). Under these conditions,
the measured decreases in nitric oxide concentrations then determines
the ozone concentrations for various generator settings and serves
to calibrate the ozone source. The ozone source may then be used
to calibrate ozone meters.
6-30
-------
OZONE METER,
NO, N9 —
N02 METER
NO METER,
MIXING AND .
REACTION VESSELS
1 MASS 1
FLOW
[ METER |
-
1 MASS I
FLOW -« °3' AIR
| METER
0.05 to 0.25 1/min
4 to 5 1/min
Figure 6-4. Gas phase titration for determining ozone
13
6-31
-------
As an option, the nitrogen dioxide concentration produced can also
be monitored at the same time. In fact, this is one way to conveniently
calibrate nitrogen dioxide meters.
Using the Dasibi ultraviolet ozone analyzer as a secondary
13
reference standard DeMore and Patapoff tested the prediction that
the observed change in nitric oxide concentration (ANO) -would equal
the ozone concentration introduced and measured by UV absorption.
They found good agreement in -which
[O ] = (1.00 ±0.05) ANO - (0.00 + 0.01). (11)
3 UV
15
Similar conclusions were recently reported both by Hodgeson and by
16
Paur. These findings provide further validation of the ultraviolet
method for calibrating air-monitoring instruments and establish gas phase
titration as an alternative primary reference method.
PRINCIPLES OF MONITORING
Most of the atmospheric oxidant and ozone data, as -well as the
experimentally determined exposure data for vegetation, animals, and
humans, have been obtained using analyzers which sample, and
record the ambient concentrations almost continuously during the period
of observation. The response times are usually acceptable for fixed
station monitoring because data describing hourly averages are sufficient.
Faster responses are needed, however, for studying chemical reaction
rates, retention upon inhalation, sampling -while in motion, (as from
6-32
-------
aircraft) and for expediting calibrations. The response times required
are therefore a function of the resolution needed.
Definitions of Resolution and Response Times
Resolution: the ability to separate two closely occurring events
in space or time at a signal-to-noise ratio of two, expressed as percent
of full scale.
Signal-to-noise ratio: the ratio of the magnitude of the response
due to the pollutant concentration and of the magnitude of unwanted,
spontaneous, short-term responses not caused by variations in pollutant
concentration.
Response times: (see Figures 6-5 and 6-6).
Lag time (initial response time), t : the interval between the
1
time t , when a step change (increase or decrease) in pollutant concen-
o
tration is made, to the time t when the instrument indicates a response
i
equal to twice the noise.
t - t - t . (12)
1 i o
Time to 95%, t : the interval between the time t , when a step
95 o
increase in pollutant concentration is made, to the time t -when the
95
instrument indicates a response equal to 95% of the step increase.
t = t - t (13)
95 95 o
Similarly, t corresponds to the time to indicate 90%, and t the time
90 100
to indicate 100% of the step increase in pollutant concentration.
6-33
-------
Time to -95%, t : the interval between the time t when a step
-95 o
decrease in pollutant concentration is made to the time t when the
-95
instrument indicates -95% of the step decrease.
t = t - t . (14)
-95 -95% o
Similarly, t corresponds: to the time to indicate -90%, and t to the
-90 -100
time to indicate -100% of the step decrease in pollutant concentration.
Rise time, t : the interval between the time of a response equal
r
to 100% of the step increase in pollutant concentration (t ) and the lag
100
time (t ).
1
t = t - t . (15)
r 100 1
Fall time, t : the interval between the time the instrument indicates
f
-100% of the step decrease in pollutant concentration (t ) and the lag
100
time (t ).
1
t = t - t . (16)
f - 100 1
The fall time does not necessarily equal the rise time.
Pulse time: the minimum time a pollutant concentration must per-
sist for the analyzer to register a peak response equal to the pollutant
concentration (see Figure 6-5).
For any event to be ziccurately recorded, it must persist for the
pulse time of the instrument. This time is equal either to the rise time
or to the time to 100% response, depending on the design of the instrument.
For accurate data from aircraft sampling plumes, for example, it is
necessary to obtain rise times of a few seconds or less. This is a
6-34
-------
2X, noise
100 -
95 -
*
i
4>
co
c.
o
a.
vt
4)
c*:
10 -
0 -
Pollutant Pollutant
i in i
T Input "
r '
/
/
/
/
/
/
/
/
/
/
_>*^
^
I
i
I
1
i
i
i
i
I
i
1
j
|
\
\
\
\
\ Output
\
\
\
\
\
\
i\^
j t • t' « j i"
-•n t»l 1
He J
•*..
* J
T
time to 100%_
l<4 tfc
1^3 C»
__LJ
f-fl , , . ^V-l
^ >j
1
J
time to 95% — 1
rise
Ian
fall
time to -95%
time to -100%
lag
Figure 6-5. Visual representation and interpretation of time delays in
-i 38
analyzer response.
6-34a
-------
c
o
•-H
•P
c
OJ
o
c.
o
Pulse time
Output
T i me
Figure 6-6. Diagram of pulse time.
28
6-35
-------
very fast response for an analyzer and has only quite recently become
possible for ozone measurements.
The analytical principles which have been applied to accumulate
air quality data are: colorimetry, amperometry, chemiluminescence and
xiltraviolet absorption. Colorimetric and amperometric continuous
analyzers employing wet chemical techniques (reagent solutions), have
been in use as ambient air monitors for many years. Chemiluminescent
analyzers, which measure the amount of chemiluminescence produced
when ozone reacts with a gets or solid, were developed to provide a
specific and sensitive analysis for ozone, and have also been field-tested.
Ultraviolet absorption analyzers are based on a physical detection principle,
the absorption of ultraviolet radiation by a substance. They do not use
chemical reagents, gases or solids in their operation and have only
recently been field-tested. Ultraviolet absorption analyzers are ideal as
transfer standards but as discussed earlier they have limitations as air
monitors because aerosols, mercury vapor, and some hydrocarbons could
interfere with the accuracy of ozone measurements made in polluted air.
Advanced electro-optical methods (e.g., laser resonance absorption),
capable of measuring avera.ge concentrations over long distances still
require extensive research and field testing to demonstrate their practical
application to ozone monitoring. Since electro-optical methods have not
to date been widely used, they will not be discussed further here.
6-36
-------
Colorimetric Analyzers
Colorimetric analyzers spectrophotometrically measure the increase
in color (absorbance) of a solution resulting from contact with a measured
volume of air. The absorbance is linearly proportional to the concentration
of the colored species, within known limits. Continuous colorimetric
analysis of total oxidants is carried out using a solution of neutral-buffered
6
potassium iodide (KI). In 1953, Littman and Benoliel developed the
first colorimetric oxidant recorder to come into general use. Instruments
of this design, using a 20% neutral-buffered potassium iodide solution,
later changed to 10%, were incorporated into the Los Angeles County
Air Pollution Control District (LAAPCD) air monitoring network in the
1
early 1950's.
The continuous colorimetric instrumental method is described in
39
detail by the Intersociety Committee and the American Society for
40
Testing and Materials, and colorimetric and amperometric analyzers
41
are discussed extensively in the papers both of Tokiwa et aL and of
2
Hodgeson.
A typical colorimetric analyzer is illustrated schematically in
41
Figure 6-7. Sample air is drawn at a metered rate into a contact
column where the air is scrubbed with a metered flow of potassium
iodide buffered at pH 6. 8. The reaction of oxidants with the potassium
iodide solution produces the yellow colored triiodide ion (I ). The
3
colored solution flows to a colorimeter cell where the absorbance of
the triiodide ion is measured at 354 nm. The photometer signal is then
electronically recorded as parts per million of oxidant. The reaction of
potassium iodide with ozone at pH 6. 8 ± 0. 2 is given by equation 17.
6-37
-------
I
J»- ^P-
CARBON •
COLUKN
REAGENT
PTIMP
IA
f
\1
n
AIR
METERING
VALVE
—O
O-(^>-
AIR
OUT
PUMP
CONTACT
COLUKN
J
I
PHOTO[METER
i
AIR
FLOWMETER
_ ^. — AIR SAMPLE IN
SCRUBBER
FOR S0£
rid
REAGENT STORAGE
(1Q% KI)
RECORDER
Figure 6-7. Colorimetric oxidant analyzer.41
6-38
-------
O + 3KI + H O + K I + 2KOH + O . (17)
3 23 2
Colorimetric oxidant analyzers operate most reliably at atmospheric
3
oxidant levels ranging from 20-2000 y g/m (0. 01 - 1. 0 ppm) of ozone or
other equivalent oxidant. A usable response can be obtained from oxidant
levels equivalent to as much as 100 ppm ozone, but at such high levels the
stability and speed of response are inferior to those attainable at lower
39
concentrations.
When sulfur dioxide (SO ) is present in the polluted air, it causes
2 40
a negative interference equal to 100% of an equimolar concentration of oxidant.
The response to the pollutant, nitrogen dioxide (NO ), varies with the reagent
2
formulation and scrubber design. For 10% potassium iodide, nitrogen
dioxide produces a positive interference of approximately 21%; for 20%
41
potassium iodide, the interference produced is approximately 30%.
Filters and scrubbers have been used for removing sulfur dioxide,
but detailed information about their performance is not available. A brief
review of performance characteristics of sulfur dioxide scrubbers is given
2
by Hodgeson. One device consists of glass fiber strips impregnated with
42
chromium trioxide and sulfuric acid. An even better scrubber consists
of a bed of small pellets of porous chromatography grade firebrick or
38
alumina impregnated with chromium trioxide. These scrubbers, that
convert only a portion of the ambient nitric oxide to nitrogen dioxide,
require an additional correction factor. Despite these drawbacks, this is
the most generally used technique for removing sulfur dioxide interference,
particularly during periods of high sulfur dioxide concentrations. When
sulfur dioxide levels are less than 20% of the nitric oxide concentration,
6-39
-------
39
this filtering system is not recommended. In this case, the correction
for nitrogen dioxide interference in the oxidant reading can be determined
by measuring the instrument response to a known stream of nitrogen
dioxide while concurrently analyzing for nitrogen dioxide in the atmosphere;
40
then the appropriate subtraction can be made.
Amperometric Analyzers
Amperometric analyzers are often referred to as "coulometric"
analyzers. Coulometry is a mode of analysis in which the quantity of
electrons (charge) necessary to oxidize or reduce a desired substance is
measured. Because it is the current and not the charge -which is measured
by these instruments, "coulometric" analyzers are more properly called
amperometric analyzers. This principle has been implemented for oxidant
47
monitoring using either a galvanic (Hersch) cell or an electrolytic
44-46
(Brewer or Schulze) cell. A schematic illustration of this instrument
using a Brewer cell is shown in Figure 6-8. (Instruments using Hersch
or Schulze cells are not widely used. )
The operational principles of the amperometric analyzer are
described in the Intersociety Committee's manual of methods for ambient
43
air sampling and analysis. The contactor-sensor consists of a rod wound
with many turns of a fine platinum wire (cathode) and two turns of a heavier
wire (anode) axially mounted in a plastic block with a hole about 0.6 cm
inside diameter and 5 cm long. The reagent is usually a mixture of
2% potassium iodide and 5% potassium bromide buffered at pH 6. 8 0. 01 M
with 0. 026 M disodium hydrogen phosphate and 0. 018 M sodium dihydrogen
41
phosphate. Sample air and reagent pass concurrently through the annular
6-40
-------
Solution pump
Liquid
waste
reservoir
Sensor stem (electrode support)
Sensor block
Air sample inlet
Contact annulus
—I'ul ti-turn wire Cathode
Double-turn
wire anode
Air exit
Air pur.p
Screws, stem support
D. C. voltage source 0.2'i-v
~i I < """
Recorder
mv.
Figure 6-8. Amperometric analyzer using Brewer type electrolytic cell.
6-41
-------
space (about 0. 2 cm) between the support rod and the cylinder wall. Ozone
in the air transferred to the reagent reacts with the iodide to produce
triiodide ions which in turn react with the hydrogen, polarizing the cathode.
As a result of these reactions), current flows to repolarize the cathode in
proportion to the amount of hydrogen removed. This current, which is
44,45
directly proportional to the ozone concentration, is recorded.
The presence of nitrogen dioxide in the sample causes a positive
interference. When the concentration of the potassium iodide reagent is
2%, then the interference is about 6% of the nitrogen dioxide concentration.
This is typical for amperometric analyzers -which have been marketed
41
specifically for ozone or oxidant measurements. The exact level of the
detector's specific response to nitrogen dioxide depends on the instrument
design, the composition of the sensing solution, the operating conditions,
43
and other unknown factors. Therefore, the specific response for each
detector should be determined experimentally. The oxidant measurements
can be corrected for nitrogen dioxide interference by subtracting the response
due to the nitrogen dioxide from the total detector response.
Sulfur dioxide in the sample causes a negative interference of
approximately 1 mole of ozone per mole of sulfur dioxide because it reduces
the iodine formed by ozone back to potassium iodide. When sulfur dioxide
levels do not exceed those of the oxidants, a method commonly used to
correct for its interference is to add the amount of sulfur dioxide determined
by an independent method to the total detector response. A second method
38,46,48
is to remove the sulfur dioxide from the sample stream by using solid
47
or liquid chromium trioxide scrubbers. Because the data on the performance
6-42
-------
of these sulfur dioxide scrubbers is inadequate, the performance for each
oxidant system must be established experimentally.
Because of the interference problems -with both colorimetric and
amperometric analyzers, they are being replaced by instruments based
on other principles. (Colorimetric analyzers are no longer commercially
available. ) One such recently improved technique developed for the specific
detection and measurement of ozone is by detecting the absorption of ultra-
violet radiation by the ozone molecule. (This method is described in the
Calibration section. ) Another, is by measuring the chemiluminescence
produced when ozone reacts with a specific gas or solid.
Chemiluminescence
In 1965, a gas-phase chemiluminescent reaction between ozone and
49
ethylene was reported by Nederbragt e^t al_. , and subsequently the sensitivity
50
of this technique was improved by Warren and Babcock. The reaction
between ozone and ethylene yields chemiluminescent emission in the 300 to
51
600 nm region with maximum intensity at 435 nm. The intensity of this
emission is directly proportional to the ozone concentration.
A diagram of a typical gas-phase (ozone-ethylene) chemiluminescent
54
ozone analyzer is shown in Figure 6-9. The detector responds linearly to
ozone concentrations between 0. 003 and 30 ppm; no interferences were
51
initally observed. More recently, however, it has been established that
when going from 0 to 60% relative humidity in the temperature range of 20
to 25 C, water vapor produces a small positive signal which results in
about an 8% increase in the ozone concentration measurement. This potential
6-43
-------
.CHEMICAL ' OZONE
FILTER GENERATOR
AIR
SAMPLE —fcJ
TETLCH
FILTER
ETJiYLENE
IN
PRESSURE
REGULATOR
REACT 10?;
I
FLOV/J.-:ETSR
PKOTOMULTIPLIEH
. FLOY/
CONTROL
VALVE
Figure 6-9. Gas-phase (ozorie-ethylene) chemlluminescent ozone monitor,
block diagram.'
6-44
-------
source of error can be minimized by using humidified rather than dry ozone in
air streams when calibrating.
Slight improvements in sensitivity can be achieved by cooling the
phototubes used to detect the emitted light or by increasing the ethylene flow
rate. Chemiluminescence produced by the reaction of ozone with ethylene
has been designated by the EPA as the reference method for the monitoring
25, 57
ozone. Several different commercially produced instruments are
available.
OTHER OXIDANTS AND PRECURSORS
Besides ozone, the main indicator of photochemical pollution, other
important concomitant products are: peroxyacetyl nitrate (PAN), hydrogen
peroxide, nitrogen dioxide, hydroxyl radicals and various aldehydes which
are both products and primary pollutants; and the particulates; sulfates,
nitrates, ammonium, chloride, water, and various types of oxygenated
organic compounds. The most important precursors of photochemical
pollution are nitric oxide and hydrocarbons. The measurement procedures
for the hydrocarbons are not as highly developed as those for ozone and the
nitrogen oxides.
Many deleterious effects have been associated with photochemically
polluted air: ozone is definitely associated with respiratory problems, plant
damage and material damage; PAN has definitely been associated with plant
damage and some other members of this class of chemical compounds have
been associated with eye irritation; the hydroxyl radical is considered to be
an important factor in the conversion of gas phase intermediates to end
6-45
-------
products, such as sulfur dioxide to particulate sulfate; the particulate
complex is responsible for haze formation and has also been associated
with eye irritation and respira,tory effects. The aldehydes have been
associated with eye irritation. Ozone and PAN itself do not cause eye
irritation. For purposes of control, much more research is needed
in order to relate the laboratory data about the levels of these various
materials which have significant effects to their formation in the atmosphere
from emissions and their atmospheric distribution. The lack of convenient
measurement methods has hindered progress in gaining this understanding.
The chief precursors for both oxidant and suspended particulate
matter formation in the atmosphere, which are directly emitted into the
atmosphere, are nitrogen oxides, hydrocarbons and their derivatives,
ammonia, and sulfur dioxide. The measurement of particulate components
is discussed in Chapter 2. This section describes briefly the measurement
of nitrogen oxides, hydrocarbons, free radicals and other precursors.
Nitrogen Oxides
The technology for the routine measurement of the nitrogen oxides
(nitrogen dioxide, NO , and nitric oxide, NO) is fairly well advanced.
2
The Environmental Protection Agency is on the verge of officially proposing
that chemiluminescence produced by the reaction of nitric oxide with ozone
57
should be the reference method for nitrogen dioxide. This method is even
more suitable for nitric oxide. Since no national air quality standard has
been promulgated for nitric oxide, no reference method will be specified.
However, its measurement in the atmosphere is crucial for establishing the
relationship of its emission to the formation of atmospheric ozone and other
photochemical oxidants.
6-46
-------
Most of the data for nitrogen oxides prior to 1970 have been obtained
by continuous measurements with a colorimetric analyzer -which was similar
41
in principle to the colorimetric oxidant analyzer shown in Figure 6-7.
The scrubbing agent is a mixture of N[-(l -Naphthyl) ethylenediamine, sulfanilic
acid and acetic acid in aqueous solution. The color is produced when both
nitrogen dioxide and nitrites react with this reagent to form an azo dye.
The color is not affected by any nitric oxide present in the air sample.
The quantification of the nitrogen dioxide depends on the acceptance
of an empirical factor which relates the response of this reagent such that
one mole of NO scrubbed from the gas phase produces the same amount of
2
color as 0. 72 mole of nitrite added in solution. Upon alterations of reagent
composition and scrubber design this factor may change. For most of the
air monitoring activities in the United States using this reagent this factor
57a
appears to have been verified. Evaluations remain to be conducted using
gas phase titration with a known ozone source.
To determine nitric oxide with this method, it is oxidized to nitrogen
dioxide by passing the air sample through a reaction vessel containing either
38
potassium permanganate or chromium trioxide. For an accurate measure-
ment, however, any nitrogen dioxide originally present must first be selectively
58
removed by passing the air sample through a triethanolamine scrubber.
This precaution has, unfortunately, not been the practice. Rather, the air
sample containing both nitric oxide and nitrogen dioxide has been passed
first through the oxidizer and then to the reagent scrubber, which gives a
measurement of total nitrogen oxides (NO ). The assumption is made that
x
no nitrogen dioxide is lost.
In another technique, the air stream exiting from the nitrogen dioxide
reagent scrubber is taken to the oxidizer and then to a second reagent scrubber.
This gives separate measurements for nitrogen dioxide and nitric oxide. Here
the assumption is that all the nitrogen dioxide is retained in the first reagent
6-47
-------
scrubber and no nitric oxide is lost. Neither of the assumptions made for
these methods of measuring nitric oxide is completely correct. For this
reason, there is a great deal of uncertainty about the amount of error in
existing data banks. The nitrogen dioxide data, on the other hand, is believed
to be reasonably accurate particularly where there has been proper instrument
38
calibration and maintenance.
The diurnal patterns of ozone, nitric oxide and nitrogen dioxide con-
centrations observed during photochemical oxidant episodes in California,
have been confirmed by smog chamber studies. There may be, however,
a decrease in reliability with decreasing concentration of values less than
0. 1 ppm that were measured by the colorimetric method. The magnitude
of these uncertainties among the various monitoring networks in the
United States has still to be assessed.
In the chemiluminescent detection of nitrogen oxides, a constant
source of ozone is reacted with a metered air sample containing nitric oxide.
52
Fontijn ^t aJ_. suggested that this method could also be used for ozone
detection by using a constant nitric oxide source for reaction with ozone in
the air sample. The ozone-nitric oxide reaction is carried out at reduced
pressure to avoid quenching the chemilumine scent reaction. Detection of
the emissions in the spectral region involved (600-3000 nm) requires using
a near-infrared sensitive photomultiplier tube. The noise of such a photo-
56
multiplier tube is reduced by cooling it to about -20 C.
To measure nitrogen dioxide with this technique it is thermochemically
converted to nitric oxide by reaction with molybdenum at about 200 C. The
extent of possible interferences at various monitoring sites from nitrogen
compounds other than ammonia, which does not interfere unless the tem-
perature is considerably higher than 200 C, remains to be assessed. The
instrumentation of this procedure is inherently more reliable than the original
6-48
-------
colorimetric analyzers. Unfortunately, the mutual equivalency under monitoring
situations of data obtained from these two techniques has not yet been evaluated.
This is particularly important for the data from California, where nitrogen
oxide measurements have been made over the past 20 years using the
colorimetric method.
At present, two primary calibration standards are available for
nitrogen oxides. One is a nitrogen dioxide generator in which the source is
a permeation tube, certified by the National Bureau of Standards, in a tem-
perature- and flow-controlled gas diluter. The amount of nitrogen dioxide
generated is determined by the loss in weight of the permeation tube.
The other standard is a cylinder of compressed nitric oxide in
oxygen-free nitrogen certified in the range of 50 ppm by the National Bureau
of Standards. To obtain concentrations smaller than one part per million, this
cylinder has to be connected to a dilution apparatus carefully regulated for
flow and temperature.
The calibration of a chemiluminescent analyzer using these standards
is verified when both the nitrogen dioxide and nitric oxide channels respond
similarly to charges from the cylinder of diluted nitric oxide. Agreement
with the response of the nitric oxide channel should also be obtained when
titrating the standard nitric oxide mixture with a previously established
source of ozone. While chemilumine scent instruments have simplified
monitoring of nitrogen oxides, achieving accurate calibration requires -well-
trained personnel.
Hydrocarbons
For the measurement of the hydrocarbon precursors of photochemical
oxidants, the naturally occurring methane must be separated from the other
so-called non-methane hydrocarbons. Only one procedure, gas chromatography
6-49
-------
coupled with flame ionization detection, is available for this separation
and measurement. Although instrumentation for routinely accomplishing this
process is commercially available, its maintenance (continued operation)
requires a degree of operational know-how which may be too costly
for most public agencies in the United States to support. Consequently,
there is at present insufficient data to relate the occurrence of photochemical
oxidants and ozone accurately to some of their most important precursors,
the non-methane hydrocarbons.
In addition, it is widely recognized, based on chamber studies,
that some of the non-methane hydrocarbon compounds are far more
important than others in causing the formation of photochemical oxidants
and aerosols. These are referred to as the reactive hydrocarbons. To date
it has only been possible to measure them by means of a more sophisticated
gas-chromatographic process than described above. Therefore, their measure-
ment in the atmosphere has been limited to research or to short-term episodic
types of studies. While these are useful for understanding the phenomena of
photochemical pollution forma.tion, they have been too costly for determining the
effectiveness of hydrocarbon control programs with respect to the changes in
the occurrence and concentration of reactive hydrocarbons in air. Consequently,
there is very little data about those hydrocarbon compounds in the atmosphere
which may be precursors for the formation of atmospheric haze.
Particulate Sulfates and Nitrates
The sulfate and nitrate content of atmospheric particles comes
primarily from the conversion of sulfur dioxide and nitrogen dioxide. Photo-
chemically initiated atmospheric reactions and transient free radicals are
6-50
-------
often associated with this process (Chapter 2). At over 3500 locations in the
United States, the total suspended particulate matter is regularly sampled
57
at least once per week for a 24 h period. The particles are usually
collected on glass-fiber filter mats which are then sent to a central laboratory.
There the sulfate and nitrate contents are extracted with hot water and deter-
mined by standardized chemical analytical procedures. Recently, there has
been a detailed review of these procedures covering their reliability, their
limitations, and the alternatives available for their improvement (emphasizing
59
sulfates).
There are several potential sources of error in these methods.
The filters routinely used have a relatively high and somewhat variable sulfate
3
content, so that at concentrations lower than 10 yg/m and sampling periods
less than 24 h, the reliability of the sulfate measurement is reduced. Several
different types of filtering media adsorb sulfur dioxide during the first few
hours of sampling, which alters the amount of sulfate observed. This inter-
ference can become critical when sampling periods are less than 24 h and
the ratio of the concentrations of sulfur dioxide to sulfate is greater than 5 to 1.
Interference can also be introduced by the hot water extraction, -when reduced
sulfur compounds such as sulfite are present, because they are oxidized to
sulfates in this process. Another possible error source is that some of the
various analytical procedures used for sulfate determination may be influenced
by other substances also present in the particulate matter.
Current developments for minimizing these limitations in the ana-
lytical procedures are: selecting filtering media for sampling -which have low
blanks and do not adsorb significant amounts of sulfur dioxide and nitrogen
oxides; and designing analytical methods which are specific for sulfates and
nitrates in the presence of other particulate substances. In addition,
6-51
-------
more sensitive procedures for studying short-term (hourly) variations are
being developed. These procedures -will permit the investigation of how the
forination and occurrence of these particulate components relate to atmospheric
reactions and to the acute effects both when inhaled and -when deposited on
surfaces.
Other Substances
As shown in Table 6-1 (and reviewed in Chapter 2), there are other
substances monitored occasionally, which are important in the photochemical
oxidant milieu. These are the peroxy alkyl and aryl nitrates (PAN's) and
the aldehydes.
The PAN's have been monitored by directly injecting the air at the
sampling site into a specifically designed gas chromatograph. Aldehydes
(formaldehyde is the major one in the atmosphere) are generally sampled in
liquid absorbers using reagents which develop a color. These are sensed
continuously with an instrument comparable in design to the colorimetric
oxidant analyzer (Figure 6-7). To obtain information about specific aldehydes,
the material collected in the absorber is sent to a central laboratory where
individual aldehydes are analyzed by gas chromatography and mass spectroscopy.
There are a number of significant oxygenated organic particulate
compounds and gas-phase free radicals formed by the reactions of gas-phase
hydrocarbons (see Table 6-1 cind Chapter 2). The measurement methods for
these substances are complicated and in the research stage. Their description
is beyond the scope of this chapter. It is of major importance to develop
methods for measuring hydroxyl and peroxyhydroxyl radicals as well as the
various oxygen species formed together with ozone (see Chapter 12).
6-52
-------
EVALUATION
In the final analysis, the purpose of measurement is to provide
the
numerical values as/basis both for making policy decisions and for enforcing
regulations. It is critical therefore to know whether measurement data is
reliable. It is also essential that all the data be intercomparable. This
includes data obtained from laboratory studies of chemical reactions, plant
and material damage, and animal and human toxicology; from field studies
of air quality, vegetation and ecosystem effects, and population exposures.
In all such studies (see Chapters 2, 4, 8, 9, H, 12, and 13), irrespective
of the measurement method used, the measurement of oxidants is based on
a "standardized" source of ozone.
Unfortunately, and probably unavoidably, from the earliest research
until the present, investigators have not-used the same standardization process.
Furthermore, standardization practices within different research groups have
only rarely followed the meticulous series of in-house calibrations, verifi-
cations and inter-laboratory comparisons long prescribed by the community
of measurement specialists. There are several quantitatively crucial studies,
however, -which have followed sound measurement procedures (e. g. , Hackney,
Chapter 9).
The chief objective of this section is to provide a perspective con-
cerning the reliability of atmospheric oxidant data. The expected performance
of atmospheric oxidant monitors is given in Table 6-7. To judge the
reliability of measurements, information about the following five factors
is required:
6-53
-------
(1) Accuracy; agreement with respect to a primary reference
standard.
(2) Reproducibility; the precision with which measurements can be
repeated both within a single measurement group and among different groups
in different laboratories.
(3) Interferences; substances that exist simultaneously with ozone
and alter the response of the measurement method.
(4) Comparisons among different measurement methods.
(5) Practices and maintenance of measurement operations.
The fundamental factors which govern the accuracy of primary
reference standards were discussed in the section of this chapter on
calibration. Even though an improved reference standard has been
advocated, most of the existing air monitoring and laboratory exposure
data have as their reference the potassium iodide procedure used by
either the California Air Resources Board (ARB), or the Los Angeles
County Air Pollution Control District (LAAPCD), or the EPA. The
relationship of these three variations of the potassium iodide procedure
to the ultraviolet method are as follows:
O (ARB) = 1.29O (UV) -0.005 (18)
3 3
O (LAAPCD) = 0.96O (UV) - 0.032 (19)
3 3
O (EPA) = 1.110 (UV) - 0.035. (20)
3 3
6-54
-------
Table 6-7
Expected Performance Specifications for Oxidant Monitors
60
Performance Characteristic
Range
Noise
Lower detectable limit
Interference equivalent
Each interferent
Total interferent
Zero drift, 12 and 24 h
Span drift, 24 h
Lag time
Rise time
Fall time
Precision, standard deviation range
as % of 0.08 air quality standard ppm
Accuracy with respect to primary
reference standard
Specification
0 to 005 ppm
0.005 ppm
0.01 ppm
±0.02 ppm
0.06 ppm
±0.02 ppm
±0.025 ppm
20 min
15 min
15 min
0.01 ppm
12.5%
not specified
6-55
-------
While the differences between these several primary reference
procedures are of some concern, the practice and performance of these pro-
cedures since 1952 has been remarkably consistent. For instance, several
studies by different investigators and laboratories have shown an average
ratio of 1. 37, between the LAAPCD and ARB methods (with extremes in
12
this value not exceeding 5%). The EPA primary reference procedure
using potassium iodide has also been evaluated recently by several laboratories.
When the concentration range of ozone was 0. 005 to 0. 5 ppm, the standard
deviation among the laboratories during a 4-day continuous measurement of
the same atmosphere was ± 6 to 10%.
With respect to stablized ozone generators supplied and calibrated by
the National Bureau of Standards, the participating laboratories obtained
values that were about 15% lower on the average, at an ozone concentration
61
of about 0. 2 ppm. Therefore, while acceptable repeatability within and
among laboratories can be achieved even with the potassium iodide bubbler
method as a reference procedure, there are unpredictable variables
inherent in this procedure which contraindicate its continued acceptance.
Differences in measurement methods include analyzer systems based
both on the same and on different measurement principles. The average
standard deviation in the performance of different chemiluminescent ozone
instruments -which are sampling the same ambient air both with and -without
an added ozone concentration of 0. 002 to 0. 5 ppm is 6 to 10%. Field studies
comparing an ultraviolet monitor with several chemilumine scent monitors
showed correlation coefficients for hourly averages of from 0. 80 to 0. 95
6-56
-------
between various pairs of instruments. Hourly averages for about 500 pairs
of values at ambient ozone concentrations of from 0. 005 to 0. 100 ppm showed
deviations of from 3 to 23% between the average values for paired instruments.
Typical colorimetric and amperometric analyzers have been compared
both in the laboratory, using ozone concentrations of 0 to 0. 6 ppm (with and
without added NO at a similar concentration range), and at field locations
2 41
where the oxidant concentration was 0 to 0. 2 ppm. When both instruments
were corrected for NO interference, the field results showed highly correlated
2
(r = 0.96) hourly averages. Nevertheless, the colorimetric readings were
consistently 6% higher than the amperometric readings. The responses
showed the following relationship:
O (amp) = 0. 942 O (col) - 0. 0038 ppm. (21)
3 3
To summarize, the results of comparing and evaluating air oxidant
analyzers indicate that when the use of similar primary reference procedures
is coupled with meticulous operational practices, agreement within about
20 to 30% can be expected. A monitoring network operated by a single
tightly managed group, however, can achieve even better agreement.
SUMMARY. CONCLUSIONS AND RECOMMENDATIONS
With the exception of calibration, the measurement problems which
were apparent in 1970, at the time of publication of the first air quality criteria
document on photochemical oxidants, have essentially been solved for ozone.
This remarkable achievement is the result of unstinting efforts by individuals
working at EPA's National Environmental Research Center, North Carolina;
the National Bureau of Standards; private research contractors sponsored
6-57
-------
primarily by EPA; private instrument manufacturers; the Jet Propulsion
Laboratory of the California Institute of Technology; the Air and Industrial
Hygiene Laboratory, California Department of Health; the Air Pollution
Research Center of the University of California at Riverside, and the California
Air Resources Board (GARB).
Due to the focus on the problem brought about by the CASE , a
significant advance in the accurate calibration of instruments for monitoring
ozone in ambient air was achieved during 1975. As a result, this agency
adopted the measurement of ozone in the ultraviolet region at 254 nm as a
primary calibration reference standard. They have also adopted state-wide,
as a transfer standard for calibrating ozone and oxidant monitoring instruments
at air monitoring stations, a commercially available instrument (coupled with
the precise controlled generation of ozone in air), which measures the dif-
ferential absorption of ultraviolet radiation.
It is important to separate conceptually, and in practice, the
calibration process from the monitoring process. Photochemical oxidants
consisting primarily of ozone were first continuously measured in Southern
California by measuring the color change of potassium iodide solutions brought
into contact with the ambient air. This measurement continues to yield valid
photochemical oxidant data in California. However, it has yielded questionable
data at ambient air monitoring sites elsewhere in the United States. For this
reason at the end of 1971, EPA officially adopted a continuous monitoring
process that measures the chemiluminescence produced when ozone in air is
brought into contact with the gas ethylene. This reference procedure when
calibrated with the primary reference procedure using ultraviolet absorption
is widely accepted.
6-58
-------
Instruments based on differential ultraviolet absorption still
need to be evaluated, and possibly modified, prior to their acceptance
for monitoring ozone in polluted atmospheres on a nationwide scale.
The California Air Resources Board and other air pollution control
agencies are currently conducting multi-year programs for evaluating
ultraviolet absorption side-by-side with chemiluminescent and potassium
Iodide based instruments, to determine their applicability and
needed modification, as well as to assure continuity in the data base
while the older monitoring instruments are being replaced.
Thus despite the remarkable progress in the monitoring for ozone,
nitrogen oxides and non-methane hydrocarbons, which has strengthened the
implementation and evaluation of control programs, substantial research
and development is still required to help resolve the uncertainties in our
knowledge which are inhibiting the actual achievement of desired air quality
standards.
The areas in which further research and development are needed,
in sequence of priority are:
• Evaluation of primary calibration procedures applicable
nationwide for ozone measurement.
of
• Development/principles and instruments which can easily track
the sources of those hydrocarbons reactive in the production of ozone and
those which are reactive in the production of particles.
• .Chemical identification of both gas and particle phase compounds
occurring in the atmosphere, which cause eye irritation and respiratory
difficulties.
6-59
-------
• Methods for the direct and continual measurement of those
chemicals in the particles of the atmospheric haze, that are known to be
formed during photochemical pollution episodes and are already suspect
as respiratory irritants. By implementing such measurements it -will be
possible to find out to what extent the occurrence of such substances can be
reduced by various emission controls. To assess actual population
exposures, it is also necessary that these measurement methods be
easily carried out indoors and in vehicles.
• Improved measurement methods suitable for observations from
airborne platforms so that the regional scale impacts of urban emissions
can be accurately assessed. This is needed because some control options
for solving the urban-scale problem have the potential of transferring
pollution from one geographic area to other areas.
6-60
-------
REFERENCES
1. U. S. Department of Health, Education, and Welfare. Public Health Service.
Environmental Health Service. National Air Pollution Control Adminis-
tration. Air Quality Criteria for Photochemical Oxidants. NAPCA Publ.
AP-63. Washington, D. C.: U. S. Government Printing Office, 1970.
2. Hodgeson, J. A. Review of analytical methods for atmospheric oxidants meas-
urements. Int. J. Environ. Anal. Chem. 2:113-132, 1972.
3- schuetzle, D., D. Cronn, and A. I. Crittenden. Molecular composition of
secondary aerosol and its possible origin. Environ. Sci. Technol. 9:
838-845, 1975.
3a. Altshuller, A. P. Evaluation of oxidant results at CAMP sites in the United
States. J. Air Pollut. Control Assoc. 25:19-24, 1975.
4. Bradley, C. E., and A. Jk Haagen-Smit. The application of rubber in the
quantitative determination of ozone. Rubber Chem. Technol. 24:750-755,
1951.
5. Smith, R. G., and P. Diamond. The microdetermination of ozone. Preliminary
studies. Amer. Ind. Hyg. Assoc. Quart. 13:235-238, 1952.
6. Littman, FTE., and R.~ Wf Benoliel. Continuous oxidant recorder. Anal. Chem.
25:1480-1483, 1953.
7. Cherniack, I,, and R. J. Bryan. A comparison study of various types of ozone
and oxidant detectors which are used for atmospheric air sampling. J.
Air Pollut. Control Assoc. 15:351-354, 1965.
8. Graedel, T. D. Private communication, Bell Telephone Labs, Murray Hills,
N. J., 1975.
6-61
-------
9. Lewis, N. Private communication, N. J. State Bureau of Environmental Control,
Trenton, N. J. , 1975.,
10. Bruntz, S. M., W.~ S." Cleveland, T. E. Craedel, B. Kleiner, and J." L." Warner.
Ozone concentrations in New Jersey and New York: Statistical association
with related variables. Science 186:257-259, 1974.
11. Environmental Instrumentation Group. Photochemical oxidant monitoring instru-
mentation, pp. .In Instrumentation on Environmental Monitoring--
Air. Berkeley: Lawrence Berkeley Laboratory, University of California,
1973.
12. DeMore, W. B. , J. C. Romanovsky, M. Feldstein, W. J. Hamming, and P. K.
Mueller. Interagency comparison of iodometric methods for ozone deter-
mination, pp. . In Proceedings of American Society for Testing
and Materials Symposium on Calibration Problems and Techniques, Boulder,
University of Colorado, Aug. 5-7, 1975. (in press) (UNVERIFIED)
13. DeMore, W. B., and M. Patapoff. Comparison of ozone determinations by ultra-
violet photometry and gas phase titration. Environ. Sci. Technol. (in
press)
14. Pitts, J. N., J. M. McAfee, W. D. Long, and A. M. Winer. Longpath infrared
spectroscopic investigation at ambient concentrations of the 2% neutral
buffered potassium iodine method for determination of ozone. Environ.
Sci. Technol. (in press)
15. Hodgeson, J. A., C. L. Bennett, H. C. Kelly, and B. A. Mitchell. Ozone
measurements by iodometry, ultraviolet photometry and gas-phase titra-
tion, pp. . In Proceedings of American Society for Testing and
Materials Symposium on Calibration Problems and Techniques, Boulder,
University of Colorado, Aug. 5-7, 1975. (in press)
6-62
-------
16. Paur, R. Comparison of uv photometry and gas phase titration as candidate
methods for absolute calibration of ozone generator output in the sub-
part-per million range, pp. . In Proceedings of American Society
for Testing and Materials Symposium on Calibration Problems and
Techniques, Boulder, University of Colorado, Aug. 5-7. 975. (in
press)
17. Hanst, P. L., E.' R." Stephens, W." E." Scott, and R. C. Doerr. Absorptivities
the infrared determination of trace amounts of ozone. Anal. Chem. 33:
1113-1115, 1961.
lg Hodgeson, jf A., R.~ E.~ Baumgardner, B." E. Martin, and R. A. Rehme. Stoichio-
metry in the neutral iodometric procedure for ozone by gas-phase titration
with nitric oxide. Anal. Chem. 43:1123-1126, 1971.
18a. Rehme, K. A., B. E. Martin, and J. A. Hodgeson. Tentative Method for the
Calibration of Nitric Oxide, Nitrogen Dioxide and Ozone Analyzers by
Gas Phase Titration. EPA-R2-73-246. Research Triangle Park: U. S.
Environmental Protection Agency, Office of Research and Monitoring,
1974. 16 pp.
19. Hodgeson, J. A., R. K. Stevens, and B. E. Martin. A stable ozone source
applicable as a secondary standard for calibration of atmospheric moni-
tors, pp. 149-158. In J. Scales, Ed. Air Quality Instrumentation. Vol.
1. Selected Papers from International Symposia Presented by the Instru-
ment Society of America. Pittsburgh: Instrument Society of America, 1972.
20. Bendix Corp., Process Instruments Div., P. 0. Drawer 477, Ronceverte, W. Va.
24970
21. Dasibi Corporation, 161 East Colorado St., Glendale, CA, 91205.
6-63
-------
22. McMillan Electronics Corporation, 7327 Ashcroft, Houston, TX 77036.
23. Monitor Labs, Inc., 4202 Sorrento Valley Blvd., San Diego, CA, 92121.
24. Ultra-Violet Products, Inc., 5114 Walnut Grove Ave., San Gabriel, CA, 91778.
25. U. S. Environmental Protection Agency. Title 40. Protection of environment.
Part 50. National primary and secondary ambient air quality standards.
Federal Register 36:22384-22397, 1971.
26. California Department of Health. Air and Industrial Hygiene Laboratory.
Calibration and Standardization of Continuous Photometric Analyzers
of Atmospheric Oxidants. Recommended Method No. 5B. Berkeley:
California Department of Health, 1970. 17 pp.
27. Los Angeles County Air Pollution Control District. Ozone Calibration Proce-
dure for Monitoring Instruments (Titrimetric Method). Los Angeles:
Air Pollution Control District, 1975. 15 pp.
28. Imada, M. Comparison of LAAPCD's Original and Modified Methods for Standar-
dizing 0.002 N Sodium Thiosulfate Solutions. Letter to W. B. Demore,
Feb. 7, 1975. (see Ref. 11.)
29. Becker, K. H. , U. Schurath, and H. Seitz. Ozone-olefin reactions in the gas
phase. 1. Rate constants and activation energies. Int. J. Chem. Kinet.
6:725-739, 1974.
30. DeMore, W. BT, and 0. Raper. Hartley band extinction coefficients of ozone iti
the gas phase and in liquid nitrogen, carbon monoxide, and argon. J. Phys.
Chem. 68:412-414, 1964.
31. Clyne, M. A." A., and J." A, Coxon. Kinetic studies of oxy-halogen radical
systems. Proc. Roy. Soc. A303:207-231, 1968.
6-64
-------
3la. Hampson, R. F., W. Braun, R. L. Brown, D. Garvin, J. T. Herron, R. E. Huie,
M. J. Kurylo, A. H. Laufer, J. D. McKinley, H. Okabe, M. D. Scheer, W.
Tsang, and D. H. Stedman. Survey of photochemical and rate data for
twenty-eight reactions of interest in atmospheric chemistry. J. Phys.
Chem. Ref. Data 2:267-311, 1973.
32. GriggS, M. Absorption coefficients of Ozone itt the ultraviolet and visible
regions. J. Chem. Phys. 49:857-859, 1968.
33. Heam, AT cT The absorption of ozone in the ultra-violet and visible regions of
the spectrum. Proc. Phys. Soc. (London) 78:932-940, 1961.
34. Inn, E. C. Y., and Y. Tanaka. Absorption coefficient of ozone in the ultra-
violet and visible regions. J. Optic. Soc. Amer. 43:870-873, 1953.
35. deVera, E. R., E. Jeung, and M. Imada. Equivalency Determination and Cali-
bration Procedure for a UV Absorption Ozone Monitor. AIHL Report No.
160. Berkeley: California State Department of Health, 1974. (UNVERIFIED)
36. Behl, B. A. Absolute Continuous Atmospheric Determination by Differential
U.V. Absorption. Presented at the 65th Annual Meeting of the Air
Pollution Control Association, June 1972. (UNVERIFIED)
37. Bowman, L., D. Horak, and F. Richard. A Continuous Ultraviolet absorption
ozone photometer. ISA AID 72430 (103-108), 1972 (UNVERIFIED)
i t f t - j * / r '
38. Mueller, P. K., Y. Tokiwa, E, R. deVera, W. J. Wehrmeister, T. Belsky, S.
Twiss, and M. Imada. A Guide for the Evaluation of Atmospheric Analyzers.
Air and Industrial Hygiene Laboratory Report No. 168. Berkeley:
California State Department of Public Health, 1973. (UNVERIFIED)
39. Intersociety Committee. Tentative method for the continuous analysis of
atmospheric oxidants (colorimetric), pp. 356-364. In Methods of Air
Sampling and Analysis. Washington, D. C.: American Public Health
Association, 1972.
6-65
-------
4°f Standard method of test for continuous analysis and automatic recording of
the oxidant content of the atmosphere. Designation: D 2011-65 (Reapproved
1972), pp. 924-928. In 1973 Annual Book of ASTM Standards. Part 23.
Water; Atmospheric Analysis. Philadelphia: American Society for Testing
and Materials, 1973.
41, Tokiwa, Y., S. Twiss, E. R. de Vera, and P. K. Mueller. Atmospheric ozone
determination by amperometry and colorimetry, pp. 109-130. In G. Mamantov,
and W. D. Shults, Eds. Determination of Air Quality. Proceedings of
the ACS Symposium, Los Angeles, 1971. New York: Plenum Press, 1972.
42. Saltzman, B. E., and A." P." Wartburg, Jr. Absorption tube for removal of inter-
fering sulfur dioxide in analysis of atmospheric oxidant. Anal. Chem. 37:
779-782, 1965.
43. Intersociety Committee. Tentative method for continuous monitoring of atmos-
pheric oxidant with amperometric instruments, pp. 341-350. In Methods
of Air Sampling and Analysis. Washington, D. C.: American Public
Health Association, 1972.
44. Brewer, A. W.', and J. R. Milford. The Oxford-Kew ozone sonde. Proc. Roy. Soc.
A256:470-495, 1960.
45, Mast, G. M., and H. E. Saunders. Research and development of the instrumenta-
tion of ozone sensing. Instrum. Soc. Amer. Trans. 1:325-328, 1962.
46. Schulze, F. Versatile combination ozone and sulfur dioxide analyzer. Anal.
Chem. 38:748-752, 1966.
47. Hersch, P., and R.~ Deuringei'. Galvanic monitoring of ozone in air. Anal. Chem.
35:897-899, 1963.
48. Wartburg, A. F., A7 W." Brewer, and jfrT Lodge, Jr. Evaluation of a coulometrit
oxidant sensor. Air Water Pollut. 8:21-28, 1964.
6-66
-------
49. Nederbragt, G. W., A. Van Der Horst, and J." Van Duijn. Rapid ozone determina-
tion near an accelerator. Nature 206:87, 1965.
50. Warren, G. J., and G. Babcock. Portable ethylene chemiluminescence ozone
monitor. Rev. Sci. Instrum. 41:280-282, 1970.
51. Stevens, R. R., and J. A. Hodgeson. Applications of chemiluminescent reactions
to the measurement of air pollutants. Anal. Chem. 45;443A-446A, 449A,
1973.
52. Fontijn, A., A." J. Sabadell, and R. J. Ronco. Homogenous ehemilumineseent
measurement of nitric oxide with ozone. Anal. Chem. 42:575-579, 1970.
53. Altshuller, A. P. Analytical problems in air pollution control, pp. 245-286.
in W. W. Meinke and J. K. Taylor, Eds. Analytical Chemistry: Key to
Progress on National Problems. Proceedings of 24th Annual Summer S /mposium
on Analytical Chemistry, National Bureau of Standards, Gaithersburg,
Maryland, 1971. National Bureau of Standards Special Publication 351.
Washington, D. C.: U. S. Government Printing Office, 1972.
54. Coloff, S. G., M. Cooke, R. J. Drago, and S. F. Sleva. Ambient air monitoring
of gaseous pollutants. Amer. Lab. 5 (7):10-22, 1973.
55. California Air Resource Board. A Study of the Effect of Atmospheric Humidity
on Analytical Oxidant Measurement Methods. Sacramento: State of
California, Air Resources Board, Report of a Interagency Study, July
9, 1975. 14 pp.
56. Altshuller, A. P. Analytical problems in air pollution control, pp. 245-286.
In W. W. Meinke and J. K. Taylor, Eds. Analytical Chemistry: Key to
Progress on National Problems. Proceedings of 24th Annual Summer Symposium
on Analytical Chemistry, National Bureau of Standards, Gaithersburg,
Maryland, 1971. National Bureau of Standards Special Publication 351.
Washington, D. C.: U. S. Government Printing Office, 1972.
6-67
-------
57. Hoffman, A. J. , T. C. Cm-ran, T. B. McMullen, W. M. Cox, and W. F. Hunt.
EPA's role in ambient air monitoring. Science 190:243-248, 1975.
57a. Intersociety Committee. Tentative method of analysis for nitrogen
dioxide content of the atmosphere (Griess-Saltzrnan Reaction), pp.
329-336. In Methods of Air Sampling and Analysis. Washington,
D. C.: American Public Health Association, 1972.
58. Levaggi, D. A., W. Siu, M. Feldstein, and E. L. Kothny. Quantitative separa-
tion of nitric oxide from nitrogen dioxide at atmospheric concentration
ranges. Environ. Sci. Technol. 6:250-252, 1972.
59. Mueller, P. K. , and G. H. Hidy. Measurement Technology of Sulfates,
Appendix D. In Electric Power Research Institute Report 485.
Menlo Park, Calif.: Electric Power Research Institute, (in press)
60. U. S. Environmental Protection Agency. Part 53--Ambient air monitoring
reference and equivalent methods. Federal Register 40:7044-4063,
1975.
61. McKee, H. C. , R. E. Childers, and V. B. Parr. Collaborative Study of
Reference Method for Measurement of Photochemical Oxidants in the
Atmosphere (Ozone-Ethylene Chemiluminescent Method) EPA 650/4-75-
016. San Antonio, Texas: Southwest Research Institute, 1975.
49 pp.
6-68
-------
Chapter 7
RESPIRATORY TRANSPORT AND ABSORPTION
This chapter discusses the transport and absorption of ozone and other
photochemical oxidants within the respiratory tract. It includes lung
morphology and flow aspects of respiratory physiology and emphasizes
methodologic approaches to modeling.
Although the theory of the uptake of inert gases is well developed,
there is no adequate theory for gases like ozone, which are reactive
and metabolized by body tissues and fluids. Thus, it is not possible
to predict local tissue dosage at critical airway sites that can be
reliably related to toxic effects. In this chapter, the background
inrormation necessary for the development of a realistic model of lung
uptake of reactive gases is presented, and subjects needing further
study are identified. A successful model would help in correlating
observations in animal and man and estimating the effective dose at
reactive sites in man. Combined with a knowledge of biologic mechanisms
of tpxicity, it allows the prediction of health effects in man outside
the range of conditions in which observations are made.
The radical or reactive intermediates discussed in Chapter 2 are
included in the term "photochemical oxidants" for the purpose of this
chapter. Of the reactive intermediates discussed, only the hydroperoxy,
H02, and singlet oxygen, 0 (a A), radicals have a lifetime long enough
to allow transport of significant quantities into the lungs before
deactivation. Modeling of oxygenated organic aerosols will not be
-------
considered in detail in this chapter. According to the discussion in
Chapter 3, most of the oxygenated organic aerosols have a diameter
less than 0.5 ym, the range in which diffusion is the controlling
90,H9g,123
mechanism of deposition. Regional particle deposition models
can account for their deposition and clearance in the tracheobronchial
67a
tree and pulmonary regions. Work by Hounam suggests that Landahl's
84
model can be used to predict the deposition of oxygenated organic
aerosols that are highly diffusive, as condensation nuclei in the
nasopharynx.
This chapter first reviews and discusses selected research on local
dose aspects of ozone toxicity, the morphology of the respiratory tract
and mucous layer, air and mucus flow, and the gas, liquid, and tissue
components of mathematical models. Next, it discusses the approaches and
60
results of the few models that exist. A similar review was recently
done to define an analytic framework for collating experiments on the
effects of sulfur oxides on the lung. Pollutant gas concentrations are
generally stated in parts per million in this chapter, because experimen-
tal uptake studies are generally quoted only to illustrate behavior
predicted by theoretical models. Chapter 5 gives a detailed discussion
of the conversion from one set of units to another.
EFFECTIVE TISSUE DOSE
In this section, sites of action in the respiratory tract will be
discussed, along with experimental studies of gas uptake in animals.
Cumulative dose and dosage at critical sites of action will be defined,
as well as the general characteristics required for modeling the trans-
port and absorption in the respiratory tract.
7-2
-------
Sites of Action
The sites of action and effects of ozone and other photochemical
38
oxidants are described in Chapters 8 and 9. Recent work with primates
has suggested that ozone is absorbed along the entire respiratory tract,
penetrates more into the peripheral nonciliated airways, and causes more
lesions in the respiratory bronchioles and alveolar ducts as the inhaled
ozone concentration increases from 0.2 to 0.8 ppm. The most common
and most severe tissue damage was observed in the respiratory bronchioles.
The ciliated cells in the terminal bronchioles and the type 1 cells in the
epithelial layer of the proximal alveoli of rats were the primary sites
117a 22
of action of ozone at 0.5 and 0.9 ppm. Recent work in rabbits and
113
rats suggests that the mucous layer in the large airways does not
completely protect the underlying cells from ozone damage. Specifically,
Boatman and Frank noted patches of desquamated ciliated cells along the
conducting airways after acute exposures to ozone at 0.25, 0.5, and 1 ppm;
these desquamated patches were most often found at bifurcations.
Table 7-1 gives an overview of various irritant and nonirritant
gases commonly found in the atmosphere, their solubility in water, and
their main sites of action. The Henry's law constant indicates the
relative solubility in waterlike lung fluid. Although most of the infor-
mation goes back to 1924, it is supported and extended by numerous studies
25,27,28,39,50,51,62,
of the effects of war gases and industrial irritants.
78,128,134
Studies that have measured the uptake of pollutant and irritant gases
in different regions of the respiratory tract of animals and man and in
experimental airway models provide the most useful data for development
l,9a,10,24,29a,33,40,41,45,46,
and verification of gas transport models.
47,68,79,94,100,116,127,138,139,140
They generally show that the very
7-3
-------
Table 7-1
Physical Properties of Pollutant Gases and Their Site of
o
Action or Absorption in the Respiratory Tract-
Gas
Ambient
Concentration,
ppm
Henry's Law Const.
at 37 C, 1 atm,
mole fraction in air
mole fraction in water
Major Site
of Action
or Absorption
Ammonia-
0.02-0.2
Sulfur dioxide- 0.01-0.5
Hydrogen sulfide~ 0.03
3.5
59.7
704
URT
URT and
large bronchi
URT and
large bronchi
Formaldehyde-
Carbon dioxide
Ozone
Nitrogen dioxide
Nitric oxide
Oxygen
Carbon monoxide
Nitrogen
0.3
330
0.05-0.5
0.05-0.5
0.05-2
209,460
1-100
780,840
791
2,190
9,700
Converts to nitric
and nitrous acids
in water
33,900
51,800
67,400
100,700
URT
PULM
TBT and
TBT and
TBT and
PULM
PULM
PULM
PULM
PULM
— Derived from Haggard.
b 68a 102
— Data from International Critical Tables and Perry et al.
c_
URT = upper respiratory tract (nose, mouth, pharynx, larynx);
TBT = tracheobronchial tree (ciliated airways);
PULM = pulmonary region (alveolated airways and alveoli).
d
~ For these highly water-soluble gases, Henry's law may be assumed to
apply over the ambient concentration range.
7-4
-------
soluble nonreactive gases—like ammonia, sulfur dioxide, and formalde-
hyde—are nearly totally removed by absorption in the nasal passages
during normal breathing. Less water-soluble gases—like ozone and nitrogen
dioxide—penetrate more peripherally into the lung and are partially exhaled.
Total uptake depends on tidal volume, respiration rate, and in some cases
139
initial inhaled concentration. For example, data from dogs show that
nearly 100% of sulfur dioxide inhaled through the nose is removed before
reaching the first bifurcation, whereas only 27-70% of the ozone is removed
in this same region.
Inhalation of gas or aerosol-gas mixtures may shift the site of action
of the separate gases. Generally, the absolute quantity of irritant gases
absorbed by suspended particles is negligible at ambient gas and aerosol
concentrations; however, the irritation may be significantly increased by
the transport of soluble gases by particles deeper into the lung. Insolu-
ble particles, like carbon, and partially water soluble particles, may
absorb ozone, sulfur dioxide, and other gases, in the water phase.
Likewise, partially soluble particles with catalytic properties may
oxidize absorbed gases and convert them to more toxic chemical forms
(e.g., sulfur dioxide oxidized to sulfur trioxide in the liquid
phase of particles with the aid of metallic catalysts). Gaseous
mixtures that are chemically unreactive in the environment may react
rapidly with each other in the special conditions of the respiratory tract
to form other toxic compounds. The synergism observed by Bates and
16 20
Hazucha and by Bell et al., when human subjects were exposed to mix-
tures of sulfur dioxide and ozone may be caused by one of these physico-
chemical interactions. Another example of synergism related to
physicochemical interaction of gases and particles was described by
93
Mezentseva, _et^ al., who observed edema in the lungs of animals exposed
7-5
-------
to hydrochloric acid derived from the hydrolysis of titanium chloride,
TiCl, , but not in those exposed to similar concentrations of pure
gaseous hydrogen chloride.
Indexes of Critical Dose
To model the uptake of ozone and other gases for establishing dose-
response relationships at specific sites, local dose must be accurately
defined. In the past, this has not been done for specific sites.
Fairchild and Graham, Stokinger et_ al_., and Stokinger a have
developed expressions for the effective dose of ozone, which depend
only on the inhaled concentration and exposure time. "Dose" may be the
mass or moles of the toxic gas delivered to the site of interest, and
the average dose may be measured in micrograms per square centimeter of
surface.
For acute exposure in a specific airway, the average rate or flux
to the epithelial tissue or mucous layer may be the critical quantity
and is measured in micrograms per square centimeter per second (in the
unit is micrograms per square centimeter per breath). Chronic effects are
probably related to the time integral over the period of exposure. When
sensory receptors are involved in the acute response, the local flux to
the small surface areas containing the receptor sites may be crucial.
Controlling Factors in Ozone Uptake
The major respiratory factors in the control of ozone uptake are the
morphology (including the mucous layer), the respiratory flow, the physical
and chemical properties oE mucus, and the physical and chemical properties
of ozone. The next two sections discuss models of the morphology and the
air and mucus flow. The physical and chemical properties of bronchial
7-6
-------
13a 26
secretions have been reviewed by Barton and Lourenco and Chairman et al.
The relevant physical and chemical properties of ozone are its solubility
and diffusivity in mucus and water and its reaction-rate constants in water,
mucus, and tissue.
Solubility data for mucus are not available, but Table 7-1 indicates
that the Henry's law constant for ozone in water in lung conditions is
9,700. Solubility data for pure ozone and other physical properties are
"7 / TOO
available from various sources. ' Air Quality Criteria for Photochemical
124
Oxidants reports an ozone solubility of 0.494 ml/100 ml of water at 0 C
122
for ozone at 760 mm Hg; extrapolation of data from Thorp indicates 1.09 g/
liter of water at 0 C and approximately 0.31 g/liter of water at 37 C for
100% ozone. The value for 37 C agrees closely with the solubility calculated
from the Henry's law constant for pure ozone at 760 mm Hg.
The diffusivities of ozone in mucus, tissue, and water are unknown.
-5 2
As an approximation, the diffusivity of oxygen in water (2.5 x 10 cm /s)
may be used for the diffusivity of ozone in water and in the mucous layer.
Although reactivities of ozone in mucus and tissue are unknown, the results
of Alder and Hill and Hoigne and Bader, a which describe the decomposition
of ozone in aqueous solutions, may be used as a first approximation.
MORPHOLOGY OF THE RESPIRATORY TRACT AND MUCOUS LAYER
Human and Animal Airway Models and the Real Lung
The airways and the tissue lining of the human respiratory tract have
a very complex structure and dynamic behavior, which vary with age, sex,
56,61 66
and state of health. Horsfield reviewed the structure and function
of the respiratory tract and described how it can be simplified into air-
103
way models for calculating gas and particle transport. Phalen has
7-7
-------
summarized the limited available data concerning the similarities and
differences between the airways of humans and several animal species.
Hausknecht and Ziskind have also reviewed airway models for gas uptake.
Most mathematical models for particle deposition are based on circular
cylindrical, rigid models of the conducting airways with dimensions repre-
sentative of the normal adult. Diameter, length, and branching angle of
the airways at the same generation of the treelike branching structures
are usually assumed equal.
43
Although the airway model of Findeisen, which was later modified by
85
Landahl, has a very unrealistic branching structure, it has been used
widely for particle-deposition. ° Weibel's model "A" has a more
realistic symmetric dichotomous branching pattern and is currently popular.
His complete model has 16 generations of conducting airways in the tracheo-
bronchial region and seven partially or completely alveolated generations
in the pulmonary zone. Table 7-2 shows the numbers and diameters of the
conducting airways that are representative of an adult with a lung volume
of 4,800 cm3 at 75% inflation.
34
Other human airway models have been developed by Davies, Horsfield
67 104 137
et al., Phalen et al., and Yeh et al. The asymmetric model of
Horsfield e_t aj^. is more realistic than Weibel's, but is more difficult to
use for calculations. Yeh et^ a^. described a Monte Carlo technique for
constructing a realistic lung model from extensive morphometric data.
Such a lung model should have statistical distributions of the geometric
characteristics similar to those of an actual lung and will permit separate
deposition calculations for each lobe of the lung. Data on the airflow
distribution between lobes of the lung can be used to verify the model
7-8
-------
structure. Phalen et_ a^. compared the airway morphology of the human,
dog, rat, and hamster. The human tracheobronchial tree was found to be
more symmetric with respect to diameter ratios and branching angles than
103
those of the other species (but closest to that of the dog). Phalen
found that the bronchial tree structure is variable from species to species,
from lobe to lobe within a given lung, and from one depth to another in
49,76
the lung. Other animal airway models have been developed for the rat,
75 77
the guinea pig, and the rabbit. If bifurcations continue to be viewed
as critical sites for gas and particle deposition and dose-response
relationships, more refined airway models will be needed that define in
more detail the structure of bifurcation regions.
Because of the complexity of the actual structures, the emphasis in
modeling has been on obtaining an average representation, and the variability
among individuals tends to be neglected. There are two experimental
studies of variability of airway dimensions in living humans as revealed
86a
by aerosol deposition studies. Lapp et^ a.1^. assessed the size of alveolar
spaces in terms of half-life of aerosol persistence during breathholding
98a
and obtained a coefficient of variation of 20-25%. Palmes and Lippmann
report the variability of a measure reflecting the influence of anatomic
factors in tracheaobronchial deposition as revealed by in vivo retention
of y-tagged microspheres in the human thorax; they found a coefficient of
60-70%. Variability of direct anatomic measurements of numbers and sizes
91a
of airways in man is given by Matsuba and Thurlbeck, Angus and
Ha I22a
Thurlbeck, and Thurlbeck and Haines. In addition to expressing one
cause of differences in the susceptibilities of individuals, the variability
in model measures indicates a bias in the results of calculations that are
7-9
-------
107
based only on average values of the measures. Proctor and Swift
described the complex anatomy of the human nose and constructed a mor-
phologic model of the nasal airways from actual casts, which may require
simplification. Similar models need to be developed for animal upper
airways.
Mucus and Alveolar Tissue, Models
Because the mucous layer or the underlying cells may serve as either
final accumulation sites of toxic gases or layers through which the gases
diffuse en route to the blood, we need simplified models of these layers.
8
Altshuler et al. have developed the only available model for these layers
that can be used in a comprehensive system for calculating tissue doses of
inhaled irritants. It asisumes that the basement membrane of the tracheo-
bronchial region is covered with three discrete layers: an inner layer of
variable thickness that contains the basal, goblet, and ciliated cells;
a 7- ym middle layer composed of waterlike or serous fluid; and a 7-ym
23
outer layer of viscous mucus. Recent work by Boatman and Luchtel in
rabbits supports the concept of a continuous fluid layer; however, airways
smaller than 1 mm in diameter do not show separate mucous and serous fluid
layers.
FLOW ASPECTS OF RESPIRATORY PHYSIOLOGY
Respiratory Airflow Patterns in Lung Models versus the Real Lung
Flow Analysis in Weibel's Model "A". Tidal volume and respiratory frequency
are used with the anatomic dimensions to model airflow patterns in the
respiratory tract.
During the respiratory cycle, the volumetric flow rate of air varies
from zero up to a maximum and back. Usually, the expiratory phase is
7-10
-------
longer than the inspiratory phase, and there may be intervening pauses
between the two, especially after expiration. Silverman et al .
studied the respiratory airflow patterns of healthy young men at rest and
under a wide range of workloads. The maximal inspiratory flow rate
increased from a mean value of 40 liters/min in sedentary subjects to
100 liters/min at an exercise workload of 622 kg/min and to 286 liters/min
at 1,660 kg/min. The corresponding values for maximal expiratory flow
rates were 32, 107, and 322 liters/min.
When a quiet breathing rate of 15 cycles/min and a tidal volume of
3
450 cm are assumed in Weibel's model "A," the time-averaged velocities
are as shown in Table 7-2. Reynolds number and entrance length are calcu-
lated by treating each branch as a straight smooth tube. Analysis of
these data suggests that there is plug (uniform) flow in the trachea through
the third generation, partially developed laminar flow in the fifth through
seventh generations, and developed Poiseuille flow for an increasing frac-
tion of each branch in the eighth through sixteenth generations. With
maximal inspiratory and expiratory effort, however, velocities and Reynolds
numbers (Re) may be 22-45 times larger than during quiet breathing. This
suggests turbulent flow in the upper generations. Furthermore, because of
variations in compliance and resistance, ventilation is not equally dis-
tributed throughout the lung; consequently, branches of the same generation
may have different flow rates.
Skewed Profiles and Secondary Flows. The preceding flow analysis neglects
the complicated flow behavior initiated at the bifurcations and transferred
112 in
to the daughter branches . Schroter and Sudlow and Schreck and Mockros
measured the velocity profiles for steady flow in the daughter branches of
airway models that were geometrically similar to Weibel's model. For
7-11
-------
CM
1
01
rH
cd
H
01
4-1
cd
PS
60
C
•H
4-1
cd
01
m
60
a
•H
4—)
CO
0)
rH
o
cd
o
~
~
->
cd
IS
0
rH
4-1
•H
rH
cd
13
•rl
cd
pj
cd
C!
•r-l
g
CO
01
rH
O
u
m
i — i
UH
0
c
Ul 0
,C 'H
1 i | j
60 Cd
pj !-l /-v
0) 01
rJ fi N
0) Q
CO O Ol
0) Pd
O MH r^ N
Co m i-J
cd o
4-14-1 O
C 60
we ^
^
'-j
CO"0 I"3 V
rO M N] *
rH 0) O 1
O ,£>
C| 0,
01 !Z ^5-
Pi
1 01 4-1 "
0) 60 vH ,-v
B td o N
•rl rl O t3 CO
H 0) rH >— ' ^-
> a) B
5
s
o
f~] r\
4—1 /~^*
60 N
C rJ
0) s-'
B
o
Ol
4J ^-v
01 N
S Q
cd s*^
•H
Q
O N
cfl
rl Cl
0) O /-.
C -H N
S ^^
1
co c*o \o m 0s
f-H f^_ 1^. ^^ t^ ^.A^ ^-J .J-L £-} .« . -» fx^ _t< »J . ^-v
^ co i — i ON o^ co r — o vsO i~o (si i — i o o o o o
°0 0 CO in ^' CM* .-I -i 0 0 0 0 t*- o ^f CT^ LD
r-oo ino^oor-icocx)ojLnvt'inTt c^5 oo o°^ m oo m r— ( xo ^~^ co t*^~ ^4^ OJ *~^ f~^ c^
sO r— 1 ^J f^ ^f LO O"** ^^5 CO (NJ t~^
C~~ t~~ IT) CO CN] i— 1
i — i vo Tt* o^ in
^^vomr-cooo^-^u^comcoofvjcr-fM
COvOmf\JocJ^n^cMcovOr-cOr-ir-i
OO^Or-iOr-ifl-^rOtMrHr-i
-
LO
or-or-^oa-r-^Din^oocotNifNarvj-H
f\l"o LO LO co co co LO co c^ 0s co r^ — ^o ^o
OO CO CO LO "^ CO fv] (\3 I-H r— < •—* i—i O O O O O
i-t^HOOOOOOOOOOOOOOO
I~^ CNJ «.^ 00 vD C'J ""^ 00 ^O C1^ **^ 00 ^O C*^ *^1 00 vO
t~H CO vO C^J UO i™^ C*^ ***J" CTN O^ 00 ViO CO
rH cNmo ooi-i cnr^m
t — 1 CM
-------
Poiseuille or plug flow in the parent tube, the flow is symmetrically
split by the carina, and the higher axial velocities are directed off
the daughter-tube axis along the inside wall. The shear rates along the inside
wall are about 4 times larger than those along the outside wall, and the
peak axial velocity is twice the average bulk flow velocity. The flow
is also skewed as a result of the lateral convection or secondary flows
generated at the bifurcation (Figure 7-1).
Although secondary flows increase the uniformity of the flow dis-
tribution distally along the branch, this is significant only in the slower
regions. The flow profiles in the daughter branches become more complicated
when this asymmetric flow reaches a second bifurcation; however, they
follow the same general trends.
Sudlow and Schroter observed secondary flows at all flow rates
(Reynolds number, 50-4,500), regardless of the shape of the entry profile.
On inspiration, a pair of symmetric vortices is formed in each daughter
branch (Figure 7-1). They are strong enough to complete one helical
cycle within three branch diameters. On expiration, a set of four vor-
tices is generated in the parent tube.
These results suggest that simple parabolic flow is ensured only
in the conductive airways where the Reynolds number is less than 1.0.
There, the fluid inertia is negligible, and the convective fluid transport
is less than the molecular transport.
Turbulence. The simplified flow analysis of Weibel's model "A" indicates
undeveloped flow with a flat profile in the trachea for Reynolds numbers
110
(Re) up to approximately 2,000. However, this does not consider
disturbances produced by the rough walls, the eccentricity of the cross
section, and the larynx.
7-13
-------
BOUNDARY
LAYER
^- SECONDARY
FLOWS
Figure 7-1. Schematic representation of flow in daughter branches
of bifurcation model for steady inspiratory flow with
flat profile in parent branch. Velocity profiles in
plane of bifurcation ( ) and in normal plane ( )
are indicated in right branch. Orientation of secondary
flows and position of laminar boundary layer are shown
in left branch. (Reprinted with permission from Bell. )
7-14
-------
For steady inspiratory flow in hollow casts of the trachea and
o£
the first bifurcation, Dekker observed that a Reynolds number of 1,800
was required for initiation of turbulence. Similarly, for stead inspira-
tory flow in casts containing a larynx with the glottis in a natural open
131
position, the required Reynolds number was 450. West observed tur-
bulence in the trachea of lung casts during exhalation at a Reynolds number
of 800, although his hollow casts did not include a larynx. As the
141
position of the vocal cords changes in the real lung, the glottis of
the larynx functions as a variable orifice. During inspiration, a jet
of turbulent air enters the trachea and is directed against its ventral
wall. Although the length of the trachea would be insufficient for the
complete development of turbulent flow, the additional turbulence created
by the jet and the corrugated walls may cause the turbulence to approach
a fully developed state by the end of the trachea for a Reynolds number
greater than 3,000.
98
A model developed by Owen indicates that turbulence will gradually
decay in any branch in which the Re is less than 3,000. Assuming a peak
3
Re of 1,865 in the trachea—corresponding to a 1.9-s, 450 -cm inhalation—
Owen's model predicts a 10% decay in the trachea and in each of the first
14
two generations of bronchi. According to Batchelor s theory for the
change in turbulent energy at regions of rapid flow contraction, decays
of 15, 16, and 10% occur in the first three generations of bifurcation,
respectively. Therefore, turbulence generated in the trachea at an Re of
1,865 will have approximately 50% of its initial intensity when it enters
the third generation of bronchi. At inspiratory flow rates with an Re
greater than 3,000, decay would be even slower.
7-15
-------
These decay calculations neglect the possible effects of the strong
secondary flows generated at the bifurcation. The regions of very high
and very low shear rate caused by the secondary flows could also be
regions of high and low turbulence and dissipation. However, Pedley,
101
Schroter, and Sudlow argued that the boundary layer remains laminar
in the daughter tube for an Re less than 15,000 and had experimental
evidence to verify this assumption for a parent-tube Re up to 10,000.
Thus, the turbulent eddies are localized in the core, and the arguments
given above are sufficient for predicting their rate of decay.
Detailed descriptions of the convective airflow patterns in cast
replicas of the human respiratory tract during steady inspiration were
97
given by Olson et al. Their results show that the effect of the larynx
is such that flow patterns typical of smooth bifurcating tubes (secondary
motions and high shear rates along the inside wall) do not occur until
the lobar bronchi. Turbulent eddies produced by flow separation below
98
the larynx do not decay as rapidly as predicted by theory. Indeed,
small eddies were observed as far down as the sublobar bronchi with
200-ml/s flows in the trachea.
Wall and Flow Oscillations. Another flow complication is the effect of
131
heartbeat. West measured flow oscillations attributed to beating of
the heart in the segmentaL bronchi and found that they were detectable
only during breathholding or during pauses between inspiration and
expiration. The peak oscillatory flow rate observed was 0.5 liter/min,
which is approximately 20% of the peak flow rate in the segmental bronchi
during quiet breathing. These oscillations will improve gas mixing.
The minor variation in airway dimensions during expansion and con-
traction of the lungs in breathing can generate radial velocities that
would be significant only in the peripheral airways with the smallest
axial velocities.
7-16
-------
112
Schroter and Sudlow estimated that the macroscopic corrugations
in the airways are below the critical protuberance height at which laminar
flow can be disturbed.
Quasisteady Flow. Most mathematical models for particle and pollutant
transport assume steady flow conditions. However, flow actually varies
approximately sinusoidally over time, and breathing frequency ranges from
eight breaths per minute for sedentary conditions to 50 breaths per
minute during sustained work and exercise.
Quasisteady flow may be a more accurate description. This means
that the pulsatile flow in the lungs can be analyzed as a continuous se-
quence of steady-flow profiles. However, there are different and sometimes
72,112,135
conflicting criteria for quasisteady flow in the airways.
-1 Q
Bell suggested that the quasisteady flow is probably a valid
approximation at quiet-breathing frequencies and that velocity and pressure
profiles in the lung during quiet breathing can be obtained from experi-
mental steady-flow data. The same conclusions cannot be generally applied
to experimental particle deposition or gas-transfer measurements.
For regions in which the flow is not quasisteady, a transient-flow
82 83
solution may be possible. For example, Lakin and Lakin and Fox
developed a two-dimensional transient-flow solution for an idealized
symmetric bifurcation during the period at the end of inspiration and
before expiration. Their finding that vorticity decreases at the carina
or bifurcation apex suggests that particle- and gas-deposition rates may
be increased at these sites in the respiratory tract. It also suggests
that reactive gas deposition rates during normal oscillatory breathing
differ significantly from those predicted for steady flow—a view
18
suggested by Bell for particles.
7-17
-------
Convective and Diffusive Transport and Mixing. Another complication
with unsteady, periodic flows is the mixing of residual air in the
respiratory tract with the tidal air. The secondary flows and turbulence
both increase the mixing of inhaled gases and particles with dead-space
air. There have been numerous theoretical and experimental studies of
convective and diffusive gas or aerosol mixing in the respiratory tract,
but there is no general agreement on their relative significance in
specific regions of the tract.
Work by Altshuler ej^ a^. with 0.4- ym particles and a tidal volume
of 500 ml showed that only about 11-27% of new air in each successive
breath actually mixes with residual air. Theoretical particle-deposition
f\ T ~7 *3 ^
models developed by Altshuler, Beeckmans, and Davies have accounted
for the mixing of inhaled aerosol with residual air.
132
Whipple, Chen, and Wang showed that the distribution of an inhaled
aerosol bolus depends on the orientation of the successive airway bifurca-
tions and the volume of the bolus. On the basis of skewed velocity pro-
files, they made theoretical calculations of the distribution of aerosol
boli in branching airways that were in fair agreement with the experimental
data. Their results suggested that slow and shallow breaths should show
greater differences in dispersion of irritant gases in the airways.
13
Baker et al. theoretically analyzed simultaneous gas flow and
diffusion in Weibel's symmetric model. They applied a time-varying flow
with simultaneous longitudinal diffusion and concluded that convective
mixing is much less important than mixing induced by molecular diffusion.
114
By analogy with heat-transfer data in curved tubes and branching
systems, the local transfer rates of easily absorbed gases are expected
to be significantly affected by convective mixing in the large conducting
7-18
-------
airways. In the terminal bronchi and pulmonary regions, where convec-
tion is very slight, molecular diffusion is clearly dominant.
120
In a recent paper by Taulbee and Yu, convective mixing of parti-
cles or gases in lung airways was defined in terms of an apparent diffu-
sion coefficient. This was derived by assuming the inhaled particles or
gases to follow the average air velocity in each airway and assuming the
average velocity to be normally distributed among airways of the same
generation. Their calculations indicated that this apparent diffusion
coefficient is dominant and accounts for the pulmonary air mixing process.
9
As pointed out by Altshuler, the flexibility and curvature of the
airway walls and the gross inhomogeneities in the expansion and contrac-
tion of the lung structure cause flow separation and vortex motion that
are directly related to convective mixing and flow irreversibility. To
estimate the transport and uptake of inert and irritant gases within the
pulmonary region better, one could use the alveolar-duct model and the
calculation methods proposed by Altshuler for convective and diffusive
mixing. He pointed out the significance of the fact that almost the
entire wall of the alveolar ducts is open to the alveoli.
133
Wilson and Lin described precisely the three transport mechanisms
that act during the flow of a nonuniform gas in a tube: pure convection,
Taylor diffusion (where radial diffusion and axial convection are coupled
to produce an effective block or plug flow), and axial diffusion. They
defined regions in the respiratory tract where each of these mechanisms
130
dominates. By using Weibel's airway model and these three mechanisms,
they developed and analyzed a model to describe the transport of inert
gases within the conducting airways. For quiet breathing, pure convec-
tion dominated the transport in the zero- through seventh-order generations,
7-19
-------
Taylor diffusion dominated in the eighth through eleventh, and axial
diffusion dominated in the twelfth through seventeenth.
Q 1
Later research by La. Force and Lewis showed that gaseous concen-
tration gradients are negligible during quiet breathing (contradicting
30
the work of Gumming et al. ). Their anatomic models of the alveolated
airways and their calculation methods should be compared with those used
9 120
or proposed by Altshuler and Taulbee and Yu in the establishment of a
model for pollutant-gas uptake in the alveolated airways.
Airflow in the Nose and Nasal Airway Models. The complex anatomic struc-
ture of the nose is ideal for humidification, temperature regulation,
107
and pollutant scrubbing of inspired air. Proctor and Swift studied
nasal airflow by observing and measuring the flow of water through a
clear plastic model of the walls of the nasal passages. They used
steady flow with a Reynolds number equivalent to that for air in the
human nose. For an inspiratory flow of 0.4 liter/s (quiet breathing),
the linear inspiratory velocity at the nasal entrance reached at least
about 4.5-5 m/s and at most 10-12 m/s. These values are significantly
larger than the peak linear velocity of 2 m/s in the bronchial tree
during quiet breathing.
As the cross-sectional area expands beyond the entrance, flow
separation occurs and results in turbulence and eddies, which continue
as the air goes through the passages around the turbinates. The linear
velocity also decreases sharply in this region, and the air stream then
bends downward into the nasopharyngeal region. Because of these complex
flow patterns and the large surface area of the nasal mucosa, the nose
effectively scrubs particles and some gases from the inspired air.
7-20
-------
Proctor and Swift's nasal passage model and their charts of
the direction and linear velocity of airflow in the model could be
used for estimating the local uptake of gaseous pollutants and the
total scrubbing efficiency of the nose. The degree of swelling of the
nasal mucosa significantly affects the scrubbing efficiency, so more
refined airway models should simulate the morphology and flow behavior
during different states of swelling.
Mucus Flow Patterns in the Respiratory Tract
The dynamic properties of the mucous, serous fluid, and epithelial
layers of the respiratory tract are important for the transport, absorp-
tion, and desorption of reactive gases. The cilia beat at a fairly
constant frequency within the stationary serous layer and cause the
outer mucous layer to move up the respiratory tract. Clearance of
deposited particles and absorbed gases in the ciliated tracheobronchial
tree depends partly on the movement of this mucous layer.
There have been a number of studies of the thickness and velocity
31
of the mucous layer, with different results. Dalham reported a thick-
4
ness of 5 Mm in the trachea of rats. Similarly, Alder et^ ajU reported
29
10 ym and less in cats, and Comroe reported 10-15 Vm. Velocity has
31
been measured at 13.5 mm/min in rats, 0-35 mm/min in cows and 5-14
64 4
mm/min in dogs, 10.5 ± 3.7 mm/min in cats, and 15 mm/min in human
64
trachea and 3.75 mm/min in human main bronchus.
Clearance in the upper, or ciliated, region is governed by the
rate of mucus transport along the airways. These rates have been
48
measured in the human nose and in dogs, rats, and other species.
Asmundsson and Kilburn, Hilding, and Iravani established that
mucociliary clearance rates increase from the distal bronchi toward
7-21
-------
the trachea. Because bronchial openings retard mucus flow, bifurcations
receive an accumulation of mucus and associated particles. The rate of
mucus production and mucus thickness and velocity vary from one person
to another. Thickness increases and velocity decreases greatly when
119b,119c,119d,119e,119f
some toxic elements are present in the airway.
New techniques have been developed for the direct measurement
135a
of mucociliary transport rates in the trachea. Yeates et_ al_. used
an external gamma camera to follow a bolus of labeled microspheres
deposited in the large airways by aerosol inhalation; they fitted a
log-normal distribution to their measurements and obtained a geometric
mean of 3.6 mm/min and a coefficient of variation of 75% among 42 healthy
nonsmokers. (The short-term coefficient of variation was considerably
less, with a value of 25%.) Santa Cruz et^ al_. used a cinebroncho-
fibroscopic method on subjects whose larger airways were anesthetized and
observed the movement of small Teflon disks (0.68 mm in diameter and 0.13
mm thick) that were blown into the trachea through the fibroscope; they
reported an arithmetic mean of 21.5 mm/min (standard deviation, 5.5 mm/
min) in 16 normal nonsmokers and a much smaller value of 1.7 mm/min in
older patients with chronic obstructive lung disease. The large discrepancy
in the values reported for normal nonsmokers has several sources: different
statistical distributions, different methods, and a complicated hetero-
geneity in local mucus velocities. It appears that the invasive aspect
of the bronchofibroscopic technique caused some of the increase in the
measured transport rate.
Mucus flows in the bronchial airways have not been directly measured.
The measurements of particle clearance for radioactively tagged particles
depend on a mixture of deposition sites and mucus flow rates. However,
7-22
-------
such measurements have shown reproducibility in the individual and a
2a
large variation among individuals.
8
Altshuler et al. developed a method for estimating the thickness
and velocity of mucus throughout the tracheobronchial region. They
matched a particle clearance time with Landahl's lung model and assumed
a constant rate of mucus production per unit surface area and a uniform
mucus thickness throughout the tract, except for terminal bronchioles.
The calculated values are given in Table 7-3. From these values, mucus
velocity in each region can be obtained by dividing the length of the
region by the corresponding transit time. Jacobi, Thomas, and
Haque and Collinson also devised mucus clearance models for estimating
the lung-tissue dosage of short-lived alpha emitters.
The calculated velocities are based on the assumption that mucus
flows axially in each region. Actually, the bronchial openings repre-
sent obstructions to this parallel flow. Some have observed that the
58
mucus stream has a spiral path sweeping over the carina. According to
63
Hilding, mucus streams move axially and parallel in each section of
o
airway. The streams that intersect the carina bend 90 , pass parallel
to the carina, and then move upward. Others continue their flow undis-
turbed. Hilding stated that the reasons for this flow behavior may be
the change in direction of cilia beat and the presence of small whirlpools
in the middle of the margin of bronchial openings.
Some studies indicate that the mucous layer is not a continuous
blanket. Direct observations of the airways of normal and bronchitic
70 125
rats in vitro by Iravani and Van As and Van As and Webster failed
to find a mucous blanket at any level of the tracheobronchial tree.
7-23
-------
cd
H
tO
CO
3
O
o
CU
0)
i-H
CO
•H
O
a
o
CU
CU
13
O
CO
co co
3 CU
o a
S o
Id
H
4J
CO -H •>
3 en cu G
U S g -H
3 cd -H 0
S H H
H
CU
rH O •>
fO CO COCNI
4-1 <4-l CU 0
O P Vl CJ
H 3 J 4J 0 CO
cu jo d So d
X i-l CU CU -H
o d to p co g
CO *H rO OO rO M
M CO O CU 3 CU
H S iJ CO CO H
oo
i-H
cd
4J
CU
CU
rH
3
.d
CO
4J
<|
e
o
il
TJ
CU
•H
M
CU
O
ml
7-24
-------
126
Further work by Van As and Webster supports the discontinuity of
mucus and shows that it is transported in well-defined streams in the
larger airways of the rat. In addition to small regions of stagnation,
local retrograde movement was observed. One must also consider the
possibility that some of the serous fluid and mucus may be reabsorbed
as it moves up the respiratory tract. This could influence the local
tissue and mucous layer concentration of absorbed gas with time. Mor-
phologic examination confirmed that mucus is present as flakes, droplets,
and plaques. Droplets 0.5-1 Mm in diameter are believed to be the pri-
mary unit of mucus, and they aggregate to form flakes, which in turn
forms plaques. The smaller flakes may be transported individually over
individual metachronal fields of cilia, but the plaques are transported
en masse by the combined action of numerous metachronal fields. The
assumption of a continuous, stationary, and uniform layer of serous
fluid also needs critical examination, especially in nonciliated areas.
23
Recent unpublished studies by Boatman and Luchtel in rabbits show
that the mucous field is indeed continuous in the medium and smaller
airways. Their morphologic techniques are currently being extended to
the large airways. Differences in techniques may account for the dis-
126
crepancy between this work and that of Van As and Webster. Further
studies of the structure of the mucous layer in animals are therefore
needed to resolve the continuity-discontinuity question.
Mucociliary clearance from the nose or upper airways of man has been
105,106,107,107b
measured and described by Proctor and co-workers,
10,11 108
Anderson et al., and Quinlin et al. Through the use of radio-
active particles as tracers, mucus has been observed to move everywhere
7-25
-------
toward the nasopharynx, although indirectly. The mucociliary stream
from the paranasal sinuses joins the nasal stream all along the middle
meatus and above the posterior end of the middle turbinate. This flow
field ensures that the region of the nose receiving the greatest desposi-
tion of toxic gases and aerosols is better protected. Mucus that reaches
the nasopharynx, where the cilia disappear, is moved downward during
swallowing, because the soft palate wipes the posterior nasopharyngeal wall.
9a
Anderson et al. found a weak positive association between tracheo-
bronchial clearance and nasal clearance with a saccharine-particle method.
A strong positive correlation would have indicated that information about
the tracheobronchial clearance rate can be derived by studying clearance
rates in the nose, which is more accessible. The saccharine method was
shown to be a useful clinical tool for evaluating the status of the nasal
mucociliary function in human subjects exposed to ambient pollutants or
to controlled concentrations of specific pollutant gases or aerosols.
Effects of Inhaled Irritants and Airway Abnormalities on Air and Mucus Flow
A particular pulmonary irritant may alter air or mucus flow, and this
in turn is one of the factors determining the local tissue dosage of the
irritant. Such positive or negative feedback effects should also be
incorporated into a dynamic model.
Acute exposure to irritant gases or particles present in urban air and
in cigarette smoke at high concentrations can change the physical and
chemical properties of mucus and cause retardation or cessation of muco-
10,31,32,73,117 ^ c ^
ciliary clearance. Chronic exposure to some of the same
irritants appears to causet hypertrophy of the mucus-secreting elements and
glands of the upper airways, with a parallel increase in the production
7-26
-------
and secretion of airway mucus, and the bronchioles may show a marked
increase in the goblet cells, resulting in excessive mucus production
91
and airway obstruction, owing to ineffective clearance.
Bronchoconstriction caused by acute exposures to ozone or sulfur
dioxide may be expected to change the ventilation distribution, local
aerodynamics, and tissue dosage. Edema resulting from exposures to
toxic gases may alter the gas-absorptive capacity of the airways, in
addition to the aerodynamics. Reaction of irritant gases with surfac-
tant material in the alveoli may alter the absorptive capacity and
physical properties of the surfactant, influence edema formation, and
alter the clearance of inhaled particles.
GAS PHASE OF MODELS FOR POLLUTANT-GAS TRANSPORT IN THE RESPIRATORY TRACT
This section is concerned mainly with the approach to modeling the
gas-phase behavior of single reactive gases. The basic approach can also
apply to sulfur dioxide, ammonia, and other pollutant gases in which
water solubility alone controls the rate of uptake. The simpler case
of inert gases has been reviewed in a conference report edited by Papper
99
and Kitz.
Boundary Conditions
The simplest boundary condition between the gas and liquid phases
is the assumption that the gas concentration at the surface of the mucus
is zero. This same boundary condition is used for particles, so the
deposition theory for highly diffusive particles may also be applicable.
It may be adequate for predicting the net uptake of highly soluble gases
like sulfur dioxide during single-breath inhalations at low gas-phase
concentrations and could also apply to cases in which a pollutant gas
undergoes rapid chemical reaction at the surface of the mucus (i.e., if
7-27
-------
the rate of chemical reaction exceeds the rate of transfer from the gas
15,139
phase). Experimental data have suggested that the rate of transfer
of ozone from the air to the mucosal lining may be partially influenced
by its capacity for undergoing chemical transformation in the liquid
phase. If this zero-conconcentration boundary condition is to be used,
data are needed on the rate of chemical transformation of ozone in mucus.
A realistic boundary condition must account for the solubility of
the gas in the mucous layer. Because ambient and most experimental
concentrations of pollutant gases are very low, Henry's law (y = Hx)
can be used to relate the gas- and liquid-phase concentrations of the
pollutant gas at equilibrium; y is the partial pressure of the pollutant
in the gas phase expressed in mole fraction at a total pressure of 1 atm,
x is the mole fraction of absorbed gas in the liquid, and H is the Henry's
law constant. Gases with high solubilities have a low value of H. When
experimental data for solubility in lung fluid are unavailable, the
Henry's law constant for t'tie gas in water at 37 C can be used (see
Table 7-1). Gas-absorption experiments in airway models lined with
68
water-saturated filter paper gave results for the general sites of
uptake of sulfur dioxide and nitrogen dioxide that agree with uptake and
histopathologic data on animals.
The Henry's law constant in water is used in the McJilton et al.
92
uptake model to determine the equilibrium concentration of ozone and
sulfur dioxide at the surface of a simulated mucous film along the air-
130
ways in Weibel's symmetric model. It is also used to determine the
concentration of absorbed gas at the surface of the mucus when the
pollutant gas undergoes a homogeneous or heterogeneous chemical reaction
within the mucous layer.
7-28
-------
An additional complexity that has not been modeled is the simul-
taneous inhalation, absorption, and chemical reaction in the gas or
liquid phase of two or more gases (e.g., sulfur dioxide and ozone).
For sufficiently dilute mixtures, Henry's law can be used for each gas.
If droplet aerosols and one or more reactive gases are simultaneously
present, absorption with or without chemical conversion in the droplets
must be considered.
Transport Equations
As noted earlier, air-velocity profiles during inhalation and
exhalation are approximately uniform and partially developed or fully
developed, depending on the airway generation, tidal volume, and respira-
tion rate. Similarly, the concentration profiles of the pollutant in
the airway lumen may be approximated by uniform partially developed or
fully developed concentration profiles in rigid cylindrical tubes. In
each airway, the simultaneous action of convection, axial diffusion,
and radial diffusion determines a differential mass-balance equation.
The gas-concentration profiles are obtained from this equation with
appropriate boundary conditions. The flux or transfer rate of the gas
to the mucous boundary and axially down the airway can be calculated
from these concentration gradients. In a simpler approach, fixed
velocity and concentration profiles are assumed, and separate mass
balances can be written directly for convection, axial diffusion, and
92
radial diffusion. The latter technique was applied by McJilton et al.
To calculate more precisely the average uptake or the local varia-
tion in uptake in each airway, the local variations in velocity and
concentration profiles must be taken into account. For example, thin
momentum and concentration boundary layers occur at bifurcations and
7-29
-------
gradually increase in thickness with distance downstream. Bell and
19
Friedlander showed that particle and gas transfer to the airway
wall is greatest where the boundary layers are thinnest, e.g., at
the carina or apex of bifurcations.
139
Experimental data (e.g., from Yokoyama and Frank ) demonstrate
that the concentrations of ozone and other gases reaching the trachea
depend heavily on whether nose or mouth breathing is used. Detailed
gas-transport equations for the nose and mouth are difficult to formulate.
on
A simple approach used by La Belle et al. assumed that the nose behaved
like a scrubbing tower used in chemical processing. By varying the num-
ber of transfer units (defined as a portion of the dose sufficient to
permit equilibration between gas and liquid phases) and the molar ratio
of inhaled air to liquid, one may control the uptake of a soluble gas
to match experimental data. Another simple approach was used by Aharonson
1
et al. A quasisteady state was considered for unidirectional flow in a
one-dimensional model of the nose. A local mass balance was made for
the transfer of the soluble vapor from the air to the mucus-tissue layer.
The local transfer rate was strictly proportional to the local partial
pressure of the vapor in the gas phase, and the average transfer rate—
i.e., the overall uptake—could be determined by integrating the local
transfer rates over the entire length of the nose.
Local and Average Mass-Transfer Coefficients
The rate of mass transfer of a pollutant gas from the gas phase to
the wall of an airway can be written as:
dm = -K A(C - C*), (1)
dt G
where K is the average mass-transfer coefficient, in centimeters per
G
second, over the entire exposed surface area, A, of the airway; C is the
7-30
-------
average pollutant mass concentration in the gas phase in the airway;
and C* is the concentration of pollutant gas at equilibrium with the
102
absorbed gas in the wall. Local mass-transfer coefficients can be
similarly defined for subsegments of an airway wall.
K can be defined as a gas-phase transfer coefficient, independent
d
of the liquid layer, when the boundary concentration of the gas is
fixed and independent of the average gas-phase concentration. In this
case, the average and local gas-phase mass-transfer coefficients for
such gases as sulfur dioxide, nitrogen dioxide, and ozone can be estimated
from theoretical and experimental data for deposition of diffusion-range
18
particles. This is done by extending the theory of particle diffusion
in a boundary layer to the case in which the dimensionless Schmidt number,
v/D, approaches 1 (vis the kinematic viscosity of the gas, and D is the
molecular diffusivity of the pollutant). Bell's results in a tubular
bifurcation model predict that the transfer coefficient depends directly
on the square root of the average airway velocity, the diffusion coefficient
raised to the 2/3 power, and the bifurcation angle.
The prediction of such "hot spots" of gas transfer at bifurcations
22
are supported by experimental data on ozone-exposed rabbits. Longitudinal
slices of airways from these rabbits showed at low magnification that
desquamation of the ciliated epithelium cells was focal and sometimes
more intense at a bifurcation.
The dependence of the local and average transfer coefficients on the
square root of the average airflow rate is supported by the experimental
data and analysis of Aharonson
-------
Values for the average vapor-transfer coefficient from the gas
phase to the airway epithelium can also be estimated from heat-transfer
data in straight, curved, or bifurcating cylindrical tubes by using
the analogy between heat transfer and mass transfer. Such an approach
136
has been used by Yeh to predict the diffusional deposition of small
particles in the conducting airways.
LIQUID-TISSUE PHASE OF MODELS FOR POLLUTANT-GAS TRANSPORT IN THE
RESPIRATORY TRACT
The important properties of the mucous and serous layers for gas-
transfer models are thickness, viscosity, velocity gradients, the
diffusion coefficients of pollutant gases in the mucous and serous fluid,
and the chemical properties of these layers in the case of gases. Non-
reactive gases like sulfur dioxide must diffuse through this liquid layer
and the underlying cellular tissue layers before being absorbed by the
blood. Figure 7-2 is an idealized cross-sectional model of part of a
conducting airway showing separate tissue- and blood-layer components
of the liquid-tissue phase. In preliminary studies designed to predict
the average uptake in each generation, complexities like detailed velocity
gradients in the liquid phase are unwarranted. Velocity gradients could
be important in predicting local dosage to tissue; however, there are
no experimental data. The mucous and serous layers may or may not be
continuous and may constitute a homogeneous layer in some airways; thus,
a single homogeneous liquid layer of constant thickness in each airway
should be assumed until a more detailed description seems justified.
Similarly, in the future it may be advantageous to subdivide the tissue
layer into cell layers to reflect the pathologic evidence of ozone
damage on some cell types and layers.
7-32
-------
MUCOUS AND SEROUS
FLUID LAYERS
GAS PHASE
BLOOD
LAYER
Figure 7-2. Cross-sectional model of part of a conducting airway
in the respiratory tract, showing a gas phase and a
liquid-tissue phase subdivided into mucous- and
serous-fluid, tissue, and blood layers. (Derived
in part from McJilton et al. )
7-33
-------
1
Aharonson et al. combined the mucous and tissue phases in their
conceptual model of the nose into one layer that separates the air and
blood. The different resistances of each interface and layer are
lumped into their local transfer coefficient.
Figure 7-2 illustrates a three-compartment structure assumed by
92
McJilton e_t^ a^. for describing radial diffusion. It consisted of a
gas phase in the lumen of the airway, a liquid layer that lined the
airway, and a tissue compartment. The rate of movement of the gas into
the liquid layer, dm-./dt., is a function of the solubility of the gas in
the liquid, as defined by the Henry's law constant. The rate of movement
of the gas molecules across the liquid layer to the tissue compartment,
dnu/dt, is a function of the diffusion coefficient of the gas in the
mucous and serous layer. The concentration of ozone was assumed to be
zero at the liquid-tissue boundary. This means that ozone is instantaneously
converted by chemical reaiction when it reaches the tissue layer, but under-
goes no chemical reaction within the mucous layer. Such a model may be
useful for ozone, if the rate of chemical reaction is very low. Ozone
3
is known to react with hydroxyl ions at a very low rate, but it probably
reacts more rapidly with organic molecules in the mucus.
The mass-transfer coefficient in Eq. 1, K , which averages over the
(j
interface between the gas phase and the mucous-fluid layer, is given by
KG = I/[(l/kg)+(H/k.)], where k is the gas-phase mass-transfer
coefficient and k is the liquid-phase coefficient. This is the two-film
x/
model for interfacial mass transfer in which a gas molecule encounters
resistance from both phases as it diffuses from the bulk vapor to the
88,102
bulk liquid. McJilton et_ a^. assumed that k »k ,/H and used
7-34
-------
empirical data to evaluate k for several gases. An alternative pro
102
cedure is to evaluate k from penetration theory:
4D
where D is the diffusivity of the absorbed gas in the mucous layer and
t is the gas-liquid contact time during inhalation or exhalation. Using the
diffusivity of the absorbed gas in water may overestimate the actual
transfer rate, because diffusivity may be much smaller in a viscous mucous
fluid. Values of k can be determined as described in the previous section.
O
If the gas is converted by chemical reaction in the liquid layer,
k is modified according to the order of the reaction and whether it is
reversible or irreversible. For example, if ozone reacts rapidly and
irreversibly with organic molecules in the mucous layer, kj could be
10-100 times higher than the estimate based on penetration theory.
To model sulfur dioxide absorption by the blood through the walls of
46
the upper airways, as demonstrated by Frank et^ a^L. , one must include
the transport rates of sulfur dioxide across a mucus-tissue interface,
a tissue layer, and a tissue-blood interface (Figure 7-2) . For the case
of release of dissolved gas back into the exhaled air, which is depleted
of gas in the lower lung, the mucous layer would still represent the
greatest resistance to transfer. Consequently, the overall transfer
coefficient, KQ, would still be given by k /H.
JO
The processes of convection, axial diffusion, radial diffusion,
and chemical reaction in the liquid and tissue layers all occur
simultaneously. A rigorous approach requires solution of several
simultaneous differential equations. To avoid this complexity in
7-35
-------
preliminary models, the transfer processes can be calculated in successive
steps, as was done by McJilton et al.
The average dose rate or mass flux to tissue in each airway genera-
tion, defined as the mass transferred per unit time to the surface area
of the generation, is given by dnu/dt, the rate of mass transfer across
the liquid-tissue interface in Figure 7-2. The dosage to tissue is found
by integrating the mass flux over time or a number of breaths. The local
dose rate and dosage are defined in analogous ways.
DISCUSSION OF RESULTS WITH VARIOUS MODELS
Most models of gas uptake in the respiratory tract have been concerned
with carbon dioxide, carbon monoxide, oxygen, and anesthetic gases like
chloroform, ether, nitrous oxide, benzene, and carbon disulfide (e.g.,
89a 99
Lin and Gumming and Tapper and Kitz. ) Unfortunately, there are
only a few preliminary models of pollutant-gas transport and uptake in
the respiratory tract.
Models of Nasal Uptake
80
La Belle et al. modeled the absorption of various gases in the
nasal passages of rats by applying principles of scrubbing-tower design.
Their important characteristics were the Henry's law constant, the molar
ratio of inhaled gas to absorbing liquid in the nose, and the number of
transfer units. One transfer unit was defined as a portion of the nose
sufficient to permit equilibration. The blood flowing through the nasal
epithelial linings, rather than the moving mucous layer, was assumed to
be the principal absorbing liquid, and the Henry's law constants for
7-36
-------
water were used. The number of transfer units, N, was guessed to be
between 1 and 10. Results of the calculations for various gases are
summarized in Table 7-4. For oxygen through acrolein (relatively
insoluble through moderately soluble gases), the penetration was controlled
largely by the Henry's law constant. For ozone, with a Henry's law con-
stant of 9,700 at 37 C, the La Belle ejt al^. model predicts 99% penetration.
For more soluble gases like sulfur dioxide and ammonia, the penetration
also depends on the number of transfer units and the molar ratio of gas
to blood.
This model appears inadequate, for a number of reasons. Although
experimental data show that less sulfur dioxide than ozone penetrates the
nasal passages in animals, as predicted by the model, much more ozone is
139
predicted to penetrate than was demonstrated by Yokoyama and Frank
in dogs. The most likely explanation is that the model does not account
for chemical reactions of ozone in the mucus and epithelial tissue.
Another problem is that the nose is believed to behave more like a
scrubbing tower with fresh liquid at each level, inasmuch as the blood
24
supply is not continuous for the entire length of the nose, as assumed
in the model. Neglecting the surface area, volume, flow, and thickness
of the mucous layer in the nose will probably also give erroneous
results for soluble gases with a small diffusion coefficient in mucus
and for single-breath inhalations of a low concentration of any gas.
89b
However, recent work by Loring and Tenney partially supported
the La Belle et al. model by suggesting that the properties of the mucous
layer in the nose may be irrelevant for modeling the absorption of
relatively water-insoluble gases, such as nitrogen, oxygen, carbon
7-37
-------
Table 7-4
o
Calculated Penetration of Gases through the Nasal Passages In Rats-
Gas
Oxygen
Nitric oxide
Nitrogen
Ozone
Nitrous oxide
Carbon dioxide
Hydrogen sulfide
Chloroform
Bromine
Ethyl ether
Acetaldehyde
Acrolein
Chlorine
Sulfur dioxide
Acetic acid
Formaldehyde
Hydrogen cyanide
Ammonia
Phenol
Hydrogen bromide
Hydrogen chloride
Hydrogen iodide
Henry's Law
Constant,
mole
fraction gas
mole fraction
in solution
50,000
31,000
10,000
9,700
2,560
1,900
610
475
160
68
25
20
12
10
7
3.8
2.5
1.5
0.4
0.003
0.001
0.0008
Fraction of Gas that
b
Penetrates to Lung,~ %
N = 1
100
100
99
99^-
98
95
90
90
80
70
60
52
28
26
25
10
8
6
5
5
5
5
N = 5
100
100
99
99^
98
95
90
90
80
70
60
50
25
18
7
4
2
1.1
0.8
0.7
0.6
0.6
N = 10
100
100
99
99£
98
95
90
90
80
70
60
50
20
9
2
0.2
0.05
0.02
0.01
0.01
0.01
0.01
a 80
— Dervied from La Belle et al.
— N = number of transfer units.
— Extrapolated values.
7-38
-------
dioxide, and nitrous oxide. They observed that the flux of these gases
from the frontal sinuses of cats was explained best by a perfusion-limited
blood absorption mechanism.
An accurate nasal model must also account for the airflow rate and
1
the concentration of the inspired gas. Aharonson et^ al^. conclusively
demonstrated that the "uptake coefficient," or average mass-transfer
coefficient, over the entire nose for acetone, ozone, sulfur dioxide,
and ether increased with increasing airflow rate.
Yokoyama and Frank, Frank et^ a^., Brain, and Egle ' largely
overlooked this flow-dependent relationship, because they did not normalize
their retention data into the average-transfer-coefficient form. In fact,
failure to do this led Yokoyama and Frank to the erroneous conclusion
that "the uptake of 0., was inversely rated to flow."
Aharonson et al. discussed four explanations for the increased
uptake with increased flow: misinterpretation of data owing to a dependence
of the average transfer coefficient on vapor concentration, decrease in
the gas-film resistance, increased perfusion of nasal tissue, and increase
in the effective surface area for uptake. For a gas like ozone, which
is fairly insoluble but probably highly reactive in the mucous layer,
the gas film or concentration boundary layer represents the major resistance
to uptake. As discussed earlier, the gas-phase average or local transfer
coefficient in airways of the tracheobronchial tree is predicted to depend
on the square root of the average airflow rate in the airway and is
independent of gas-phase concentration. Because the data analyzed by
Aharonson et al. agree roughly with this square-root dependence, the
7-39
-------
properties of the boundary layers in the nose may be similar to those
in the tracheobronchial tree. The vapor concentration enters only in
calculation of the average flux to the tissue from the product of the
transfer coefficient and the concentration gradient between the gas
phase and the liquid-tissue phase. The overall transfer coefficient,
Kp, in Eq. 1 may be concentration-dependent, if the vapor is reacting
reversibly in the liquid layer or reacting reversibly with a second
dissolved vapor in an inert liquid layer.
Models of Tracheobronchial Uptake
There are no published models that adequately describe ozone uptake
in the tracheobronchial tree. To present the methodology, a few published
and unpublished models of the uptake of various gases will be reviewed.
The McJilton et al. model of ozone uptake has been widely cited,
although not formally published. It is described here because it was
the first attempt to model the absorption of pollutant gases in each
generation of the tracheobronchial tree.
They used mass-balance expressions in finite-difference form to
approximate the convection and diffusion of the pollutant gas in a
25-segment airway model that started at the trachea and was a modified
version of Weibel's model. A sinusoidal breathing cycle and uniform
plug flow were assumed in each of the first 20 segments. Beyond the
twentieth segment, where the segmental volume was greater than 5 ml,
uniform convective mixing was assumed. The cylindrical airways were
assumed to be lined with a stationary mucous-fluid layer and a tissue
layer as shown in Figure 7-2. The mucus was assumed to have the
properties of water and thicknesses of 10 ym in the upper generations,
3-5 ym in the alveolar ducts, and 0.3 ym in the alveoli. There were
no chemical reactions in the mucus.
7-40
-------
Finite-difference techniques were also used to calculate the
rate of diffusion of the pollutant gas from the airway to the mucous
layer and through the mucous layer to a perfectly absorbing sink at
the mucus-tissue interface. The mass of pollutant lost from the
airway or transferred across the air-mucus interface during each
breathing cycle was divided by the segmental surface area to obtain
the dosage, in micrograms per square centimeter per breath.
Although convection, axial diffusion, and radial diffusion actually
occur simultaneously, a multistep procedure was adopted in the finite-
3
difference calculation. For each 5-cm increment in tidal volume and
for each time increment, At, the differential mass-balance equations were
solved for convection, axial diffusion, and radial diffusion in that
order. This method may slightly underestimate the dosage for weakly
soluble gases, because the concentration gradient in the airway may be
decreased.
Although the authors view their results with this model as only pre-
liminary, a few of the results are presented here to contrast the expected
behavior of water-soluble gases (e.g., sulfur dioxide) with that of fairly
insoluble gases (e.g., ozone).
Figure 7-3 shows the percentage of total gas uptake for the steady
3
state (after five or six breaths of 500-cm tidal volume with 2-s
inspirations). Uptake increases from about 75% for a relatively insoluble
4
gas with a Henry's law constant of about 10 to a peak of 95% for soluble
gases with a Henry's law constant of 20 or less. Figure 7-4 shows why
there is only a 20% variation in uptake over a wide range of the Henry's
4
law constant (10-10 ). The model predicts that the dosage of gases of
7-41
-------
100
90-
80 -
70-
a-
60-
DECREASING SOLUBILITY
N
SO-
H?S C02 0
10
10'
10'
H,, Henry's Law Constant at 37 C
Figure 7-3. Uptake of pollutant gases in the entire tracheobronchial
tree and pulmonary region at steady state as a function
of Henry's law constant. (Modified from the model results
of McJilton et al.92)
7-42
-------
10
-3
V
!
8 io-
10
-8,
H* 100
H=500
H515000
labor
10 ^ ' < I I I < ( l~j "T"» I l I ' i ' t ' I ' I '
1 3 5 7 9 }} 13 15 IT 19 2) 23 25
MODEL SEGMENTS
Figure 7-4. Uptake or dose predicted for each model segment by
McJilton et al. for gases of different solubilities
at 37 C. Tidal volume = 500 cm . Inspiration time =
2 s. (Modified from McJilton et al. )
7-43
-------
low solubility is fairly uniform throughout most of the airways until
the alveoli are reached, whereas gases of high solubility are predomi-
nantly removed in the upper airways of the tracheobronchial tree.
The model also predicts an increase in uptake as tidal volume
increases over a constant breathing period. As the breathing period
increases at a constant tidal volume, the uptake also increases. In
the former case, increased ventilation of peripheral airways with a high
surface:volume ratio increases uptake. In the latter, the period for
radial diffusion is increased in every segment.
Figure 7-5 contrasts the steady-state uptake of sulfur dioxide and
ozone per breath in each segment of the McJilton et al. model when the
inhaled gas concentration at the entrance to the trachea is 1,000 y g/m .
The patterns for the uptake of the highly soluble sulfur dioxide and
the relatively insoluble ozone are strikingly different. The segmental
-3 , 2
dosage of sulfur dioxide peaks at 1.2XLO y g/cm -breath in segments 11
and 12 (immediately beyond the lobar bronchi). Ozone dosage is fairly
-6 , 2
uniform around 8.5x10 yg/cm -breath in segments 1-18. It then dips
-6 '2
sharply to 5x10 yg/cm -breath in segment 22 and peaks again at
-6 2
8.3x10 yg/cm -breath in respiratory bronchioles. A much smaller
-7 3
dosage of 2x10 yg/m -breath is calculated in the alveoli, mainly
because of the large surface area.
There are no experimental data to verify the detailed dosage dis-
tribution among airway segments; however, experimental data on mouth-
breathing animals and man support the general concept of rapid absorption
24,46,140
of sulfur dioxide and greater peripheral absorption of ozone.
7-44
-------
29a
Recently, Corn et al. measured an overall mass-transfer co-
efficient for sulfur dioxide and nitrogen dioxide in the upper airways
of the tracheobronchial tree in cats during controlled respiratory
cycles. Their measured transfer coefficient for sulfur dioxide was
nearly 100 times larger than the average transfer coefficient predicted
between the trachea and segmental bronchi from Figure 7-5. Their co-
efficients for sulfur dioxide were also slightly dependent on concen-
tration, but they were independent of concentration for nitrogen dioxide.
At a fixed tidal volume, the measured transfer coefficients were consistent
with the square root of the breathing frequency or average flow rate, as
predicted by the previously discussed theory based on the boundary layer
model, and as shown by the uptake data for sulfur dioxide in the naso-
pharyngeal region.
In general, the McJilton et al. model appears to be useful for
estimating the uptake in the tracheobronchial tree and pulmonary region
of water-soluble and relatively water-insoluble gases that are nonreactive
with the mucous layer. Nonreactive gases that are only partially soluble
in tissue or blood (sulfur dioxide, oxygen, carbon dioxide, and anesthetic
gases) may exert a backpressure that inhibits the gas uptake from the air-
ways. Modifications, including the local blood flow rate and the tissue
thickness (Figure 7-2) are required to handle these gases properly.
The major weakness is the requirement of nonreactivity of gases in the
mucous layer. Very weakly reactive gases may be treated as nonreactive.
However, the uptake of ozone, which is known to decompose in water and
is expected to react rapidly with biopolymers and other organic molecules
in the mucous layer, is probably underestimated in the upper airways and
7-45
-------
CO
^^^
w
01 X 40 H3d 6n 3SOQ
2 O. CD
I
I
I
10
! '
ro CM
1 i
O
t:
in
ro
CD
N-
rO
o
to
O
< _J
CJ
O
CO
r
o
^D
o
r
o
o
D
7111
O
IT)
O
T i I
CM
TO
CM
Ol
IO
to
ro
CJ
"o "^
g 21
4J — /
T3 B
C a) •
n o en
0) O CM
T3 O
•O e8 -H
•S ^
^ fi
3 H B
q_, O
"4-1 .'^
0 SB-
0) 0)
M
CB
CD
O
T3 -H
-
II
0
B
O
en
in
a
0 4-1
c
>•>
4J ca
B >-)
a) 1-1
Hi »
-------
overestimated in the terminal airways of their model. Thus, their model
represents a worst-case estimate of dosage of ozone to the terminal
airways, which are unprotected by mucus. Too little is known of the
chemical and physical properties of the mucous layer, and there is great
uncertainty in the values of the diffusivity of ozone or other gases to
be used in the liquid-tissue phase of gas uptake models.
Gases that do not react irreversibly with epithelial tissue, such
as anesthetic gases, may diffuse into the bloodstream and will ultimately
be eliminated from the body. A different and earlier model developed by
37
DuBois and Rogers estimates the rate of uptake of inhaled gas from the
tracheobronchial tree in terms of diffusion through the epithelial tissue,
rate of blood flow, and solubility of the gas in blood. The rate of uptake
from the airway lumen is determined by the equation:
V
o (xQ+DA)760
where V is the rate of uptake of gas from the lumen, P is the partial
pressure of the gas in the lumen, a is the solubility of the gas in blood,
D is the coefficient of gas diffusion in tissue, A is the surface area of
•
the bronchial segment, Q is the bronchial blood flow, and x is the thickness
of the bronchial epithelium between lumen and blood flow. DuBois and
Rogers used this equation to calculate the uptake of several gases of
different solubilities from the first 16 generations of Weibel's model
during acute exposures.
7-47
-------
The absorption distribution between generations of Weibel's model
had the same trend for acetone, nitrous oxide, and sulfur hexafluoride,
which are soluble, moderately soluble, and insoluble, respectively, in
blood or water. Absorption decreased by one-third from the trachea
through the third generation and then increased rapidly and continuously
with depth in the tracheobronchial tree; absorption in the sixteenth
generation was 15-25 times that in the trachea. The relative magnitude
of absorption at each generation was directly related to solubility.
The distribution for nitrous oxide, whose solubility is slightly greater
than that of ozone in water, differs radically beyond the third generation
from the distribution predicted by McJilton et al. for ozone (Figure 7-5).
Similarly, the uptake distribution for acetone, whose solubility is close to
that of sulfur dioxide, differs significantly from the prediction for
sulfur dioxide (Figure 7-5) . This is plausible, because the tissue thick-
ness decreases and the blood flow per generation increases with depth
"beyond the third generation in the DuBois and Rogers model, whereas, in
the McJilton et al. model, blood absorption is neglected, the mucous-layer
thickness decreases only slightly with depth, and the coefficient of
transfer across the air-mucus interface decreases rapidly.
An improved gas-uptake model should incorporate the features of
the DuBois and Rogers model and the McJilton et^ al^. model. As shown in
Figure 7-2, the model for gas uptake in the airways should include separate
layers for mucous-serous fluid, epithelial tissue, and blood. Development
of such a model awaits reliable data and methods for predicting the
coefficient of diffusion of pollutant gases in tissue and information on
the rates of local perfusion of blood and lymph in the bronchial epithelium.
7-48
-------
52,53,54,138
Experimental data from humans and animals on the rate of
sulfur dioxide absorption in blood could be used to make improved estimates
of the tissue diffusion coefficients in vivo.
New or improved methods are needed to measure local uptake experi-
mentally. Such data can be used to verify the detailed dosage distribution
predicted by the models. For example, the retrograde catheter and tracheal
29a
cannula system used by Corn et al. appears promising for transfer-
coefficient measurements within segments of the tracheobronchial tree. A
I6a
similar method was used by Battista and Goyer to measure the absorption
of acetaldehyde vapor in the dog lung.
Radioactive lung-scanning techniques that use tagged irritant gases
could give regional uptake data similar to scans obtained from radioactive-
particle deposition and clearance studies. With methods of chemical
separation and quantitation of the radioactivity in compounds isolated
from the mucous layer, the reactions of ozone with biopolymers may be
determined. Autoradiographic methods may also be useful for measuring the
local uptake of tagged soluble gases within specific airways.
18
Bell described how deposition of particles by convective diffusion
in the respiratory tract can be used to estimate the average and local
rates of gas transfer. For example, local inhomogeneities in pollutant-
gas transfer can be included in uptake models. This is done by multiplying
the local-transfer coefficients for 0.088-ym-diameter particles (Figure 7-6)
2/3
by (D2as/Dparticie) '•> D is tne coefficient of diffusion of the gas or
particle in the gas phase. When the boundary is perfectly absorbing, the
local gas-phase transfer coefficients are to be multiplied by the gas
concentration, the surface area between appropriate contours in Figure 7-6,
and the inhalation time to determine the local gas dosage. Figure 7-7
7-49
-------
Dp=0.088/jm 0=100 cm/sec
8.0-
Y 0-
-2.0 -
-4.0 -
-6.0
-8.0 -
- 0 Y
4.4
--2.0
-J-4.0
--6.0
--8.0
4.8
TRANSFER.
COEFFICIENT
CONTOURS
(cm/sec) x |Q3
A
8
C
D
E
F
G
H
I
J
K
L
M
N
0
= 8.42
= 7.36
= 6.31
= 5.26
= 4.21
= 3.79
- 3.37
= 2.95
= 2.53
= 2.10
= 1 .68
= 1.26
= 0.842
= 0.631
= 0.421
X cm
Figure 7-6. Transfer-coefficient contours for 0.088-ym-diameter particle
deposition in the daughter branch of the three-dimensional
bifurcation model shown in Figure 7-1 during inhalation. The
time-averaged velocity in the parent branch, U, is 100 cm/sec.
Total surface area = 18.6 cm. X = distance (in centimeters)
downstream from the carina; Y = deposition locations around
the branch circumference with Y = 0 at the carina. Each unit
of the ordinate corresponds to a distance of 0.215 cm.
(Adapted from Bell. )
7-50
-------
ro
O
o
d>
o
o
10 20 30 40 50 60 70 80 90 100
% AREA w/kloc > kloc
Figure 7-7. Cumulative surface-area distribution for Figure
7-6, showing the fraction of surface area of daugh-
ter branch with local transfer coefficient equal to
or greater than stated value (d = 0.088 ym, U =
100 cm/sec). (Reprinted with permission from Bell. )
7-51
-------
allows a rapid determination of the area between contours. When the
boundary is not perfectly absorbing, the local gas-phase transfer co-
efficient must be substituted for k in the expression 1/K = 1/k +
. Here, Kp is redefined as the coefficient of local transfer
across the air-liquid interface. Local nonuniformities of gas transfer
would be most prevalent in single-breath experiments, during the transient
periods before equilibrium is attained, and in exposures with pollutant
gases that react rapidly with the mucous layer.
Dose-Response Correlations
Modeling of gas transport is also useful for correlating dose-response
24
data obtained under different conditions. Brain suggested that the
total dose of an inhaled gas is related to ventilation rate, duration of
44
exposure, and gas concentration before inhalation. Folinsbee et al.
exposed human subjects to ozone at 0.37, 0.5, or 0.75 ppm for 2 h while
they were at rest or exercising intermittently. The primary response of
the subjects was an alteration in the exercise ventilatory pattern. They
showed an increase in respiratory rate and a decrease in tidal volume that
were correlated with the total dose of ozone (expressed as the volume of
ozone inspired during exposure). Other pulmonary function data, like the
flow at 50% of vital capacity, also appeared to be related to the volume
of ozone inspired.
42 118 119a
Fairchild and Graham, Stokinger et al. , and Stokinger showed
that the toxic effect of ozone in experimental animals is cumulative.
They found that the effective dose depends on the product of ozone concen-
tration and exposure duration for short-term single exposures .
7-52
-------
When the pulmonary response is activated by irritant receptors in
the nose, response for different flows and concentrations would not be
expected to correlate with the volume of inspired gas but rather with
regional dosage (e.g., nasal) or the local dosage of gas to irritant
2
receptors lining the airway.
SUMMARY
This chapter has discussed the general approach required to model
the transport and absorption of ozone and other pollutant gases in the
respiratory tract. For unreactive or very weakly reactive gases, there
are a few models that are qualitative descriptions for assessing total
dosage and dosage in major regions. However, there is no adequate model
for gases like ozone, which are strongly reactive within the mucous
and tissue layers.
RECOMMENDATIONS
The development of models requires more knowledge about the chemical,
physical, morphologic, and flow properties of the mucous layer; the
kinetics of the reactions of ozone in the mucous and tissue layers; and
the molecular diffusivity of ozone in these layers. Similar information
is needed for the hydroperoxy, HO , and singlet oxygen, 02 (a A), free
radicals, which are reactive intermediates in photochemical smog.
Furthermore, a realistic model based on such knowledge needs to be
verified by measurements of uptake and tissue dosage in the various
regions of the respiratory tract. These are currently difficult to make,
but are required to establish accuracy and reliability. New methods of
sampling and techniques using tagged gases should be developed, so that
local uptake can be measured.
7-53
-------
An extensive effort is needed in studies of pollutant-gas transfer,
absorption, and reaction in the respiratory tract. After some of the
experimental questions about behavior of ozone in the mucous layer and
adjacent tissue are answered, available methods for calculating the local
dosage to critical airway sites can be used in new uptake models. Gas-
absorption and particle-deposition models for the upper respiratory
tract (nose, mouth, pharynx,, larynx) also need to be improved. Experimen-
tal data now available can be used to develop semiempirical relations for
gas uptake in the nose in a procedure analogous to that used to model
particle deposition. Development of a more refined model for nonreactive
gases requires data on gas diffusivities in the mucus and tissue,
local blood perfusion rates in the nasal epithelium, and physiologic and
pharmacologic factors affecting the mucosa and local blood flow rates.
Models need to be developed for mixtures of gases that may interact
chemically in the gas phase,, in the mucus, or in aerosol droplets to form
other species. This requires theoretical and experimental studies of
dissolution, absorption, adsorption, and desorption of gases in or on
aerosols in the respiratory tract.
Improved modeling is needed for the design and interpretation of
animal experiments and controlled human studies, and for the collation
of diverse data from animal and human exposures to ozone. Calculations
of local dose at reactive tissue sites can help to explain the mechanisms
of toxicity and are needed to extrapolate animal and human data for
assessing population risks under different environmental conditions.
7-54
-------
REFERENCES
1. Aharonson, E. F., H. Menkes, G. Gartner, D. L. Swift, and D. F. Proctor.
Effect of respiratory air flow rate on removal of soluble vapors by
the nose. J. Appl. Physiol. 37:654-657, 1974.
2. Alarie, Y. Sensory irritation by airborne chemicals. CRC Crit. Rev.
Toxicol. 2:299-363, 1973.
2a. Albert, R. E., M. Lippmann, H. T. Peterson, Jr., J. Berger, K. Sanborn,
and D. Bohning. Bronchial deposition and clearance of aerosols.
Arch. Intern. Med. 131:115-127, 1973.
3. Alder, M. G., and G. R. Hill. The kinetics and mechanism of hydroxide ion
catalyzed ozone decomposition in aqueous solution. J. Amer. Chem. Soc.
72:1884-1886, 1950.
4. Alder, K. B., 0. Wooten, and J. M. Dulfano. Mammalian respiratory muco-
ciliary clearance. Arch. Environ. Health 27:364-369, 1973.
5.' Altman, E.~ 1., and 2.~~ C7 Men. Aerodynamical characteristics o£ the bronchial
tree of the right human lung. Biofizika 20:303-307, 1975. (in Russian,
summary in English)
6. Altshuler, B. Calculation of regional deposition of aerosol in the respir-
atory tract. Bull. Math. Biophys. 21:257-270, 1959.
7. Altshuler, B., E. D. Palmes, L. Yarmus, and N. Nelson. Intrapulmonary mix-
ing of gases studied with aerosols. J. Appl. Physiol. 14:321-327, 1959.
8. Altshuler, B., N. Nelson, and M. Kuschner. Estimation of lung tissue dose
from the inhalation of radon and daughters. Health Phys. 10:1137-
1161, 1964.
7-55
-------
9. Altshuler, B. Behaviour of airborne partides in the respiratory tract, pp.
215-231. In G. E. W. Wolstenholme and J. Knight, Eds. Circulatory and
Respiratory Mass Transport. A CIBA Foundation Symposium. Boston:
Little, Brown and Co., 1969.
9a. Andersen, I., P. Camner, P. L. Jensen, K. Philipson, and D. F. Proctor.
A comparison of nasal and tracheobronchial clearance. Arch. Environ.
Health 29:290-293, 1974.
10. Andersen, I., G. R. Lundquist, P. L. Jensen, and D. F. Proctor. Human
response to controlled levels of sulfur dioxide. Arch. Environ.
Health 28:31-39, 1974.
11. Andersen, I., G. R. Lundquist, and D. F. Proctor. Human nasal mucosal func-
tion in a controlled climate. Arch. Environ. Health 23:408-420, 1971.
lla. Angus, G. E. , and W. M. Thrulbeck. Number of alveoli in the human lung.
J. Appl. Physiol. 32:483-485, 1972.
12. Asmundsson, T., and K. H. Kilburn. Mucociliary clearance rates at various
levels in dog lungs. Amer. Rev. Resp. Dis. 102:388-399, 1970.
13. Baker, L. G., J. S. Ultman, and R. A. Rhoades. Simultaneous gas flow and
diffusion in a symmetric airway system: A mathematical model. Respir.
Physiol. 21:119-138, 1974.
13a. Barton, A. D., and R. V. Lourenco. Bronchial secretions and mucociliary
clearance. Biochemical characteristics. Arch. Intern. Med. 131:140-
144, 1973.
14. Batchelor, G. K. The Theory of Homogeneous Turbulence. London: Cambridge
University Press, 1953. p. 74. (UNVERIFIED)
15. Bates, D." V.' Air pollutants and the human lung. The James Waring Memorial
Lecture. Amer. Rev. Resp. Dis. 105:1-13, 1972.
7-56
-------
16. Bates, D., and M. Hazucha. The short-term effects of ozone on the human
lung, pp. 507-540. In National Research Council.. Assembly of Life
Sciences. Proceedings of the Conference on Health Effects of Air
Pollutants, October 3-5, 1973. Senate Committee on Public Works Print
Serial No. 93-15. Washington, D. C. : U. S. Government Printing
Office, 1973.
I6a. Battista, S. P., and M. M. Goyer. studies of the pulmonary absorption of
acetaldehyde vapor in the dog. Fed. Proc. 33:569, 1974. (abstract)
17. Beeckmans, J. M. The deposition of aerosols in the respiratory tract. I.
Mathematical analysis and comparison with experimental data. Can. J.
Physiol. Pharmacol. 43:157-172, 1965.
18. Bell, K. A. Aerosol Deposition in Models of a Human Lung Bifurcation.
Ph.D. Thesis. Pasadena: California Institute of Technology, 1974.
426 pp.
19. Bell, Tt. A., and S. K. Friedlander. Aerosol deposition in models of a human
lung bifurcation. Staub-Reinhalt Luft. 33:178-182, 1973.
20. Bell, £. A., tf. S. linn, and J. D. Hackney. Effects on pulmonary function
of humans exposed to mixtures of ozone, sulfur dioxide, and their
reaction products. Fed. Proc. 34:428, 1975. (abstract)
Delete 2"l--omitted
22. Boatman, E. S., and R. Frank. Morphologic and ultrastructural changes in
the lungs of animals during acute exposure to ozone. Chest 65(Suppl.)
9S-18S, 1974.
23. Boatman. E. s., and D. Luehte. University of Washington, Seattle, Washing-
ton. Personal communication, August 1975.
7-57
-------
24. Brain, J. D. The uptake of inhaled gases by the nose. Ann. Otol. 79:529-
539, 1970.
25. Cameron, G. R., J. H. Gaddum, and R. H. D. Short. The absorption of war
gases by the nose. J. Path. 58:449-455, 1946.
26. Charman, J., M. T. Lopez-Vidriero, E. Real, and L. Reid. The physical and
chemical properties of bronchial secretion. Brit. J. Dis. Chest 68:
215-227, 1974.
27. Clay, JTR., and Rf G? Ross ing. Histopathology of exposure to phosgene. An
attempt to produce pulmonary emphysema experimentally. Arch. Path. 78:
544-551, 1964.
28. Coman, t).~ R.", H," D.~ Bruner, R." C." Horn, Jr., M." Friedman, RT D." Boehe, M. I).
McCarthy, M? Gibbon, and J? Schultz. Studies on experimental phosgene
poisoning: I. The pathologic anatomy of phosgene poisoning with special
reference to the early and late phases. Amer. J. Pathol. 23:1037-1073,
1947.
29. Comroe, J. H., Jr. Physiology of Respiration. An Introductory Text.
Chicago: Yearbook Medical Publisher, Inc., 1965. p. 27
29a. Corn, M., N. Kotsko, and D. Stanton. Mass transfer coefficient for sulfur
dioxide and nitrogen dioxide removal in cat upper respiratory tract,
pp. .In Proceedings of the 4th International Symposium on
Inhaled Particles and Vapors, British Occupational Hygiene Society,
Edinburgh, September, 1975.
30. Curtttnirtg, GT, J." Crank, Jf Horsfield, and l.~ Parker. Gaseous diffusion in the
airways of the human lung. Respir. Physiol. 1:58-74, 1966.
7-58
-------
31. Dalhamn, T. Mucous flow and ciliary activity in the trachea of healthy
rats exposed to respiratory irritant gases (S02, H3N, and HCHO).
A functional and morphologic (light microscopic and electron micro-
scopic) study with special reference to technique. Acta Physiol.
Scand. 36(Suppl. 123):1-161, 1956.
32. Dalhamn, T. A method for studying the effect of gases and dusts on ciliary
activity in living animals, pp. 315-319. In C. N. Davies, Ed. Inhaled
Particles and Vapours. Proceedings of an International Symposium organ-
ized by the British Occupational Hygiene Society, 1960. New York: Per-
gamon Press, 1961.
33. Davies, C. N. Filtrations of droplets in the nose of the rabbit. Proc.
R. Soc. Lond. Ser. B 133:282-299, 1946.
34. Davies, C. N. A formalized anatomy of the human respiratory tract, pp. 82-
91. In C. N. Davies, Eds. Inhaled Particles and Vapours. Proceedings
of an International Symposium Organized by the British Occupation
Hygiene Society, 1960. New York: Pergamon Press, 1961.
35. Davies, C. N. An algebraical model for the deposition of aerosols in the
human respiratory tract during steady breathing. J. Aerosol Sci. 3:
297-306, 1972.
36. Dekker, E. Transition between laminar and turbulent flow in human trachea.
J. Appl. Physiol. 16:1060-1064, 1961.
37. DuBois, A. B., And R. M. Rogers. Respiratory factors determining the tissue
concentrations of inhaled toxic substances. Respir. Physiol. 5:34-52,
1968.
38. Dungworth, D. L. , W. L. Castleman, C. K. Chow, P. U. Mellick, M. G. Mustafa,
B. Tarkington, and W. S. Tyler. Effect of ambient levels of ozone on
monkeys. Fed. Proc. 34:1670-1674, 1975.
7-59
-------
39. Durlacher, S. H., and H. Bunting. Pulmonary changes following exposure to
phosgene. Amer. J. Path. 23:679-693, 1947.
40. Egle, J. L., Jr. Single-breath retention of acetaldehyde in man. Arch.
Environ. Health 23:427-433, 1971.
41. Egle, JfL., Jr. Retention of inhaled acetone and ammonia in the dog. Amer.
Ind. Hyg. Assoc. J. 34:533-539, 1973.
42. Fairchild, E. J., and S. L. Graham. Thyroid influence on the toxicity of
respiratory irritant gases, ozone and nitrogen dioxide. J. Pharmacol.
Exp. Ther. 139:177-184, 1963.
H
^•^ Findeisen, W. Uber das Absetzen kleiner, in der Luft suspendierter Teilchen
in der menschlichen Lunge bei der Atmung. Pfluegers Arch. 236:367-379,
1935.
44. Folinsbee, L. J. , F. Silverman, and R. J. Shephard. Exercise responses follow-
ing ozone exposure. J. Appl. Physiol. 38:996-1001, 1975.
45 ^ Frank, NT R. Studies on the effects of acute exposure to sulphur dioxide in
human subjects. Proc. R. Soc. Med. 57:1029-1033, 1964.
46. Frank, N." R., R.~ Ef Voder, E. Yokoyama, and F.~ E." Speizer. The diffusion of
35
S0? from tissue fluids into the lungs following exposure of dogs to
35S02- Health Phys. 13:31-38, 1967.
— 35
47. Frank, N.~"R., RT E." Yoder;> J." DT Brain, and E. Yokoyama. S02 ( S labeled)
absorption by the nose and mouth under conditions of varying concentration
and flow. Arch. Environ. Health 18:315-322, 1969.
4g Fry, F. A. , and A. Black. Regional deposition and clearance of particles in
the human nose. J. Aerosol Sci. 4:113-124, 1973.
49. Granite, S. M. Calculated Retention of Aerosol Particles in the Rat Lung.
M. S. Thesis. Chicago: University of Chicago, 1971. (UNVERIFIED)
7-60
-------
50. Gross, P., W. E. Rinehart, and T. Hatch. Chronic pneumonitis caused by phos-
gene. An experimental study. Arch. Environ. Health 10:768-775, 1965.
51. Gross, P., W. E. Rinehart, and R. T. P. deTreville. The pulmonary reactions
to toxic gases. Amer. Ind. Hyg. Assoc. J. 28:315-321, 1967.
52. Gunnison, AT F., and A. W7 Benton. Sulfur dioxide; Sulfite. Interaction with
mammalian serum and plasma. Arch. Environ. Health 22:381-388, 1971.
53. Gunnison, A. P., and E. D. Palmes. Persistence of plasma S-sulfonates follow-
ing exposure of rabbits to sulfite and sulfur dioxide. Toxicol. Appl.
Pharmacol. 24:266-278, 1973.
54. Gunnison, A. P., and E. D7 Palmes. S-sulfonates in human plasma following in-
halation of sulfur dioxide. Amer. Ind. Hyg. Assoc. J. 35:288-291, 1974.
55. Haggard, H. W. Action of irritant gases upon the respiratory tract. J. Ind.
Hyg. 5:390-398, 1924.
56. Ham, A. W. The respiratory system, pp. 744-771. In Histology. (5th ed.)
Philadelphia: J. B. Lippincott Co., 1965.
57. Haque, AT Kf M." V.7, and AT J." 17 Collinson. Radiation dose to the respiratory
system due to radon and its daughter products. Health Phys. 13:431-443,
1967.
58. Hatch, T. P., and P. Gross. Pulmonary Deposition and Retention of Inhaled
Aerosols. New York: Academic Press, 1964. pp. 12, 74, 82, 88.
59. Hatch, Tf F7 Significant dimensions of the dose-response relationship. Arch.
Environ. Health 16:571-578, 1968.
60. Hausknecht, D. P., and R. A. Ziskind. Definition of an Analytic Framework
for Collating Experiments on the Effects of Sulfur Oxides on the Lung.
Final Report by Science Applications, Inc. (Contract RP-205-2) Pre-
pared for the Electric Power Research Institute, Palo Alto, California,
February 28, 1975. (UNVERIFIED)
7-61
-------
61. Hayek, H. von. The "Human Lung. (Translated by V. E. Krahl) New York:
Hafner Publishing Company, Inc., 1960. 372 pp.
62. Henderson, Y., and H. H. Haggard. Noxious Gases and the Principles of Respir-
ation Influencing Their Action. New York: Reinhold Publishing Corpora-
tion, 1943. 294 pp.
63. Hilding, A. C. Ciliary streaming in the lower respiratory tract. Amer. J.
Physiol. 191:404-410, 1957.
64, Hilding, A. C. Ciliary streaming through the larynx and trachea. Relation
to direction of ciliary beat and significance in sites of respiratory
disease. J. Thorac. Surg. 37:108-117, 1959.
65. Hilding, AT C. Phagocytosis, mucous flow, and ciliary action. Arch. Environ.
Health 6:61-73, 1963.
- - (
65a. Hoigne^ J. > and H. Bader. Ozonation of water: Role of hydroxyl radicals
as oxidizing intermediates. Science 190:782-784, 1975.
66. Horsfield, K. The relation between structure and function in the airways of
the lung. Brit. J. Dis. Chest 68:145-160, 1974.
67. Horsfield, K. , G. Dart, D. E. Olson, G. F. Filley, and G. Gumming. Models
of the human bronchial tree. J. Appl. Physiol. 31:207-217, 1971.
67a. Hounaiu, R. F. The deposition of atmospheric condensation nuclei in the
nasopharyngeal region of the human respiratory tract. Health Phys.
20:219-220, 1971. (letter)
68 Ichioka, M. Model experiments on absorbability of the airway mucous mem-
brane of S0? and N02 gases. Bull. Tokyo Med. Dent. Univ. 19:361-375,
1972. (UNVERIFIED)
68a. Washburn, E. W., Ed. National Research Council of the U. S. A. Inter-
national Critical Tables of Numerical Data, Physics, Chemistry and
Technology. Vol. 3. New York: McGraw Hill, 1928. 444 pp.
7-62
-------
69. Iravani, J. Clearance function of the respiratory ciliated epithelium in
normal and brochitic rats, pp. 143-148. In W. H. Walton, Ed. Inhaled
Particles III. Vol. 1. Proceedings of an International Symposium
organized by the British Occupational Hygiene Society, 1970. Surrey:
Unwin Brothers Limited, 1970.
70. Iravani, J., and A. van As. Mucus transport in the tracheobronchial tree of
normal and bronchitic rats. J.~ Path. 106:81-93, 1972.
71. Jacobi, W. The dose to the human respiratory tract by inhalation of short-
222 ??n
lived Rn- and Rn-decay products. Health Phys. 10:1163-1174, 1964.
72. Jaffrin, M. Y. , and T. V. Hennessey, Jr. Pressure distribution in a model of
the central airways for sinusoidal flow. Bull. Physiopath. Respir. 8:
375-390, 1972.
73. Kilburn, K. H., and R. Rylander. An interpretive summary of the Bermuda
worshop. Arch. Intern. Med. 126:508-511, 1970.
74. Kirk, R. E., and D. F. Othmer, Eds. Encyclopedia of Chemical Technology.
Vol. 9. .Metal Surface Treatment to Penicillin. New York: The Inter-
science Encyclopedia, Inc., 1952. 943 pp.
75. Kliment, V., J. Libich, and V. Kaudersova". Geometry of guinea pig respira-
tory tract and application of Landahl's model of deposition of aerosol
particles. J. Hyg. Epidemiol. Microbiol. Immunol. 16:107-114, 1972.
76. Kliment, V. Similarity and dimensional analysis, evaluation of aerosol depo-
sition in the lungs of laboratory animals and man. Folia Morphol. 21:
59-64, 1973.
77. Kliment, V. Dichotomical model of respiratory airways of the rabbit and
its significance for the construction of deposition models. Folia
Morphol. 22:286-290, 1974.
7-63
-------
78. Kowitz, T. A., R. C. Reba, R. T. Parker, and W. S. Spicer, Jr. Effects of
chlorine gas upon respiratory function. Arch. Environ. Health 14:545-
558, 1967.
79. Kulle, T. J., and G. P. Cooper. Effects of formaldehyde and ozone on the
trigeminal nasal sensory system. Arch. Environ. Health 30:237-243, 1975,
80. La Belle, C. W., J. E. Long, and E. E. Christofano. Synergistic effects of
aerosols. Particulates as carriers of toxic vapors. A.M.A. Arch. Ind.
Health 11:297-304, 1955.
g^ La Force, R. C. , and B. M. Lewis. Diffusional transport in the human lung.
J. Appl. Physiol. 28:291-298, 1970.
82. Lakin, M. B. Oscillatory Flow Simulation in an idealized Bifurcation.
Ph.D. Thesis. Denver: University of Denver, 1973. 223 pp.
83. Lakin, M. B. , and V. G. Fox. Transient flow characteristics in an ideal-
ized bronchial bifurcation. Respir. Physiol. 21:101-117, 1974.
84. Landahl, tt. D. On the removal of air-borne droplets by the human respiratory
tract. Il7 The nasal passages. Bull. Math. Biophys. 12:161-169, 1950.
85. Landahl, H.~ D. On the removal of air-borne droplets by the human respiratory
tract: 17 The lung. Bull. Math. Biophys. 12:43-56, 1950.
86. Langhaar, tt. L. Steady flow in the transition length of a straight tube.
J. Appl. Mech. 9:A55-A58, 1942.
86a. Lapp, K. L. , J. L. Hankinson, H. Amandus, and E. D. Palmes. Variablitity
in the size of airspaces in normal human lungs as estimated by
aerosols. Thorax 30:293-299, 1975.
87> Levich, V. G. Physicochemical Hydrodynamics. Englewood Cliffs, N. J.:
Prentice Hall, 1962. 700 pp. (Translated by Scripta Technica, Inc.,)
7-64
-------
88. Lewis, W. K., and W. G. Whitman. Principles of gas adsorption. Ind. Eng.
Chem. 16:1215-1220, 1924.
89a. Lin, K. H. , and G. Gumming. A model of time-varying gas exchange in the
human lung during a respiratory cycle at rest. Respir. Physiol. 17:
93-112, 1973.
89b. Loring, 5. H., and §. M. Tenney. Gas absorption form frontal sinuses.
Arch. Otolaryngol. 97:470-474, 1973.
90. National Research Council. Contmittee on Medical and Biologic Effects of
Environmental Pollutants. Airborne Particles. Washington, D. C. :
National Academy of Sciences, (in press)
91 Macklem, P., and K. H. Kilbum. Tracheobronchial response to insult, pp. 31-
35. In D. H. K. Lee, Ed. Environmental Factors in Respiratory Disease.
New York: Academic Press, 1972.
91a. Matsuba, K. , and W. M. Thurlbeck. The number and dimensions of small
airways in nonemphysematous lungs. Amer. Rev. Resp. Dis. 104:516-
524, 1971.
92. McJilton, C., J. Thielke, and R. Frank. Ozone uptake model for the respira-
tory system. Paper No. 45. In Abstracts of Technical Papers. American
Industrial Hygiene Conference, May 14-19, 1972.
93> Mezentseva, M. V., E. A. Melnikova, and 6. Ya. Mogilevskaya. Titanium, pp.
35-44. In Z. I. Izrael'son, Ed. Toxicology of the Rare Metals.
Jerusalem: Israel Program for Scientific Translations, Ltd., 1967.
(Translated by Y. Halperin, edited by E. Lieber)
94. Moorman, W. JT, J. J. Chmiel, and j.~ F.~ Stara. Comparative decomposition of
ozone in the nasopharynx of beagles. Arch. Environ. Health 26:153-155,
1973.
7-65
-------
95. Morrow, P. E. Airborne contaminants, pp. 71-90. In D. H. K. Lee, Ed.
Environmental Factors in Respiratory Disease. New York: Academic
Press, 1972.
96. Washburn, E. W., Ed. National Research Council of the U. S. A. Interna-
tional Critical Tables of Numerical Data, Physics, Chemistry and
Technology. Vol. 3. New York: McGraw Hill, 1928. 444 pp.
97. Olson, D. E., M. F. Sudlcw, K. Horsfield, and G. F. Filley. Convective pat-
terns of flow during, inspiration. Arch. Intern. Med. 131:51-57, 1973.
98. Owen, P. R. Turbulent flow and particle deposition in the trachea, pp. 236-
252. In G. E. W. Wolstenholme and J. Knight, Eds. Cirulatory and Respir-
atory Mass Transport. A CIBA Foundation Symposium. Boston: Little,
Brown and Co., 1969.
98a. Palmes, E. D., and M. Lippmann. Influence of respiratory air space dimen-
sions on aerosol deposition, pp. . In Proceedings of the 4th
International Symposium on Inhaled Particles and Vapors, British Occupa-
tional Hygiene Society, Edinburgh, Sept. 1975.
99. Papper, E. M., and R. J. Kitz, Eds. Uptake and Distribution of Anesthetic
Agents. New York: McGraw-Hill Book Co., 1963. 321 pp.
100,, Pattle, R. E. The retention of gases and particles in the human nose, pp.
302-311. In C. N. Davies, Ed. Inhaled Particles and Vapours. Pro-
ceedings of an International Symposium organized by the British Occupa-
tional Hygiene Society, Oxford, 1960. Oxford: Pergamon Press, 1961.
101. Pedley, T. J. , R. C. Schroter, and M. F. Sudlow. Flow and pressure drop
in systems of repeatedly branching tubes. J. Fluid Mech. 46:365-
383, 1971.
7-66
-------
102. Emmert, R. E., and R. I. Pigrdrd. Mass-transfer fundamentals, pp. 14-13--
14-19. In R. H. Perry, C. H. Chilton and S. D. Kirkpatrick, Eds.
Chemical Engineers' Handbook. (4th ed.) New York: McGraw-Hill Book
Company, 1963.
103. Phalen, R. F. Summary of a Respiratory Tract Morphology Conference.
Lovelace Foundation for Medical Education and Research, Albuquerque,
New Mexico. Report LF-47, June 1974.
104. Phalen, R. F. , H. C. Yeh, and D. J. Velasquez. Bronchial tree structure in
the human, beagle, rat, and hamster, LF-49, pp. 289-292. In 1973-1974
Annual Report of Inhalation Toxicology Research Institute. Lovelace
Foundation for Medical Education and Research, 1974.
105. Proctor, D. F. Physiology of the upper airway, pp. 309-345. In W. 0. Fenn
and H. Rahn, Eds. Handbook of Physiology. Section 3: Respiration. Vol.
1. Washington, D. C.: American Physiological Society, 1964.
106. Proctor, D.~ F., and H? N.~ Wagner. Clearance of particles from the human nose.
Preliminary report. Arch. Environ. Health 11:366-371, 1965.
107. Proctor, D. F., and D. L. Swift. The nose--a defense against the atmospheric
environment, pp. 59-70. In W. H. Walton, Ed. Inhaled Particles III.
Proceedings of an International Symposium organized by the British Occupa-
tional Hygiene Society in London, 1970. Surrey: Unwin Brothers Limited,
1971.
107a. --deleted
107b. Proctor, D. F. , I. Andersen, and G. Lundqvist. Clearance of inhaled
particles from the human nose. Arch. Intern. Med. 131:132-139, 1973.
108. Quinlan, M. F., S. D. Salman, D. L. Swift, H. N. Wagner, Jr., and D. F.
Proctor. Measurement of mucociliary function in man. Amer. Rev.
Respir. Dis. 99:13-23, 1969.
7-67
-------
109 deleted—omitted
109a. Santa Cruz, R. , J. F. Landa, J. Hirsch, and M. A. Sackner. Tracheal
mucous velocity in normal man and patients with obstructive lung
disease; effects of terbutaline. Amer. Rev. Respir. Dis. 109:458-
463, 1974.
110,. Schlichting, H. Boundary-Layer Theory. (6th ed.) New York: McGraw-Hill,
1968. 747 pp. (translated by J. Keatin)
111. Schreck, R. M., and L. P. Mockros. Fluid Dynamics in the Upper Pulmonary
Airways. American Institute of Aeronautics and Astronautics Paper
No. 70, 1970. p. 788.
112,, Schroter, R. C. , and M. R. Sudlow. Flow patterns in models of the human
bronchial airways. Respir. Physiol. 7:341-355, 1969.
113. Schwartz, L. W., It. S. Tyler, and D. L. Dungworth. Pulmonary changes induced
by ambient levels of ozone: A morphological study, pp. 367-383. In
Proceedings of the 5th Annual Conference on Environmental Toxicology.
Wright-Patterson Air Force Base, Ohio, Sept. 24-26, 1974. AMRL-TR-74-125.
Wright-Patterson AFB, Ohio: Aerospace Medical Research Laboratory, 1974.
114. Seban, R. A., and E. F. McLaughlin. Heat transfer in tube coils with
laminar and turbulent flow. Int. J. Heat Mass Transfer 6:387-395,
1963.
115. Silverman, L." G." Lee, T." Plotkin, L.~ A." Sawyers, and A. R. Yancey. Air flaw
measurements on human subjects with and without respiratory resistance at
several work rates. A.M.A.. Arch. Ind. Myg. Occup. Med. 3:461-478, 1951.
116. Speizer, F."E., and N.~ 'RT Frank. The uptake and release of SO by the human
ncse. Arch. Environ. Health 12:725-728, 1966.
7-68
-------
117. Spiegelman, J. R., G. D. Hanson, A, Lazarus, RT J. Bennett, M. Lippmann, and
R7 A7 Albert. Effect of acute sulfur dioxide exposure on bronchial clear-
ance in the donkey. Arch. Environ. Health 17:321-326, 1968.
117a. Stephens, R. J. , M. F. Sloan, M. J. Evans, and G. Freeman. Early response
of lung to low levels of ozone. Amer. J. Path. 74:31-44, 1974.
117b. Stevens, R. J. , M. F. Sloan, M. J. Evans, and G. Freeman. Alveolor type 1
cell response to exposure to 0.5 ppm 0- for short periods. Exp. Molec.
Path. 20:11-23, 1974. (UNVERIFIED)
118. Stokinger, H.~ E.~, W.~ t>T Wagner, and P." G.~ Wright. Studies of ozone toxicity.
I. Potentiating effects of exercise and tolerance development. A.M.A.
Arch. Ind. Health 14:158-162, 1956.
119a. Stokinger, HT E. Evaluation of the hazards of ozone and oxides of nitrogen.
Factors modifying toxicity. A.M.A. Arch. Ind. Health 15:181-190, 1957.
119b. Sullivan R. J. Preliminary Air Pollution Survey of Arsenic and Its Com-
pounds. A Literature Review. National Air Pollution Control Admin-
istration Publication No. APTC 69-26. Bethesda, Md.: Litton Systems,
Inc., 1969. 60 pp.
119c. Sullivan, R. J. Preliminary Air Pollution Survey of Chromium and Its Com-
pounds. A Literature Review. National Air Pollution Control Adminis-
tration Publication No. APTD 69-34. Bethesda, Md.: Litton Systems,
Inc., 1969. 75 pp.
119d. Sullivan, R. J. Preliminary Air Pollution Survey of Iron and Its Compounds.
A Literature Review. National Air Pollution Control Administration
Publication No. 69-38. Bethesdas Md. : Litton Systems, Inc., 1969.
94 pp.
7-69
-------
119e. Sullivan, R. J. Preliminary Air Pollution Survey of Nickel and Its Com-
pounds. A Literature Review. National Air Pollution Control Admin-
istration Publication No. 69-41. Bethesda, Md.: Litton Systems,
Inc., 1969. 69 pp.
119f. Sullivan, R. J. Preliminary Air Pollution Survey of Manganese and Its Com-
pounds. A Literature Review. National Air Pollution Control Adminis-
tration Publication No. 69-39. Bethesda, Md.: Litton Systems, Inc.,
1969. 54 pp.
119g. Task Group on Lung Dynamics. Deposition and retention models for internal
dosimetry of the human respiratory tract. Health Phys. 12:173-207, 1966.
120. Taulbee, D.~ Bk\ and C," PT Yu. A theory of aerosol deposition in the human res-
piratory tract. J. Appl. Physiol. 38:77-85, 1975.
121. Thomas, J. A method for calculation of the absorbed dose to the epithelium
of the respiratory tract after inhalation of daughter products of radon.
Ann. Occup. Hyg. 7:271-284, 1964.
122 Thorp, C. E. Bibliography of Ozone Technology. Vol. 2. Physical and
Pharmacological Properties. Chicago: Armour Research Foundation,
Illinois Institute of Technology, 1955. (UNVERIFIED)
122a. Thurlbeck, W. M., and J. R. Haines. Bronchial dimension and stature.
Amer. Rev. Respir. Dis. 112:142-145, 1975. (UNVERIFIED)
123. tf- s- Department of Health, Education, and Welfare. Public Health Service.
Air Quality Criteria for Particulate Matter. National Air Pollution
Control Administration Publ. No. AP-49. Washington, D. C.: U. S. Govern-
ment Printing Office, 1969. 211 pp.
7-70
-------
|