Course Hero - We put you ahead of the curve!

You have requested the below document.

Coursehero >> Texas >> Texas >> DECHAPANYA 029

by Copyright Wipawee Dechapanya 2002 Kinetic and Physic Models of Secondary Organic Aerosol Formation and their Application to Houston Conditions by Wipawee Dechapanya, M.S. Dissertation Presented to the Faculty of the Graduate School of the University of Texas at Austin in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy The University of Texas at Austin December 2002 I would like to dedicate this dissertation to Raweewan Dechapanya, my mother. ACKNOWLEDGMENTS I would like to thank my advisor, Prof. David T Allen, for his support, guidance and patience throughout this research project. I also would like to thank my committee: Dr Elena McDonald Buller and Dr. Richard Corsi of the Department of Civil and Environmental Engineering, and Dr. Gary Rochelle and Dr. Buddy Mullins of the Department of Chemical Engineering, for their continued support and advice. Thanks also to Dr. Alexandra Eusebi for assisting me the guidance, knowledge, and mental support. A very special thank to Dr. Yosuke Kimura, this work would not have been done without his help and patience. I also would like to thank my friends at center for energy and environmental resources (CEER). I am deeply indebted to the Royal Thai Government who provides me with the financial support throughout my studies in the United States. I would like to express my gratitude to my mother; Raweewan, my sister; Wipada, my cats; Nu-Meaw and especially Midas, and all my friends in Thailand for their love and greatest support. Finally, a very special thank you to my dearest husband, Attaso Khamwichit, for his continued encouragement, support, and love. v Kinetic and Physic Models of Secondary Organic Aerosol Formation and their Application to Houston Conditions Publication No.______________ Wipawee Dechapanya, Ph.D. The University of Texas at Austin, 2002 Supervisor: David T Allen Atmospheric reactions of volatile organic compounds can produce low volatility species that condense onto atmospheric particles (secondary organic aerosol), and these particles have significant impact on public health. This work develops quantitative kinetic and physical phase partitioning models of secondary organic aerosol (SOA) formation. These mechanisms were integrated into a state of the art mechanism for gas phase reactions (SAPRC). Using the resulting model, a series of sensitivity analyses were performed. Analyses of the sensitivity of SOA formation to several parameters (e.g., VOC/NOx ratio, rate parameters) were performed. Results indicated that aerosol yield (SOA formed per amount of hydrocarbons reacted) depends on the extent of conversion of parent hydrocarbons, partitioning coefficient (Kom), initial aerosol mass concentration (Mint), and rate vi parameters. Based on the sensitivity studies, empirical models for SOA yield were developed for both individual and lumped hydrocarbon species. The models were used to examine a number of case studies relevant to the formation of SOA in Houston. In general, the analyses indicated that strategies effective in reducing ozone concentrations will also be effective in reducing SOA. Emission reductions that reduce ozone mixing ratios by O.1 ppm reduces SOA concentration by approximately 2.5 g/m3. The models developed in this study are effective prognostic tools for analyzing SOA production under a variety of conditions, and these models can be readily implemented into 3D air quality models, and modified easily if more experimental data on SOA formation become available. vii Table of Contents Abstract Table of Contents List of Tables List of Figures Chapter 1 Introduction 1.1Background and Motivation 1.2 Research objective 1.3 Dissertation overview Chapter 2 Literature Review 2.1 Organic Particulate Matter 2.1.1 Sources of atmospheric PM 2.1.2 Organic PM Formation 2.1.3 Composition and Sources of Secondary Organic Aerosol 2.1.4 Incremental Ozone Reactivity 2.2 PM modeling 2.2.1 Background 2.2.2 PM Models 2.2.3 Treatment of SOA in PM models 2.3 Summary vi viii xii xvi 1 1 3 4 6 6 8 10 11 18 21 21 23 26 30 viii Chapter 3 Quantitative models of SOA formation for individual compounds development 3.1 Introduction 3.2 SAPRC model 3.3 Quantitative models of SOA formation 3.3.1 Previous smog chamber investigations 3.3.2 Development of quantitative models of SOA formation 3.4 G/P Partitioning model 3.4.1 Theoretical Background 3.4.2 Estimates of Model Parameters 3.4.3 Sensitivity Analysis of SOA yields: Case study for 1,3,5trimethylbenzene 31 31 31 57 57 58 75 75 78 84 3.5 Quantitative models of SOA formation for various aromatic hydrocarbons 103 3.5.1 Quantitative mechanistic models for aromatic hydrocarbons 3.5.2 Comparison of SOA yields 3.5.3 Correlation between M and Mint for aromatic hydrocarbons 3.6 Summary Chapter 4 Quantitative models of SOA formation for lumped aromatic speciess development 4.1 Analysis of incremental aerosol reactivity 4.2 SOA formation in Eulerian photochemical models 4.3 Quantitative models of SOA formation for lumped species 4.3.1 Theoretical Background ix 104 115 118 126 129 129 132 133 133 4.3.2 Lumping Scheme 4.3.3 Estimates of Model Parameters 4.3.4 Correlation between SOA yields and chemical and physical parameters 4.4 Summary Chapter 5 Characterization of ambient SOA formation in Houston area 5.1 Introduction 5.2 Case Study Scenario 5.2.1 September 2000 Episode 5.2.2 Case Study Description 5.3 Case Study Results 5.3.1 Case A: Box model 5.3.2 Case B: Box model with NOx emissions 5.3.3 Case C: Box model with NOx and VOC emissions 5.4 Summary Chapter 6 Summaries and Future Work 6.1 Summaries Key Conclusions 6.2 Future Work Nomenclatures Appendix A: Data from smog chamber experiment 135 144 161 164 166 166 167 171 173 178 178 186 191 197 199 199 199 200 201 205 Appendix B: Houston atmospheric hydrocarbon composition, obtained from the CARB Report (Cater, 1998) 207 x Appendix C: Reaction pathways and mechanistic models of SOA formation of aromatic hydrocarbons modeled in SAPRC-99 Appendix D: Conditions used for SAPRC simulation for case study of aromatic precursors Appendix E: Aerosol mass changes of PM3-PM8 from lumped APR2-APR4 Bibliography Vita 210 309 310 312 324 xi List of Tables Table 2.1: VOC Emissions Inventory and Secondary Aerosol Yields (Grosjean, 1992) Table 2.2: Summary of Major Characteristics of the Primary Carter Reactivity Scale 15 20 Table 2.3: Incremental ozone reactivities for selected hydrocarbons, evaluated using three different initial mixtures of air pollutants (Carter, 1994) 20 Table 2.4: Host Air Quality Model, Meteorological Model and Applications to Date of Episodic PM Models (Seigneur, 1997) 25 Table 2.5: Treatment of SOA formation in PM models (Seigneur, 1997) Table 3.1: Detailed species that are modeled in SAPRC 99 (Carter, 2000) 27 32 Table 3.2: Summaries of lumped classes and lumped molecule employed in SAPRC99 model 54 Table 3.3: VOC Emission Inventory and Secondary Aerosol Yields of Hydrocarbon Precursors Selected for Study 59 Table 3.4: Kinetic parameters of 1,3,5-trimethybenzene reactions with OH radicals forming aerosol products and their sources 65 Table 3.5: Kinetic parameters of 1,2,4-trimethylbenzene reactions with OH radicals forming aerosol products and their sources 69 Table 3.6: Condensed mechanisms and rate constants for 1,3,5-and 1,2,4trimethylbenzene Table 3.7:Partitioning Parameters used in the estimates of SOA formation from reactions of hydrocarbon precursors Table 3.8: The Estimated Partitioning Coefficients and Vapor Pressures of Condensable Products from trimethylbenzenes 74 78 81 xii Table 3.9: A set of simulation conditions for 1,2,4-trimethylbenzene, obtained from Odum's experiments 82 Table 3.10: Initial conditions for box simulation for 1,3,5-trimethybenzene in the presence of Houston air pollutants. 85 Table 3.11: Condensed mechanisms of SOA formation for 15 aromatic precursors developed in this study 108 Table 3.12: The Estimated Partitioning Coefficients and Vapor Pressures of Condensable Products from aromatic hydrocarbons 114 Table 3.13: Aerosol yields for low- and high-yield aromatics: results from simulations and observation 116 Table 3.14: Coefficients a, b, and c for the correlation equation between %DFi and Mint for 11 aromatic hydrocarbons 126 Table 3.15: Chemical and physical model parameters for the estimate of secondary aerosol formation 127 Table 4.1: Incremental aerosol reactivity for 11 aromatic compounds Table 4.2: Lists of speciess represented by model species ARO1 and ARO2 130 136 Table 4.3: Lists of detailed species and their OH rate constants represented by model species APR1, APR2, APR3, and APR4 137 Table 4.4: Model Parameters Used to Simulate SOA Formation from Lumped Aromatic Species 144 Table 4.5: The estimated parameters for lumped species from fitting method (FT) and weighting method (WF) 160 Table 4.6: The estimated molecular weights of lumped species from weighting method Table 4.7: Summaries of model parameters for lumped species 160 161 Table 4.8: Coefficients a, b, and c for the correlation equation between %DFi and Mint for lumped species APR1-APR4 164 xiii Table 5.1: Case studies for characterization and sensitivity analysis of SOA formation Table 5.2: Initial conditions used in box model simulations for the September episode 171 174 Table 5.3: Emission rates of NOx at Houston Ship Channel area, fraction of NO is equal 85% 176 Table 5.4: SOA yields of each aerosol products on the September episode Table A.1: Relative molar loading 183 205 Table A.2: Predicted compositions of SOA from the photooxidation of hydrocarbon precursors 205 Table B.1: Houston atmospheric hydrocarbon comppsitions used in the case study for sensitivity analysis 207 Table C.1: Kinetic parameters of 1,2,3-trimethylbenzene reactions with OH raidicals forming aerosol products and their sources 215 Table C.2: Kinetic parameters of 1,2,4-trimethylbenzene reactions with OH raidicals forming aerosol products and their sources 222 Table C.3: Kinetic parameters of o-xylene reactions with OH raidicals forming aerosol products and their sources Table C.4: Kinetic parameters of p-xylene reactions with OH raidicals forming aerosol products and their sources Table C.5: Kinetic parameters of m-xylene reactions with OH raidicals forming aerosol products and their sources 229 233 239 Table C.6: Kinetic parameters of toluene reactions with OH raidicals forming aerosol products and their sources 244 Table C.7: Kinetic parameters of benzene reactions with OH raidicals forming aerosol products and their sources 247 Table C.8: Kinetic parameters of ethylbenzene reactions with OH raidicals forming aerosol products and their sources 254 xiv Table C.9: Kinetic parameters of n-propylbenzene reactions with OH raidicals forming aerosol products and their sources 260 Table C.10: Kinetic parameters of iso-propylbenzene reactions with OH raidicals forming aerosol products and their sources 266 Table C.11: Kinetic parameters of sec-butylbenzene reactions with OH raidicals forming aerosol products and their sources 274 Table C.12: Kinetic parameters of p-ethyltoluene reactions with OH raidicals forming aerosol products and their sources 281 Table C.13: Kinetic parameters of o-ethyltoluene reactions with OH raidicals forming aerosol products and their sources 294 Table C.14: Kinetic parameters of m-ethyltoluene reactions with OH raidicals forming aerosol products and their sources 307 Table D.1: Simulation conditions for aromatic precursors, obtained from Odum's experiments 309 xv List of Figures Figure 2.1: Sources of atmospheric particulate matter (Seinfeld, et al., 1994) Figure 2.2: Spatial distribution of major chemical components during the Texas PM2.5 Sampling and Analysis Study between 03/11/97 and 03/12/98 Figure 2.3: Schematic Representation of a PM Source Model (Seigneur, 1997) 8 10 22 Figure 3.1: Overview of relationships between data files and programs used in the preparation of chemical mechanism for airshed model calculation (Carter, 1988) 56 Figure 3.2: Hydrogen abstraction from 1,3,5-trimethybenzene by OH radical Figure 3.3: OH addition to 1,3,5-trimethybenzene reaction pathways Figure 3.4: Semi-volatile product reaction pathways from aerosol precursors Figure 3.5: The H-abstraction from aromatic ring by the reaction of OH and 1,2,4-TMB Figure 3.6: The OH-addition from aromatic ring by the reaction of OH and 1,2,4 TMB Figure 3.7: The ring fragmentation by the reaction of OH and 1,2,4-TMB Figure 3.8: Aerosol product formation from oxidation reaction of 1,2,4-TMB 60 60 61 65 66 66 66 Figure 3.9: SOA yields for 1,2,4-trimethylbenzene, NOx = 359-1178 ppb, C3H6 = 300 ppb, and HC/NOx = 4.3-7.7 ppbC/ppb 83 Figure 3.10: SOA yields for 1,3,5-trimethylbenzene represented as %conversion and C.D.T, Houston conditions 87 Figure 3.11: SOA yields for 1,3,5-trimethylbenzene at different base hydrocarbon composition represented against %conversion and C.D.T., Houston conditions 89 Figure 3.12: SOA yields for 1,3,5-trimethylbenzene for rate constant of aerosol precursor varies from 1.0E-9 to 1.0E-13 cm3/molc-sec, Houston conditions 90 xvi Figure 3.13: SOA yields for 1,3,5-trimethylbenzene for organic particulate mass (Mint) equals 5 and 15 g/m3, Houston conditions 92 Figure 3.14: SOA yields for 1,3,5-trimethylbenzene for the original, double, and triple values of Kom,1 and Kom,2. Simulation conditions are based on the CARB report (Carter, 1998). 93 Figure 3.15: versus % conversion of parent hydrocarbons from Odum model with 4 parameters and modified SAPRC model with single parameter 96-97 Figure 3.16: %DF and M as a function of Mint at 25%, 50%, and 75% 1,3,5trimethylbeznene conversions. Simulation conditions were obtained from the CARB report (Carter, 1998) 102 Figure 3.17: Hydrogen abstraction from o-xylene by OH radical reaction pathway 104 Figure 3.18: Addition of OH to o-xylene forming methylcresol Figure 3.19: OH addition to o-xylene reaction pathways Figure 3.20: Semi-volatile product reaction pathways from aerosol precursors Figure 3.21: SOA yields for high- and low-yield aromatics from simulation calculation compared to chamber experiments 104 105 105 116 Figure 3.22: %DFi as a function of Mint at 25, 50, and 75% hydrocarbon conversions for 11 aromatics 119-123 Figure 3.23: M for toluene presented against Pi 125 Figure 4.1: Lumping scheme for aromatic species and aerosol precursors modified in SAPRC 142 Figure 4.2: Concentrations of aerosol precursors as a function of time, three sets of data: (1) summation of concentrations of individual species in the group, (2) obtained by using fitting method, and (3) estimated using weighting factor 151 Figure 4.3: Concentrations of condensable products, three sets results: summation of concentrations from individual species, concentration of lumped species obtained by using the fitting method, and by using weighting factor 154 xvii Figure 4.4: SOA mass changes from four lumped aerosol species: three sets of data point: Mi from individual species in the groups, M for lumped products obtained from fitting and weighting methods 157 Figure 4.5: %DFi as a function of Mint at 25, 50 and 75% conversions for lumped species APR1-APR4 163 Figure 5.1: Average concentrations of constituents found in PM2.5 collected between 3/97 and 3/98 (Tropp et. al., 1998) for various Texas cities. 167 Figure 5.2: The annual average OC, EC, and OC/EC ratio at Houston areas on February 4, 1998, data was extracted from the Desert Research Institute PM2.5 study of 1997/1998 169-170 Figure 5.3: Hydrocarbon concentrations observed from TexAQS at the Clinton site on September 1 and 2, 2000 at 4:00 pm 172 Figure 5.4: Monitoring speciated concentrations of aromatic hydrocarbons on September 1 and 2, 2000 at the Clinton site at 4:00 pm Figure 5.5: NOx and VOC point sources in Houston Ship Channel area 1998) Figure 5.6: VOC point sources in Ship Channel area from 1996 inventory 173 (TNRCC, 175 178 Figure 5.7: SOA mass changes (Mo) at the Clinton site at 4:00 pm on the September 179 episode, aerosol seed varies from 5-15 g/m3. Figure 5.8: Speciated aerosol mass changes of eight aerosol products (PM1-PM8) during the September episode, Mint = 5 g/m3 181 Figure 5.9: Ozone concentrations at the Clinton site on the September episode from 4:00 pm, data points are from case study simulation, gas-phase chemistry simulation, and monitors 184 Figure 5.10: Ozone concentrations on September 1, 2000 around 3:30 to 4:30 pm, at the Ship Channel area, Houston TX 185 Figure 5.11: SOA mass changes at the Clinton site from 4:00 on September 1,2000: results from varying NOx emissions by 10-100% from base case, Mint = 5 g/m3, fraction of NO = 0.85 188 xviii Figure 5.12: Ozone concentrations at the Clinton site from 4:00 on September 1, 2000: results from varying NOx emissions by 10-100% from base case, Mint = 5 189 g/m3, fraction of NO = 0.85 Figure 5.13: SOA yields at the Clinton site on September 1, 2000 from 4:00 pm, results from changing NOx emissions 50% from base case, and varying fraction of NO from 0.85-0.95, Mint = 5 g/m3 190 Figure 5.14: Aerosol mass changes as the puff of air from the Clinton site move along the Houston Ship Channel area for case C2, results from varying NOx emissions by 10-95% from base case, SOA seed = 5 g/m3, and fraction of NO = 0.85 193 Figure 5.15: Ozone concentrations as the puff of air from the Clinton site move along the Houston Ship Channel area for case C2, results from varying NOx emissions by 10-95% from base case, SOA seed = 5 g/m3, and fraction of NO = 0.85 194 Figure 5.16: Aerosol mass changes as the puff of air from the Clinton site move along the Houston Ship Channel area for case C1, results from varying NOx emissions by 10-95% from base case, SOA seed = 5 g/m3, and fraction of NO = 0.85 195 Figure 5.17: Ozone concentrations as the puff of air from the Clinton site move along the Houston Ship Channel area for case C1, results from varying NOx emissions by 10-95% from base case, SOA seed = 5 g/m3, and fraction of NO = 0.85 196 Figure 5.18: O3 versus SOA compared to basecase NOx emissions when NOx emissions were decreased to 100%, and increased to 70% from the basecase 197 Figure E.1: Amount of PM3 - PM8 Produced and parition into particulate phase as a fuinction of time 310 xix Chapter 1 Introduction 1.1Background and Motivation New National Ambient Air Quality Standards (NAAQS) for fine airborne particles smaller than 2.5 m, were issued in July 1997 by the administrator of the U.S. Environmental Protection Agency (EPA). This legislation was brought about due to recent epidemiological evidence that indicates a strong relationship between elevated concentrations of fine particulate matter and increased mortality (Schwartz, 1996). Studies from EPA (1999) indicate that approximately 76 million people (29% of the total U.S. population) lived in areas in 1996 where long term ambient PM2.5 levels were predicted to be at or above 16 g/m3, the low end of the range of long term average PM2.5 concentrations which may cause serious health effects. Reducing PM2.5 concentrations to meet the new standards is likely to require substantial reductions in emissions leading to fine particulate matter, and therefore a thorough understanding of fine particulate matter sources has become necessary. Airborne particulate matter may be liquid or solid and can contain acids, salts, heavy metals, organics, and biological or biogenic material. Sources of ambient particles that are emitted directly to the atmosphere include forest fires, wind erosion, agricultural activities, fossil fuel combustion, industrial manufacturing, and construction of buildings and roads. These particles emitted directly into atmosphere 1 are referred to as "primary" particulate matter. In contrast, some particulate matter is formed as a result of reactions that occur in the atmosphere, and are referred to as "secondary" particulate matter. This secondary particulate matter includes sulfate produced by the reactions of SO2 emissions, nitrates produced as a result of the reactions of NOx, and low volatility organics produced as a result of the reactions of hydrocarbons. Ambient particulate matter is made up of both inorganic and organic, primary and secondary particles. The chemical and physical compositions of ambient particulate matter vary depending on geographical location, time of year, and weather. While the sources and formation pathways for inorganic aerosols are reasonably well known, very little is understood concerning the sources and chemistry of secondary organic aerosol (SOA). Previous investigators (e.g., Seinfeld 1992) presented several approaches to understanding the compositions and sources of particulate matter, yet it is not entirely clear what fraction of atmospheric organic aerosol is secondary, and more detailed studies are required to investigate their sources. Currently the most commonly employed methodology for estimating SOA formation is to assume that gas phase hydrocarbons react, in the gas phase, to form products that condense into the particle phase. The selectivity of gas phase reactions to SOA is assumed to be constant, and fixed fractional aerosol coefficients (FACs), which were developed based on smog chamber experiments and expert judgment, are employed (Grosjean and Seinfeld, 1989). The concept of incremental ozone 2 reactivities (described in more detail in Chapter 2), which reveal that different hydrocarbons lead to different rates of ozone production in the atmosphere, suggests that FAC will not be constant, and that FAC will be a function of the concentrations of hydrocarbons, concentrations of NOx, and other parameters. This study extends the concept of incremental ozone reactivities to PM formation. Quantitative mechanistic models of SOA formation were developed based on data from smog chamber experiments. The sensitivity of SOA formation to changes of NOx, changes in the mixture of hydrocarbon precursors, and other parameters were investigated. This work will be one of the first attempts to quantitatively model SOA formation using data from chamber experiments, and one of the first attempts to incorporate chemical mechanisms of SOA formation into a comprehensive gas phase photochemical model. 1.2 Research objective The overall objectives of this study were to develop quantitative models of SOA formation, and to use these models to characterize SOA formation in Houston. The specific objectives were to: Use data available from previous chamber experiments to develop quantitative photochemical mechanistic models for individual and lumped aromatic hydrocarbons. 3 Use a phase partitioning model from the literature to estimate the partitioning of individual compounds and lumped species between gas and aerosol phases. Incorporate chemical and physical models into a detailed gas phase photochemical mechanism developed by the Statewide Air Pollution Research Center (SAPRC). Apply the modified SAPRC mechanisms to characterize the pathways important for SOA formation in Houston, and to investigate the sensitivity of SOA formation in Houston to changes in NOx concentrations, and other parameters. 1.3 Dissertation overview This dissertation discusses the development of quantitative mechanistic models of SOA, for individual and lumped species, and application of these models to characterize SOA formation in Houston. A literature review is provided in Chapter 2. Chapter 3 discusses the methodology and procedure used to develop quantitative photochemical mechanistic models for individual aromatic hydrocarbons. This chapter also includes the procedure and methodology employed to describe partitioning between gas and particle phases. Chapter 4 discusses lumping procedures and development of chemical models and partitioning models for lumped aromatic species. Chapter 5 addresses application of models developed in Chapter 4 4 to Houston conditions, and describes the sensitivity of SOA formation to the changes of NOx emissions and other parameters. Summaries and conclusions are presented in chapter 6. 5 Chapter 2 Literature Review 2.1 Organic Particulate Matter Atmospheric aerosols have significant impacts on our health and quality of life. While many studies have focused on role of inorganic species in atmospheric aerosol, new interest has been given to the organic portion. The organic component of PM is hypothesized to cause adverse health effects (Schwartz, 1996), and organics may be particularly important because approximately 95% of organics present in urban aerosols are found in particles with diameters less than 3 microns, which can be deposited deeply into the human respiratory system (Isidorov, 1990). Because of their optical characteristics, organic aerosols also are a major cause of reductions in visibility (EPA, 1997). In addition, as will be demonstrated in this thesis, SOA formation is associated with the photochemical reactions leading to ozone production. Further, deposition of particles containing organic acid species may change the nutrient balance and acidity of land, plant, or water bodies. Finally, organic PM may change the chemical, optical, and hygroscopic behavior of inorganic aerosols (Saxena et al., 1995). Legislation has been placed in effect to reduce the amount of fine particles that are emitted directly into atmosphere, and control the emissions of gases that are precursors of condensable gases that lead to PM formation (secondary PM). National 6 Ambient Air Quality Standards (NAQQS) for particulate matter were issued by U.S. EPA in 1997. These include standards or proposed standards for both fine particulate matter smaller than 10 microns in diameter (PM10), and fine particulate matter smaller than 2.5 microns in diameter (PM2.5). These standards are: A PM10 annual average standard of 50 g/m3. A PM10 24-hour average standard of 150 g/m3. A proposed PM2.5 annual average standard of 15 g/m3. A proposed PM2.5 24-hour average standard of 65 g/m3. For the 24-hour standard, the concentrations measured during a year are rank ordered. If the concentration at the 98th percentile is greater than the 24 hour standard, a region is designated non-attainment. For an annual standard, concentrations measured throughout a year are averaged. To determine whether an area meets the annual standard, samples are collected on a regular basis (daily, every third day or 6th day) over a 24-hour period. Samples are collected using a filter-based sampler with an inlet that removes large particles. The filters are weighted and using the airflow rate through the filter, a concentration is determined. Developing effective strategies to reduce concentrations of PM to meet these standards will require a better understanding of PM sources, including the mechanisms responsible for their production. 7 2.1.1 Sources of atmospheric PM The sources of particulate matter are summarized in Figure 2.1. Figure 2.1 indicates that approximately two-thirds of the mass of atmospheric particles is inorganic. The composition and sources of these inorganic species are reasonably well understood. In contrast to the inorganics, the composition and sources of carbonaceous species, which are significant contributors to total particle mass, are poorly understood. Figure 2.1: Sources of atmospheric particulate matter (Seinfeld, et al., 1994) The distribution of ambient particulate matter in Figure 2.1 is specific to Southern California. This distribution can vary depending on geographical location, time of year, and meteorology. Figure 2.2 illustrates the distribution of ambient PM 8 that is found in several locations in Texas (Tropp et al., 1998). These distributions were identified during a Texas PM2.5 sampling and analysis study performed between 03/11/97 and 03/12/98. While the distributions shown in Figure 2.2 do not explicitly indicate the relative percent of secondary versus primary species, they do show the relative contribution of inorganic and organic compounds present in ambient aerosol. The results from the Texas PM2.5 study indicate slightly different ratios of organic to inorganic species than those noted in Southern California. In Southern California organic species contribute roughly 33 % of ambient aerosol. In Texas organic species were found to range between 25 % and 50 % in the study locations. Organics contribute approximately 25 % of airborne loadings in San Antonio and Corpus Christi. In contrast, roughly 40 % of ambient PM in Houston and Dallas are organics, and up to 50 % of PM-2.5 in El Paso is organic. 9 Figure 2.2: Spatial distribution of major chemical components during the Texas PM2.5 Sampling and Analysis Study between 03/11/97 and 03/12/98 2.1.2 Organic PM Formation Organic particulate matter is both primary and secondary in nature. Sources of primary organic aerosol include combustion processes, forest fires, and agricultural activities. Secondary organic aerosols (SOA) are produced by the gas-phase photochemical oxidation reactions of volatile organic compounds (VOCs). These reactions produce low volatility products which partition into the aerosol phase. VOCs are emitted from a variety of sources, and include alkanes, alkenes, aromatics, aldehydes, terpenes, and oxygenated species. Although photooxidation reactions to form SOA may proceed via reactions of VOCs with NO3 and ozone, OH radicals 10 initiate the majority of the reactions. It is not entirely clear what fraction of atmospheric organic aerosol is secondary; however, in most urban areas, secondary organic aerosol is believed to dominate total organic loadings. 2.1.3 Composition and Sources of Secondary Organic Aerosol Several previous investigations have attempted to determine the compositions and sources of secondary organic aerosol. Most have relied on measurements of the fraction of volatile carbon (also called organic carbon, OC) and non-volatile carbon (also called graphitic or elemental carbon, EC) in the particulate matter. In these studies, it is assumed that all EC is primary in origin and that all primary emissions will have a characteristic organic (volatile) carbon to elemental carbon ratio (OC/EC). OCsecondary = OCtotal OCprimary OCprimary = EC* (OC/EC)primary Equation 1 Equation 2 Turpin and Huntzicker (1991), among others, have examined the ratios of elemental and organic carbon in aerosol to estimate the extent of primary and secondary aerosol. They calculated secondary organic carbon concentrations in Southern California using primary OC/EC ratios between 1.4 and 2.9, and by assuming that only primary organic carbon is present in the morning. Based on these assumptions, the authors could determine what fraction of OC in the afternoon was secondary. Correlations based on Turpin and Huntzicker's estimates of SOA indicate that the concentration of secondary organic aerosol depends on the O3, NOx, and 11 organic precursor concentrations. It was estimated that approximately 30-50 % of the carbon present in aerosol during August in Los Angeles is secondary. More recent studies by Allen and co-workers (Pickle et al., 1990; Mylonas et al., 1991; Eusebi 1996) have utilized infrared spectroscopy to characterize the chemical structure and compound classes present in organic aerosols, as a function of particle size. Results from these studies show a good correlation between carbonyls and organonitrates with ozone concentrations; therefore, it can be concluded that the carbonyls and organic nitrates in atmospheric aerosol are mainly the results of atmospheric reactions. Statistical analysis indicated that approximately 95 % of aerosol carbonyl loadings in Los Angeles are secondary (Pickle et al, 1990). The limitations of these studies are that no individual organic molecules, only compound classes can be identified in ambient aerosol, and that distinguishing between primary and secondary organics based on these profiles is difficult. An alternative approach to understanding the composition and sources of SOA is to predict the extent and composition of secondary aerosol based on the estimated yields of products resulting from atmospheric reactions. Grosjean and Seinfeld (1989) and Grosjean (1992) predicted average aerosol composition by using emission inventories and their knowledge of the photochemistry of gas-phase hydrocarbon emissions. Using this approach, the formation rate for a single condensable organic product formed in the oxidation of one hydrocarbon precursor with hydroxyl radical is assumed to be given by: 12 d[SOA]/dt = ad[P]/dt = -a*d[VOC]/dt = a*kOH[OH][VOC] where: kOH [P] Equation 3 = rate constant of oxidation reaction of VOC by OH radicals = concentration of reaction product [SOA] = concentration of SOA produced by oxidation reaction [OH] = concentration of OH radicals in the reaction a = fraction of VOC reacted that produced product P = fraction of the product that is present in the condensed phase Currently the most commonly employed methodology for estimating SOA formation is to assume the selectivity to SOA (parameters "a and " in Equation 3) are constant and to employ a fractional aerosol coefficient. The fractional aerosol coefficient (FAC) is the fraction of SOA that would result from the emissions of a particular gas phase hydrocarbon (VOC). FAC Where: = [SOA] / [VOC]0 * Fraction of VOC reacted Equation 4 [SOA] = concentration of SOA ( g/m3) [VOC]0 = initial concentration of VOC ( g/m3) The estimation of FAC is based on the assumption that the oxidation reactions of each VOC lead to a fixed fraction of secondary organic aerosol product. For example, the data in Table 2.1 indicate that 2.9 % of the products formed from the reaction of 1,3,5-trimethybenzene with OH radicals will be SOA (Grosjean, 1992). 13 Table 2.1 displays the typical amount of individual organic species emitted into the atmosphere in the Los Angeles area on a kilogram per day basis. Grosjean (1992) estimated FAC of each organic species and the amount of each species that is transformed to secondary aerosol. Examination of Table 2.1 provides a rough characterization of the roles of various hydrocarbon compound classes in the formation of secondary aerosol. 14 Table 2.1: VOC Emissions Inventory and Secondary Aerosol Yields (Grosjean, 1992) Amount of VOC emitted** (kg / day) Fractional aerosol coefficient (%) Amount of Fraction of VOC reacted aerosol produced (kg / day) Hydrocarbon Precursors Olefins Alkenes: 1-Heptene C7 terminal alkenes 1-Octene C8 terminal alkenes Cis-2-octene 2,4,4-trimethyl-1-pentene Methyl heptene 1-Nonene C9 terminal alkenes Ethyl heptene 1-Decene 1-Undecene C11 terminal alkenes Trimethyldecene Cyclic Olefins: Cyclopentene 1-Methylcyclopentene 3-Methylcyclopentene Cyclohexene 1-Methylcyclohexene 3-Methylcyclohexene 4-Methylcyclohexene Aromatics Toluene o-Xylene m-Xylene p-Xylene Ethylbenzene 1,3,5-Trimethylbenzene 1,2,4-Trimethylbenzene 1,2,3-Trimethylbenzene n-Propylbenzene Isopropyl benzene o-Ethyltoluene m-Ethyltoluene p-Ethyltoluene n-Butylbenzene Sec-butylbenzene Tert-butylbenzene 1,3-Diethylbenzene 1,4-Diethylbenzene 1,2-Diethylbenzene Tetra methylbenzene Substituted benzenes: tri, C10 mono, C11 di, C11 tri, C11 di, C12 tri, C12 Miscellaneous 2,891 7,259 3,619 381 131 159 272 1,024 7,076 492 1,712 153 5,477 517 412 165 412 824 459 494 494 105,480 19,614 12,186 12,057 8,481 4,667 7,029 4,133 591 208 1,282 4,126 1,345 174 174 1,631 558 168 168 1,373 11,380 1,100 10,180 10,180 2,850 2,850 2.0 1.0 1.0 1.0 2.0 4.9 2.0 6.0 3.0 60 9.0 12.0 6.0 15.0 4.0 5.0 5.0 7.0 10.6 3.1 3.1 5.4 5.0 4.7 1.6 5.4 2.9 2.0 3.6 1.6 4.0 5.6 6.3 2.5 2.6 2.6 2.6 6.3 6.3 6.3 6.3 6.3 7.5 7.5 7.5 8.5 8.5 0.52 0.52 0.55 0.55 0.64 0.64 0.59 0.57 0.57 0.57 0.57 0.57 0.57 0.57 0.94 0.95 0.95 0.79 0.76 0.53 0.65 0.12 0.26 0.4 0.28 0.15 0.74 0.58 0.51 0.12 0.13 0.23 0.31 0.21 0.18 0.19 0.19 0.47 0.47 0.47 0.58 0.47 0.47 0.47 0.47 0.47 0.47 30.2 38.0 19.9 2.2 1.9 5.1 3.0 34.8 120.8 16.5 87.8 10.3 187.5 49.9 15.4 7.8 19.6 45.6 37.0 8.1 10.0 683.5 255.0 229.1 54.0 68.7 100.1 81.5 75.9 1.1 1.1 16.5 80.6 7.1 0.9 1.0 8.0 16.4 5.2 5.2 49.9 337.0 38.5 358.6 358.6 113.7 113.7 15 Table 2.1 (continued) Amount of VOC emitted** (kg / day) 630 487 Fractional aerosol coefficient (%) 30.0 30.0 Amount of Fraction of VOC reacted 1.0 1.0 aerosol produced (kg / day) 189.0 146.0 Hydrocarbon Precursors -Pinene -Pinene Parafins Alkanes: n-Heptane n-Octane 3-Methylheptane 2-Methylheptane 2,4,4-Trimethylpentane n-Nonane 3,5,5-Trimethylhexane 2,2,5-Trimethylhexane 2,6-Dimethylheptane n-Decane Branched C10 alkanes n-Undecane Branched C11 alkanes n-Dodecane Branched C12 alkanes n-Tridecane n-Tetradecane n-Pentadecane Cycloalkanes: Methylcyclopentane Cyclohexane Methylcyclopentane Ethylcyclopentane Trimethylcyclopentane Dimethylcyclohexane Ethylcyclohexane Trimethylcyclohexane n-Propyl cyclohexane Diethylcyclohexane n-Butylcyclohexane n-Pentylcyclohexane n-Hexylcyclohexane Oxygenated aliphatics 2-Methyl 3 hexanone C8 aldehyde Heptanone C4 subst. cyclohexanone Propyl cyclohexanone Dibutyl ether 2-Butyl tetrahydrofuran Miscellaneous compounds 2-Methylnaphthalene Naphthalene Indan Ethyl indan Phenol Decalin Ethyl decalin Carvomethol 27,233 10,284 5,791 3,540 4,719 10,975 6,927 5,928 5,626 25,840 17,500 3,016 7,700 1,932 1,950 1,101 431 0 5,024 12,370 8,695 1,558 307 2,488 2,016 1,157 815 1,734 776 769 287 3,990 2,034 881 219 156 36 23 939 667 427 1,032 417 0 362 166 0.06 0.06 0.5 0.5 0.73 1.5 0.5 0.5 0.65 0.2 2.0 2.5 2.5 3.0 3.0 3.5 4.0 5.0 0.17 0.17 2.7 2.7 2.7 2.7 2.7 3.0 3.0 4.0 4.0 5.0 6.0 0.12 0.24 0.12 0.17 0.17 0.65 0.65 5.0 4.0 2.5 2.5 5.0 3.0 4.0 30.0 0.14 0.17 0.1 0.1 0.16 0.2 0.19 0.19 0.16 0.22 0.22 0.25 0.25 0.26 0.26 0.28 0.28 0.28 0.10 0.14 0.20 0.12 0.12 0.10 0.12 0.12 0.12 0.12 0.12 0.12 0.12 0.26 0.47 0.26 0.19 0.19 0.26 0.19 0.68 0.32 0.19 0.19 0.46 0.35 0.38 0.35 2.3 1.0 2.9 1.8 5.4 33.0 6.6 5.7 5.8 113.8 77.0 18.7 48.0 15.1 15.1 10.6 4.8 0.0 0.8 2.9 46.9 5.0 1.0 6.7 6.5 4.2 2.9 8.3 3.7 4.6 2.0 1.3 2.3 0.3 0.1 0.1 0.1 0.1 32.0 8.3 2.1 4.9 9.7 0.0 5.3 17.5 16 Table 2.1 (continued) Amount of VOC emitted** (kg / day) 166 30 299 970 450827 Fractional aerosol coefficient (%) 30.0 10.0 30.0 30.0 522.76 Amount of Fraction of VOC reacted 0.35 1.0 0.35 1.0 37.75 aerosol produced (kg / day) 17.5 3.0 24.1 291.0 5030.1 Hydrocarbon Precursors Carvone Indene Isopulegone + terpinene Total Using these fraction aerosol coefficients, studies by previous investigators (Pandis et al., 1992; Grosjean and Seinfeld, 1989) have concluded that certain hydrocarbon classes contribute much more substantially to secondary organic aerosol formation than other compound classes. Pandis et al., (1992) found aromatic precursors contribute 65 % of the SOA in Los Angeles, while biogenics, alkanes, and alkenes contribute 16 %, 15 %, and 4 % respectively. Grosjean and Seinfeld (1989) suggested that individual hydrocarbons within these compound classes can dominate overall secondary organic aerosol formation as displayed in Table 2.1. For example, the photooxidation reaction of toluene leads to over 600 kg/day of SOA, accounting for 13.6% of the total SOA yield Currently, the best available estimates of fractional aerosol coefficients are the fixed values, listed in Table 2.1. The problem associated with this approach is that FACs are not constant, and can vary based on the conditions of the reacting mixture. Therefore, the estimates of SOA formation assuming FACs are constant can result in inaccuracies in estimating SOA. To understand the variability in SOA as a function of hydrocarbon mixture effects and other variables, it is useful to examine a related concept: incremental ozone reactivity. 17 2.1.4 Incremental Ozone Reactivity Carter (1994) has used a detailed photochemical model to examine how changes in hydrocarbon emissions influence ozone formation. He defined the concept of incremental ozone reactivity as IRi =[O3}/[VOCi] Equation 5 Where [O3] are grams of additional ozone formed per gram of hydrocarbon added to a reacting mixture ([VOC]). Carter (1993,1994) developed 18 separate reactivities scales using a detailed gas phase mechanism in a single-cell trajectory model. Three of these 18 scales are commonly used in quantifying the reactivities of a variety of VOCs. The first scale is the maximum incremental reactivity (MIR). MIR is the additional grams of ozone formed per gram of hydrocarbon added to a mixture with an initial ratio of VOC to NOx that generates the largest average incremental reactivity. For this scenario, ozone formation is inclined to effect greatly by emissions of VOC. As a result, this generally occurs at a low VOC to NOx ratio. MIR is mathematically expressed as [O3 ] MIRi = MAX [VOCi ] Where MIRi is maximum incremental reactivity of species i. Carter estimated the MIRi by fixing the VOC concentrations and adjusting the NOx until the reactivity for specific model run was maximized. 18 The second scale is the maximum ozone reactivity (MOIR). MOIR are grams of ozone formed per gram of hydrocarbon added to a mixture with an initial ratio of VOC to NOx that generates the maximum amount of ozone. This commonly represents conditions in which the VOC to NOx ratio is moderate (transition region between VOC limitation and NOx limitation). MOIR for compound i is given by [O ] 3 MOIRi = [VOCi ] for [ O3 ] max imized MOIR was evaluated by setting the NOx concentration in the trajectory to give the maximum ozone formation, and then assessing the sensitivity of the ozone for individual VOCs. The third scale is the equal benefit incremental reactivity (EBIR), which are grams ozone per gram of hydrocarbon added to a mixture in which both NOx and VOC levels limit ozone formation. EBIR scale is determined for conditions that are midway between VOC limitation and NOx limitation. The National Research Council (1999) reviewed the major characteristics of these three scales as detailed in Table 2.2. 19 Table 2.2: Summary of Major Characteristics of the Primary Carter Reactivity Scale Scale Type of Scenarios Used Low VOC to NOx ratio conditions in which ozone is most sensitive to VOC changes Moderate VOC to NOx ratio condition in which highest ozone yield are formed Higher VOC to NOx ratio conditions in which VOC an NOx control are equally effective in reducing ozone Derivation of Scale from Individual Scenario Reactivity Average of incremental reactivities in the MIR scenarios Average of incremental reactivities in the MOR scenarios Average of incremental reactivities in the EBIR scenarios Ozone Quantification Maximum ozone Reflects Effect of VOC on Ozone formation rates Maximum incremental reactivity (MIR) Maximum ozone reactivity (MOR) Maximum ozone Ultimate ozone yields Equal benefit incremental reactivity (EBIR) Maximum ozone Ultimate ozone yields Three incremental reactivity scales for selected hydrocarbons evaluated by Carter (1994) are listed in Table 2.3. Table 2.3: Incremental ozone reactivities for selected hydrocarbons, evaluated using three different initial mixtures of air pollutants (Carter, 1994) Hydrocarbon Octane Octene Toluene o-Xylene Benzadehyde Isoprene -Pinene Maximum Incremental Reactivity 0.19 0.85 0.88 2.11 -0.18 2.9 1.04 Incremental reactivity at maximum ozone levels 0.34 0.90 0.53 1.60 -1.08 2.9 1.08 Incremental reactivity for nitrogen oxide limited Chemistry 0.33 0.89 -0.023 1.26 -2.7 3.3 1.23 20 It is evident from Table 2.3 (more incremental reactivities are available from Carter, 1994) that incremental ozone reactivities are different between compounds and can vary for any given compound, or even change in sign based on the initial conditions of the reacting mixture. This principle should also extend to incremental particle formation since SOA formation, like ozone formation, is controlled by OH radical reactions. The goal of this thesis was to develop quantitative mechanistic models and fractional aerosol yield models for a variety of hydrocarbon precursors, accounting for this phenomenon. 2.2 PM modeling 2.2.1 Background For decades, researchers have been developing modeling tools that are used to simulate particulate matter concentrations in the atmosphere. Modeling systems required for simulating particulate matter formation includes an air quality model, a PM module incorporated in the air quality model, and a meteorological model. The major inputs for an air quality PM model are represented in Figure 2.3. 21 Figure 2.3: Schematic Representation of a PM Source Model (Seigneur, 1997) As shown in Figure 2.3, there are three major components required to simulate atmospheric PM concentrations. The first component identifies initial and boundary conditions such as concentrations of gases and PM. The second component is meteorological data, e.g., wind field and temperature. Emissions from point and area sources are also required. In addition to these input components are understanding of gas phase chemistry and PM chemistry and physics, i.e., chemical mechanisms and deposition rates are required. This thesis focuses on the integration of gas phase chemistry with PM chemistry and physics. 22 2.2.2 PM Models PM models can be categorized into two groups: Lagrangian models, and Eulerian three-dimensional gridded models. Lagrangian models simulate the formation of PM in a parcel of air that is convected along a mean wind trajectory. The several types of Langrangian models can further be separated into plume models and trajectory models. Although these two model categories are similar in their formulation, which is based on the Langragian framework, plume models (e.g., the Industrial source complex model, ISC (EPA, 1992)) focus on the simulation of PM formation from a single source plume emission. On the other hand, trajectory models focus on the input of all emission sources along the air parcel trajectory. This thesis will utilize primarily trajectory models. Eulerian models simulate PM formation in a three-dimensional gridded domain. The simulation of PM formation for Eulerian models is based on the solution of the mass continuity equation. Major components required to run Eulerian simulations include: initial and boundary conditions of chemical species in model, emission rates of these chemical species, meteorological data, physiographic information, and time and date of the scenario being simulated. Results from simulations include the concentrations of specific chemical species as a function of time and location. Eulerian models are able to simulate the evolution of PM in a threedimensional domain. Eulerian models can further be grouped into two major 23 categories: episodic, and long-term models. Episodic models include a detailed mechanism of atmospheric chemistry. These models are generally limited in their application to a short period of simulation time (e.g., a few days) because of their computational cost. More long-term models can simulate longer time periods, but required simplified mechanism for atmospheric chemistry. Seigneur (1997) identified eight Eulerian episodic PM air quality models. The CIT model, developed at the California Institute of Technology. The Denver Air Quality Model (DAQM), developed by the State University of New York at Albany and Science & Policy Associates for the Denver Air Quality Council. The Gas, Aerosol, Transport, and Radiation model (GATOR), developed at the University of California at Los Angeles and at Stanford University. The Regional Particulate Model (RPM), developed at the U.S. EPA office of Research and Development (ORD) as a postprocessor to the Regional Acid Deposition Model (RADM) framework. The SAPMAP Air Quality Model with aerosols (SAQM-AERO) developed at the State University of New York at Albany and modified for PM by the California Institute of Technology, Sonoma Technology, Inc. and others for the California Air Resources Board. 24 The Urban Airshed Model version IV with aerosols (UAM-AERO), using UAM-IV as the host air quality model, developed by the California Institute of Technology, Sonoma Technology, Inc. and others for the California South Coast Air Quality Management District and the California Air Resources Board. The Urban Airshed Model version IV with an aerosol module based on the Aerosol Inorganic Model (UAM-AIM), using UAM-IV as the host air quality model, developed at the University of Delaware. The Comprehensive Air Quality Model with Extensions (CAMx) version III, developed by Environ International Cooperation. Table 2.4 presents a list of PM episodic models along with their host air quality model, the meteorological models used, and the geographical areas where these models have been applied. Table 2.4: Host Air Quality Model, Meteorological Model and Applications to Date of Episodic PM Models (Seigneur, 1997) PM Model CIT DAQM GATOR PRM SAQM-AERO UAM-AERO Air Quality Model CIT RADM GATOR RADM SAQM UAM-IV Meteorological Model Diagnostic model MM4* MMTD MM4* MM5 Various models Applications to Date Los Angeles Basin, CA Denver, CO Los Angeles Basin, CA Eastern North America San Joaquin Valley, CA Los Angeles Basin and San Joaquin Valley, CA 25 PM Model Air Quality Model Meteorological Model Applications to Date UAM-AIM CAMx UAM-IV CAMx Various models MM5, SAIMM, RAMS Los Angeles Basin Eastern Texas * Further applications are planned with MM5 From Table 2.4, two PM models use UAM-IV as host air quality model, the other two models use RADM, and the rest use the CIT, GATOR, CAMx, and SAQM, respectively. There are several meteorological models employed for these PM models, the most commonly used model is MM4. This model, however, will be replaced by MM5. 2.2.3 Treatment of SOA in PM models Since this research focuses on developing a mechanistic model of SOA formation, this section discusses the current treatment of SOA in PM models. The current PM models use various methods to predict and treat SOA formation. Four gas phase mechanisms are used among the seven episodic PM models. These mechanisms are the Statewide Air Pollution Research Center (SAPRC) gas phase chemistry model developed by Carter (1990, 1995), the Carbon-Bond Mechanism version IV (CBM-IV) of Gery et al., (1989), the RADM mechanism version 2 (RADM2) developed by Stockwell et al., (1990), and a mechanism developed 26 explicitly for GATOR (Jacobson, 1994). Table 2.5 details current treatment of SOA formation in PM models. Table 2.5: Treatment of SOA formation in PM models (Seigneur, 1997) PM Model CIT DAQM GATOR RPM SAQM-AERO UAM-AERO UAM-AIM VISHWA SOA Treatment Gas Phase Chemical SOA formation Mechanism Extended LCC Absorption mechanism with mass transfer RADM2 Lumped aerosol yields of Pandis et al., 1992 GATOR Solubility RADM2 Lumped aerosol yields of Pandis et al., 1992 CBM-IV and SAPRC Lumped aerosol yields of Pandis et al., 1992 CBM-IV and SAPRC Lumped aerosol yields of Pandis et al., 1992 CBM-IV and SAPRC Lumped aerosol yields of Pandis et al., 1992 Condensed chemistry 2-component approach (7 reactions and 7 with aromatic and species) terpene lumped aerosol yields of Pandis et al., 1992 adjusted downward to match ambient observations CBM-IV and SAPRC Simplified parameterized aerosol thermodynamics package Reference Odum et al., 1996; Meng et al., 1997 Moucheron and Milford, 1996 Jacobson, 1997 Binkowski and Shankar, 1995 NA Kumar et al., 1996 Sun and Wexler, 1997 Venkatram et al., 1997 CAMx Environ, user's guide, 2000 As noted in Table 2.5, DAQM, RPM, UAM-AERO, UAM-AIM, and SAQM use the lumped SOA yields approach of Pandis et al., (1992). This method makes use of volatile organic compounds (VOCs) specific fractional aerosol coefficients (FACs) of Grosjean and Seinfeld (1989). For CIT, SOA is modeled by the mass transport with absorption mechanism between gas and particulate phases. In the VISHWA model, the lumped SOA yields of Pandis et al., for lumped aromatics and terpenes 27 have been adjusted downward to match experimental observations. GATOR describes gas-particle partitioning of condensable and soluble organic species. There are a number of limitations in the atmospheric gas phase chemical mechanisms currently employed in urban and regional scale models to predict SOA formation. These mechanisms do not account for the chemistry of organic precursors that lead to the formation of condensable products in sufficient detail to estimate the SOA generation. In addition, many of the higher molecular weight organics that are not important in ozone production, but are sources of SOA formation, are neglected in most current mechanisms. Finally, gas phase mechanisms such as CBM-IV in CAMx, cannot estimate a production of semi-volatile products from each parent compound, because species are lumped into carbon bond groups based on their chemical structures. For example, decene will be represented by eight paraffinic carbons and two carbons associated with the double bond. Heptene would be represented by five paraffin carbons and two carbons associated with the double bond. So, in a mixture of heptene and decene, the CBM follows only total concentrations of paraffinic and olefinic carbons, not molecular concentrations. Yet heptene and decene have very different SOA yields because of the differences in molecular weights. So, if CBM is used as the chemical mechanism, the SOA formation calculated from simulations will not be specifically from the amount reacted of decene or heptene, but will be from the emissions of decene or heptene. Since there is a large uncertainty for FACs and extent of conversion of hydrocarbons in ambient air, the estimated SOA formation may be inaccurate. 28 Unlike CBM-IV mechanism, SAPRC mechanism retains the information about molecular weight of hydrocarbons in the reacting mixture, because different parts of molecule are not broken down and treated differently. Compounds in the reacting mixture are lumped as a whole molecule into different groups based on their chemical structures and reactivities. For example, benzene will be lumped in a group of slower aromatic compounds and represented by model species ARO1. Parameters, product yields, and molecular weight of ARO1 are weighted averages from every compound in the group (more details about lumping are described in Chapter 4). Further, the SAPRC model allows the user to modify or update a detailed mechanism including SOA formation for both individual and lumped hydrocarbon precursors. Because of these advantages, SAPRC99 was used in this thesis to quantify the SOA formation for individual and lumped compounds at a variety of initial conditions. As described earlier, several PM models employ a fixed FAC in evaluating SOA formation for any given compounds. However, by examining the concept of incremental ozone reactivity (Carter, 1994), the hypothesis that FACs are not constant, but can vary depending on the VOC to NOx ratio and the composition of VOCs emerges. This hypothesis was examined in this thesis. 29 2.3 Summary This chapter has summarized the current literature on organic particulate matter and PM modeling. The sources and compositions of inorganic aerosol are reasonably well understood, on the other hand, better understanding of sources and compositions of organic particulate matter are needed. Previous investigators have estimated SOA formation using several approaches. The most common approach, which is employed in a variety of PM models, is to use a fixed fractional aerosol coefficient (FAC) for the yields of SOA form any given hydrocarbon (Pandis et al., 1992). The concept of incremental ozone reactivity, however, raises the argument that FAC is not constant, and can vary depending on the initial conditions and other parameters. This research builds on this concept to develop a suitable approach for estimating SOA formation. 30 Chapter 3 Quantitative models of SOA formation for individual compounds development 3.1 Introduction Two major topics are addressed in this chapter. The first is the development of quantitative mechanistic models and physical phase partitioning models of secondary organic aerosol (SOA) formation for individual compounds, using data available from smog chamber experiments. The second topic is the use of these models to examine the SOA formation potential for various aromatic hydrocarbons. Since the chemical mechanisms, kinetic models and partitioning models will be incorporated into an existing model of gas phase chemistry (the SAPRC model), the chapter will begin with a discussion of the baseline, gas-phase model. Then the approaches used for modeling the chemistry and phase partitioning of SOA will be described. 3.2 SAPRC model Aerosol formation is the result of a set of series/parallel reactions of hydrocarbon precursors, and aerosol precursors with OH radicals. The concentration of OH radical is critical in determining the aerosol yields. In atmospheric oxidations, OH radical concentration is highly dependent on concentrations of nitrogen oxides and the concentrations of all hydrocarbons in the reacting mixture, and other parameters. To accurately predict aerosol yields, the mechanism for SOA formation 31 for a compound, (e.g. trimethylbenzene) must be integrated into a comprehensive gas phase model. The comprehensive model of gas phase chemistry used in this work has been developed by the Statewide Air Pollution Research Center (SAPRC) of the University of California. SAPRC has developed software, SAPRC 99, to model the gas phase chemistry of mixtures of more than 350 common air pollutants. These mechanisms have been evaluated extensively against environmental chamber data for gas phase species (Carter and Lurmann, 1991). The SAPRC model contains a gas phase reaction mechanism for the atmospheric photoreaction of over 350 hydrocarbon precursors, including alkenes, alkanes, aromatics, alcohols, ethers and other compounds representative of the range of reactive organics emitted into atmosphere. Table 3.1 lists all species that are modeled in SAPRC. Table 3.1: Detailed species that are modeled in SAPRC 99 (Carter, 2000) Compounds Carbon Monoxide Methane Alkanes Ethane Propane n-Butane n-Pentane n-Hexane n-Heptane n-Octane n-Nonane n-Decane n-Undecane n-Dodecane n-Tridecane n-Tetradecane n-Pentadecane n-C16 Isobutane Iso-Pentane Neopentane 2,2-Dimethyl Butane Model Name CO METHANE ETHANE PROPANE N-C4 N-C5 N-C6 N-C7 N-C8 N-C9 N-C10 N-C11 N-C12 N-C13 N-C14 N-C15 N-C16 2-ME-C3 2-ME-C4 22-DM-C3 22-DM-C4 Cs 1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 4 5 5 6 32 MWt 28 16 30 44 58 72 86 100 114 128 142 156 170 184 198 212 226 58 72 72 86 kOH (300) (cm3/molec-sec) 2.1e-13 6.6e-15 2.6e-13 1.1e-12 2.5e-12 4.0e-12 5.5e-12 7.0e-12 8.8e-12 1.0e-11 1.1e-11 1.3e-11 1.4e-11 1.6e-11 1.8e-11 2.1e-11 2.3e-11 2.2e-12 3.7e-12 8.6e-13 2.4e-12 Compounds 2,3-Dimethyl Butane 2-Methyl Pentane 3-Methyl Pentane 2,2,3-Trimethyl Butane 2,2-Dimethyl Pentane 2,3-Dimethyl Pentane 2,4-Dimethyl Pentane 2-Methyl Hexane 3,3-Dimethyl Pentane 3-Methyl Hexane 2,2,3,3-Tetramethyl Butane 2,2,4-Trimethyl Pentane 2,2-Dimethyl Hexane 2,3,4-Trimethyl Pentane 2,3-Dimethyl Hexane 2,4-Dimethyl Hexane 2,5-Dimethyl Hexane 2-Methyl Heptane 3-Methyl Heptane 4-Methyl Heptane 2,2,5-Trimethyl Hexane 2,3,5-Trimethyl Hexane 2,4-Dimethyl Heptane 2-Methyl Octane 3,3-Dimethyl Pentane 3,5-Dimethyl Heptane 4-Ethyl Heptane 4-Methyl Octane 2,4-Dimethyl Octane 2,6-Dimethyl Octane 2-Methyl Nonane 3,4-Diethyl Hexane 3-Methyl Nonane 4-Methyl Nonane 4-Propyl Heptane 2,6-Dimethyl Nonane 3,5-Diethyl Heptane 3-Methyl Decane 4-Methyl Decane 2,6-Diethyl Octane 3,6-Dimethyl Decane 3-Methyl Undecane 5-Methyl Undecane 3,6-Dimethyl Undecane 3,7-Dietyl Nonane 3-Methyl Dodecane 5-Methyl Dodecane 3,7-Dimethyl Dodecane 3,8-Diethyl Decane 3-Methyl Tridecane Model Name 23-DM-C4 2-ME-C5 3-ME-C5 223TM-C4 22-DM-C5 23-DM-C5 24-DM-C5 2-ME-C6 33-DM-C5 3-ME-C6 2233M-C4 224TM-C5 22-DM-C6 234TM-C5 23-DM-C6 24-DM-C6 25-DM-C6 2-ME-C7 3-ME-C7 4-ME-C7 225TM-C6 235TM-C6 24-DM-C7 2-ME-C8 33-DE-C5 35-DM-C7 4-ET-C7 4-ME-C8 24-DM-C8 26DM-C8 2-ME-C9 34-DE-C6 3-ME-C9 4-ME-C9 4-PR-C7 26DM-C9 35-DE-C7 3-ME-C10 4-ME-C10 36-DE-C8 36DM-C10 3-ME-C11 5-ME-C11 36DM-C11 37-DE-C9 D-ME-C12 5-ME-C12 37DM-C12 38DE-C10 3-ME-C13 Cs 6 6 6 7 7 7 7 7 7 7 8 8 8 8 8 8 8 8 8 8 9 9 9 9 9 9 9 9 10 10 10 10 10 10 10 11 11 11 11 12 12 12 12 13 13 13 13 14 14 14 33 MWt 86 86 86 100 100 100 100 100 100 100 114 114 114 114 114 114 114 114 114 114 128 128 128 128 128 128 128 128 142 142 142 142 142 142 142 156 156 156 156 170 170 170 170 184 184 184 184 198 198 198 kOH (300) (cm3/molec-sec) 5.8e-12 5.3e-12 5.4e-12 4.3e-12 3.4e-12 1.4e-11 5.0e-12 1.4e-11 6.0e-12 1.4e-11 1.1e-12 3.6e-12 4.8e-12 7.1e-12 1.7e-11 1.7e-11 1.7e-11 1.7e-11 1.7e-11 1.7e-11 1.2e-11 7.9e-12 2.0e-11 1.0e-11 4.9e-12 2.1e-11 2.1e-11 9.7e-12 2.3e-11 1.3e-11 1.3w-11 7.4e-12 2.3e-11 2.3e-11 2.4e-11 2.6e-11 2.8e-11 2.6e-11 2.6e-11 3.1e-11 2.9e-11 2.9e-11 2.9e-11 3.2e-11 3.4e-11 3.1e-11 3.1e-11 3.5e-11 3.6e-11 3.4e-11 Compounds 6-Methyl Tridecane 3,7-Dimethyl Tridecane 3,9-Dimethyl Undecane 3-Methyl Tetradecane 6-Methyl Tetradecane 3-Methyl Pentadecane 4,8-Dimethyl Tetradecane 7-Methyl Pentadecane Cyclopropane Cyclobutane Cyclopentane Cyclohexane Isoporpyl Cyclopropane Methylcyclopentane 1,3-Dimethyl Cyclopentane Cycloheptane Ethyl Cyclopentane Methylcyclohexane 1,3-Dimethyl Cyclohexane Cyclooctane Ethylcyclohexane Propyl Cyclopentane 1,1,3-Trimethyl Cyclohexane 1-Eth.-4-Meth. Cyclohexane Propyl Cyclohexane 1,3-Diethyl Cyclohexane 1,4-Diethyl Cyclohexane 1-Methyl-3-Isopr, Cyclohex. Butyl Cyclohexane 1,3-Diethyl 5-ME. Cyclohex. 1-Ethyl-2-Propyl Cyclohex. Pentyl Cyclohexane 1,3,5-Triethyl Cyclohex. 1-Meth.-4-Pentyl Cyclohex. Hexyl Cyclohexane 1,3-Dieth-5-Pent Cyclohex. 1-Meth.2-Hexyl-Cyclohex. Heptyl Cyclohexane 13-Diprop-5-Eth Cyclohex. 1-Meth.-4-Heptyl Cyclohex. Octyl Cyclohexane 1,3,5-Tripropyl Cyclohex. 1-Methyl-2-Octyl Cyclohex. Nonyl Cyclohexane 1,3-Prop,-5-Butyl Cyclohex. 1-Methyl-4-Nonyl Cyclohex. Decyl Cyclohexane Alkenes Ethene Model Name 6-ME-C13 37DM-C13 39DM-C13 3-ME-C14 6-ME-C14 3-ME-C15 48DM-C14 7-ME-C15 CYCC3 CYCC4 CYCC5 CYCC6 IRP-CC3 ME-CYCC5 13DMCYCC5 CYCC7 ET-CYCC5 ME-CYCC6 13DMCYC6 CYCC8 ET-CYCC6 PR-CYCC5 113MCYC6 1E4MCYC6 C3-CYCC6 13DECYC6 14DECYC6 1M3IPCY6 C4-CYCC6 13E5MCC6 1E2PCYC6 C5-CYCC6 135ECYC6 1M4C5CYC6 C6-CYCC6 13E5PCC6 1M2C6CC6 C7-CYCC6 13P5ECC6 1M4C7CC6 C8-CYCC6 135PCYC6 1M2C8CC6 C9-CYCC6 13P5BCC6 1M4C9CY6 C10CYCC6 ETHENE Cs 14 15 15 15 15 16 16 16 3 5 6 6 6 7 7 7 7 7 8 8 8 8 9 9 9 10 10 10 10 11 11 11 12 12 12 13 13 13 14 14 14 15 15 15 16 16 16 2 34 MWt 198 212 212 212 212 226 226 226 42 56 70 84 84 84 98 98 98 98 112 112 112 112 126 126 126 140 140 140 140 154 154 154 168 168 168 182 82 182 196 196 196 210 210 210 224 224 224 28 kOH (300) (cm3/molec-sec) 3.4e-11 3.8e-11 3.9e-11 3.7e-11 3.7e-11 4.0e-11 4.0e-11 4.0e-11 8.4e-14 1.5e-12 5.1e-12 7.3e-12 2.7e-12 1.1e-11 1.4e-11 1.3e-11 1.5e-11 1.0e-11 2.4e-11 1.4e-11 2.4e-11 1.7e-11 8.7e-12 2.7e-11 2.7e-11 3.1e-11 3.1e-11 3.0e-11 3.0e-11 3.4e-11 3.4e-11 3.3e-11 3.8e-11 3.6e-11 1.8e-11 4.1e-11 3.9e-11 3.8e-11 4.4e-11 4.2e-11 4.1e-11 4.7e-11 4.4e-1 4.4e-11 4.9e-11 4.7e-11 4.7e-11 9.2e-12 Compounds Propene 1-Butene 1-Pentene 3-Methyl 1-Butene 1-Hexene 3,3-Dimethyl-1Butene 3-Methyl-1-Pentene 4-Methyl-1-Pentene 1-Heptene 1-Octene 1-Nonene 1-Decene 1-Undecene 1-Dodecene 1-TRIdecene 1-Tetradecene 1-Pentadecene Isobutene 2-Methyl-1-Butene 23-Dimethyl-1-Butene 2-Ethyl-1-Butene 2-Methyl-1-Pentene 2,3,3-trimethyl-1-Butene 3-Methyl-2-Isopropyl-1Butene Cis-2-Butene Trans-2-Butene 2-Methyl-2Butene Cis-2-Pentene Trans-2-Pentene 2,3-Dimethyl-2-Butene 2-Methyl-2 Pentene Cis-2-Hexene Cis-3-Hexene Cis-3-Methyl-2-Hexene Trans-3-Methyl2-Hexene Trans-4-Methyl-Hexene Trans-2-Hexene Trans-3-Hexene 2,3-Dimethyl-2Hexene Cis-3-Heptene Trans 4,4-dimethyl-2-Pentene Trans-2-Heptene Trans-3-Heptene Cis-4-Octene Trans 2,2-Dimetyl 3-Hexene Trans 2,5-Dimethyl 3-Hexene Trans-3-Octene Trans-4-Octene 2,4,4-trimethyl-2-Pentene Model Name PROPENE 1-BUTENE 1-PENTEN 3M-1-BUT 1-HEXENE 33M1-BUT 3M1-C5E 4M1-C5E 1-HEPTEN 1-PCTENE 1-C9E 1-C10E 1-C11E 1-C12E 1-C13E 1-C14E 1-C15E ISOBUTEN 2M-1-BUT 23M1-BUT 2E1-BUT 2M1-C5E 23M1BUT 3M2I1C4E C-2-BUTE T-2-BUTE 2M-2-BUT C-2-PENT T-2-PENT 23M2-BUT 2M-2-C5E C-2-C6E C-3-C6E C3M2-C5E T3M2-C5E T4M2-C5E T-2-C6E T-3-C6E 23M2-C5E C-3-C7E T44M2C5E T-2-C7E T-3-C7E C-4-C8E T22M3C6E T25M3C6E T-3-C8E T-4-C8E 244M2C5E Cs 3 4 5 5 6 6 6 6 7 8 9 10 11 12 13 14 15 4 5 6 6 6 7 8 4 4 5 5 5 6 6 6 6 6 6 6 6 6 7 7 7 7 7 8 8 8 8 8 8 35 MWt 42 56 70 70 84 84 84 84 98 112 126 140 154 168 182 196 210 6 70 84 84 84 98 112 56 56 70 70 70 84 84 84 84 84 84 84 84 84 98 98 98 98 98 112 112 112 112 112 126 kOH (300) (cm3/molec-sec) 3.1e-11 3.6e-11 3.6e-11 3.7e-11 4.2e-11 3.0e-11 6.6e-11 6.8e-11 4.5e-11 7.0e-11 7.2e-11 6.8e-11 7.2e-11 7.2e-11 7.2e-11 7.2e-11 7.2e-11 5.8e-11 7.1e-11 1.2e-10 1.2e-10 7.2e-11 1.2e-10 1.2e-10 1.1e-10 1.5e-10 3.1e-10 1.7e-10 1.7e-10 8.7e-10 4.8e-10 2.3e-10 2.0e-10 4.6e-10 5.1e-10 1.6e-10 2.3e-10 2.1e-10 1.2e-9 2.3e-10 1.6e-10 1.7e-10 2.3e-10 1.7e-10 1.5e-10 2.3e-10 1.3e-10 3.2e-10 2.3e-10 Compounds Trans-4-Nonene 3,4-Diethyl-2-Hexene Cis-5-Decene Trans-4-Decene Trans-5-Undecene Trans-5-Dodecene Trans-5-Tridecene Trans-5-Tetradecene Trans-5-Pentadecene Cyclopentene 1-Methyl Cyclopentene Cyclohexene 1-Methyl Cyclohexene 4-Methyl Cyclohexene 1,2-Dimethyl Cyclohexene 1,3-Butadiene Isoprene 3-Carene -Pinene -Pinene d-Limonene Sabeiene Styrene Aromatics Benzene Toluene Ethyl Benzene Isopropyl Benzene n-Propyl Benzene s-Butyl Benzene m-Xylene o-Xylene p-Xylene 1,2,3-Trimethyl Benzene 1,2,4-Trimethyl Benzene 1,3,5-Trimethyl Benzene Naphthalene Tetralin Methyl Naphthalene 2,3-Dimethyl Naphth. Alkynes Acetylene Methyl Acetylene 2-Butyne Ethyl Acetylene Alcohols, Glycols and Eters Methanol Ethanol Isopropyl Alcohol n-Propyl Alcohol Model Name T-4-C9E 34E2-C6E C-5-C10E T-4-C10E T-5-C11E T-5-C12E T-5-C13E T-5-C14E T-5-C15E CYC-PNTE 1M-CC5E CYC-HEXE 1M-CC6E 4M-CC6E 12M-CC6E 13-BUTDE ISOPRENE 3-CARENE A-PINENE B-PINENE D-LIMONE SABINENE STYRENE BENZENE TOLUENE C2-BENZ I-C3-BEN N-C3-BEN S-C4-BEN M-XYLENE O-XYLENE P-XYLENE 123-TMB 124-TMB 135-TMB NAPHTHAL TETRALIN ME-NAPH 23-DMN ACETYLEN ME-ACTYL 2-BUTYNE ET-ACTYL MEOH ETOH I-C3-OH N-C3-OH Cs 9 10 10 10 11 12 13 14 15 5 5 6 7 7 8 4 5 10 10 10 10 10 8 6 7 8 9 9 10 8 8 8 9 9 9 10 10 11 12 2 3 4 4 1 2 3 3 36 MWt 128 140 140 140 154 168 182 196 210 68 82 82 96 96 110 54 68 136 136 136 136 136 104 78 92 106 120 120 134 106 106 106 120 120 120 128 132 142 156 26 40 54 54 32 46 60 60 kOH (300) (cm3/molec-sec) 2.6e-10 2.6e-10 1.8e-10 2.3e-10 2.3e-10 2.3e-10 2.3e-10 2.3e-10 2.3e-10 3.1e-10 5.6e-10 1.1e-10 3.3e-10 1.7e-10 8.3e-10 6.9e-10 1.1e-10 1.5e-10 1.2e-10 9.7e-11 3.1e-10 2.0e-10 6.6e-11 1.2e-12 5.9e-12 7.1e-12 6.5e-12 6.0e-12 6.0e-12 2.4e-11 1.4e-11 1.4e-11 3.3e-11 3.3e-11 5.8e-11 2.1e-11 3.4e-11 5.2e-11 7.7e-11 9.2e-13 5.9e-12 2.7e-11 8.0e-12 9.3e-13 3.3e-12 5.3e-12 5.5e-12 Compounds Isobutyl Alcohol n-Butyl Alcohol s-Butyl Alcohol t-Butyl Alcohol Cyclopentanol 2-Pentanol 3-Pentanol Pentyl Alcohol Cyclohexanol 1-Hexanol 2-Hexanol 1-Heptanl 1-Octanol 2-Ethyl-1-Hexanol 2-Octanol 3-Octanol 4-Octanol 8-Methyl-1-Nonanol Ethylene Glycol Propylene Glycol 1,2-Butandiol Glycetol 1,2-Dihydroxy Hexane 2-Methyl-2,4-Pentanediol Dimethy Ether Trimethylene Oxide Tetrahydrofuran Diethyl Ether Dimethoxy methane Alpha-Methytetrahydrofuran Tetrahydropyran Ethyl Isopropyl Ether Methyl n-Butyl Ether Methyl t-Butyl Ether Di n-Propyl Ether Ethyl n-Butyl Ether Ethyl t-Butyl Ether Methyl t-Amyl Ether 2-Butyl Tetrahydrofuran Di-Isobutyl Ether Di n-Butyl Ether Di n-Pentyl Ether 2-Methoxyethanol 1-Methoxy-2-Propanol 2-Ethoxyethanol 2-Methoxy-1-Propanol 1-Ethoxy-2-Propanol 2-Propoxyethanol 3-Ethoxy-1-Propanol 3-Methoxy-1-Butanol Model Name I-C4-OH N-C4-Oh S-C4-OH T-C4-OH CC5-OH 2-C5OH 3-C5OH C5OH CC6-OH 1-C6OH 2-C6OH 1-C7OH 1-C8-OH 2-ETC6OH 2-C8-OH 3-C8-OH 4-C8-OH 1-C10-OH ET-GLYCL PR-GLYCL 12-C4OH2 GLYCERL C6-GLYCL 2M24C5OH ME-O-ME TME-OX THF ET-O-ET METHYLAL AM-THF THP ET-O-IPR MNBE METBE PR-O-PR ENBE ETBE MTAE 2BU-THF IBU2-O BU-O-BU C5-O-C5 MEO-ETOH MEOC30H ETO-ETOH 2MEOC3OH ETOC3OH 2PROETOH 3ETOC3OH 3MEOC4OH Cs 4 4 4 4 5 5 5 5 6 6 6 7 8 8 8 8 8 10 2 3 4 3 6 6 2 3 4 4 3 5 5 5 5 5 6 6 6 6 8 8 8 10 3 4 4 4 5 5 5 5 37 MWt 74 74 74 74 86 88 88 88 100 102 102 11 130 130 130 130 130 158 62 76 90 92 118 118 46 58 72 74 76 86 86 88 88 88 102 102 102 102 128 130 130 158 76 90 90 90 104 104 104 104 kOH (300) (cm3/molec-sec) 1.4e-11 8.6e-12 2.0e-11 1.1e-12 1.1e-11 1.2e-11 1.2e-11 1.1e-11 3.5e-11 1.3e-11 1.2e-11 1.4e-11 2.0e-11 2.7e-11 2.5e-11 3.1e-11 2.9e-11 3.1e-11 1.5e-11 2.2e-11 3.2e-11 3.7e-11 3.7e-11 1.1e-11 3.0e-12 1.0e-11 1.6e-11 1.3e-11 4.9e-12 2.2e-11 1.4e-11 4.9e-11 1.5e-11 2.9e-12 1.8e-11 2.1e-11 8.8e-12 7.9e-12 5.5e-11 2.6e-11 2.9e-11 3.5e-11 1.3e-11 2.0e-11 1.9e-11 5.1e-11 5.2e-11 4.9e-11 2.2e-11 2.4e-11 Compounds Diethylene Glycol 1-Prppoxy-2-Propanol 2-Butoxyethanol 3 methoxy-3-methyl-Butanol 2-(2-Methoxyethoxy) Ethanol 1-tert-Butoxy-2-Propanol 2-tert-Butoxy-1-Propanol n-Butoxy-2-Propanol 2-(2-Ethoxyethoxy) EtOH Dipropylene Glycol 2-Hexyloxyethanol 2-(2-Propoxyethoxy)ethanol Dipropylene Glycol Methyl Ether 2-(2-Buthoxyethoxy)-EtOH 2-(2-(2-Methoxyethoxy) ethoxy) ethanol 2-(2-Ethylhexyloxy)ethanol 2-(2-(2-Ethoxyethoxy) ethoxy) ethanol 2-(2-Prppoxyethoxy) ethanol 2-(2-(2-Butoxyethoxy) ethoxyl )ethanol 2-2-(2-Butoxyethoxy) ethoxyl ) ethanol Tripropylene Monomethyl ether 2,5,8,11-Tetraoxatridecan-13ol 3,6,9,12-Tetraoxahexadecan1-ol Methyl Formate Ethyl Formate Methyl acetate Ethyl Acetate Methyl Propinate n-propyl Formate Ethyl Propionate Isopropyl Acetate Methyl Butyrate Methyl Isobutyrate n-Butyl Formate Propyl Acetate Ethyl Butyrate Isobutyl Acetate Methyl Pivalate n-Butyl Aceetate n-Propyl Propionate s-Butyl Acetate Model Name DET-GLCL PROXC3OH BUO-ETOH 3MOMC4OH MOEOETOH PG-1TB-E PG-3TB-E BUOC3OH CARBITOL DPR-GLCL EGHE DGPE DPRGOME C8-CELSV TGME EGEHE TGEE DGHE TGPE TGBE TPRGOME TETRAGME TETRAGBE ME-FORM ET-FORM ME-ACETE ET-ACET ME-PRAT C3-FORM ET-PRAT IPR-ACET ME-BUAT ME-IBUAT C4-FORM PR-ACET ET-BUAT IBU-ACET ME-PVAT BU-ACET PR-PRAT SBU-ACET Cs 4 6 6 6 5 7 7 7 6 6 8 7 7 8 7 10 8 10 9 10 10 9 12 2 3 3 4 4 4 5 5 5 5 5 5 6 6 6 6 6 6 38 MWt 106 118 118 118 120 132 132 132 134 134 146 148 148 162 164 174 178 190 192 206 206 208 250 60 74 74 88 88 88 102 102 102 102 102 102 116 116 116 116 116 116 kOH (300) (cm3/molec-sec) 5.5e-11 2.9e-11 2.6e-11 1.4e-11 6.8e-11 3.7e-11 4.9e-11 6.1e-11 5.1e-11 7.3e-11 5.8e-11 8.8e-11 9.8e-11 9.0e-11 1.1e-10 6.5e-11 1.3e-10 9.6e-11 1.3e-10 1.3e-10 1.6e-10 1.4e-10 1.7e-10 2.3e-13 1.0e-12 3.5e-13 1.6e-12 1.0e-12 2.4e-12 2.1e-12 3.4e-12 3.0e-12 1.7e-12 3.1e-12 3.4e-12 4.9e-12 9.2e-12 1.3e-12 4.2e-12 4.0e-12 5.5e-12 Compounds t-Butyl Acetate Butyl Propionate Amyl Acetate n-Propyl Butyrate 2,3-Dimethylbutyl Acetate 2-Methylbutyl Acetate 3-Methylbutyl Acetate 4-Methylbutyl Acetate Isobutyl Acetate n-Butyl Butyrate n-Hexyl Acetate Ethyl 3-Ethoxy Propionate 2,4-Dimethylpentyl Acetate 2-Methylhexyl Acetate 3-Ethylpentyl Acetate 3-Methylhexyl Acetate 4-Methylhexyl Acetate 5-Methylhexyl Acetate Isoamyl Isobutyrate n-Heptyl Acetate 2,4-Dimethylhexyl Acetate 2-Ethyl-Hexyl Acetate 3,4-Dimethylhexyl Acetate 3,5-Dimethylhexyl Acetate 3-Ethylhexyl Acetate 3-Methylheptyl Acetate 4,5-Dimethylhexyl Acetate 4-Methylheptyl Acetate 5-Methylheptyl Acetate n-Octyl Acetate 2,3,5-Trimethylhexyl Acetate 2,3-Dimethylheptyl Acetate 2,4-Dimethylheptyl Acetate 2,5-Dimethylheptyl Acetate 2-Methyloctyl Acetate 3,5-Dimethylheptyl Acetate 3,6-Dimethylheptyl Acetate 3-Ethylheptyl Acetate 4,5-Dimethylheptyl Acetate 4,6-Dimethylheptyl Acetate 4-Methyloctyl Acetate 5-Methyloctyl Acetate n-Nonyl Acetate 3,6-Dimethyloctyl Acetate 3-Isopropylheptyl Acetate 4,6-Dimethyloctyl Acetate 3,5,7-Tirmethyloctyl Acetate e-Ethyl-6-Methyloctyl Acetate 4,7-Dimethylnonyl Acetate Model Name TBU-ACET BU-PRAT AM-ACET PR-BUAT 23MC4ACT 2MC5-ACT 3MC5-ACT 4MC5-ACT IBU-IBTR BU-BUAT NC6-ACET E3DOC3OH 24MC5ACT 2MC6-ACT 3EC5-ACT 3MC6-ACT 4MC6-ACT 5MC6-ACT IC5IBUAT NC7-ACET 24MC6ACT 2ETHXACT 34MC6ACT 35MC6ACT 3EC6-ACT 3MC7-ACT 45MC6ACT 4MC7-ACT 5MC7-ACT NC8-ACET 235M6ACT 23MC7ACT 24MC7ACT 25MC7ACT 2MC8-ACT 35MC7ACT 36MC7ACT 3EC7-ACT 45MC7ACT 46MC7ACT 4MC8-ACT 5MC8-ACT NC9-ACET 36MC8ACT 3IPC7ACT 46MC8ACT 357M8ACT 3E6M8ACT 47MC9ACT Cs 6 7 7 7 8 8 8 8 8 8 8 7 9 9 9 9 9 9 9 9 10 10 10 10 10 10 10 10 10 10 11 11 11 11 11 11 11 11 11 11 11 11 11 12 12 12 13 13 13 39 MWt 116 130 130 130 144 144 144 144 144 144 144 144 158 158 158 158 158 158 158 158 172 172 172 172 172 172 172 172 172 172 186 186 186 186 186 186 186 186 186 186 186 186 186 200 200 200 214 214 214 kOH (300) (cm3/molec-sec) 4.3e-12 5.1e-12 6.1e-12 7.4e-12 1.5e-11 1.5e-11 1.5e-11 1.5e-11 1.1e-11 1.1e-11 1.5e-11 3.9e-11 1.8e-11 1.8e-11 1.9e-11 1.8e-11 1.8e-11 1.8e-11 1.4e-11 1.8e-11 2.2e-11 2.2e-11 2.2e-11 2.1e-11 2.2e-11 2.1e-11 2.1e-11 2.1e-11 2.1e-11 2.1e-11 2.4e-11 2.5e-11 2.5e-11 2.5e-11 2.4e-11 2.5e-11 2.4e-11 2.5e-11 2.5e-11 2.4e-11 2.4e-11 2.4e-11 2.3e-11 2.7e-11 2.8e-11 2.7e-11 3.0e-11 3.1e-11 3.0e-11 Compounds 2,3,5,7-Tetramethyloctyl Acetate 3,5,7-Tirmethylnonyl Acetate 3,6,8-Trimethylnonyl Acetate 2,4,6,8-Tetramethylnonyl Acetate 3-Ethyl-6,7-Dimethylnonyl Acetate 4,7,9-Trimethyldecyl Acetate 2,3,5,6,8-Pentaamethylnonyl Acetate 3,5,7,9-Tetramethyldecyl Acetate 5-Ethyl 3,6,8-Trimethylnonyl Acetate Dimethyl Carbonate Propylene Carbonate 2-Methyl Lactate 2-Methoxyehtyl Acetate Ethyl Lactate Methyl Isopropyl Carbonate 1-Methoxy-2-Propyl Acetate 2-Ethoxyethyl Acetate 2-Methyoxy-1-propyl Acetate Dimethyl Succinate Ethylene Glycol Diacetate Diisopropyl Carbonate Dimethyl Glutarate 2-Butoxyethyl Acetate Dimethyl Adipate 2-(2-Ethoxyethoxy) ethyl acetate 2-(2-Butoxyethoxy)acetate 1-Hydroxy-2,2,4Trimethylpentyl-3Isobutyrate 3-Hydroxy-2,2,4Trimethylpentyl-3Isobutyrate Oxides Ethylene Oxide Propylene Oxide 1,2-Epoxybutane Acids Formic Acid Acetic Acid Acylic Acid Propionic Acid Methyl acrylate Vinyl Acetate Model Name 2357M8AC 357M9ACT 368M9ACT 2468M8AC 3E67M9AC 479MIOAC 23568M9A 3579MIOA 5E368M9A DMC PC ME-LACT MCSVACET ET-LACT MIPR-CB PGME-ACT CSV-ACET 2PGMEACT DBE-4 ETGLDACT DIPR-CB DBE-5 2BUETACT DBE6 DGEEA DGBEA TEXANOL2 TEXANOL1 Cs 14 14 14 15 15 15 16 16 16 3 4 4 5 5 5 6 6 6 6 6 7 7 8 8 8 10 12 12 MWt 228 228 228 242 242 242 256 256 256 99 102 104 118 118 118 132 132 132 146 146 146 160 160 174 176 204 216 216 kOH (300) (cm3/molec-sec) 3.4e-11 3.4e-11 3.3e-11 3.6e-11 3.7e-11 3.6e-11 4.0e-11 3.9e-11 4.0e-11 3.3e-13 6.9e-13 2.8e-12 2.5e-11 3.9e-12 2.6e-12 1.4e-11 3.9e-11 4.6e-11 1.5e-12 7.6e-12 1.4e-11 3.5e-12 4.8e-11 8.8e-12 7.7e-11 8.6e-11 2.6e-11 3.2e-11 ETOX PROX 12BUOX FORMACID ACETACID ACYRACID PROPACID ME-ACRYL VIN-ACET 2 3 4 1 2 3 3 4 4 40 44 58 72 46 60 72 74 86 86 7.6e-14 5.2e-13 1.9e-12 4.5e-13 8.0e-13 3.7e-11 1.2e-12 6.6e-11 7.2e-11 Compounds 2-Methyl-2Butene-3-ol Ethyl Acrylate Methyl Methacrylate Butyl Methacrylate Isobutyl Methacrylate Formaldehyde Acetaldehyde Propinaldehyde 2-Methylpropanal Butanol 2,2-Dimethylpropanal 3-Methylbutanal Pentanal Glutaraldehyde Hexanal Haptanal Octanal Glyoxal Methyl Glyoxal Acrolein Crotonaldehyde Methacrolein Hydroxy Methacrolein Benzaldehyde Ketones Acetone Cyclobutanone Methyl Ethyl Ketone Cyclopentanone 2-Pentanone 3-Pentanone Cyclohexanone 4-Methyl-2-Pentanone Methyl n-Butyl Ketone Methyl-Butyl Ketone 2-Heptanone 2-Methyl-3-Hexanone Di-Isopropyl Ketone 2-Octanone 2-Nonanone Di-isobutyl ketone 2-Decanone Biacetyl Methyl vinyl ketone Hydroxy acetone Methoxy Acetone Diacetone Alcohol Phenol o-Cresol Nitrobenzene Model Name MBUTENOL ET-ACRYL ME-MACRT BU-MACRT IBUMACRT FORMALD ACETALD PROPALD 2MEC3AL IC4RCHO 22DMC3AL 3MC4RCHO IC5RCHO GLTRALD IC6RCHO IC7RCHO IC8RCHO GLYOXAL MEGLYOX ACROLEIN CROTALD METHACRO HOMACR BENZALD ACETONE CC4-KET MEK CC5-KET MPK DEK CC6-KET MIBK MNBK MTBK C7-KET-2 2M-3-HXO DIPK C8-KET-2 C9-KET-2 DIBK C10-K2 BIACETYL MVK HOACET MEOACET DIACTALC PHENNOL O-CRESOL NO2-BENZ Cs 4 5 5 8 8 1 2 3 4 4 5 5 5 5 6 7 8 2 3 3 4 4 4 7 3 4 4 5 5 5 6 6 6 6 7 7 7 8 9 9 10 4 4 3 4 6 6 7 6 41 MWt 86 100 100 142 142 30 44 58 72 72 86 86 86 100 100 114 128 58 72 56 70 70 86 106 58 70 72 84 86 86 98 100 100 100 114 114 114 128 142 142 156 86 70 74 88 116 94 108 123 kOH (300) (cm3/molec-sec) 6.6e-11 6.6e-11 6.2e-11 6.2e-11 6.2e-11 2.5e-11 2.1e-11 2.8e-11 3.4e-11 3.1e-11 3.4e-11 3.5e-11 3.6e-11 9.1e-11 5.6e-11 5.9e-11 6.2e-11 4.0e-10 1.8e-10 9.2e-11 1.1e-10 1.1e-10 1.2e-10 1.5e-10 2.2e-12 2.9e-12 3.3e-12 5.0e-12 6.6e-12 4.1e-12 8.4e-12 1.6e-11 1.1e-11 3.3e-12 1.4e-11 1.6e-11 7.4e-12 1.3e-11 1.4e-11 3.0e-11 1.5e-11 2.8e-10 9.3e-11 5.1e-12 8.8e-12 5.0e-12 2.6e-11 4.2e-11 1.5e-13 Compounds Nitrogen-Containing Compounds Para Toluene Isocyanate Toluene Diisocyanate Ethylene Diphenylene Diisocyanate Dimethyl Amine Ethyl Amine Trimethyl Amine Ethanolamine Dimethylaminoethanol Diethanol Amine Triethanolamine N-Methyl-2-Pyrrolidone Halogen-Containing Compounds Methyl Chloride Vinyl Chloride Ethyl Chloride Dichloromethane Methyl Bromide 1,1-Dichloroethane 1,2-Dichloroethane Ethyl Bromide Chloroform n-Propyl Bromide 1,1,1-Trichloroethane 1,1,2-Trichloroethane n-Butyl Bromide 1,2-Dibromoethane Trans 1,2-Dichloroethene 2-(Cl-methyl)3-Cl-Propene Trichloroethylene Perchloroethylene Monochlorobenzene Benzotrifluoride p-Dichlorobenzene p-Trifluoromethyl-ClBenzene Model Name P-TI TDI MDI DM-AMINE ET-AMINE TM-AMINE ETOH-NH2 DMAE ETOH-HN ETOH-N NMP CH3-CL CL-ETHE C2-CL CL2-ME ME-BR 11CL2-C2 12CL2-C2 C2-BR CHCL3 C3-BR 111-TCE 112CL3C2 C4-BR 11BR2-C2 T-12-DCE CL2IBUTE CL3-ETHE CL4-ETHE CL-BEN CF3-BEN CL2-BEN PCBTF Cs 7 9 15 2 2 3 2 4 4 6 5 1 2 2 1 1 2 2 2 1 3 2 2 4 2 2 4 2 2 6 7 6 7 MWt 134 174 250 45 45 59 1 89 105 149 99 50 62 64 84 95 99 109 119 119 123 133 133 137 187 97 125 131 165 112 146 147 180 kOH (300) (cm3/molec-sec) 5.9e-12 7.4e-12 1.2e-11 6.6e-11 2.8e-11 6.1e-11 3.2e-11 9.0e-11 9.4e-11 1.2e-10 2.2e-11 4.5e-14 6.9e-12 4.2e-13 1.5e-13 4.1e-14 2.6e-13 2.5e-13 3.1e-13 1.1e-13 1.2e-12 1.2e-14 2.0e-13 2.5e-12 2.3e-12 2.3e-12 3.2e-11 2.3e-12 1.7e-13 7.7e-13 4.6e-13 5.6e-13 2.4e-13 The SAPRC mechanism includes the reactions of aromatics, terpenes and alkanes with OH radicals, alkenes with O3, NO3, OH radicals and O3P, and organic radicals with NO2. Inorganic reactions are also included in SAPRC model. 42 Hydrocarbons reacting with OH radicals can undergo hydrogen abstraction or OH addition, depending on whether the group has an abstractable hydrogen or a double bond. If the group has an abstractable hydrogen, the reaction is (Carter, 2000) RH + OH. R. + H2O Where RH is any H-containing group and R. is the corresponding radical formed when the H atom is removed. If the compound has a double bond, an additional reaction can occur: >C=C< + OH. >C(OH)-C[.]- For each molecule that reacts with OH, one reaction is generated for each group in the molecule that can react in this way. Though reactions of VOCs primarily occur by OH radical initiation, under high NOx conditions, reactions of alkenes and aldehydes with NO3 radical are not negligible. Reactions with NO3 essentially proceed through the same pathways as reaction with OH radicals. For compounds that have an abstractable hydrogen, the reaction is RH + NO3 R. + HNO3 And for molecules with a double bond, the reaction is given by >C=C< + NO3 >C(ONO2)-C[.]- Besides reactions with OH radicals and NO3, reactions of some classes of VOCs (e.g., alkenes) with O3 are included in SAPRC mechanism. The reactions 43 involve O3 addition across the double bond, followed by breaking the bond and the formation of a carbonyl and an excited Crigee biradical. Reactions are given by >C=C< + O3 >CO2[excited] + >C=O Two reactions are generated for each C=C bond, forming biradical from each of the two groups around the bond. The final type of reactions is reaction with O3P radicals. O3P reactions have found to be non-negligible in some environmental chamber experiments. Reactions of O3P proceed through the addition of O3P to C=C bonds, forming an excited adduct that may decompose or undergo collisional stabilization. The important features of the chemical mechanisms for each compound class included in SAPRC model are described below Inorganic reactions Inorganic reactions represented explicitly in SAPRC include the formation and reactions of ground state O3P oxygen atoms, the oxidation reaction of NO to NO2, the reactions between HO2 and NO3, and the reaction of OH radicals with HO2. Reaction of O3P and oxidation of NO are important under relatively high NOx conditions. The reaction of HO2 with NO3 cannot be neglected under low NOx conditions and at nighttime. The reaction of OH with HO2 affects the prediction of H2O2 levels, so it is included in base mechanism. In addition to these reactions, the reaction of OH with HONO, which was omitted in the previous SAPRC mechanism is included in SAPRC-99, because of its important as a radical source. A second 44 photolysis channel for HONO., forming H. + NO2 is also added. This channel estimated to account for approximately 10% of the reaction of HONO under typical atmospheric conditions (Carter, 1990). The removal process of OH radicals by NO3 is included to account for some nighttime conditions. Organic reactions While over 300 organic reactants are represented explicitly in SAPRC, the level of characterization of reaction products is more limited. Nineteen species are employed to represent the reactions of the major organic photooxidation products. The oxygenated products include formaldehyde (HCHO), acetaldehyde (CCHO), propionaldehyde and lumped higher aldehydes (RCHO), acetone (ACET), methyl ethyl ketone and lumped higher ketone (MEK), glyoxal (GLY), methylglyoxal (MGLY), cresols (CRES), peroxyacetyl nitrate (PAN), peroxypropionyl nitrate and higher PAN analogues (PPN), and the PAN analogue formed from glyoxal (GPAN). This mechanism also includes a lumped alkyl nitrate species (RNO3), phenol (PHEN), nitrophenols (NPHE), benzaldehyde (BALD) and its PAN analogue (PBZN). Three species are used to represent uncharacterized aromatic fragmentation products (DCB1, DCB2, and DCB3). The mechanism and rate constants for the reactions for some of these oxygenated products are based on the chamber data by Atkinson (1988,1990). For compounds that do not have sufficient data from chamber experiments, the mechanisms and rate constants are estimated using data from surrogate compounds. Organic peroxy and acyl peroxy radicals 45 The photooxidations of most organics involve the formation of intermediate peroxy or acyl peroxy radical species. The major depletion routes for these peroxy radicals are either through the reactions with NO, NO2, HO2, or with other peroxy or alkyl peroxy radicals. For example, the removal processes of alkyl peroxy radicals (RO2.) proceed through these reactions: RO2. + NO RO2. + NO RO2. + NO2 RO2. + HO2 RO2. + RO2. NO2 + RO. RONO2 RO2NO2 RO2H + O2 products Reaction 1 Reaction 2 Reaction 3 Reaction 4 Reaction 5 Polluted urban atmospheres include a large number of organics that can result in a large number of radical species formed. To reduce the size of mechanism and the number of species, surrogate compounds are used to represent these individual radical species. In addition, the reaction of peroxy + peroxy and peroxy + HO2 radicals are neglected based on the assumption that under high NOx conditions, these reactions are less important than reaction of peroxy radicals with NO. Finally, organic radicals that either react rapidly or whose reactions do not depend on other reacting species are replaced by the set of products they ultimately form. A few organic radicals are represented explicitly because their reactions are sufficiently different than the reactions of other organics. These compounds include methyl peroxy radicals, acyl peroxy radicals, T-Butoxy radicals, phenoxy radicals, and nitro phenoxy radicals. 46 In the computational aspects of the model, radicals are categorized into active and steady state species. Active species such as acyl peroxy radicals and PAN analogues are model species whose concentrations need to be calculated by integrating their time rate of change. In contrast, steady state species are model species for which the steady state approximation can be employed. Radicals that react rapidly such as alkyl and alkoxy are classified as steady state species. Peroxy radical operators To reduce the potentially large number of peroxy radical mechanisms in the model, the approach of representing organic peroxy radicals with the set of products they would ultimately form, along with a set of chemical "operators" is employed. These operators include RO2R, R2O2, and RO2N. Operator RO2R Operator RO2R represents peroxy radical reactions with NO that result in NO to NO2 conversion and formation of HO2 radical. For example, the oxidation reaction of 1,3,5-trimethylbenzene with OH radical initiated by H abstraction from the aromatic ring in the presence of O2 and NO, will result in 3,5-di-methyl-benzaldehyde and formation of HO2 radical with the conversion of NO to NO2 (pathway 1 in Section 3.3.2). This full mechanism can be summarized by the use of operator RO2R. Full mechanism: 135-TMB + HO. + 2O2 + NO = NO2 + HO2. + 3,5dimethylbenzaldehyde Summarized mechanism: 135-TMB + HO. = BALD + 2C + RO2R 47 Thus the RO2R operator represents an NO to NO2 conversion, with the creation of an HO2 radical. The product 3,5-dimethyl-benzaldehyde is represented by the surrogate species benzaldehyde and two carbons are added to the products (2C) to provide a mass balance. Operator RO2R is treated as zero carbon species. The RO2R operator is also used in reactions such as: RO2R + HO2. ROOH + O2 -3C This reaction is a sink for RO2R. This reaction consumes 2 free radicals (HO2 and RO2R, which normally results in the creation of HO2) makes one less NO to NO2 conversion possible and produces an organic peroxide (ROOH). When the operator is used, however, the left hand side of the equation has no carbon present, while the right hand side has, on average, 3 carbons (the number of carbons assumed for ROOH). So, to make the reaction balance a carbon adjustment must be included by employing "lost carbon" (C) species in reactions (Carter, 2000). Still other RO2R reactions are used, but will not be described here. Operator R2O2 Operator R2O2 represents the effect of NO to NO2 conversion without HO2 formation. For example, the reaction of acetone with OH radical, O2, and NO to form HCHO, CCO-O2 and NO2. The operator R2O2 can be used to represent this full mechanism. Full mechanism: ACET + HO. + O2 + NO = HCHO + CCO-O2 + NO2 Summarized mechanism: ACET + HO. =HCHO + CCO-O2 + R2O2 48 Unlike other operators, R2O2 does not imply the formation of a radical (such as RO2R implies the formation of HO2.). Operator RO2N Operator RO2N represents the reactions of peroxy radicals with consumption of NO and various types of organic nitrate formation. For instance, aerosol precursor (APT1) reacts with OH radical via the addition at the carbon double bond, and the OH adduct then reacts with O2 and NO to form organonitrate (TPM1) (pathway 4a in Section 3.2.2). This reaction pathway can be summarized by using operator RO2N. Full mechanism: Summarized mechanism: APT1 + HO. + O2 + NO = TPM1 APT1 + HO. = TPM1 + RO2N In contrast to operator RO2R, RO2N conserves carbon. This operator is also used for reactions of peroxy radicals with species other than NO, but these reactions will not be described in detail here. Alkane reactions Gas-phase photooxidation reactions of alkanes are initiated primarily by OH radicals. The net effect of these reactions is represented by a single lumped reaction OH. + alkane a1P1 + a2P2 +... Where P1 and P2 refer to the set of organic products formed and the various chemical operators, and a1 and a2 represent yields of these products or operators. Rate constants and product yields depend on the individual alkanes represented. The OH radical reaction rate constants and the product yields parameters for individual 49 compounds were derived based on results from chamber experiments by Carter and Atkinson (1985); Atkinson (1987, 1988, 1990). Alkene reactions Unlike alkane hydrocarbons, alkenes can react with O3, NO3, and OH radicals. SAPRC mechanism contains reactions of alkenes with all these radicals. In addition to primary initiation reactions of alkenes by those radicals, reactions of alkenes with O3P are not negligible under high NOx conditions such as in plumes. Therefore, they are included in SAPRC mechanism to improve its range of validity. Aromatic reactions Mechanisms for aromatic hydrocarbons are based on reactions of OH radicals with aromatics involving hydrogen abstraction or addition of OH radical to the aromatic ring. H abstraction reactions result primarily in aromatic aldehydes, ketones, and small yields of aromatic nitrate. OH-addition to the ring forms phenols or cresols, and various ring fragmentation products. Aromatic + OH. abstraction Aromatic aldehyde or ketone or aromatic nitrate + operators + HO2. Aromatic + OH. addition Phenol or cresol + -dicarbonyls and other fragmentation products + operators + HO2. Reactive aromatic fragmentation products that do not undergo significant photodecomposition to radicals are represented by DCB1. Reactive aromatic fragmentation products which photolyze with alpha-dicarbonyl-like structures are 50 denoted by DCB2. DCB3 signifies reactive aromatic fragmentation products which photolyze with acrolein like reactivity. Overall reactions of aromatics are represented as: OH. + aromatic yRHHO2. + YRRRO2R + YNRRO2N + yQ2RCOO2 + yPHPHEN + yCRCRES + yBLBALD + yKGPROD2 + yGLGLY + yMGMGLY + yBABACL + yD1DCB1 + yD2DCB2 + yD3DCB3 Where yRH, yRR,..... are stoichiometric parameters. Mechanisms represented in the SAPRC model are assumed to be gas phase. A maximum of four alphanumeric characters are allowed to represent model species . For example, CRES is notation refers to cresol, and BALD is used to represent benzaldehyde. Rate constants used in SAPRC are based on recommendations derived from Atkinson's atmospheric chamber experiments (1989, 1994, 1997). Unless referred to the falloff expression, these rate constants are given by the expression k = A(T/300)B exp(-Ea/RT) Equation 6 Where k and A are in units of cm3/molc-sec , T is the temperature in K, and R is 0.0019872 kcal/ K-mole. For the fall off expression, the rate constant is both temperature and pressure dependent and is given by k = [(k0*M)/(1+(k0*M/k))]*fg g = 1/[1+(log10[k0*M/k]/n)2] 51 Equation 7 Equation 8 M is the pressure in molecules cm-3, unless indicated otherwise f = 0.6 and n=1. k0 and k are the rate constants at the low and high pressure limiting values. Lumped species in SAPRC-99 model To reduce the size of the mechanism and the number of species modeled in SAPRC, a group of compounds that have similar chemical and physical properties are lumped together and represented by surrogate compounds. The lumping approaches employed in SAPRC model are the variable lumped parameter and the lumped molecular method. For the first approach, the kinetic and product yield parameters of lumped model species are weighted by reactivity or molar concentration of the mixture of VOCs presented. The details of weighting methods applied in this approach are elaborated in the next chapter. For the second approach, compounds are represented by a model species on a molecule-for molecule basis. More details are addressed in the next chapter. Primary alkane hydrocarbons are lumped and represented by five model species: ALK1, ALK2, ALK3, ALK4, and ALK5, depending on their OH rate constant. Alkenes are separated and represented by two model species: OLE1 and OLE2 based on their OH rate constant. Model species ARO1 and ARO2 represent groups of slow and fast reactivity aromatic hydrocarbons, corresponding to their OH rate constant. Terpenes except isoprene, which is the dominant biogenic, are represented by TRP1. VOCs are lumped into these nine model species using the variable parameter approach. The reactive products of VOCs such as unsaturated aldehydes and ketones are lumped into product structure. 52 Reactive ketones whose reactions cannot be represented using lumped alkane classes are represented by MEK and PROD2. Emitted species that are represented explicitly include methene, ethene, isoprene, formaldeyde, acetone, methanol, and phenol. Table 3.2 summarizes lumped classes and lumped molecules employed in SAPRC model. 53 Table 3.2: Summaries of lumped classes and lumped molecule employed in SAPRC-99 model Model Species Description Emitted Compounds Represented Explicitly CH4 Methane ETHENE Ethene ISOPRENE Isoprene HCHO Formaldehyde ACET Acetone MEOH Methanol PHEN Phenol Lumped Molecule Groups CCHO Acetaldehyde and Glycolaldehyde RCHO Lumped C3+ Aldehydes MEK Ketones that react with OH radicals slower than 5.0E-12 cm3/molec2-sec PROD2 Ketones that react with OH radicals faster than 5.0E-12 cm3/molec2-sec CRES Cresols BALD Aromatic aldehydes (e.g., benzaldehyde) METHACRO Methacrolein and acrolein ISOPROD Unsaturated aldehydes other than acrolein and methacrolein MVK Unsaturated ketones Unreactive Compounds INERT Compounds other than CO or methane that do not react, or react only with OH with a rate constant less than approximately half that of ethane, or 2.0E+2 ppm-1min-1 Lumped Parameter Groups (Lumped using molar weighting except as indicated) ALK1 ALK2 ALK3 ALK4 ALK5 ARO1 Alkanes and other non-aromatic compound that are react only with OH, and have an OH rate constant between 2.0E+2 and 5.0E+2 ppm-1 min-1 Alkanes and other non-aromatic compound that are react only with OH, and have an OH rate constant between 5.0E+2 and 2.5E+3 ppm-1 min-1 Alkanes and other non-aromatic compound that are react only with OH, and have an OH rate constant between 2.5E+3 and 5.0E+3 ppm-1 min-1 Alkanes and other non-aromatic compound that are react only with OH, and have an OH rate constant between 5.0E+3 and 1.0E+4 ppm-1 min-1 Alkanes and other non-aromatic compound that are react only with OH, and have an OH rate constant greater than 1.0E+4 ppm-1 min-1 Aromatics with kOH less than 2.oE+4 ppm-1 min-1 (Primarily toluene and other monoalkybenzenes) Benzene and slower reacting aromatics such as halobenzenes are lumped with reactivity weighting based on their OH rate constant relative to that of toluene, all others are lumped using molar Aromatics with kOH greater than 2.oE+4 ppm-1 min-1 (Primarily xylems and polyalkyl benzenes) 54 ARO2 Model Species OLE1 OLE2 TRP1 Description Alkenes (other than ethane) with kOH less than 7.0E +4 ppm-1 min1 (Primarily terminal alkenes) Alkenes with kOH greater than 7.0E+4 ppm-1 min-1 (Primarily internal or disubstituted alkenes) Biogenic alkenes other than isoprene (primarily terpenes) Inputs required for SAPRC model Inputs required to perform SAPRC simulations include four groups of parameters. The first group represents initial conditions, which are the initial composition of reacting mixture, temperature, and relative humidity. The second group describes the geographical data such as locations and date and time the simulations are performed which set solar intensity for photolysis rate calculations. The third group includes emission rates of emitted species. In addition to these three groups, parameters to characterize the loss mechanisms such as chamber wall loss coefficients, and deposition rates are optional inputs. Along with these inputs, the lumped control definitions can be modified and incorporated into the model, in case the lumped model species are assigned differently from the existing definition in SAPRC model. 55 Figure 3.1: Overview of relationships between data files and programs used in the preparation of chemical mechanism for airshed model calculation (Carter, 1988) Figure 3.1 demonstrates the data preparation process required in modeling simulations by the SAPRC model. All data required to prepare mechanisms and run simulation within SAPRC are provided in file model.prp. SAPRC uses these data to generate the file model.mod and model.ftn, which are required for implementation of the mechanisms in an airshed model. The file model.pro is also generated in the preparation process. This file contains the mechanisms generated by SAPRC, which are suitable for use by airshed models, a list of the mechanism species, and kinetic parameters. After all parameters are prepared, the file model.ftn, which contains Fortran source code, implements the chemical mechanisms. SAPRC is able to 56 calculate: 1) rates of all reactions, 2) concentrations of the steady state species, and 3) concentrations and formation and destruction rates of the active species. 3.3 Quantitative models of SOA formation 3.3.1 Previous smog chamber investigations In order to develop chemical mechanisms for secondary organic aerosol formation it is necessary to identify hydrocarbons that are aerosol precursors and the rates at which they are emitted. Table 2.1 (in Chapter 2) illustrates an emission inventory of reactive organic compounds released into the Southern California air basin, which was organized by Grosjean and Seinfeld (1989). Grosjean and Seinfeld then used expert judgment to estimate aerosol yields from each of these precursors. The goal of the thesis is to convert the fractional aerosol coefficients developed by Grosjean and Seinfeld based on expert judgment, to a more fundamental mechanistic basis. The analysis is based on a series of smog chamber experiments that were performed by previous investigators (e.g., Eusebi, 1996). These prior investigations examined the products formed in the photooxidation of hydrocarbon precursors, which are important in the formation of SOA. Compounds were selected for the smog chamber experiments based on their relative contribution to secondary loading in Los Angeles and the assumption that similar compounds (e.g., alkenes) would yield similar reaction products (Eusebi, 1996). 57 The experiments were performed by injecting the selected hydrocarbon precursors along with nitrogen oxides into batch reactors with Teflon walls, illuminated by ambient sunlight. The composition of aerosol produced in the chambers was monitored by collecting aerosol samples using a low pressure impactor and analyzing the collected samples using infrared microscopy. The IR data were used to estimate the relative molar loading of functional groups in the aerosol, and from these data, average product structures were calculated (Appendix A). This information was then used by Eusebi and others to develop qualitative photochemical mechanisms for organic aerosol formation (e.g., Holes et al. 1997, Dekermenjial et al., 1999). A major goal of this work is to convert these qualitative mechanisms into quantitative models of aerosol formation. This is specifically accomplished by developing quantitative mechanistic models of the chemical reactions that convert gas phase hydrocarbons into low volatility particulate matter. 3.3.2 Development of quantitative models of SOA formation Pandis et al., (1992) found that a majority of secondary organic aerosol is attributed to aromatic precursors. In Los Angeles, aromatics are estimated to contribute 65% of the SOA. The significance of aromatic precursors in SOA formation indicates the importance for further studies into the reaction pathways and products from the atmospheric reactions of these species. For this reason, aromatic hydrocarbon precursors were selected for this study. Aromatic hydrocarbons selected 58 for this study are species modeled in SAPRC-99. The investigated precursors are listed in Table 3.3. Table 3.3: VOC Emission Inventory and Secondary Aerosol Yields of Hydrocarbon Precursors Selected for Study Hydrocarbon Precursor Amount of VOC emitted (Southern California) (kg / day) (%) (kg / day) (%) Fractional aerosol coefficient Fractions of VOC reacted Amount of aerosol produced Contribution in the total SOA Toluene o-Xylene m-Xylene p-Xylene Ethylbenzene 1,3,5-Trimethylbenzene 1,2,4-Trimethylbenzene 1,2,3-Trimethylbenzene n-Propylbenzene Isopropyl benzene o-Ethyltoluene m-Ethyltoluene p-Ethyltoluene Sec-butylbenzene 105,480 19,614 12,186 12,057 8,481 4,667 7,029 4,133 591 208 1,282 4,126 1,345 174 5.4 5.0 4.7 1.6 5.4 2.9 2.0 3.6 1.6 4.0 5.6 6.3 2.5 2.6 0.12 0.26 0.4 0.28 0.15 0.74 0.58 0.51 0.12 0.13 0.23 0.31 0.21 0.19 683.5 255.0 229.1 54.0 68.7 100.1 81.5 75.9 1.1 1.1 16.5 80.6 7.1 1.0 13.6 5.0 4.6 1.1 1.4 2.0 1.7 1.6 0.02 0.02 0.3 1.6 1.4 0.02 Conversion of gas phase aromatic hydrocarbon species to low volatility particulate matter occurs via a series of oxidation steps. Although ozone and NO3 can participate, the majority of the oxidation reactions are initiated by hydroxyl radical. As shown below for trimethylbenzene, these reactions can either proceed through abstraction of hydrocarbon or via OH addition. 59 (1) Figure 3.2: Hydrogen abstraction from 1,3,5-trimethylbenzene by OH radical (2) (3) Figure 3.3: OH addition to 1,3,5-trimethylbenzene reaction pathways 60 (4) Figure 3.4: Semi-volatile product reaction pathways from aerosol precursors Pathway 1 in figure 3.2 denotes initiation through H abstraction and pathways 2 and 3 show initiation via OH addition (Atkinson, 1990; 1994). Perry (1977) suggested that only 2.0 to 3.5% of the 1,3,5-trimethybenzene that reacts proceeds through the hydrogen abstraction route. The majority of gas-phase 1,3,5trimethybenzene photooxidation reaction results in the dicarbonyl ring cleavage product (denoted aerosol precursor from trimethylbenzene, APT1). APT1 further reacts with OH radical to form condensable products TPM1 and TPM2 (trimethylbenzene-derived particulate matter), pathways 4a and 4b. Thus, this overall mechanism is a series/parallel network: A+B C+B Where: A = 1,3,5-trimethybenzene B = OH radical 61 C D C = APT1 (dicarbonyl) D = condensable product In order to convert the current trimethylbenzene reaction in SAPRC, which only accounts for the initial reaction of 1,3,5-trimethylbenzene, into a form that accounts for SOA, a specific format is required. Gas phase reaction in SAPRC: 135-TMB + HO. = .18HO2. + .804RO2R. + .01RO2N + .621MGLY + .18CRES + .03BALD + .569DCB1 + .097DCB2 + .114DCB3 + 2.273XC Based on the distribution suggested by Perry, 3% of products resulted from initiation by hydrogen abstraction, and 97% resulted from OH radical addition. In SAPRC notation, the hydrogen abstraction route of 1,3,5-trimethybenzene results in a dimethylbenzaldehydye (represented as benzaldehyde and two extra carbon atoms, BALD + 2C), along with an NO to NO2 conversion and the production of HO2 radical. Notation BALD is used to represent benzaldehyde, which has the structure as shown below. Since the product from the hydrogen abstraction of 1,3,5trimethybenzene has similar structure with benzaldehyde but with two extra carbon atoms (as shown below), and because every single product cannot be included in the SAPRC mechanism, the dimethylbenzaldehyde is represented by BALD + 2XC. 62 3%: 135-TMB + HO. = BALD + 2XC + RO2R Reaction 6 This route is assumed to account for 3% of the reacted 1,3,5-trimethybenzene (SAPRC99). Addition of the OH radical to trimethylbenzene occurs at any of the three unsaturated aromatic carbons. All are equivalent. Initial reaction with OH radical will produce cresol with two extra carbon atoms (CRES +2XC) and HO2 radical. This pathway accounts for 18.6% of the reacted trimethylbenzene (SAPRC99). 18%: 135-TMB + HO. = CRES + 2XC + HO2. Reaction 7 In the original SAPRC mechanism, the remainder of the OH-adduct (79% of total reaction) further reacts with O2 with conversion of NO to NO2 to form uncharacterized reactive aromatic ring fragmentation products (DCB1, DCB2, and DCB3), MGLY, and peroxy radical. 79%: 135-TMB + HO. = DCB1 + DCB2 + DCB3 + MGLY + RO2R In the modified mechanism accounting for aerosol formation, the rest of the OH radical addition products further react with O2 with conversion of NO to NO2 to generate methyglyoxal (MGLY), peroxy radical, and unsaturated dicarbonyls (referred to as Aerosol Precursor (APT1)). This pathway accounts for 79% of all reactions (Bandow, 1985). 79%: 135-TMB + HO. = MGLY + APT1 + RO2R Net: 135-TMB + HO. = .03 BALD + .18 CRES + .81 RO2R + .18 HO2. + .79 MGLY + .79 APT1 + .42 XC 63 Reaction 9 Reaction 8 The species identified as BALD (pathway 1 in Figure 3.2), CRES (pathway 2 in Figure 3.3), MGLY (pathway 3 in Figure 3.3), TPM1, TPM2 and APT1 (pathway 4 in Figure 3.4) represent actual chemical species. The identifications RO2R, RO2 and RO2N represent operators, which indicate product formation pathways. Aerosol precursor (APT1) further reacts with OH radicals via addition at the carbon double bond to form two types of condensable products (TPM1 and TPM2). TPM1 is the aerosol product containing organonitrate group (as shown in pathway 4a, Figure 3.4) and TPM2 (pathway 4b in Figure 3.4) is the product without the organonitrate group. TPM1 is produced through OH radical addition to the double bond in APT1. The OH-adduct reacts with O2 and NO to form organonitrate. Experimental results from Eusebi (1996) suggest that 25% of APT1-OH addition reactions produce TPM1. The remaining APT1-OH-adduct reacts with O2 and NO to form TPM2, NO2, and HO2 radical. APT1 + HO. = .25 TPM1 + .75 TPM2 + .25 RO2N + .75 RO2R Reaction 10 Reactions 13 and 14 describe the condensed mechanisms of SOA formation for 1,3,5-trimethybenzene which proceed through two steps: gas phase reaction of 1,3,5-trimethybenzene with OH radicals to produce aerosol precursor (APT1), and reaction of APT1 with OH radical to form two types of potential aerosol products (TPM1 and TPM2). Stoichiometric coefficients in reactions 10-14 are obtained from the evaluation of atmospheric chamber experiments. Table 3.4 summarizes these coefficients and cites the source of the data. 64 Table 3.4: Kinetic parameters of 1,3,5-trimethybenzene reactions with OH radicals forming aerosol products and their sources Notations CRES BALD MGLY HO2. RO2R APT1 RO2N TPM1 TPM2 XC Descriptions Cresol Benzaldehyde Methylglyoxal Hydroperoxy radicals Operator RO2R Aerosol precursor Operator RO2N Aerosol product species 1 Aerosol product species 2 Extra carbon Stoichiometric Coefficients (SC) 0.18 0.03 0.79 0.18 0.82 (reaction 13) 0.75 (reaction 14) 0.79 0.25 0.25 0.75 0.42 Sources of SC SAPRC-99 SAPRC-99 Bandow, 1985 SAPRC-99 Bandow. 1985, SAPRC-99 Eusebi, 1996 Bandow, 1985 Eusebi, 1996 Eusebi, 1996 Eusebi, 1996 SAPRC-99 Besides 1,3,5-trimethybenzene, 1,2,4-trimethybenzene is a significant source of secondary aerosol formation as well. Odum (1996) performed a set of smog chamber experiments for 1,2,4-trimethybenzene to study the formation of secondary aerosol yield in terms of organic mass concentration. In this research, 1,2,4trimethybenzene was studied by using information from Odum's experiments, the knowledge of chemical reaction pathways examined by Bandow (1985), and the application of 1,3,5-trimethylbenzene smog chamber results by Eusebi (1996). The reaction pathways for 1,2,4-trimethybenzene are shown in Figures 3.5-3.7. Figure 3.5: The H-abstraction from aromatic ring by the reaction of OH and 1,2,4-TMB 65 Figure 3.6: The OH-addition from aromatic ring by the reaction of OH and 1,2,4 TMB Figure 3.7: The ring fragmentation by the reaction of OH and 1,2,4-TMB Figure 3.8: Aerosol product formation from oxidation reaction of 1,2,4-TMB Similar to 1,3,5-trimethybenzene, oxidation of 1,2,4-trimethybenzene either proceeds through OH addition to the aromatic ring, or abstraction of H from the methyl group. It was found that 96% of products resulted from the addition of OH 66 radical to the aromatic ring, and 4% resulted from initiation by H abstraction (SAPRC99). The H abstraction route occurs primarily from the methyl substituents. In SAPRC notation, hydrogen abstraction reaction route of 1,2,4-trimethybenzene results in benzaldehyde and two extra carbon atoms (BALD + 2XC), with an NO to NO2 convension and the production of an HO2 radical (Figure 3.5). 4%: 124-TMB + HO. = BALD + 2XC + RO2R Reaction 11 This route is assumed to account for 4% of the 1,2,4-trimethybenzene that reacts. Initial reaction with OH radical can produce cresol with two extra carbon atoms (CRES + 2XC) and HO2 radical (Figure 3.6). This pathway accounts for 18.6% of the reacted 1,2,4-trimethybenzene. 18%: 124-TMB + HO. = CRES + 2XC + HO2. Reaction 12 The remaining OH radical addition products further react with O2 and NO to NO2 conversion to generate methyglyoxal (MGLY), glyoxal (GLY), biacetyl (BACL), and unsaturated dicarbonyls (refered to as Aerosol Precursors (APT5-9, and AP10). The addition of O2 to the OH-aromatics adduct occurs at a position which yields conjugated double bonds. The ring cleavage reaction of the O2-OH aromatics adduct can proceed through six possible positions on the aromatic ring. All of these potential reaction proceed through two similar steps: the abstraction of the O atom by NO, and the abstraction of the H atom by O2. Figure 3.7 demonstrates the reaction pathway of APT5 production. Although the chemical structures of each speciess are different, the reaction pathway of APT5 production is similar to those for APT6-APT9 and 67 AP10 production. The cleavage reactions produce 25.4 % of APT5, 8.4 % of APT6, 13.1 % of APT7, 14.6 % of APT8, 4.6 % of APT9, and 10.7% of AP10. In addition to these unsaturated speciess, methyglyoxal (MGLY), glyoxal (GLY), and biacetyl (BACL) are also produced with the yields of 0.469, 0.192, and 0.107, respectively (Bandow, 1985). Formation of APT5-APT9 and AP10 accounts for 78% of all reactions (Bandow, 1985). 79% 124-TMB + HO. = .593MGLY + .243GLY + .135BACL + .321APT5 + .106APT6 + .166APT7 + .184APT8 + .058APT9 + .137AP10 + 1.26RO2R Reaction 13 Net: 124-TMB + HO. = .044BALD + .186CRES + .812RO2R + .18HO2. + .469MGLY + .192GLY + .107BACL + .254APT5 + .084APT6 + .131APT7 + .146APT8 + .046APT9 + .107AP10 + .46XC Reaction 14 Aerosol precursors (APT5-APT9, and AP10) react with OH radicals via addition at the carbon double bond to form two types of aerosol products, TPM6 and TPM7, which can be classified by the same criteria as for 1,3,5-trimethybenzene products. TPM5 is produced through OH radical addition, similar to TPM1. The OH adduct then reacts with O2 and NO to form an organonitrate speciess. Results from Eusebi suggest that 25% of APT5-OH addition reaction leading to formation of TPM5 (Eusebi, 1996). The remaining APT5-OH reacts with O2 and NO to form TPM6, NO2 and HO2 radical. APT5 + HO. = .75TPM6 + .75RO2R + .25TPM5 + .25RO2N Reaction 15 68 Analogous to 1,3,5-trimethylbenzene, reactions 18 and 19 describe the condensed mechanisms of SOA formation for 1,2,4-trimethylbenzene, which proceed through two steps: gas phase reaction of 1,2,4-trimethylbenzene with OH radicals to produce aerosol precursor (APT5-APT9, and AP10), and reaction of aerosol precursors with OH radical to produce two types of condensable products (TPM5 and TPM6). Stoichiometric coefficients in reaction 18-19 are estimated based on smog chamber experiments. Table 3.5 lists these coefficients and cites the source of the data. Table 3.5: Kinetic parameters of 1,2,4-trimethylbenzene reactions with OH radicals forming aerosol products and their sources Notations BALD CRES GLY BACL MGLY HO2. RO2R Descriptions Benzaldehyde Cresol Glyoxal Biacetyl Methylglyoxal Hydroperoxy radicals Operator RO2R Stoichiometric coefficients (SC) 0.04 0.18 0.192 0.107 0.469 0.18 0.812 (reaction 18) 0.75 (reaction 19) APT5 APT6 APT7 APT8 APT9 AP10 RO2N TPM5 TPM6 XC Aerosol precursor Aerosol precursor Aerosol precursor Aerosol precursor Aerosol precursor Aerosol precursor Operator RO2N Aerosol product species 1 Aerosol product species 2 Extra carbon 69 0.254 0.084 0.131 0.146 0.046 0.107 0.25 0.25 0.75 0.46 Sources of SC SAPRC-99 SAPRC-99 Bandow, 1985 Bandow, 1985 Bandow, 1985 SAPRC-99 Bandow, 1985, SAPRC-99 Estimated from 1,3,5trimethylbenzene Bandow, 1985 Bandow, 1985 Bandow, 1985 Bandow, 1985 Bandow, 1985 Bandow, 1985 Estimated from 1,3,5trimethylbenzene Estimated from 1,3,5trimethylbenzene Estimated from 1,3,5trimethylbenzene SAPRC-99 As described above, formation of secondary organic aerosol proceeds through a series/parallel network. To analyze this network quantitatively, all reaction rates must be determined. Many of the reactions are well studied (such as trimethybenzene + OH radical) while for others rate constants must be estimated (e.g., APT1 + OH radical). The reaction rate constants for the reaction of aerosol precursors, such as APT1 with OH radical, were calculated by using the Atmospheric Oxidation Program. The Atmospheric Oxidation Program for Microsoft Windows 3.1 (AOPWIN) estimates the rate constant for atmospheric gas-phase reactions between photochemically produced hydroxyl radicals and organic chemicals (Meylan, 1998). The estimation methods used by AOP are based on the structure-activity relationship (SAR) method developed by Dr. Roger Atkinson (Atkinson, 1985, 1986, 1987, 1991; Atkinson and Carter, 1984; Biermann et al., 1985; Kwok et al., 1992, Kwok and Atkinson, 1995; Kwok et al., 1996). This method uses the fact that the reactions of gas phase organic compounds with OH radical can proceed by four possible pathways. These pathways are: H-atom abstraction from C-H and O-H group, OH radical addition to >C=C< and -CC- bonds, OH radical addition to aromatic rings, and OH radical reaction with N-, S-, and P-atoms. H-atom abstraction from C-H and O-H bonds The rate constants for H-atom abstraction from CH3, -CH2, >CH- groups depend on the identity of the substituents attached to these groups, with 70 k(CH3-X) k(X-CH2-Y) k(X-CH Y Z = kprimF(X) = ksecF(X)F(Y) )) = ktertF(X)F(Y)F(Z) Equation 9 Equation 10 Equation 11 Where kprim, ksec, ktert are the group rate constants for H-abstraction from CH3, -CH2, and >CH- groups, respectively, and F(X), F(Y), and F(Z) are the substituent factors for the substituent groups X, Y, and Z, respectively. Methyl group is chosen to be the standard substituent group, with F(-CH3) = 1.00. Temperature dependence of the substituent factors F(X) is expressed as (Atkinson, 1995) F(X) = eEx/T . Group rate constants kprim, ksec, and ktert were obtained from available database for the alkanes. The group rate constants at a wide range of temperatures were derived using the correlation k= CT2e-D/T. For compound classes other than alkanes, F(X) were derived from non-linear least square analysis based on the rate constants given by Atkinson (1989, 1994). The comparison between the calculated and experimentally measured rate constants presents an excellent agreement within a factor of 1.37 over the temperature ranges. Only 36 out of total 290 organic compounds (12%) show discrepancies over a factor of 2 between the calculated and experimental values. For example, the calculated and the experimental rate constants for H abstraction for cyclobutanone are 4.5E-12 and 0.87E-12 cm3/molec-sec, respectively. OH radical addition to >C=<C and -CC- bonds 71 Rate constants for OH radical addition to carbon-carbon unsaturated bonds depend on the number, identity, and position of substituent groups around these bonds. Conjugated double bond systems are dealt with by considering both double bonds as a single unit. Rate constants for addition to CH=CH2-, CH2=C<, c-and t- CH=CH-, -CH=C<, and >C=C< groups are based on those for propane, 2methylpropene, c-and t-2-butene, 2-methyl-2-butene, and 2,3-dimehtyl-2-butene, respectively. If the substituent groups are not alkyl groups, group substituent factors C(X) are utilized. For example, CH2=CHX k(CH2=CHX) = k(CH2=CH)C(X) Equation 12 C(X) were derived by Atkinson from the available database for haloalkens, nitriles, and oxygenated organic compounds containing >C=C< bonds. The calculated and experimental rate constants of reactions involve OH radical addition to carbon-carbon unsaturated bonds are reliable within a factor of 2, and only six of total 98 compounds (6%) are different more than a factor of 2. For instance, the calculated and the experimental rate constants for OH radical addition to >C=C< for CH2=CCl2 are 2.67E-12 and 10.9E-12 cm3/molc-sec, respectively. OH radical addition to aromatic ring Rate constants for OH addition to aromatic rings are estimated using the correlation between the OH radical addition rate constant and the sum of the electrophilic substituent constants +. A least square analysis of monocyclic aromatics and biphenols yield the correlation 72 Log10kadd = -11.71-1.34+ Equation 13 The electrophilic substituent factors (+) were obtained from experimental database by Atkinson (1991). The estimated rate constants from AOP programs for 12 out of 66 for monocyclic aromatics and biphenyls (18%) differ from the experimental values by greater than a factor of 2. For example, the calculated and the experimental rate constants for OH radical addition to o-xylene are 6.51E-12 and 13.7E-12 cm3/molcsec, respectively. The upper bound rate constant of ~2.0E-10 cm3molc-1sec-1 is recommended for reactions involving OH addition to aromatic rings. The estimation of rate constants of PAH and alkyl-substituted PAH by AOP is uncertain due to a large number of assumptions such as ionization potential. OH radical interaction with N-, S-, and P- containing groups The group rate constants and substituent factors for reactions involving OH radical interaction with N-, S-, and P- containing groups were derived from experimental data by Atkinson, 1989. The calculated rate constants at room temperature for 34 N-, S-, and P- containing organic compounds are consistent with the experimental values to within a factor of 2. These rate constant estimation methods were not used in this work, however, so they will not be described here. Condensed mechanisms and rate constants for 1,3,5-and 1,2,4trimethylbenzene, calculated from the AOP program, are reported in the following table. 73 Table 3.6: Condensed mechanisms and rate constants for 1,3,5-and 1,2,4-trimethylbenzene Aromatic Precursors 1,3,5Trimethybenzene Aerosol Precursors Condensed Mechanisms 135-TMB + HO. = 0.03BALD + 0.18CRES + 0.82RO2R + 0.18HO2. + 0.79MGLY + 0.79APT1+ .42XC APT1 + HO. = 0.25TPM1 + 0.75TPM2 + 0.25RO2N + 0.75RO2R 124-TMB + HO. = 0.04BALD + 0.18CRES+ 0.812RO2R + 0.18HO2. +0.469MGLY + 0.192GLY +0.107BACL + 0.254APT5 + 0.084APT6 + 0.131APT7 + 0.146APT8 + 0.046APT9 + 0.107AP10 + 0.46XC APT5 + HO. = 0.25TPM5 + 0.75TPM6 + 0.25RO2N + 0.75RO2R APT6 + HO. = 0.25TPM5 + 0.75TPM6 + 0.25RO2N + 0.75RO2R APT7 + HO. = 0.25TPM5 + 0.75TPM6 + 0.25RO2N + 0.75RO2R APT8 + HO. = 0.25TPM5 + 0.75TPM6 + 0.25RO2N + 0.75RO2R APT9 + HO. = 0.25TPM5 + 0.75TPM6 + 0.25RO2N + 0.75RO2R AP10 + HO. = 0.25TPM5 + 0.75TPM6 + 0.25RO2N + 0.75RO2R Rate Constants (cm3/molc-sec) 5.79e-11 Source of Rate Constant SAPRC-99 APT1 1,2,4Trimethylbenzene 2.60e-11 3.35e-11 AOP Program SAPRC-99 APT5 APT6 APT7 APT8 APT9 AP10 4.08e-11 2.60e-11 2.60e-11 3.30e-11 1.20e-11 1.00e-11 AOP program AOP program AOP program AOP program AOP program AOP program 74 3.4 G/P Partitioning model 3.4.1 Theoretical Background In addition to understanding the chemical mechanisms leading to secondary organic aerosol formation, it is also necessary to understand the partitioning of semi volatile secondary organic aerosol (SOA) species between gas and particle phases. This partitioning has been described by two conceptual frameworks. The first involves gas/particle partitioning through nucleation and condensation (McMurry and Grosjean, 1985). This principle assumes that once a vapor phase product exceeds its saturation concentration, it begins to condense onto existing seed particles. Aerosol mass then grows as a result of condensation of a condensable species. The second concept involves gas/particle partitioning according to simple thermodynamic principles such as Raoults' law (Pankow, 1994; Odum 1996). In this case, the condensable product will partition to the particulate phase at all gas-phase concentrations. The extent of particle partitioning can be described by the equilibrium between the gas-phase concentration of the condensable products and its concentration in the condensed phase. The basis for work presented in this thesis is the framework described by Odum. Odum et al. (1996) expressed SOA formation using the gas/particle partitioning absorption mechanism of Pankow (1994). In this framework, the amount of condensable product that is in the particulate phase is 75 related to the total particle mass concentration. A partitioning coefficient for species i (Kom,i) can be defined in terms of the organic mass concentration (Odum, 1996). Kom, i = Fi, om Ai * Mo Equation 14 Pi = Ai + Fi,om Where: Equation 15 Fi,om = concentration of compound i in the absorbing organic material (om) phase ( g/m3) Mo Ai Pi = total organic mass concentration ( g/m3) = gas phase concentration of compound i ( g/m3) = total concentration of product i that is formed ( g/m3) Kom,i = partitioning coefficient of compound i (m3/ g) To examine the dependence of SOA yield on the total organic aerosol mass concentration, Odum assumed that the concentrations of individual products produced from photooxidation of reactive organic gases (ROG) are proportional to the amount of ROG that reacts in which i = Where: Pi ROG Equation 16 ROG = the amount of reactive organic gas reacts i = proportionality constant relating the amount of ROG reacted to Pi 76 Combining the above equation with the definition of partitioning coefficient (equation 14) and equation 15 yields the expression for the overall SOA yield as a function of Mo as Yi = Mo (1 + Kom, i * Mo ) i i i * Kom, i Equation 17 Odum examined how SOA smog chamber data can be used to identify important sources of SOA in the urban environment using the expressions for secondary aerosol yields (Yi) (equation 17). The expression was applied to over 30 chamber experiments performed in the summer of 1995. To fit the observed yield to equation (17), oxidation products for each parent hydrocarbon have been represented by two empirical products. Odum fit his experimental data for each compound with 4 adjustable parameters. He assumed four parameters: Kom,1, Kom,2, 1, and 2 that would fit the results from model calculations and experiments the best. While this study will use the Odum framework, the 4 adjustable parameters will be condensed to a single adjustable parameter. Alpha values (1 and 2) will be obtained from the chemical mechanism described in the previous section. The partitioning parameters Kom,1 and Kom,2 will be assumed to have the same ratio as the vapor pressure of the aerosol products, i.e., Kom,1/Kom,2 = Vapor pressure of other products /Vapor pressure of organonitrate products. Therefore, compared to Odum study (which used 4 adjustable parameters 1, 2, Kom,1, and Kom,2) in this work only one parameter (Kom,1) is adjusted to fit the SOA mass changes (Mo). 77 The goal of this section of thesis is to employ both chemical mechanisms and the absorption partitioning model in determining SOA formation and estimates of SOA yields. This is accomplished by combining results from modified SAPRC mechanisms, as described in the preceding section, with partitioning between gas and particle phases model. 3.4.2 Estimates of Model Parameters Once the quantitative models of SOA formation were developed and integrated into the SAPRC model, a series of model parameters were required. These parameters can be separated into two groups. The first group includes the kinetic parameters, which were described in the previous section. The second group involves the partitioning parameters, which are based on a single adjustable parameter, Kom,1. Table 3.7 lists the partitioning parameters, including their descriptions, and the procedures used to estimate them. Table 3.7:Partitioning Parameters used in the estimates of SOA formation from reactions of hydrocarbon precursors Parameters Kom,i i Pi ROG VPi Descriptions Partitioning coefficients of compound i Proportionality constant relating the amount of ROG reacted to the total concentration of product i Total concentration of product i Concentration of ROG that reacts Vapor pressure of product i Calculation Procedures Estimated from chamber data Estimated from SAPRC rate parameters and branching ratios from chamber data SAPRC simulation based on ambient Houston data SAPRC simulation based on ambient Houston data Antoine method and modified Grain method 78 A box model has been set up to examine the yields of secondary aerosol from individual hydrocarbon precursors. The box model uses the full SAPRC mechanism, including the additions of pathway to account for aerosol formation. The simulations were conducted under similar conditions to those performed in Odum's studies (1996, 1997). Initial conditions used in Odum's experiments, such as the concentration of parent hydrocarbons, NOx, propene, temperature, and relative humidity, were used in the simulations. Simulation results include the amount of total product 1 and 2 (P1 and P2) produced from the amount of hydrocarbons reacted (ROG). For example, for 1,3,5trimethylbenzene, P1 and P2 represent the total amount of organonitrate product (TPM1) and other product (TPM2) produced, respectively. Chemical structures of TPM1 and TPM2 are shown after Table 3.8. Since P1/ROG and P2/ROG are known from the SAPRC chemistry, as a function of ROG, 1 and 2, as a function of conversion, can be calculated. The partitioning coefficients of the two products (Kom,1 and Kom,2) were then estimated by 1 * Kom,1 (1 + Kom,1 * Mo) + 2 * Kom,2 (1 + Kom 2 * Mo) = 1 ROG Equation 18 Kom,2 = Kom,1 * VP1 VP2 Equation 19 Where: VP1 = vapor pressure of product 1 (mm Hg) VP2 = vapor pressure of product 2 (mm Hg) 79 Kom,1 was determined by iteration In this calculation, Kom,1 was the only adjustable parameter, because Kom,2 was expressed as a function of Kom,1, VP1, and VP2 as shown in Equation (19). The vapor pressures of the condensable products from the individual hydrocarbon precursors were estimated by the MPBPWIN program (developed by Syracuse Research Corporation, 1996). MPBPWIN estimates vapor pressure based on three separate methods. The first is the Antoine method, which expressed vapor pressure as: ln Pvp = A - B T +C Equation 20 In which T is in Kelvins. The second method is the modified Grain method. This method is applicable to solids, liquids, and gases (Meylan, 1996). The third method employs a correlation suggested by Mackey, in which vapor pressure is given by ln P = -(4.4 + ln Tb)[1.803( Tb Tb Tm - 1) - 0.803 ln( ] - 6.8( - 1) T T) T Equation 21 Where Tb is the normal boiling point (K), T is the VP temperature (K), and Tm is the melting point (K). This expression is derived from aliphatic and aromatic classes and halogenated compounds (also aliphatic and aromatic classes) (Meylan, 1996). MPBPWIN gives all three vapor pressures from the three methods, along with a "suggested" VP. A suggested VP is the average of the Antoine and the modified Grain estimates for liquids and gases. For solids, the modified grain estimate is the 80 suggested VP. The vapor pressure of the target compounds can be predicted by providing the chemical structure into MPBPWIN program. The estimated partitioning coefficients and vapor pressures of condensable products from the reactions of 1,2,4- and 1,3,5-trimethylbenzene are reported in Table 3.8. Table 3.8: The Estimated Partitioning Coefficients and Vapor Pressures of Condensable Products from trimethylbenzenes Parent Hydrocarbon 1,2,4trimethybenzene 1,3,5trimethybenzene Kom,1 (m3/ g) 0.00609 0.00337 Kom,2 (m3/ g) 0.00046 0.00029 VP1 (mm Hg) 0.000113 0.00026 VP2 (mm Hg) 0.0015 0.0030 Temp. ( K) 297 302 Condensable products from oxidation of 1,3,5-trimethylbenzene Condensable products from oxidation of 1,2,4-trimethylbenzene Comparison of SOA yields for 1,2,4-trimethylbenzene 81 As described in the previous section, chamber data from studies performed by Odum (1996, 1997) and vapor pressure data were used in the estimating of partitioning coefficients (Kom,1 and Kom,2). Six different initial conditions (Odum, 1996) were employed to simulate SOA formation and to estimate partitioning coefficients of low-volatile products for 1,2,4-trimethybenzene. These conditions are shown in Table 3.9. The Estimated partitioning coefficients were reported in Table 3.8. Table 3.9: A set of simulation conditions for 1,2,4-trimethylbenzene, obtained from Odum's experiments Date 10/17B 11/02A 11/02B 11/07A 11/07B 11/09B ROGo ( g/m3) 2391 3367 1607 1932 1237 1745 ROG ( g/m3) 1996 2282 1198 1533 1020 1309 Mo ( g/m3) 113 155 43 78 26.5 53 NOx (ppb) 975 1178 490 590 359 528 C3H6 (ppb) 300 300 300 300 300 300 HC/NOx (ppb of C/ppb) 4.7 4.3 6.3 6.3 7.7 6.2 Y (%) 5.66 6.79 3.59 5.09 2.60 4.05 SOA yields for 1,2,4-trimethylbenzene from modified SAPRC simulation calculations using a partitioning model were then compared to those from Odum's chamber experiments. Figure 3.9 shows SOA yields against total aerosol mass concentration (Mo). 82 0.1 0.08 SOA yield 0.06 0.04 0.02 Best fit of this work Odum et al. ( 1996 ) chamber data Odum et al. ( 1996 ) model 0 0 50 100 M ( g/m ) 3 o 150 200 Figure 3.9: SOA yields for 1,2,4-trimethybenzene, NOx = 359-1178 ppb, C3H6 = 300 ppb, and HC/NOx = 4.3-7.7 ppbC/ppb 83 Figure 3.9 illustrates SOA yields from the simulations (labeled best fit of this work), from the observed SOA yields from Odum (labeled as "Odum et al. (1996) chamber data"), and from the prediction of SOA yields from empirical model of Odum (labeled as "Odum et al. (1996) model" in figure). The comparison shows good agreement between two sets of simulations and the observations. This suggests that the quantitative models of SOA formation used in this thesis, relying on only one empirical partitioning parameter are in good agreement with chamber experiments. 3.4.3 Sensitivity Analysis of SOA yields: Case study for 1,3,5trimethylbenzene The model developed in this work, based on a chemical reaction mechanism and a single partitioning parameter showed good agreement with Odum's chamber experiments for 1,2,4-trimethylbenzene. The model therefore can be used to investigate the sensitivity of SOA formation to parameters such as VOC/NOx ratio, hydrocarbon composition, and others. These sensitivity analyses of SOA formation were initially performed for 1,3,5-trimethylbenzene (estimates for approximately a dozen compounds will be presented in the next section). 1,3,5-trimethylbenzene was selected for initial examination because of the availability of smog chamber experiments that provided estimates for the branching ratio for organonitrate formation (and hence accurate values of 1 and 2) (Holes et al., 1997; Eusebi, 1996). The sensitivity of SOA yields on the ratio of VOCs to NOx, the base 84 hydrocarbon composition, and other parameters were examined for 1,3,5trimethybenzene in the presence of a typical urban mix of air pollutants using a box model. The box model used the full SAPRC mechanism. Initial conditions and composition of hydrocarbons used in the simulations are similar to those seen in the Houston area (Carter, 1998). Table 3.10 lists these conditions, while the hydrocarbon composition is displayed in Table B.1 (Appendix B). Table 3.10: Initial conditions for box simulation for 1,3,5-trimethybenzene in the presence of Houston air pollutants. Simulation time (hrs) 8 Reporting time (hr) 1 NO (ppm) 0.20806 NO2 (ppm) 0.01095 CH4 (ppm) 2.5606 NMOC (ppmC) 6.8367 CO (ppm) 2.1457 Temperature (K)* 312 * Temperature is a function of time; the value reported is at the beginning of the simulation. Temperature is based on ambient condition plus 10 degree to account for heating in chamber. SOA yield as a function of conversion and VOC/NOx Shown in Figures 3.10a and 3.10b are the yields of aerosol products as a function of 1,3,5-trimethybenzene conversion and time when VOC/NOx ranges from 4 to 31 (31 is the ratio for base case) ppmC/ppm. While concentration of nonmethane organic carbons (NMOC) was fixed, concentrations of NO2 and NO were changed to vary the VOC/NOx ratio between 4 and 31. In revising to the NOx concentration, the initial ratio of NO to NO2 was kept constant. Figure 3.10a 85 suggests that SOA yield (Y), which is the ratio of the amount of condensable products partitioning to particulate phase to the amount reacted of primary hydrocarbon (Mo/ROG), depends on percent 1,3,5-trimethybenzene conversions. The yield is a parabolic function of conversion. SOA yields at various VOC/NOx ratios presented against percent conversion are essentially collapsed onto a single line. In contrast, figure 3.10b shows that VOC/NOx ratio affects overall VOC reaction rates by effecting OH radical concentration. Therefore, when plotted versus reaction time, the SOA yield depends on VOC/NOx ratio. M = f (% primary VOCs conversion) 86 0.007 0.006 0.005 VOC/NOx ratio varies from 4 to 31 SOA yields 0.004 0.003 0.002 0.001 0 0 20 40 60 80 100 120 (a) 0.007 0.006 0.005 135-TMB conversion (%) VOC/NOx ratio varies from 4 to 31 SOA yields 0.004 0.003 0.002 0.001 0 10:00 (b) 12:00 14:00 16:00 18:00 Time of day ratio =16 ratio = 12 ratio=4 ratio=20 ratio=8 ratio=31 Figure 3.10: SOA yields for 1,3,5-trimethybenzene represented as %conversion and C.D.T, Houston conditions 87 This has implications in formulating SOA yield parameters for photochemical grid models, as will be described later in this thesis. SOA yields as a function of hydrocarbon composition To examine the effect of the base hydrocarbon composition on SOA yield, the composition of alkenes was reduced by 50% from the base case, and the reduced mass was added to alkanes. Then the box model simulation was run again and the SOA yield for 1,3,5-trimethylbenzene was compared to the SOA yield for the organic base hydrocarbon mixture. The calculation was repeated with 50% of the aromatics in the base mixture converted to alkanes. Shown in Figure 3.11 are SOA yields represented as a function of %conversion of 1,3,5-trimethybenzene compared between the organic base case hydrocarbon composition and the modified base hydrocarbons. There is no significant difference among results for all cases in Figure 3.11. These results suggest that SOA yield, expressed as a function of conversion, is independent of base hydrocarbon composition. 88 0.007 0.006 0.005 SOA yields 0.004 0.003 0.002 0.001 0 0 20 40 60 80 100 135-TMB Conversion (%) convert 50% of alkenes to alkanes base case convert 50% of aromatics to alkanes Figure 3.11: SOA yields for 1,3,5-trimethybenzene at different base hydrocarbon composition represented against %conversion and C.D.T., Houston conditions SOA yields as a function of rate parameters The sensitivity of SOA yield to the values of rate parameters was examined. The rate constant for the reaction of aerosol precursor for 1,3,5-trimethybenzene (kAPT1) was varied from 1.00E-9 to 1.00E-13 cm3/molc-sec. SOA yields are illustrated against percent 1,3,5-trimethybenzene reacted in Figure 3.12. As is evident from the Figure, SOA yield is rate parameter dependent. Increasing rate 89 constant of the reaction of aerosol precursor with OH enhances the extent of reaction. However, ultimately SOA yields for the OH rate constant of aerosol precursor converge at the same point for 100% conversion. At kAPT1 below 1.0E-12 cm3/molcsec the production of organic particulate products is barely detectable at most levels of conversion. 0.007 0.006 0.005 0.004 0.003 0.002 0.001 0 0 20 40 60 80 100 SOA yields 135-TMB Conversion (%) kAPT1 =1.0E-9 k APT1 =2.6E-11 k APT1 =1.0E-10 k APT1 =1.0E-11 k APT1 =1.0E-12 k APT1 =1.0E-13 Figure 3.12: SOA yields for 1,3,5-trimethybenzene for rate constant of aerosol precursor varies from 1.0E-9 to 1.0E-13 cm3/molc-sec, Houston conditions The behavior shown in Figure 3.12 is due to the series nature of the aerosol (TPM) formation chemistry. Hydrocarbon precursor reacts with OH radical to form aerosol 90 precursor (APT), then APT further reacts with OH radical to form semivolatile product species TPM. kTMB TMB APT kAPT TPM As the value of kAPT is reduced, APT accumulates as trimethylbenzene reacts, rather than forming aerosol species (TPM). SOA yields as a function of organic particulate mass The dependence of gas/particle partitioning on the organic particulate mass was examined for 1,3,5-trimethybenzene. The initial amount of aerosol mass that was used in the box model (Mint) was varied from 5 to 15 g/m3. SOA yields for both cases were compared as shown in Figure 3.13. The SOA yields for Mint equal 5 and 15 g/m3 are different by a factor of 2 for the build-up period of aerosol precursor concentrations (up to 40% conversion). Beyond 40% conversion, the differences gradually increase and then reach a constant factor of 3 after 50% conversion. The dependence of aerosol yield on Mint can vary depending on the reactivity of parent hydrocarbon and the aerosol precursor compounds. In this thesis, the relationship between the aerosol yield and Mint will be assumed to also depend on vapor pressure, which can be expressed by: Y = f(Mint, VP) Where VP is vapor pressure of condensable products. 91 0.02 0.015 SOA yields 0.01 int M = 15 ug/m 0.005 3 M = 5 ug/m int 3 0 0 20 40 60 80 100 135-TMB conversion (%) Figure 3.13: SOA yields for 1,3,5-trimethybenzene for organic particulate mass (Mint) equals 5 and 15 g/m3, Houston conditions SOA yields as a function of partitioning coefficient (Kom) Partitioning coefficients (Kom,1 and Kom,2) are employed to describe the physical partitioning of semi volatile products onto available organic medium (Mint). To examine how SOA yield is sensitive to the values of these parameters, Kom,1 was varied by factors of 2 and 3 from the original value. Because Kom,1 and Kom,2 are related through the vapor pressures of the compounds, changing Kom,1 will result in proportional changes in Kom,2. SOA yields represented as a function of 1,3,5trimethylbenzene conversion at the base case, double the base case, and triple the base case Kom are shown in Figure 3.14. Results reveal that at the early stage of 92 reaction (up to 40% conversion), the differences of SOA yields for three cases are small. Beyond 40% conversion the differences in SOA yields begin to increase in magnitude. At 60% conversion the yields at triple and double Kom are approximately 2.3 and 1.6 times higher than yield at original Kom. These results lead to the conclusion that the amount of condensable species partitioning onto the existing organic mass is sensitive to the values of partitioning coefficients (Kom,1 and Kom,2). 0.025 0.02 original K om double Kom triple Kom SOA yields 0.015 0.01 0.005 0 0 20 40 60 80 100 135-TMB conversion (%) Figure 3.14: SOA yields for 1,3,5-trimethylbenzene for the original, double, and triple values of Kom,1 and Kom,2. Simulation conditions are based on the CARB report (Carter, 1998). Correlation between SOA yield and chemical and physical parameters 93 Previous work (Odum et al, 1996; 1997) suggested that secondary organic aerosol formation is best described by a gas/particle partitioning model. This model is described in section 3.4.1. Odum et al. showed that SOA yields for individual hydrocarbon precursors are a function of Mo. Experimental data were fit to a 4 parameter model where the 4 parameters are Kom1, Kom2, 1, and 2. In Odum's work these parameters were determined empirically and were assumed to be constant. In contrast, the previous sections in this thesis have shown that while SOA yield is a function of Komi, the values of these variables have physical and chemical constraints and the i's are not constant but depend on extent of conversion. Thus, the Odum approach and the approach described in this work have the following basic differences Odum assumes aerosol yield depends only on Mo while this work assumes SOA yield depends on Mo and conversion of the parent hydrocarbon. The conversion dependence in this work is determined by the dependence of i on conversion. Odum uses four adjustable parameters to described experimental data while this work uses one adjustable parameter with other parameters determined from reaction branching ratios and relative vapor pressure. Comparison of the results of applying these two approaches was performed for m-xylene and n-propylbenzene. The amount of products partitioning into aerosol 94 phase (Mo) as a function of percent conversions of hydrocarbon were estimated for Odum's model with 4 parameters and for the modified SAPRC model with a single adjustable parameter (this work). Figure 3.15 illustrates the comparison. Figure 3.15a is the results of simulation using m-xylene as a representative of the low-yield and high reactivity hydrocarbons. It shows that Odum's model (blue) is more linear than compared with the modified SAPRC model (red). The modified SAPRC model does not predict yield of secondary organic aerosol until conversion of parent hydrocarbon proceeds to approximately 70%, and beyond that point, SOA yield dramatically increases. This represents the fact that loss of parent hydrocarbon is not directly related to condensable product formation, but is a consequence of both formation of aerosol precursor and condensable products. As a result Odum's model predicts more SOA yield at lower percentage of conversions and the modified SAPRC model predicts more yield at higher conversion. It also is noted that slope of SAPRC based model is significantly steeper at higher conversion. This might require extra caution in designing experiments to determine SOA yield. Estimates of yield may be affected significantly by the extent of reaction. Figure 3.15b is the results for n-propylbenzene, which represents species with high PM yield and low reactivity. In this plot, the difference between two models is less distinctive compared to the results for m-xylene. 95 Results for both m-xylene and n-propylbenzene at Mint equal 5 and 15 g/m3 show no significant differences. However when available aerosol mass concentration at the initial gets higher (more than 25 g/m3), the differences become more evident. This advocates that the condensation of low volatile products relate to initial aerosol mass concentration (Mint). 140 120 100 M ( g/m ) 3 Modified SAPRC w/single parameter Odum model w/4 parameters 80 60 40 20 0 0 20 40 60 80 100 conversions (%) (a) o 96 640 560 480 M ( g/m ) 400 320 240 160 80 0 0 20 40 60 80 100 120 3 o conversion (%) Modified SAPRC w/single parameter Odum model w/4 parameters (b ) Figure 3.15: versus % conversion of parent hydrocarbons from Odum model with 4 parameters and modified SAPRC model with single parameter Evidently SOA yields are associated to physical and chemical properties. To develop a quantitative relationship between the amount of condensable species partitioning to aerosol phase and these physical parameters, it is useful to start with a mass balance. Mass balance in aerosol phase: M tot = M int + massfractioni * M tot i Equation 22 Where Mtot = total amount of organic aerosol mass ( g/m3) Mint = amount of organic mass served as medium ( g/m3) 97 massfractioni = mass fraction of species i in aerosol phase Rearranging Equation 22 yields M tot = M int (1 - massfraction i ) i Equation 23 Mass Balance for species i in two phases: Pi = Ai + massfractioni*Mtot Where Pi = total amount of semivolatile particulate matter species i generated from reaction ( g/m3) Ai = mass concentration of semivolatile particulate matter species i in gas phase ( g/m3) At equilibrium: Komi = Fi,om/Ai*Mo massfrationi / Komi = Ai Substituting Ai into Equation 24 gives Pi = massfractioni /Komi + massfractioni * Mtot Equation 27 Equation 25 Equation 26 Equation 24 massfracti oni = Pi 1 / Komi + M tot Equation 28 Substituting Equation 28 into Equation 23 and rearranging yields: M tot - i M tot * Pi = M int 1 / Komi + M tot 98 Equation 29 From Equation 29 Mtot can be solved for numerically based on the known parameters (Kom,i, and Pi). In this work, equation 29 will be solved numerically using an equation solver in Microsoft excel. Equation 29 relates a total aerosol mass to an initial aerosol mass through an equilibrium parameter Kom, the total mass of semivolatile product Pi, and the volume of the gas phase. Equations 22-28 also hold for a change in Pi, Pi, and a change in aerosol mass Mi . This type of relationship, shown in Equation 30, will be particularly effective for use in 3D Eulerian air quality models. In these grid models, the goal will be to calculate a change in aerosol mass in a grid cell for each time step, based on the amount of semivolatile product produced during that time step. The expression will be of the form: Mi = (%DF)i * (Pi) Equation 30 Where %DFi = distribution factor (expressed as a percentage) of species i in the aerosol phase (and is a function of Mint or Mtot) Note that this form for Mi differs from the form used in most current grid models, where Mi = Yi * ROG Equation 31 In the traditional form, the aerosol yield depends on the amount of hydrocarbon precursor that has reacted. In this formulation, the yield of aerosol (Y) will depend on conversion, partitioning parameters and initial aerosol composition (i, Kom,i, and 99 Mo in the model of Odum (1996). In contrast, equation 30 separates the aerosol yield into a conversion dependent component (Pi) and a component dependent on Kom and Mint. This formulation is useful in a grid model because SOA yield (Mi/ROG) strongly depends on conversion (see Figure 3.11). In a grid cell, it is impossible to determine whether a given amount of ROG (e.g., amounting to 10% conversion) leads to a large amount of semivolatile product (e.g., conversion increasing from 90% to 100%) or to a small amount of product (e.g., conversion increasing from 0% to 10%) and it is very difficult to account for accumulated conversion as material advects between grid cells. The proposed formulation requires that the concentration of individual aerosol products be followed, but (equation 30) makes the physical partitioning independent of conversion, as demonstrated below. Demonstrating that %DFi depends only Mint and is independent of %conversion was done for 1,3,5-trimethylbenzene. Mint was changed from 5 to 15 and 25 g/m3, and %DFi at 25%, 50%, and 75% conversions were then calculated for each Mint level. Shown in Figures 14a and b are %DF and M for condensable products formed from reaction of 1,3,5-trimethylbenzene with OH radicals against Mint at 25%, 50%, and 75% conversions, respectively. Figure 3.16 shows that %DFi depends on Mint. The values of %DFi are independent of % conversion. The polynomial best fit of these results yields the empirical function of %DFi and Mint for 1,3,5-trimethylbenzene as expressed below. 100 %DFi = -0.0005(Mint)2 + 0.13(Mint) 0.0375 Equation 32 Therefore at any given Mint, %DFi can be estimated for 1,3,5-trimethylbenzene, and consequently M can be predicted by using Equation 30. 101 3 2.5 2 % DF i 1.5 1 0.5 0 5 10 15 M ( g/m ) int 3 20 25 30 0.6 0.5 0.4 M ( g/m ) 3 0.3 0.2 0.1 0 0 5 10 15 M ( g/m ) int 3 o 20 25 30 25% conversion 50% conversion 75% conversion Figure 3.16: %DF and M as a function of Mint at 25%, 50%, and 75% 1,3,5-trimethylbeznene conversions. Simulation conditions were obtained from the CARB report (Carter, 1998) 102 3.5 Quantitative models of SOA formation for various aromatic hydrocarbons Section 3.4 described how Mi could be calculated for a single hydrocarbon and showed that a function of the form: Mi = %DFi * Pi could be developed where %DFi depended only on Kom and Mint. However, because Kom,i is compound specific, the expression for %DFi will be compound specific. Section 3.3.2 identified aromatic precursors to SOA formation, and 15 of these aromatic hydrocarbons were selected for examination in this work. This section will give an example of the development of a model for o-xylene. The analogous procedures were reapplied in developing mechanistic models for 1,2,3trimethylbenzene, m-xylene, p-xylene, m-ethyltoluene, p-ethyltoluene, oethyltoluene, ethylbenzene, benzene, n-propylbenzene, iso-propylbenzene, secbutylbenzene, and toluene. Condensed mechanisms and kinetic parameters for these aromatic hydrocarbons are also presented in this section. SOA yields from simulations were compared against experimental values (Odum, 1997). 103 3.5.1 Quantitative mechanistic models for aromatic hydrocarbons Chemistry for o-Xylene In addition to trimethybenzene, xylene also produces a significant amount of secondary organic aerosol in the urban atmosphere. Reaction pathways of o-xylene with OH radicals are shown in Figures 3.17-3.19. (5) Figure 3.17: Hydrogen abstraction from o-xylene by OH radical reaction pathway (6) Figure 3.18: Addition of OH to o-xylene forming methylcresol 104 (7) Figure 3.19: OH addition to o-xylene reaction (8) Figure 3.20: semi-volatile product reaction pathways from aerosol precursors The major daytime loss process for atmospheric aromatic hydrocarbons is through reaction with OH radicals. As with 1,3,5-trimethybenzene, this reaction can proceed through either the hydrogen abstraction (pathway 5), or addition to the aromatic ring (pathways 6 and 7). For o-xylene, hydrogen abstraction accounts for 105 4.5% of all reactions (SAPRC99), occurring primarily at the substituent methyl groups. The product of this reaction is benzaldehyde with one extra carbon atom (BALD + 1C), along with an NO to NO2 conversion and the production of HO2 radical. 4%: O-XYLENE + HO. = BALD +1C + RO2R Reaction 16 Addition of the OH radical will occur at one of the three ring positions, but predominantly adds at the aromatic carbon ortho to one of the carbons attached to the methyl group (Grovenstein, 1970; Atkinson et al., 1980). The initial addition of OH radical to o-xylene results in cresol with one extra carbon atom (CRES + 1C) and HO2 radical. This pathway accounts for 16.1% of all reactions of o-xylene (SAPRC99). 16%: O-XYLENE + HO. = CRES + 1C + HO2. Reaction 17 The remainder of the OH radical addition products from o-xylene will further react with O2, again with conversion of NO to NO2. The ring cleavage reaction of the O2OH-aromatic-adduct proceeds through four possible positions on the aromatic ring. Figure 3.19 illustrates the reaction pathway leading to formation of aerosol precursor (APX2). Although chemical structures of all aerosol precursors from o-xylene, APX1-APX3 are different, their formation pathways are similar. The cleavage reactions produce 45% of APX1, 14% of APX2, and 20% of APX3 (Bandow, 1985). 106 In addition to these, MGLY, GLY, and BACL are generated. In total, these pathways account for 79.4% of all reactions of o-xylene with OH radical. 79%: O-XYLENE + HO. = .569 APX1 + .569 MGLY + .17 APX2 + .17 GLY + Reaction 18 .253 APX3 + .253 BACL Net: O-XYLENE + HO. = .045 BALD + .161 CRES + .835 RO2R + .161 HO2. + .45 APX1 + .14APX2 + .2 APX3 + .2 BACL + .14 GLY + .45 MGLY + .206 C Reaction 19 Aerosol precursors (APX1-APX3) generate two types of aerosol products (XPM1 and XPM2) via the OH-addition at the carbon double bond. Results from previous studies suggested that 25% of APX2-OH addition reaction produce XPM1 (Eusebi, 1996). The remaining APX2-OH reacts with O2 and NO to form XPM2, NO2, and HO2 radicals (Figure 3.20). APX1-APX3 + HO. = .25 XPM1 + .75 XPM2 +.25 RO2N +.75 RO2R Reaction 20 By applying the same procedure used in developing mechanisms for 1,3,5trimethybenzene and o-xylene, condensed mechanisms for the individual aromatic precursors were developed. Reaction pathways of the photooxidation reactions and detailed mechanisms for each species are given in Appendix C. Condensed mechanisms of SOA formation and rate constants, which were estimated from the AOP program, for these 15 compounds are presented in Table 3.11. 107 Table 3.11: Condensed mechanisms of SOA formation for 15 aromatic precursors developed in this study Aromatic Precursors 1,3,5Trimethybenzene APT1 1,2,3Trimethybenzene Aerosol Precursors Condensed Mechanisms 135-TMB + HO. = 0.03BALD + 0.18CRES + 0.81RO2R + 0.18HO2. + 0.79MGLY + 0.79APT1+ .42XC APT1 + HO. = 0.25TPM1 + 0.75TPM2 + 0.25RO2N + 0.75RO2R 123-TMB + HO. = 0.044BALD + 0.186 CRES + 0.809RO2R + 0.186HO2. + 0.254MGLY + 0.085GLY + 0.431BACL+ 0.431APT2+ 0.254APT3 + 0.085APT4 + 0.46XC APT2 + HO. = 0.25TPM3 + 0.75TPM4 + 0.25RO2N + 0.75RO2R APT3 + HO. = 0.25TPM3 + 0.75TPM4 + 0.25RO2N + 0.75RO2R APT4 + HO. = 0.25TPM3 + 0.75TPM4 + 0.25RO2N + 0.75RO2R 124TMB + HO. = 0.044BALD + 0.186CRES + 0.812RO2R + 0.186HO2. + 0.469MGLY + 0.192GLY + 0.107BACL + 0.254APT5 + 0.084APT6 + 0.131 APT7 + 0.146APT8 + 0.046APT9 + 0.107AP10 + 0.46XC APT5 + HO. = 0.25TPM5 + 0.75TPM6 + 0.25RO2N + 0.75RO2R APT6 + HO. = 0.25TPM5 + 0.75TPM6 + 0.25RO2N + 0.75RO2R APT7 + HO. = 0.25TPM5 + 0.75TPM6 + 0.25RO2N + 0.75RO2R APT8 + HO. = 0.25TPM5 + 0.75TPM6 + 0.25RO2N + 0.75RO2R Rate Constants (cm3/molc-sec) 5.79e-11 Source of rate constants SAPRC-99 2.60e-11 3.27e-11 AOP program SAPRC-99 APT2 APT3 APT4 1,2,4Trimethybenzene 1.85e-11 2.66e-11 3.36e-11 3.35e-11 AOP program AOP program AOP program SAPRC-99 APT5 APT6 APT7 APT8 4.80e-11 2.60e-11 2.60e-11 3.30e-11 AOP program AOP program AOP program AOP program 108 Aromatic Precursors Aerosol Precursors APT9 AP10 Condensed Mechanisms APT9 + HO. = 0.25TPM5 + 0.75TPM6 + 0.25RO2N + 0.75RO2R AP10 + HO. = 0.25TPM5 + 0.75TPM6 + 0.25RO2N + 0.75RO2R O-XYLENE + HO. = 0.045BALD + 0.161CRES + 0.835RO2R + 0.161HO2. + 0.45MGLY + 0.146GLY + 0.45APX1 + 0.14APX2 + 0.2APX3 + 0.206XC + 0.2BACL APX1 + HO. = 0.25XPM1 + 0.75XPM2 + 0.25RO2N + 0.75RO2R APX2 + HO. = 0.25XPM1 + 0.75XPM2 + 0.25RO2N + 0.75RO2R APX3 + HO. = 0.25XPM1 + 0.75XPM2 + 0.25RO2N + 0.75RO2R P-XYLENE + HO. = 0.083BALD +0.188CRES + 0.812RO2R + 0.188HO2. + 0.24MGLY + 0.489GLY + 0.489APX4 + 0.24APX5 + 0.271XC APX4 + HO. = 0.25XPM3 + 0.75XPM4 + 0.25RO2N + 0.75RO2R APX5 + HO. = 0.25XPM3 + 0.75XPM4 + 0.25RO2N + 0.75RO2R M-XYLENE + HO. = 0.037BALD + 0.21CRES + 0.789RO2R + 0.21HO2. + 0.564 MGLY +0.188GLY + 0.376APX6 + 0.188APX7 + 0.188APX8 + 0.247XC APX6 + HO. = 0.25XPM5 +0.75XPM7 + 0.25RO2N +0.75RO2R APX7 + HO. = 0.25XPM5 +0.75XPM7 + 0.25RO2N +0.75RO2R 109 Rate Constants (cm3/molc-sec) 1.20e-11 1.00e-11 1.37e-11 Source of rate constants AOP program AOP program SAPRC-99 o-Xylene APX1 APX2 APX3 p-Xylene 1.85e-11 2.66e-11 6.96e-12 1.43e-11 AOP program AOP program AOP program SAPRC-99 APX4 APX5 m-Xylene 4.88e-11 1.00e-11 1.37e-11 AOP program AOP program SAPRC-99 APX6 1.80e-11 AOP program AOP program APX7 2.60e-11 Aromatic Precursors Aerosol Precursors APX8 Condensed Mechanisms APX8 + HO. = 0.25XPM5 +0.75XPM7 + 0.25RO2N +0.75RO2R P-C2-TOL + HO. = 0.49APET1 + 0.2APET2 + 0.04APET3 + 0.24MGLY + 0.49GLY + 0.188HO2. + 0.188CRES + 0.083BALD + 0.856RO2R + 0.702XC + 0.042CO2 + 0.084H2O 0.042HO. APET1 + HO. = 0.25ETPM1 + 0.75ETPM2 + 0.25RO2N + 0.75RO2R APET2 + HO. = 0.25ETPM1 + 0.75ETPM2 + 0.25RO2N + 0.75RO2R APET3 + HO. = 0.25ETPM1 + 0.75ETPM2 + 0.25RO2N + 0.75RO2R O-C2-TOL + HO. = 0.329APET4 + 0.085APET5 + 0.026APET6 + 0.051APET7 + 0.186APET8 + 0.395MGLY + 0.162GLY+0.186BACL + 0.188HO2. + 0.188CRES + 0.054BALD + 0.8404RO2R + 0.7255XC + 0.0659APET9 +0.0325CO2+0.065H2O 0.0325HO. APET4 + HO. = 0.25ETPM3 + 0.75ETPM4 + 0.25RO2N + 0.75RO2R APET5 + HO. = 0.25ETPM3 + 0.75ETPM4 + 0.25RO2N + 0.75RO2R APET6 + HO. = 0.25ETPM3 + 0.75ETPM4 + 0.25RO2N + 0.75RO2R APET7 + HO. = 0.25ETPM3 + 0.75ETPM4 + 0.25RO2N + 0.75RO2R APET8 + HO. = 0.25ETPM3 + 0.75ETPM4 + 0.25RO2N + 0.75RO2R 110 Rate Constants (cm3/molc-sec) 1.05e-11 Source of rate constants AOP program p-ethyltoluene 1.43E-11 SAPRC-99 APET1 APET2 APET3 o-ethyltoluene 4.50e-11 1.00e-11 1.00e-11 1.37e-11 AOP program AOP program AOP program SAPRC-99 APET4 APET5 APET6 APET7 APET8 1.88e-11 1.00e-11 2.70e-11 2.70e-11 6.90e-12 AOP program AOP program AOP program AOP program AOP program Aromatic Precursors Aerosol Precursors APET9 Condensed Mechanisms APET9 + HO. = 0.25ETPM3 + 0.75ETPM4 + 0.25RO2N + 0.75RO2R M-C2-TOL + HO. = 0.054APET10 + 0.134APET11 + 0.08APET12 + 0.106APET13 + 0.226APET14 +0.567MGLY + 0.186GLY + 0.21HO2. + 0.21CRES + 0.037BALD + 0.8085RO2R + 0.6825XC + 0.153APET15+0.0185CO2 + 0.037H2O 0.0185HO. APET10 + HO. = 0.25ETPM5 + 0.75ETPM6 + 0.25RO2N + 0.75RO2R APET11 + HO. = 0.25ETPM5 + 0.75ETPM6 + 0.25RO2N + 0.75RO2R APET12 + HO. = 0.25ETPM5 + 0.75ETPM6 + 0.25RO2N + 0.75RO2R APET13 + HO. = 0.25ETPM5 + 0.75ETPM6 + 0.25RO2N + 0.75RO2R APET14 + HO. = 0.25ETPM5 + 0.75ETPM6 + 0.25RO2N + 0.75RO2R APET15 + HO. = 0.25ETPM5 + 0.75ETPM6 + 0.25RO2N + 0.75RO2R TOLUENE + HO. = 0.234CRES + 0.515RO2R + 0.234HO2. + 0.167MGLY + 0.238GLY + 0.238APTO1 + 0.167APTO2 + 0.11BALD + 0.25NBEN + 0.25H2O + 0.25NO2 + 0.25XC APTO1 + HO. = 0.25TOPM1 +0.75TOPM2 + 0.25RO2N + 0.75RO2R APTO2 + HO. = 0.25TOPM1 +0.75TOPM2 + 0.25RO2N + 0.75RO2R Rate Constants (cm3/molc-sec) 1.80e-11 1.37e-11 Source of rate constants AOP program SAPRC-99 m-ethyltoluene APET10 APET11 APET12 APET13 APET14 APET15 Toluene 1.00e-11 1.00e-11 2.70e-11 2.70e-11 1.80e-11 1.80e-11 5.95e-12 AOP program AOP program AOP program AOP program AOP program AOP program SAPRC-99 APTO1 1.85e-11 AOP program APTO2 6.96e-12 AOP program 111 Aromatic Precursors Benzene Aerosol Precursors Condensed Mechanisms BENZENE + HO. = 0.236PHEN + 0.236HO2. + 0.566NBEN + 0.2RO2R + 0.566H2O + -0.566NO2 + 0.2APB1 + 0.2GLY APB1 + HO. = 0.25BPM1 + 0.75BPM2 + 0.25RO2N + 0.75RO2R C2-BEN + HO. = 0.19CRES + 0.574RO2R + 0.19HO2. + 0.167MGLY + 0.237GLY + 0.167APEB1 + 0.158APEB2 + 0.079APEB3 + 0.999XC + 0.296NBEN + 0.416H2O + 0.03CO2 + -0.06HO. + 0.296NO2 + 0.11BALD +0.03OCH2 APEB1 + HO. = 0.25EBPM1 + 0.75EBPM2 + 0.25RO2N + 0.75RO2R APEB2 + HO. = 0.25EBPM1 + 0.75EBPM2 + 0.25RO2N + 0.75RO2R APEB3 + HO. = 0.25EBPM1 + 0.75EBPM2 + 0.25RO2N + 0.75RO2R N-C3-BENZ + HO. = 0.19CRES + 0.634RO2R + 0.19HO2. + 0.167MGLY + 0.238GLY + 0.296NBEN + 0.167APPB1 + 0.158APPB2 + 0.079APPB3 + 1.702XC + 0.476H2O + 0.12CO2 + 0.12HO. + -0.296NO2 + 0.11BALD + 0.12OCH2 APPB1 + HO. = 0.25PBM1 + 0.75PBM2 + 0.25RO2N +0.75RO2R APPB2+ HO. = 0.25PBM1 + 0.75PBM2 + 0.25RO2N +0.75RO2R APPB3 + HO. = 0.25PBM1 + 0.75PBM2 + 0.25RO2N +0.75RO2R Rate Constants (cm3/molc-sec) 1.23e-12 Source of rate constants SAPRC-99 APB1 6.96e-12 AOP program SAPRC-99 Ethyl Benzene 7.10e-12 APEN1 APEB2 APEB3 n-Propyl Benzene 6.96e-12 1.85e-11 1.00e-11 6.00e-12 AOP program AOP program AOP program SAPRC-99 APPB1 APPB2 APPB3 6.96e-12 1.85e-11 1.85e-11 AOP program AOP program AOP program 112 Aromatic Precursors Isopropyl Benzene Aerosol Precursors Condensed Mechanisms I-C3-BEN + HO. = 0.11BALD + -0.11HO. + 0.19CRES +0.19HO2. + 0.167APB2 + 0.167MGLY + 0.158APB3 + 0.079APB4 + 0.237MGLY + 0.624RO2R +0.296NO2 + 0.296NBEN + 0.516 H2O + 1.712XC + 0.11CO2 APB2 + HO. = 0.25BPM3 + 0.75BPM4 + 0.25RO2N +0.75RO2R APB3 + HO. = 0.25BPM3 +0.75BPM4 +0.25RO2N +0.75RO2R APB4 + HO. = 0.25BPM3 + 0.75BPM4 + 0.25RO2N +0.75RO2R S-C4-BEN + HO. = 0.11BALD + -0.16HO. + 0.19CRES + 0.19HO2. + 0.167APB5 + 0.167MGLY + 0.119APB6 + 0.119APB7 +0.238GLY + 0.675RO2R + 0.296NO2 + 0.296NBEN + 0.566H2O + 2.425XC + 0.16CO2 APB5 + HO. = 0.25BPM5 +0.75BPM6 +0.25RO2N + 0.75RO2R APB5 + HO. = 0.25BPM5 +0.75BPM6 +0.25RO2N + 0.75RO2R APB5 + HO. = 0.25BPM5 +0.75BPM6 +0.25RO2N + 0.75RO2R Rate Constants (cm3/molc-sec) 6.50e-12 Source of rate constants SAPRC-99 APB2 APB3 APB4 s-Butyl Benzene 6.96e-12 1.85e-11 1.00e-11 6.00e-12 AOP program AOP program AOP program SAPRC-99 APB5 APB6 APB7 6.96e-12 1.85e-11 1.00e-11 AOP program AOP program AOP program After the chemistry of SOA formation for these aromatic hydrocarbons were implemented into SAPRC-99, results from simulations were used to estimate the partitioning of condensable products. Quantitative mechanistic models of SOA formation were developed for 15 aromatics, but the partitioning coefficients were 113 estimated for only 11, due to the lack of information from chamber experiments for the remaining 4: benzene, sec-buthylbenzene. Iso-propylbenzene, and 1,2,3trimethylbenzene. The prediction of SOA formation for these 11 compounds were done by employing partitioning coefficients for individual species (as previously elaborated for example of 1,3,5-trimethylbenzene). Partitioning coefficients for individual compounds listed in Table 3.12 were estimated based on results for SAPRC simulations and Odum's experimental results. Table 3.12: The Estimated Partitioning Coefficients and Vapor Pressures of Condensable Products from aromatic hydrocarbons Parent Hydrocarbons 1,3,5-trimethybenzene 1,2,4-trimethybenzene o-xylene p-xylene m-xylene o-ethyltoluene m-ethyltoluene p-ethyltoluene toluene n-propylbenzene ethylbenzene Kom,1 (m3/ g) Kom,2 (m3/ g) VP1 (mm Hg) VP2 (mm Hg) 0.00337 0.00609 0.00513 0.00451 0.00203 0.00844 0.00490 0.00690 0.01230 0.07220 0.02600 0.00029 0.00046 0.000412 0.000394 0.000175 0.000735 0.000433 0.000569 0.001090 0.001980 0.002310 0.00026 0.000113 0.00097 0.00158 0.00066 0.00049 0.00038 0.00045 0.00066 0.000096 0.00039 0.0030 0.0015 0.0121 0.0181 0.0076 0.0056 0.0043 0.0054 0.0074 0.0035 0.0044 114 3.5.2 Comparison of SOA yields In addition to 1,2,4- and 1,3,5-trimethylbenzene, the yields of SOA for 9 aromatic compounds were examined individually. The conditions used in these simulations were identical to those used by Odum, 1997. These conditions are reported in Table D.1, Appendix D. Aromatic precursors were classified into two categories: high-yield aromatics, and low-yield aromatics (Odum, 1997). The highyield aromatic species are those species containing one or fewer methyl substituent and one or fewer ethyl substituent (i.e., toluene, ethylbenzene, and ethyltoluene). The low-yield aromatics are those species that contain two or more methyl substituents (i.e., xylenes, trimethylbenzenes). Figure 3.21 demonstrates SOA yields of these aromatic species as a function of Mo (or Mtot). While there is disagreement between some data points from simulations and observations, the overall trend shows good correspondence in the results of the quantitative models and chamber experiments. 115 0.2 0.15 SOA yield 0.1 0.05 0 0 100 200 o 300 3 400 500 M ( g/m ) low-yiled aromatics, Odum low-yield aromatics, simulation high-yield aromatics, Odum high-yield aromatics, simulation Figure 3.21: SOA yields for high- and low-yield aromatics from simulation calculation compared to chamber experiments Observed and predicted aerosol yields presented in Figure 3.21 for low- and highyield aromatics are reported in Table 3.13. 116 Table 3.13: Aerosol yields for low- and high-yield aromatics: results from simulations and observation Chamber experiments by Odum (1997) Compounds 135-TMB m-xylene m-xylene m-xylene o-xylene o-xylene p-xylene p-xylene Toluene Toluene Toluene toluene o-ethyltoluene p-ethyltoluene m-ethyltoluene m-ethyltoluene m-ethyltoluene ethylbenzene ethylbenzene ethylbenzene ethylbenzene n-propylbenzene n-propylbenzene n-propylbenzene ROG 1029 1891 1571 1528 1117 1082 823 1063 1413 1268 1710 923 789 708 1927 971 334 434 3176 1872 1169 1314 657 1790 M 31 106 46 48 35 23 16 32 133 111 171 68 49 38 208 66 13 13 394 185 104 103 39 190 Y 0.030 0.056 0.029 0.031 0.031 0.021 0.019 0.030 0.094 0.088 0.100 0.074 0.062 0.054 0.108 0.068 0.039 0.030 0.124 0.099 0.089 0.078 0.059 0.106 ROG 1030 1870 1450 1530 1170 1170 865 1090 1440 1300 1700 960 833 751 1780 999 356 465 3060 1870 1260 1320 781 1780 SAPRC simulations Simulating time (hrs) M Y 31 0.030 2.0 124 0.066 5.0 46 0.032 10.0 31 0.020 10.0 100 0.085 3.0 9 0.008 2.5 6 0.007 2.5 81 0.074 3.0 149 0.103 7.0 99 0.076 7.0 252 0.148 8.0 NA NA 3.0 49 0.059 7.0 38 0.051 7.0 318 0.179 10.0 99 0.099 6.0 NA NA 4.0 NA NA 2.5 419 0.136 9.0 230 0.122 8.0 70 0.055 3.0 76 0.058 6.0 54 0.069 4.0 133 0.075 6.0 Results in the above table show aerosol mass change and SOA yields from the SAPRC simulation and observations (Odum, 1996). The simulations were conducted under similar conditions as in Odum's chamber experiments, and continue until the amount of ROG is relatively equal to ROG from experiment was achieved (e.g., 135-TMB requires 2 hours of simulating time to the reach optimal ROG similar to Odum's experiment. The total amount of semivolatile products generated from 117 reaction (Pi), obtained from simulation, and M from Odum results were used in the calculation to estimate partitioning coefficients. The estimated partitioning coefficients were then used to determine the estimated aerosol mass change. The estimated mass changes were compared with those from Odum observation. As seen from Figure 3.21 and Table 3.13, SOA yields from the estimation are comparable within the factor of 0.5 to 2.0 to those from the experimental results, 19% of the results from the simulation are outside the range (i.e., results for o- and p-xylene). The relatively poor agreement for these 4 runs of o- and p-xylene results from the relatively small aerosol mass changes. 3.5.3 Correlation between M and Mint for aromatic hydrocarbons The motivation and procedure for establishing the correlation between M and Mint was described in Section 3.4.3 (for 1,3,5-trimethylbezene). With the same purpose and by employing analogous procedures, the correlation between M and Mint in terms of %DFi for the remaining aromatics was developed. The results are shown below. Figures 3.22a-j demonstrate the relationship between %DFi and Mint and as a function of hydrocarbon conversion for 1,2,4-trimethylbenzene, ethylbenzene, m-ethyltoluene, o-ethyltoluene, p-ethyltoluene, m-xylene, o-xylene, pxylene, n-propylbenzene, and toluene, respectively. 118 6 124-TMB 5 y=-0.0045x^2 + 0.34x - 0.5875 4 %DF i 3 2 1 25% conversion 50% conversion 75% conversion 4 6 8 10 3 int 0 12 14 16 M ( g/m ) (a) 20 ethylbenzene 16 y= - 0.0135x^2 + x +1.3375 12 %DF 8 4 i 25% conversion 50% conversion 75% conversion 0 4 6 8 10 12 3 14 16 M ( g/m ) int (b) 119 6 m-ethyltoluene 5 y = -0.001x^2 + 0.2x + 0.025 4 %DF i 3 2 1 25% conversion 50% conversion 75% conversion 4 6 8 10 12 3 0 14 16 M ( g/m ) int (c) 8 o-ethyltoluene 7 6 y = -0.0015x^2 + 0.3x + 0.1375 %DF 5 4 3 2 1 4 6 8 int i 25% conversion 50% conversion 75% conversion 10 12 3 14 16 M ( g/m ) (d) 120 6 p-ethyltoluene 5 y = -0.002x^2 + 0.26x - 0.05 4 %DF 3 2 i 25% conversion 50% conversion 75% conversion 4 6 8 10 12 3 1 14 16 M ( g/m ) int (e) 2.5 m-xylene 2 y = 1E-17x^2 + 0.07x + 0.05 1.5 %DF 1 0.5 i 25% conversion 50% conversion 75% conversion 4 6 8 10 12 3 0 14 16 M ( g/m ) int (f) 121 4.5 4 3.5 3 %DF i o-xylene y = - 0.001x^2 + 0.19x - 0.025 2.5 2 1.5 1 0.5 4 6 8 int 25% conversion 50% conversion 75% conversion 10 12 3 14 16 M ( g/m ) (g) 4.5 p-xylene 4 3.5 y = -0.003x^2 + 0.24x - 0.225 3 %DF i 2.5 2 1.5 1 0.5 4 6 8 10 12 3 25% conversion 50% conversion 75% conversion 14 16 M int ( g/m ) (h) 122 22 n-propylbenzene 20 18 y = -0.013x^2 + 0.9x +6.625 %DF i 16 14 12 10 8 4 6 8 10 12 3 25% conversion 50% conversion 75% conversion 14 16 M ( g/m ) int (i) 12 toluene 10 y = -0.0115x^2 + 0.69x - 0.2625 8 %DF 6 4 i 25% conversion 50% conversion 75% conversion 2 4 6 8 10 12 3 14 16 M int ( g/m ) (j) Figure 3.22: %DFi as a function of Mint at 25, 50, and 75% hydrocarbon conversions for 11 aromatics 123 Results for most of species reveal the same behavior as seen for 1,3,5trimethylbenzene, in which %DFi is strongly dependent on Mint but not influenced by %conversion (i.e., %DFi = f(Mint) f(%conversion)). The exceptions are toluene, m-xylene and m-ethyltoluene. For these three compounds, the percentage of product that partitions into the aerosol phase (%DFi ) is similar at 50 and 75% conversions, , but is different for 25% conversion. The deviation occurs because Mi for these 3 species is comparable to Mtot. As the partitioning of condensable species onto Mint goes on, total organic mass increases, and this increase can be a substantial fraction of the total aerosol mass, especially at low values of Mint. This creates a dependence of %DFi on conversion especially when (M > 0.6Mint). This is shown in Figure 3.23, M for toluene is plotted against total amount of semivolatile products (Pi) produced from reaction of OH radical with toluene. As seen from the figure, below the upper limit (M = 0.6Mint), amount of semivolatile products partitioning to aerosol phase occurs at a relatively constant rate. Above this point, aerosol mass change (M) substantially increases so that its relationship with Pi changes from linear to parabolic function. For relatively low volatility compounds (e.g., products from toluene reaction) at the high level of conversion (such as 75% conversion), the partitioning occurs rapidly so that M is comparable with Mint, and the partitioning rate is not constant anymore. Fortunately, 124 for most aromatics, and in most air quality modeling applications, M is only a small fraction of Mint. Figure 3.23: M for toluene presented against Pi Table 3.14 summarizes the results of the evaluation of %DFi , as a function of Mint. %DFi = a*Mint2 + b*Mint + c Equation 33 Equation 34 M = %DFi * Pi 125 Table 3.14: Coefficients a, b, and c for the correlation equation between %DFi and Mint for 11 aromatic hydrocarbons initial HCs @ 1998 CARB report Compounds a b c 1,2,4-trimethylbenzene -0.005 0.340 -0.588 1,3,5-trimethylbenzene -0.001 0.130 -0.038 ethylbenzene -0.014 1.000 1.338 m-ethyltoluene -0.001 0.200 0.025 m-xylene 0.000 0.070 0.050 n-propylbenzene -0.013 0.900 6.625 o-ethyltoluene -0.002 0.300 0.138 o-xylene -0.001 0.190 -0.025 p-ethyltoluene -0.002 0.260 -0.050 p-xylene -0.003 0.240 -0.225 toluene -0.012 0.690 -0.263 Note: These coefficients were estimated based on the Houston conditions obtained from the CARB Report (Carter, 1998), which represent a typical atmospheric environment. 3.6 Summary Quantitative models of SOA formation were developed for 15 aromatic hydrocarbons, but the partitioning coefficients were determined individually for only 11 compounds due to the lack of information from chamber experiments for 4 aromatics: iso-propylbenzene, sec-butylbenzene, benzene, and 1,2,3triemethybenzene. Thus full SOA models were performed for only 11 compounds, though the chemical mechanisms of SOA formation for 15 compounds were implemented into SAPRC. Data available from previous smog chamber experiments were employed in the development. These models were then implemented into a comprehensive gas-phase chemistry which includes reactions of over 350 ambient 126 species (SAPRC99). A box model was set up to examine SOA yields for individual aromatic hydrocarbon precursors. The gas/particle partitioning absorption mechanism from previous studies was applied with results from the simulations to estimate SOA yields and model parameters. All model parameters estimated in this study are presented in Table 3.15. Table 3.15: Chemical and physical model parameters for the estimate of secondary aerosol formation Parameters Chemistry kOH OH radical rate constant SAPRC-99, Literature review, and structure activity relationship ai Physics Kom,i Partitioning coefficients of compound i Estimated from chamber data i Proportionality constant relating the amount of ROG reacted to the total concentration of product i Estimated from SAPRC rate parameters and branching ratios from chamber data Pi ROG VPi M Mtot Total concentration of product i Concentration of ROG that reacts Vapor pressure of product I Amount of secondary aerosol produced and portioning into particulate phase Total amount of secondary aerosol in particulate phase SAPRC simulation based on ambient Houston data SAPRC simulation based on ambient Houston data Antoine method and modified Grain method SAPRC simulation calculation SAPRC simulation calculation Stoichiometric coefficient of species i Estimated from chamber data Descriptions Calculation Procedures 127 SOA yields from the simulation calculations were compared with the observed yields from chamber experiments performed by Odum. The comparison shows a good agreement of SOA formation between the quantitative models and chamber experiments. After testing the reliability of these SOA formation models, the sensitivity analysis was conducted for 1,3,5-trimethylbenzene to investigate which parameters affect the formation and partitioning of SOA. Results from the analysis leading to the correlation between the amount of secondary aerosol produced and partitioning into particulate phase (M) and chemical and physical parameters as expressed below: M = %DFi * Pi Where %DFi = aMint2 + bMint + c Coefficients a, b, and c are compound specific. In conclusion, SOA formation depends on three parameters (%conversion, Mint, and Komi). Parameters affecting the formation and partitioning of condensable products are compound specific. 128 Chapter 4 Quantitative models of SOA formation for lumped aromatic species development Chapter 3 presented the development of detailed chemical and physical models for secondary organic aerosol (SOA) formation for individual aromatic species. Use of these models requires a detailed gas phase photochemical model coupled with a compound specific phase partitioning model (SAPRC). Recognizing that in many applications more empirical approaches will be required, this chapter develops two additional modeling approaches: an incremental aerosol reactivity and a lumped photochemical mechanism. 4.1 Analysis of incremental aerosol reactivity As discussed in Chapter 2, the concept of incremental reactivity has been used in modeling ozone formation to provide an assessment of the relative ozone formation potential for hydrocarbons. The same concept can be applied to particulate matter formation. This section will give the results of the analysis of incremental aerosol reactivity (IAR). A first step in this analysis is to define an IAR. The IAR to be used in this work is defined in Equation 34. IAR = ( PMproduced ) ( HCadded ) Equation 35 Units of both nominator and denominator are in g/m3. 129 In this work, the IAR was calculated by increasing the concentration of a single hydrocarbon, such as 1,3,5-trimethylbenzene, from the base case by an amount equivalent to 5% (based on mole basis) of total hydrocarbon concentration. Carter (1994) found that incremental ozone reactivities depended on the base hydrocarbon mixture, the VOC/NOx ratio, and other parameters. The results presented in Chapter 3 suggest that IAR, as defined in Equation 34, will depend on the base hydrocarbon mixture, the VOC/NOx ratio, and the initial aerosol concentration. If however, a modified definition of IAR is used IAR = PMproduced (assuming 50% conversion of HC) HCadded Equation 36 then, as shown in Chapter 3, IAR will depend only on the initial aerosol concentration and conversion of hydrocarbon precursors. Sample sets of IAR (as defined in equation 35) at initial aerosol mass (Mint) equals 5 and 15 g/m3 and the values of Fractional Aerosol Coefficient (FAC) for 11 aromatic hydrocarbons are given in Table 4.1. Table 4.1: Incremental aerosol reactivity for 11 aromatic compounds IAR at 25% conv. IAR at 50% conv. IAR at 75% conv. IAR at 25% conv. IAR at 50% conv. IAR at 75% conv. Mint = 5 g/m3 Mint = 5 g/m3 Mint = 15 g/m3 Mint = 15 g/m3 Mint = 15 g/m3 Compounds Mint = 5 g/m3 1,2,4-trimethylbenzene 0.0006 0.002 0.007 0.0014 0.006 0.0192 1,3,5-trimethylbenzene 0.00007 0.001 0.001 0.0002 0.002 0.0053 ethylbenzene 0.00173 0.009 0.041 0.0040 0.02 0.0582 m-ethyltoluene 0.00043 0.003 0.006 0.0012 0.007 0.0170 m-xylene 0.00018 0.001 0.002 0.0005 0.002 0.0062 n-propylbenzene 0.00344 0.019 0.068 0.0061 0.027 0.0705 o-ethyltoluene 0.00076 0.002 0.013 0.0018 0.009 0.0304 o-xylene 0.00045 0.002 0.006 0.0012 0.005 0.0165 p-ethyltoluene 0.00093 0.004 0.013 0.0024 0.011 0.0311 p-xylene 0.00067 0.003 0.008 0.0007 0.007 0.0209 toluene 0.00095 0.006 0.019 0.0024 0.012 0.0322 FAC* 0.02 0.029 0.054 0.063 0.047 0.016 0.056 0.05 0.025 0.016 0.054 * FAC = mass of aerosol produced/ mass of hydrocarbon reacted (Grosjean , 1992) 130 Total of four sets of IAR are calculated at 25 and 75% conversion for Mint of 5 g/m3, and at 50% conversion for Mint of 5and 15 g/m3. At 50% conversion, incremental aerosol reactivity ranges from 0.001 (1,3,5-trimethylbenzene) to 0.019 (n-propylbenzene) for Mint equal 5 g/m3, and from 0.002 (1,3,5-trimethylbenzene) to 0.027 (n-propylbenzene) when Mint is 15 g/m3. For the same amount of initial aerosol concentration (e.g., 5 g/m3), IAR increases as the conversion of parent hydrocarbons gets higher. For example, IAR of toluene are 0.00095, 0.006, and 0.019 at 25, 50, and 75%, respectively. These values are somewhat lower than the Fractional Aerosol Coefficients (Grosjean, 1992) reported in Table 4.1. As noted in Chapter 2, these FACs were based on knowledge of the reactions of these compounds and expert judgement. The most probable reason for the FACs values being consistently higher than the models used in this work is due to partitioning. Early models of SOA formation generally assumed all condensable species would reside in the aerosol phase, while current models recognize that a gas phase-condensed phase equilibrium will be established. IARs of all investigated compounds (shown in Table 21) are positive, which means that adding more hydrocarbon precursors results in higher amounts of aerosol products. This contrasts with incremental reactivities for ozone, which have both positive and negative values (Carter, 1994). IARs for Mint equal 15 g/m3 are a factor of 2-3 higher than those for Mint equal 5 g/m3, although the ratio is not consistent for all compounds. These incremental aerosol reactivities may be useful for estimating 131 aerosol formation potentials from emission ratios. A better approach, as outlined in Chapter 3 would be to estimate aerosol yield based on condensable product formation. If individual reaction products are tracked, then the results of Chapter 3 can be employed. In many cases, however, only lumped reacting species are used. Therefore, the remainder of this chapter will present a lumped mechanism for predicting SOA yield from aromatic precursors. 4.2 SOA formation in Eulerian photochemical models Airshed model applications include simulations of mixtures of a large number of reacting species. The information and chamber data required for determining the exact kinetic behaviors of individual species are often unavailable. To solve this problem, species with similar reaction rates and mechanisms are commonly grouped together and represented by lumped model species. The lumped model species are treated as a pseudo compound. One of the main benefits of lumping is that the chemistry of a large number of volatile organic compounds (VOCs) can be modeled without complete information for each individual species. The previous chapter described the development of quantitative models of SOA formation for individual hydrocarbon precursors. This chapter discusses the lumping scheme used for aromatic species and the estimates of model parameters for lumped species, which were developed based on information for groups of individual compounds. 132 4.3 Quantitative models of SOA formation for lumped species 4.3.1 Theoretical Background The aggregation system of lumping compounds into groups based on their chemical and physical similarities has been studied by researchers for over three decades. Wei and Kuo (1969) were pioneers in studying aggregation of many monomolecular reacting species. Goliken (1972) investigated the effect of temperature on the reaction rate of group model species. Results revealed that although the activation energies of individual species can be described by the Arrhenius expression, this equation does not adequately describe the activation energies of lumped species. The behavior of the lumped species can differ from that of a single compound. The properties of the grouped species may depend on the compositions and total concentrations of the feed (Golikeri and Luss, 1974). To diminish errors associated with representing individual species by lumped model species, it may be necessary to lump compounds into sub-groups. There are various types of lumping approaches used in applications of airshed modeling. Carter (2000) recommended several lumping methods. The first method is the lumped molecule approach. In this approach, VOCs are represented by model species in the base mechanisms on a molecule-for-molecule basis (i.e., Clump = Ci,species)(Carter, 2000). For example, in SAPRC the lumped higher aldehyde species, RCHO, are used to represent all aldehydes. This approach is appropriate for compounds with similar chemical characteristics. 133 The second approach is the variable weighting lumped parameter. In this method, a group of VOCs with similar reactivities are represented by a lumped model species. Kinetic and product yield parameters of the lumped species depend on the concentrations of VOCs in the reacting mixture. Kinetic parameters for the lumped species are weighted averages of emissions of VOCs within the mixture (e.g., klump = Wi*ki, species, where Wi is a weighting factor). This approach is the most accurate lumping method for following reasons. First, compared to the lumped molecule approach (which sum up the concentrations of participating VOCs in the mixture regardless of their reactivities), this approach weights kinetic and product yield parameters depending on reactivities of VOCs in the mixture. Second, model parameters are evaluated based on the characteristics of VOCs in the mixture, while those for the fixed parameter approach (discussed below) are derived using a typical ambient mixture, and used for all applications regardless of the actual concentrations. There are two weighting methods currently used to estimate the kinetic and product yield parameters of model species in SAPRC. The first method is based on reactivity weighting, and second is based on molar weighting. For reactivity weighting, the contribution of a given VOC is related to the amount of VOC that is estimated to react (Carter, 2000). Amount reacted Fraction Reacted = Amount emitted * Fraction reacted Equation 37 = Kinetic reactivity (1-exp(-kOH*IntOH)) Equation 38 134 Where kOH is the OH radical rate constant and IntOH is an effective integrated OH radical rate constant. Reactivity weighting is appropriate for slower reacting species (kOH < 2.0E +04 ppm-1min-1). For slowly reacting species, 100% conversion is unlikely to happen in a limited simulation time. Therefore, a fraction reacted is utilized to determine the actual amount of each VOC that reacts. In molar weighting, the contribution of a given VOC to the parameters of lumped model species is proportional to the amount of VOC emitted or input. This method is more suitable for faster species (kOH > 2.0E+04 ppm-1min-1) (Carter, 2000), since the fast reacting species are more likely to react completely in a shorter period of time. The third lumping approach is a fixed parameter method. Parameters for the lumped species are estimated from a representative ambient mixture or emission profile. These parameters are then applied in all model simulations irrespective of the actual emissions involved. This approach is accurate if the compositions of VOCs emitted are well represented by the compositions used to derive the mechanisms. 4.3.2 Lumping Scheme The lumping approach used in SAPRC is the variable weighting lumped parameter method. Aromatic species are lumped into two groups: ARO1 and ARO2 based on their reactivities. ARO1 represents a group of species that react slowly (kOH < 2.0E +04 ppm-1min-1 at 300 K). Species that react fast (kOH > 2.0E +04 ppm-1min-1) are represented by ARO2. Table 4.2 lists species in ARO1 and ARO2. 135 Table 4.2: Lists of species represented by model species ARO1 and ARO2 VOC Name Benzene Toluene Ethyl Benzene n-Propyl Benzene Isopropyl Benzene (cumene) C9 Monosub. Benzenes s-Butyl Benzene C10 Monosub. Benzenes C11 Monosub. Benzenes C12 Monosub. Benzenes o-Xylene p-Xylene m-Xylene C9-Disub. Benzenes C10-Disub. Benzenes C11-Disub. Benzenes C12-Disub. Benzenes 1,3,5-Trimethylbenzene 1,2,3-Trimethylbenzene 1,2,4-Trimethylbenzene C9 Trisub. Benzenes C10 Trisub. Benzenes C11 Trisub. Benzenes C12 Trisub. Benzenes C10 Tetrasub. Benzenes Represented By* BENZENE TOLUENE C2-BENZ N-C3-BEN I-C3-BEN N-C3-BEN S-C4-BEN N-C3-BEN N-C3-BEN N-C3-BEN O-XYLENE P-XYLENE M-XYLENE 0.34 M-XYLENE + 0.33 OXYLENE + 0.33 P-XYLENE 0.34 M-XYLENE + 0.33 OXYLENE + 0.33 P-XYLENE 0.34 M-XYLENE + 0.33 OXYLENE + 0.33 P-XYLENE 0.34 M-XYLENE + 0.33 OXYLENE + 0.33 P-XYLENE 135-TMB 123-TMB 124-TMB 0.34 135-TMB + 0.33 123-TMB + 0.33 124-TMB 0.34 135-TMB + 0.33 123-TMB + 0.33 124-TMB 0.34 135-TMB + 0.33 123-TMB + 0.33 124-TMB 0.34 135-TMB + 0.33 123-TMB + 0.33 124-TMB 0.34 135-TMB + 0.33 123-TMB + 0.33 124-TMB O-C2-TOL P-C2-TOL M-C2-TOL Lumped As 0.295ARO1** ARO1 ARO1 ARO1 ARO1 ARO1 ARO1 ARO1 ARO1 ARO1 ARO2 ARO2 ARO2 ARO2 ARO2 ARO2 ARO2 ARO2 ARO2 ARO2 ARO2 ARO2 ARO2 ARO2 ARO2 o-Ethyl Toluene ARO2 p-Ethyl Toluene ARO2 m-Ethyl Toluene ARO2 * represented in SAPRC99 **one mole of benzene is presented by 0.295 moles of ARO1, because kinetic reactivity of benzene is less than 1/3 that of toluene and ARO1 is dominated by toluene and alkylbenzenes Integrating the quantitative SOA formation into SAPRC requires that aerosol species be lumped as well. Aerosol precursor species (APRs) are separated into four groups: APR1-APR4, based on the reactivities of aromatic reactants and aerosol precursors. 136 The same criteria used in lumping aromatic reactants are also applied in the second step aggregation (APRs). APR1 are aerosol precursors that react slowly that are formed by aromatic precursors that react slowly (ARO1) APR2 are aerosol precursors that react slowly that are formed by aromatic precursors that react rapidly (ARO2) APR3 are aerosol precursors that react rapidly that are formed by aromatic precursors that react slowly (ARO1) APR4 are aerosol precursors that react rapidly that are formed by aromatic precursors that react rapidly (ARO2) The lists of detailed species of aerosol precursors, along with their rate constants for reactions with OH radicals for each lumped groups are presented in Table 4.3. Table 4.3: Lists of detailed species and their OH rate constants represented by model species APR1, APR2, APR3, and APR4 Precursors APRs kOH 3 Kom1 3 Kom2 3 Stoichiometric Coefficients* (cm /molc-sec) (m / g) (m / g) APR1 (slow reactivity hydrocarbon precursors and slow reactivity APRs) Benzene APB1 O O 6.96E-12 0.20 O I-C3-Ben APB2 O 6.96E-12 0.167 CH3 H3C O O APB4 1.00E-11 137 0.079 Precursors APRs kOH (cm /molc-sec) 3 Kom1 (m / g) 3 Kom2 (m / g) 3 Stoichiometric Coefficients* 0.167 S-C4-Ben APB5 O O 6.96E-12 CH3 O O APB7 H3C 1.00E-11 0.119 Toluene APTO2 O O 6.96E-12 0.0123 0.00109 0.167 Ethylbenzene APEB1 O O 6.96E-12 0.036 0.00322 0.167 APEB3 CH3 O O 1.00E-11 0.036 0.00322 0.079 O C3-Ben APPB1 O 6.96E-12 0.0722 0.00198 0.167 APR2 (fast reactivity hydrocarbon precursors and slow reactivity APRs) 124-TMB APT9 AP10 H3C O O H3C 1.27E-11 1.00E-11 0.00708 0.00708 0.000164 0.000164 0.046 0.107 H3C O O H3C p-xylene APX4 O O 1.00E-11 0.0042 0.000367 0.489 138 Precursors APRs kOH (cm3/molc-sec) Kom1 (m3/ g) 0.00459 Kom2 (m3/ g) 0.000369 Stoichiometric Coefficients* 0.20 o-xylene APX3 O O 6.99E-12 H3C m-xylene p-C2-Tol APX8 H3C O O 1.00E-11 1.00E-11 1.00E-11 0.00203 0.0069 0.0069 0.000175 0.000569 0.000569 0.188 0.20 0.04 APET2 CH3 O O APET3 O O o-C2-Tol m-C2-Tol APET8 H3C O O 6.99E-12 O O 0.00844 0.0049 0.000735 0.000433 0.186 0.054 APET10 CH3 O 1.00E-11 APET11 O 1.00E-11 0.0049 0.000433 0.134 APR3 (slow reactivity hydrocarbon precursors and fast reactivityAPRs) H3C CH3 I-C3-Ben APB3 1.85E-11 0.158 O O H3C S-C4-Ben APB6 O O CH3 1.88E-11 0.119 H3C Toluene APTO1 O O 1.85E-11 0.0123 0.00109 0.238 139 Precursors Ethylbenzene APRs APEB2 CH3 kOH (cm3/molc-sec) 1.85E-11 CH3 Kom1 (m3/ g) 0.036 Kom2 (m3/ g) 0.00322 Stoichiometric Coefficients* 0.158 O O O C3-Ben APPB2 H3C O 1.85E-11 1.85E-11 0.0722 0.0722 0.00198 0.00198 0.158 0.079 APPB3 O O APR4 (fast reactivity hydrocarbon precursors and fast reactivity APRs) CH3 O O H3C 135-TMB APT1 2.66E-11 0.00337 0.000294 0.79 123-TMB APT2 CH3 O O 1.85E-11 0.431 APT3 CH3 H3C O O 2.66E-11 0.254 CH3 APT4 H3C O O 3.66E-11 0.085 H3C CH3 124-TMB APT5 O O CH3 4.88E-11 0.00708 0.000164 0.254 CH3 APT6 H3C O O 2.66E-11 0.00708 0.000164 0.084 APT7 H3C CH3 O O 2.66E-11 0.00708 0.000164 0.131 140 Precursors APRs CH3 kOH (cm3/molc-sec) Kom1 (m3/ g) 0.00708 Kom2 (m3/ g) 0.000164 Stoichiometric Coefficients* 0.146 APT8 H3C O O 3.36E-11 H3C CH3 O p-xylene APX5 CH3 O CH3 4.88E-11 0.0042 0.000367 0.24 o-xylene APX1 O O 1.85E-11 0.00459 0.000367 0.45 CH3 APX2 m-xylene APX6 H3C O O 2.66E-11 1.85E-11 0.00459 0.00203 0.000369 0.000175 0.14 0.376 CH3 O O CH3 APX7 H3C O O 2.66E-11 0.00203 0.000175 0.188 CH3 p-C2-Tol APET1 O O CH3 4.57E-11 0.0069 0.000569 0.49 CH3 o-C2-Tol APET4 CH3 O O 1.85E-11 0.00844 0.000735 0.329 O APET5 O H3C 1.00E-11 0.00844 0.000735 0.085 CH3 CH3 O O APET7 H3C 2.70E-11 0.00844 0.000735 0.051 CH3 O APET6 APET9 O 2.70E-11 O O 0.00844 0.00844 0.000735 0.000735 0.026 0.0659 H3C 1.85E-11 141 Precursors APRs CH3 kOH (cm3/molc-sec) Kom1 (m3/ g) 0.0049 Kom2 (m3/ g) 0.000433 Stoichiometric Coefficients* 0.08 m-C2-Tol APET12 H3C O O 2.70E-11 H3C APET13 CH3 O O 2.70E-11 0.0049 0.000433 0.106 CH3 APET14 H3C O O 1.85E-11 1.85E-11 0.0049 0.0049 0.000433 0.000433 0.226 0.153 APET15 O O * moles of APRs produced per one mole of precursor reacted Figure 4.1 depicts lumping scheme of aromatic hydrocarbons and aerosol precursors that is used in the lumped model developed in this thesis. These hydrocarbon and aerosol precursors were lumped using the variable lumped parameter approach. Figure 4.1: Lumping scheme for aromatic species and aerosol precursors modified in SAPRC 142 As shown in Figure 4.1, the first tier of lumping is for aromatic hydrocarbon precursors. Groups of aromatic species are lumped and represented by two model species: ARO1 and ARO2 based on their reactivities. In the second tier, aerosol precursors are grouped and represented by four lumped model species: APR1-APR4. For example, APR1 represents a group of less reactive aerosol precursors that are produced from slow reactive aromatic reactants. Quantitative models of SOA formations were developed for these six lumped groups. ARO1 + HO. = gas-phase products + a1APR1 + a2APR2 ARO2 + HO. = gas-phase products + a3APR3 + a4APR4 APR1 + HO. = a5PM1 + a6PM2 + a6RO2R + a5RO2N APR2 + HO. = a7PM3 + a8PM4 + a8RO2R + a7RO2N APR3 + HO. = a9PM5 + a10PM6 + a10RO2R + a9RO2N APR4 + HO. = a11PM7 + a12PM8 + a12RO2R + a11RO2N Reaction 21 Reaction 22 Reaction 23 Reaction 24 Reaction 25 Reaction 26 Where a1-a11 are stoichiometric coefficients. Coefficients a1-a4 were estimated using either the reactivity or molar weighting method from SAPRC. Coefficients a5-a11 were derived using the reactivity weighting method and simulations. These estimates will be described in the next section. PM1 through PM8 are low-volatility products produced from reactions of aerosol precursors with OH radicals. 143 4.3.3 Estimates of Model Parameters After the models of SOA formation for lumped aromatic species were developed and incorporated into SAPRC, the model parameters for each lumped species listed in Table 4.4 were estimated. Analogous to model parameters for individual compounds, model parameters for lumped species can be classified into two categories. The first group of parameters is used to describe the reactions. This group contains rate constants, stoichiometric coefficients, and molecular weights of the lumped species. The second group of parameters is used to estimate the phase partitioning of semivolatile products. This group contains partitioning coefficients. Table 4.4: Model Parameters Used to Simulate SOA Formation from Lumped Aromatic Species Parameters Chemical model ki ai MWi Physical model Kom,i SOA production Ai Fi,om Mo Yi Gas phase concentration of product i Concentration of product i in the absorbing om phase Organic aerosol mass concentration produced for a given amount of ROG reacted Fractional aerosol yields for product i Simulation calculation Simulation calculation Simulation Calculation Partitioning coefficient of compound, i Fitting method, weighting method Rate constant of compound i Stoichiometric coefficient of compound i Molecular weight of compound i Fitting method, weighting method Fitting method, weighting method Weighting method Descriptions Calculation Procedures 144 There are two methods used in this study to estimate model parameters: the weighting method and the simulations. As described in the preceding section, the weighting methods can further be categorized into two groups: reactivity weighting and molar weighting. Model parameters for ARO1, which is the model species that represents a group of slower aromatic compounds, were estimated using reactivity weighting (based on data in SAPRC). Parameters for ARO2, which represents a group of faster aromatic compounds, were determined using molar weighting (based on data in SAPRC). Model parameters for APR1-APR4 were estimated based on reactivity weighting. Although the molar weighting can be used for fast reacting species (APR2 and APR4), it is more appropriate to use the reactivity weighting for the following reasons. First, these model species represent groups of aerosol precursors with widely varying kinetic reactivities. Second, the reactivity weighting is reduced to the molar weighting for fast reacting species. Finally reactivities of some of aerosol precursors are relatively close to the reactivity used to separate molar and reactivity weighting in SAPRC (2.0E+4 ppm-1min-1). Thus, employing the reactivity weighting to estimate model parameters for APR1-APR4 covers the entire range of slow and fast reacting species. To describe the expressions used to estimate the model parameters, the consecutive reactions resulting in the condensable products must be considered. A + OH. B + OH. k1 k2 a1 B a2 C 145 Where A is aromatic reactant, B is aerosol precursor, and C is condensable product. The concentration at any given time is expressed by the following equations. CA(t) = CA0e-k1IntOH CB(t) = a1CA0k1/(k2-k1)[e-k1IntOH e-k2IntOH] Equation 39 Equation 40 CC(t) = CA0-CA-CB = a2k2a1k1CA0/(k2-k1)[(e-k2IntOH)/k2 (e-k1IntOH)/k1 (1/k2-1/k1)] Where: CA(t) = concentration of hydrocarbon precursor at time t (ppm) CB(t) = concentration of aerosol precursor at time t (ppm) CC(t) = concentration of semivolatile product at time t (ppm) k1 k2 a1 = OH rate constant of hydrocarbon precursor (ppm-1min-1) = OH rate constant of aerosol precursor (ppm-1min-1) = stoichiometric coefficient of aerosol precursor produced from Equation 41 oxidation reaction of hydrocarbon precursor a2 = stoichiometric coefficient of semivolatile product produced from aerosol precursor reaction CA0 = initial concentration of hydrocarbon precursor (ppm) IntOH = an effective integrated OH radical rate constant, which is the concentration of OH integrated during the specific time. For the purpose of lumping, OH concentration is assumed to be constant. (IntOH =1.0E-4 ppm-min for regional model applications, Carter, 2000) 146 The weighting factor (WF) is based on the amount for hydrocarbon and aerosol precursors that react. Weighting factor of A and B at time t are given by WFA = CA(t)R = CA0 CA(t) = CA0(1-e-k1*IntOH) Equation 42 WFB = 1/a2*CC(t) = CB(t)R = CBgenerated(t) CB(t) = k2a1k1CA0/(k2-k1)[(ek IntOH )/k2 2 (e-k1IntOH)/k1 (1/k2-1/k1)] Equation 43 Where: CA(t)R = amount of hydrocarbon precursor that reacts (ppm) CB(t)R = amount of aerosol precursor that reacts (ppm) CAgenerated(t) = amount of hydrocarbon precursor generated (ppm) CBgenerated(t) = amount of aerosol precursor generated (ppm) The weighting factor used to estimate model parameters for ARO1 was determined using equation 42. The weighting factor for ARO2 (fast reacting group) was simply the emitted amount of VOCs in the group (i.e., WFARO2 = CARO20 , where CARO20 is the initial concentration of model species ARO2 (ppm)). Equation 43 is applied to evaluate weighting factors for aerosol precursors (APR1-APR4). Besides using weighting factors, empirical fitting was also employed in the parameter estimates for lumped model compounds (APR1-APR4). The concentrations of the lumped model species were compared to the summed concentrations of individual species in that group. Parameters such as OH rate constant of lumped aerosol precursors (klump) were adjusted to improve the fit. For example, OH rate constant of model species APR1 was estimated by comparing the 147 concentration of APR1 with the summed concentrations of individual species it represents (i.e., fitting CAPR1 = Ci, species by adjusting kOH,lump of APR1). Estimates of Rate Constants of lumped aerosol precursors (APR1-APR4) with OH Radical The rate constants of lumped species (APR1-APR4) representing aerosol precursors can be determined using weighting method as expressed below: k OH ,lumped k * WFi = WFi OH ,i i i Equation 44 Where WFi is expressed as equation 43 and kOH,i is rate constant of aerosol precursor species i represented by lumped species. In addition to weighting method, kOH of the lumped species (APR1-APR4) can be estimated by fitting the rate constant to the results from simulations using single compound. The concept of this method is to fit the concentration of a lumped aerosol precursor to the summed concentrations of individual aerosol precursor (equation 45). The summation of concentrations of individual species is denoted as: C APR ,i = a 1,i C A0,i k 1,i /(k 2,i - k 1,i )[e i -k1, i IntOH - e -k 2, i IntOH ] Equation 45 Where: CAPRi = concentration of aerosol precursor species i (ppm) CA0,i = initial concentration of hydrocarbon precursor responsible in producing aerosol precursor species i (ppm) 148 k1,i = OH rate constant of hydrocarbon precursor responsible in producing aerosol precursor species i (ppm-1min-1) k2,i a1,i = OH rate constant of aerosol precursor species i (ppm-1min-1) = stoichiometric coefficient of aerosol precursor species I produced from reaction of hydrocarbon precursor The concentration of lumped aerosol compounds is given by Clumped APR = a1,lumped*CA0,lumped*k1,lumped/(k2,lumped-k1,lumped)[e-k1,lumped*IntOH e-k2,lumped*IntOH] Equation 46 Where: Clumped APR = concentration of lumped aerosol precursor (APR1-APR4) (ppm) CAO,lumped = initial concentration of lumped hydrocarbon precursor (ARO1 or ARO2) responsible in producing lumped species APR1-APR4 (ppm) k1,lumped = OH rate constant of lumped hydrocarbon precursor (ARO1 or ARO2) (ppm-1min-1) k2,lumped (ppm-1min-1) a1,lumped = stoichiometric coefficient of lumped aerosol precursor (APR1= OH rate constant of lumped aerosol precursor (APR1-APR4) APR4) produced from reaction of lumped hydrocarbon precursor (ARO1 or ARO2) Figure 4.2 shows the results of applying fitting and weighting methods to calculate kOH for lumped aerosol precursors as well as the summation of concentrations of individual aerosol precursors. Results for APR1, APR2, and APR4 149 show a good fit among the three sets of data points. These results indicate that the estimated rate constants of lumped APR species are a good representative for those of individual species in the groups. The concentration of lumped APR2 estimated from the weighting method differs from the summation of concentrations of individual aerosol precursors (CAPRi) by approximately 5-30%. This reveals that the reactivity of lumped APR2 estimated using the weighting factor is overestimated. However, OH rate constant of lumped APR2 obtained from the fitting method agrees better with the summation of concentrations of individual species (CAPRi), therefore this OH rate constant can be used for APR2. The estimated rate constants for lumped APR1 thorough APR4 are reported in Table 4.5. The disadvantage of using the rate parameter derived from the fitting method is that these results depend on the specific composition of the base mixtures in ways that are not precisely defined. The gas hydrocarbon composition used in the work is defined in appendix and is based on data for Houston, Texas. 150 Figure 4.2: Concentrations of aerosol precursors as a function of time, three sets of data: (1) summation of concentrations of individual species in the group, (2) obtained by using fitting method (Clump, fitting method), and (3) estimated using weighting factor (Clump, weighting method) 151 Estimates of Stoichiometric Coefficients Semivolatile products produced from lumped APR1-APR4 are denoted as PM1-PM8. Semivolatile products with organonitrate groups were classified as PM1, while products without organonitrate groups were denoted as PM2. PM1 and PM2 are formed from the reaction of APR1, and so on for PM3-PM8. Stoichiometric coefficients of PM1-PM8 can be estimated using either the weighting or fitting method. By using the weighting method, stoichiometric coefficients for lumped condensable species (PM1-PM8) are determined by alumpedPM a , *WFi = WFi 2 i i i Equation 47 Where: a2,i = stoichiometric coefficient of semivolatile product species i being represented by lumped model species alumpedPM = stoichiometric coefficient of lumped semivolatile product (PM1PM8) WFi = weighting factor for species i as expressed by Equation 43. For empirical fitting, the concentration of lumped semivolatile product (CPM1 CPM8), expressed by equation 48, was fitted to the summed concentration of detailed products in the groups (equation 49) by adjusting the stoichiometric coefficient for the lumped species (parameter "a2,lumped in equation 48). Clumped PM = a2,lumped*k2,lumped*a1,lumped*k1,lumped*CA0lumped/(k2,lumped-k1,lumped)* [(e-k2,lumped*IntOH)/k2,lumped (e-k1,lumped*IntOH)/k1,lumped (1/k2,lumped-1/k1,lumped)] 152 Equation 48 C PM , i = a 2,i k 2,i a 1,i k 1,i C A0,i /(k 2,i - k 1,i )[(e i -k 2, i IntOH )/k 2,i - (e -k1, i IntOH )/k 1,i - (1/k 2,i - 1/k 1,i )] Equation 49 Where: Clumped PM = concentration of lumped semivolatile product (ppm) a2,lumped a2,i = stoichiometric coefficient of lumped semivolatile product = stoichiometric coefficient of semivolatile product species i Seen in Figure 4.3 are the estimates of the concentration of lumped semivolatile products for APR1 (PM1 and PM2) using the weighting and fitting method, and the summation of concentrations of semivolatile products for individual aerosol precursors. Results of all products from four lumped aerosol species are displayed in Appendix E. As shown in Figure 4.3, concentrations of PM1 and PM2 fit extremely well among the three sets of data points. Stoichiometric coefficients of condensable products generated from the individual species in each group are well represented by those of products produced from lumped aerosol precursors. The estimated stoichiometric coefficients of semivolatile products for the four lumped aerosol precursors (APR1-APR4) are reported in Table 4.5. 153 Figure 4.3: Concentrations of condensable products, three sets results: summation of concentrations from individual species, concentration of lumped species obtained by using the fitting method, and by using weighting factor Estimates of Partitioning Coefficients A gas/particle partitioning absorption mechanism developed by Pankow (1996) was employed in this study to estimate the amount of semivolatile products partitioning to the aerosol phase. The partitioning coefficient for compound i (Kom,i) is expressed in terms of total aerosol mass concentrations as: Kom,i = Fi,om/Ai*Mo And Where: Ai Fi,om phase ( g/m3) 154 Equation 50 Equation 51 Ai = Pi Fi,om = gas phase concentration of product species i ( g/m3) = concentration of semivolatile product species i in the absorbing om Pi Mo = total concentration of semivolatile product species i ( g/m3) = total aerosol mass concentration ( g/m3) The estimated Kom,i for lumped species obtained using the weighting method is given by Kom, i *WFi Kom, lumped = WFi i i Equation 52 Where Kom,lumped is partitioning coefficient of lumped semivolatile product (m3/ g), and WFi is given by Equation 43. The estimated partitioning coefficients are then applied in equation (53) and (54) to determine SOA mass changes (M) for lumped products. In addition to using the weighting method, Kom for lumped products can also be calculated using the empirical fitting. Kom for lumped species (parameters "Kom,lumped" in equation (54)) was adjusted until the difference between M of lumped species and the summed aerosol mass change from individual products in the group (Mi) is minimized. Mi and Mlumped are given by the following equations: M i = i K om ,1i * Mo * P1i 1 - Mo * K om ,1i + K om , 2i * Mo * P2i 1 - Mo * K om , 2i Equation 53 M lumped = K om ,lumped 1 * Mo * P1lumped 1 - Mo * K om ,lumped 1 + K om ,lumped 2 * Mo * P2lumped 1 - Mo * K om ,lumped 2 Equation 54 155 Where: Kom1,i = partitioning coefficient of product type 1 produced from aerosol precursor species i (m3/ g) Kom2,i = partitioning coefficient of product type 2 produced from aerosol precursor species i (m3/ g) P1,i species i ( g/m3) P2,i species i ( g/m3) Mi = aerosol mass changes for aerosol precursor species i ( g/m3) = total semivolatile product type 2 produced from aerosol precursor = total semivolatile product type 1 produced from aerosol precursor Mlumped = aerosol mass changes for lumped aerosol precursor ( g/m3) Komlumped,1 = partitioning coefficient of lumped product type 1 (e.g., Kom for PM1) produced from lumped aerosol precursor (m3/ g) Komlumped,2 = partitioning coefficient of lumped product type 2 (e.g., Kom for PM2) produced from lumped aerosol precursor (m3/ g) P1,lumped = total semivolatile lumped product type 1 (e.g., PM1)produced from lumped aerosol precursor ( g/m3) P2,lumped = total semivolatile lumped product type 2 (e.g., PM2) produced from lumped aerosol precursor ( g/m3) 156 Figure 4.4: SOA mass changes from four lumped aerosol species: three sets of data point: Mi from individual species in the groups, M for lumped products obtained from fitting and weighting methods Results of comparison among M for the lumped products using the fitting methods, and the summation of M of individual products indicate good 157 correspondence between them for the four lumped aerosol precursors. The calculated SOA mass changes determined using the weighting method differ from Mi in the ranges of 5-17%. It can be implied from these results that the partitioning coefficients determined for the lumped products are good representatives for those of individual products. The estimated partitioning coefficients are presented in Table 4.5. Estimates of Molecular Weights of Lumped Species Molecular weights of the lumped species were obtained by using the weighting method. This method averages number of hydrogen and carbon atoms in the product species for each lumped group. Aerosol precursors from the individual parent hydrocarbon in the same groups share the same base structure, but can differ in the total numbers of carbon and hydrogen atoms. For example, if APT2 and APX8 will be lumped together in the same group, and their structures are O O APT1 O O APX8 Therefore, APT1 and APX8 share base structure, which is C5H6O2. APT1 has one carbon and two hydrogen atoms more than base structure. In turn, the molecular weights of lumped species are the combination of the molecular weight of base 158 structure and the weighted averages of the molecule weight of the excess carbon and hydrogen atoms from the individual aerosol precursors in these groups. molecular weights of the lumped species are given by MWlumped = MWbase structure + XC2*12 + XH2 Equation 55 The Where XC2 and XH2 are the weighted averages of the extra carbon and hydrogen atoms from the individual aerosol precursors in the groups. The weighting of these parameters is based on the reactivities of the aerosol precursors and parent hydrocarbons, and the emissions of hydrocarbons. Tables 4.5 and 4.6 enumerate the estimated parameters for the lumped species obtained from both the fitting and weighting methods. Ambient conditions for Houston, TX (Carter, 1998) were used in the estimates of model parameters reported in Table 4.5. While the estimated molecular weights of lumped species, as listed in Table 4.6, were determined for the ambient mixtures at the Clinton site observed on the First and Second of September 2000. 159 Table 4.5: The estimated parameters for lumped species from fitting method (FT) and weighting method (WF) APR1 FT kAPR (cm /molcsec) Stoichiometric coefficients Stoichiometric coefficients of semivolatile product type 1 produced from Stoichiometric coefficients of semivolatile product type 2 produced from Partitioning coefficients of semivolatile product type 1 produced from Partitioning of coefficient of semivolatile product type 2 produced from 0.00144 0.00156 0.00026 0.00032 0.00264 0.00137 0.000233 0.000328 0.016 0.023 0.003 0.00359 0.0224 0.0214 0.0027 0.0039 0.75 0.75 0.70 0.75 0.75 0.75 0.76 0.75 0.25 0.25 0.30 0.25 0.25 0.25 0.24 0.25 0.179 0.260 0.209 0.499 3 APR2 FT 7.83E-12 APR3 WF FT 1.82E-11 APR4 WF FT 2.8E-11 WF 7.19E-12 WF 3.0E-11 7.17E-12 1.05E-11 1.85E-11 Table 4.6: The estimated molecular weights of lumped species from weighting method Date Sep01 Sep02 MW ARO1 89.9 ARO2 106.0 APR1 84.0 APR2 98.0 APR3 98.0 APR4 109.7 PM1 163.0 PM2 117.0 PM3 177.0 PM4 131.0 PM5 177.0 PM6 131.0 ARO1 87.8 ARO2 114.8 APR1 85.2 APR2 99.0 APR3 100.0 APR4 110.0 PM1 178.0 PM2 132.0 PM3 179.0 PM4 133.0 PM5 190.0 PM6 178.0 160 As discussed in Chapter 3, Odum fit adjustable 4 parameters, Kom,1, Kom,2, 1, and 2, that provided the best fit of SOA yields from his model to observations made in environmental chamber. These values are 0.038, 0.042, 0.167, and 0.0014 for 1, Kom,1, 2, and Kom,2, respectively for high-yield aromatics. Odum found 1, Kom,1, 2, and Kom,2 for low-yield aromatics are 0.071, 0.053, 0.138, 0.0019, respectively. In this study, high-and low-yield aromatics are represented by model species ARO1 and ARO2. ARO1 produces aerosol precursors represented by lumped model species APR1 and APR3. Model species APR2 and APR4 represent aerosol precursors generated from ARO2. Thus Kom,i for semivolatile products produced from APR1 and APR3 (PM1,PM2,PM5,and PM6) are compared with Kom,i for high-yield aromatics of Odum, and Kom,i for products formed from APR2 and APR4 (PM3, PM4, PM7, and PM8) are compared with those for low-yield aromatics. Table 4.7 recaps model parameters for lumped species estimated from this work and from Odum's study. Table 4.7: Summaries of model parameters for lumped species Low-yield aromatics Parameters Estimates* Odum's study Kom1 0.0033 0.042 Kom2 0.000286 0.0014 *Values were estimated from this work High-yield aromatics Estimates* Odum's study 0.0207 0.0018 0.053 0.0019 4.3.4 Correlation between SOA yields and chemical and physical parameters 161 Chapter 3 presented empirical correlations for SOA yield for individual aromatic compounds. This section will present correlations for lumped aromatic compounds. The following expressions, developed for individual compounds, were also applied for lumped aromatics: Mi = (%DFi) * (Pi) Equation 56 %DFi = a*Mint2 + b*Mint + c Equation 57 In these expressions, %DFi is distribution factor of lumped semivolatile product species i in the aerosol phase. In Chapter 3 it was shown that %DFi depends only on Mint and is independent of % conversion (for about a dozen aromatic compounds). This is expected be observed for lumped compounds as well. To assess this hypothesis, Mint was changed from 5 to 15 and 25 g/m3, and %DFi at 25, 50, and 75% conversions were then calculated for each Mint level. 162 Figure 4.5: %DFi as a function of Mint at 25, 50 and 75% conversions for lumped species APR1APR4 163 Seen in Figure 4.5 are %DFi as a function of Mint at 25, 50, and 75% conversions for lumped aerosol precursors APR1- APR4. Results for these lumped species show the dependency of %DFi on Mint but little dependence on % conversion. The exceptions are APR2 and APR3. For APR2 and APR3, the percentage of product that partitions into the aerosol phase (%DFi) is similar at 50 and 75% conversions, but is different for 25% conversion. The biggest deviations between %DFi at 25 and 50% conversion for APR2 and APR4 are 13 and 33%, respectively. Similar reason causing the digression for individual compounds, which is the amount of Mi is comparable to Mtot. The polynomial fitting was performed for %DFi at 50% conversion for these lumped species. Table 4.8 lists the results of the evaluation of %DFi, as a function of Mint (as expressed in Equation 56). Table 4.8: Coefficients a, b, and c for the correlation equation between %DFi and Mint for lumped species APR1-APR4 Lumped species a* b* c* APR1 -0.0075 0.69 0.3375 APR2 -0.002 0.13 -0.1 APR3 -0.0105 0.94 0.3625 APR4 -0.005 0.09 0.625 * These coefficients were evaluated based on the base hydrocarbon composition and concentrations as seen in Houston, TX on September 2, 2000 4.4 Summary The variable parameter lumping approach was used to lump aromatic hydrocarbon precursors into two groups. The aerosol precursors produced from the two lumped aromatic species (ARO1 and ARO2) were further separated into four groups (APR1-APR4) based on how rapidly these aerosol precursors react with OH 164 radical. Quantitative models of SOA formation for six lumped model species were developed and integrated into SAPRC99. Rate constants, stoichiometric coefficients, partitioning coefficients, and molecular weight of the lumped species were estimated from information on the individual species in the groups. Table4.5 summarizes the source of data for the parameters used in the model. The approaches used in these calculations were the fitting and weighting methods. These parameters were estimated based on emissions of the parent hydrocarbon in the reacting mixtures and the reactivities of the hydrocarbon and aerosol precursors. An ambient hydrocarbon mixture for Houston,TX area from 1998 data (Carter, 1998) was used in the estimates of the fixed model parameters, with the exception of the estimates of molecular weights of lumped species. Molecular weights were determined for the Houston conditions observed on the First and Second of September 2000. Overall results indicate that when an appropriate group of hydrocarbons is aggregated together, the accuracy of representing individual compounds by the lumped model species is enhanced. The SOA formation model developed in this study gives researchers an effective procedure to predict ambient SOA formation independent of complete information and understanding of individual ambient species. The models and reaction parameters developed in this chapter are applied to characterize SOA formation in Houston, TX in Chapter 5. 165 Chapter 5 Characterization of ambient SOA formation in Houston area 5.1 Introduction Regional air quality models currently in use often neglect the formation of SOA in the aerosol module. This is due to the lack of complete understanding of secondary organic aerosol and its sources. With the implementation of proposed daily and annual PM-2.5 standards (NAAQS), many areas such as Houston, TX would likely to be declared non-attainment areas. Since SOA is a major contributor to ambient PM-2.5, meeting the proposed standards requires the reduction of emissions of hydrocarbons leading to SOA formation. In order to develop an effective strategy to reduce PM-2.5 concentrations, the sources and mechanisms describing SOA formation must be integrated into current air quality models. Quantitative models of SOA formation for lumped aromatic hydrocarbons, developed in this study, were employed to characterize ambient SOA formation in the Houston, TX region. This chapter reports on the sensitivity of SOA formation in Houston to changes in concentrations of NOx, as well as the effect of SOA formation on ozone formation chemistry. 166 5.2 Case Study Scenario PM_2.5 in Texas Figure 5.1 reveals the average constituent concentration ( g/m3) found in PM_2.5 in several cities in Texas. High levels of organic compounds have been found in PM_2.5, range from 4-5.6 g/m3 in, Houston, Mauriceville, and Dallas. Figure 5.1: Average concentrations of constituents found in PM2.5 collected between 3/97 and 3/98 (Tropp et. al., 1998) for various Texas cities. The Houston area was selected for investigation due to the potential for high daily average concentrations of PM_2.5. Figure 5.2 shows the annual average organic carbon (OC), elemental carbon (EC), and OC/EC ratios in the Houston area in 1998. These data was taken from the Desert Research Institute PM2.5 study of 1997/1998 (Tropp, et al., 1998). Monitoring sites are denoted in green. EC ranges 167 from 0.2-1.5, region that has the relatively highest EC in the map is downtown Houston. OC map shows that OC ranges about 3.0-3.9 spatially, except in the south east of downtown Houston. During 1997/1998, the OC/EC ratio for this region ranges from 0.7 to 4.0. The higher ratios are observed outside of the urban or industrial area. Fractions of secondary organic carbons cannot be estimated, because the primary OC/EC ratios are not known. However, if it is assumed that the OC/EC ratio from primary emissions is relatively constant spatially, then the results in figure would suggest that there is more OC relative to EC outside the urban area and that in these areas SOA formation is significant. Thus these data indicate that SOA formation cannot be ignored in air quality simulations in the Houston area. 168 EC OC 169 OC/EC Figure 5.2: The annual average OC, EC, and OC/EC ratio at Houston areas on February 4, 1998, data was extracted from the Desert Research Institute PM2.5 study of 1997/1998 Sensitivity of SOA formation with respect to changes in concentration of NOx and primary hydrocarbons were performed under three case study scenarios. The first case or base case characterizes ambient SOA formation at the Clinton site, Houston, TX on the first and the second of September 2000, assuming that it a closed system. This case is denoted as case A. The second case, case B, analyzes the sensitivity of SOA formation to changes of emission rates of NOx. The third case, comprising of two sub-cases (case C1 and C2), examines the effects of NOx emission rates on SOA formation when the system is dominated by NOx emissions (i.e., VOC limiting), and by VOC emissions (i.e., NOx limiting), respectively. Table 5.1 summarizes these case studies. 170 Table 5.1: Case studies for characterization and sensitivity analysis of SOA formation Case A Summarized description Closed system: ambient pollutants at the Clinton site, Houston, TX on the First and the Second of September 2000 at 4:00 pm Case A with the emission of NOx Objectives to examine Characterize and compare formation of ambient SOA NOx emission VOC emission NO: NO2 Investigate the effect *1050 of emission rate and tons/day/155km2 composition of NOx on SOA formation C1 Case A with the Examine the *1050 emission of NOx sensitivity of SOA tons/day/155 and VOC: VOC formation to km2 limiting changes of NOx emission with the presence of VOC emission for higher emission of NOx than VOC C2 Case A with the Analyze the effect of **150 tons/day/ emission of NOx emission rate of 155km2 and VOC: NOx NOx on SOA limiting formation with the presence of VOC emission for higher emission of VOC than NOx *Data obtained from 1998 inventory in Houston Ship Channel area **Data obtained from 1996 inventory in Houston Ship Channel area B - 0.85:0.15 0.90:0.10 0.95:0.05 0.85:0.15 **275 tons/day/155 km2 **145 tons/day/155 km2 0.85:0.15 5.2.1 September 2000 Episode Data observed during September 1-2, 2000 (during the Texas Air Quality Study (TEXAQS) period) at the Clinton site were chosen to define the base case. This episode was selected due to high concentrations of m-and p-xylene at the site on September 1, 2000. This event occurs suddenly between 3:00- 4:00 pm is believed to be due to a large release of m-and p-xylene directly into the atmosphere. Data on 171 September 1, 2000 represents non-routine release of VOCs. Data on September 2, 2000, on the other hand, represents typical conditions. Bar graphs in Figure 5.3 illustrate the concentrations of three compound classes: aromatics, alkenes, alkanes, and NOx during the September episode at the Clinton site between 3:00- 4:00 pm. Total emission on the first of September is approximately 17 times higher than that on the second. Concentration at the site on September 1, 2000 are dominated by aromatic hydrocarbons. Figure 5.4 shows a speciated distribution of aromatic hydrocarbon concentrations during the September episode. Data indicate that the combination of p-and m-xylene is a major component of the emissions on September 1, 2000. Quantitative models were used to characterize formation of SOA on this episode. Figure 5.3: Hydrocarbon concentrations observed from TexAQS at the Clinton site on September 1 and 2, 2000 at 4:00 pm 172 Figure 5.4: Monitoring speciated concentrations of aromatic hydrocarbons on September 1 and 2, 2000 at the Clinton site at 4:00 pm 5.2.2 Case Study Description Case A: Closed system Box model simulations were performed to characterize SOA formation at the Clinton site for the September 1-2, 2000 episode. Monitoring data collected during this episode were used as the initial conditions for the box model simulations, and are reported in Table 5.2. The goal of the case study is to characterize SOA formation under typical and non-routine hydrocarbon releases into atmosphere. Simulations started at 4:00 pm and ended at 10:00 pm. Results were reported every half hour. 173 The estimated SOA mass changes, yields, and ozone concentrations on each day are compared. These results are presented in the next section. Table 5.2: Initial conditions used in box model simulations for the September episode Date Sep 1 Sep 2 TOT.HCs (ppbvC) 1705 106 NO (ppb) 4.5 0.1 NO2 (ppb) 25.2 8.4 T (F) 102.5 102.3 % RH 48.1 46.7 Case B: Box model with NOx emissions Case study B investigated the effects of NOx emission rates on SOA formation. The concentrations within a plume of air traveling form the Clinton site along the Ship Channel area were investigated. Case B used initial conditions from the first of September, 2000, with the addition of NOx emission at a constant rate. Figure 5.5 shows point sources of NOx in the Houston Ship Channel area. NOx emission data were extracted from the TNRCC point source database for 1997 or 1998. The emissions are reported as NO2. The NOx emission for the entire 155 km2 Houston Ship Channel area is 1050 tons/day. Assuming that mixing height is 500 meters, NOx emission rate is 1.37E-8 tons/m3/day. 174 Figure 5.5: NOx and VOC point sources in Houston Ship Channel area (TNRCC, 1998) A sensitivity analysis of SOA formation to changes in emissions of NOx was examined. Rates of NOx emissions were reduced from the base case by 10-100%, and also increased by 10-70%. The fraction of NO was also varied from 0.85-0.95 simultaneously with the changes of NOx emission to investigate the effect of NOx composition on SOA formation. Emission rates of NOx estimated for NO: NO2 equal to 0.85:0.15 are listed in Table 5.3. 175 Table 5.3: Emission rates of NOx at Houston Ship Channel area, fraction of NO is equal 85% Changes from base case (%) Base case -100 (no NOx emission) -95 -90 -80 -70 -50 -30 -10 +10 +30 +50 +70 Emission rates of NO (mmol/m3/min) 1.75E-04 0.00E-00 8.77E-06 1.75E-05 3.51E-05 5.26E-05 8.77E-05 1.23E-04 1.58E-04 1.93E-04 2.28E-04 2.63E-04 2.98E-04 Emission rates of NO2 (mmol/m3/min) 2.79E-05 0.00E-00 1.55E-06 3.10E-06 6.19E-06 9.29E-06 1.55E-05 2.17E-05 2.79E-05 3.41E-05 4.02E-05 4.64E-05 5.26E-05 Case C: Box model with VOC and NOx emissions Case C1: VOC limiting Case C1 examines the sensitivity of SOA formation to changes in NOx emission in the presence of VOC emissions under VOC limiting conditions (i.e., higher emission of NOx than emission of VOC). Initial conditions and NOx emission rates as described in Case B were used in this case with the addition of VOC emissions along Houston Ship Channel area. VOC point sources are illustrated in Figure 5.6. The estimated emission rate of VOC over the entire area is 257 tons/day. The emissions are reported on a carbon and hydrogen basis. Using the C/H mass ratio of 12/14, the mass basis emission was converted to mole basis, which is 1.66E-04 mmolC/m3/min. The ratio between VOC and NOx emissions for this case is 0.80 mmolC/mmol, which is generally regarded as VOC limited. The composition of the hydrocarbon emissions was assumed to be the same as reported by Carter (1998). For 176 the sensitivity analyses, NOx was reduced by 10-100%, and increased by 10-70% from the base case. Emissions of VOCs were kept constant. Case C2: NOx limiting Case study C2 investigates the effect of emission rate of NOx on SOA formation for the NOx limiting system (i.e., VOC emission is higher than NOx emission). In this scenario, NOx and VOC point sources at Houston Ship Channel area from the 1996 inventory were employed. VOC point sources are illustrated in Figure 5.6. The emission rate of VOC is approximately 145 tons/day, and using C/H ratio of 12/14, this rate can be converted to 9.37E-05 mmolC/m3/min. The rate of NOx emission, which is reported as NO2, is 155 tons/day. Assuming that the fraction of NO is 0.85, emission rates of NO and NO2 are 2.78E-5 and 4.90E-6 mmol/m3/min, respectively. The ratio between VOC and NOx emission is 2.86 mmolC/mmol. To examine the sensitivity analysis of SOA formation, base case emission of NOx was changed to 10-100 %. The composition of hydrocarbons for emission is identical to that of case C1. 177 Figure 5.6: VOC point sources in Ship Channel area from 1996 inventory 5.3 Case Study Results Quantitative models of SOA formation, developed in this study, were employed to characterize SOA production during the September episode. The goal is to investigate SOA formation when high concentrations of reactive parent hydrocarbon are present. The analysis will help identifying an important role of sources and physical and chemical processes of SOA formation. Results for each case study are displayed in the following sections. 5.3.1 Case A: Box model Box model simulations were conducted under initial conditions as described previously. The comparison of SOA mass changes (Mo) on September 1 and 2, 2000 are shown in Figure 5.7. Mo were estimated for Mint equals 15 g/m3. As 178 seen in the Figure, Mo on the first date are 13-16 times higher than those on the second date. The high production of Mo on the first date results from a massive release of reactive hydrocarbon (m-and p-xylene) into ambient air. Figure 5.7: SOA mass changes (Mo) at the Clinton site at 4:00 pm on the September episode, aerosol seed varies from 5-15 g/m3. Each lumped model aerosol species will produce two types of condensable products that will partition into aerosol phase. The first type is compound with organonitrate groups, and the second is compounds with carboxylic acid groups. Products from the four lumped aerosol precursors (APR1-APR4) are referred to as: PM1, PM2, PM3. PM4, PM5, PM6, PM7, and PM8. Partitioning of individual products depend on their vapor pressures and molecular weights (i.e., lower vapor 179 products can relatively partition more into aerosol phase). Mass concentrations of these semivolatile products in particulate phase (PM1-PM8) during the September episode are illustrated in Figure 5.8. 180 Figure 5.8: Speciated aerosol mass changes of eight aerosol products (PM1-PM8) during the September episode, Mint = 5 g/m3 181 For September 1, 2000, there was non-routine release of m-and p-xylene, which was represented by lumped species ARO2 in SAPRC-99. ARO2 produces aerosol precursors represented by lumped species APR2 and APR4. From the figure, aerosol mass change (M) of PM7 dominates the total mass changes in aerosol phase. This can be explained by following reasons. First, the reactivities of both ARO2 and APR4 are faster than the other lumped species. Thus, rate of production of semivolatile product species PM7 and PM8 from lumped species APR4 are higher than that of the others. Secondly, the partitioning coefficient of product PM7 is higher than that of product PM8, therefore PM 7 partitions more into particulate phase. However, results from September 1, 2000 are not typical ambient conditions. When there are no unusual emissions of individual compounds (e.g., September 2, 2000), the major components of total aerosol mass are products PM1 and PM5, because partitioning coefficients of PM1 and PM5 are higher than those of the others. Generally, the simulations suggest that fractions of products with organonitrate group are higher than those of products with carboxylic acid group. Mass changes of individual aerosol products (Mo of PM1-PM8) are normalized by the reacted amount of hydrocarbons responsible for their production (i.e., Yields). Table 5.4 lists the yields of each product, and the ratios of yield for each product to the minimum yield among these 8 products (Yi/Ymin). The ratios indicate that PM5 has the highest yield among all products independent of the initial ambient conditions. 182 Table 5.4: SOA yields of each aerosol products on the September episode Date Products PM1 PM2 PM3 PM4 PM5 9-1-2000 Yields 9.9E-04 1.9E-04 3.0E-04 5.8E-05 3.0E-03 Yi/Ymin 17.0 3.3 5.1 1.0 51.2 9-2-2000 Yields 5.1E-03 9.8E-04 1.3E-03 2.5E-04 5.7E-03 Yi/Ymin 20.4 4.0 5.2 1.0 22.8 PM6 7.8E-04 13.4 1.5E-03 6.0 PM7 1.2E-03 20.8 3.2E-03 12.8 PM8 2.6E-04 4.4 6.8E-04 2.7 Besides secondary aerosol formation, ozone formation is another important issue to be examined in this study. Box model simulations with only gas phase chemistry (SAPRC 99) were performed under conditions from the September episode. The ozone concentrations from gas phase chemistry and from case study, which includes SOA formation chemistry, are illustrated in Figure 5.9. The comparison shows that ozone concentrations from gas phase chemistry are higher than those from aerosol case study. The difference is approximately 21% on September 1, 2000 and 7% on September 2, 2000. Causes of these differences include as: introducing chemistry of SOA formation into a gas phase mechanism consume reactive hydrocarbons (APR) that might otherwise contribute to ozone formation by photolysis reaction. 183 Figure 5.9: Ozone concentrations at the Clinton site on the September episode from 4:00 pm, data points are from case study simulation, gas-phase chemistry simulation, and monitors The ground monitoring ozone concentrations at the Clinton site from 4:00 pm are also presented in Figure 5.9. Results from simulations are not comparable with monitoring data, because simulations were performed for box model (closed system). 184 The parcel containing the high concentrations of air was not followed. Further, the loss mechanisms, such as the depositions, are not handled in SAPRC simulation. It is evident from aircraft monitoring data (Twin Otto, afternoon flight on September 1, 2000) that there are plumes of high ozone concentrations (120-165 ppb) in Houston area from 3:45-4:21 pm (Figure 5.10). This elevated ozone concentrations may due to the enormous release of reactive hydrocarbons at the Clinton site between 3:00 and 4:00 pm. These high concentrations could not be observed from the ground monitors. Therefore, this explains the discrepancy between simulation and ground monitoring data. Figure 5.10: Ozone concentrations on September 1, 2000 around 3:30 to 4:30 pm, at the Ship Channel area, Houston TX 185 5.3.2 Case B: Box model with NOx emission NOx emissions were added into the box model to investigate the sensitivity of SOA yields to changes in the concentrations of NOx. Mint is equal 5 g/m3, and the fraction of NO is equal 85% in case study B, C1 and C2. The simulations were conducted under conditions as described in the case study scenario section. Results from simulations are depicted in Figures 5.11 and 5.12. The estimated SOA mass changes decrease 5-60% from base case when NOx emissions are increased 10-70%. Yields (Y) and Mo increase when NOx emissions are reduced up to 80%. Beyond 80%, aerosol masses and yields begin to decline. To understand this phenomenon, it is useful to consider these following reactions: hv NO2 ROO* + NO NO2 O + O2 O3 + NO O3 + H2O NO + O* NO2 + RO* O + NO O3 NO2 + O2 2OH. + O2 Reaction 27 Reaction 28 Reaction 29 Reaction 30 Reaction 31 Reaction 32 Where ROO* is hydrocarbon radical. NO plays the role of an ozone sinks, while NO2 is ozone source. In the case study, NOx emissions are dominated by NO (85%). Therefore, increasing NOx emissions will enhance the ozone depletion by NO, which consequently decreases the formation of OH radicals and SOA. On the other hand, when NOx emissions decrease up to 80% from base case, the ozone titration by NO 186 decreases. Therefore, ozone concentration and SOA formation is getting higher as the reduction of NOx emissions increases. However, when the reduction of NOx emission goes beyond 80% from base case, the depletion of ozone by NO does not play an important role in ozone production and destruction. Therefore, SOA formation and ozone concentration decline at the high reductions of NOx emissions (i.e., > 80%). 187 Figure 5.11: SOA mass changes at the Clinton site from 4:00 on September 1,2000: results from varying NOx emissions by 10-100% from base case, Mint = 5 g/m3, fraction of NO = 0.85 188 Figure 5.12: Ozone concentrations at the Clinton site from 4:00 on September 1,2000: results from varying NOx emissions by 10-100% from base case, Mint = 5 g/m3, fraction of NO = 0.85 189 Impact of composition of NOx on SOA formation was also examined. The fraction of NO was varied from 85-95%. SOA mass changes for 50% NOx emissions from base case at all levels of fraction of NO are shown in Figure 5.13. Results indicate that varying composition of NOx does not influence formation of SOA significantly. Figure 5.13: SOA yields at the Clinton site on September 1, 2000 from 4:00 pm, results from changing NOx emissions 50% from base case, and varying fraction of NO from 0.85-0.95, Mint = 5 g/m3 190 5.3.3 Case C: Box model with NOx and VOC emissions There are two sub-cases for this scenario: VOC limiting and NOx limiting. Simulations were conducted under conditions as described in the preceding section for cases C1 and C2. For these cases, the amount of hydrocarbon that reacts is estimated by Rdt = 0 t t E dt - dC V C0 0 t C Equation 58 Where: R = the amount of hydrocarbon that reacts ( g/m3/min) E = emission of hydrocarbon ( g/min) V = Volume of system (m3) C0 = Initial concentration of hydrocarbon ( g/m3) Results of these two cases are illustrated in Figures 5.14 - 5.17. For case C1 (VOC limiting), results are similar to Case B, in which increasing NOx emissions results in declining SOA formation and ozone concentration. Decreasing NOx emissions 70-80% from the base case enhances SOA formation and ozone concentration. The causes of this behavior are similar to case B. Case C2 (NOx limiting) shows behavior opposite to case C1. As seen in Figure 5.16 and 5.17, when NOx emissions are raised by 10-70% from base case, formation of SOA and ozone increases. In this case the formation of SOA and ozone are strongly influenced by the concentration of NOx in the system. 191 The sensitivity analyses from case C1 and C2 reveal that the strategy to control SOA formation depends strongly on the base case scenarios (i.e., NOx or VOC limiting). 192 Case C2: NOx limiting Figure 5.14: Aerosol mass changes ( g/m3) as the puff of air from the Clinton site moves along the Houston Ship Channel area for case C2, results from varying NOx emissions by 10-95% from base case, SOA seed = 5 g/m3, and fraction of NO = 0.85 193 Figure 5.15: Ozone concentrations as the puff of air from the Clinton site moves along the Houston Ship Channel area for case C2, results from varying NOx emissions by 10-95% from base case, SOA seed = 5 g/m3, and fraction of NO = 0.85 194 Case C1: VOC limiting Figure 5.16: Aerosol mass changes ( g/m3) as the puff of air from the Clinton site moves along the Houston Ship Channel area for case C1, results from varying NOx emissions by 10-95% from base case, SOA seed = 5 g/m3, and fraction of NO = 0.85 195 Figure 5.17: Ozone concentrations as the puff of air from the Clinton site moves along the Houston Ship Channel area for case C1, results from varying NOx emissions by 10-95% from base case, SOA seed = 5 g/m3, and fraction of NO = 0.85 Results from sensitivity for cases C1 and C2 indicate that the impact of changing NOx emissions on aerosol formation is similar to that on ozone production. Figure 5.18 shows a correlation between changes in ozone concentration and changes 196 in SOA concentrations for the scenarios shown in Figures 5.14 5.17. Changes of 0.1 ppm in ozone concentration lead to changes in SOA concentrations of approximately 2.5 g/m3. Figure 5.18: O3 versus SOA compared to base case NOx emissions when NOx emissions were decreased to 100%, and increased to 70% from the base case 5.4 Summary Quantitative models of SOA formation for lumped aromatic hydrocarbons were applied to characterize the formation of SOA in the Houston area. A number of case studies were performed to examine the key parameters controlling SOA formation. Monitoring data for a September 2000 episode were employed as initial conditions for all case studies. Results indicate that the incorporation of SOA 197 formation chemistry into gas phase chemistry (SAPRC99) strongly influences ozone formation chemistry The concentration of NOx is another essential aspect in SOA formation, because it influences the generation of OH radical. Therefore, the concentration of NOx is believed to be the one of the limiting factors in SOA formation. For VOC limited scenario, decreasing NOx emissions will enhance the formation of SOA and ozone, while raising NOx emissions will deplete ozone and SOA formation. For NOx limited case, formation of ozone and SOA is proportional to the concentration of NOx in the system. The development of quantitative models of SOA formation by this study provides an effective tool in characterizing SOA production more accurately and directly from its sources, as well as in identifying the key features in controlling SOA formation. 198 Chapter 6 Summaries and Future Work 6.1 Summaries Quantitative models of secondary organic aerosol (SOA) formation for 15 individual aromatic hydrocarbons and two lumped aromatic species were developed based on data available from smog chamber experiments. These models were used to characterize the SOA formation in Houston areas, and to examine the key components in controlling SOA formation. The finds are summarized below. Key Conclusions Improving the modeling of secondary organic aerosol needs to account for the fact that SOA is produced from the second generation products of hydrocarbons with OH radical (HCs OH. . Aerosol precursor OH SOA). Fractional Aerosol Coefficients (FACs) are not constant for each VOC, but can vary depending on several factors especially the extent of conversion and the amount of seed aerosol. The concentration of NOx is a determining factor in SOA formation. Sensitivity analyses indicate that changing the concentration of NOx can have significant positive and negative effects on SOA formation. The effect of NOx is represented by a direct correlation of ozone and SOA formation. Aerosol yields of organonitrate products are higher than those of carbonyl products. Because the organonitrate aerosol products have lower vapor 199 pressure than the carbonyl products but lower reaction yields. The vapor pressure effect is stronger. SOA formation chemistry strongly influences ozone formation chemistry. Production of SOA competes the consumption of hydrocarbon precursors leading to ozone formation. 6.2 Future Work Further studies that are recommended are: Apply models with more extensive chamber data. Extend the development of SOA module to other hydrocarbon classes: alkenes, alkanes, and terpenes. Incorporate SOA mechanisms into regional air quality models (e.g., CAMx) so the module can be employed to estimate the formation of SOA in regional and urban ambient conditions Integrate SOA mechanism with chlorine chemistry. There is evidence that chlorine chemistry promotes ozone formation (Tanaka, 2000), and the reactivity of alkanes with chlorine radical is much higher than that with OH radical. This may enhance the production of OH radicals in urban environments, and consequently increase production of SOA. 200 Nomenclatures *Listed as presented chronically in the dissertation NAAQS EPA SOA FACs SAPRC PM OC EC VOCs MIR MOIR EBIR RO2R National Ambient Air Quality Standards Environmental Protection Agency Secondary Organic Aerosol Fractional Aerosol Coefficients Statewide Air Pollution Research Center Particulate Matter Organic Carbon (Volatile carbon) Elemental Carbon (non-volatile carbon) Volatile Organic Carbons Maximum Incremental Reactivity Maximum Ozone Incremental Reactivity Equal Benefit Incremental Reactivity Operator represents peroxy radical reactions with NO that result in NO to NO2 conversion and formation of HO2 radical R2O2 Operator represents the effect of NO to NO2 conversion without HO2 formation RO2N Operator represents the reactions of peroxy radicals with consumption of NO and various types of organic nitrate formation 135-TMB 1,3,5-trimethylbenzene 201 APT1 BALD CRES MGLY TPM1 Aerosol precursor from 1,3,5-trimethylbenzene Benzaldehyde Cresol Methylglyoxal Aerosol product with organonitrate groups from 1,3,5trimethylbenzene TPM2 Aerosol product with carbonyl groups from 1,3,5trimethylbenzene 124-TMB GLY BACL APT5 APT6 APT7 APT8 APT9 AP10 TPM5 1,2,4-trimethylbenzene Glyoxal Biacetyl Aerosol precursor from 1,2,4-trimethylbenzene Aerosol precursor from 1,2,4-trimethylbenzene Aerosol precursor from 1,2,4-trimethylbenzene Aerosol precursor from 1,2,4-trimethylbenzene Aerosol precursor from 1,2,4-trimethylbenzene Aerosol precursor from 1,2,4-trimethylbenzene Aerosol product with organonitrate groups from 1,2,4trimethylbenzene TPM6 Aerosol product with carbonyl groups from 1,2,4trimethylbenzene Kom,i G/P Partitioning Coefficient of species I 202 Fi,om Ai Pi Mo or Mtot ROG i Concentration of compound i in the absorbing om phase Gas phase concentration of compound i Total concentration of condensable product i that is formed Total organic aerosol mass concentration Amount of reactive organic gas reacts Proportionality constant relating the amount of ROG reacted to form Pi Yi Mo or Mi Aerosol yield Aerosol mass formed from photolysis reaction and partitioning into aerosol phase Mint %DFi Initial amount of aerosol mass Distribution factor of species i in the aerosol phase (expressed as a percentage) IAR HC ARO1 Incremental Aerosol Reactivity Hydrocarbon Lumped species represents a group of aromatics that react slowly ARO2 Lumped species represents a group of aromatics that react rapidly APR1 Lumped species represents a group of slow reactive aerosol precursors produced from slow reactive aromatic precursors 203 APR2 Lumped species represents a group of fast reactive aerosol precursors produced from slow reactive aromatic precursors APR3 Lumped species represents a group of slow reactive aerosol precursors produced from fast reactive aromatic precursors APR4 Lumped species represents a group of fast reactive aerosol precursors produced from fast reactive aromatic precursors PM1 PM2 PM3 PM4 PM5 PM6 PM7 PM8 Aerosol product species 1 produced from APR1 Aerosol product species 2produced from APR1 Aerosol product species 1 produced from APR2 Aerosol product species 2 produced from APR2 Aerosol product species 1 produced from APR3 Aerosol product species 2 produced from APR3 Aerosol product species 1 produced from APR4 Aerosol product species 2 produced from APR4 204 Appendix A : Data from smog chamber experiment Table A.1: Relative Molar Loading (Eusebi, 1996) Hydrocarbon Aliphatic Aromatic Ketone + Precursors C-H bonds C-H bonds Aldehyde 1-Decane 0.79 0.07 1-Dodecane 0.96 0.01 o-Xylene 0.63 ND 0.19 1,3,5-Trimethybenzene 0.58 0.006 0.22 0.80 0.06 -Pinene R(+)-Limonene 0.65 0.23 0.85 0.10 -Caryophyllene Organic Acid 0.00 0.00 0.04 0.05 0.04 0.02 0.006 Alkyl Nitrate 0.01 0.01 0.03 0.02 0.005 ND ND Nitro % of Initial Aromatics Hydrocarbon 0.62 2.7 ND 0.38 0.01 0.27 13 1.8 1.7 Alcohol 0.13 0.03 0.12 0.12 0.09 0.09 0.04 Table A.1 lists the relative molar loadings of the observed functional groups for each of the smog chamber experiments by Eusebi (1996), and the total carbon mass of the secondary aerosol. Eusebi suggested some patterns for the aerosol generated by different hydrocarbon precursors. Photooxidation of the aromatic hydrocarbons leads to products that have strong carbonyl and organonitrate bands. They also shows moderate C-H absorbances. The aerosol produced from oxidation reaction of olefins exhibit strong C-H stretch absorbances and organonitrate bands. The carbonyl absorbance is relatively weak in these spectra. Table A.2: Predicted composition of SOA from the photooxidation of hydrocarbon precursors (Eusebi, 1996) Hydrocarbon Number of Aliphatic Aromatic Ketone + Precursors Carbons C-H bonds C-H bonds Aldehyde 1-Decane 10 16.24 1.44 1-Dodecane 12 24.47 0.25 o-Xylene 5 5.91 ND 1.78 1,3,5-Trimethybenzene 6 6.21 3.21 9 5.27 0.055 2.73 10 14.10 1.06 -Pinene R(+)-Limonene 9 10.32 3.65 14 21.37 2.51 -Caryophyllene Organic Acid 0.00 0.00 0.38 0.07 0.06 0.7 0.32 0.20 Alkyl Nitrate 0.21 0.25 0.28 0.27 0.23 0.09 0.00 0.00 Nitro Aromatics ND 0.06 - Alcohol 2.67 0.76 1.13 0.89 0.76 1.59 1.43 1.01 205 The method developed by Eusebi (1996) for calculating the predicted functional group distributions per average secondary aerosol molecule aonsists of five simple computational steps. The first step is to predict the number of carbons per average product structure based on dominant gas-phase products and oxidation mechanisms of original hydrocarbon precursors. For exaple, 1,3,5-trimethylbenzene has nine carbons in average product. The second step is to calculate the number of potential bonding positions accessible to functional group bonding on the carbon chain, 1,3,5-trimethylbenzene has 12 potential bonding positions. Then the next step is to determine the number of bonding position required by each non-aliphatic C-H functional group per molecule, per mole of aliphatic C-H. This is done by multiplying the relative molar loadings reported in Table A.1 by the number of bonding position required by each functional group, and dividing by the relative molar loading of aliphatic C-H. The fourth step is to calculate the number of aliphatic bonds on the carbon chain by deviding the toal number of potential bonding positions by the total number of bond required per alophatic C-H bond. The final step is to back calculate the functional gorup distribution based on the relative molar loading by multiplying the ratios of the relative molar loading by the number of alophatic C-H bons per average molecule (calculated in step 4). 206 Appendix B : Houston atmospheric hydrocarbon composition, obtained from the CARB Report (Cater, 1998) Table B1: Houston atmospheric hydrocarbon composition used in the case study for sensitivity analysis Compounds ETHANE PROPANE N-C4 N-C5 N-C6 N-C7 N-C8 N-C9 N-C10 N-C11 N-C12 N-C13 2-ME-C3 2-ME-C4 2-ME-C5 3-ME-C5 22-DM-C4 23-DM-C4 23-DM-C5 24-DM-C5 3-ME-C6 CYCC5 ME-CYCC5 CYCC6 ME-CYCC6 ET-CYCC6 BR-C6 BR-C7 BR-C8 BR-C9 BR-C10 BR-C11 BR-C12 BR-C13 CYC-C7 ppb/ppbC 1.69E-02 1.41E-02 1.81E-02 6.13E-03 1.32E-03 1.20E-03 7.45E-04 7.45E-04 1.84E-03 1.62E-04 3.24E-04 1.91E-05 7.89E-03 1.52E-02 3.56E-03 2.53E-03 4.58E-04 9.56E-04 6.03E-04 1.27E-03 1.12E-03 7.07E-04 1.60E-03 6.88E-04 6.78E-04 1.82E-04 2.39E-04 2.09E-03 4.04E-03 1.71E-03 1.56E-03 1.62E-04 3.24E-04 1.91E-05 1.24E-04 207 Compounds ETHENE PROPENE 1-BUTENE C4-OLE1 3M-1-BUT 1-PENTEN 1-HEXENE ISOBUTEN 2M-1-BUT T-2-BUTE C-2-BUTE 2M-2-BUT 13-BUTDE ISOPRENE CYC-HEXE A-PINENE 3-CARENE C5-OLE1 C6-OLE1 C7-OLE1 C8-OLE1 C9-OLE1 C10-OLE1 C11-OLE1 C4-OLE2 C5-OLE2 C6-OLE2 C7-OLE2 C8-OLE2 C9-OLE2 C10-OLE2 C11-OLE2 C7-OL2D BENZENE TOLUENE C2-BENZ N-C3-BEN I-C3-BEN C9-BEN1 S-C4-BEN C10-BEN1 C11-BEN1 ppb/ppbC 1.35E-02 3.18E-03 1.15E-03 1.43E-04 3.24E-04 8.03E-04 3.35E-04 1.15E-03 9.16E-04 1.15E-03 9.08E-04 5.17E-04 6.21E-04 1.30E-03 1.72E-04 5.06E-04 1.91E-04 4.39E-04 2.22E-03 1.18E-03 2.39E-04 5.17E-04 9.56E-05 1.91E-04 1.43E-04 3.18E-03 1.00E-03 4.39E-04 2.20E-04 2.49E-04 9.56E-05 1.91E-04 1.91E-04 3.31E-03 9.25E-03 1.28E-03 3.64E-04 1.91E-04 1.62E-04 2.30E-04 1.82E-04 6.51E-04 208 Compounds C12-BEN1 O-XYLENE P-XYLENE M-XYLENE C9-BEN2 C10-BEN2 C11-BEN2 C12-BEN2 135-TMB 123-TMB C9-BEN3 C10-BEN3 C11-BEN3 C12-BEN3 C10-BEN4 C9-STYR C10-STYR ACETYLEN FORMALD ACETALD PROPALD C4-RCHO C5-RCHO C6-RCHO BENZALD ACETONE MEK ppb/ppbC 2.87E-05 1.82E-03 2.20E-03 2.20E-03 2.47E-03 1.54E-03 9.56E-05 8.60E-05 7.26E-04 7.55E-04 2.36E-03 1.60E-03 9.56E-05 8.60E-05 4.21E-04 4.77E-04 3.64E-04 9.75E-03 7.93E-03 4.77E-03 6.97E-04 3.16E-04 1.07E-03 7.37E-04 1.62E-04 3.10E-03 1.10E-03 209 Appendix C : Reaction pathways and mechanistic models of SOA formation of aromatic hydrocarbons modeled in SAPRC-99 1,2,3-trimethylbenzene Oxidation of 1,2,3-trimethybenzene proceeds either through the abstraction of hydrogen, or by addition of the OH radicals to the aromatic ring. Most of reactions proceed through the addition of OH radicals to the aromatic ring (pathway 1). Hydrogen abstraction which accounts for only 4.4 % of all reactions with the OH radicals (SAPRC 99) results in a products represented in SAPRC 99 as benzaldehyde plus two carbon atoms (BALD + 2XC) with an NO to NO2 conversion and the production of HO2 radical (pathway 1). (1) 4.4%: 123-TMB + HO. = BALD + 2XC +RO2R Reaction 1 When OH is added to the ring, the OH radicals add primarily to the ortho position of each side chain. Initial reaction with OH radical addition can produce cresol plus two 210 carbon atoms (CRES + 2XC) as a gas phase product, which accounts for 18.6 % of all reactions (SAPRC 99). (2) 18.6%: 123-TMB + HO. = CRES + 2XC + HO2. Reaction 2 Another reaction pathway that can occur following the formation of the OH-aromatic adduct is bridging of O2 across the ring as shown in pathway 3. Another O2 addition occurs to the OH-aromatic-O2 adduct, then with the abstraction of O atom by NO, the alkoxyl-type radicals are formed. The alkoxyl-type radicals undergoes hydrogen abstraction by O2, and unsaturated dicarbonyls or aerosol precursors are produced. Experimental results from Bandow (1985) indicates that 43 % of the addition of O2 to the OH-aromatics adduct results in biacetyl (BACL) and unsaturated dicarbonyl compound, denoted as APT2, 25 % results in methygyloxal (MGLY) and APT3 and approximately 8.5 % results in glyoxal (GLY) and APT4. Pathways 3-5 show the reaction pathways leading to the formation of APT2-APT4. (3a) 211 (3b) (4a) (4b) (5) 123-TMB + HO. = 0.431APT2 + 0.254APT3 + 0.085APT4 + 0.431BACL + 0.254MGLY + 0.085GLY + 0.77RO2R Reaction 3 Net: 123-TMB + HO. = 0.044BALD + 0.186CRES + 0.809 RO2R + 0.186 HO2. +0.254MGLY + 0.085GLY + 0.431BACL + 0.431APT2 + 0.254APT3 + 0.085APT4 + 0.46XC Reaction 4 212 Aerosol precursors (APT2-APT4) react further with OH radicals via addition at the carbon bond to form two types of aerosol products, TPM3 and TPM4, TPM3 is the aerosol product with organonitrate group (pathways 6a-8a), and TPM4 is the product with hydroxy carbonyl group (pathways 6b-8b). Relative yields of TPM3 and TPM4 were based on data for products for 1,3,5-trimethylbenzene (Eusebi, 1996). (6a) (6b) 213 (7a) (7b) (8a) (8b) APT2 + HO. = 0.25TPM3 + 0.75TPM4+ 0.25RO2N + 0.75RO2R APT3 + HO. = 0.25TPM3 + 0.75TPM4 + 0.25RO3N + 0.75RO2R Reaction 5 Reaction 6 APT4 + HO. = 0.25TPM3 + 0.75MGLY + 0.75BACL + 0.25RO2N +0.75RO2R Reaction 7 214 Reactions 4-7 describe the condensed mechanisms of SOA formation for 1,2,3-trimethylbenzene which proceed through two steps: gas phase reaction of 1,2,3trimethylbenzene with OH radicals to produce aerosol precursors (APT2-APT4), and reactions of APT2-4 with OH radical to form two types of semivolatile products (TPM3 and TPM4). Stoichiometric coefficients in reactions 1-7 are obtained from the evaluation of atmospheric chamber experiments. Table C1 summarizes these coefficients and cites the source of the data. Table C.1: Kinetic parameters of 1,2,3-trimethylbenzene reactions with OH radical forming aerosol products and their source Notations CRES BALD MGLY GLY HO2. BACL RO2R APT2 APT3 APT4 TPM3 TPM4 RO2N XC Descriptions Cresol Benzaldehyde Methylglyoxal Glyoxal Hydroperoxy radical Biacetyl Operator RO2R Aerosol precursor APT2 Aerosol precursor APT3 Aerosol precursor APT4 Aerosol product species 1 Aerosol product species 2 Operator RO2N Balance extra carbon Stoichiometric Coefficients (SC) 0.186 0.044 0.254 0.085 0.186 0.431 0.809 (reaction 4) 0.75 (reaction 5-7) 0.431 0.254 0.085 0.25 0.75 0.25 0.46 Source of SC SAPRC-99 SAPRC-99 Bandow, 1985 Bandow, 1985 SAPRC-99 Bandow, 1985 Bandow, 1985 Estimated from 1,3,5trimethylbenzene Bandow, 1985 Bandow, 1985 Bandow, 1985 Estimated from products from 1,3,5trimethylbenzene Estimated from products from 1,3,5trimethylbenzene Estimated from products from 1,3,5trimethylbenzene SAPRC-99, Bandow, 1985 1,2,4-trimethylbenzene 215 Oxidation of 1,2,4-trimethybenzene either proceeds through the OH addition to the aromatic ring, or the abstraction of H from the methyl group. Most of the reactions, 77%, proceed through the addition of OH radical to the aromatic ring (SAPRC 99). Hydrogen abstraction occurs primarily from the methyl substituents. The hydrogen abstraction reaction results in benzaldehyde plus two carbon atoms (BALD + 2XC), pathway 9. This accounts for 4.4 % of all reactions of 124trimethybenzene with OH radical (SAPRC 99). (9) 4.4%: 124-TMB + HO. = BALD +2XC + RO2R Reaction 8 When OH adds to the ring, the OH radicals add primarily to the ortho position of each side chain. 18.6 % (SAPRC 99) of initial reactions with OH radicals result in cresol plus two carbon atoms (CRES + 2XC), pathway 10. (10) 18.6%: 124-TMB + HO. = CRES +2XC + HO2. Reaction 9 216 Another reaction pathway that can occur following the formation of the OH-aromatic adduct is bridging of O2 across the ring as shown in pathway 11, which leads to the ring clevage reactions. The ring cleavage reaction of the O2-OH aromatics adduct can proceed through ten possible pathways (pathways 11-16). All of these ten reaction pathways proceed through two similar steps, the abstraction of O atom by NO, and the abstraction of H atom by O2. The cleavage reactions produce 25.4 % unsaturated dicarbonyl species or aerosol precursor APT5, 8.4 % of APT6, 13.1 % of APT7, 14.6 % APT8, 4.6 % APT9, and 10.7% AP10 (Bandow, 1985). Along with those unsaturated species, methyglyoxal (MGLY), glyoxal (GLY), and biacetyl (BACL) are also produced with the yields of 0.515, 0.146, and 0.107, respectively (Bandow, 1985). (11a) (12a) 217 (12b) (13a) (13b) (14a) (14b) (15) 218 (16a) (16b) 80.4%: 124-TMB + HO. = 0.254APT5 + 0.084APT6 + 0.131APT7 + 0.146APT8 + 0.046APT9 + 0.107AP10 + 0.469MGLY + 0.192GLY + 0.107BACL + 0.804RO2R Reaction 10 Net: 124-TMB + HO. = 0.044BALD + 0.186CRES + 0.8122RO2R + 0.186HO2. + 0.469MGLY + 0.192GLY + 0.107BACL + 0.254APT5 + 0.084APT6 + 0.131 APT7 + 0.146APT8 + 0.046APT9 + 0.107AP10 + 0.46XC Reaction 11 Aerosol precursors (APT5-APT9, and AP10) can further oxidize to form two types of aerosol products, TPM5 and TPM6. TPM5 is the aerosol product with organonitrate group and TPM6 is a hydroxylic carbonyl. The formation of TPM5 proceeds through the addition of O2 and NO to the OH-dicarbonyl adduct (pathways 17a-22a) , while the production of TPM6 undergoes the addition of O2, and the abstraction of O atom by NO (pathways 17b-22b). Yields of TPM5 (25%) and TPM6 (75%) were estimated from experimental results by Eusebi (1996) for 1,3,5trimethylbenzene. 219 (17a) (17b) (18a) (18b) (19a) (19b) 220 (20a) (20b) (21a) (21b) 221 (22a) (22b) APT5 + HO. = 0.25TPM5 + 0.75TPM6 + 0.25RO2N + 0.75RO2R APT6 + HO. = 0.25TPM5 + 0.75TPM6 + 0.25RO2N + 0.75RO2R APT7 + HO. = 0.25TPM5 + 0.75TPM6 + 0.25RO2N + 0.75RO2R APT8 + HO. = 0.25TPM5 + 0.75TPM6 + 0.25RO2N + 0.75RO2R Reaction 12 Reaction 13 Reaction 14 Reaction 15 APT9 + HO. = 0.25TPM5 + 0.75(2MGLY) + 0.25RO2N + 0.75RO2R Reaction 16 AP10 + HO. = 0.25TPM5 + 0.75TPM6 + 0.25RO2N + 0.75RO2R Reaction 17 Formation of semivolatile products for 1,2,4-trimethylbenzene proceed through steps similar to those for 1,2,3-trimethylbenzene. Stoichiometric coefficients and sources of the data are summarized in Table C2. Table C.2: Kinetic parameters of 1,2,4-trimethylbenzene reactions with OH radical forming aerosol products and their source Notations CRES BALD MGLY GLY HO2. BACL RO2R Descriptions Cresol Benzaldehyde Methylglyoxal Glyoxal Hydroperoxy radical Biacetyl Operator RO2R Stoichiometric Coefficients (SC) 0.186 0.044 0.469 0.192 0.186 0.107 0.8122 (reaction 11) 222 Source of SC SAPRC-99 SAPRC-99 Bandow, 1985 Bandow, 1985 SAPRC-99 Bandow, 1985 Bandow, 1985 0.75 (reaction 12-17) APT5 APT6 APT7 APT8 APT9 AP10 TPM3 TPM4 RO2N XC Aerosol precursor APT5 Aerosol precursor APT6 Aerosol precursor APT7 Aerosol precursor APT8 Aerosol precursor APT9 Aerosol precursor AP10 Aerosol product species 1 Aerosol product species 2 Operator RO2N Balance extra carbon 0.254 0.084 0.131 0.146 0.046 0.107 0.25 0.75 0.25 0.46 Estimated from 1,3,5trimethylbenzene Bandow, 1985 Bandow, 1985 Bandow, 1985 Bandow, 1985 Bandow, 1985 Bandow, 1985 Estimated from 1,3,5trimethylbenzene Estimated from 1,3,5trimethylbenzene Estimated from 1,3,5trimethylbenzene SAPRC-99, Bandow, 1985 o-xylene In addition to trimethylbenzene, xylene also produces a significant amount of secondary organic aerosol in the urban atmosphere. The major daytime loss process for atmospheric aromatic hydrocarbons is through reaction with hydroxyl radical. As with 1,3,5-trimethybenzene, this reaction can proceed through either hydrogen abstraction, or addition to the aromatic ring. For o-xylene, hydrogen abstraction accounts for 6.5% of all reactions (SAPRC99), occurring primarily at the substituent methyl groups. The product of this reaction is benzaldehyde with one extra carbon atom (BALD + 1XC), along with an NO to NO2 conversion and the production of HO2 radical (pathway 23). 223 O2 (23) 4.5%: O-XYLENE + HO. = BALD +1XC + RO2R reaction 18 Addition of the OH radical will occur at one of the three ring positions, but predominantly adds at the aromatic carbon ortho to one of the carbons attached to the methyl group (Grovenstein, 1970; Atkinson et al., 1980). The initial addition of OH radical to o-xylene results in cresol with one extra carbon atom (CRES + 1XC) and HO2 radical (pathway 24). This pathway accounts for 18% of all reactions of oxylene (SAPRC99). (24) 16.1%: O-XYLENE + HO. = CRES + 1XC + HO2. Reaction 19 224 The remainder of the OH radical addition products from o-xylene will further proceed through the addition of O2, forming the bridge of O2 across the ring. This structure then reacts further with O2, with conversion of NO to NO2 and formation of HO2 radical as shown in pathway 25 to produce the ring cleavage products. The ring cleavage reaction of the O2-OH-aromatic-adduct proceeds through four possible positions on the aromatic ring (pathways 25-27). Although chemical structures of all aerosol precursors from o-xylene, APX1-APX3 are different, their formation pathways are similar. The cleavage reactions produce 45% of APX1, 14% of APX2, and 20% of APX3 (Bandow, 1985). In addition to these, MGLY, GLY, and BACL are generated. In total, these pathways account for 74% of all reactions of o-xylene with OH radical. (25a) 225 (25b) (26) 226 (27) 79 % O-XYLENE + HO. = .45 APX1 + .45 MGLY + .14 APX2 + .14 GLY + .20 APX3 + .20 BACL + 0.79RO2R Reaction 20 Net: O-XYLENE + HO. = .045 BALD + .161 CRES + .835 RO2R + .161 HO2. + .45 APX1 + .14APX2 + .2 APX3 + .2 BACL + .14 GLY +.45 MGLY + .206 XC Reaction 21 Aerosol precursors (APX1-APX3) further react to generate two types of aerosol products (XPM1 and XPM2) via the OH-addition at the carbon double bond. Results from previous studies suggested that 25% of APX2-OH addition reaction produce XPM1 (pathways 28a -30a) (Eusebi, 1996). The remaining APX2-OH reacts with O2 and NO to form XPM2, NO2, and HO2 radicals (pathways 28b-30b). 227 (28a) (28b) (29a) (29b) (30a) (30b) 228 APX1 + HO. = .25 XPM1 + .75 XPM2 + .25 RO2N + .75 RO2R Reaction 22 APX1 + HO. = .25 XPM1 + .75 XPM2 + .25 RO2N + .75 RO2R Reaction 23 APX1 + HO. = .25 XPM1 + .75 XPM2 + .25 RO2N + .75 RO2R Reaction 24 Reactions 21-24 describe the condensed mechanisms of SOA formation for oxylene, which proceed through steps similar to 1,3,5-trimethylbenzene. Table C3 summarizes stoichiometric coefficients in reaction 21-24 and cites sources of the data. Table C.3: Kinetic parameters of o-xylene reactions with OH radical forming aerosol products and their source Notations CRES BALD MGLY GLY HO2. BACL RO2R APX1 APX2 APX3 XPM1 XPM2 RO2N XC Descriptions Cresol Benzaldehyde Methylglyoxal Glyoxal Hydroperoxy radical Biacetyl Operator RO2R Aerosol precursor APX1 Aerosol precursor APX2 Aerosol precursor APX3 Aerosol product species 1 Aerosol product species 2 Operator RO2N Balance extra carbon Stoichiometric Coefficients (SC) 0.161 0.045 0.450 0.146 0.188 0.200 0.835 (reaction 21) 0.75 (reaction 22-24) 0.450 0.140 0.200 0.25 0.75 0.25 0.24 Source of SC SAPRC-99 SAPRC-99 Bandow, 1985 Bandow, 1985 SAPRC-99 Bandow, 1985 Bandow, 1985 Eusebi, 1996 Bandow, 1985 Bandow, 1985 Bandow, 1985 Eusebi, 1996 Eusebi, 1996 Eusebi, 1996 SAPRC-99, Bandow, 1985 p-xylene Reaction of p-xylene with OH radical can proceed through either the hydrogen abstraction, or addition to the aromatic ring. The hydrogen abstraction, accounting for 8.3 % of the reactions (SAPRC 99), occurs primarily at the substituent 229 methyl groups. The product of this reaction is benzaldehyde plus one carbon atom (BALD + 1XC), pathway 31. (31) 8.3%: P-XYLENE + OH. = BALD + 1XC + RO2R Reaction 25 Addition of OH radical can occur at one of the three ring positions, but predominantly attaches at the aromatic carbon ortho to one of the carbons attached to the methyl group (Grovenstein, 1970; Atkinson. et al., 1980). The initial addition of OH radical to the aromatic can result in the gas phase product, cresol plus one carbon atom (CRES + 1XC), which accounts for 18.8 % of all reactions of o-xylene with OH radical (SAPRC 99). (32) 18.8%: P-XYLENE + OH. = CRES + HO2. + 1XC Reaction 26 Another reaction pathway that can occur following the formation of the OH-aromatic adduct is bridging of O2 across the ring. Another O2 addition occurs to the OH230 aromatic-O2 adduct, then with the abstraction of O atom by NO, the alkoxyl-type radicals are formed. The alkoxyl-type radical undergoes hydrogen abstraction by O2, leading to the production of ring-cleavage products. The xylene-ring cleavage reactions generate individually 48.9% of APX4 and GLY, and 24% of APX5 and MGLY (pathways 33-34). (34) (35) 231 72.9 % P-XYLENE + OH. = 0.489APX4 + 0.240APX5 + 0.489GLY + 0.240MGLY + 0.729RO2R Reaction 27 Net: P-XYLENE + OH. = 0.489APX4 +0.240APX5 + 0.489GLY + 0.240MGLY + 0.188HO2. + 0.188CRES + 0.083BALD + 0.812RO2R + 0.271XC Reaction 28 Aerosol precursors (APX4-APX5) react further with OH radicals through addition at the carbon double bond to produce two types of semivolatile products (XPM3 and XPM4), where XPM3 are the products with organonitrate groups, and XPM4 is a product with a carbonyl. Pathways 35a-36a and 35b-36b demonstrate the formation of XPM3 and XPM4, respectively. Yields of XPM3 (25%) and XPM4 (75%) were estimated from experimental results for o-xylene from Eusebi (1996). (35a) (35b) 232 (36a) (36b) APX4 + OH. = 0.25XPM3 + 0.75XPM4 + 0.25RO2N + 0.75RO2R APX5 + OH. = 0.25XPM3 + 0.75XPM4 + 0.25RO2N + 0.75RO2R Reaction 29 Reaction 30 Analogous to 1,3,5-trimethylbenzene, reactions 28-30 describe the formation of two types of semivolatile products formed from the oxidation reaction of p-xylene through two steps. The first step is the reaction of p-xylene with OH radical to form aerosol precursors, and the second step is the reaction of aerosol precursors with OH radical to produce semivolatile products. Table 4 summarizes stoichiometric coefficients in reactions 28-30, and cites sources of the data. Table C.4: Kinetic parameters of p-xylene reactions with OH radical forming aerosol products and their source Notations CRES BALD MGLY GLY HO2. RO2R APX4 APX5 Descriptions Cresol Benzaldehyde Methylglyoxal Glyoxal Hydroperoxy radical Operator RO2R Aerosol precursor APX4 Aerosol precursor APX5 Stoichiometric Coefficients (SC) 0.188 0.083 0.240 0.489 0.188 0.812(reaction 28) 0.75 (reaction 29-30) 0.489 0.240 233 Source of SC SAPRC-99 SAPRC-99 Bandow, 1985 Bandow, 1985 SAPRC-99 Bandow, 1985 Estimated from o-xylene Bandow, 1985 Bandow, 1985 XPM3 XPM4 RO2N XC Aerosol product species 1 Aerosol product species 2 Operator RO2N Balance extra carbon 0.25 0.75 0.25 0.271 Estimated from products from o-xylene, Eusebi, 1991 Estimated from products from o-xylene, Eusebi, 1991 Estimated from products from o-xylene, Eusebi, 1991 SAPRC-99, Bandow, 1985 m-xylene Reaction of m-xylene proceeds through either hydrogen abstraction, or addition of OH radical to the aromatic ring. Hydrogen abstraction accounts for 3.7 % (SAPRC 99) of the reaction, and occurs primarily at the substituent methyl groups. The main product of this reaction is benzaldehyde plus one carbon atom (BALD + 1XC), pathway 37. O2 (37) 3.7%: M-XYLENE + OH. = BALD + 1XC + RO2R Reaction 31 234 When OH is added to the ring, the OH radicals add primarily to the ortho position of each side chain. Initial addition of OH radical to the aromatic can result in the gas phase product, cresol plus one carbon atom (CRES + 1XC), which accounts for 21 % of all reactions of o-xylene with OH radical (SAPRC 99). (38) 21%: M-XYLENE + OH. = CRES + HO2. + 1XC Reaction 32 Another reaction pathway that can occur following the formation of the OH-aromatic adduct is bridging of O2 across the ring. The OH-O2- aromatic adduct further reacts with O2 with conversion of NO to NO2 and formation of HO2 radical to generate ring cleavage products. The xylene-ring cleavage reactions produce 37.6% APX6, 18.8% APX7, and 18.8% APX8 (pathways 39-41), Bandow, 1985. Along with those aerosol precursors MGLY and GLY are also formed with yields of 56.4% and 18.8%, respectively (Bandow, 1985). 235 MGLY (39) (40) 236 (41) 75.2 % M-XYLENE + OH. = 0.376APX6 +0.188APX7 + 0.188APX8 + 0.564MGLY + 0.188GLY + 0.752RO2R Reaction 33 Net: M-XYLENE + OH. = 0.21HO2. + 0.21CRES + 0.037BALD + 0.376APX6 +0.188APX7 + 0.188APX8 + 0.564MGLY + 0.188GLY + 0.789RO2R + 0.247XC Reaction 34 Aerosol precursors (APX6-APX8) react further with OH radicals through addition at the carbon double bond to produce two types of semivolatile products (XPM5 and XPM6), where XPM5 is a product with organonitrate group, and XPM6 is a product with carbonyl groups. Pathways 42a-44a and 42b-44b demonstrate the formation of XPM5 and XPM6, respectively. Yields of XPM5 (25%) and XPM6 (75%) were estimated from experimental results for o-xylene from Eusebi (1996). 237 (42a) (42b) (43a) (43b) (44a) (44b) 238 APX6 + OH. = 0.25XPM5 + 0.75XPM6 + 0.25RO2N + 0.75RO2R APX7 + OH. = 0.25XPM5 + 0.75XPM6 + 0.25RO2N + 0.75RO2R APX8 + OH. = 0.25XPM5 + 0.75XPM6 + 0.25RO2N + 0.75RO2R Reaction 35 Reaction 36 Reaction 37 Reactions 34-37 describe the formation of two types of semivolatile products from the oxidation reaction of p-xylene through two steps. The first step is the reaction of p-xylene with OH radical to form aerosol precursors, and the second step is the reaction of aerosol precursors with OH radical to produce semivolatile products. Table C5 summarizes stoichiometric coefficients in reactions 34-37, and cites sources of the data. Table C.5: Kinetic parameters of m-xylene reactions with OH radical forming aerosol products and their source Notations CRES BALD MGLY GLY HO2. RO2R APX6 APX7 APX8 XPM5 XPM6 RO2N XC Descriptions Cresol Benzaldehyde Methylglyoxal Glyoxal Hydroperoxy radical Operator RO2R Aerosol precursor APX6 Aerosol precursor APX7 Aerosol precursor APX8 Aerosol product species 1 Aerosol product species 2 Operator RO2N Balance extra carbon Stoichiometric Coefficients (SC) 0.21 0.037 0.564 0.188 0.21 0.789(reaction 34) 0.75 (reaction 35-37) 0.376 0.188 0.188 0.25 0.75 0.25 0.247 Source of SC SAPRC-99 SAPRC-99 Bandow, 1985 Bandow, 1985 SAPRC-99 Bandow, 1985 Estimated from o-xylene Bandow, 1985 Bandow, 1985 Bandow, 1985 Estimated from products from o-xylene, Eusebi, 1996 Estimated from products from o-xylene, Eusebi, 1996 Estimated from products from o-xylene, Eusebi, 1996 SAPRC-99, Bandow, 1985 239 Toluene Similar to other aromatic precursors, the oxidation reaction of toluene primarily occurs via addition and abstraction by OH radical. The hydrogen abstraction accounts for approximately 11% of all reactions (Bandow, 1985). The reaction proceeds through the abstraction of hydrogen atom by OH radical to form H2O , then the radical reacts further with O2 and with conversion of NO to NO2 and formation of HO2 radical to produce benzaldehyde (BALD). (45) 11%: TOLUENE + OH. = BALD + RO2R Reaction 38 When OH adds to the ring, the OH radicals add primarily to the ortho position. The addition of OH radical to toluene produces either cresol or nitrotoluene (NTOL), which accounts for 23.4 (SAPRC-99) and 25% (Bandow, 1985) of all reactions, respectively. While cresol is formed through the reaction of OH addition to toluene in the presence of O2 to form HO2 radical (pathway 46), nitrotoluene is generated via the addition of NO2 to OH-aromatic adduct and the formation of H2O (pathway 47). 240 (46) (47) 23.4%: TOLUENE + OH. = CRES + HO2. 25.0%: TOLUENE + OH. = NBEN + H2O NO2 + XC Reaction 39 Reaction 40 The remainder of OH-aromatic ring adduct leads to the formation of ring-cleavage products. The OH-aromatic adduct reacts with O2 to from bicyclo compounds, then with the abstraction of O atom by NO, alkoxyl-type radicals are formed. The alkoxyl-type radicals undergo hydrogen abstraction by O2 with the formation of HO2 radical, and the ring-cleavage product (denoted as aerosol precursors) are generated. Experimental results from Smith (1998) indicates that 23.8% of the addition of O2 to the OH-aromatic adduct resuts in glyoxal and unsaturated dicrabonyl compound, denoted as APTO1 in this work, and 16.7% results in methylglyoxal and APTO2 (pathways 48 and 49). 241 (48) (49) 23.8%: TOLUENE + OH. = APTO1 + RO2R + GLY 16.7%: TOLUENE + OH. = APTO2 + MGLY + RO2R Reaction 41 Reaction 42 Net: TOLUENE + OH. = 0.234CRES + 0.234HO2. + 0.11BALD + 0.238APTO1 + 0.167APTO2 + 0.238GLY + 0.167MGLY + 0.515RO2R + 0.25H2O + 0.25NBEN 0.25NO2 + 0.25XC Reaction 43 Aerosol precursors (APTO1 and APTO2) undergo on oxidation reaction with OH radical to form two types of semivolatile products (TOPM1 and TOPM2). Yields of TOPM1 (25%) and TOPM2 (75%) were estimated relatively from products for 1,3,5-trimethylbenzene from chamber experiements by Eusebi (1996). Reactions leading to formation of TOPM1 and TOPM2 are deplicted by reaction pathways 50a and 51a, and 50b and 51b, respectively. 242 (50a) (50b) (51a) (51b) APTO1 + OH. = 0.25TOPM1 + 0.75TOPM2 +0.25RO2N + 0.75RO2R APTO2 + OH. = 0.25TOPM1 + 0.75TOPM2 +0.25RO2N + 0.75RO2R Reaction 44 Reaction 45 Reactions 43-45 show the formation of semivolatile products for toluene which undergo two oxidation steps with OH radical. The stoichiometric coefficients and sources of the data are summarized in Table C6. 243 Table C.6: Kinetic parameters of toluene reactions with OH radical forming aerosol products and their source Notations CRES BALD MGLY GLY HO2. H2O NBEN RO2R APTO1 APTO2 TOPM1 TOPM2 RO2N XC Descriptions Cresol Benzaldehyde Methylglyoxal Glyoxal Hydroperoxy radical Water Nitrobenzene Operator RO2R Aerosol precursor APTO1 Aerosol precursor APTO2 Aerosol product species 1 Aerosol product species 2 Operator RO2N Extra carbon Stoichiometric Coefficients (SC) 0.234 0.11 0.167 0.238 0.234 0.25 0.25 0.515(reaction 43) 0.75 (reaction 44-45) 0.238 0.167 0.25 0.75 0.25 0.25 Source of SC SAPRC-99 Bandow, 1985 Bandow, 1985 Bandow, 1985 SAPRC-99 Bandow, 1985 Bandow, 1985 Bandow, 1985 Estimated from 1,3,5trimethylbenzene Smith, 1998 Smith, 1998 Estimated from 1,3,5trimethylbenzene Estimated from 1,3,5trimethylbenzene Estimated from 1,3,5trimethylbenzene Bandow, 1985 Benzene The photooxidation reaction of benzene predominantly occurs via the addition of OH radical. Experimental resutls from Bandow (1985) suggest that approximaetly 56.6% of all reactions result in nitrobenzene, 23.6% form phenol, and the remainder produce ring-cleavage products. The addition of NO2 to the OH-aromatic adduct, followed by the abstraction of hydrogen by O2 to form H2O lead to the formation of nitrobenzene (pathway 52). Phenol is produced via the abstraction of hydrogen by molecular oxygen from the OH-aromatic adduct (pathway 53). 244 (52) (53) 56.6%: Benzene + OH. = NBEN + H2O NO2 23.6%: Benzene + OH. = PHEN + HO2. Reaction 46 Reaction 47 The remainder of OH-aromatic adduct reacts with O2 to form on OH-O2-aromatic adduct. As for toluene, the OH-O2-aromatic adduct of benzene undergoes the cyclization reaction, with the conversion of NO to NO2 and formation of HO2 radical to produce a ring-cleavage product, denoted as APB1 (pathway 54). This pathway accounts for 20% of all reaction of benzene with OH radicals (Bandow, 1985 (54) 245 20%: BENZENE + OH. = APB1 + RO2R + GLY Reaction 48 Net: BENZENE + OH. = 0.236PHEN + 0.236HO2. +0.566NBEN + 0.566H2O 0.566NO2 + 0.2APB1 +0.2GLY + 0.2RO2R Reaction 49 Aerosol precursor (APB1) further reacts with OH radical to produce two types of condensable products (BPM1 and BPM2). The product with an organonitrate functional group is presented by BPM1, while the product with hydroxy carbonyl functional groups is represented by BPM2. BPM1 is formed through the addition of NO to OH-O2 aromatic adduct (pathway 55a). BPM2 is produced via the conversion of NO to NO2 and the abstraction of hydrogen atom by O2 from the OH-O-aromatic adduct (pathway 55b). Yields of BPM1 (25%) and BPM2 (75%) were relatively estimated from the products for 1,3,5-trimethylbenzene, which were obtained from chamber experiments (Eusebi, 1996). 246 (55a) (55b) APB1 + OH. = 0.25BPM1 + 0.75BPM2 +0.25RO2R +0.75RO2N Reaction 50 The condensed mechainsims of formation of semivolatile products (BPM1 and BPM2) for benzene is expressed by reactions 49 and 50. Table C7 summarizes the stoichiometric coefficients and cites sources of the data in those reactions. Table C.7: Kinetic parameters of benzene reactions with OH radical forming aerosol products and their source Notations PHEN GLY HO2. H2O NBEN RO2R APB1 BPM1 Descriptions Phenol Glyoxal Hydroperoxy radical Water Nitrobenzene Operator RO2R Aerosol precursor APB1 Aerosol product species 1 Stoichiometric Coefficients (SC) 0.236 0.20 0.236 0.566 0.566 0.20(reaction 49) 0.75 (reaction 50) 0.20 0.25 Source of SC SAPRC-99 Bandow, 1985 SAPRC-99 Bandow, 1985 Bandow, 1985 Bandow, 1985 Estimated from 1,3,5trimethylbenzene Bandow, 1985 Estimated from product for 1,3,5trimethylbenzene, Eusebi, 1996 247 1996 BPM2 Aerosol product species 2 0.75 RO2N Operator RO2N 0.25 Estimated from product for 1,3,5trimethylbenzene,Eusebi, 1996 Estimated from 1,3,5trimethylbenzene, Eusebi, 1996 Ethylbenzene Similar to toluene, the oxidation reactions of ethylbenzene primarily occur via the addition and abstraction by OH radical. The hydrogen abstraction accounts for approximately 11% of all reactions estimated from the yield of BALD for toluene, Bandow, 1985. The reaction proceeds through two possible pathways. The first pathway, hydrogen atom is abstracted by OH radical forming secondary radical and H2O, then the secondary radical reacts further with O2 and with coversion of NO to NO2 and formation of HO2 radical to produce benzaldehyde plus one carbon atom (BALD + 1XC) as shown in pathway 56. This pathway accounts for 5% of all reaction (estimated from toluene (Smith, 1998)). (56) 5%: C2-BENZ + OH. = BALD + 1XC + RO2R Reaction 51 248 The second possible pathway occurs through the abstaction of hydrogen atom to form a primary radical. The primary radical further oxidizes, along with conversion of NO to NO2 and HO2 radical formation to produce C6H5C2H3O. H atom from C6H5C2H3O is then abstracted by OH radical to form C6H5C2H2O. and H2O. C6H5C2H2O. undergoes decomposition to form C6H5CH2. by either two possible pathways as shown in the following figure. C6H5CH2. undergoes oxidization with the presence of NO to generate NO2, HO2 radical , and benzaldehyde (BALD). These reactions account for 6% of all reactions (estimated from toluene (Smith, 1998)), and are expressed by pathway 57. (57) 249 6%: C2-BENZ + 2OH. = BALD + 2RO2R + 0.5CO2 + 0.5HCHO + 2H2O Reaction 52 When OH is added to the ring, the OH radicals add primarily to the ortho position. Initial reaction with OH radical addition can produce cresol plus one extra carbon atom (CRES + 1XC) or nitrobenzene plus two extra carbon atoms (NBEN + 2XC) which account for 19% (SAPRC-99) and 29.6% (estimated from toluene (Smith, 1998)), respectively. (58) CRES + 1C (59) NBEN + 2C 19%: C2-BENZ + OH. = CRES +1XC + HO2. 29.6%: C2-BENZ + OH. = NBEN +2XC + H2O NO2 Reaction 53 Reaction 54 The remainder of the OH reactions further oxidizes produce ring-cleavage products. Ring-cleavage products are produced through four possible pathways (pathways 6062). All of these four possible pathway proceed through similar steps, the abstraction of O atom by NO forming NO2, and the abstraction of H atom by O2. The cleavage 250 reactions produce 16.7% APEB1 and MGLY +1C, 15.8% APEB2, 7.9% of APEB3, and 23.7% GLY (estimated from toluene (Smith, 1998)). (60) (61a) (61b) (62) 40.4 % C2-BENZ + OH. = 0.167APEB1 + 0.158APEB2 + 0.079APEB3 + Reaction 55 0.167MGLY + 0.167C + 0.237GLY + 0.404RO2R Net: C2-BENZ + OH. = 0.11BALD + 0.03CO2 0.06OH. + 0.03HCHO +0.19CRES + 0.296NBEN + 0.416H2O 0.296NO2 + 0.19HO2. + 0.167APEB1 + 0.167MGLY + 0.158APEB2 + 0.079APEB3 + 0.237GLY +0.574RO2R + 0.999XC Reaction 56 251 Aerosol precursors (APEB1-APEB3) react further to generate two types of condensable products (EBPM1 and EBPM2). The addition of OH radical occurs at carbon double bonds to form dicarbonyl radicals. Dicarbonyl radicals either react with NO forming EBPM1 (pathway 63a-65a), or further oxidizes in the presence of NO to generate NO2, HO2 radical, and EBPM2 (pathway 63b-65b). Yields of EBPM1 (25%) and EBPM2 (75%) were estimated from semivolatile products for 1,3,5-trimethylbenzene. (63a) (63b) 252 (64a) (64b) (65a) (65b) APEB1 + OH. = 0.25EBPM1 +0.75EBPM2 +0.25RO2N + 0.75RO2R APEB2 + OH. = 0.25EBPM1 +0.75EBPM2 +0.25RO2N + 0.75RO2R APEB3 + OH. = 0.25EBPM1 +0.75EBPM2 +0.25RO2N + 0.75RO2R Reaction 57 Reaction 58 Reaction 59 The stoichiometric coefficients in reactions 56-59 for ethyl benzene are summarized in Table 8. 253 Table C.8: Kinetic parameters of ethylbenzene reactions with OH radical forming aerosol products and their source Notations CRES BALD MGLY GLY HO2. H2O NBEN CO2 HCHO RO2R Descriptions Cresol Benzaldehyde Methylglyoxal Glyoxal Hydroperoxy radical Water Nitrobenzene Carbon dioxide Formaldehyde Operator RO2R Stoichiometric Coefficients (SC) 0.19 0.11 0.167 0.237 0.19 0.416 0.296 0.03 0.03 0.574(reaction 56) 0.75 (reaction 57-59) APEB1 APEB2 APEB3 EBPM1 Aerosol precursor APEB1 Aerosol precursor APEB2 Aerosol precursor APEB3 Aerosol product species 1 0.167 0.158 0.079 0.25 Source of SC SAPRC-99 Estimated from toluene, Bandow, 1985 Estimated from product for toluene, Smith 1998 Estimated from product for toluene, Smith 1998 SAPRC-99 Estimated from product for toluene, Smith 1998 Estimated from product for toluene, Smith 1998 Estimated from product for toluene, Smith 1998 Estimated from product for toluene, Smith 1998 Estimated from product for toluene, Smith 1998 Estimated from 1,3,5trimethylbenzene (Eusebi, 1996) Estimated from product for toluene, Smith 1998 Estimated from product for toluene, Smith 1998 Estimated from product for toluene, Smith 1998 Estimated from product for 1,3,5trimethylbenzene (Eusebi, 1996) Estimated from product for 1,3,5trimethylbenzene (Eusebi, 1996) Estimated from 1,3,5trimethylbenzene (Eusebi, 1996) SAPRC-99 EBPM2 Aerosol product species 2 0.75 RO2N XC Operator RO2N Balance carbon atom 0.25 0.999 254 n-propylbenzene The oxidation reactions of n-propylbenzene are primarily initiated by OH radical. Reaction can either proceed through the addition of OH to the aromatic ring or the abstraction of an H atom by OH radical. As for the reactions for ethylbenzene, the abstraction of H atom by OH radical for n-propylbenzene can occur via two possible pathways. In the first pathway, the abstraction can occur at the carbon adjacent to the aromatic ring, then oxidation occurs in the presence of NO to form benzaldehyde plus two extra carbon atoms (BALD + 2XC) (pathway 66). Yiled of this pathway (5%) was estimated from that for toluene (Smith, 1998). (66) 5%: N-C3-BEN + OH. = BALD + 2XC + RO2R Reaction 60 The second pathway for the abstraction proceeds similary to that for ethylbenzene. H atom at the methyl substituent group is abstracted by OH radical. The oxidization then occurs in the presence of NO to form NO2, HO2 radical, and C6H5C3H5O. The abstraction of H atom from carboxylic acid group by OH radical occurs, followed by the decomposition to form either CO2 or OCH2 with polysubstituent-aromatic radical. The polysubstituent-aromatic radical undergoes oxidization in the presence of NO to form NO2, HO2 radical, and C6H5C2H3O. The consecutive reactions of H atom 255 abstraction, decomposition, and the oxidization then occur again to finally form benzaldehyde. Pathway 67 elaborates the steps of reactions. Yield of this pathway (6%) was estimated from that for toluene (Smith, 1998). (67) 6%: N-C3-BEN + OH. = BALD + 3RO2R + CO2 + HCHO - 2OH.+ 3H2O Reaction 61 When OH adds to the ring, the OH radicals add primarily to the ortho position. Initial reaction with OH radical addition can produce cresol plus two extra carbon atoms (CRES + 2XC) or nitrobenzene plus three extra carbon atoms (NBEN + 3XC) which account for 19% (SAPRC-99) and 29.6% (estimated from toluene (Smith, 1998)), respectively. 256 (68) (69) 19%: N-C3-BEN + OH. = CRES + 2XC + HO2. 29.6%: N-C3-BEN + OH. = NBEN + 3XC + H2O NO2 Reaction 62 Reaction 63 The remainder of the OH additional reaction further oxidizes producing ring-cleavage products. Ring-cleavage products are produced through four possible pathways (pathways 70-72). All of these four possible pathway proceed through two similar steps, the abstraction of O atom by NO forming NO2, and the abstraction of H atom by O2. The cleavage reactions produce each 16.7% of APPB1 and MGLY + 2XC, 15.8% of APPB2, 7.9% of APPB3, and 23.7% GLY (relatively estimated from toluene (Smith, 1998)). (70) 257 (71a) (71b) (72) N-C3-BEN + OH. = 0.167APPB1 + 0.167MGLY + 0.167(2XC) + 0.158APPB2 + 0.079APPB3 + 0.238GLY + 0.404RO2R Reaction 64 Net: N-C3-BEN + OH. = 0.11BALD + 0.06CO2 + 0.06HCHO 0.12OH. + 0.19CRES + 0.19HO2. + 0.296NBEN - 0.296NO2 +0.167APPB1 + 0.158APPB2 + 0.079APPB3 + 0.238GLY + 0.634RO2R + 0.476H2O + 1.702XC + 0.167MGLY Reaction 65 Aerosol precursors (APPB1-APPB3) react further with OH radical to form two types of condensable products (PBPM1 and PBPM2). PBPM1 is a product with organonitrate group, PBPM2 is product with hydroxy carbonyl groups. Reactions leading to formation of PBPM1 and PBPM2 from aerosol precursors are shown in 258 pathways 73-75. Yields of PBPM1 (25%) and PBPM2 (75%) were estimated from those for 1,3,5-trimethylbenzene, obtained from chamber experiments (Eusebi, 1996). (73a) (73b) (74a) (74b) 259 (75a) (75b) APPB1 +OH. = 0.25PBPM1 +0.75PBPM2 +0.25RO2N + 0.75RO2R APPB2 +OH. = 0.25PBPM1 +0.75PBPM2 +0.25RO2N + 0.75RO2R APPB3 +OH. = 0.25PBPM1 +0.75PBPM2 +0.25RO2N + 0.75RO2R Reaction 66 Reaction 67 Reaction 68 The stoichiometric coefficients and sources of the data in reactions 65-68 are summarized in Table 9. Table C.9: Kinetic parameters of n-propylbenzene reactions with OH radical forming aerosol products and their source Notations CRES BALD MGLY GLY HO2. H2O Descriptions Cresol Benzaldehyde Methylglyoxal Glyoxal Hydroperoxy radical Water 260 Stoichiometric Coefficients (SC) 0.19 0.11 0.167 0.238 0.19 0.476 Source of SC SAPRC-99 Estimated from toluene, Bandow, 1985 Estimated from product for toluene, Smith 1998 Estimated from product for toluene, Smith 1998 SAPRC-99 Estimated from product NBEN CO2 HCHO RO2R Nitrobenzene Carbon dioxide Formaldehyde Operator RO2R 0.296 0.06 0.06 0.634(reaction 65) 0.75 (reaction 66-68) APPB1 APPB2 APPB3 PBPM1 PBPM2 RO2N XC Aerosol precursor APPB1 Aerosol precursor APPB2 Aerosol precursor APPB3 Aerosol product species 1 Aerosol product species 2 Operator RO2N Balance carbon atom 0.167 0.158 0.079 0.25 0.75 0.25 1.702 for toluene, Smith 1998 Estimated from product for toluene, Smith 1998 Estimated from product for toluene, Smith 1998 Estimated from product for toluene, Smith 1998 Estimated from product for toluene, Smith 1998 Estimated from 1,3,5trimethylbenzene Estimated from product for toluene, Smith 1998 Estimated from product for toluene, Smith 1998 Estimated from product for toluene, Smith 1998 Estimated from product for 1,3,5trimethylbenzene Estimated from product for 1,3,5trimethylbenzene Estimated from toluene, Eusebi 1996 SAPRC-99 Isopropylbenzene Photochemical reactions of iso-propylbenzene leading to the formation of semivolatile products are similar to those for n-propylbenzene. The oxidation is initiated by OH radical. Reaction can either proceed through the addition of OH to aromatic ring or the abstraction of H atom by OH radical. The abstractions predominantly occur at either one of the methyl groups, then undergoes oxidation with the conversion of NO to NO2 and formation of HO2 radical forming C6H5C3H5O. The abstraction of H atom from the carbonyl group by OH radical occurs again. The 261 product from this reaction then proceeds through the consecutive reactions of decomposition generating CO2, and oxidization in the presence of NO to produce NO2, HO2 radical, and benzaldehyde plus one extra carbon atom (BALD +1XC). Yield of this pathway (11%) was relatively estimated from that for toluene (Smith, 1998). (76) 11%: I-C3-BEN + OH. = BALD + 1XC + 2RO2R + CO2 OH. + 2H2O Reaction 69 The addition of OH radicals to the aromatic ring primirily occurs at the ortho position. Initial reaction of iso-propylbenzene with OH radical addition can produce cresol plus two extra carbon atoms (CRES + 2XC) or nitrobenzene plus three extra carbon atoms (NBEN + 3XC) which account for 19% (SAPRC-99) and 29.6% (estimated from toluene (Smith, 1998)), respectively. 262 (77) (78) 19%: I-C3-BEN + OH. = CRES + 2XC + HO2. 29.6%: I-C3-BEN + OH. = NBEN + 3XC + H2O NO2 Reaction 70 Reaction 71 Another reaction pathway that can occur following the formation of the OH-aromatic adduct is bridging of O2 across the ring, leading to ring cleavage reactions. Ringcleavage products are produced through four possible pathways (pathways 79-81). All of these four possible pathways proceed through two similar steps, the abstraction of O atom by NO forming NO2, and the abstraction of H atom by O2. The cleavage reactions produce 16.7% APB2 and MGLY + 2XC, 15.8% APB3, 7.9% APB4, and 23.7% GLY (estimated from toluene (Smith, 1998)). (79) 263 (80a) (80b) (81) 40.4 % I-C3-BEN + OH. = 0.167APB2 + 0.158APB3 + 0.079APB4 + 0.167MGLY + 0.334XC + 0.404RO2R Reaction 72 Net: I-C3-BEN + OH. = 0.11BALD + 0.11CO2 0.11OH. + 0.19CRES + 0.19HO2. + 0.167APB2 + 0.167MGLY +0.158APB3 + 0.079APB4 + 0.237GLY + 0.624RO2R 0.296NO2 + 0.516H2O + 1.712XC + 0.296NBEN Reaction 73 Aerosol precursors (APB2-APB4) react further with OH radical to form two types of condensable products (BPM3 and BPM4). BPM3 is a product with an organonitrate group, BPM4 is product with hydroxy carbonyl groups. Reactions leading to formation of BPM3 and BPM4 from aerosol precursors are shown in pathways 82-84. Yields of PBPM1 (25%) and PBPM2 (75%) were relatively estimated from those for 1,3,5-trimethylbenzene, obtained from chamber experiments (Eusebi, 1996). 264 (82a) (82b) (83a) (83b) 265 (84a) (84b) APB2 + OH. = 0.25BPM3 +0.75BPM4 +0.25RO2N + 0.75RO2R APB3 + OH. = 0.25BPM3 +0.75BPM4 +0.25RO2N + 0.75RO2R APB4 + OH. = 0.25BPM3 +0.75BPM4 +0.25RO2N + 0.75RO2R Reaction 74 Reaction 75 Reaction 76 Reactions 73-76 show the mechanistic models leading to SOA formation for iso-propylbenzene which proceed via two steps, similar to other aromatic precursors. Table C10 lists stoichiometric coefficients in reactions 73-76, and cite sources of the data. Table C.10: Kinetic parameters of iso-propylbenzene reactions with OH radical forming aerosol products and their source Notations CRES BALD MGLY GLY Descriptions Cresol Benzaldehyde Methylglyoxal Glyoxal 266 Stoichiometric Coefficients (SC) 0.19 0.11 0.167 0.237 Source of SC SAPRC-99 Estimated from toluene, Bandow, 1985 Estimated from product for toluene, Smith 1998 Estimated from product for toluene, Smith 1998 HO2. H2O NBEN CO2 RO2R Hydroperoxy radical Water Nitrobenzene Carbon dioxide Operator RO2R 0.19 0.516 0.296 0.11 0.624(reaction 73) 0.75 (reaction 74-76) APB2 APB3 APB4 BPM3 BPM4 RO2N XC Aerosol precursor APB2 Aerosol precursor APB3 Aerosol precursor APB4 Aerosol product species 1 Aerosol product species 2 Operator RO2N Balance carbon atom 0.167 0.158 0.079 0.25 0.75 0.25 1.712 SAPRC-99 Estimated from product for toluene, Smith 1998 Estimated from product for toluene, Smith 1998 Estimated from product for toluene, Smith 1998 Estimated from product for toluene, Smith 1998 Estimated from 1,3,5trimethylbenzene Estimated from product for toluene, Smith 1998 Estimated from product for toluene, Smith 1998 Estimated from product for toluene, Smith 1998 Estimated from product for 1,3,5trimethylbenzene Estimated from product for 1,3,5trimethylbenzene Estimated from 1,3,5trimethylbenzene SAPRC-99 Sec-butylbenzene Photooxidation reactions of sec-butylbenzene are initiated by OH radical. Reactions either occur via addition or abstraction. Abstraction of H atom by OH radical can proceed through two possible pathways. In the first pathway, H atom at one of the four of alkyl groups is abstracted by OH radical, then oxidation occurs with conversion of NO to NO2 and formation of HO2 radical forming C6H5C4H7O. The abstraction of H atom from carbonyl group of C6H5C4H7O by OH radical occurs again. The product from this reaction then proceeds through consecutive reactions of 267 decomposition producing CO2, and oxidization in the presence of NO to produce NO2, HO2 radical, and C6H5C3H5O. Consecutive reactions of H atom abstraction, decomposition, and oxidation are repeated again to produce benzaldehyde plus one extra carbon atom (BALD +1XC), pathway 85. Yield of this pathway (5%) was estimated from that for toluene (Smith, 1998). (85) 5%: S-C4-BEN + 3OH. = BALD + 1XC + 3RO2R + 3H2O + 2CO2 Reaction 77 268 In the second pathway, the abstraction of H atom occurs at position two (secondary carbon) of polysubstituent group (pathway 86). Proceeding through the similar consecutive reactions as in first pathway, benzaldehyde plus two extra carbon atoms, NO2, and HO2 radical are formed (pathway 86). (86) 6%: S-C4-BEN + OH. = BALD + 2C + 2RO2R + CO2 OH. + 2H2O Reaction 78 The addition of OH radicals primarily occur at the othro position. Initial addtion of OH radical to sec-butylbenzene produces 19% cresol plus three extra carbon atoms (pathway 87, Smith, 1998), and 29.6% nitrobenzene plus four extra carbon atoms (pathway 88, Smith, 1998). (87) 269 (88) 19%: S-C4-BEN + OH. = CRES + 3XC + HO2. 29.6%: S-C4-BEN + OH. = NBEN + 4XC + H2O NO2 Reaction 79 Reaction 80 The remainder of OH-aromatic adduct undergoes ring-cleavage reactions, which can proceed through six possible pathways (pathway 89-91). All of these six possible pathway occur via two steps, the abstraction of O atom by NO forming NO2, and the abstraction of H atom by O2. The cleavage reactions produce 16.7% APB5 and MGLY + 2XC, 11.9% APB6, 11.9% APB7, and 23.7% GLY (estimated from toluene (Smith, 1998)). (89a) (89b) 270 (90a) (90b) (91a) (91b) 40.5% S-C4-BEN + OH. = 0.167APB5 + 0.167MGLY + 0.167(3)XC +0.119APB6 + 0.119APB7 + 0.238GLY + 0.404RO2R Reaction 81 Net: S-C4-BEN + OH. = 0.19CRES + 0.19HO2. + 0.11BALD + 0.296NBEN 0.296NO2 + 0.167APB5 + 0.119APB6 + 0.119APB7 + 0.167MGLY + 0.238GLY + 0.675RO2R + 0.566H2O 0.16OH. + 2.425XC + 0.16CO2 Reaction 82 Aerosol precursors (APB5-APB7) further react with OH radical. OH radical adds to the carbon double bonds. Two types of semivolatile products (BPM5 and 271 BPM6) are produced from reactions of aerosol precursors. BPM5 is produced through the addition of NO to the O2-OH aromatic adduct (pathways 92a-94a). BPM6 is generated via the abstraction of O atom by NO forming NO2, and followed by the abstraction of H atom by O2 producing HO2 radical (pathways 92b 94b). Yields of these two products, 25% for BPM5 and 75% for BPM6 were estimated from those for 1,3,5-trimethylbenzene (Eusebi, 1996). (92a) (92b) 272 (93a) (93b) (94a) (94b) APB5 + OH. = 0.25BPM5 + 0.75BPM6 + 0.25RO2N +0.75RO2R 273 Reaction 83 APB6 + OH. = 0.25BPM5 + 0.75BPM6 + 0.25RO2N +0.75RO2R APB7 + OH. = 0.25BPM5 + 0.75BPM6 + 0.25RO2N +0.75RO2R Reaction 84 Reaction 85 Reactions 82-85 describe the mechanistic models leading to SOA formation for iso-propylbenzene which proceed via two steps, similar to other aromatic precursors. Table 11 enumerates stoichiometric coefficients in reactions 82-85, and cites sources of the data. Table C.11: Kinetic parameters of sec-butylbenzene reactions with OH radical forming aerosol products and their source Notations CRES BALD MGLY GLY HO2. H2O NBEN CO2 RO2R Descriptions Cresol Benzaldehyde Methylglyoxal Glyoxal Hydroperoxy radical Water Nitrobenzene Carbon dioxide Operator RO2R Stoichiometric Coefficients (SC) 0.19 0.11 0.167 0.238 0.19 0.566 0.296 0.16 0.675(reaction 82) 0.75 (reaction 83-85) APB5 APB6 APB7 BPM5 Aerosol precursor APB5 Aerosol precursor APB6 Aerosol precursor APB7 Aerosol product species 1 0.167 0.119 0.119 0.25 Source of SC SAPRC-99 Estimated from toluene, Bandow, 1985 Estimated from product for toluene, Smith 1998 Estimated from product for toluene, Smith 1998 SAPRC-99 Estimated from product for toluene, Smith 1998 Estimated from product for toluene, Smith 1998 Estimated from product for toluene, Smith 1998 Estimated from product for toluene, Smith 1998 Estimated from 1,3,5trimethylbenzene (Eusebi, 1996) Estimated from product for toluene, Smith 1998 Estimated from product for toluene, Smith 1998 Estimated from product for toluene, Smith 1998 Estimated from product for 1,3,5trimethylbenzene, (Eusebi 1996) Estimated from product for 1,3,5- BPM6 Aerosol product species 2 0.75 274 RO2N XC Operator RO2N Balance carbon atom 0.25 2.425 trimethylbenzene (Eusebi, 1996) Estimated from 1,3,5trimethylbenzene, Eusebi 1996 SAPRC-99 p-ethyltoluene Similar to photooxidation reactions of p-xylene, reactions of p-ethyltoluene are predominantly initiated by OH radical. Reactions with OH radical can either proceed through abstraction of H atom, or addition of OH. Abstraction by OH radical is likely to occur via two possible pathways. In the first pathway, H atom at the position terminal methyl group of the ethyl chain is abstracted by OH radical, then the oxidation occurs with the conversion of NO to NO2 and formation of HO2 radical, forming C6H5C2H3OCH3 (pathway 95) . The abstraction of H atom from the carbonyl group of C6H5C2H3OCH3 by OH radical occurs again. The product from this reaction then proceeds through the consecutive reactions of decomposition forming CO2, and oxidization in the presence of NO to produce NO2, HO2 radical, and benzaldehyde plus one extra carbon atom (pathway 95). estimated from p-xylene. Yield of this pathway (4.2%) was 275 (95) 4.2%: P-C2-TOL + OH. = BALD + 1XC + CO2 + 2RO2R + 2H2O OH. Reaction 86 In another abstraction pathway, an H atom at the alpha position in the ethyl group is extracted. The structure with secondary radical is oxidized in the presence of NO generating NO2, HO2 radical, and benzaldehyde plus two extra carbon atoms (pathway 96). This pathway accounts for 4.2% of all reactions (relatively estiamted from p-xylene). (96) 4.2%: P-C2-TOL + OH. = BALD + 2XC + RO2R Reaction 87 276 Addition of the OH radical to the p-ethyltoluene ring can occur at multiple ring positions, but predominantly adds at the aromatic ring carbon ortho to one of the cabons attached to the methyl group (Grovenstein, 1970; Atkinson et al., 1980). The initial addition of OH radical to p-ethyltoluene partially results in cresol with two extra carbon atoms and HO2 radical (pathway 97). This pathway acounts for 18.8% of all reaction (estimated from p-xylene). (97) 18.8%: P-C2-TOL + OH. = CRES + 2XC + HO2. Reaction 88 The remainder of the OH-aromatic adduct reacts further forming ring-cleavage products. Ring-cleavage reactions can proceed through five likely pathways, each undergoes two steps. The first step is the addition of O2 to the OH-aromatic radical, and the abstraction of O atom by NO forming NO2. The second step is the abstraction of H atom by O2 to produce HO2 radical and ring-cleavage products (unsaturated dicarbonyl species denoted as aerosol precursors). Estimates of yields of products for p-xylene suggest that 49% of all reactions produce aerosol precursor species APET1, 20% result in APET2, and 4% form APET3 (pathways 98-100). 277 Methylglyoxal and glyoxal are also generated with the yields of 24% and 49%, respectively (pathways 98-100). (98a) (99a) 278 (99b) (100) 73.0% P-C2-TOL + OH. = 0.49APET1 + 0.2APET2 +0.04APET3 + 0.24MGLY + 0.49GLY + 0.73RO2R + 0.2XC Reaction 89 Net: P-C2-TOL + OH. = 0.188CRES + 0.188HO2. + 0.084BALD + 0.042CO2 + 0.084H2O + 0.49APET1 + 0.2APET2 + 0.04APET3 + 0.24MGLY + 0.49GLY + 0.702C + 0.856RO2R 0.0402OH. Reaction 90 Aerorol precursors (APET1-APET3) further react with OH radical in the presence of O2 and NO2 to form two types of condensable products (ETPM1 and ETPM2). 279 ETPM1, aerosol products with organonitrate, is produced through first the addition of OH to carbon double bonds, then the addition of O2 and NO2 to the radical (pathways 101a 103a). This pathway accounts for 25% (estimated from o-xylene). The remainder of the OH-ring clevage adduct (75%) further oxidizes in the presence of NO forming NO2, HO2 radical, and ETPM2 (pathways 101b-103b). (101a) (101b) (102a) (102b) 280 (103a) (103b) APET1 + OH. = 0.25ETPM1 +0.75ETPM2 +0.25RO2N +0.75RO2R APET2 + OH. = 0.25ETPM1 +0.75ETPM2 +0.25RO2N +0.75RO2R APET3 + OH. = 0.25ETPM1 +0.75ETPM2 +0.25RO2N +0.75RO2R Reaction 91 Reaction 92 Reaction 93 Reactions 90-93 describe condensed mechanisms of p-ethyltoluene with OH radical leading to SOA formation. The stoichiometric coefficients are summarized in Table C12. Table C.12: Kinetic parameters of p-ethyltoluene reactions with OH radical forming aerosol products and their source Notations CRES BALD MGLY GLY HO2. RO2R Descriptions Cresol Benzaldehyde Methylglyoxal Glyoxal Hydroperoxy radical Operator RO2R Stoichiometric Coefficients (SC) 0.188 0.084 0.24 0.49 0.188 0.856(reaction 90) 0.75 (reaction 91-93) 281 Source of SC Estimated from p-xylene (Eusebi, 1996) Estimated from p-xylene (Eusebi, 1996) Estimated from p-xylene (Eusebi, 1996) Estimated from p-xylene (Eusebi, 1996) Estimated from p-xylene (Eusebi, 1996) Estimated from p-xylene (Eusebi, 1996) Estimated from o-xylene (Eusebi, 1996) APET1 APET2 APET3 ETPM1 ETPM2 RO2N XC CO2 H2O Aerosol precursor APET1 Aerosol precursor APEt2 Aerosol precursor APET3 Aerosol product species 1 Aerosol product species 2 Operator RO2N Balance extra carbon Carbon dioxide Water 0.49 0.20 0.04 0.25 0.75 0.25 0.702 0.042 0.084 Estimated from p-xylene (Eusebi, 1996) Estimated from p-xylene (Eusebi, 1996) Estimated from p-xylene (Eusebi, 1996) Estimated from products from o-xylene (Eusebi, 1996) Estimated from products from o-xylene (Eusebi, 1996) Estimated from products from o-xylene (Eusebi, 1996) SAPRC-99, estimated from p-xylene, (Eusebi, 1996) Estimated from p-xylene (Eusebi, 1996) Estimated from p-xylene (Eusebi, 1996) o-ethyltoluene Condensed mechanisms describing SOA formation from o-ethyltoluene were developed analogous to those of other aromatics. Reactions of o-ethyltoluene with OH radical can either proceed through the abstraction of H atom, or the addition of OH radical to aromatic ring. Abstraction of H atom by OH radical is likely to occur via two possible pathways. The first pathway, H atom at position terminal methyl group in the ethyl chain is abstracted by OH radical, then the oxidation occurs with the conversion of NO to NO2 and formation of HO2 radical forming C6H5C2H3OCH3 (pathway 104) . The abstraction of H atom from the carbonyl group of C6H5C2H3OCH3 by OH radical occurs again. The product from this reaction then proceeds through the consecutive reactions of decomposition producing CO2, and 282 oxidization in the presence of NO to produce NO2, HO2 radical, and benzaldehyde plus one extra carbon atom (pathway 104). Yield of this pathway (3.25%) was estimated from o-xylene. (104) 3.25%: O-C2-TOL + OH. = BALD + 1XC + CO2 + 2RO2R + 2H2O OH. Reaction 94 The second pathway involves abstraction of H atom at the alpha-carbon in the ethyl chain. The radical is oxidized in the presence of NO generating NO2, HO2 radical, and benzaldehyde plus two extra carbon atoms (pathway 105). accounts for 3.25% of all reactions (estiamted from o-xylene). This pathway (105) 3.25%: O-C2-TOL + OH. = BALD + 2XC + RO2R 283 Reaction 95 Addition of OH radical predominantly occurs at the aromatic carbon ortho to the carbon attached to the methyl group. The initial addition of OH radical to o- ethyltoluene results in cresol with two extra carbon atoms and HO2 radical (pathway 106). This pathway acounts for 18% of all reaction (estimated from o-xylene). (106) 18.8%: O-C2-TOL + OH. = CRES + 2XC + HO2. Reaction 96 Another reaction pathway that can occur following the formation of the OH-aromatic adduct is bridging of O2 atom across the ring. This reaction then leads to the formation of ring-cleavage products. Ring-cleavage reactions can proceed through 14 likely pathways, each of them undergoes two similar steps. The first step is the addition of O2 to the OH-aromatic radical, and the abstraction of O atom by NO forming NO2. The second step is the abstraction of H atom by O2 to produce HO2 radical and ring-cleavage products (unsaturated dicarbonyl species denoted as aerosol precursors). The yield of aerosol precursor species APET4 is 32.9%, APET5 is 8.5%, APET6 is 2.6%, APET7 is 5.1%, APET8 is 18.6%, and APET9 is 6.6%. These yields were estimated from yields of products for o-xylene. Along with the production of 284 aerosol precursors, methylglyoxal, biacetyl, and glyoxal are also generated with the yields of 39.5, 18.6, and 16.2%, respectively (pathways 107-112). (107a) (107b) 285 (108a) (108b) 286 (108c) (109a) 287 (109b) (110a) (110b) 288 (110c) (111a) 289 (111b) (112a) 290 (112b) 74.29% O-C2-TOL + OH. = 0.329APET4 + 0.085APET5 +0.026APET6 +0.051APET7 +0.186APET8 +0.0659APET9 + 0.395MGLY + 0.186BACL +0.162GLY + 0.7429RO2R + 0.252XC Reaction 97 Net: O-C2-TOL + OH. = 0.188CRES + 0.188HO2. + 0.065BALD + 0.0325CO2 + 0.065H2O + 0.329APET4 + 0.085APET5 +0.026APET6 +0.051APET7 +0.186APET8 +0.0659APET9 + 0.395MGLY + 0.186BACL +0.162GLY + 0.7255XC + 0.8404RO2R 0.0325HO. Reaction 98 Aerorol precursors (APET4-APET9) further react with OH radical in the presence of O2 and NO2 to form two types of condensable products (ETPM3 and ETPM4). ETPM3, containing organonitrate, is produced through the addition of OH to carbon double bonds, then the addition of O2 and NO2 to the radical (pathways 113a 118a). This pathway accounts for 25% (estimated from o-xylene) of the products. The 291 remainder of the OH-ring clevage adduct (75%) further oxidizes in the presence of NO forming NO2, HO2 radical, and ETPM4 (pathways 113b-118b). (113a) (113b) (113b) (114a) (114b) (115a) (115b) 292 (116a) (116b) (117a) (117b) (118a) (118b) 293 APET4 + OH. = 0.25ETPM3 +0.75ETPM4 +0.25RO2N +0.75RO2R Reaction 99 APET5 + OH. = 0.25ETPM3 +0.75(2MGLY + 1C) +0.25RO2N +0.75RO2R Reaction 100 APET6 + OH. = 0.25ETPM3 +0.75ETPM4 +0.25RO2N +0.75RO2R APET7 + OH. = 0.25ETPM3 +0.75ETPM4 +0.25RO2N +0.75RO2R APET8 + OH. = 0.25ETPM3 +0.75ETPM4 +0.25RO2N +0.75RO2R APET9 + OH. = 0.25ETPM3 +0.75ETPM4 +0.25RO2N +0.75RO2R Reaction 101 Reaction 102 Reaction 103 Reaction 104 Reactions 98-104 describe condensed mechanisms of o-ethyltoluene with OH radical leading to SOA formation. The stoichiometric coefficients are summarized in Table 13. Table C.13: Kinetic parameters of o-ethyltoluene reactions with OH radical forming aerosol products and their source Notations CRES BALD MGLY BACL GLY HO2. RO2R Descriptions Cresol Benzaldehyde Methylglyoxal Biacetyl Glyoxal Hydroperoxy radical Operator RO2R Stoichiometric Coefficients (SC) 0.188 0.065 0.395 0.186 0.162 0.188 0.8404(reaction 98) 0.75 (reaction 99-104) APET4 APET5 APET6 APET7 Aerosol precursor APET4 Aerosol precursor APET5 Aerosol precursor APET6 Aerosol precursor APET7 294 0.329 0.085 0.026 0.051 Source of SC Estimated from o-xylene (Eusebi, 1996) Estimated from o-xylene (Eusebi, 1996) Estimated from o-xylene (Eusebi, 1996) Estimated from o-xylene (Eusebi, 1996) Estimated from o-xylene (Eusebi, 1996) Estimated from o-xylene (Eusebi, 1996) Estimated from o-xylene (Eusebi, 1996) Estimated from o-xylene (Eusebi, 1996) Estimated from o-xylene (Eusebi, 1996) Estimated from o-xylene (Eusebi, 1996) Estimated from o-xylene (Eusebi, 1996) Estimated from o-xylene (Eusebi, 1996) APET8 APET9 ETPM3 ETPM4 RO2N XC CO2 H2O Aerosol precursor APET8 Aerosol precursor APET9 Aerosol product species 1 Aerosol product species 2 Operator RO2N Balance extra carbon Carbon dioxide Water 0.186 0.066 0.25 0.75 0.25 0.7255 0.0325 0.065 Estimated from o-xylene (Eusebi, 1996) Estimated from o-xylene (Eusebi, 1996) Estimated from products from o-xylene (Eusebi, 1996) Estimated from products from o-xylene (Eusebi, 1996) Estimated from products from o-xylene (Eusebi, 1996) SAPRC-99, estimated from o-xylene (Eusebi, 1996) Estimated from o-xylene (Eusebi, 1996) Estimated from o-xylene (Eusebi, 1996) m-ethyltoluene Condensed mechanisms describing SOA formation of m-ethyltoluene were developed analogous to those used for other aromatics. Reactions of m-ethyltoluene with OH radical can either proceed through abstraction, or addition. Similar ring to p-and o-ethyltoluene, abstraction of H atom by OH radical is likely to occur via two possible pathways. In the first pathway, H atom from the methyl group in the ethyl chain is abstracted by OH radical, then the oxidation occurs with the conversion of NO to NO2 and formation of HO2 radical forming C6H5C2H3OCH3 (pathway 119). The abstraction of H atom from the carbonyl group of C6H5C2H3OCH3 by OH radical occurs. The product from this reaction then proceeds through the consecutive reactions of decomposition producing CO2, and oxidization in the presence of NO to 295 produce NO2, HO2 radical, and benzaldehyde plus one extra carbon atom (pathway 119). Yield of this pathway (1.85%) was estimated from m-xylene. (119) 1.85%:M-C2-TOL + OH. = BALD + 1XC + CO2 + 2RO2R + 2H2O OH. Reaction 105 The second pathway involves abstraction of H. The aromatic-radical reacts with O2 in the presence of NO generating NO2, HO2 radical, and benzaldehyde plus two extra carbon atoms (pathway 120). This pathway accounts for 1.85% of all reactions (estimated from m-xylene). (120) 1.85%: M-C2-TOL + OH. = BALD + 2XC + RO2R 296 Reaction 106 Addition of OH radical occurspredominantly adds at the aromatic carbon ortho to the methyl group. The initial addition of OH radical to o-ethyltoluene results in cresol with two extra carbon atoms and HO2 radical (pathway 121). This pathway acounts for 21% of all reactions (estimated from m-xylene). (121) 21%: M-C2-TOL + OH. = CRES + 2XC + HO2. Reaction 107 Another reaction pathway that can occur following the formation of the OH-aromatic adduct is bridging of O2 across the ring, leading to the formation of ring-cleavage products. Ring-cleavage reactions proceed through 12 pathways, each of them undergoes two steps. The first step is the addition of O2 to the OH-aromatic radical, and the abstraction of O atom by NO forming NO2. The second step is the abstraction of H atom by O2 to produce HO2 radical and ring-cleavage products (unsaturated dicarbonyl species denoted as aerosol precursors). The yield of aerosol precursor species APET10 is 5.4%, APET11 is 13.4%, APET12 is 8%, APET13 is 10.6%, APET14 is 22.6%, and APET15 is 15.3%. These yields were estimated from yields of products for m-xylene. Ring-cleavage reactions also produce 297 methylglyoxal, and glyoxal with the yields of 57.7 and 18.6%, respectively (pathways 122-127). (122a) (122b) 298 (123a) (123b) 299 (123c) (124a) 300 (124b) (125a) 301 (125b) (126a) 302 (126b) (127) 75.3% M-C2-TOL + OH. = 0.054APET10 + 0.134APET11 +0.08APET12 +0.106APET13 +0.226APET14 +0.153APET15 + 0.567MGLY +0.186GLY + 0.753RO2R + 0.207XC Reaction 108 303 Net: M-C2-TOL + OH. = 0.21CRES + 0.21HO2. + 0.037BALD + 0.0185CO2 + 0.037H2O + 0.054APET10 + 0.134APET11 +0.08APET12 +0.106APET13 +0.226APET14 +0.153APET15 + 0.567MGLY +0.186GLY + 0.6825XC + 0.8085RO2R 0.0185OH.. Reaction 109 Aerorol precursors (APET10-APET15) further react with OH radical in the presence of O2 and NO2 to form two types of condensable products (ETPM5 and ETPM6). ETPM5, with organonitrate, is produced through the consecutive reactions: first the addition of OH to carbon double bonds, then the addition of O2 and NO2 to the radical (pathways 128a 133a). This pathway accounts for 25% of the products(estimated from o-xylene). The remainder of the OH-ring clevage adduct (75%) further oxidizes in the presence of NO forming NO2, HO2 radical, and ETPM6 (product without organonitrate group), pathways 128b-133b. (128a) (128b) 304 (129a) (129b) (130a) (130b) (131a) (131b) 305 (132a) (132b) (133a) (133b) APET10 + OH. = 0.25ETPM5 +0.75ETPM6 +0.25RO2N +0.75RO2R Reaction 110 APET11 + OH. = 0.25ETPM5 +0.75ETPM6 +0.25RO2N +0.75RO2R Reaction 111 APET12 + OH. = 0.25ETPM5 +0.75ETPM6 +0.25RO2N +0.75RO2R Reaction 112 APET13 + OH. = 0.25ETPM5 +0.75ETPM6 +0.25RO2N +0.75RO2R Reaction 113 APET14 + OH. = 0.25ETPM5 +0.75ETPM6 +0.25RO2N +0.75RO2R Reaction 114 APET15 + OH. = 0.25ETPM5 +0.75ETPM6 +0.25RO2N +0.75RO2R Reaction 115 306 Reactions 109-115 describe condensed mechanisms of m-ethyltoluene with OH radical leading to SOA formation. Table 14 summarizes the stoichiometric coefficients in reactions 109-115, and cites sources of the data. Table C.14: Kinetic parameters of m-ethyltoluene reactions with OH radical forming aerosol products and their source Notations CRES BALD MGLY GLY HO2. RO2R Descriptions Cresol Benzaldehyde Methylglyoxal Glyoxal Hydroperoxy radical Operator RO2R Stoichiometric Coefficients (SC) 0.21 0.037 0.567 0.186 0.21 0.8085(reaction 109) 0.75 (reaction 110-115) APET10 APET11 APET12 APET13 APET14 APET15 ETPM3 ETPM4 RO2N XC CO2 Aerosol precursor APET10 Aerosol precursor APET11 Aerosol precursor APET12 Aerosol precursor APET13 Aerosol precursor APET14 Aerosol precursor APET15 Aerosol product species 1 Aerosol product species 2 Operator RO2N Balance extra carbon Carbon dioxide 307 0.054 0.134 0.08 0.106 0.226 0.153 0.25 0.75 0.25 0.6825 0.0185 Source of SC Estimated from m-xylene (Eusebi, 1996) Estimated from m-xylene (Eusebi, 1996) Estimated from m-xylene (Eusebi, 1996) Estimated from m-xylene (Eusebi, 1996) Estimated from m-xylene (Eusebi, 1996) Estimated from m-xylene (Eusebi, 1996) Estimated from o-xylene (Eusebi, 1996) Estimated from m-xylene (Eusebi, 1996) Estimated from m-xylene (Eusebi, 1996) Estimated from m-xylene (Eusebi, 1996) Estimated from m-xylene (Eusebi, 1996) Estimated from m-xylene (Eusebi, 1996) Estimated from m-xylene (Eusebi, 1996) Estimated from products from o-xylene (Eusebi, 1996) Estimated from products from o-xylene (Eusebi, 1996) Estimated from products from o-xylene (Eusebi, 1996) SAPRC-99, estimated from m-xylene (Eusebi, 1996) Estimated from m-xylene H2O Water 0.037 (Eusebi, 1996) Estimated from m-xylene (Eusebi, 1996) 308 Appendix D : Conditions used for SAPRC simulation for case study of aromatic precursors Table D1: Simulation conditions for aromatic precursors, obtained from Odum's experiments Species m-ethyltoluene p-ethyltoluene o-ethyltoluene m-xylene p-xylene o-xylene 1,3,5trimethylbenze ne ethylbenzene Date 06/06a 06/12b 06/17b 07/08a 07/08b 06/06b 06/19a 06/19b 06/28b 07/01b 09/09 09/11 06/12a ROGo ( g/m3) 2141 1091 415 998 1131 2017 1596 1596 1163 1399 1403 1779 1029 ROG ( g/m3) 1927 971 334 708 789 1891 1571 1528 823 1063 1117 1082 1029 Mo ( g/m3) 208 66 13 38 49 106 46 48 16 32 35 23 31 NO (ppb) 510 265 96 166 166 415 306 287 228 210 368 511 280 NO2 (ppb) 240 97 47 110 113 185 136 127 142 171 188 266 105 Tavg (K) 307 304 306 308 308 307 304 304 303 315 312 310 304 Y (%) 0.108 0.068 0.039 0.054 0.062 0.056 0.029 0.031 0.019 0.030 0.031 0.021 0.031 06/17a 4011 434 13 989 460 306 0.030 06/21a 6733 3176 394 484 260 301 0.124 06/24b 3696 1872 185 255 167 302 0.099 07/10a 3514 1169 104 185 119 307 0.089 n07/05b 4314 1314 103 507 316 312 0.078 propylbenzene 07/17a 1862 657 39 260 158 307 0.059 07/19a 5398 1790 190 409 259 312 0.106 toluene 06/28a 2560 1413 133 443 242 303 0.094 07/01a 2415 1268 111 200 145 315 0.088 07/05a 2656 1710 171 488 305 312 0.100 07/10b 3065 923 68 149 104 307 0.074 Note: 1.Simulations were performed using temperature as given in the above table plus extra 10 K for smog chamber 2. Light intensities were multiplied by factor of 2 to obtain the desirable amount of ROG within simulating time of 10 hours Where: ROGo is initial concentration of hydrocarbon precursors (e.g., 1,2,4trimethylbenzene) ROG is amount of hydrocarbon that reacts Mo is amount of secondary organic aerosol formed Y is secondary organic aerosol yield (Mo/ROG) 309 Appendix E : Aerosol mass changes of PM3-PM8 from lumped APR2-APR4 Lumped aerosol precursors APR2-APR4 produce secondary organic aerosol species PM3-PM8. Figure E.1 demonstrates amount of PM3-PM8 produced from precursor reactions and partitions into particulate phase. 310 Figure E1: amount of PM3-PM8 produced and partition into particulate phase as a function of time 311 Bibliography Atkinson et al., "Evaluated Kinetic, Photochemical and Heterogeneous Data for Atmosphere" 11, 45, 1997 Atkinson, "Gas Phase Tropospheric Chemistry of Organic Compounds" J. Phys. Chem. Ref. Data, Monograph no. 2, 1994 Atkinson et al., "Formation of Ring-Retaining Products from the OH RadicalInitiated Reactions of o-,m-, and p-Xylene", Int. J. Chem. Kint., 23, 77, 1991 Atkinson, "Gas-Phase Tropospheric Chemistry of Organic Compounds: A Review", Atmos. Envion., 24A, 1-41, 1990 Atkinson et al., "Kinetics of the Reactions of Acenaphthalene and Acecnaphthylene and Structurally-Related Aromatic Compounds with OH and NO3 Radicals, N2O5 and O3 at 296 + 2 K", Int. J. Chem. Kinet., 20, 513, 1988 Atkinson et al., "Kinetics of Gas Phase Reactions of Alkylnaphthalenes with O3, N2O5, and OH radicals at 298 2 K" Atmos. Environ, 21, 2323, 1987 312 Atkinson et al., "Kinetic of the Reactions of Naphthalene, 2-Methylnaphthalene, and 2,3-Dimehtylnephthalene with OH radicals, and with O3 at 295 + 1K", Int. J. Chem. Kint, 18, 569, 1986 Atkinson et al., "Atmospheric Gas Phase Loss Processes for Chlorobenzene, Benzotrifluoride, and 4-Chlorobenzotrifluoride, and Generalization of Predictive Techniques for Atmospheric Lifetimes of Aromatic Compounds" Arch Environ. Contamin. Toxical. 14, 417, 1985 Atkinson and Carter, "Kinetic and Mechanisms of the Gas-Phase Reactions of Ozone with Organic Compounds under Atmospheric Conditions", Chem. Rev., 84, 437-470, 1984 Atkinson, Carter, "A Smog Chamber and Modeling Study of the Gas Phase NO-Air Photooxidation of Toluene and the Cresols" Int. J.Chem. Kin., 12, 779-836, 1980 Bandow, "Ring-Clevage Reactions of Aromatic Hydrocarbons Studied by FT-IR Spectroscopy. I. Photooxidation of Toluene and Benzene in the NOx-Air System", Bull. Chem.Soc. Jpn., 58, 2531-2540, 1985, Bandow Hiroshi, Washida Nobuaki, "Ring-cleavage Reactions of Aromatic Hydrocarbons Studies by FT-IR Spectroscopy. II. Photooxidation of o-, m-, and p313 Xylene in the NOx-Air System", The Chemical Society of Japan, 58, 2541-2548 (1985) Bandow Hiroshi, Washida Nobuaki, "Ring-cleavage Reactions of Aromatic Hydrocarbons Studies by FT-IR Spectroscopy. III. Photooxidation of 1,2,3-,1,2,4-, and 1,3,5-Trimethybenzene in the NOx-Air System", The chemical Society of Japan, 58, 2549-2555 (1985) Barthelmie, R. J., Pyor, S. C., "Secondary Organic Aerosols: Formation Potential and Ambient Data", The Science of the Total Environment, 205, 167-178, 1997 Biermann et al., "Kinetics of the Gas-Phase Reactions of the Hydroxyl Radical with Baogtgakebe, Phenanthrene, and Anthracecne", Environ. Sci., 19, 244-248, 1985 Binkowski and Shankar, "The Regional Particulate Matter Model", 1: Model Description and Preliminary Results, J. Geophy. Res., 100, 26, 191-209, 1995 Bowman, F. M., Standley, I. J., "Natural and Anthropogenic Organic Compounds in the Global Atmosphere", Atmospheric Acidity, Elsevier Applied Science, 313-381, 1992 314 Carter P. L., "Documentation of the SAPRC-99 Chemical Mechanism for VOC Reactivity Assessment" Report to California Air Resources Board Contract 92-329, 95-308, 2000 Carter, P. L., "Development and Application of an Up-To-Date Photochemical Mechanism for Airshed Modeling and Reactivity Assessment", Air Resources Board Research Division, Final Report Contract No. A932-094, 1998 Carter, P. L., "Development of Ozone Reactivity Scales for Volatile Organic Compounds", Journal of Air and Waste Management Assoc., 44, 881-899, 1994 Carter, P. L., "Computer modeling of environmental chamber studies of maximum incremental reactivities of volatile organic compounds, Atmos. Environ., 29, 2, 513-2, 527, 1995 Carter, P. L., "A Detailed Mechanism of the Gas Phase Atmospheric Reactions of Organic Compounds", Atmospheric Environment, 24A, 481-518, 1990 Carter, P. L., "Documentation for the SAPRC Atmospheric Photochemical Mechanism Preparation and Emissions Processing Programs for Implementation in Airshed Models", California Air Resources Board, Contact No. A5-122-32, 1988 315 Dekermenjian M., Allen, D. T., Atkinson R., Arey J., "FTIR Analysis of Aerosol Formed in the Photooxidation of Naphthalene", Aerosol Science and Technology, 1997 Environ, user's guide, "Comprehensive Air Quality Model with Extensive (CAMx) version 2, 2000 Eusebi A., "Composition of Aerosol Formed by the Reactions of Hydrocarbons in Urban Atmospheres: Smog Chamber and Field Measurements", University of California, 1996 Fujiwara M., Mishima K., Tamai K., Tanimoto Y., "Spectroscopic Studies on Photochemical of o-Xylene in Solution", Journal of Chem. 101, 4912-4915, 1997 Gery et al., "A Photochemcial Kinetics Mechanism for Urban and Regional Scale Computer Modeling" J. Geophys,. Res., 94, 12, 925-12, 956, 1989 Golikeri and Luss, "Aggregation of Many Coupled Consecutive First Order Reactions", Chem. Engineering Sci., 29, 845-855, 1974 316 Griffin, R. J., Cocker, D. R., "Organic Aerosol Formation from the Oxidation of Biogenic Hydrocarbons", Journal of Geo., 104, 3555-3567, 1999 Grosjean D., and Seinfeld, J. H., "Parametrization of the Formation Potential of Secondary Organic Aerosols", Atmos Environ. 23. 1733-1747 (1989) Grosjean D., "In Situ Organic Aerosol Formation During a Smog Episode Estimated Production and Chemical Functionality", Atmos. Environ. 26A. 953-963 (1992) Grovenstein et al., "Reaction of Atomic Oxygen with Atomic Hydrocarbons", J. Amer. Chem. Soc., 92, 3810-3812, 1970 Heintzenberg J., "The Life Cycle of the Atmospheric Aerosol", Topics in Atmospheric and Interstellar Physics and Chemistry, 251-270, 1994 Holes A., Eusebi A., Allen D.T., "FTIR Analysis of Aerosol Formed in the Photooxidation of 1,3,5-Trimethybenzene", Aerosol Science and Technology, 1997 Isidorov V.A., "Organic Chemistry of the Earths Atmosphere", 215, 1990 Jacobson, "Developing, Couple and Applying a Gas, Aerosol, Transport and Radiation Model to Study Urban and Regional Air Pollution, PhD. Thesis, 317 Department of Atmospheric Sciences, University of California, Los Angeles, California, 1994 Jacobson, "Development and Application of a New Air Pollution Modeling System, part II: Aerosol Module Structure and Design", Atmos. Environ., 31, 131-144, 1997 Kumar, "Development and Application of a Three-Dimensional Aerosol Model, Sonoma Technology, Inc., Santa Rosa, California, 1996 Kwok, "Rate Constants for the Gas-Phase Reaction of Selected Carbamates and Lactates", Environ. Sci. Tchono., 30, 329-334, 1996 Kwok and Atkinson, "Estimation of Hydroxyl Radical Reaction Rate Constants for Gas-Phase Organic Compounds Using a Structure-Reactivity Relationship: and Update.", Atmos. Environ., 29, 1685-1695, 1995 Kwok et al., "Gas Phase Atmospheric Chemistry of Selected Thiocarbamates", Environ. Sci. Techno., 26, 1798-1807, 1992 McMurry and Grosjean, "Photochemical Formation of Organic Arosols: Growth Laws and Mechanisms", Atmos. Environ., 19, 1, 445-1, 451, 1985 318 Meylan "Atmospheric Oxidation Program, Group Rate Constants and Substituent Values", Syracuse Research Corporation, 1998 Meylan, "User's Guide for MPBPVP version 1.2" Syracuse Research Corporation, 1996 Meng et al., "Gas/Aerosol Distribution of Formic and Acetic Acids, Aerosol Sci. Technol., 23, 561-578, 1997 Moucheron and Milford, "Development and Testing of a Process Model for Secondary Organic Aerosol, Air & Waste Management Association Annual Meeting, Paper 96-FA1308.03, Nashville, Tennessee, 1997 MPBPWIN programming software, developed by Syracuse Research Corporation, 1996 Mylonas, T. D., Allen, D. T., Ehrman, S. H., and Pratsinis, S. E., "The Sources and Size Distributions of Organonitrates in Los Angeles", Atmos. Envir. 25A. 12, 28552861 (1991) 319 The National Research Council, "Ozone-Forming Potential of Reformulated Gasoline", National Academic Press, Washington D. C., 1999 Noakov T., Penner J.E., "Large Contribution of Organic Aerosols to CloudCondensation Nuclei Concentration", Nature, 365, 823-826, 1993 Odum et al., "The Atmospheric Aerosol-Forming Potential of Whole Gasoline Vapor", Science, 276, 96-99, 1997 Odum et al, "Gas'Particle Partitioning and Secondary Organic Aerosol Yields", Environ. Sci. Technol., 30, 2,580-2,585, 1996 Pandis, S. N., and Seinfeld, J. H., "Atmospheric Photochemical Oxidation of 1Octene: OH, O3, and O(3P) Reactions", Environ. Sci. Technol., 26, 1165-1173 (1992). Pankow, "An Absorption Model of Gas/Particle Partitioning of Organic Compounds in the Atmosphere, Atmos. Environ., 28, 185-188, 1994 Perry R. A., Atkinson R., and Pitts J. N., Jr., "Kinetics and Mechanisms of the Gas Phase Reactions of OH Radicals with Aromatic Hydrocarbons Over the Temperature Range 296 473 K", J. Phys. Chem., 81, 296-304 (1977) 320 Pickle T. and Allen D. T., "The Sources and Size Distributions of Aliphatic and Carbonyl Carbon in Los Angeles Aerosol", Atmos. Envir. 24A, 8, 2221-2228 (1990) Saxena P., Hildemann L. M., McMurry P.H., Seinfeid J. H., "Organics Alter Hygroscopic Behavior of Atmospheric Particles", Journal of Geophysic Research, 100, 18-755, 1995 Schwartz J., Dockery D.W., Neas L.M., "Is Daily Mortality Associated Specifically with Fine Particles", J. Air & Waste Management Association, 46, 927-939, 1996 Siegneur et al.,, "Review of Air Quality models for Particulate Matter" Document Number CP015-97-1b, 1997 Seinfeld J.M., Bowman F.M., Andino J.M., "Tropospheric Chemistry Advances in Chemical Engineering", J. Wei, ed., 19, 325-407, 1994 Stockwell et al., "The Second-Generation Regional Acid Deposition Model Chemical Mechanism for Regional Air Quality Modeling, J. Geophys. Res., 95D, 16, 343-16, 367, 1990 321 Sun and Wexler, "Am implicit-Explicit Hybrid Solver for a System of Stiff Kinetic Equations, Air & Waste management. Assoc. 87th Annual Meeting, Cincinnati, Ohio, 1997 Tropp, R. J., Kohl, S. D., Chow, J. C., and Frazier, C. A., "Final Report for the Texas PM2.5 Sampling and Analysis Study", prepared for Bureau of Air Quality Control, City of Houston, TX, 1998 Turpin B. J., and Huntzicker J. J., "Secondary Formation of Organic Aerosol in the Los Angeles Basin: A Descriptive Analysis of Organic and Elemental Carbon Concentrations", Atmos. Envir., 25A, 2, 207-215 (1991) U.S. Environmental Protection Agency, "Summary of Model-Adjusted Estimates of Fine Particulate Matter for 2020 Base and Control Cases", 1999 U.S. Environmental Protection Agency, http://www.epa.gov/oar/aqtrnd97/chapter2.pdf", 1997 U.S. Environmental Protection Agency, "National Ambient Air Quality Standard for Particulate matter: Proposed Decision", 40 CFR Part 50, 1996 322 U.S. Environmental Protection Agency, User's guide for the industrial source complex dispersion models, Vol.1, User Instruction, EPA-450/14-92-008, 1992 Venkatram, "The Development of a Model to Examine Source-Receptor Relationships for Visibility on the Colorado Plateau, J. Air & Waste Manage. Assoc., 47, 286-301, 1997 Wei and Kuo, Ind. Eng. Chem. Fundamentals, 8, 114-124, 1969 323 VITA Wipawee Dechapanya was born in Bangkok, Thailand. She is the youngest sister of her family. She has one twin sister, Wipada Dechapanya, who is currently a graduate student in the Department of Civil and Environmental Engineering at Michigan Technological University (MTU). She attended the primary school and high school in Bangkok, Thailand. After graduating from high school, she became an undergraduate student in the Department of Chemical Engineering at Khonkean University, Khonkean, Thailand. On the senior year of her undergraduate study, she applied to be a lecturer in the faculty of Engineering at Ubon-Ratchathani University, Thailand. While she was working as a lecturer, she passed the qualified examination, held by the Royal Thai Government, and received the Royal Thai Scholarship for studying abroad. She had become a graduate student in the Department of Chemical Engineering at MTU since September 1996, and received the degree of Master of Science in July 1998. She has pursued her doctoral study at the University of Texas at Austin in the Fall, 1998. She is going to finish her Ph.D in May 2002. Once she receives her doctoral degree, she will return to Thailand and become a faculty member at Ubon-Ratchathani University, Thailand. Permanent Address: 15/2 Moo.1 Tonsai Bangyao, Pharpradang Samutprakarn 10130, Thailand This dissertation was typed by Wipawee Dechapanya. 324

Find millions of documents here - Study Guides, Homework Solutions, Papers, Exam Answer Keys and more. Course Hero has millions of course related materials that will enable you to learn better, faster and get an A in all your courses.
Below is a small sample set of documents:

shoemakerdb042.pdf
Path: Texas >> SHOEMAKERD >> 042 Fall, 2009

johnsonam71217.pdf
Path: Texas >> JOHNSONAM >> 71217 Fall, 2009
sampselld77810.pdf
Path: Texas >> SAMPSELLD >> 77810 Fall, 2009
complex.txt
Path: CSU San Bernardino >> CS >> 330 Fall, 2009
cadenheadjk046.pdf
Path: Texas >> CADENHEADJ >> 046 Fall, 2009
benjaminsmr042.pdf
Path: Texas >> BENJAMINSM >> 042 Fall, 2009
simpsonal13317.pdf
Path: Texas >> SIMPSONAL >> 13317 Fall, 2009
hamiltont84490.pdf
Path: Texas >> HAMILTONT >> 84490 Fall, 2009
kotrlaka518287.pdf
Path: Texas >> KOTRLAKA >> 518287 Fall, 2009
harrisont86130.pdf
Path: Texas >> HARRISONT >> 86130 Fall, 2009
brandonjc99738.pdf
Path: Texas >> BRANDONJC >> 99738 Fall, 2009
MATH107A46024536.doc
Path: MD University College >> ASIA >> 2092 Fall, 2009
crawforda65881.pdf
Path: Texas >> CRAWFORDA >> 65881 Fall, 2009
achacosom07761.pdf
Path: Texas >> ACHACOSOM >> 07761 Fall, 2009
jarroldwl86380.pdf
Path: Texas >> JARROLDWL >> 86380 Fall, 2009
sharyginany026.pdf
Path: Texas >> SHARYGINAN >> 026 Fall, 2009
goncalvesac026.pdf
Path: Texas >> GONCALVESA >> 026 Fall, 2009
zieglerkj47418.pdf
Path: Texas >> ZIEGLERKJ >> 47418 Fall, 2009
burtnerjc90760.pdf
Path: Texas >> BURTNERJC >> 90760 Fall, 2009
alvarezla07232.pdf
Path: Texas >> ALVAREZLA >> 07232 Fall, 2009
MATH012A46124534.doc
Path: MD University College >> ASIA >> 2092 Fall, 2009
bonningew86532.pdf
Path: Texas >> BONNINGEW >> 86532 Fall, 2009
CMIS141AA44024445.doc
Path: MD University College >> ASIA >> 2092 Fall, 2009
CMIS102AA42086692.doc
Path: MD University College >> ASIA >> 2088 Fall, 2009
STAT200A42186896.doc
Path: MD University College >> ASIA >> 2088 Fall, 2009
kulkarnis86095.pdf
Path: Texas >> KULKARNIS >> 86095 Fall, 2009
chapmanbg60287.pdf
Path: Texas >> CHAPMANBG >> 60287 Fall, 2009
slattonkc78713.pdf
Path: Texas >> SLATTONKC >> 78713 Fall, 2009
michalskylo026.pdf
Path: Texas >> MICHALSKYL >> 026 Fall, 2009
batemanmt33508.pdf
Path: Texas >> BATEMANMT >> 33508 Fall, 2009
lodowskid97061.pdf
Path: Texas >> LODOWSKID >> 97061 Fall, 2009
raichlend29983.pdf
Path: Texas >> RAICHLEND >> 29983 Fall, 2009
perkinsjd44616.pdf
Path: Texas >> PERKINSJD >> 44616 Fall, 2009
mehdiabadinj026.pdf
Path: Texas >> MEHDIABADI >> 026 Fall, 2009
borisovasa86653.pdf
Path: Texas >> BORISOVASA >> 86653 Fall, 2009
Abuhakema504399.pdf
Path: Texas >> ABUHAKEMA >> 504399 Fall, 2009
hw03_solution.doc
Path: Penn State >> ME >> 581 Fall, 2009
oestreichj19588.pdf
Path: Texas >> OESTREICHJ >> 19588 Fall, 2009
evstatieve01477.pdf
Path: Texas >> EVSTATIEVE >> 01477 Fall, 2009
paschvaldesg042.pdf
Path: Texas >> PASCHVALDE >> 042 Fall, 2009
alvaradocg86236.pdf
Path: Texas >> ALVARADOCG >> 86236 Fall, 2009
martinssonpj026.pdf
Path: Texas >> MARTINSSON >> 026 Fall, 2009
makowitza504694.pdf
Path: Texas >> MAKOWITZA >> 504694 Fall, 2009
andersonmw81540.pdf
Path: Texas >> ANDERSONMW >> 81540 Fall, 2009
martinezrs39334.pdf
Path: Texas >> MARTINEZRS >> 39334 Fall, 2009
elshayebta87380.pdf
Path: Texas >> ELSHAYEBTA >> 87380 Fall, 2009
cowmeadowr17589.pdf
Path: Texas >> COWMEADOWR >> 17589 Fall, 2009
schougaardsb029.pdf
Path: Texas >> SCHOUGAARD >> 029 Fall, 2009
kordoskyma87090.pdf
Path: Texas >> KORDOSKYMA >> 87090 Fall, 2009
metcalfets016-x.pdf
Path: Texas >> METCALFETS >> 016 Fall, 2009
bocknackbm84986.pdf
Path: Texas >> BOCKNACKBM >> 84986 Fall, 2009
mahdjoubid26824.pdf
Path: Texas >> MAHDJOUBID >> 26824 Fall, 2009
vanderveenaa029.pdf
Path: Texas >> VANDERVEEN >> 029 Fall, 2009
crabtreejc17037.pdf
Path: Texas >> CRABTREEJC >> 17037 Fall, 2009
steubingdm73657.pdf
Path: Texas >> STEUBINGDM >> 73657 Fall, 2009
johnsonhl692102.pdf
Path: Texas >> JOHNSONHL >> 692102 Fall, 2009
quintopozosd022.pdf
Path: Texas >> QUINTOPOZO >> 022 Fall, 2009
micklerpj516685.pdf
Path: Texas >> MICKLERPJ >> 516685 Fall, 2009
00000011.pdf
Path: Carnegie Mellon >> TERA >> 05102571 Fall, 2009
00000011.pdf
Path: Carnegie Mellon >> DISK >> 05102571 Fall, 2009
strycharskiat042.pdf
Path: Texas >> STRYCHARSK >> 042 Fall, 2009
podorozhnyr48572.pdf
Path: Texas >> PODOROZHNY >> 48572 Fall, 2009
alexandermw25054.pdf
Path: Texas >> ALEXANDERM >> 25054 Fall, 2009
batchelorme80690.pdf
Path: Texas >> BATCHELORM >> 80690 Fall, 2009
franzosajw504611.pdf
Path: Texas >> FRANZOSAJW >> 504611 Fall, 2009
brumbaughms81936.pdf
Path: Texas >> BRUMBAUGHM >> 81936 Fall, 2009
abplanalpbs52539.pdf
Path: Texas >> ABPLANALPB >> 52539 Fall, 2009
notes_05_02.doc
Path: Penn State >> ME >> 481 Fall, 2009
ulloabarbaran022.pdf
Path: Texas >> ULLOABARBA >> 022 Fall, 2009
poppendieckdg026.pdf
Path: Texas >> POPPENDIEC >> 026 Fall, 2009