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

-------