DEVELOPMENT OF OZONE REACTIVITY SCALES FOR VOLATILE ORGANIC COMPOUNDS

Methods for ranking photochemical ozone formation reactivities of volatile organic compounds (VOCs) are discussed. Photochemical mechanisms for the atmospheric reactions of 118 VOCs were used to calculate their effects on ozone formation under various NOx conditions in model scenarios representing 39 different urban areas. Their effects on ozone were used to derive 18 different ozone reactivity scales, one of which is the Maximum Incremental Reactivity (MIR) scale used in the new California Low Emission Vehicle and Clean Fuel Regulations. These scales are based on 3 different methods for quantifying ozone impacts and on 6 different approaches for dealing with the dependencies of reactivity on NOx. The predictions of the scales are compared, the reasons for their similarities and differences are discussed, and the sensitivities of the scales to NOx and other scenario conditions are examined. Scales based on peak ozone levels were highly dependent on NOx, but those based on integrated ozone were less sensitive to NOx and tended to be similar to the MIR scale. It is concluded that the MIR scale or one based on integrated ozone is appropriate for applications requiring use of a single reactivity scale.

The effects of 26 individual volatile organic compounds (VOCs) on ozone formation, NO oxidation, and OH radical levels were measured by adding them to reactive organic gas (ROG) - NOx - air environmental chamber irradiations representing a simplified model photochemical smog system. These experiments had relatively low ROG/NOx ratios to represent conditions where ozone formation is most sensitive to VOC additions. The compounds studied included representative alkanes, alkenes, aromatic hydrocarbons, aldehydes, alcohols, and CO. The addition of formaldehyde, methylbenzenes, alkenes and methanol all caused increased integrated OH radical concentrations, and caused the most NO oxidation and ozone formation per molecule reacted. The C6+ n-alkanes had the most inhibiting effects on OH radicals, and caused reduced NO oxidation and ozone formation in these experiments. The other compounds had smaller negative effects on OH radicals, but moderate positive effects on ozone formed and NO oxidized. The implications of these results in terms of the atmospheric reaction mechanisms of these compounds are discussed.

A detailed atmospheric photochemical mechanism which had been previously used in model calculations for developing ozone reactivity scales for volatile organic compounds (VOCs) was evaluated by comparing its predictions with measurements of incremental reactivities in an environmental chamber system. An updated version of this mechanism is also described and evaluated. The experiments consisted of determining the effects of adding representative alkanes, alkenes, aromatic hydrocarbons, aldehydes or CO on NO oxidation, ozone formation and radical levels in a simplified model photochemical smog system representing conditions where ozone formation is most sensitive to VOCs. The published mechanism correctly simulated the observed qualitative reactivity trends, but overpredicted the effect of adding formaldehyde early in the experiments, performed poorly in simulating reactivities of branched alkanes, tended to underpredict the reactivities of alkenes, and did not simulate differences in reactivities of aromatic isomers. The updates to the mechanism improved the simulation results for the branched alkanes and the alkanes, but not for formaldehyde and the aromatics. The implications of these results concerning the development of atmospheric mechanisms for VOCs are discussed.

A detailed atmospheric photochemical mechanism for the atmospheric reactions of isoprene its major oxidation products in the presence of NOx, which incorporates the most recent laboratory results and our current understanding of the system, is described. It is evaluated by comparing its predictions against results of NOx-air irradiations of isoprene and its two major products, methacrolein, and methyl vinyl ketone (MVK), in five different types of environmental chambers at two different laboratories. In most cases it simulated experimental results within the uncertainty of the data and the chamber and run characterization model. However, the photodecomposition quantum yields of methacrolein and MVK and the organic nitrate yield from the OH + isoprene reaction had to be adjusted to obtain satisfactory simulations of the data. The major discrepancy observed was that the model tended to underpredict PAN by ~40% in the isoprene experiments, despite the fact that the model predicted PAN from methacrolein and MVK reasonably well. The uncertainties and additional data needed to completely characterize the isoprene atmospheric photooxidation system are discussed.

Two condensed mechanisms for the atmospheric reactions of isoprene, which differ in the number of species used to represent isoprene's reactive products, have been developed for use in ambient air quality modeling. They are based on a detailed isoprene mechanism that has recently been developed and extensively evaluated against environmental chamber data. The new condensed mechanisms give very close predictions to those of the detailed mechanism for ozone, OH radicals, nitric acid, H2O2, formaldehyde, total PANs, and for incremental effects of isoprene on for ozone formation in one day simulations. The effects of the condensations become somewhat greater in multi-day simulations, particularly in cases where NO3 reactions are important at nighttime, but the ozone predictions are still very close. On the other hand, the SAPRC-90, RADM-2, and Carbon Bond IV isoprene mechanisms give quite different predictions of these quantities. It is recommended that the new mechanisms replace those currently used in airshed simulations where isoprene emissions are important.

Rate constants for the gas phase reactions of O(3P) atom with a series of monoterpenes have been determined at ambient temperature (~302-309K) and atmospheric pressure using a relative rate technique. Using the literature rate constants for O(3P) + isobutene, cis and trans-2-butene, 3-methyl-1-butene, 2-methyl-2-butene and 2,3-dimethyl-2-butene as the standards, the O(3P) rate constants derived for the terpenes (in units of 10^-11 cm^3 molecule^-1 s^-1) are 2.8+/-0.4 for alpha-pinene, 2.8+/-0.5 for beat-pinene, 3.1+/-0.5 for delta-3-carene, 3.5+/-0.5 for 2- carene, 2.6+/-0.5 for camphene, 7.6+/-1.2 for d-limonene, 9.0+/-1.6 for gamma-terpinene and 10.7+/-1.6 for terpinolene. The relative rate constants in this work agreed with literature values to within +/-10% for the standard alkenes, and to within +/-~35% for the terpenes.

This report documents software which has been developed for the preparation of chemical mechanisms and the processing of emissions input data for use in airshed modeling applications, and which has subsequently implemented into the CALGRID model developed for the California ARB. Although the programs are now somewhat out-of-date and are superseded by more recent work of Lurmann and co-workers (see below), it still provides the best available description of certain features of the SAPRC modeling software.

See Lurmann, F. W., M. Gery, and W. P. L. Carter (1991): "Implementation of the 1990 SAPRC Chemical Mechanism in the Urban Airshed Model," Final Report to the California South Coast Air Quality Management District, Sonoma Technology, Inc. Report STI- 99290-1164-FR, Santa Rosa, CA.

Also, a separate report describing more recent updates, entitled "Development of the Flexible Chemical Mechanism Version of the Urban Airshed Model, by N. Kumar and F. W. Lurmann of Sonoma Technology, Inc., is now available from the California Air Resources Board.

This report describes a method for analyzing the impacts on ozone formation of emissions of CO and volatile organic compounds (VOCs) from motor vehicles, given the results of detailed speciated analyses of exhaust and evaporative emissions. The analysis is based on estimating maximum amounts of additional ozone formed caused by CO and VOC emissions from the motor vehicles in idealized scenarios representing photochemical smog formation. The ozone impacts are given as grams of ozone formed per gram exhaust or evaporative VOC emitted, and as grams of ozone formed per vehicle mile traveled. The method is illustrated using results of analyses recently carried out by the California Air Resources Board of emissions from several motor vehicles utilizing gasoline and various alternative fuels. Contributions of the various individual compounds to the overall reactivity of the emissions are calculated, and the relative reactivities of the emissions from these vehicles calculated in this work are compared with results of analyses based only on considerations of the OH radical rate constants of the individual compounds. General issues involved in the estimation of ozone reactivities of VOC emissions from motor vehicles and other sources are briefly discussed.

Note: This is the first distributed report where it was proposed to use a maximum reactivity scale for alternative fuel applications. It has always been referenced as "draft", but the draft is the final version. The word "Draft" was deleted from the cover page, but the date is the same, indicating it was not changed.

A series of environmental chamber experiments were conducted to measure the effects of isoprene on ozone formation, NO oxidation, and OH radical levels in a simplified model photochemical smog system. The experiments consisted of repeated 6-hour irradiations of a simplified mixture of smog precursors, alternating with runs with varying amounts of isoprene added. The experiments were conducted at relatively low ROG/NOx ratios to simulate conditions where VOCs have the greatest effect on ozone formation, and were carried out in conjunction with a larger program where similar data was obtained for 35 other types of VOCs. The amount of ozone formed and NO oxidized per isoprene reacted increased with reaction time, being approximately two molecules of ozone formed and NO oxidized per molecule of isoprene reacted in the first hour, and approximately four molecules in six hours, under the conditions of these experiments. Approximately half of this is estimated to be due directly to the reactions of isoprene and its oxidation products, while the other half is estimated to be due to the fact that isoprene increases the radical levels present in the system, causing additional ozone formation from the other VOCs present. Current atmospheric chemical mechanisms for isoprene, including a preliminary detailed isoprene mechanism, could not correctly simulate the results of these experiments.

The effects of 36 representative volatile organic compounds (VOCs) and CO on ozone formation, NO oxidation, and OH radical levels were measured in a series of environmental chamber experiments representing conditions where VOCs have the greatest effect on photochemical ozone formation. The experiments consisted of repeated 6-hour indoor chamber irradiations of a simplified mixture of ozone precursors with NOx in excess, alternating with runs with varying amounts of a test VOC added. The VOCs studied included representative alkanes, alkenes, aromatic hydrocarbons, aldehydes, alcohols, ethers, alcohol ethers, and siloxanes. CO, Acetone and 2-chloromethyl-3-chloropropene were also studied. Reactions of formaldehyde, acetone, the methylbenzenes and the alkenes had the largest positive effects on OH radical levels, and because of this they caused the most NO oxidation and ozone formation per molecule reacted. Reactions of the siloxanes and the C6+ n-alkanes had the most inhibiting effects on radicals, causing them to inhibit NO oxidation and ozone formation under the conditions of these experiments. The other compounds had smaller and usually negative effects on OH radicals, and had moderate but positive effects on ozone formed and NO oxidized. Information was also obtained on amounts of NO oxidation caused directly by the reactions of the added VOCs or their products.

The results are compared with model calculations using a detailed atmospheric photochemical reaction mechanism, and new or refined mechanisms for isobutene, isooctane, MTBE and alcohol ethers were developed. The model fit the data to within the experimental uncertainties for approximately half the VOCs, and generally predicted the observed qualitative reactivity trends. However, the results indicate that refinements to the mechanisms for alkenes and aromatics are needed. Additional data needs, some of which we will be addressing in the next phase of our ongoing studies, are discussed.

A series of environmental chamber experiments and computer model simulations were carried out to assess the tendency of acetone to promote ozone formation in photochemical smog. The experiments consisted of NOx - air photolysis of acetone by itself, and determinations of the effect of adding acetone on ozone formation in model photochemical smog systems. Indoor chambers using either fluorescent blacklight or xenon arc light sources and an outdoor chamber utilizing sunlight were employed. Similar experiments utilizing acetaldehyde were carried out for comparison and control purposes. The gas-phase photochemical mechanism for the atmospheric reactions of acetone was updated and was evaluated by model simulations of the results of these experiments. The mechanism was found to overpredict the effect of acetone on ozone formation and radical levels in the indoor chamber experiments with the blacklight light source and in some of the outdoor chamber runs, but fit the results of other outdoor chamber runs and the experiments runs using the xenon arc light source reasonably well. An adjusted mechanism which gave better agreement with the blacklight experiments and with some of the outdoor runs was developed.

The updated and adjusted acetone mechanisms were then used in model calculations to assess the effects of acetone on ozone formation under atmospheric conditions. This was done by calculating its incremental reactivity (defined as amount of additional ozone formed caused by adding acetone to the emissions, divided by the amount added) in model scenarios representing ozone episodes in 39 urban areas around the United States. The incremental reactivities of ethane and a mixture representing total emissions reactive organic gases from all sources were also calculated for comparison. The results indicate that acetone forms 10-15% as much ozone on a per mass basis as total ROG emissions, while ethane forms 6-20% as much ozone, depending on conditions. The implications of these results on the question of whether acetone should be exempt from regulation as an ozone precursor are discussed.

This report describes a calculation of the Maximum Incremental Reactivity (MIR) and the Maximum Ozone Incremental Reactivity (MOIR) scales using an updated version of the Carbon Bond IV chemical mechanism which is being used in the air quality modeling in Phase II of the Auto/Oil Air Quality Improvement Research Program. Except for the mechanism, the methodology and model scenarios employed are essentially the same as used to calculate the MIR and MOIR scales used in the California Air Resources Board "Clean Fuels/Low Emissions Vehicle" regulations.

Note that the development of the MIR scale using the SAPRC-90 mechanism is documented in the published paper: W. P. L. Carter, "Development of Ozone Reactivity Scales for Volatile Organic Compounds," J. Air and Waste Manage. Assoc., 44, 881-899, 1994, and in a draft ARB report which has not yet been released. Until the release of the ARB report, this Auto/Oil report serves as the most complete documentation of the details of the reactivity calculation methodology which are not given in the JA&WMA paper.

A series of indoor environmental chamber experiments were conducted to measure incremental reactivities of representative volatile organic compounds (VOCs) in irradiations of various reactive organic gas (ROG) surrogate - NOx - air mixtures designed to represent or approximate conditions of urban photochemical smog. Incremental reactivities are defined as the change in to ozone formation or OH radical levels caused by adding the VOC to a "base case" experiment, divided by the amount added. The base case included irradiations, at both relatively high and low NOx levels, of a surrogate mixture of 8 VOCs which model calculations predicted would yield the same results as use of a full ambient ROG mixture, and high NOx experiments where ethylene alone represented the ambient ROGs. The test VOCs included carbon monoxide, n-butane, n- hexane, n-octane, ethylene, propene, trans-2-butene, benzene, toluene, m-xylene, formaldehyde, and acetaldehyde. The data obtained show that VOC have a greater range of incremental reactivities when simplified base case ROG surrogates are used than with the more realistic 8-component surrogate. Reducing NOx reduced incremental reactivities by differing amounts for different VOCs, with ozone reactivities of propene, trans-2-butene, acetalde- hyde, and the aromatics becoming negative in the low NOx experiments. These results are consistent with model predictions. The model simulated reactivities in experiments with the more complex surrogate reasonably well, though it was more variable in the simulations of the simpler systems, which are more sensitive to differences among the VOCs. Model calculations indicated that experimentally measured incremental reactivities may correlate well with those in the atmosphere under high NOx conditions, but not when NOx is low. Thus the best use for data from incremental reactivity experiments is evaluating the models used to predict reactivities in the atmosphere.

This two report describes the data base of University of California, Riverside, environmental chamber experiments for use when evaluating photochemical mechanisms for urban and regional airshed models. This includes data obtained using the Statewide Air Pollution Research Center (SAPRC) Evacuable Chamber (EC), Indoor Teflon Chamber #1 (ITC), Indoor Teflon Chamber #2 (ETC), Dividable Teflon Chamber (DTC), and Xenon arc Teflon Chamber (XTC) between September of 1975 through November of 1993. This document provides backing information and data for that data set as well. This document lists and summarizes the experiments, summarizes the facility and procedures employed, documents the analytical and monitoring methods and their calibration data and associated uncertainties, assigns and documents the input data needed to conduct model simulations of the experiments in the present data base, and describes the format of the data sets which are distributed with this document on computer diskettes. Files are included in the distribution to permit modeling of the experiments in the present data base using the SAPRC-90 and the Carbon Bond IV chemical mechanisms, though a full mechanism evaluation procedure is beyond the scope of this report. Recommendations are made concerning the steps that need to be taken before using these data to evaluate chemical mechanisms.

This report consists of two volumes. Volume 1 contains the main body of the text documenting the data base, and Volume 2 contains the three appendices. Appendix A contains printouts of spreadsheets containing summaries of the runs in the data base. Appendix B contains tabulations of the NOx and GC calibration data, which are too lengthy to include in the main body of the report. Appendix C describes how to install the distributed data files and software on a computer and how to conduct initial model simulations of the runs using the SAPRC modeling software and the SAPRC-90 and Carbon Bond IV mechanisms.

An experimental and modeling study was conducted to assess how chemical mechanism evaluations using environmental chamber data are affected by the light source and other chamber characteristics. Xenon arc light lights appear to give the best artificial representation of sunlight currently available, and experiments were conducted in a new Teflon chamber constructed using such a light source. Experiments were also conducted in an Outdoor Teflon Chamber using new procedures to improve the light characterization, and in Teflon chambers using blacklights. These results, and results of previous runs other chambers, were compared with model predictions using an updated detailed chemical mechanism. The magnitude of the chamber radical source assumed when modeling the previous runs were found to be too high; this has implications in previous mechanism evaluations. Temperature dependencies of chamber effects can explain temperature dependencies in chamber experiments when T >300K, but not at temperatures below that. The model performance had no consistent dependence on light source for experiments not containing aromatics, but the model tended to underpredict O3 in the new xenon arc and blacklight chamber runs. This is despite the fact that such biases are not seen in modeling runs in the older xenon arc chamber or in preliminary modeling of University of North Carolina outdoor chamber runs. The reasons for this are not clear, and additional studies are planned as part of our ongoing program.

This quarterly report to an ongoing California Air Resources Program describes the reoptimization of the aromatics portions of the SAPRC detailed mechanisms to perform better in representing differences among aromatic isomers, and in simulations of new chamber experiments using a xenon arc light source. The updated mechanism is compared with the SAPRC-93 mechanisms in simulations of the chamber data and in predictions of atmospheric reactivities in the MIR and the MOIR ozone reactivity scales. The updates must be considered to be preliminary, and these results are made available primarily as a status report. The future schedule for this program is briefly summarized.

An experimental and modeling study was conducted to assess the atmospheric impacts of chloropicrin emissions. Chloropicrin absorption cross-sections were measured in the ~270-390 nm wavelength region, and its overall photodecomposition quantum yield under simulated sunlight conditions was found to be 0.87ñ0.26. In environmental chamber experiments, chloropicrin significantly enhanced rates of NO oxidation, O3 formation, and consumptions of alkanes and other organic reactants. This is attributed to the formation of Cl atoms and NOx in its photodecomposition. A previously developed atmospheric chemical mechanism was expanded to include chloropicrin and Cl atom reactions. It gave reasonably good simulations of the chamber experiments. This mechanism predicted that when emitted into polluted urban atmospheres, chloropicrin would have between 0.4 and 1.5 times the ozone impact of the average of emitted VOCs on a mass emitted basis. This value varied depending on environmental conditions and assumptions made concerning the photodecomposition mechanism. The data obtained in this study were inconsistent with a previous study of chloropicrin's photodecomposition in air, probably due to differences in the light sources employed.

A complete listing of the chemical mechanism used, which is not in the version submitted to Atmospheric Environment, is given in Appendix A.

A series of environmental chamber experiments and computer model calculations were carried out to asses the atmospheric ozone formation potential of methyl acetate. The experiments consisted of determining the effects of adding methyl acetate on NO oxidation, ozone formation and integrated OH radical levels in simulated model photochemical smog systems. Experiments were carried out using two different surrogate mixtures to represent the reactive organic gases (ROGs) present in the atmosphere, and using differing ROG/NOx ratios. It was found that methyl acetate has a small but positive effect on ozone formation, and that its it does not significantly enhance or inhibit OH radical levels. The results are well fit by model predictions if it assumes that methyl acetate reacts primarily to form a relatively unreactive product, with two NO to NO2 being involved before OH radicals are regenerated. This suggests a mechanism involving formation of a ùCH2O(CO)CH3 radical intermediate which primarily reacts with O2 to form HO2 and H(CO)O(CO)CH3, though the latter product could not be detected by the analytical techniques we employed. The alternative mechanism, that the ùCH2O(CO)CH3 radical decomposes to form formaldehyde and CO, significantly overpredicted the observed ozone impact of methyl acetate, and was inconsistent with the observation that the addition of methyl acetate had no significant effect on formaldehyde formation.

The experimentally-validated methyl acetate mechanism was then used to estimate its ozone impacts for a variety of atmospheric conditions. These were compared with ozone impacts calculated for ethane, the compound the EPA uses as the borderline for determining "negligible" ozone reactivity) and the mixture of all emitted VOCs. The results indicated that methyl acetate was approximately 1/3 to 1/2 as reactive as ethane, on an ozone per gram basis, with relatively little variation from scenario to scenario. Therefore, it is concluded that for regulatory purposes methyl acetate can be considered to have a lower ozone impact than that of ethane.

A series of environmental chamber experiments and computer model calculations were carried out to asses the atmospheric ozone formation potentials of the C12 through C16 n-alkanes. The experiments consisted of determining the effects of adding n-C12, n-C14, n-C15, or n-C16 on NO oxidation, ozone formation and integrated OH radical levels in simulated model photochemical smog systems. Either blacklights or xenon arc lamps were used as the light source, two different surrogate mixtures were used to represent the reactive organic gases (ROGs) present in the atmosphere, and the ROG/NOx ratio was also varied. The C12+ n-alkanes were found to inhibit integrated OH levels and initial rates of NO oxidation and ozone formation in all experiments, but the inhibiting effect on ozone decreased with time in the runs with the more realistic ROG surrogate and the lower NOx levels, and in some cases slightly higher levels of ozone were eventually formed by the end of the experiment.

An updated detailed gas-phase mechanism gave good fits to data in model simulations of the experiments with n-C12, but tended to underpredict the radical and ozone inhibition in some experiments with the C14+ n-alkanes. Somewhat better fits to those runs were obtained in a modified mechanism where ~20% higher alkyl nitrate yields, and thus more radical and NOx inhibition, was assumed. It is concluded that for the C14+ n-alkanes the likely atmospheric reactivities will be somewhere between those predicted by the standard mechanism and the mechanism assuming the ~20% higher nitrate yields.

These mechanisms were then used to estimate the ozone impacts of the higher n-alkanes for a variety of atmospheric conditions. These were compared with ozone impacts calculated for ethane (the compound the EPA uses as the borderline for determining "negligible" ozone reactivity) and the mixture of all emitted VOCs. The results indicated that, regardless of which mechanism was used, the relative ozone impacts of the C12+ n-alkanes were highly dependent on atmospheric conditions, and also on whether ozone impacts were measured by effects on peak ozone levels, or on integrated ozone over the standard. Under some conditions the C12+ n-alkanes inhibited ozone, while under others they formed up to 2 or 3 times more ozone than ethane, depending on how ozone was quantified. Overall, for a majority of the atmospheric conditions examined the C12+ n-alkanes were calculated to be slightly more reactive than ethane in terms of their effects on maximum ozone yields, and to be of comparable or lesser reactivity than ethane in terms of their effects on integrated ozone over the standard.

A series of indoor environmental chamber experiments were conducted to fill gaps in the data base needed for evaluation of gas-phase photochemical mechanisms for assessing the effects of emissions of volatile organic compounds (VOCs) on ambient air quality. Two large dual-mode indoor Teflon bag chambers, one irradiated by blacklights and the other by xenon arc light, were employed. Alternative methods for measuring light intensity in these chambers were evaluated.

It was found that quartz tube NO2 actinometry provides satisfactory data for the blacklight chamber, but that Cl2 - n-butane irradiations provide a better method for the xenon arc chamber. The effects of varying humidity on results and reproducibility of chamber experiments were examined. It was found that differences between dry runs and runs at ~50% RH were minor and should not significantly affect mechanism evaluation results, but that runs with humidities approaching 100% RH may have problems.

Incremental reactivity experiments, where a test compound is added to NOx-air irradiations of reactive organic gas (ROG) surrogates representing ambient pollution, were conducted for representative compounds using the xenon arc chamber. These were needed to supplement the much larger data base of incremental reactivity experiments in blacklight chambers. The results were consistent with model predictions and the large data base of incremental reactivity experiments in blacklight chambers, if the aromatics mechanisms was modified to account for differences in light source.

An extensive series of single aromatic - NOx experiments were carried out using both light sources to provide data needed to develop, adjust, and evaluate mechanisms for benzene, toluene, ethylbenzene, o-, m-, and p-xylenes, and all three trimethylbenzene isomers. The current version of the detailed SAPRC mechanism (SAPRC-93) did not correctly account for isomeric differences and tended to underpredict reactivities in the xenon arc chamber. Much better fits could be obtained if yields of two lumped fragmentation products, representing different photodecomposition action spectra, are optimized separately for each isomer. However, such adjustments still did not provide satisfactory fits for benzene, and could not simulate the data for the other aromatics in all respects.

The data base obtained in this study will be an important resource for evaluating updated mechanisms which are under development.

A series of environmental chamber experiments and computer model calculations were carried out to assess the atmospheric ozone formation potential of propylene glycol (PG). The experiments consisted of determining the effects of adding PG on NO oxidation, ozone formation and integrated OH radical levels in simulated model photochemical smog systems. Experiments were carried out using two different surrogate mixtures to represent the reactive organic gases (ROGs) present in the atmosphere, and using differing ROG/NOx ratios. It was found that PG has a positive effect on ozone formation, and that it does not significantly enhance or inhibit OH radical levels. The rates of consumption of PG in the chamber experiments relative to those of m-xylene corresponded to OH radical rate constant of (2.8+/-0.6) x 10-11 cm3 molec-1 s-1, which is almost a factor of three higher than the previously reported value of Wiedelmann and Zetzch (1982), but is within the experimental uncertainty of the value recently obtained by Aschmann and Atkinson (1997). The observed effects of PG on NO oxidation and O3 formation were inconsistent with the previously determined lower OH + PG rate constant, and were much better predicted by model simulations using the higher rate constants of Aschmann and Atkinson (1997) as determined in this work.

The PG mechanism with the higher OH + PG rate constant was then used to estimate its ozone impacts for a variety of atmospheric conditions. These were compared with ozone impacts calculated for representative VOCs and for the mixture of all emitted VOCs. The results indicated that the ozone impact of PG was comparable to the average for all VOC emissions.

The current practice for the U.S. EPA is to consider exempting volatile organic compounds (VOCs) from regulation as tropospheric ozone precursors if they can be shown to have lower or similar ozone impacts as does ethane. This paper discusses a procedure to estimate upper limit ozone impacts of compounds relative to ethane (or any other compound chosen to define the low-reactivity standard), if only the atmospheric reaction rate constants are known. This is based on deriving upper limits for the two factors which determine the ozone impact of a compound in a pollution episode: its kinetic reactivity, or the fraction of emitted VOC which reacts during the episode, and its mechanistic reactivity, the amount of ozone formed per molecule of VOC which reacts. Upper limit relative kinetic reactivities are derived from the ratio of the OH radical rate constants, with corrections being made if the VOC reacts with O3, NO3 radicals, or by photolysis. Upper limit relative mechanistic reactivities are obtained from calculated relative mechanistic reactivities for a variety of VOCs for various one day ozone episodes. The minimum information needed for making such estimates is discussed, and examples are given for several representative low reactivity compounds.

A procedure is presented for estimating upper limit Maximum Incremental ozone Reactivities (MIR's) for volatile organic compounds (VOCs) whose atmospheric ozone impacts are uncertain. This might be applicable for including compounds of unknown reactivity in reactivity-based regulations which employ the MIR scale for quantifying ozone impacts. The procedure is based on deriving upper limits for the two factors which determine the ozone impact of a compound in pollution episode: the fraction of emitted VOC which reacts during the episode (the kinetic reactivity), and the amount of ozone formed per molecule of VOC which reacts (the mechanistic reactivity). Upper limit kinetic reactivities are derived from the rate constants for the atmospheric reactions of the VOCs if they are known, otherwise a maximum kinetic reactivity of unity is assumed. Upper limit mechanistic reactivities are obtained from calculated relative mechanistic reactivities for a variety of VOCs for MIR scenarios, with lower upper limits being obtained for compounds known not to photolyze, or for alkanes or saturated compounds containing only alcohol, ether or ester groups. The information needed for making such estimates are discussed, and examples are given for representative VOCs whose reactivities are reasonably well known.

Environmental chamber experiments and computer model calculations were conducted to assess the atmospheric ozone formation potentials of four mineral spirits samples. Analyses of the four samples by high-resolution GC-MS, FIA type analysis, carbon number fractionation, and elemental composition indicated that they consisted primarily of C8-C15 normal (5-26% by weight), branched (23-42%), and cyclic (44-52%) alkanes. Three of the samples were >98% alkane, while one sample also contained ~6% aromatics and ~2% olefins. The chamber experiments consisted of blacklight irradiations, in a dual ~5000-liter chamber, of simulated photochemical smog mixtures with and without the sample added. They employed two different reactive organic gas (ROG) surrogate mixtures to represent other organic pollutants in the atmosphere, and two different ROG/NOx levels. All four samples inhibited OH radical levels in all experiments and inhibited rates of O3 formation and NO oxidation in the simplified surrogate runs which are more sensitive to radical inhibition effects. However, the inhibition was somewhat less for the sample containing the aromatics and olefins than the samples consisting entirely of alkanes. The all-alkane mineral spirits had relatively small effects on ozone in the experiments using the more realistic ROG surrogate, while the aromatic and olefin-containing sample had a positive effect on ozone in the run with this surrogate at the higher NOx levels, though it had no effect on the final ozone yield in the lower NOx run. The results of the experiments with the all- alkane samples were similar to experiments with n-alkanes which were carried out in a previous program.

The analytical data were sufficient to determine the set of model species needed to calculate their ozone reactivities in environmental chamber and airshed simulations. However, the model underpredicted the O3 inhibition in the runs with the simplified ROG surrogate. and overpredicted the O3 reactivities in the runs with the more realistic surrogate. Much better simulations were obtained if the model represented the branched and cyclic alkane constituents as if they were normal alkanes. This is despite the fact that current estimation methods for atmospheric reactions of alkanes predict that branched and cyclic alkanes have mechanisms which are significantly more favorable for ozone formation than those for normal alkanes. This indicates that current reactivity scales [such as the Maximum Incremental Reactivity (MIR) scale] might be overestimating the ozone impacts of mineral spirits and similar petroleum-based mixtures by a factor of 2 or more. On the other hand, the model performed reasonably well in simulating the increase in reactivity caused by the presence of aromatics or alkenes in the sample, once it was suitably adjusted to correctly simulate all-alkane sample reactivities. It is concluded that the current methods for estimating mechanisms for the branched and cyclic alkanes are unsatisfactory and need to be studied. It is also concluded that more information is needed concerning the representativeness of the samples studied in this program to mineral spirits in general, and data are needed to improve our ability to model the atmospheric reactions of branched and cyclic alkanes.

A series of environmental chamber experiments were conducted to measure the effects of isoprene on ozone formation, NO oxidation, and OH radical levels in photochemical smog systems. The experiments consisted of irradiations of isoprene by itself in NOx- air systems, and isoprene in the presence of other organics to determine their effects on simplified model smog systems. An updated isoprene atmospheric reaction mechanism was developed, and its predictions were compared with those of published isoprene mechanisms and with the results of these and previous experiments.

A series of environmental chamber experiments were conducted to measure the effects of selected terpenes compounds on ozone formation, NO oxidation, and OH radical levels in photochemical smog systems. The compounds studied included alpha- and beta- pinene, sabinene, delta-3-carene, and d-limonene. The experiments consisted of irradiations of the compounds by themselves in NOx-air systems, and irradiations of these compounds in the presence of other organics to determine their effects on simplified model smog systems. The experiments indicated that different terpene isomers can have quite different reactivities, even after the effects of differences in their reaction rates are taken into account. Current mechanisms used to represent the atmospheric reactions of terpenes in airshed models perform poorly in simulating these results. New mechanisms for representing the reactions of the terpenes in airshed models, which are more consistent with the results of these experiments and other recent laboratory studies, are discussed.

These are the slides used at the presentation. The atmospheric chemistry of ozone formation from VOCs and NOx and its implications to control strategies and assessing alternative fuel use are summarized. The factors involved in the quantification of VOC impacts on ozone formation are summarized, examples of reactivity scales are discussed, and the current status of chemical mechanisms used to calculate VOC reactivity and assess alternative fuel impacts on ozone are discussed. Recommendations for future research are summarized.

An extended abstract and the overheads used in the presentation are included in the downloadable file. The presentation focuses primarily on the updates to the SAPRC-90 mechanism and the results of a reevaluation of the chamber effects model. The updates to generally cause a more "reactive" mechanism. Inappropriately high assignments of chamber radical sources may have masked biases in previous mechanisms for underprediction of ozone.

An extended abstract and the overheads used in the presentation are included in the downloadable file. The University of California, Riverside environmental chamber data base for evaluating photochemical mechanisms is described. The data are now available at ../chdata.

The factors involved in evaluations of effects of volatile organic compound (VOC) and oxides of nitrogen emissions from motor vehicles on ground-level ozone are reviewed. The chemical basis for ozone formation, the computer models and chemical mechanisms used for predicting ozone impacts, and methods for quantifying differences among VOCs in their ozone impacts, are discussed. It is recommended that evaluations of effects of changing fuel and vehicle technology on ozone include an analysis of the effects of uncertainties in the model calculations and of the variability of ozone impacts with atmospheric conditions. Other related research needs are summarized.

Ozone in photochemical smog is formed from the gas-phase reactions of oxides of nitrogen (NOx) and volatile organic compounds (VOCs). VOCs differ in their effects on ozone formation, and methods for quantifying these differences would aid in the development of cost- effective ozone control strategies. In this presentation, the role of VOCs and NOx in ozone formation and the process of developing and evaluating chemical mechanisms for simulating these processes in the atmosphere will be summarized. Following that, the problems and recent progress in developing quantitative reactivity scales measuring ozone impacts of VOCs will be discussed.

An overview of VOC reactivity with respect to ground level ozone formation, and the mechanistic and environmental factors affecting it, are discussed. Methods of quantifying reactivity are described, and the roles of chemical mechanisms and environmental chamber data in reactivity quantification are discussed. A brief summary of ongoing reactivity-related research is presented, which can be classified as basic kinetic and mechanistic studies, environmental chamber studies, and chemical mechanism development. The current status of the reactivity-related environmental chamber and mechanism development research being carried out by the author is summarized, and illustrative examples of recent research results are presented. The development of an updated SAPRC mechanism for VOC reactivity assessment calculations is summarized.

A background discussion will be given on the definition of reactivity, the factors affecting reactivity, and how reactivity is measured or calculated. The specific areas of research, then be discussed, with emphasis on the uncertainties involved and the areas where more research is needed. The status of chemical mechanisms the author has used to calculate reactivity, problems and uncertainties for various VOC classes, methods for uncertainty analysis, and the data needs for mechanism evaluation will be discussed. Following that, a brief discussion will be given concerning airshed model uncertainties as they relate to reactivity, and the problem of how to deal with the dependence of reactivity on airshed conditions.

A series of laboratory studies, environmental chamber experiments, and computer model calculations were carried out to asses the atmospheric ozone formation potential of t-butyl alcohol (TBA), N-methyl pyrrolidinone (NMP), and propylene carbonate (PC). The OH radical rate constants at 296-298øK were measured to be (1.43ñ0.36) x 10-12, (2.15ñ0.36) x 10-11, and (6.9ñ1.5) x 10-13 cm3 molec-1 s-1, for TBA, NMP, and PC, respectively, and the NO3 radical rate constant for NMP was measured to be (1.26ñ0.40) x 10-13 cm3 molec-1 s-1. Photolysis and reaction with ozone were found not to be significant loss processes for NMP and PC under atmospheric conditions. NMP was observed to form N-methylsuccinimide and 1-formyl-2- pyrrolidinone with yields of 44% and 41%, respectively in the OH radical reaction, and of 59% and ~4% in the reaction with NO3. The chamber data confirmed that the major products from TBA are formaldehyde and acetone, and ruled out mechanisms predicting that formaldehyde is a major product from PC.

The chamber experiments consisted of determining the effects on NO oxidation, ozone formation and integrated OH radical levels of adding these compounds to simulated model photochemical smog systems. Mechanisms were developed for TBA and PC which were consistent with the results of these experiments, though to obtain acceptable fits it was necessary to assume ~7% organic nitrate yields in the reactions of NO with the peroxy radicals formed from TBA, and ~8% for those formed from PC. NMP was found to inhibit NO oxidation during the initial stages of the experiments, then accelerate O3 formation once O3 formation began. This is attributed to the effect of its NO3 reaction, and was well simulated by the estimated mechanism provided a ~15% nitrate yield in the NO + peroxy radical reactions was assumed. However, the estimated NMP mechanism overpredicted its effect on final ozone yields by up to 50%, and to obtain improved fits in this regard it was necessary to assume fewer NO to NO2 conversions are involved in the photooxidation reactions than could be rationalized. This adjustment reduced NNP' predicted atmospheric reactivities by 15- 20%.

The experimentally evaluated mechanisms were then used to predict the ozone reactivities of these compounds under various atmospheric conditions. TBA and PC were calculated to have ozone impacts very close to those of ethane, with reactivities relative to ethane varying from ~1.4 to 1.0ñ0.2 for TBA and ~1.4 to 0.9ñ0.2 for PC, depending on conditions and how O3 impacts were quantified. NMP was calculated to form 5-6 times more ozone than ethane, though it was somewhat less reactive than the average of all emissions, and its Maximum Incremental Reactivity was less than half that of toluene.

An experimental and computer modeling study was carried out to estimate the atmospheric ozone formation potentials of the dibasic esters dimethyl succinate [CH3OC(O)CH2CH2C(O)OCH3, DBE-4], dimethyl glutarate [CH3OC(O)CH2CH2CH2C(O)OCH3, DBE-5], and dimethyl adipate [CH3OC(O)CH2CH2CH2CH2C(O)OCH3, DBE-6]. The rate constants for their reactions with OH radicals at 298 K were measured, using a relative rate method, to be (1.5 ñ 0.6), (3.5 ñ 1.1), and (8.8 ñ 2.6) x 10-12 cm3 molec-1 s-1, for DBE-4, DBE-5 and DBE-6, respectively. These are relative to an OH + cyclohexane rate constant of 7.49 x 10-12 cm3 molec-1 s-1.

A series of environmental chamber experiments were carried out to determine the effects of adding DBE-4 and DBE-5 on NO oxidation, ozone formation, and OH radical levels in simulated model photochemical smog systems. The experiments used two different NOx levels and used two different surrogate mixtures to represent the reactive organic gases (ROGs) present in the atmosphere. Both DBEs caused reduced OH radical levels in all the experiments, caused reduced rates of NO oxidation and O3 formation in the experiments using the simpler ROG surrogate, but caused enhanced O3 formation in experiments using a more complex ROG surrogate which is representative of atmospheric conditions. Similar results are observed in experiments with the higher alkanes, with the radical inhibition in both cases being attributed to alkyl nitrate formation in the reactions of proxy radicals with NO.

Atmospheric reaction mechanisms for the DBEs were estimated by analogy with known atmospheric reactions of other compounds. They had a number of uncertainties, but the estimated DBE-4 and DBE-5 mechanisms were found to be consistent with the results of the chamber experiments if the overall alkyl nitrate yields of ~12% was assumed for DBE-4 and ~25% for DBE-5. These were respectively ~33% lower and about the same as expected based on nitrate yields for alkanes with the same number of carbons. Alternative mechanisms assuming formation of formaldehyde or highly photoreactive à- dicarbonyl products were found not to be consistent with the chamber data.

The estimated and adjusted DBE mechanisms were then used to calculate ozone impacts of these compounds for a variety of atmospheric conditions. The predicted ozone impacts, relative to an equal mass of ethane were: ~0.65 - 0.8 for DBE-4; ~1 - 1.25 for DBE-5; and ~3 - 4.5 for DBE-6. These relative reactivities were only slightly affected by variations in scenario conditions or how ozone was quantified. The chamber data suggests that the model might be slightly overestimating the ozone impacts for DBE-4 and DBE-5 under lower NOx conditions, with the estimated uncertainty being ~30%. The predictions for DBE-6 are more uncertain because no chamber data were obtained to test the assumed mechanism for this compound.

Environmental chamber experiments and computer model calculations were carried out to assess the atmospheric ozone formation potential of acetylene. The experiments consisted of irradiations, using a xenon arc light source simulating ground-level sunlight, of acetylene - NOx - air mixtures and simulated photochemical smog mixtures with and without added acetylene. The latter employed two different reactive organic gas surrogate mixtures to represent other organic pollutants in the atmosphere, and used different surrogate/NOx levels. Acetylene was found to enhance both ozone formation and OH radical levels to a much greater extent than predicted by the previously assumed atmospheric reaction mechanism for acetylene. The data could be satisfactorily simulated by model calculations only it was assumed that the photolysis of glyoxal, acetylene's major photooxidation product, involved significantly more radical formation than previously assumed, that the overall glyoxal photodecomposition quantum yield in simulates sunlight is almost a factor of 2 higher than previously reported, and that most of the glyoxal formation in the OH + acetylene reaction comes from the decomposition of the OH + acetylene + O2 adduct to glyoxal + OH radicals.

The modified acetylene and glyoxal mechanisms were then used to estimate ozone impacts of acetylene in one-day box model simulations of a variety of urban ozone pollution episodes. It was found that acetylene caused ~20-25% as much ozone formation on a per gram basis as the sum of all measured VOCs in urban atmospheres, and that acetylene caused somewhere between 25% to three times more ozone formation than an equal mass of ethane, the compound the EPA has been using as the standard to determine VOC exemption. The impact of acetylene relative to ethane tending to be highest in the higher NOx scenarios representing maximum incremental reactivity (MIR) conditions, while the impact of acetylene relative to the total of all emissions tending to be much less variable. It is concluded that while acetylene has a much lower ozone impact than average, its ozone impact should be considered to be somewhat higher than that of ethane.

A series of environmental chamber experiments and computer model calculations were carried out to asses the atmospheric ozone formation potential of trichloroethylene (TCE). The experiments consisted of determining the effects of adding TCE on NO oxidation, ozone formation and integrated OH radical levels when irradiated in the presence of NOx, NOx and ethane, or in simulated model photochemical smog systems using differing surrogate mixtures to represent the reactive organic gases (ROGs) present in the atmosphere, and using differing ROG/NOx ratios. The addition of relatively small amounts of ethane slowed down the rates of NO oxidation, O3 formation, and TCE consumption in TCE - NOx irradiations. This is explained by chain reactions involving chlorine atoms. TCE had a positive effect on NO oxidation, O3 formation, and radical levels in simulated photochemical smog systems, though the positive effect on ozone declined to zero in experiments with sufficiently low NOx/ROG ratios. TCE also enhanced the rate of alkane consumption, also indicating the role of chlorine atoms.

Two chemical mechanism, with differing assumptions concerning the relative importance of the TCE + ozone reaction, were developed which could predict the effect of TCE on rates of NO oxidation and initial rates of ozone formation in these experiments. However, the mechanisms could not account for the leveling off of O3 in the added TCE, high ROG/NOx experiments, the acceleration of ozone formation and leveling off and rapid TCE consumption at the end of one TCE - ethane - NOx experiment, and they tended to underpredict the apparent Cl atom production in many experiments. It is concluded that there are other secondary reactions occurring in the TCE photooxidation system which are not presently understood.

Atmospheric ozone model calculations were conducted to estimate TCE's ozone impacts for a variety of atmospheric conditions, though the chamber experiments indicated that the model predictions are probably unreliable except for the higher NOx Maximum Incremental Reactivity (MIR) conditions. The MIR of TCE was found to be approximately 3 times that of ethane on an ozone per gram emitted basis, or approximately 1/4 that of the mixture representing all reactive organic gas emissions. The model predicted that the relative ozone impacts of TCE did not change significantly as NOx levels are reduced, though based on the chamber experiments it is likely, though not certain, that its relative impacts decline. More laboratory data concerning the identity and reactions of TCE's atmospheric oxidation products are needed before models can be developed which can reliably predict TCE's ozone impacts under the full range of atmospheric conditions.

A series of environmental chamber experiments and computer model calculations were carried out to asses the atmospheric ozone formation potential of t-butyl acetate. The experiments consisted of determining the effects of adding t-butyl acetate on NO oxidation, ozone formation and integrated OH radical levels in simulated model photochemical smog systems. Experiments were carried out using two different surrogate mixtures to represent the reactive organic gases (ROGs) present in the atmosphere, and using differing ROG/NOx ratios. It was found that t-butyl acetate caused reduced OH radical levels in all experiments, had a small but positive effect on ozone formation when an ROG surrogate which is representative of ambient atmospheres is employed, but had slightly negative effects on ozone when a simpler ROG surrogate, which is more sensitive to radical inhibition effects, is employed. Several alternative mechanisms were developed, using results of recent product studies and various estimation methods. The results of chamber experiments were reasonably well simulated by these mechanisms if relatively high organic nitrate yields of ~120% are assumed, and if the OH radical + t-butyl acetate rate constant used is in the extreme low end of its uncertainty range. Using the mechanisms which were most consistent with the chamber data, the atmospheric ozone impact of t-butyl acetate was calculated to be approximately half that of ethane, with relatively little dependence on atmospheric conditions and how ozone impacts were quantified. Although the OH radical rate constant and some other aspects of the mechanism were uncertain, making alternative assumptions did not have large effects on the atmospheric reactivity predictions, if the model was adjusted to be approximately consistent with the chamber data. Regardless of the mechanism uncertainties, it can be concluded that t-butyl acetate can be considered to have a lower ozone impact than that of ethane.

Environmental chamber experiments and computer model calculations were conducted to assess the atmospheric ozone formation potentials of 1-propyl and 1-butyl bromides. The experiments consisted of blacklight irradiations, in a dual ~5000-liter chamber, of simulated photochemical smog mixtures with and without propyl or butyl bromide added. They employed two different reactive organic gas surrogate mixtures to represent other organic pollutants in the atmosphere, and used different surrogate/NOx levels. The results indicated that the alkyl bromides cause accelerated rates of ozone formation once ozone formation begins in the higher NOx experiments, but have smaller effects on O3 formation rates in lower NOx experiments, and cause ozone levels to peak and then decline in the later stages of the runs. A chemical mechanism was developed to represent the atmospheric reactions for these alkyl bromides, and the bromine atom, BrOx species, and bromine- containing carbonyl products they are predicted to form. This mechanism could simulate the results of the chamber experiments only if it is assumed that there is a rapid reaction between O3 and HBr, forming HOBr, which photolyzes rapidly to form OH + Br. However, separate measurements in another chamber indicate that this reaction is relatively slow. This indicates that there is some unknown process which is occurring in our chamber experiments which has the same effect as a rapid O3 + HBr reaction. The calculated ozone impacts of the bromides under atmospheric conditions were found to depend significantly on NOx conditions and whether the unknown process, represented by the rapid O3 + HBr reaction, is assumed to occur in the atmosphere as it does in our chamber. Propyl bromide is calculated to be an ozone inhibitor or have very low ozone impact under low to moderate NOx conditions regardless of whether the rapid O3 + HBr reaction is included in the model. The magnitude of this inhibition depends on whether the reaction is included. Under higher NOx, Maximum Incremental Reactivity (MIR) conditions propyl bromide is calculated to have comparable ozone impact (on a mass basis) as ethane if the reaction is not included, but it is calculated to form about three times more ozone than ethane if the O3 + HBr reaction is included. Butyl bromide calculated to have 1.5 - 2 times more ozone impact than calculated for propyl bromide using the same assumptions about the O3 + HBr reaction under high NOx MIR conditions, and correspondingly higher, or less negative, ozone impacts under lower NOx conditions. It is concluded that these compounds can be reasonably be expected to have low or negative ozone impacts under low NOx conditions, but that their atmospheric ozone impacts under higher NOx conditions are uncertain. In the case of propyl bromide, the range of uncertainty is between approximately the reactivity of ethane to approximately one-fourth of the average of the reactivity of all reactive VOC emissions.

Environmental chamber experiments and computer model calculations were conducted to assess the atmospheric ozone formation potentials of 2,4- and 2,6-toluene diisocyanate. The experiments consisted of determining the effects of adding the TDI isomer on ozone formation, NO oxidation, and integrated OH radical levels in three different simulated photochemical smog systems in ~5000-liter, blacklight-irradiated Teflon environmental chambers. TDI was found to inhibit radical levels and ozone formation rates and yields in all experiments. This is apparently due to the presence of both radical sinks and NOx sinks in TDI's overall atmospheric reactions. Although the mechanism for TDI's atmospheric reactions is unknown, the results of the experiments could be fit by simple parameterized models assuming 70% radical inhibition, no NO to NOx conversions, and formation of significant yields of products, such as cresols or nitrophenols, whose subsequent reactions cause NOx removal. The NOx sinks are clearly more important than any NOx source which might result from the oxidation of the NCO groups. The lack of apparent NO to NO2 conversions in TDI's overall mechanism suggests that TDI will not cause significant ozone formation even under conditions which are not sensitive to its radical and NOx inhibition effects. An 18-hour irradiation indicated that the ozone inhibition effects of TDI continue over at least a two day time period. Both 2,4-TDI and 2,6-TDI were simulated with essentially the same mechanism, indicating no major isomeric differences affecting TDI's reactivity.

The TDI photooxidation models which fit the chamber data were used to estimate TDI's impacts on ozone formation in one-day EKMA model scenarios representing various ozone pollution episodes throughout the United States. TDI was predicted to have a negative effect on ozone formation in all those episodes. Similar results were obtained in a limited number of calculations using multi-day scenarios. It is concluded that emissions of TDI are unlikely to have a positive effect on ozone formation under any atmospheric conditions, and thus it should not be considered to be an ozone precursor.

.The use of natural gas as a vehicle fuel has benefits over petroleum-based fuels in reducing the primary emission of volatile organic compounds (VOC) and particulate matter (PM). The objective of this research was to provide information concerning the relative PM forming potential of exhaust emissions from vehicles fueled with natural gas compared with that from gasoline and diesel-fueled vehicles. A dual-reactor, large (~20 m³ each) indoor environmental chamber was constructed to allow for growth and measurement under controlled conditions simulating the atmosphere. A chassis dynamometer facility was used to generate and characterize exhausts from a natural gas, a gasoline, and from a diesel-fueled vehicle. The exhaust gases were added to propene - NO_x - air mixtures to simulate gas-phase pollutants from other sources in polluted urban atmospheres. Cold-start exhaust was injected into one reactor and hot start exhaust from the same vehicle was injected into the other, and the two mixtures were irradiated with blacklights for 4-6 hours, with gas-phase species measured before and during the irradiations. A scanning electrical mobility spectrometer (SEMS) was used to obtain particle number and volume information throughout the experiments, and the final aerosol mass was also determined by passing the contents of the reactors through filters which were weighed at the end of the experiments. Significant small particle formation was observed in control experiments due to nucleation by background contaminants, which prevented quantitative particle growth information to be obtained from the SEMS data. The particle mass emitted or formed from the natural gas vehicle was found to be not significantly different from the control experiments, while significantly higher particle mass was formed from the vehicle fueled by conventional gasoline. The data from the diesel vehicle were inconclusive because of the small amount added. Additional experiments using an improved air purification system and including separate particle loss measurements in the chamber are needed before quantitative information can be obtained concerning differences in PM forming potentials in vehicle emissions. Nevertheless, the data obtained indicate that emissions from natural gas vehicles are likely to have lower PM forming potentials than those from gasoline vehicles

Environmental chamber experiments and computer model calculations were conducted to assess the atmospheric ozone formation potentials of para toluene isocyanate (PTI), which is an analogue for methylene diphenylene diisocyanate (MDI), a compound of commercial importance which might be emitted into the atmosphere. MDI could not be studied directly because of its low vapor pressure, but PTI is considered to be a good analogue to MDI because of structural similarities. The experiments consisted of determining the effects of adding PTI on ozone formation, NO oxidation, and integrated OH radical levels in three different simulated photochemical smog systems in ~5000-liter, blacklight-irradiated Teflon environmental chambers. The rate constant for the reaction of PTI with hydroxyl radicals, its major expected atmospheric fate, was measured, and found to be essentially the same as that for toluene. PTI was found to have very small effects on ozone formation and OH radical levels under higher NOx conditions, but was found to cause reduced ozone yields and OH radical levels in lower NOx experiments. Although the mechanism for PTI's atmospheric reactions is unknown, the results of the experiments could be fit by simple parameterized models assuming significant formation of a cresol-like product, approximately 10% radical inhibition, some NO to NO2 conversion, and some formation of a photoreactive product which is represented in the model by methyl glyoxyl.

The parameterized model which best fit the PTI chamber data and the PTI OH radical rate constant were used to estimate a mechanism to predict the atmospheric impact of MDI, which is essentially a dimer of PTI. It is concluded that MDI will probably have negative impacts on ozone formation under atmospheric conditions where NOx limited or most favorable for ozone formation, which was the case in most of the scenarios used in this study to represent ozone exceedence episodes in various areas of the United States. However, MDI was calculated to have positive impacts and form comparable or more ozone than ethane under higher NOx, MIR conditions where ozone is sensitive to VOC emissions. Therefore, MDI is expected to be a less general ozone inhibitor than TDI, which based on a previous study is expected to be an ozone inhibitor under all conditions.

Environmental chamber experiments and computer model calculations were conducted to assess the impacts of the gas-phase reactions of styrene on the atmospheric formation of ozone and styrene's known oxidation products. The experiments consisted of determining the effects of adding styrene on ozone formation, NO oxidation, integrated OH radical levels, and formation of benzaldehyde, formaldehyde, and peroxybenzoyl nitrate (PBzN) in various simulated photochemical smog systems in a dual ~2500-liter, xenon-arc- irradiated Teflon environmental chamber. The gas-phase mechanism for the atmospheric reactions of styrene was updated based on available literature data, and model predictions using the mechanism were compared with the results of the chamber experiments. The model predictions were consistent with the observed effects of styrene on benzaldehyde and formaldehyde and not inconsistent with the qualitative PBzN data, but the effect of styrene on radical levels and ozone could only be simulated if it was assumed that the reaction of styrene with ozone does not result in radical formation. The mechanisms which fit the chamber data was used to used to predict the impacts of styrene on ozone formation under various atmospheric conditions. The changes to the ozone + styrene mechanism caused the Maximum Incremental Reactivity (MIR) for styrene, relative to the average of all VOC emissions, to decrease from ~0.7 to ~0.6, on a mass basis. Uncertainties in the mechanism for the styrene + NO3 reactions were found not to significantly impact ambient ozone impact predictions. The impact of styrene on ozone formation depended significantly on NOx conditions, being greater than that of ethane (but less than the average of all emissions on a mass basis and therefore not "highly reactive") under the high NOx conditions of the MIR scale, but becoming significantly negative in scenarios with lower NOx levels. The reactivities in the MIR scenarios were the same regardless of whether the ozone impact was defined in terms of peak ozone or the maximum 8-hour average, but the 8-hour average ozone reactivities declined much more slowly as NOx was reduced than reactivities in terms of effects on peak ozone yields. It is concluded that models with an appropriate styrene mechanism (and an explicit representation of benzaldehyde) can probably reliably represent the effects of styrene on the formation of ozone, benzaldehyde, formaldehyde, PBzN, and overall radical levels in the atmosphere. However, the known products from styrene's reactions account for only ~60% of the carbon reacted, and atmospheric impacts of the unknown products may not be well represented in current models.

Although some characterization problems and model discrepancies were observed, the results of most of the experiments with LPG, M100, M85, CNG and RFG exhausts were consistent with results of experiments using synthetic exhausts derived to represent them, and were generally consistent with model predictions. The major exception to this was the one experiment with diesel exhaust, where a complete analysis was not conducted and where it was clear that the major reactive species have not been identified. The results with the other exhausts indicate that the major constituents contributing to their ozone impacts have probably been identified, and that current chemical mechanisms are reasonably successful in predicting these impacts. There was no evidence for a contribution of nitrites or other contaminates or artifacts to the reactivities of any of these exhausts. There was some evidence, albeit inconclusive, that the model may be underpredicting the ozone impacts of some of the constituents of exhausts from the two highest mileage RFG-fueled vehicles in some experiments. This would require further studies with other vehicles before any conclusions can be made. However, the model gave reasonably good simulations of effects of adding these to realistic ambient VOC - NOx mixtures, as was the case for all the other exhausts for which complete analyses were conducted.

A detailed mechanism for the gas-phase atmospheric reactions of volatile organic compounds (VOCs) and oxides of nitrogen (NOx) in urban and regional atmospheres is comprehensively documented in this report. This can be used in airshed models to determine absolute and relative ozone impacts (reactivities) of the many types of VOCs that can be emitted into the atmosphere, and for other control strategy and research applications. This mechanism, designated SAPRC-99, represents a complete update of the SAPRC-90 mechanism of Carter (1990), and incorporates recent reactivity data from a wide variety of VOCs. The mechanism has assignments for ~400 types of VOCs, and can be used to estimate reactivities for ~550 VOC categories. A condensed version was developed for use in regional models. A unique feature of this mechanism is the use of a computerized system to estimate and generate complete reaction schemes for most non-aromatic hydrocarbons and oxygenates in the presence of NOx, from which condensed mechanisms for the model can be derived. The mechanism was evaluated against the results of approximately 1700 environmental chamber experiments carried out at the University of California at Riverside, including experiments to test ozone reactivity predictions for over 80 types of VOCs. The mechanism was used to update the various ozone reactivity scales developed by Carter (1994a), including the widely used Maximum Incremental Reactivity (MIR) scale. However, the reactivity estimates for many VOC classes are uncertain, which must be taken into account when using these data for regulatory applications. For this reason, uncertainty classifications have been assigned to all VOCs, and upper limit MIRs for VOCs with uncertain mechanisms are presented.

This report documents the files, software and procedures needed to implement the SAPRC-99 detailed chemical mechanism into the Models-3 software framework. The SAPRC-99 mechanism is a detailed mechanism for the gas-phase atmospheric reactions of volatile organic compounds (VOCs) and oxides of nitrogen (NOx) in urban and regional atmospheres, and represents the state-of-the-art as of mid-1999. It is a completely updated and expanded version of the earlier SAPRC mechanisms, and is comprehensively documented in a report to the California Air Resources Board (Carter, 1999). It has the capability of separately representing the atmospheric reactions of ~400 types of VOCs, and can be used to estimate reactivities for ~550 VOC categories. Condensed versions of this mechanism have been developed for use in regional models, using a more limited number of lumped VOC classes whose mechanistic parameters depend on the mixture of compounds they represent. Different versions can be used depending on which VOC mixture is used to derive the mechanisms and parameters for these lumped VOC classes. This report describes the implementation of two condensed versions of SAPRC-99 into the Models-3 framework, one where the lumped VOC classes are derived from VOCs measured in ambient air, and one where the lumped VOC classes are derived from VOCs present in a recent EPA emissions inventory. Methods for deriving versions of the mechanism representing other mixtures or emissions inventories, and for explicitly representing selected VOCs for reactivity assessment and other purposes, are discussed. The procedures for obtaining, installing, and using the software and files needed to implement this mechanism are described.

Because the major atmospheric reaction pathways and products for aromatic hydrocarbons are uncertain, they are represented in air quality models using parameterized mechanisms derived by modeling environmental chamber data. Uncertainties in rate constants, experimental conditions and chamber artifacts affect the parameter estimates derived in this manner. The SAPRC-97 mechanism represents aromatic ring fragmentation products by model species MGLY (alpha-dicarbonyls) and AFG2 (other photoreactive products) with yields derived from aromatics-NOx experiments conducted in indoor chambers with blacklight or xenon arc light sources. This study explores how experimental and modeling uncertainties affect these chamber-derived aromatics parameters, and in turn the reactivity estimates calculated for the aromatic compounds.

The uncertainty levels (1 sigma relative to the mean) for the aromatics oxidation parameters range from about 29% for the MGLY yield from 135-trimethylbenzene oxidation to 71% for the MGLY yield from p-xylene. Major causes are uncertainties in rate constants for the aromatics + OH and NO2 + OH reactions, and the light intensity, chamber radical source parameters and initial aromatic concentrations in the experiments.  The chamber radical source parameters are estimated from CO-NOx and n-butane NOx experiments, and are sensitive to uncertainties in the rate constants for n-butane or CO + OH, NO2 + OH, HONO photolysis and the experimental light intensity.

More than 100 parameters of the SAPRC-97 mechanism, including the chamber-derived aromatics parameters, are propagated through incremental reactivity calculations using Monte Carlo analysis with Latin hypercube sampling. The uncertainty levels found for the maximum incremental reactivities (MIRs) of the aromatic compounds range from 27 to 32%, and are about the same as those for other volatile organic compounds with relatively well-established mechanisms. The uncertainty levels for the maximum ozone incremental reactivities (MOIRs) and equal benefit incremental reactivities (EBIRs) of the aromatics range from 38 to 75% and 30 to 520%, respectively. Uncertainties in relative reactivities for the aromatic compounds range from 13 to 25%, 20 to 63% and 21 to 360% under MIR, MOIR and EBIR conditions. Uncertainties in the relative reactivities of most, but not all of the VOCs studied are smaller than the uncertainties in their absolute incremental reactivities.  The exceptions include some slowly reacting compounds under MIR, MOIR and EBIR conditions, and some of the aromatic compounds under EBIR conditions.

From 30% to 70% of the uncertainty in the relative MIRs of the aromatic compounds is contributed by their chamber-derived parameters. Similarly, from 14% to 60% of the uncertainties in the relative MOIRs and from 3% to 56% of the uncertainty in the relative EBIRs of the aromatics is attributed to their chamber-derived parameters.  Although the chamber-derived parameters are influential, the rate constant for the reaction CRES (cresol) + NO3 is the largest contributor to the relatively high uncertainty in the EBIRs of toluene, p-xylene and ethylbenzene.

As long as incremental reactivity estimates for aromatic compounds have to rely on chamber-derived parameters, uncertainty in these estimates could be reduced most by improving the characterization of radical sources, light intensity and initial concentrations in environmental chamber studies, and by reducing uncertainty in the rate constants for NO2 + OH, aromatics + OH, and CRES + NO3.  Future chamber studies of aromatics chemistry should emphasize low-NOx conditions to reduce the relatively high uncertainties in MOIR and EBIR estimates.

Download Document(1099 Kbytes)

Incremental reactivity estimates for many high molecular weight hydrocarbons and oxygenated compounds used in consumer products, coatings and solvents are viewed as uncertain because of limited data on their reaction mechanisms and products. This study performs a systematic uncertainty analysis for two solvents of interest for use in consumer products: 2-butoxy ethanol and n-butyl acetate. 2-butoxy ethanol provides an example of a relatively well-studied compound for which product data are available for most reaction pathways and for which incremental reactivity data are available from environmental chamber experiments. In contrast, n-butyl acetate is an example of a compound for which environmental chamber studies have been conducted but for which there are essentially no product data.

As a first step in the study, key mechanistic parameters for 2-butoxy ethanol and n-butyl acetate were estimated from experiments conducted in the University of California at Riverside chambers. The organic nitrate yield for 2-butoxy ethanol in the SAPRC-97 mechanism is estimated to be 0.13 ± 0.02 (mean ± 1s). For n-butyl acetate, the organic nitrate yield is estimated to be 0.13 ± 0.05 and the probability of the intermediate CH₃-CH₂-CH₂-CH[O.]-O-CO-CH₃ undergoing ester rearrangement to be 0.72 ± 0.22.

Along with uncertainty estimates for other parameters of the SAPRC-97 mechanism, the uncertainties for these chamber-derived parameters were propagated through incremental reactivity calculations using Monte Carlo analysis. The maximum incremental reactivity (MIR), maximum ozone incremental reactivity (MOIR) and equal benefit incremental reactivity (EBIR) are estimated to be 1.12 ± 0.27, 0.59 ± 0.14 and 0.40 ± 0.11 ppm O₃/ppmC respectively for 2-butoxy ethanol and 0.41 ± 0.16, 0.29 ± 0.10 and 0.20 ± 0.08 ppm O₃/ppmC for n-butyl acetate. The corresponding relative reactivities compared to the reactivity of a base mixture are 0.90 ± 0.14, 1.08 ± 0.16 and 1.20 ± 0.19 for 2-butoxy ethanol and 0.34 ± 0.13, 0.53 ± 0.16 and 0.60 ± 0.18 for n-butyl acetate. Uncertainties in the 2-butoxy ethanol reactivity estimates are lower than those estimated previously for many other VOCs. In contrast, the uncertainties in the n-butyl acetate reactivity estimates are at the upper end of the range of uncertainties estimated previously for other VOCs.

The uncertainty estimates for the incremental reactivities take into account the available kinetic data for 2-butoxy ethanol and n-butyl acetate, the product data for 2-butoxy ethanol, and the chamber-derived mechanistic parameter estimates for both compounds. Measurements of the 2-butoxy ethanol + OH rate constant and product yields and chamber-derived estimates of key mechanistic parameters are estimated to have reduced the uncertainty in the relative MIR of 2-butoxy ethanol by 40%, compared to the uncertainty level that was estimated assuming they were not available. The availability of measurements of the n-butyl acetate + OH rate constant is estimated to have reduced the uncertainty in the relative MIR for n-butyl acetate by about 25%.

The relative reactivity estimates for 2-butoxy ethanol are strongly influenced by uncertainty in the rate constants for its reaction with OH and the reaction of higher reactivity ketones (PROD2 in SAPRC) with OH. Rate parameters for n-butyl acetate + OH, O₃ photolysis and NO₂ + OH contribute most to the uncertainty in n-butyl acetate relative reactivity estimates.

From 2 to 3% of the total uncertainty in the relative reactivities of 2-butoxy ethanol is attributable to uncertainty in the chamber-derived yield of organic nitrate from its reaction. About 4 to 7% of the uncertainty in the n-butyl acetate relative reactivities is attributable to the organic nitrate yield in its mechanism. The most influential sources of uncertainty in the 2-butoxy ethanol organic nitrate yield are the parameters that quantify the unknown radical sources in the chamber experiments. The initial concentration of m-xylene and rate constants for PPN formation and decompostion are the largest contributors to the uncertainty in the organic nitrate yield from n-butyl acetate.

Overall, the uncertainty analysis indicates that improved quantification of chamber radical sources and initial conditions would help reduce uncertainty in the n-butyl acetate and 2-butoxy ethanol reaction mechanisms and in turn their incremental reactivity estimates. Reducing uncertainty in chamber experiments would also improve estimates of incremental reactivity for other compounds with chamber-derived parameters in their mechanisms. However, most of the uncertainty in the incremental reactivity estimates for n-butyl acetate and 2-butoxy ethanol is attributable to other parameters of the base SAPRC mechanism and to the rate constants for their primary reactions with OH. Among the parameters of the base mechanism, improving the representation of higher reactivity ketones and reducing the uncertainty in the ozone action spectra and PAN and PPN rate parameters would have the greatest effect on the relative reactivity estimates.

# Jana Milford: Department of Mechanical Engineering, University of Colorado, Boulder, CO  80309-0427, email milford@spot.colorado.edu

A series of environmental chamber experiments and computer model calculations were carried out to assess the atmospheric ozone formation potentials of selected organic compounds representative of those emitted from consumer products. This information is needed to reduce the uncertainties of ozone reactivity scales for stationary source emissions. The compounds studied were cyclohexane, cyclohexane, isopropyl alcohol, the three octanol isomers, diethyl ether, methyl ethyl ketone, cyclohexanone, methyl isobutyl ketone ethyl acetate, methyl isobutyrate, n-butyl acetate, and propylene glycol methyl ether acetate. “Incremental reactivity” experiments were carried out to determine the effect of each compound on O3 formation, NO oxidation and integrated OH radical levels when added to irradiations of reactive organic gas (ROG) - NOx representing simplified polluted urban atmospheres. Differing ROG surrogates and ROG/NOx ratios were employed to test how the impacts of the compounds vary with chemical conditions. In addition, single compound - NOx irradiations were carried out for the various ketones, OH radical rate constants were measured for the octanol isomers and propylene glycol methyl ether acetate, and the yields of the C8 carbonyl products were determined for each of the octanol isomers.

The results of these experiments were used in the development and testing of the SAPRC-99 mechanism that is documented in detail in a separate report (Carter, 2000). The data obtained, in conjunction with results of industry-funded studies of related compounds, has resulted in significantly reduced uncertainties in estimates of ozone impacts of the wide variety of oxygenated compounds present in consumer product emissions inventories. However, uncertainties still remain, and information is still inadequate to estimate ozone impacts for other classes of emitted compounds, such as amines and halogenated organics

The representation of aromatic hydrocarbons in the recently developed SAPRC-99 chemical mechanism for airshed models is discussed. Because of lack of available information, parameterized mechanisms adjusted to fit chamber data must still be used. The chamber data base employed consists of NOx-air irradiations of benzene, toluene, ethylbenzene, all three xylene and all three trimethylbenzene isomers in both blacklight and xenon-arc irradiated chambers, and of naphthalene, dimethylnaphthalene and tetralin in blacklight chambers alone. Significant differences in ozone formation potentials were observed that are not accounted for by differences in OH radical rate constants. These are accounted for in the model by using differing yields of model species used to represent the various photoreactive aromatic ring-opening products, with use of products with two different action spectra being necessary to fit data using differing light sources. Lower yields of photoreactive products are needed to fit the data for p-xylene or 1,2,4-trimethylbenzene compared to the other isomers, suggesting the formation of higher yields of less photoreactive unsaturated ketones compared to presumably more photoreactive unsaturated aldehydes. The parameterization that was found to successfully simulate the data for the alkylbenzenes was found to be less successful for benzene and the naphthalenes and tetralin, and a different parameterization performed better for the latter two. Implications of these results on predictions of O3 impacts in the atmosphere are discussed.

A detailed mechanism for the gas-phase atmospheric reactions of volatile organic compounds (VOCs) and oxides of nitrogen (NOx) in urban and regional atmospheres is summarized. This can be used in airshed models to determine ozone impacts (reactivities) of the many types of VOCs that can be emitted into the atmosphere, and for other control strategy and research applications. This mechanism, designated SAPRC-99, represents a complete update of previous mechanisms developed by the author, and incorporates recent reactivity data from a wide variety of VOCs. The mechanism has assignments for ~400 types of VOCs, and can be used to estimate reactivities for ~550 VOC categories. A unique feature of this mechanism is the use of a computerized system to estimate and generate complete reaction schemes for most non-aromatic hydrocarbons and oxygenates in the presence of NOx, from which condensed mechanisms for the model can be derived. The mechanism was evaluated against the results of almost 1700 environmental chamber experiments, including experiments to test ozone reactivity predictions for over 80 types of VOCs. The mechanism was used to update the various ozone reactivity scales developed previously, including the Maximum Incremental Reactivity (MIR) scale. Uncertainties and areas where additional work is needed are briefly summarized.

Evaluations of air quality impacts from vehicle emissions are based on the assumptions that all the important reactive species in the exhaust have been identified and quantified and that they are accurately represented in airshed models. To test this, environmental chamber experiments were carried out with exhaust from vehicles fueled by LPG, M100, M85, CNG, diesel, and Phase 2 reformulated gasoline (RFG), with the latter including vehicles representing a range of mileages, types, and pollution levels. Baseline FTP tests and speciation analyses were carried out for all vehicles studied but diesel. The chamber experiments consisted of irradiations of the exhausts themselves, “incremental reactivity” experiments with the exhaust added to two different surrogate VOC - NOx mixtures simulating conditions of photochemical smog, and irradiations of synthetic exhaust mixtures based on the results of the exhaust analyses. Computer model simulations of the experiments were also carried out. The results of most of the experiments were consistent with results of experiments using synthetic exhausts derived to represent them, and were generally consistent with model predictions. This indicates that the major constituents contributing to their ozone impacts have probably been identified, and that a current chemical mechanism is reasonably successful in predicting these impacts. The major exception to this was the one experiment with diesel exhaust, where a complete analysis was not conducted. There was some evidence, albeit inconclusive, that the model may be underpredicting the ozone impacts of some of the constituents of exhausts from the two highest mileage RFG-fueled vehicles in some experiments, but the model gave reasonably good simulations of effects of adding these to realistic ambient VOC - NOx mixtures, as was the case for all the other exhausts for which complete analyses were conducted.

A series of environmental chamber experiments and computer model calculations were carried out to assess the atmospheric ozone formation potential of dimethyl sulfoxide (DMSO). The experiments consisted of measuring ozone formation, NO oxidation and DMSO consumption rates in irradiations of DMSO - NOx mixtures and determining the effects of DMSO on O3 formation, NO oxidation and integrated OH radical levels when added to various simulated photochemical smog systems. The results indicated that DMSO is highly reactive towards ozone formation under all conditions examined. High yields of formaldehyde were observed, and approximately half of the sulfur in DMSO reacts to form products that are not detected by a total gas phase sulfur analyzer (i.e., not SO2). In addition, an upper limit rate constant of 3 x 10-20 cm3 molec-1 s-1 was determined for the reaction of DMSO with O3. The information available from previous studies is not sufficient to determine the mechanism for DMSO’s atmospheric reactions, and a number of alternative mechanisms were examined for consistency with the data obtained in this study. The best results are obtained using a mechanism where 75% of the reaction of DMSO with OH results in the formation of SO2 and two formaldehyde molecules after conversions of two molecules of NO to NO2; with the remaining 25% involving the formation of CH3S(O)2CH3 (DMSO2) and HO2. Although this mechanism underpredicts the effects of DMSO on NO oxidation and O3 formation in some experiments, it generally gives good simulations of the experiments most closely representing polluted urban atmospheres. This mechanism predicted that DMSO emissions form about twice as much ozone on a mass basis then emissions of the mixture of reactive VOCs representing emissions from all sources.

A series of environmental chamber experiments and computer model calculations were carried out to assess the atmospheric ozone formation potentials of methyl pivalate. The experiments consisted of determining the effects of this compound on NO oxidation, ozone formation, OH radical levels, and other measures of reactivity when added to varying simulated model photochemical smog systems. Methyl pivalate was found to slightly enhance O3 formation under conditions most representative of the atmosphere, but to inhibit radical levels and inhibit O3 in experiments that are sensitive to effects on radical levels. A mechanism for the atmospheric reaction for methyl pivalate was developed that is consistent with available laboratory data and current estimation methods, and that could simulate the results of these experiments after only minor adjustments of the overall nitrate yield to within its range of uncertainty. This mechanism was then incorporated in the overall SAPRC-99 mechanism and used to predict the atmospheric ozone impacts of this compound under various atmospheric conditions. The ozone formation potentials of methyl pivalate on a mass basis was found to be about 10-20% that of the mixture used to represent reactive VOC emissions from all sources. Its ozone formation potentials were found to be very similar to those of ethane, the compound used by the EPA as the basis for determining exemptions of compounds from regulation as ozone precursors.

A series of environmental chamber experiments and computer model calculations were carried out to assess the atmospheric ozone formation potentials of dimethyl carbonate and methyl isopropyl carbonate. The experiments consisted of determining the effects of these compounds on NO oxidation, ozone formation, OH radical levels, and other measures of reactivity when added to three types of simulated model photochemical smog systems. Both compounds were found to enhance NO oxidation and O3 formation rates under all conditions, though relatively large amounts of dimethyl carbonate had to be added to have measurable effects. The addition of methyl isopropyl carbonate caused increased formation of formaldehyde and acetone, while dimethyl carbonate had no effects on formaldehyde or other organic products that could be monitored. The rate constant for the reaction of OH radicals with dimethyl carbonate was measured to be 2.55±0.64 x 10^-12 cm³ molec^-1 s-¹, using a relative rate method with propane as the reference compound. Atmospheric reaction mechanisms for these compounds were developed using the measured OH radical rate constants and estimates for the subsequent reactions. These mechanisms were found to give good predictions of the effects of these compounds on O3 formation, NO oxidation, and product formation observed in the chamber experiments. These mechanisms were then incorporated in the overall SAPRC-99 mechanism and used to predict the atmospheric ozone impacts of these compounds under various atmospheric conditions. The ozone impacts of dimethyl carbonate were found to be no more than ~30% that of ethane on a mass basis, suggesting that it may be appropriate to exempt this compound from regulation as a VOC ozone precursor under the current standards used by the EPA. The ozone impacts of methyl isopropyl carbonate were found to be about twice those of ethane on a mass basis, comparable to that of propane, and about 1/5 to 1/3 that of the mixture used to represent reactive VOC emissions from all sources.

A series of environmental chamber experiments and computer model simulations were carried out to assess the atmospheric ozone formation potentials of the Exxsol® D95, Isopar® M and the Exxate® fluids (These names are trademarks of ExxonMobil Chemical Company). D95 is a petroleum-derived mixture of C12-C15 normal, branched, and cyclic alkanes, Isopar-M is a mixture of primarily C11-C16 branched alkanes made using an isomerization process, and the Exxate fluids are acetates of various branched alcohol mixtures in narrow weight ranges from C6 to C~13. The experiments consisted of determining the effects on NO oxidation, ozone formation and OH radical levels when adding D95, Isopar-M, or Exxate 1000 to varying simulated model photochemical smog systems. The reactivity characteristics of these materials were very similar to those for other alkane mixtures and individual higher molecular weight alkanes in this weight range that have been studied. They were found to be inhibitors of radical levels under all conditions and to inhibit rates of NO oxidation and O3 formation in experiments that are sensitive to radical effects, but to have relatively small effects on ozone in experiments more representative of atmospheric conditions. The results were used to determine whether we could accurately simulate the effects of these compounds on ozone formation and other manifestations of photochemical smog. Compositional information provided by ExxonMobil were used to derive compositions in terms of individual representative compounds, though assumptions had to be made concerning the appropriate compounds to represent the complex mixtures of branched and cyclic constituents. The models using these compositions and the SAPRC 99 mechanisms for the constituents gave reasonably good simulations to the results of most of the experiments, though the Isopar-M experiments were not simulated as well as the others. The models were then used to calculate ozone impacts of these materials in various urban photochemical smog scenarios. The impacts were on peak ozone yields were variable, but generally were 20-40% those of an equal mass of VOC emissions from all sources. The relative impacts on maximum 8-hour average ozone levels were generally less than their relative impacts on peak ozone levels, especially in scenarios with relatively low NOx conditions. However, making alternative assumptions concerning the appropriate compounds to represent the constituents of Isopar-M affected atmospheric ozone yield impacts by ~40%, even though they did not affect the simulations of the chamber data. Therefore, it is important that the types of compounds present in these fluids be represented as accurately as possible, even when environmental chamber data are available to test model predictions.

As the second phase of our study of the atmospheric ozone formation potentials of constituents of aluminum rolling mill emissions, a series of environmental chamber experiments and computer model simulations were carried out to assess the atmospheric ozone formation potentials of the selected cycloalkanes hexyl cyclohexane and octyl cyclohexane. The experiments consisted of determining the effects on NO oxidation, ozone formation and OH radical levels when adding the compounds to varying simulated model photochemical smog systems. The chamber experiments employed both blacklight and xenon arc light sources. These data supplemented similar experiments carried out previously on C12 - C16 normal alkanes. Like the C³12 normal alkanes, these cycloalkanes were found to be strong inhibitors of radical levels, inhibited rates of NO oxidation and O3 formation in experiments that are sensitive to radical effects, but had only relatively small effects on O3 in experiments more representative of atmospheric conditions.

Model simulations using the recently developed SAPRC 99 atmospheric chemical mechanism was gave good simulations of the results of the C³12 normal and cyclic alkane data obtained in both phases of this project. This mechanism was then used to calculate that impacts of these compounds on ozone formation in a one day box model scenarios representing various urban areas around the United States. These were compared with ozone impacts calculated for ethane, the compound that has been used by the EPA to determine “negligible” ozone reactivity for regulatory exemption purposes. The ozone impacts for the C³12 alkanes were found to be highly dependent on scenario conditions and how the ozone impact was quantified. In most cases the C³12 alkane ozone impacts on a mass basis were less than 25% those of the mixture representing VOCs from all sources, but the impacts relative to ethane varied greatly. In general, the impacts of the C³12 alkanes on maximum 8-hour average ozone levels were found to less than those of ethane under most conditions, while their impacts on peak ozone yields tend to be greater than ethane, except for the lowest NOx scenarios. However, the results are highly variable and dependent on other scenario conditions besides NOx.

A series of environmental chamber experiments and computer model calculations were carried out to assess the atmospheric ozone formation potentials of isopropyl acetate, 2-pentanone, and 2-heptanone. The experiments consisted of determining the effects the compounds on NO oxidation, ozone formation and OH radical levels when added to varying simulated model photochemical smog systems, and also included ketone - NOx - air irradiations. Isopropyl acetate was found to enhance O3 formation under all conditions examined, while the effects of the ketones on ozone depended on conditions. Atmospheric reaction mechanisms for the compounds were developed based on product data for isopropyl acetate and 2-pentanone and on various estimation methods, and these were evaluated by simulating the chamber experiments. The mechanism for isopropyl acetate fit the chamber data with only minor adjustment of the estimated overall nitrate yield, and the mechanisms for the ketones fit the data after adjusting the overall quantum yield for photolysis to form radicals. The best fit overall quantum yields were 0.1 for 2-pentanone and 0.02 for 2-heptanone, which are consistent with data for other compounds in indicating a monotonic decrease with the size of the molecule. Te mechanisms were used to predict the ozone impacts for these compounds under simulated atmospheric conditions. Isopropyl acetate was predicted to form about 1/3 to 1/2 as much ozone on a mass basis as the average of reactive VOC emissions from all sources, but about 3-4 times more ozone than ethane, the compound used by the EPA to define “negligible” ozone reactivity. The ketones were found to have comparable ozone impacts on a mass basis as the average reactive VOC emissions, with 2-heptanone being about 10% as reactive as 2-pentanone. The data obtained in this study were used in part in the development of estimation methods for the SAPRC-99 chemical mechanism.

Experiments were carried out to investigate the gas-phase reactions of simple bromine compounds under atmospheric conditions to provide data aimed at reducing uncertainties in model calculations of the atmospheric ozone impacts of propyl bromide and other bromine containing compounds. The reactions of HBr with O3, HBr with O3 and formaldehyde with NO or NO2 in the dark and upon irradiation, and the irradiations of Br2 in the presence of formaldehyde and/or NOx were studied. The experiments were carried out in air at atmospheric pressure and ambient temperature in a 5870 liter evacuable Teflon-coated chamber with quartz end windows and a xenon arc solar simulator light source. In-situ FT-IR spectroscopy was used to monitor most reactants and products. Most experiments used dry air, except for the HBr + O3 runs with varying humidity.

Ozone was found to react with HBr with an apparent rate constant of 4 x 10^-19 cm3 molec-1 s-1, with the data suggesting that HOBr is formed in the initial reaction, but that it reacts with HBr to form Br2, with a rate constant of at least 3.5 x 10^-15 cm3 molec-1 s-1. Humidity affected the HBr wall loss rate but did not affect the apparent O3 + HBr rate constant, suggesting that the reaction of HBr with O3 may not be surface dependent. In the presence of NOx the reactions of Br2 or HBr caused consumption or transformations of NO and NO2 and formation of varying amounts of BrNO and BrNO2. Attempts to model the experiments in the presence of NOx were unsuccessful.  The data obtained indicated that the mechanism used in the previous study of Carter et al (1997) to account for the observed ozone reactivity of propyl bromide was incorrect, and no mechanism was found that is consistent with the data. The data are also ambiguous concerning the role of wall effects in this system. It is concluded that considerable fundamental research is required on the reactions of bromine containing species in the presence of NOx before we can predictively model the atmospheric reactions of bromine containing organics.

The California Air Resources Board (CARB)’s air quality modeling procedures that are underway or being proposed for the 2003 ozone SIP are critically reviewed, and areas of concern and recommendations are summarized. This review is based on a reading of the September 18, 2000 draft of the CARB’s ozone SIP modeling protocol document, discussions and meetings with the CARB staff in June 2001, and the reviewer’s experience in chemical mechanism development and VOC reactivity modeling. This document gives a summary of what this reviewer sees as an ideal ozone SIP modeling procedure given the current state of knowledge and data availability, and discusses general and specific aspects of the current CARB modeling plan in light of these considerations. It is concluded that for the most part the models and modeling procedures being proposed for use represent the state of the art and incorporate significant improvements over past SIP modeling, but there are potentially significant concerns and recommendations. The major recommendations concern using more episodes in the control strategy modeling to represent the distribution of relevant conditions, discontinuing use of model components, such as the Carbon Bond mechanism, that are out-of-date and have known errors and biases, more comprehensive analysis of uncertainty and bias, and improved documentation of the process and results, especially to policymakers. Recommendations are also made concerning the process for review and external input.

The objective this project is to develop and employ the next-generation environmental chamber facility needed for evaluating gas-phase and gas-to-particle atmospheric reaction mechanism under more realistic conditions and with lower pollutant concentrations than previously has been possible. Progress was made towards this objective in several areas during the first two years of this project. A successful international workshop was held in Riverside California concerning the atmospheric chemistry of ozone and particle formation and environmental chamber research, and useful input concerning this project was obtained from environmental chamber and other researchers from the United States and Europe. Although the construction of the new facility is behind schedule because of a number of unanticipated delays, the new facility, which is housed in a new laboratory building designed primarily for this purpose, is now mostly completed and is expected to become operational in early 2002. Near-term needs for analytical instrumentation were assessed, and equipment was purchased and evaluated. A quality assurance project plan for experiments to be carried out in the new facility is being developed and a draft should be submitted to the EPA before the new facility becomes operational.

A series of experiments were carried out using smaller reactors to evaluate NOx offgasing effects in Teflon bag reactors, since this is expected to be the main factor limiting how low pollutant concentrations can be usefully employed in the new facility. These experiments were also used to evaluate analytical methods to be employed to monitor NOx species and formaldehyde at low concentrations. The results of the NOx offgasing tests indicated that minimum NOx offgasing rates of ~0.5 ppb per 24-hour day can be obtained in these small reactors if steps are taken to avoid contamination. This suggests that that useful mechanism evaluation data can be obtained at NOx levels as low as 2-5 ppb in these small reactors, and probably at lower levels in the larger reactors. No reaction wall material was found to be significantly better than the type of FEP Teflon® that is generally employed. The magnitude of the chamber radical source was measured in these Teflon reactors and was found to be dependent on the average NO2 concentration, with the results being consistent with the radical source measured in previous Teflon bag reactors at higher NOx levels. Several evaluation experiments were carried out at low NOx levels using the small reactors, and some inconsistencies with model predictions were found that will need to be investigated.

A proposed research plan for work to be carried out in the facility was developed and is presented in this report. This includes experiments that might be carried out through mid 2005 using funds not only from this project but also from current and anticipated projects for the California ARB to assess atmospheric impacts of coatings VOCs. The number of experiments that appears to be needed is sufficiently large that it may be difficult to conduct all of them during the period remaining in this project, and external input and review is needed to prioritize the research as well as to critically review the proposed research plan. A proposal for utilizing the Reactivity Research Working Group and paid peer reviewers for providing external input to this project is presented.

This project was aimed at developing improved and lower-cost alternative experimental procedures for evaluating chemical mechanisms for predicting ozone impacts of volatile organic compounds (VOCs). More precise measurements of effects of VOCs on OH radicals in chamber experiments could be obtained if 1,3,5-trimethylbenzene is used instead of m-xylene as the radical tracer, but our ability to model reactions of 1,3,5-trimethylbenzene need improvement before the data will reduce uncertainties in evaluations. The use of HONO + VOC irradiations to provide an alternative reactivity measurement to reduce ambiguities in mechanism evaluations, and extend the range of compounds that can be studied, was investigated. Calculations indicate that such experiments are more sensitive to the direct effects of VOCs on ozone formation than environmental chamber experiments. Plug flow experiments with short reaction times should provide the best measurement of direct reactivity of the VOC itself, while stirred flow experiments with longer reaction times are more useful for evaluating effects of the VOC's reactive products. A HONO generator was constructed that produces a continuous and stable output of HONO at the needed levels. Use of static, stirred flow, and plug flow systems were examined, with best results being obtained with a plug flow system using a 0.7" x 3' quartz tube reactor with a residence time of ~30 seconds. This was tested using most of the homologous n-alkanes through n?hexadecane, with results generally in good agreement with model predictions. However, low volatility compounds could not always be reliably and reproducibly injected and measured. Experiments with CO, 2-2-4-trimethylpentane, methyl ethyl ketone, ethyl acetate, propene, benzene, toluene, and 1,3,5-trimethylbenzene were also conducted, with the results indicating potential problems with the mechanisms for 2,2,4-trimethyl pentane and the aromatics. Improvements are needed to the system used to inject low volatility compounds, and a total carbon analysis system needs to be integrated into the experiment before the HONO flow method can reliably applied to low volatility compounds. Work on improving this method is underway as part of a new project for the CARB.

Cooperative efforts between the University of California at Riverside through the College of Engineering-Center for Environmental Research and Technology (CE-CERT) and the Mexican Petroleum Institute (IMP) have been carried out. The areas of collaboration were in air quality modeling simulation, environmental chamber experimentation, and chemical reaction mechanisms. During the two-year UCMEXUS project the IMP researchers visited CE-CERT on two separate occasions. Extensive discussions concerning these research areas were held and direct participation on experiments and training courses were undertaken. The bilateral technical effort between IMP and CE-CERT resulted in a number of achievements. Input was obtained on the design of an indoor environmental gas chamber laboratory to conduct reactivity studies of Volatile Organic Compounds (VOC) at IMP.  This effort materialized in the construction of a smog chamber facility designed to operate as a Dividable Teflon Chamber (DTC), similar to the gas chambers employed at CE-CERT. The second important area of collaboration focused on air quality modeling using the most recent atmospheric modeling system developed by the USEPA, known as Models-3/CMAQ. New meteorological and air quality data for the Mexico City valley that have become available will be used to conduct collaborative efforts on ozone and aerosol formation with CE-CERT researchers by October of this year. A third important area of interaction concerns the application of the SAPRC-99 chemical mechanism for model evaluation. SAPRC-99 is one of the most advanced explicit photochemical mechanisms developed by CE-CERT that can be used to study VOC reactivity in the airshed. The chamber database from the indoor environmental gas chamber simulations will be used to evaluate the SAPRC-99 chemical mechanism for predicting the effects of VOCs on ozone formation in Mexico City.

As the second phase of our study of the atmospheric ozone formation potentials of mineral spirits, which consist primarily of mixtures of C8 - C15 alkanes, a series of environmental chamber experiments and computer model simulations were carried out to assess the atmospheric ozone formation potentials of the selected representative branched alkanes, and an updated modeling analyses of previous experiments on representative mineral spirits samples was carried out. The experiments consisted of determining the effects on NO oxidation, ozone formation and OH radical levels when adding 2-methyl nonane, 2,6-dimethyl octane, or 3,4-diethyl hexane to varying simulated model photochemical smog systems. The OH radical rate constants for these compounds were measured to be 1.28 x 10^-11, 1.29 x 10^-11, and 7.96 x 10^-12 cm³ molec^-1 s^-1, respectively, using a relative rate method. These branched alkanes were found to be inhibitors of radical levels, and rates of NO oxidation and O3 formation in experiments that are sensitive to radical effects, but to a somewhat lesser extent than observed for normal alkanes and alkyl cyclohexanes in the same molecular weight range. The results were used to determine whether the current SAPRC-99 atmospheric chemical mechanism could accurately simulate the effects of these compounds on ozone formation and other manifestations of photochemical smog. The SAPRC-99 mechanism gave satisfactory simulations of the data for 2-methyl nonane, but tended to slightly underpredict the inhibiting characteristics of the more branched alkanes. The current mechanism also gave very good simulations of the results the previous experiments with actual mineral spirits samples, in contrast with the poor performance of the earlier mechanism used when modeling the results of these experiments when they were first reported. The current mechanism also performed significantly better in simulating the results of the branched alkane experiments. The reasons for the improved performance of the updated mechanism are discussed. The current mechanism was then used to calculate ozone impacts of the representative alkanes and mineral spirits samples in various urban photochemical smog scenarios. The impacts of the all-alkane mineral spirits samples on peak ozone yields were variable, but generally were 20-30% those of an equal mass of VOC emissions from all sources. The ozone impacts of the mineral spirits sample with 8% aromatics and alkenes were about 2-3 times those of the all-alkane samples. The relative impacts on maximum 8-hour average ozone levels were generally less than their relative impacts on peak ozone levels, especially in scenarios with relatively low NOx conditions.

Note: This report reflects work carried out through August, 2000. It was not finalized until July 2002 because of contractual problems.

The CAMx grid model was used to assess ozone reactivity effects for Carbon Bond (CB4) VOC species and ethane using the CRC-NARSTO database for the July 12-15, 1995 NARSTO-NE episode in the Eastern United States. The ozone sensitivities to emissions changes in NOx, total VOCs, total anthropogenic VOCs, CO, ethane and the 8 CB4 species used to represent major anthropogenic VOC emissions were calculated using DDM sensitivity analysis. A number of different ozone reactivity scales were derived using various methods to quantify the ozone impacts of the VOC species on the regional scale. These were based on effects of VOCs on daily maximum 1-hour averages in four different episode days, on effects on daily maximum 8-hour averages in three different episode days, and on using six different methods or metrics to derive regional reactivity scales from the varying impacts throughout the modeling domain. The results were compared to relative reactivities calculated with the same chemical mechanism in an EKMA box model used previously to derive the Carter reactivity scales.

The CAMx DDM results showed that there are differences in relative ozone impacts of VOC species with time and location, and that these differences are reflected in the effective ranges of the reactivity scales derived by the various methods. The effective range is the ratio of relative reactivities for the most reactive VOC species to ethane, the compound currently used by the EPA to define “negligible” reactivity. This varies from ~60 (on a carbon basis) for MIR and other scales that represent primarily urban impacts to ~20 to for regional average ozone and other scales that weigh impacts over large regions more equally. However, except for the CB4 TOL model species, whose relative O3 impacts are much more sensitive to NOx conditions than comparable model species in other mechanisms, the ordering of reactivity rankings are generally preserved regardless of which region of the domain or quantification method are employed. The relative reactivities derived from regional model results were generally consistent with the results using the EKMA scenarios when derived using comparable metrics. Using 8-hour vs. 1-hour ozone averaging time does not significantly affect relative reactivity scales.

A series of large-scale substitution calculations also were carried out where all anthropogenic VOC emissions were removed or replaced with varying amounts of ethane. The results were generally consistent with expectations based on the DDM first-order sensitivities. Replacing all anthropogenic VOCs (AVOCs) with equal mass or moles of ethane resulted in ozone reductions comparable to, but somewhat less than, removing all AVOCs. If ethane was added back to replace the AVOCs on a “reactivity neutral” basis, ozone tended to increase in the non-urban regions but decrease in VOC-sensitive areas dominated by urban emissions.

The ability SAPRC-99 atmospheric chemical mechanism to predict photochemical smog formation under low NOx conditions was evaluated by comparing model predictions to results experiments from three different environmental chamber facilities. These included new experiments from our UCR EPA environmental chamber, and previous experiments at from the Tennessee Valley Authority (TVA) and the Australian Commonwealth Scientific and Industrial Research Organization (CSIRO) chambers. The facility and procedures for the new UCR EPA experiments, and the procedures for modeling data from all three chambers, are discussed.

The results indicated no apparent low NOx mechanism performance problem for SAPRC-99 for simple chemical systems and for ambient surrogate reactive organic gas (ROG) - NOx experiments with ROG/NOx ratios high enough for maximum ozone formation potentials to be achieved. However, a consistent underprediction bias for NO oxidation and O3 formation rates was found in simulations ambient surrogate ROG - NOx experiments at low ROG/NOx ratios. The widely used Carbon Bond 4 mechanism was even worse in this regard. Furthermore, new aromatic - CO - NOx experiments indicate problems with current aromatic mechanisms that may be the cause of the low ROG/NOx underpredictions. Integrated reaction rate calculations indicate that increasing the accuracy in representing combination reactions of organic peroxy radicals will probably have an insignificant effect on model predictions. It is concluded that at a minimum new aromatic mechanisms need to be developed for model predictions to be consistent with available data.

An experimental and modeling study was carried out to reduce uncertainties in atmospheric ozone impacts of architectural coatings VOCs. The focus of this project was Texanol® (isobutyrate monoesters of 2,2,4-trimethyl-1,3-pentanediol), which is widely used in water-based coatings, and various hydrocarbon solvents representative of those used in solvent-based coatings. The hydrocarbon “bin” reactivity assignments developed by the CARB for hydrocarbon solvents were evaluated using compositional data from 124 different solvents, and a new methodology was developed for deriving such assignments that can be used when reactivity scales are updated or modified. Progress was made towards developing a direct reactivity measurement method that does not require gas chromatographic analyses, but additional work is needed before data can be obtained for solvents of interest. Environmental chamber experiments were carried out to evaluate the abilities of mechanisms of Texanol® and six different types of hydrocarbon solvents to predict their atmospheric ozone impacts, and comparable experiments were carried out with m-xylene and n-octane for control and comparison purposes. Chamber data were also used to derive rate constants for the reactions of OH radials with the Texanol® isomers that are in excellent agreement with current estimates. The UCR EPA environmental chamber was employed, and the experiments were carried out at NOx levels of 25-30 ppb and at ROG/NOx ratios representing maximum incremental reactivity (MIR) and NOx-limited conditions. The current SAPRC-99 mechanism was found to simulate the results of the experiments with Texanol® and the primarily alkane petroleum distillate solvents reasonably well, though uncertainties exist because of problems with the model simulating results of the MIR base case experiments. The mechanism also simulated the effects of Aromatic 100 on O3 formation in the MIR experiments, but underpredicted the tendency for the aromatics to inhibit O3 in NOx -limited experiments and had other problems. The results of the experiments with the synthetic C10-C12 isoparaffinic mixture were not well simulated by the model, and suggest that current mechanisms may underpredict their atmospheric reactivities by 25-75% depending on the source of the discrepancy. Recommended needs for additional research are discussed.

A new state-of-the-art indoor environmental chamber facility for the study of atmospheric processes leading to the formation of ozone and secondary organic aerosol (SOA) has been constructed and characterized. The chamber is designed for atmospheric chemical mechanism evaluation at low reactant concentrations under well-controlled environmental conditions. It consists of two collapsible 90 m3 FEP Teflon film reactors on pressure-controlled moveable frameworks inside a temperature-controlled enclosure flushed with purified air. Solar radiation is simulated with either a 200 kW Argon arc lamp or multiple blacklamps. Results of initial characterization experiments, all carried out under dry conditions, concerning NOx and formaldehyde offgasing, radical sources, particle loss rates, and background PM formation are described. Results of initial single organic - NOx and simplified ambient surrogate - NOx experiments to demonstrate the utility of the facility for mechanism evaluation under low NOx conditions are summarized and compared with the predictions of the SAPRC-99 chemical mechanism. Overall, the results of the initial characterization and evaluation indicate that this new environmental chamber can provide high quality mechanism evaluation data for experiments with NOx levels as low as ~2 ppb, though the results indicate some problems with the gas-phase mechanism that need further study. Initial evaluation experiments for SOA formation, also carried out under dry conditions, indicate that the chamber can provide high quality secondary aerosol formation data at relatively low hydrocarbon concentrations.

Environmental chamber experiments were carried out to assess the atmospheric ozone and particle matter (PM) impacts of selected representative VOCs emitted from architectural coatings. The UCR EPA environmental chamber was employed for the ozone and PM impact experiments, and most consisted of incremental reactivity experiments carried out at NO_x levels of 25-30 ppb and at ambient surrogate ROG/NO_x ratios representing maximum incremental reactivity (MIR) and NO_x-limited conditions. The compounds studied included the representative water-based coatings VOCs ethylene and propylene glycol, 2-(2-butoxyethoxy)-ethanol (DGBE), and benzyl alcohol. In addition, measurements of PM formation were made in experiments for these compounds and also in experiments with Texanol® (isobutyrate monoesters of 2,2,4-trimethyl-1,3-pentanediol), and several representative hydrocarbon solvents that were carried out for a separate project for the California Air Resources Board. Information was also obtained about PM background effects in the environmental chamber experiments.

The results of the chamber experiments were used to evaluate the predictions of the SAPRC-99 mechanism. The existing mechanism for DGBE was found to simulate the ozone reactivity data adequately, and a new mechanism for benzyl alcohol was developed that simulated the chamber data as well as mechanisms for other aromatics. The existing mechanisms for ethylene and propylene glycols were found to underpredict their ozone impacts by ~20% and 25-30% in some, but not all, experiments, but no scientifically acceptable basis was found to modify the mechanisms to improve these predictions. It is possible that that this is due to problems with the mechanisms for the aromatics present in the base ROG. The results of the experiments were also used to derive rate constants for the reactions of OH radicals with DGBE and benzyl alcohol relative to that for m-xylene, of 5.04 x 10^-11 and 2.56 x 10^-11, respectively. The rate constant for DGBE is in good agreement with the estimated value used by SAPRC-99, and that for benzyl alcohol is in good agreement with another measurement in the literature.

In terms of PM impacts in the incremental reactivity experiments, the relative ordering was found to be benzyl alcohol >> DGBE > petroleum distillates > a synthetic hydrocarbon solvent consisting mainly of branched alkanes " Texanol® > ethylene and propylene glycols. The benzyl alcohol was found to have a surprisingly high PM impact compared to other aromatics, and the glycols were found to actually reduce PM levels in the experiments, probably due to reducing rates of reactions of other VOCs present in the incremental reactivity experiment. No clear correlation between aromatic content and PM formation potential in the hydrocarbon solvents was seen. Background PM formation was observed in the chamber that will need to be characterized before these data can be used for PM model evaluation. Exploratory availability experiments were carried out to assess whether the presence of (NH₄)₂SO₄ and NH₄HSO₄ seed aerosol at levels up to ~10 µg/m³ and humidity up to ~10% RH affected the gas-phase loss rates or ozone formation potentials of ethylene and propylene glycol, but effects were seen.

Experiments with updated ambient reactive organic gas (ROG) surrogate mixtures that represents current emissions from mobile and other sources were also planned for this project, but they could not be carried out because of a lack of time and resources to derive a target composition for a new ROG surrogate within the timescale needed for this project.

Air quality simulation models are essential tools for assessing effects of emission controls on the formation of secondary pollutants such as ozone (O3). Volatile organic compounds (VOCs) can differ significantly in their effects on O3 formation, and control strategies aimed at reducing the O3 formation potentials (“reactivities”) of the emitted VOCs are being considered in addition to mass-based controls. The Statewide Air Pollution Research Center (SAPRC) chemical mechanism (Carter, 2000), which is well suited for VOC reactivity assessment, has been implemented in EPA’s Community Multiscale Air Quality (CMAQ) (Mebust, 2003) model, but additional work is needed to allow SAPRC to be used in CMAQ for assessment of reactivity controls.

Two major challenges must be addressed before CMAQ and SAPRC can be used effectively for reactivity assessment. First, the design of current emissions speciation databases is insufficient and needs to be corrected, because the databases do not clearly distinguish among actual chemical species, complex or poorly characterized mixtures, and model species used by specific mechanisms. Second, the emission inventory data processing to support SAPRC needs to be created. Both of these challenges are addressed by this work.

We integrated the improved speciation database and the SAPRC chemical mechanism with the Sparse Matrix Operator Kernel Emissions (SMOKE) (Houyoux and Vukovich, 2000) modeling system for CMAQ. Termed the Reactivity Modeling System, we will test this SMOKE/SAPRC system to ensure support of reactivity controls and the ability to create correct emissions used in the CMAQ model. Additionally, we will run CMAQ and analyze the results in the context of simulating reactivity substitutions.

Meeting the study’s objectives will result in the capability to process emission inventories to use in air quality modeling for evaluating the impact of reactivity controls. Specifically, the Reactivity Modeling System will enable changing the chemical makeup of the emissions using SMOKE and modeling the impact of this change on air quality. The specific goals and objectives of the project are as follows.

Create the speciation database organization and assignments that appropriately represent relationships between mixtures, actual chemical compounds, and the model species in SAPRC-99 and other mechanisms.

Develop the software and essential data to update both emissions processing methods and SMOKE for SAPRC-99 and other mechanisms using the reorganized speciation database and existing emissions datasets.

Test SMOKE with the redesigned speciation database for SAPRC-99 and other mechanisms, and for reactivity controls using data available from one previous modeling application.

Demonstrate application of reactivity controls to emissions using SAPRC-99 and the CMAQ modeling system.

Download Document
Link to Emissions Speciation Project Page

---

An experimental and modeling study was carried out to assess the ground-level atmospheric ozone impacts of representative pesticide-related VOCs. Environmental chamber experiments were carried out to develop and evaluate mechanisms for methyl isothiocyanate (MITC), S-ethyl N,N-di-n-propyl thiocarbamate (EPTC), 1-3-dichloropropenes, kerosene, and carbon disulfide. The first four are important compounds in the California pesticide emissions inventory where ozone impact data are not available, and carbon disulfide is a known pesticide degradation product. In addition, results of previous experiments on the pesticide chloropicrin were used to evaluate an updated mechanism for this compound. Chamber data were also used to derive rate constants for the reactions of OH radials with the MITC and EPTC, with the results indicating rate constants of 1.72 x 10-12 and 2.21 x 10-11 cm3 molec-1 s-1, respectively, which are both somewhat different from previously measured values. The UCR EPA environmental chamber was employed, and most experiments were "incremental reactivity" experiments to determine effects of adding the test compounds to experiments simulating representative ambient chemical conditions. These employed NOx levels of 25-30 ppb and at reactive organic gas (ROG)/NOx ratios representing maximum incremental reactivity (MIR) and NOx-limited conditions.

Mechanisms for the compounds studied were developed based on available information in the literature and the results of the chamber experiments, and the mechanism for chloropicrin was updated. Mechanisms for other thiocarbamates were estimated based on the mechanism derived for EPTC. The SAPRC-99 mechanism was used as the starting point, to which an updated chlorine mechanism was added so the reactions of the chlorine-containing compounds could be modeled. It was necessary to adjust uncertain portions of the mechanisms for MITC, CS2, and EPTC to give predictions that were consistent with the chamber data, and it was also necessary for the 1,3-dichloropropene mechanism to include an explicit representation of chloroacetaldehyde undergoing photolysis to form chlorine atoms at near-unit quantum yields to simulate the reactivity data for these compounds. The experiments with kerosene were found to be consistent with the predictions of the model derived from the available compositional data without the need for adjustments. The mechanisms were then used to derive quantitative ozone impacts for these and other pesticide compounds in the MIR and other incremental reactivity scales. The MIR values derived (in units of grams O3 per gram VOC) were 1,3-dichloropropenes: 5.4; chloropicrin: 2.2; EPTC and pebulate: 1.8; kerosene and molinate: 1.7; thiobencarb: 0.7; MITC: 0.35; and CS2: 0.25-0.28. (For comparison, the mixture used to represent reactive VOCs from all sources has an MIR of 3.7, and ethane has an MIR of 0.3).

In addition, relative impacts of these compounds on particulate matter (PM) formation were determined, with kerosene having the greatest PM impact on a mass basis, followed by the sulfur-containing compounds, and with the 1,3-dichloropropenes having no PM impact.

Areas of uncertainties and needs for future research are discussed.

An experimental and modeling study was carried out to assess the impacts of methyl iodide on ground-level ozone formation compared to other chemicals that are emitted into the atmosphere. The experiments consisted of environmental chamber irradiations of methyl iodide and NOx with and without added CO, methyl iodide and O3 with CO, and incremental reactivity environmental chamber experiments to determine the effect of adding methyl iodide to irradiations of reactive organic gas (ROG) surrogate - NOx mixtures representing ambient conditions. The results were modeled using the SAPRC-07 mechanism with the reactions of methyl iodide and iodine species added. The data were reasonably well simulated after adjusting uncertain parameters concerning photolysis rate of INO2 and the formation of IxOx oligomers from the reactions of IO radicals. This mechanism was then used to calculate the atmospheric impact of methyl iodide in the box model scenarios to derive the Maximum Incremental Reactivity (MIR) and other reactivity scales. Methyl iodide was calculated to inhibit ozone under all conditions except those with very high NOx levels, where its impact is comparable to that for ethane. It is concluded that methyl iodide should not be regulated as contributing to ground level ozone formation.

A completely updated version of the SAPRC-99 chemical mechanism, designated SAPRC-07, has been developed and is documented in this report. This includes a complete update of the rate constants and reactions based on current data and evaluations, reformulated and less parameterized aromatics mechanisms, a representation of chlorine chemistry, a reformulated method to represent peroxy reactions that is more appropriate for modeling secondary organic aerosol formation, and improved representations for many types of VOCs. This mechanism was evaluated against the result of ~2400 environmental chamber experiments carried out in 11 different environmental chambers, including experiments to test mechanisms for over 120 types of VOCs. The performance of the mechanism in simulating the chamber data was comparable to SAPRC-99, with generally satisfactory results for most types of VOCs but some increases in biases in simulations of some mixture experiments. The mechanism was used to derive an update to the MIR and other ozone reactivity scales for almost 1100 types of VOCs. The average changes in relative MIR values was about 10%, with >90% of the VOCs having changes less than 30%, but with larger changes for some types of VOCs, including halogenated compounds. Recommendations are given for future mechanism development research.

The mechanism documentation includes some large tabulations that are being provided only in electronic form. Links to downloading these tabulations are available at /~carter/SAPRC.

An experimental and modeling study was conducted to assess the ground-level atmospheric ozone impacts of several types of consumer product compounds Environmental chamber experiments were carried out for the representative amines 2-amino-2-methyl-1-propanol (AMP), ethanolamine, isopropyl amine and t-butyl amine and also for d-limonene. AMP and t-butyl amine were found to inhibit ozone formation, but the others enhanced ozone, and most were found to significantly enhance formation of secondary particle matter (PM). Methods to estimate mechanisms for amines that were qualitatively consistent with the chamber data were developed. However, the amine chamber data were not useful for quantitative valuation because the amount of amine available for gas-phase reaction could not be measured, and appeared to be significantly less than the amount injected. Estimates of atmospheric ozone impacts for the amines are also very uncertain because the amount of amines removed by reaction with HNO3 in the atmosphere cannot be predicted. The chamber data obtained for d-limonene were well simulated by the existing d-limonene mechanism.

Representations of the atmospheric reactions of for 15 amines and 30 other types of consumer product VOCs were added to the SAPRC-07 mechanism and its MIR and other reactivity scale tabulations, which are included with this report.
.

Download Document

Condensed versions of the SAPRC-07 mechanism, designated CS07A and CS07B, have been developed and are documented. They are derived directly from detailed SAPRC-07, which serves as the basis for their chemical validity and evaluation against chamber data. Both incorporate condensations involving removing or lumping less reactive compounds, lumping some product species in isoprene or aromatic mechanisms with other species with similar mechanisms using reactivity weighting, removing some compounds and reactions that are rapidly reversed, and using fewer model species to represent emitted alkanes and similar species. Mechanism CS07A is comparable in size to CB05 and incorporates the more condensed and approximate peroxy radical lumped operator method employed in SAPRC99, CB4, and CB05. It gives predictions of O3, total PANs and OH radicals that are very close to the uncondensed mechanism, but overpredicts H2O2 by about 15%. Mechanism CS07B retains the more detailed peroxy radical representation of uncondensed SAPRC-07, giving it ~40% more species than CS07A, and giving it much agreements in predictions of H2O2. Use of CS07A is suitable for models where the priority is O3 formation, while CS07B should be used if more accurate hydroperoxide predictions are a priority. Files for implementing these mechanisms are available at /~carter/SAPRC and /~carter/SAPRC/files.htm.
.

Download Document

---

INVESTIGATION OF ATMOSPHERIC OZONE IMPACTS OF TRANS 1,3,3,3-TETRAFLUOROPROPENE

William P. L. Carter

Final Report to Honeywell International Inc

February 9, 2009

An experimental and modeling study was carried out to assess the impacts of trans 1,3,3,3-tetrafluoropropene on ground-level ozone formation compared to other chemicals that are emitted into the atmosphere. The experiments consisted of incremental reactivity environmental chamber experiments to determine the effect of adding the tetrafluoropropene to irradiations of reactive organic gas (ROG) surrogate - NOx mixtures representing ambient conditions. The results were modeled using the SAPRC-07 mechanism with the reactions of the tetrafluoropropene and its oxidation products added. The data were reasonably well simulated after adjusting, to within its level of uncertainty, the overall nitrate yield in the reactions of NO with the peroxy radical intermediates. This mechanism was then used to calculate the atmospheric impact of the tetrafluoropropene in the box model scenarios to derive the Maximum Incremental Reactivity (MIR) and other ozone reactivity scales. Trans 1,3,3,3-tetrafluoropropene was calculated to have an ozone impact on a mass basis that was less than ethane under all the conditions simulated. The average ratio of mass-based incremental reactivities relative to ethane was 0.39±0.07 and the MIR ratio was 0.34. It is concluded that if ethane is used as the standard to define "negligible" ozone impact for the purpose of determining VOC exemptions for ozone precursors, then trans 1,3,3,3-tetrafluoropropene will meet this standard. The yields of fluorine-containing products formed in the oxidation of the tetrafluoropropene under various atmospheric conditions were also calculated.

.

Download Document

---

INVESTIGATION OF ATMOSPHERIC OZONE IMPACTS OF 2,3,3,3-TETRAFLUOROPROPENE

William P. L. Carter

Final Report to Honeywell International Inc

June 2, 2009

An experimental and modeling study was carried out to assess the impacts of 2,3,3,3-tetrafluoropropene on ground-level ozone formation compared to other chemicals that are emitted into the atmosphere. The experiments consisted of incremental reactivity environmental chamber experiments to determine the effect of adding the tetrafluoropropene to irradiations of reactive organic gas (ROG) surrogate - NO_x mixtures representing ambient conditions. The results were modeled using the SAPRC-07 mechanism with the reactions of the tetrafluoropropene added. The data were reasonably well simulated if it is assumed that nitrate formation from the reactions of peroxy radicals with NO was negligible. This mechanism was then used to calculate the atmospheric impact of the tetrafluoropropene in the box model scenarios to derive the Maximum Incremental Reactivity (MIR) and other ozone reactivity scales. 2,3,3,3-Tetrafluoropropene was calculated to have an ozone impact on a mass basis that is the same as that of ethane to within the variability of the model for atmospheric conditions. The average ratio of mass-based incremental reactivities relative to ethane was 0.9±0.2 and the MIR ratio was 1.0±0.1. It is concluded that if ethane is used as the standard to define “negligible” ozone impact for the purpose of determining VOC exemptions for ozone precursors, then 2,3,3,3-tetrafluoropropene might meet this standard. 2,3,3,3-Tetrafluoropropene was also found to have no significant effect on particle formation in the incremental reactivity chamber experiments..

Download Document

---

DETERMINATION OF ALDEHYDE AND PAN FORMATION POTENTIALS FOR VOLATILE ORGANIC COMPOUNDS

William P. L. Carter

Final Report to the California Office of Environmental Health Hazard Assessment Contract No. 00-E0027

September 23, 2001

The SAPRC 99 chemical mechanism and the one-day EKMA model scenarios we employed previously for calculating the MIR and other ozone reactivity scales were used to derive numerical factors quantifying the impacts of different types volatile organic compounds (VOCs) on formation of selected VOC oxidation products. Formation potentials for 12 oxidation products, including formaldehyde, acetaldehyde, lumped higher aldehydes, PAN, higher PAN analogues, PBzN, acrolein, lumped organic nitrates, and lumped aromatic product species, were calculated for 694 types of VOCs and mixtures and for 348 of the 373 emissions profiles used in the current California Air Resources Board emissions inventory. The tabulated results give averages of the direct and total effects of the VOCs on hourly average concentrations of the products in the 39 base case scenarios employed in the study. The use of maximum or final product concentrations as impact quantification methods were found to be much more dependent on scenario conditions and did not provide as consistent a basis for comparison as using effects on average concentrations for this purpose. The scenario-to-scenario variability of the direct impacts on average concentrations in the base case scenarios for the VOCs with the highest was 10-15% for the aldehyde products, and 25-35% for PAN and PAN analogues, with higher variability for compounds with lower impacts. However, greater variability may result from use of a comprehensive set of scenarios. The effects of chemical mechanism uncertainty, the need to update the scenarios representing atmospheric conditions, and considerations involving use of alternative quantification methods are discussed.

Because of their sizes, the tables giving the mechanisms of the individual VOCs, the compositions of the emissions input and profiles, and the product formation potential results for the individual VOCs and emissions profiles, are available only in the electronic version of the Appendix to this report. This is available at /~carter/pubs/aldrpt.xls.

.

Download Document

---

PROTOCOLS FOR EVALUATING OXIDANT MECHANISMS USED IN URBAN AND REGIONAL MODELS

H. E. Jeffries, M. W. Gery, W. P. L. Carter

Final Report for U.S. EPA Cooperative Agreement 815779

June, 1992

In this project, a task force of three chamber operators and three modelers was assembled to address needs raised at a prior workshop on the procedures and practices that should be followed when evaluating photochemical reaction mechanisms for their suitability for use in EPA's urban and regional air quality models. In addition to the members of the task force, two workshops were held that were attended by 18-20 researchers in the field including scientists from England and Australia. At these meetings, issues raised by the task force were discussed and commentary on the approach was provided.

In this report, based on our work and the community input, we describe how to create a protocol for the evaluation of photochemical reaction mechanisms. Rather than prescribe a set of specific actions, we present instead criteria that influence decisions without specifying what those decisions must be. These specify what the evaluator must consider, what is and is not relevant, and what must be reported as the basis of decisions made.

Based on general scientific principles, we describe five characteristics that reaction mechanisms must have if they are to be acceptable. Mechanism evaluation is more than just establishing agreement of the model with observations. Our approach is based on a complex argument in the form of a cascaded inference chain showing how to proceed to establish that a candidate mechanism might exhibit all five characteristics. The evidence in this argument is mostly chamber data. The argument rests on the credibility of the chamber data and thus, the data's credibility must be communicated by the evaluator by reporting how he was convinced the data are credible.

In five chapters the report details the elements that must be considered and describes the general content of reports the evaluator must produce.

While we now have a well reasoned and logically supported procedure for evaluating a mechanism, making the case that a particular mechanism is accurate is presently a somewhat difficult undertaking mainly because of the data incompleteness problem (discussed in Chapter 6). Although a large number of experiments are available that were produced over more than 15 years by three different chamber groups, the number of experiments that are capable of providing a compelling evaluation of the various mechanism components is quite limited. Thus, at present, it is not possible to make a compelling case for accepting a mechanism for ambient air use, only to vindicate its use as having been evaluated the best that can be currently done, given the available data. Clearly, additional chamber data are needed for the "standard chamber data base."

Download Document

---

INVESTIGATION OF ATMOSPHERIC OZONE IMPACTS OF
TRANS 1-CHLORO-3,3,3-TRIFLUOROPROPENE

William P. L. Carter

Final Report

June 8, 2009

An experimental and modeling study was carried out to assess the impacts of trans 1-chloro-3,3,3-trifluoropropene on ground-level ozone formation compared to other chemicals that are emitted into the atmosphere. The experiments consisted of incremental reactivity environmental chamber experiments to determine the effect of adding the halopropene compound to irradiations of reactive organic gas (ROG) surrogate - NOx mixtures representing ambient conditions. The results were modeled using the SAPRC-07 mechanism with the reactions of the halopropene added. The data were reasonably well simulated after adjusting, to within its level of uncertainty, the overall nitrate yield in the reactions of NO with the peroxy radical intermediates. This mechanism was then used to calculate the atmospheric ozone impact of this compound in the box model scenarios to derive the Maximum Incremental Reactivity (MIR) and other ozone reactivity scales. Trans 1-chloro-3,3,3-trifluoropropene was calculated to have an ozone impact on a mass basis that is the that was less than ethane under all the conditions simulated. The average ratio of mass-based incremental reactivities relative to ethane for all the box model scenarios was 0.19±0.03 and the MIR ratio was 0.16±0.02. It is concluded that if ethane is used as the standard to define “negligible” ozone impact for the purpose of determining VOC exemptions for ozone precursors, then trans 1-chloro-3,3,3-trifluoropropene will meet this standard. The yields of halogen-containing products formed in the oxidation of this compound under various atmospheric conditions were also calculated. Trans 1-chloro-3,3,3-trifluoropropenewas also found to have no significant effect on particle formation in the incremental reactivity chamber experiments.

Download Document

---

INVESTIGATION OF ATMOSPHERIC OZONE IMPACTS OF 3-METHOXY-3-METHYL-1-BUTANOL

William P. L. Carter, Wendy S. Goliff, Roger Atkinson, Janet Arey and Sara M. Aschmann

Final Report

July 8, 2010

An experimental and modeling study was carried out to assess the impacts of 3-Methoxy-3-methyl-1-butanol (MMB) on ground-level ozone formation. The rate constant for the reaction of MMB with OH radicals was measured to be (1.64 ± 0.06) x 10^-11 cm³ molecule^-1 s^-1 at 296 ± 2 K, which is over a factor of 2 higher than previously estimated. Acetone, methyl acetate, and 3 methoxy-3-methylbutanal were identified as products of this reaction, with yields of 3 ± 1%, 34 ± 8, and approximately 33-47%, respectively. Glycolaldehyde was also observed but was not quantified. These data were used to derive a mechanism for the atmospheric reactions of MMB, for use in calculating its atmospheric ozone impacts using the SAPRC-07 mechanism. To evaluate the predictive capability of this mechanism, incremental reactivity environmental chamber experiments were carried out to determine the effects of adding MMB to irradiations of reactive organic gas (ROG) surrogate - NOx mixtures representing ambient conditions. The results were reasonably consistent with the new MMB mechanism derived in this work. This mechanism was then used to calculate the atmospheric ozone impact of MMB in the MIR and other SAPRC-07 ozone reactivity scales. The MIR calculated for MMB was 2.78 grams O₃ per gram VOC, which is higher than calculated for it previously using an estimated mechanism, but slightly lower than the MIR for the mixture used to represent anthropogenic VOCs from all sources in the reactivity calculations. The environmental chamber experiments also indicated that MMB does not form measurable secondary organic aerosol under the conditions of our experiments.

Download Document

---

ENVIRONMENTAL CHAMBER STUDIES OF OZONE IMPACTS OF COATINGS VOCs

William P. L. Carter

Final Report for California Air Resources Board Contract 07-339

May 11, 2011

An environmental chamber and modeling study was conducted to reduce uncertainties in atmospheric ozone impacts for volatile organic compounds (VOCs) emitted from coatings. Some coatings VOCs (Texanol® and low-aromatic petroleum distillates) have near-zero or negative incremental ozone reactivities in chamber experiments, but calculations show positive ozone impacts in the atmosphere. Modeling indicated that experiments with increased light intensity and added H2O2 should give reactivities that better correlate with those in the atmosphere. After upgrading our chamber's light source, experiments to test the new method performed as expected, and gave good correlations between experimental and atmospheric MIR values for the VOCs tested. These experiments also appear to be more sensitive to effects of VOCs on secondary organic aerosol (SOA) formation than previous experiments. Such experiments should be included in future environmental chamber reactivity studies, though other types of experiments are also needed for adequate mechanism evaluation.

Experiments were also conducted to assess ozone impacts of ethyl methyl ketone oxime (EMKO), and soy ester solvents. The EMKO results indicated it has both radical sinks and NOx sources in its mechanism, and has no measurable impact on SOA formation. The EMKO mechanism that simulated the data gave a negative MIR of -1.27 gm O3 /gm VOC, but positive MOIR and EBIR values of 0.41 and 1.14, respectively. This has implications about the use of the MIR scale for such compounds. The experiments to assess soy ester reactivity were not successful because of the low volatility of the constituents, but uncertainties concerning atmospheric availability are probably more important.

Download Document

---

DEVELOPMENT OF REVISED SAPRC AROMATICS MECHANISMS

William P. L. Carter and Gookyoung Heo

Final Report to California Air Resources Board Contracts No. 07-730 and 08-326

April 12, 2012

The representation of the gas-phase atmospheric reactions of aromatic hydrocarbons in the SAPRC-07 mechanism has been updated and revised to give better simulations of recent environmental chamber experiments. The SAPRC-07 mechanism consistently underpredicted NO oxidation and O3 formation rates observed in recent aromatic - NOx environmental chamber experiments carried out using generally lower reactant concentrations than the set of experiments used to develop SAPRC-07 and earlier mechanisms. The new aromatics mechanism, designated SAPRC-11, was evaluated against the expanded chamber database and gave better simulations of ozone formation in almost all experiments, except for higher (>100 ppb) NOx benzene and (to a lesser extent) toluene experiments where O3 formation rates were consistently overpredicted. This overprediction can be corrected if the aromatics mechanism is parameterized to include a new NOx dependence on photoreactive product yields, but that parameterization was not incorporated in SAPRC-11 because it is inconsistent with available laboratory data. The new version incorporates a few minor updates to the base mechanism concerning acetylene, glyoxal and acyl peroxy + HO2, has new parameterized mechanisms for phenolic compounds, and incorporates modifications and readjustments to the parameterized mechanisms representing reactive ring-opening products, but otherwise is the same as SAPRC-07. The new mechanism gives up to ~15% higher ozone concentrations under maximum incremental reactivity (MIR) conditions and gives ~0-50% higher MIR values for most aromatic compounds, and much higher reactivities for benzene and phenolic compounds. However, the mechanism revision has relatively small effects on O3 predictions under NOx-limited conditions, and the MIR values for non-aromatic compounds are not significantly affected.

NOTE: The version of SAPRC-11 described in this report has been superceded by the version in the journal article Carter, W. P. L. and G. Heo, “Development of Revised SAPRC Aromatic Mechanisms,” Atmos. Eviron., 77, 404-414, 2013. This should be used if you want to conduct model simulations using this mechanism.

Download Document

---

SOA FORMATION: CHAMBER STUDY AND MODEL DEVELOPMENT

William P. L. Carter, Gookyoung Heo, David R. Cocker III, and Shunsuke Nakao

Final Report to California Air Resources Board Contract 08-326

May 21, 2012

An experimental and mechanism development study was carried out to enhance the recently developed SAPRC-11 gas phase aromatic mechanism so it can predict secondary organic aerosol (SOA) formation from the atmospheric reactions of aromatics. This phase of the project covered dry conditions and 300K. A total of 158 dual reactor chamber experiments were carried out using the UCR-EPA environmental chamber, and their results were combined with previous data from this chamber to provide a database of 315 separate reactor irradiations for mechanism evaluation. A total of 14 representative aromatic hydrocarbons and 7 representative phenolic compounds were studied with varying reactant and NOx levels and in some cases with different light sources and other added reactants Methods were developed and evaluated to represent gas-particle partitioning, nucleation, and chamber effects when modeling the experiments. Alternative mechanisms were examined and SOA yield and gas-particle partitioning parameters were optimized to simulate the available chamber data. The model simulated most of the data without large biases but with larger run-to-run variability in model performance than observed in ozone mechanism evaluations, and potential evaluation problems were observed for some compounds. It is concluded that this new mechanism reflects the current state of the science. Recommendations are given for the next phase of SOA mechanism development and other needed research.

Download Document

---

[Downloadable Documents] [Carter Publications] [Carter Home Page]

[Top of Page]

This page created by

Send comments or questions to carter@cert.ucr.edu

Last Modified May 29, 2013