This item is only available as the following downloads:
1 UV VISIBLE LIGHT ABSORPTION PROPERTIES OF ORGANIC CARBON AEROSOL IN ATMOSPHERE By MIN ZHONG A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE D EGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2013
2 2013 Min Zhong
3 To my beloved parents and family
4 ACKNOWLEDGMENTS This work was supported by a grant from the National Science Foundation (AT M 0852747) and by a UF Alumni Scholarship. I would like to first give thanks to my advisor, Dr. Myoseon Jang, for offering me the opportunit y to study the exciting research topic providing me countless support of my research and writing of this dissertat ion. Besides, I would like to thank Dr. Nicolo Omenetto, Dr. Jennifer S. Curtis Dr. Ben Koopman and Dr. Barron H. Henderson for being my committee members and giving me insightful comments and hard questions. Next, I would like to give thanks to all my current and past lab mates. Together with these bright guys, I had many unforgettable moments in Florida such as nighttime camping beach fishing, and river tubing I want to thank my husband who always stand s by me and cheer s me up when I have difficulti es and my daughter who g ives me endless happiness. Last but not the least; I would also like to thank my dad, who is in China, for his love and spiritual support s over the past years and also my mother, who gave me life and provide d me the financial sup port s to cover the cost of education. I would also like to thank my parents in law, who came to the United States to help me tak ing care of my daughter when I had to work.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ .......... 9 LIST OF ABBREVIATIONS ................................ ................................ ........................... 11 ABSTRACT ................................ ................................ ................................ ................... 12 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 14 Atmospheric Aerosols and Climate Impact ................................ ............................. 14 Organic Carb on Aerosol ................................ ................................ ......................... 15 Optical Properties of Organic Aerosol ................................ ................................ ..... 16 Direct Climate Impact of Organic Carbon ................................ ............................... 18 Motivation and Objectives ................................ ................................ ....................... 18 2 MASS ABSORPTION CROSS SECTION MEASUREMENT OF SOA USING A UV VISBILE SPECTROMETER CONNECTED WITH AN INTEGRATING SPHERE ................................ ................................ ................................ ................. 21 Background ................................ ................................ ................................ ............. 21 Experimental Section ................................ ................................ .............................. 23 S OA Formation ................................ ................................ ................................ 23 Light Absorption Measurement ................................ ................................ ......... 24 Results and Discussion ................................ ................................ ........................... 25 Methodology Developme nt ................................ ................................ ............... 25 Theory of t ransmittance for t he p article f ilter s ample ................................ 25 Calibration of TUV IS d ata ................................ ................................ ......... 27 Relationship b etween TUV IS and RUV IS ................................ ................ 28 Aerosol Mass Absorption Cross Section Calculation ................................ ........ 29 App lication on SOA ................................ ................................ .......................... 31 Absorption s pectra and MAC of SOA ................................ ......................... 31 Effect of i norganic s eeds ................................ ................................ ............ 33 Effect of light source ................................ ................................ .................. 34 Conclusion and Atmospheric Implication ................................ ................................ 35 3 T HE SOA FORMATION MODEL COMBINED WITH SEMIEMPIRICAL QUANTUM CHEMISTRY TO PREDICT UV VIS ABSORPTION OF SECONDARY ORGANIC AEROSOLS ................................ ................................ ... 47
6 Background ................................ ................................ ................................ ............. 47 Experimental Sectio n ................................ ................................ .............................. 49 SOA Formation ................................ ................................ ................................ 49 SOA UV Visible Spectra Recording ................................ ................................ 50 Result s and Discussion ................................ ................................ ........................... 51 Measurement of SOA UV V isible S pectra ................................ ........................ 51 Theoretical C alculations of SOA UV V isible S pectra ................................ ........ 52 SOA F ormation M odel ................................ ................................ ...................... 52 SOA P roducts ................................ ................................ ................................ ... 55 UV V isible S pectrum P rediction ................................ ................................ ....... 55 Simulation R esults ................................ ................................ ............................ 57 Light a bsorption s pectra of t oluene SOA ................................ ................... 57 Lig ht absorpt pinene SOA ................................ ................. 59 Conclusion ................................ ................................ ................................ .............. 60 4 DYNAMIC LIGHT ABSORPTION OF BIOMASS BURNING ORGANIC AEROSOL PHOCHEMICALLY AGED UN DER NATURAL SUNLIGHT ................. 71 Background ................................ ................................ ................................ ............. 71 Experimental Section ................................ ................................ .............................. 73 Ou tdoor Chamber Experimental Setup ................................ ............................ 73 Wood S moke C haracterization ................................ ................................ ......... 74 Light A bsorption of A mbient Organic Carbon ................................ ................... 76 Results and D iscussion ................................ ................................ ........................... 76 Light Absorption of O C ................................ ................................ ..................... 76 Effect o f Photochemical Agi ng o n Light Absorption o f OC ............................... 78 Effect of RH on L ight A bsorption of O C ................................ ............................ 79 Effect of NO x on L ight A bsorption of O C ................................ ........................... 80 Chemical E volution of Organic Carbon Aerosol ................................ ............... 80 Levoglucosan decay ................................ ................................ .................. 81 PAHs decay ................................ ................................ ............................... 82 FTIR spectra of wood burning aerosol ................................ ....................... 82 H ygroscopic p roperties of wood burning aerosol ................................ ....... 83 Conclusion and A tmospheric I mplication ................................ ................................ 83 5 RADIATIVE IMPACT OF ORGANIC CARBON AEROSOL ................................ .... 92 Bac kground ................................ ................................ ................................ ............. 92 Method ................................ ................................ ................................ .................... 92 Mie Scattering Model ................................ ................................ ........................ 93 Effect of RH on Par ticle Size ................................ ................................ ............ 94 Simple Radiative Efficiency Estimation ................................ ............................ 94 Results and Discussion ................................ ................................ ........................... 95 Optical Parameters ................................ ................................ ........................... 95 Effect of RH on Optical Parameters ................................ ................................ 96 Radiative Impact ................................ ................................ ............................... 96 Conclusion ................................ ................................ ................................ .............. 96
7 6 CONCLUSIONS ................................ ................................ ................................ ... 101 7 FUTURE STUDIES ................................ ................................ ............................... 103 APPENDIX A SUPPLEMENTARY METERIALS FOR CHAPTER 2 ................................ ........... 105 B SUPPLEMENTARY METERIALS FOR CHAPTER 3 ................................ ........... 108 C SUPPLE MENTARY METERIALS FOR CHAPTER 4 ................................ ........... 120 D SUPPLEMENTARY METERIALS FOR CHAPTER 5 ................................ ........... 128 LIST OF REFERENCES ................................ ................................ ............................. 134 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 146
8 LIST OF TABLES Table page 2 1 Experimental conditions and the resulting SOA data for p hotooxidation of toluene using the 2 m 3 Teflon film chamber ................................ ........................ 37 2 2 Experimental conditions and the resulting SOA data for oxidation of d limonene and a pinene using the 2 m 3 Teflon film cham be r ............................... 38 3 1 Experimental conditions and the resulting SOA data for photooxidation of pinene using the 2 m 3 Teflon film chamber ................................ .. 61 3 2 Selected products of toluene SOA and their mass percentages at different NO x conditions ................................ ................................ ................................ .... 62 3 3 pinene SOA and their mass percentages in SOA at different NO x conditions ................................ ................................ ......... 63 3 4 Comparisions between model predict ed abs orbance and literature values ........ 64 4 1 Summary of experimental conditions of fresh wood smoke for photochemical oxidation ................................ ................................ ................................ ............. 85 A 1 Indoor Teflon film chamber experiments involving the aerosol of known composition ................................ ................................ ................................ ...... 105 B 1 Chemical structure of toluene SOA products ................................ .................... 108 B 2 Representative products of toluene SOA and their mass percentages at different NO x conditions ................................ ................................ .................... 111 B 3 pinene SOA from MCM mechnism ............................. 112 B 4 pinene SOA and their mass percentages in SOA at different NO x conditions ................................ ................................ ....... 116 D 1 Input parameters in Mie code for dry SOA aerosol ................................ ........... 128 D 2 Input parameters in Mie code for SOA aerosol at RH of 50% .......................... 129 D 3 Input parameters in Mie code for d ry POA ................................ ....................... 130 D 4 Input parameters in Mie code for POA at RH of 50% ................................ ....... 132 D 5 Input parameters in Mie code for dry sulfate ................................ .................... 133
9 LIST OF FIGURES Figure page 1 1 Composition of ambient particular matter ................................ ........................... 20 2 1 Schematic diagram for the beam pathway in both transmittance and reflectance mode ................................ ................................ ............................... 39 2 2 ln(1/T) plotted vs. metanil yellow aerosol mass collected on filter at three different wavelengths (422 nm, 300 nm, and 500 nm) ................................ ....... 40 2 3 Measured ln(1/ T ) vs theoretical ln(1/ T ) ................................ ............................... 41 2 4 Reflectance is plotted against transmittance at 422~680 nm for MY NaCl aerosol with different particle mass ................................ ................................ .... 42 2 5 UV visible spectra of SOA freshly generated from photooxidation of different hydrocarbons ................................ ................................ ................................ ..... 43 2 6 and 450 nm ................................ ................................ ................................ ......... 44 2 7 E ffect of light on MAC of d limonene and toluene SOA ................................ ..... 45 2 8 Proposed example of conjugated compound formation through aerosol phase reaction in d limonene SOA ................................ ................................ .... 46 3 1 Schematic diagram for the measurem ent of UV visible spectra of SOA ............ 65 3 2 S tructure of the SOA light absorption model. ................................ ..................... 66 3 3 Comparisons of the observed and cal culated spectra of testing compounds using Gaussian band shape function ................................ ................................ 67 3 4 FWHM for two systems. ................................ ............ 68 3 5 Comparison of the predicted absorption spectra and the measured absorption spectra of toluene SOA under three different NO x conditions. .......... 69 3 6 Comparison of the predicted absorption spectra and the measured spectra of pinene SOA under different NO x conditions ................................ .................... 70 4 1 UV visible light absorption spectra and MAC OC of wood smoke OC. .................. 86 4 2 Comparison of light absorption of wood OA photochemically oxidized at different humidity conditions. ................................ ................................ .............. 87
10 4 3 Comparison of light absorption of wood OA photochemic ally oxidized at different NO x conditions. ................................ ................................ ..................... 88 4 4 Decay of levoglucosan and PAHs ................................ ................................ ..... 89 4 5 FTIR spectra and hygroscopic growth pro file of fresh and aged wood burning particles ................................ ................................ ................................ ............. 90 4 6 Light absorption of ambient biomass burning OA sampled during the country line wildfire event at different date ................................ ................................ ...... 91 5 1 Optical parameters of SOA POA and sulfate estimated using Mie code : ( a) extinction cross section area, ( b) aerosol asymmetry factor, and ( c) single scattering albedo ................................ ................................ ............................... 98 5 2 Effect of RH on optical parameters ................................ ................................ .... 99 5 3 Comparison of radiative efficiency of SOA, POA and sulfate as a function of wavelength ................................ ................................ ................................ ...... 100 A 1 M easured molar absorptivity of Metanil Yellow as a function of wavelength (280 680 nm). ................................ ................................ ................................ ... 1 06 A 2 UV visible absorption spectra of d limonene SOA collected on the filter at different exposure time in air ................................ ................................ ........... 107 B 1 Comparison of model simulated and measured concentrations of toluene, O 3 NO x and NO for experiments at different NO x levels ................................ ....... 117 B 2 Comparison of model simulated and measured concentrations of pinene, O 3 NO x and NO for experiments at different NO x levels ................................ 118 B 3 Comparison of the predicted OM T and the measured OM T for different systems ................................ ................................ ................................ ............ 119 C 1 Time profile of sunlight total ultra violet radiation (TUVT), temperature and relative humidity measured in the UF APHOR E ast chamber on October 30, 2012. ................................ ................................ ................................ ................ 122 C 2 Mass spectra of BSTFA derivatives of levoglucosan oxidation products in EI mode. ................................ ................................ ................................ ............... 123 C 3 Mass spectra of BSTFA derivatives of levoglucosan oxidation products in CI mode. ................................ ................................ ................................ ............... 125 C 4 Reaction pathways for levoglucosan decompositio n in the presence of OH radical. ................................ ................................ ................................ .............. 127
11 LIST OF ABBREVIATIONS AP alpha P inene ( AP ) is a major biogenic volatile organic compou nd mainly emitted from pine trees, with molecular formula of C 10 H 16 BC Black Carbon ( BC) is a type of carbonaceous material formed in flames during combustion of carbon based fuels. DL D Limonene (DL) is a common biogenic volatile organic compound emitt ed from c itrus trees, with molecular formula of C 10 H 16 MAC Mass absorption cross section ( MAC m 2 g 1 ) is a measurement for light absorption property of aerosol. MCM The master chemical mechanism (MCM) is a near explicit chemical mechanism which describes t he gas phase chemical oxidation of volatile organic compounds. O C Organic Carbon (OC) aerosol is a type of organic aerosol contains only carbon and hydrogen, usually oxygen. POA Primary Organic Aerosol (POA) is the aerosol which directly emitted from comb ustion sources. SOA Secondary Organic Aerosol (SOA) is a type of organic aerosol that formed through condensation of semivolatile organic compounds in gas phase. SVOC Semi volatile organic compounds (VOC) is any organic compound having a boiling point be tween 250 and 400 measured at a standard atmospheric pressure of 101.3 kPa TOL Toluene (TOL) is a major anthropo genic volatile organic compound with molecular formula of C 7 H 8 UV Ultraviolet ( UV) light is electromagnetic radiation with a wavelength shorter th an that of visible light, but longer than X rays. UV light reaches to troposphere has wavelength ranging from 280nm to 400nm. VOC Volatile organic compounds (VOC ) is any organic compound having a boiling point less than or equal to 250 measured at a stan dard atmospheric pressure of 101.3 kPa
12 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy UV VISIBLE LIGHT ABSORPTIO N PROPERTIES OF ORGANIC CARBON AEROSOL IN ATMOSPHERE By Min Zhong August 2013 Chair: Myoseon Jang Major: Environmental Engineering Sciences Atmospheric aerosols play an import role in climate change through direct interaction with sunlight. Organic car bon (OC) aerosol which absorbs ultraviolet and short visi ble wavelength sunlight, is one of the least understood factors in climate change To quantify the climate effect of organic aerosol the first step is to obtain its fundamental optical parameters. This study explored a solvent free method to measure the mass absorption cross section ( MAC ) of OC and applied this method to both secondary organic aerosol ( SOA) and primary organic aerosol ( POA ) SOA was produced through photooxidation of different p recursor hydrocarbons such as toluene, d limonene and pinene using a 2 m 3 indoor Teflon film chamber. Of the se three precursor hydrocarbons the MAC value of toluene SOA was the highest compared with MAC of d limonene and pinene SOA. To understand how the composition of SOA affects its MAC a new model to predict the UV visible absorption spectrum of SOA was developed. The model predict ed that, with higher level of NO x (~100ppb), absorption of toluene SOA would be higher due to the formation of more nitrocatechols. This method was also used to measure M AC of PO A Wood burning OC produced under smoldering conditions were photochemically aged using a 104 m 3 dual outdoor chamber under
13 natural sunlight. Overall, the mass absorption cross section of OC increased by 11 54% in the morning and then gradually d ecreased by 19 68% in the afte rnoon A similar trend in the change of light absorption was observed in ambient smoke aerosol originating from the 2012 County Line Wildfire in Florida. To quantify the climate forcing of OC, a simple forcing efficiency was used to provide a first estimate of the climate impact of SOA and PO A Our results confi r m that SOA can be treated as cooling aerosol with negative climate forcing similar to inorganic sulfates. POA has the potential to release heat to atmosphere with net positive forcing.
14 CHAPTER 1 INTRODUCTION Atmospheric Aerosols and C limate I mpact Atmospheric aerosols are generally defined as those liquid, semi solid or solid particles suspended in air, having diameters in the range of 1 nm to 10 (Seinfeld and Pandis, 1998) Based on the ir formation mechanism s a tmospheric aerosols can be divided in to two categories: secondary aerosols and primary aerosols. Secondary particles are formed through gas to particle conversion. For example, ammonium sulfate, (NH 4 ) 2 SO 4 is formed through the ammonia titration with sulfuric acid. Primary particles are di rectly emitted from sources such as biomass burning, sea salt, volcano eruption and soil Elemental carbon and most inorganic species are usually the components of primary particles. Figure 1 summarizes typical atmospheric aerosol compositions based on current ambient studies. Concerns on anthropogenic climate change have drawn attention to the role of atmospheric aerosol in the earth s climate system. Atmospheric aerosols have a significant impact on climate through their direct and indirect influence s on climate (Pschl, 2005;Rosenfeld et al., 2008) Aerosols have a direct climate forcing effect because they scatter and absorb sunlight. They also change the formation and precipitation efficiency of c loud thu s causing an indirect radiative forcing. To measure the influence of aerosol on climate, radiative fo rcing has been typically used. Radiative forcing (W/m 2 ) is the rate of energy c hange per unit area of the glob e as measured at the top of the atmosphere. The positive and negative sign of radiative forcing indicate warming and cooling influence of aerosols respectively
15 Unlike greenhouse gases, which are well mixed in the atmosphere and have a long life time in atmosphere, aerosol has much shorter life t ime, typically on the order of 6 days. The mass concentration, size distribution, and physiochemical properties of aerosols are highly variable in space and time. According to Intergovernmental Panel on Climate Change ( IPCC ) Assessment (2007) atmospheric aerosols are the largest source of uncertainty in climate forc ing estimation. Among various atmospheric aerosols, inorganic salts such as sea salt and sulfate can cool down the atmosphere by scattering solar radiation (Randles et al., 2004;Tegen et al., 1996) O nly a few type s of aerosol s such as carbonaceous aerosol and some mineral dust s have the potential to cause climate warming by directly absorbing solar radiation (Menon et al., 2002) Carbonaceo us aerosol includes both organic carbon (OC) and bla c k carbon (BC) aerosol. Light absorption of OC has been poorly understood mainly due to its complex chemical composition as well as evolution of chemical physical, and optical properties O rganic C arbo n Aerosol Organic carbon ( OC) aerosol is defined as the aerosol which is composed of primarily carbon containing compound s that include hydrogen and, usually oxygen. On a global scale O C is a significant fraction of ambient aerosol, accounting for 20 50 % to the total fine aerosol mass at continent (Putaud et al., 2004) an d as much as about 90% in tropical forest areas (Andreae and Crutzen, 1997) Compared with other atmospheric aerosols, O C has some unique physical an d chemical properties : 1. I t absorbs sun light in certain wavelength region particularly in the UV and short visible range. 2. I t has very complex chemical composition, with thousands of chemical compounds present simultaneously in the particle phase.
16 3. I t ca n undergo a wide range of chemical transformation ( oxidation, photolysis, polymerization, etc.) under atmospheric conditions. O C has been categorized as primary organic aerosol ( POA) and secondary organic aerosol (SOA ) depending on the source and formatio n mechanism POA is directly emitted from sources such as bioma ss burning and fuel combustion. In the global emission budget, the total PO A emission is about 35 Tg C per year (Hallquist et al., 2009) whic h accounts f or 23% of total O C flux. SOA is formed from the ph o tochemical oxidation of volatile (VOCs) or semivolatile organic compounds ( SVOCs) Biogenic terpenes from vegetation and aromatics from anthropogenic sources can react with oxidant s such as OH radical an d O 3 to produc e less volatile compounds, which can condense on preexisting particles or self nucleate to form SOA. The total flux of biogenic SOA is 115 Tg C per ye ar, contributing to 77% of total OC Among SOA, 76 % is from biogenic VOCs, such as isopren e, pinene, and d limonene. Optical P roperties of O rganic Aerosol Radiative transfer models require the input of optical properties of aerosol, including asymmetry factor, single scattering albedo, and extinction coefficients ( m 2 m 3 or m 1 ). The estimation of these parameters is needed in all wavelengths particularly in the UV visible wavelength range between 280nm to 7 00nm For longer wavelength in the infrared range, scattering and absorption of particles are less important than gas molecules. With giv en particle size distribution and complex refractive index, it is possible to calculate the optical parameters of aerosol using Mie scattering theory. T he complex refractive index ( m ) is expressed as: m = n + k i. For organic aerosol the real part, n ha s a value betwee n 1.3 1. 7 in the wavelength range of 280 800nm (Nakayama et al ., 2013) The real refractive index decreased slightly with
17 increasing wavelength The imaginary part, k is directly related to the light absorption capacity of aeroso l. L ight absorption of O C is wavelength dependent (Kirchstetter et al., 2004;Schnaiter et al., 2005) dramatically increasing towards shorter wavelengths. Since n does not change too much and k is the optical property most relevant to the positive forcing, t his study is fo cus ed on light absorptio n of organic aerosol This study uses mass absorption cross section ( MAC m 2 g 1 ) to quantitatively describe the light absorption propert ies of organic aerosol. MAC of O C can be calculated from the light absorption coefficient of O C divided by O C mass conc entration. MAC is related to k through the following equation (1 1) w here is the density (g/m 3 ) of O C is the wavelength. POA from c omb ustion organic carbon has been suggested to a main source of brown carbon an important light absorbing particulate matter (Andreae and Gelencser, 2006) There is growing evidence (Nozire and Esteve, 2005;Shapiro et al., 2009;Sareen et al., 2010;Laskin et al., 2010b;Bones e t al., 2010a;Nakayama et al., 2010a;Zhong and Jang, 2011) to show that SOA potentially would contribute to brown carbon. The recent study by Noziere and Esteve (2005) has demonstrated the formation of light absorbing organic matter due to aldehydes and ketones absorbed by sulfuric acid aqueous solution. Shapiro et al. (2009) and Sareen et al. (2010) also reported the production of chromophores when ammonium ions were added to glyoxal and methylglyoxal The formation of colored S OA has been observed in eit her the ozonolysis of biogenic terpenes (Laskin et al., 2010b;Bones et al., 2010a) or the photooxidation of toluene (Nakayama et al., 2010a;Zhong and Jang, 2011)
18 Direct Climate Impact of O rganic Carbon A relatively limited number of studies have investigated the direct radiative forcing of organic carbon The estimated value of radiative forcing of organi c carbon was 0.3 ~ 0.09 w/m 2 (IPCC, 2007;Maria et al., 2004;Chung and Seinfeld, 2002;Myhre et al., 2009;Hoyle et al., 2009) Both Maria et al. (2004) and Myher et al. (2009) reported the radiative forcing of SOA was 0.1w/m 2 For POA, the average value was 0.14 w/m 2 (Schulz et al., 2006) The results of current model simulation show that both SOA and POA ha ve a direct cooling effect on climate. However, one big uncertaint y in the above models is the optical properties of organic carbon. Models which predict the radiative forcing of SOA took the optical properties of POA for SOA. But the optical properties of POA were treated as the same as those of inorganic aerosols. For example, in the work of Chuang and Seinfeld (2002) the optical properties of O C were similar to sulfate and nitrates. Myh er et al (2009) treat ed POA equivalent to sulfate. These inorganic aerosols are known to be cooling aerosols with negligible absorption in the UV visible range. However, organic carbon might absorb both UV and vis ible light. Considering the large source of OC in the atmosphere, q uantification of the light absorption properties of organic aerosol is necessary to reduce the uncertainty in climate models. Motivation and O bjectives Although OC is a major component in atmospheric aerosol, their impact on climate is not well evaluated due to the lack of optical properties of O C Thus detailed measurement and evaluation of the optical properties of both SOA and POA are necessary to accurately model the climate impact of organic aerosols in atmosphere. The objectives of this work are there fore to:
19 1. D evelop a new methodology to quantify the light absorption properties of organic aerosol; 2. M easure the mass absorption cross section ( MAC) of SOA; 3. I nvestigate the contribution of chemicals to SOA light absorption; 4. Q uantify MAC of POA; 5. E valuate the forcing efficiency of SOA and POA
20 Figure 1 1 Composition of a mbient particular matter
21 CHAPTER 2 MASS ABSORPTION CROSS SECTION MEASUREMENT OF SOA USING A UV VISBILE SPECTROMETER CONNECTED WITH AN INTEGRATING SPHERE Background A substantial fraction (~70% ) (Hallquist et al., 2009) of atmospheric organic carbon (20 50% of the total fine aerosol mass ) (EPA, 2004) is in the form of secondary organic aerosol (SOA), which is produced from atmospheric photochemical reactions of volatile organic compounds with atmospheric oxidants (e.g., OH, O 3 and NO x ). The effect of SOA on climate forcing, however, is poorly understood due to the complexity of chemical compositions of SOA limitation in product analysis, and the lack of measurement methods for aerosol optical properties To date, t here is a discrepanc y between mode treatment and observation for light absorption properties of SOA. Most direct radiative forcing models consider the light absorption of org anic carbon a s negligible (Forster, 2007) In contrast, l aboratory and field st udies have show n that SOA potentially would contribute to brown carbon as an important light absorbing particulate matter. For example, laboratory studies (Bones et al., 2010b;Laskin et al., 2010a) suggest that SOA formed from th e o zonolysis of terpenes can be transformed to brown carbon in the presence of ammonium ion or ammonia gas In the recent study of PM 2.5 sampled from fifteen southeastern monitoring sites in the United States Hecobian et al (2010) has reported that SOA is the one of the major sources of chromophores contributing 20 % to 30% of the light absorption coefficient for the total water so luble OC at 365nm The determination of the mass absorption cross section ( MAC ) of SOA is essential to develop a predictive model for radiative forcing of organic aerosol. However, conventional methods for measuring MAC of aerosol are limited in either
22 wa velength or the aerosol samp le preparing proce dure For example, typical optical instruments such as an optical photoacoustic spectrometer (Moosmuller et al., 1998) a particle soot absorption p hotometer ( Radiance Research, Seattle, WA), an a e thalometer ( Magee Scientific, Berkeley CA) an integrating plate photometer (Lin et al., 1973) and an integrating sphere photometer (Campbell et al., 1995) are able to measure MAC of aerosol filter samples or the aeros ol suspended in the air but operated only at a fixed wavelength (e.g., one to seven wavelength s ). L igh t absorption of actual atmospheric organic aerosol is wavelength dependent due to various functional groups present in the aerosol Thus, it is necessar y to determine MAC of organic aerosol over UV visible wavelength s covering the sunlight spectrum. As an alternative approach, an aerosol filter sample has been extracted with different solvents and measured for light absorbance using a UV visible spectrome ter covering a wide range of wavelengths. The weakness of the solvent extraction method is the modification of the chemical and physical propert ies of the aerosol. For example, the esterification of humic like aerosol products stored in alcohol solvents has been reported (McIntyre and McRae, 2005) Researchers have also shown that carbonyls in SOA react with an alcohol (e.g., methanol) which is used as a solvent forming hemiacetals and acetals (Bateman et al., 2008) It is known that organic compounds react with sulfuric acid and produce organosulfates in aerosol (Iinuma et al., 2007;Surratt et al., 2008;Liggio et al., 2005a) Acetals, hemiacetal, and organosulfates present in aerosol can also return to original parent compounds during the solvent extraction I n addition the solubili ty of aerosol products varies depending upon types of solvents. Therefore a technique for measuring the light absorption coefficient of
23 aerosol directly using filter samples covering the full spectrum of UV v isible light is needed. We proposed here to evaluate the feasibility of a UV visible spectrometer equipped with an integrating sphere for measuring MAC of SOA collected on the filter. UV visible spectrometry is not new, but is a user friendly technique based on a fundamental light absorption theory The use of a filter for collecting aerosol is also very common and efficient. Yet, it is acknowledged that the UV visible spectroscopic technique associated with the aerosol filter sample has not been fully developed for the study of an aerosol absorpt ion coefficient. In this study, both the transmittance mode (TUV IS) and reflective mode (RUV IS) have been demonstrated for the measurement of MAC of chamber generated SOA covering a wide range of wavelengths (280 800nm). T oluene (TOL) as a major anthro pogenic precursor hydrocarbon and d limonene (DL) and pinene (AP) as major biogenic precursor hydrocarbons, were photochemically oxidized in the presence of NO x and inorganic seed aerosol using a 2 m 3 Teflon indoor chamber. The influence of the precurso r hydrocarbon types seed acidity, and light source (no light vs. light) on light absorption spectra of SOA has also been investigate d Experimental Section S OA Formation SOA was generated using a 2 m 3 Teflon film chamber equipped with UV and visible lam ps. The indoor chamber operation has been reported previously (Cao and Jang, 2007). The chamber is surrounded by 16 lamps with total light emission between 280 900nm. Seed aerosol was generated using a constant output atomizer (TSI, model 3076). Inorga nic seed aerosol was made from 0.01M aqueous solution of (NH 4 ) 2 SO 4 and a mixture of 1:1 volume ratio of 0.01M solution of H 2 SO 4 and NH 4 HSO 4 solutions.
24 After injection of the seed, a known amount of NO x and HC were injected into the chamber using a gentle stream of the clean air through a manifold and the lights were turned on, marking the starting point of the SOA experiment. For the ozonolysis experiment, ozone was introduced into the chamber by passing clean air through a photolytic ozone generator (Jel ight Model 600, Irvine, CA). The particle population was measured using a scanning mobility particle sizer (TSI, SMPS Model 3080, Shoreview, MN) together with a condensation nuclei counter (TSI, Model 3025A). The gas phase concentrations of hydrocarbons were measured with an HP 5890 GC FID. The experimental conditions and the resulting data for SOA formed from the oxidation of TOL AL and AP are shown in Table 2 1 and Table 2 2 For light absorption analysis, the aerosol was collected on a 13mm diamete r filter (Borosilicate microfibers reinforced with a woven glass cloth and bonded with Teflon, Gelman Science Pallflex, Type: TX40H120 WW) using a pump (Gast, DOA P704 AA) at 13 L/min. To determine the aerosol density, the aerosol volume was measured with SMPS data and the aerosol mass collected on the filter was obtained by weighing the filter mass before and after sampling using an analytical balance (MX5 Mettler Toledo Ltd., England). Light A bsorption M easurement The light absorption spectra of aerosol collected on the filter were measured using a Perkin Elmer lambda 35 UV visible spectrophotometer equipped with a Labsphere RSA PE 20 diffuse reflectance accessory. The deuterium lamp generated light between 190 1100 nm. The wavelength interval of UV v isible spectral data was 1.0 nm and the slit width was 1.0 nm. The filter sample was supported by an in house developed holder. For the TUV IS mode, the particle sample filter was placed at the
25 entrance of the integrating sphere, and a barium sulfate pac ket for the reflection of transmitted light was attached at the exit of the integrating sphere as shown in Figure 2 1A. For the RUV IS mode, the filter sample was located at the exit of the integrating sphere (Figure 2 1C). A blank filter was used to est ablish a zero baseline. In order to calibrate the measured light absorption coefficients of SOA, the r eference aerosol made of Metani l yellow (MY, Aldrich) or the internal mixture of NaCl and MY was introduced into the chamber using a constant output atomi zer. MY aerosol was made from 3.510 5 M aqueous solution of MY. The internal mixed MY NaCl aerosol is made of the 2:8 volume ratio of 3.510 4 M MY solution to 0.01 M NaCl solution. The absorption cross section of the reference aerosol was determined f rom the MY solution of various concentrations. The resulting data for the MY aerosol is summarized in Table A 1 Results and Discussion Methodology D evelopment Theory of t ran smittance for t he p article f ilter s ample The attenuation of a light beam that pass es through a solution can be traditionally expressed by the Beer Lambert Law as: ( 2 1) w here a bs(a q ) is the absorbance of solution, is the molar absorptivity of an analyte, c is the concent ration of an analyte, is the attenuation coefficient and d is the beam path length through the cuvette For the particle filter sample, light scattering, however, arises from both the filter matr ix and particles which have discontinuity for the refractive index among the fibers of the filter materials, the air void, and the aerosol (Figure 2 1B). This light scattering
26 between interfaces contributes to the light attenuation. A mathematical formul a for the light attenuation ( OD ) of the particle in the matrix, as a sum of the optical density ( OD scat ) by scattering and the optical density ( OD abs ) by absorption, is expressed as the following form (Anderson and Sekelj, 1967) : ( 2 2) ( 2 3) ( 2 4) where I 0 and I are intensity of the incident light and the transmitted light, e and q are constants depending on particle size, wavelength, instrument, filter fiber and aerosol, H is the volume fraction of aerosol in the filter sample d is the beam path depth, and a is the attenuation coefficient as shown in e q uation 2 1 OD abs which is treated by the traditional Beer Lambert Law is linear to the concentration of ab sorbers. Equations 2 2 decouple the total OD into two distinct and independent parts, OD abs and OD scat Thus, we can evaluate the contribution of aerosol scattering to the total light attenuation. In our aerosol filter sample, H is very small compared to the filter volume fraction and is almost negligible. OD scat in e q uation 2 4 is then close to zero when H and ed H (1 H ) approach zero. The calculated H values for most aerosol samples in our study were less than 110 3 Thus, the measured transmittance ( T = I / I 0 ) bears an exponential relationship to the aerosol mass on the filter as an analog of the aerosol concentration ( e quations 2 2 and 2 3): ( 2 5)
27 In this study, the MY aerosol and the internally mixed MY NaCl particles were us ed as reference particles to test the feasibility of both TUV IS and RUV IS. The strong linear relation (R 2 : 0.95 0.97) appears in plotting of l n (1/ T ) vs. the mass of MY aerosol in the TUV IS mode at different wavelengths as shown in Figure 2 2. This re sult suggests that aerosol absorption coefficient can be mainly expressed by OD abs (Eq uation 2 3). Calibration of TUV IS d ata Absorption spectra for the aerosol collected on the filter require a calibration for the increase in path length caused by multipl e scattering in the filter fiber. In this study, MY was used for the calibration. In general, the atmospheric organic aerosol could absorb visible light up to 500 nm (Baduel e t al., 2009) The absorbing peak of MY is near 422 nm and can cover the light absorption of atmospheric organics in the visible range. The reference absorbance of MY aerosol is obtained as follows: ( 2 6) where is the cross section (m 2 ) for absorption of MY, L is the length (m) of the tube of the air sample, which is estimated from V/A n is the molecular number concentration (#/m 3 ) of MY suspended in the air and calculated as: ( 2 7) where V a is th e aerosol volume concentration (nL/m 3 ) obtained from SMPS data, is the density of the MY aerosol( 1.43 g/nL) or MY NaCl aerosol mixture (2.11 g/nL in this study), f is the mass fraction of MY in the MY NaCl aerosol N A number and Mw is the molecular weight of MY (375.38 g/mol). is related to its mo lar absorptivity (M 1 cm 1 ) described as (Lakowicz, 2006) :
28 ( 2 8) where is obtained from a plot of the aqueous MY absorbance vs the concentrations usi ng e q uation 2 1 (when d = 1 cm). Figure A 1 illustrates the value of the MY as a function of The resulting values obtained from the MY aqueous solution using e q uatio n 2 8 are then used to calculate ln(1/ T ) in e q uation 2 6. Figure 2 3A illustrates both the estimated ln(1/ T ) and the measured ln(1/ T ) for the MY NaCl aerosol when MY NaCl max ), which occurred at 434nm for the reference aerosol, is 422nm for the measured aerosol as shown in Figure 2 3A. This difference between two max indicates that the solvent (water) used for the MY solution can induce the bathochromic shift. Figure 2 3B shows the plot of t he theoretically estimated ln(1/ T ) at max = 434 nm using equation 2 6 vs. the measure ln(1/ T ) at max = 422 nm with various aerosol sample mass: slope = 1.4845 with a strong linearity ( R 2 = 0.9511). The resulting slope was used as correction factor, C in e q uation 2 11, to correct the measured ln(1/ T ) for SOA. Relationship b etween TUV IS and RUV IS Since transmittance bears an exponential relationship to the molar absorptivity ( ) of the absorber, it follows that reflectance also has the same relationshi p to The filter material used in this study has a fabric texture. The beam through a filter medium can travel typically by the diffuse reflectance mode (Figure 2 1B). The decrease in diffusion reflectance of the filter sample is caused by the absorpt ion of the sample. The linear relationship between reflectance ( R ) and transmittance ( T ) may be expressed by (Anderson and Sekelj, 1967) :
29 ( 2 9) C 1 and C 2 are constants at a given aerosol constituent and sampl e depth. Rearrange e q uation 2 9, then ( 2 10) The MY NaCl aerosol with different sample mass (e.g., 14.5, 29.18 and 41.86 g) was used to test the relation between R and T Plotting of R against T between 422 nm to 600 nm shows a strong linearity (R 2 = 0.9935, Figure 2 4) using e q uation 2 9. The resulting C 2 value is 0.0038 and is negligible which can be demonstrated by showing no change in slope and R 2 when C 2 = 0. The resulting C 1 value was 1.01 at both C 2 = 0.0038 and C 2 = 0. C 1 is very close to one, indicating that only a slight difference appears in measured T and R The C 1 and C 2 are measurement method dependent constants mainly depending on the reflection of the filter surface. Since the integrat ing sphere is able to collect most of the reflected and scattered light from the filter sample the spectrum originated from RUV IS that measures log (1/ R ) is mo re robust and less sensitive to environmental errors, compared to the TUV IS data. Consequently, RUV IS data include s a smaller error contribution than TUV IS Hence, in this study, the RUV IS data was used to calculate the absorption coefficient of SOA. Aerosol M ass A bsorption C ross S ection C alculation The volume absorption coefficient ( b v in unit of m 1 ) of aerosol in the filter is generally calculated in the transmittance mode following as: ( 2 11)
30 Here C is the correction factor and obtained from the calibration experiment, A is the aerosol sampling area (7.8510 5 m 2 in this sampling syst em) on the filter, V is the volume (m 3 ) of the air sample passing through the filter during a given sampling time, L is the length (m) of the tube of the air sample, which is estimated from V/A I 0 and I are the transmitted intensity for the blank filter with no aerosol and the filter loaded with aerosol respectively abs is the absorbance typically measured from a spectrophotometer In this study, abs is measured in both TUV IS and RUV IS modes. In order to compare TUV IS data among different SOA samp les, the mass absorption cross section (m 2 /g), MAC can be estimated by normalizing b v with the aerosol mass concentration ( M V ) (g/m 3 ) : ( 2 12) MAC is also described as (Patterson and Marshall, 1982) ( 2 13) w here m a is the absorption coefficient (m 1 ) of a bulk material, is the density (g/m 3 ) of the absorbing material. T he density of SOA is estimated from the aerosol mass and the SMPS aerosol volume data, assuming that the measured aerosol density is similar to the density of the absorbing matter in aerosol. The densit y of the aerosol is 1.07 g/cm 3 for AP SOA, 1.2 7 g/cm 3 for DL SOA and 1.3 4 g/cm 3 for TOL SOA. T he imaginary component k for a refractive index which is used for the Mie scattering calculation is proportional to m a at a given wavelength, shown in the f ollowing equation: ( 2 14)
31 Application on SOA Absorption s pectra and MAC of SOA The chamber SOA generated from the photooxidation of three different precursors (TOL, DL, and AP) in both the absence and the presence of seed aeroso l was collected on the filter and analyzed to measure the absorption spectrum between 280 nm and 800 nm. Figure 2 5 illustrates the absorption spectra of various SOA. Because there was no significant absorption beyond 580 nm for all SOA, absorption spect ra were mainly focused on the wavelengths between 280 and 580 nm. All SOA filter samples strongly absorb UV light and the absorption intensities dramatically increase as the wavelength becomes shorter. In the visible range, the intensity of SOA light abs orption varies with the types of precursors. The color of the TOL SOA filter sample was yellowish indicating that TOL SOA can absorb visible light while no color was observed for the AP SOA in this study. The fresh DL SOA showed no color for the first 20 minutes after collection on the filter, but it gradually turned to yellow during the next three hours standing in room air. The absorption spectra o f DL SOA were shown in Figure A 2. The UV visible spectrum of SOA is originated from a variety of oxyg enated products. The chemical functional groups present in SOA provide useful insight for SOA spectra. For example, both AP SOA (Jang and Kamens, 1999;Camredon et al., 2010) and DL SOA (Grosjean et al., 1993;Glasius et al., 2002;Leungsakul et al., 2005) contain non conjugated oxygenated compounds. AP SOA includes pinonaldehyde, norpinonic acid and hydroxyl pinonadehydes, pinonic acid, and pinic acid and DL SOA includes ke to limononaldehyde, keto limonomic acid and keto limonalic acid. The n transitions absorption of such non conjugated carbonyl and carboxylic acid products
32 typically appears between 280 and 300nm. The absorption of TOL SOA (Figure 2 5 A ) between 400 and 500 nm is most likely contributed from transitions from conjuga ted double bond (e.g. aromatic ring) and n transitions from chromophoric groups (e.g., NO 2 C=O) and auxochromic groups (e.g., OH,). For example, 4,6 dinitro o cresol, one of the major TOL SOA products (Jang and Kamens, 2001b) can show a yellow color corresponding with the visible absorption spectrum beyond 400nm (Howard, 1991) Figure 2 5D shows the UV visible spectrum of the DL SOA produced from the ozonolysis of DL in the absence of inorganic seeds. The absorption peak at 428nm gradually increased for 7 hours during the experiment A poss ible explanation for such an absorption peak is the formation of conjugated products through particle phase reaction s. A proposed scheme is shown in F igure 2 8 Both limononaldehyde and 5 oxo limononaldehyde are major products from ozonolysis of d limone ne and can produce colored products through the acid catalyzed aldo l condensation between these two products, followed by dehydration and a double bond rearrangement. Figure 2 6 illustrates the comparison of the mass absorption cross section ( MAC ) at 350 nm and 450 nm among three different SOA (TOL DL and AP SOA). Overall, photooxidat ion of TOL gives the highest MAC followed by DL SOA and AP SOA for both neutral and acid seeds. S ince the SOA absorption property is directly related to oxygenated prod ucts originated from the oxidation of precursors, the chemical structure of precursors are expected to affect MAC of SOA. TOL is expected to produce SOA with more conjugated functional groups and higher MAC due to three double bonds in TOL. AP has an end ocyclic double bond and DL has two non conjugated double bonds.
33 Data in Figure 2 6 show s that MAC of TOL SOA at 350nm is 15 times higher than that of fresh DL SOA and 20 times higher than AP SOA. Effect of i norganic s eeds T he effect of the preexisting i norganic seed on SOA absorption coefficients is shown in Figure 2 6 In the UV range, the presence of both neutral and acid ic inorganic seeds considerably increases light absorption of TOL SOA and DL SOA. For example, with (NH 4 ) 2 SO 4 seed aerosol, MAC of TOL SOA at 350 nm increase d by 3.1 times compared to a no seed condition. In the same way, MAC of TOL SOA with acidic seed aerosol (NH 4 HSO 4 :H 2 SO 4 = 1:1) is 2.1 times higher than that of TOL SOA without seeds. The effect of the seed aerosol on MAC is the highest with TOL. For both TOL SOA and DL SOA, the MAC value is higher in the presence of (NH 4 ) 2 SO 4 than that in the presence of acidic seed aerosol. No significant difference appears between the seed and no seed condition for AP SOA. The brown or yellow color products have been reported for SOA from ozonolysis of d limonene in the presence of amino acids, (NH 4 ) 2 SO 4 or the mixed vapor of NH 3 and HNO 3 (Bones et al., 2010b;Laskin et al., 2010a) B rown color product s ha ve also been reported by reaction between glyoxal aerosol and amino acids (Galloway et al., 2009) Such studies suggest that light absorbing materials would be originated from C N containing compounds (Galloway et al., 2009) The possible pathways to form the light absorbing specie s include aldol condensation, imidazole formation, Leuckart / Mannich reactions and pyridinim ions formation (Nozie`re and Esteve 2007;De Haan et al., 2009;Bones et al., 2010b) However, explanations for the SOA co lor change due to inorganic species are still unclear and require the future investigation.
34 Effect of light s ource To investigate the effect of light source on SOA light absorption, several sets of SOA experiments have been conducted with and without lig ht for both TOL and AP systems. Figure 2 7 summarizes the impact of light exposure on MAC of SOA under various experimental conditions of different oxidation, precursor HCs and preexisting seed aerosol. For DL SOA, the time profiles of MAC ( = 428nm) of the photoirradiated SOA in the presence of NO x were compared to those of the SOA produced through ozonolysis under no light (Figure 2 7A). MAC values of DL ozone SOA at 428 nm ( the small peak in the visible range, shown in Figure 2 5D) contin uously increases, while those of photochemically irradiated DL NO x SOA gradually decrease. For TOL SOA, to determine the effect of light source on MAC values, the MAC value at 120 minutes (photooxidation with light) was subtracted from MAC values at 240 m inutes (oxidation both with and without light) under the same experimental conditions. For oxidation experiments without light, the light source had been turned off immediately after SOA was sampled at 120min. The reduction of MAC ( = 350nm) for TOL SOA was observed with light exposure compared with the MAC without a light source, as shown in Figure 2 7B. We conclude that the light source influences the light absorption of atmospheric organic compounds in two opposite ways. One is t he formation of light absorbing products from the atmospheric oxidation of precursor HCs. For example, the colored products such as nitrophenols in TOL SOA are formed from toluene photooxidation with OH radicals in the presence of NO x Opposite to the col or production in SOA, the bleaching effect of the light source progresses in SOA through further photoirradiation of light absorbing organics. For example, the conjugated organic compound as an
35 oxygenated SOA product can be further photodissociated with the UV visible light which excites organic compounds through a transition. The n transitions in carbonyls is also able to decompose carbonyls via Norri sh type I and II mechanisms (Calvert and Pitts, 1966.) As previously reported by Mang et al (2008, photodegradation of the DL SOA collected on a glass fiber filter produced large amounts of small molecules such as CO, CH 4 acetaldehyde, acetaone and other VOCs, resulting in off gassing of those organics from SOA into the air. Conclusion and Atmospheric I mplication In this study, light absorption of a variety of SOA ( e q uation 2 12) generated using the indoor chamber has been measured using the TUV IS and the RUV IS. The resulting MAC value enables estimation of absorption coefficient ( m a in e q uation 2 13) of SOA. Consequently, the imaginary part k of the reflective index of SOA can be estim ated from m a values using e q uation 2 14. For the chamber generated SOA, k values are in the order of 10 4 to 10 2 at 350 nm. The k values of SOA are relatively smaller than those for black carbon which ranges between 0.63~0.79 at 550 nm (Bond and Bergstromb, 2006) and atmospheric brown carbon which is about 0.27 at 550 nm (Alexander et al., 2008) Compared with black carbon and brown carbon, t he light absorption of freshly chamber produced SOA of this study is relatively weak even in the UV range. This study has focus ed on MAC values of SOA formed within 1 7 hours after the reaction began using the 2 m 3 chamber that facilitates the artificial UV v isible light source However, MAC values of the ambient SOA that is photoirradiated under actual sunlight for longer time (on average 6 days of aerosol life) would be different than those of the indoor chamber SOA In the future, the outdoor chamber or the ambient field
36 studies using TUV IS and the RUV IS methods are necessary to investigate the light absorption of aged SOA Both TUV IS and the RUV IS techniques are simple because they do not require solvent extraction and allow direct measurement of the MAC value of aerosol filter samples The TUV IS and the RUV IS techniques are reproducible and sensitive even to a small amount of aerosol mass. The detection limit of this method not only depends on molar absorptivity of a variety of analytes but also is a function of wavelength showing a higher detection limit in the longer wavelength. For example, the lowest detectable aerosol mass is 5 ng at 422 nm for ambient particle collected at the sampling site at Gainesville, FL. The major uncertainty in both the TUV IS and the RUV IS methods is the scattering of light caus ed by the filter material Different filter materials can produce different scattering effects on the calibration curve (Figure 2 3).
37 Table 2 1. Experimental conditions and the resulting SOA data for photooxidation of toluene usi ng the 2 m 3 Teflon film chamber a Precursor HCs No. Seed RH % Initial NO x (ppb) Initial O 3 (ppb) Initial HC (ppb) 3 ) f b SOA mass conc. 3 ) L c (m) ln(1/R) b (Mm 1 ) MAC (m 2 g 1 ) m 10 Toluene T1 None 42 62 NA 207 149 NA 46.0 4377 7. 5 E 02 26.43 0.574 0.214 T2 AS 70 62 NA 291 197 0.54 37.6 3925 1. 7 E 01 66.43 1.767 0.660 T 3 Acid 40 70 NA 193 155 0.60 34.5 4118 1.1 E 01 41.62 1.208 0.451 T4 None 43 70 NA 202 188 NA 47.2 5045 8. 9 E 02 27.23 0.577 0.215 T5 AS 72 69 NA 195 322 0.63 44.1 4304 1. 9 E 01 66.73 1.514 0.565 T6 Acid 35 68 NA 193 131 0.65 30.9 4503 1.0E 01 34.85 1 .128 0.421 a. Temperature: 294 298 K. b. The seed fraction of SOA was calculated from SMPS data at a given reaction time, 2h for T1 T6. c. The length of the tube for the air sample that includes the suspended aerosol is obtained by dividing the air sampl e volume with the sample area on a filter.
38 Table 2 2 Experimental conditions and the resulting SOA data for oxidation of d limonene and a pinene using the 2 m 3 Teflon f ilm c hamber a Precursor HCs No. Seed RH (%) Initial NO x (ppb) Initial O 3 (ppb) In itial HC (ppb) HC 3 ) f b SOA mass conc. 3 ) L c (m) ln(1/R) b (Mm 1 ) MAC (m 2 g 1 ) m 10 d Limone L1 None 35 62 NA 188 1058 NA 409.9 1047 1.06E 02 15.69 0.038 0.014 L2 AS 70 70 NA 183 803 0.64 362.5 924 2.25E 02 37.82 0.104 0.037 L3 Acid 37 72 NA 193 986 0.49 308.7 706 8.82E 03 19.39 0.063 0.022 L4 None 56 NA 126 189 1061 NA 588.5 936 8.73E 03 14.48 0.025 0.009 L5 AS 69 NA 107.3 158 708 0.48 243.6 1539 8.20E 03 8.26 0.034 0.012 L6 Acid 49 NA 140 207 1163 0.46 449.2 1698 1.2 7E 02 11.64 0.026 0.009 Pinene A1 None 37 68 NA 142 710 NA 133.5 2078 5.18E 03 3.86 0.029 0.009 A2 AS 70 71 NA 159 772 0.5 164.6 1030 5.42E 03 8.15 0.05 0.015 A3 Acid 38 62 NA 142 685 0.68 140.7 1059 3.71E 03 5.44 0.039 0.012 a Temperature : 294 29 8 K. b. The seed fracti on of SOA was calculated from SMPS data at a given reaction time, 1h 40 min for L1 L6, 2h for A1 A3. c. The l ength of the tube for the air sample that includes the suspended aerosol is obtained by dividing the air sample volume with the sample area on a fi lter.
39 Figure 2 1. Schematic diagram for the beam pathway in both transmittance and reflectance mode : A) transmittance mode B) beam interaction with particles and filter matrix and C ) reflectance mode
40 Figure 2 2 ln(1/T) plotted vs. metanil yello w aerosol mass collected on filter at three different wavelengths (422 nm, 300 nm, and 500 nm)
41 Figure 2 3 Measured ln(1/ T ) vs theoretical ln(1/ T ) : (A) for the MY NaCl aerosol as a function of wavelength (B) for different MY NaCl mixture at 434 n m
42 Figure 2 4 Reflectance is plotted against transmittance at 422~680 nm for MY NaCl aerosol with different particle mass (41.86, 29.18, and 14.5). The linear regression lines with an intercept (C1 = 1.0108) and with intercept = 0 (C1 = 1.0064).
43 Figure 2 5 UV visible spectra of SOA freshly generated from photooxidation of different hydrocarbons: (A) toluene using data T1 T3 in Table 2 1 (B) d limonene using data L1 L3 in Table 2 2 and (C) pinene using data A1 A3 in Table 2 2 (C). AS deno tes the (NH 4 ) 2 SO 4 seed condition and acid denotes the acidic seed (NH 4 HSO 4 :H 2 SO 4 =1:1) condition. UV VIS spectra of DL SOA (L4 in Table 2 2) formed from ozonolysis of d limonene without seed at different oxidation time (D). For comparison, the specific abs orbance is reported by dividing ln(1/ R ) with the mass of organic aerosol collected on the filter.
44 Figure 2 6 MAC values of SOA formed from three hydrocarbon precursors at =350 nm and 450 nm: (A) t oluene (B) d limonene and (C) pinene under different seed conditions. AS: (NH 4 ) 2 SO 4 Acid: seed aerosol made of NH 4 HSO 4 : H 2 SO 4 =1:1
45 Figure 2 7 E ffect of light on MAC of d limonene and toluene S OA. (A) The time profile of MAC (at =428 nm) for DL SOA formed from ozonolysis and phorooxidation of d limonene in the presence of NO x without seed (L4) ( B ) The reduction ratio of MAC for TOL SOA due to the exposure to light source at di fferent seed co nditions (T1 T6) The reduction ratio is obtained by dividing the MAC ( MAC = MAC .at 240 minutes oxidation MAC at 120 minutes photooxidation) with MAC at 120 min photooxidation.
46 Figure 2 8 P roposed example of conjugated compound formation through a erosol phase reaction in d limonene SOA
47 CHAPTER 3 T HE SOA FORMATION MODEL COMBINED WITH SEMIEMPIRICAL QUANTUM CHEMISTRY TO PREDICT UV VIS ABSORPTION OF SECONDARY ORGANIC AEROSOLS B ackground Light absorption of a erosol is mainly determined by its chemic al composition. The chemical composition of SOA is complex due to the multi generation al gas phase reactions of VOCs with atmospheric oxidants (e.g., O 3 OH, NO 2 and NO 3 ) and aerosol phase reactions Very few SOA products have been identified although m uch effort has been expended in this direction. For example, in a study by Forstner et al. (1997) identified SOA products from the photooxidation of aromatic hydrocarbons were only 15~3 0% of the total extract able SOA. Also, i n a recent study by Sato et al. (2007) using liquid chromatography mass spectrometry ( LC MS ) only ~1wt% of the total mass of aerosol products from the photooxidation of toluene were identified. For pinene SOA, relatively complete product identification has been achieved. Jaoui and Kamens (2001) have identified more than 80wt% of pinene SOA products The possible difficulty in characterizing SOA product s is the unidentified aerosol phase reactions which lead to the form ation of oligomeric compounds H igh molecular weight structures are ubiquitous in SOA s compositions and t hey may account for a significant portion of the SOA mass (Gao et al., 2004;Tolocka et al., 2004;Kalberer et al., 2004) T he great complexity of the SOA composition inhibits an improved understanding of the light a bsorption properties of SOA. Thus, the development of a model to overcome this difficulty would enhance our knowledge of how the composition of SOA a ffects its optical properties.
48 A f ew modeling approaches have been developed to predict the light absorpti on spectrum of organic aerosols Lund Myhre and Nielsen (2004) calculated the absorption index of synthetic organic aqueous mixtures by an empirical approach using the measured absorption index of individual acids Sun et al. (2007) employed the Band gap an d Urbach tail relationships to predict UV visible absorption spectra of various organic aerosols. Moosmulle r et al. (2011) calculated the absorption index of brown carbon using a damped simple harmonic oscillator model. However, t o the best of our knowledge, development of a model for light absorption of SOA based on the chemical composit ion has not yet been explored. T he prediction of the UV v isible a bsorption s pectrum of SOA on the basis of first principles begins with spectr a of individual organic compounds. To calculate the light absorption spectra of organic compounds, quantum chem istry has been frequently employed P opular computational approaches include ab initio quantum chemistry methods (Runge and Gross, 1984) and semiempirical methods (Zerner, 1991;Dewar et al., 1985) In this study, prediction o f UV visible spectra of a variety of organic compounds was approached with the AM1 ( Austin Model 1 ) method (Dewar et al., 1985) based on the NDDO ( N eglect of D iatomic D ifferential O verlap) semiempirical quantum chemistry approximation since it provide s spect ral predictions with reasonable accuracy and inexpensive computation. The NDDO based methods have been prove n to be a useful tool using configuration interaction to calculate the spectroscopic properties of many organic compounds (Fabian et al., 2002;Matsuura et al., 2008) T he SOA chemical composition used for the SOA light absorption model was predicted from t he SOA formation model by explicitly analyzing the gas phase kinetics
49 of a precursor VOC Jang et al. (2006) in their recent study incorporated both heterogeneous reaction s of organic products and gas particle partitioning into a SOA formation model named PHRCSOA (Partitioning Heterogeneous Reaction Consortium Secondary Organic Aerosol Model) The PHRCSOA model has been evaluated for pinene ozonolysis (Jang et al., 2006) and toluene photo oxidation (Cao and Jang, 2010) in the presence of inorganic seeds. This model tends to reasonably predict precursor hydrocarbon decay, NO NO 2 conversion, ozone formation and SOA mass. In this study, a n NDDO based s emi empirical quantum chemistry method and the PHRCSOA model are implemented to predict the UV visible absorption spectrum of SOA. The absorption spectrum of SOA is calculated by taking the sum mation of spectr um of each individual SOA produc t. The m odel was tested and evaluated by chamber generated SOA. Toluene, which is a major anthropogenic precursor, or pinene a major biogenic precursor, was photochemically oxidized in the presence of inorganic seed aerosol using a 2 m 3 indoor Teflon chamber. The influence of NO x on light absorption of SOA was also investigated. Experimental Section SOA Formation All experiments were conducted in a 2 m 3 indoor Teflon film chamber under UV visible light A detailed description of the procedure s for c hamber experiments and sample measurements has been provide d elsewhere (Cao and Jang, 2010) T he chamber is s urrounded by 16 lamps with light emission in a range between 280nm and 900nm. Prior to each SOA experiment, the chamber was flushed with clean air from a clean air generator. The inorganic seed aerosols were injected into the chamber by nebulizing an ino rganic aqueous solution using a Constant Output Atomizer (TSI, Model
50 3076, Shoreview, MN). The inorganic seed was made from a 0.01 mol L 1 aqueous solution of Mg SO 4 After injection of the seed, a known amount of NO was added into the chamber from a certi fi ed NO tank Once the NO x concentration became stabilized, a known volume of volatile hydrocarbon (HC) precursor was injected and the lamps were turned on, which marked the starting point of the SOA experiment. T he HC s used in this study include t oluene (99%, Aldrich) and pinene (98%, Aldrich) The particle population was measured continuously using a scanning mobility particle sizer (TSI, SMPS Model 3080, Shoreview, MN) together with a condensation nuclei counter (TSI, Model 3025A). The gas phase con centrations of HCs were measured with a HP 5890 GC FID. Ozone and NO x concentrations were monitored using a photometric ozone detector analyzer (model 400E, Teledyne Instruments, San Diego, CA) and a chemiluminescence NO x detector (model 200E, Teledyne In struments, San Diego, CA). Temperature and humidity were measured with an electronic thermo hygrometer (Hanna Instruments, Italy). The humidity of the chamber experiments ranged between 43 ~ 47 %, while temperature was 294 298 K with an increase of around 4 K for the duration of the SOA experiment. SOA UV V isible S pectra R ecording The resulting aerosol was collected on a 13mm diameter filter (Borosilicate microfibers reinforced with woven glass cloth and bonded with Teflon, Gelman Science Pallflex, Type: TX 40H120 WW) using a pump (Gast, DOA P704 AA). The UV visible spectra of aerosol s collected on the filter were directly measured using a Perkin Elmer lambda 35 UV visible spectrophotometer equipped with a Labsphere RSA PE 20 diffuse reflectance accessory. The deuterium lamp generated light between 190 1100 nm. The wavelength interval of UV v isible spectral data was 1.0 nm T he filter sample
51 was supported in a home made filter holder and was located at the exit of the integrating sphere upon which light ab sorptions measurements were performed (see Figure 3 1). For the zero baseline, a blank filter was used. The aerosol absorption measurement using filter samples is simple because this technique does not require solvent extraction. The detection limit of this method not only depends on molar absorptivity of a variety of analytes but also is a function of wavelength showing a higher detection limit in the longer wavelength. The lowest detectable aerosol mass is 6 nm for chamber generated toluene SOA. The detailed measurement procedure of SOA light spectra can be found in our previous study (Zhong and Jang, 2011) The experimental conditions and the resulting data are shown in Table 3 1. Results and Discussion Measu rement of SOA UV V isible S pectra The light absorbance of organic aerosol, A SOA is defined as: ( 3 1) where I 0 and I are the transmitted intensities of light at a specific wavelength ( ) without and with aerosol, respectively. A SOA can be obtained by converting the experimentally measure d absorbance abs SOA : ( 3 2) where C is a correction factor with a value of 1.4845 obtained from our previous study (Zhong and Jang, 2011) f o r the purpose of eliminat ing the absorbance caused by filter material scattering.
52 Theoretical C alculatio ns of SOA UV V isible S pectra The light absorbance of SOA ( A SOA ) can also be calculated by taking the sum of each individual spectr um of a SOA product a nd described as : ( 3 3) where A k is the absorbance of individual species ( k ) of SOA. The light absorbance of individual species A k is calculated as (Lakowicz, 2006) ( 3 4 ) In this expression, is the cross section area for absorption in m 2 V is the volume (m 3 ) of air passing through the filter during a given sampli ng time, and A is the filter surface area exposed to the sample (7.85 10 5 m 2 in this sampling system ) k is the molar absorptivity (L mol 1 cm 1 ) which is predicted using a semi empirical quantum chemistry method n k is the molecular number concentratio n(# m 3 ), which is calculated from the mass percentage of component k : (5) MW k is the molecular weight (g mol 1 ), N A is the Avogadro s number and F k is the mass percentage. F k is determined by the mass balance of organic compou nds in SOA OM T is the total SOA mass concentration and is predicted by the SOA formation model SOA F ormation M odel The SOA formation model, (i.e., the PHRCSOA model ), which includes a heterogeneous reaction model and a partitioning model, is used to pre dict the SOA mass from heterogeneous reaction ( OM H ) and partitioning ( OM P ). The detailed description of the PHRCSOA model can be found in the work of Cao and Jang (Cao and
53 Jang, 2010) The model starts from the gas phase oxidation of VOCs described by the master chemical mechanism ( MC M 3.2 http://mcm.leeds.ac.uk/MCM ). A chemical kinetic solver ( Morpho ) was used to run the gas phase kinetic reactio n model (Jeffries et al., 1998) The comparisons between model simulations and experimental data for toluene or pinene decay, ozone formation and NO x consumption at different NO x levels are shown in Figures B 1and B 2 (supporting information). Among the products predicted from the MCM model ( Figure 3 2) the oxygenated products that can significantly contribute SOA mass are chosen A total of 147 products from the oxidation of toluene and 1 29 products from the oxidation of pinene were used to predict the SOA mass. The selected gas products are lumped into 20 groups according to their vapor pressures and reactivity in the aerosol phase The vapor pressure ( bar ) is estimated by the following equation (Schwarz enbach et al., 1993) : ( 3 6) where vap is the entropy of vaporization ( Jmol 1 K 1 ) which is calculated by the rule modified by Zhao et al (1998) with parameters related to molecular geometry. T b is the boiling point (K) which is estimated from a group contribution metho d originally developed by Joback and Reid (1987) with a modifi ed equation and group contribution parameters (Stein and Brown, 1994) T is the ambient temperature (K). R is the gas constant (8.314 JK 1 mo l 1 ) The vapor pressures of gas phase organic products are categorized into five groups labeled by i with i = 1 through 5 corresponding to 1.3 10 4 1.3 10 3 1.3 10
54 2 1.3 10 1 and 1.3 Pa respectively. Each group is then divided into four subgroups ( j ) according to the reactivity in the aerosol phase ; j = PO, H s, H m, and H f correspond to partitioning only, slow reactivity, middle reactivity and fast reactivity. The reactivity can be ranked by the order of functional groups: multifunctional aldehydes > aldehydes > ketones > carboxylic acids. The photochemical reactions then can be expressed as, Precursor + OH 1, PO S 1, PO + 2 PO S 2 PO + + 1, H f S 1, H f + 2 H f S 2 H f + + 1, H m S 1, H m + 2 H m S 2 H m + + 1, H s S 1, H s + 2 H s S 2 H s where i,j is the mass based stoichiometric coefficient for the lumped group S i,j The aerosol phase reacti ons producing OM H th r ough dimerization of organic compounds are assumed to be second order reactions and t he reaction rate is estimated from the semiempirical model which has been addressed in the previous study (Jang et al., 2006) The resulting OM H is then taken as the preexisting absorbing material and integrated into a SOA partitioning model to calculate OM P The SOA partitioning model employed in this study was developed by Shell et al. (2001) and has b een used in the regional air quality model ( CMAQ 4.7). The summation of OM H and OM P gives the total mass of SOA ( OM T ): ( 3 7) T he detailed derivation and mathematical expressions for OM H and OM P can be found in the previous work (Cao and Jang, 2010) of our group. Figure B 3 shows the simulated OM T levels under different NO x conditions. T he error range for the difference between the predicted and observed aerosol mass concentration is 7.3% at the 95% confidence level for 20 samples (Cao and Jang, 2010)
55 SOA P roducts Based on the contribution s to the total SOA mass m ajor SOA species ( k ) are selected among the products predicted using the m aster c hemical m echanism kinetic model. The distributio n of the selected SOA products is presented in Table 3 2 for toluene SOA and in Table 3 3 for pinene SOA In our calculation, the selected SOA compounds contribute more than 97% of total SOA mass for both toluene pinene SOA. For calculating light absorption spectra of glyoxal ( GLYOX ) and methlyglyoxal ( MGLYOX ) their oligomeric forms are used based on the structures known in literature (Liggio et al., 2005b;Hu et al., 2007) 2 Dihydroxymethyl [1,3] dioxol ane 4,5 diol is used for glyoxal, and 2 Dihydroxymethyl 2,4 dimethyl [1,3] dioxolane 4,5 diol for methylglyoxal. For UV v isible spectra of other compounds that contribute to OM H their monomeric structures are used The chemical structure of toluene SOA products and that of pinene SOA products can be found in Table B 1 and Table B 3. UV V isible S pectrum P rediction All g eometr y optimization s and calculations of spectroscopic parameters for individual molecules were performed using the NDDO based AM1 Hamiltonian in conjunction with the pair excitation configuration interaction (PECI) method. The calculations were done with the semiempirical program VAMP (Clark et al., 2002) (Accelrys Inc) In the PECI method, excited states are calculated by including all single and double excitations in which a complete electron pair i s promoted. The calculated wavelength of the maximum absorption ( max ) and oscillator strength ( f ) were used to predict the UV visible absorption intensity. f is related to the molar absorptivity (L mol 1 cm 1 ), by the following equation (Belay, 2010) :
56 ( 3 8) where (cm 1 ) is wavenumber ( =1/ ). UV visible spectra generally have a Gaussian band shape (Barker and Fox 1980) Using the Gaussian function, the spectrally integrated molar absorptivity can be expressed as: ( 3 9) where max (L mol 1 cm 1 ) is the maximum molar absorptivity and FWHM (cm 1 ) is the full width at half maximum T he values of FWHM of many SOA products are not available because these compounds are either difficult to synthesize or chemically too unstable The sensitivity analysis for the predicted spectrum due to variation of FWHM ( 56~80nm ) is performed for both toluene SOA and pinene SOA The model is highly sensitive to FWHM as shown in Figure 3 4 When the same or a similar structure with a SOA p roduct is found in the spectral data base, literature values were used for FWHM Otherwise, FWHM values were fixed at 6 8 nm to minimize the difference between the calculation and observation. For selected nine compounds, the predict ed spectroscopic parameters using the DNNO based AM1 method are compared with those known in the literature (see Table 3 4 and Figure 3 3). At the 95% confidence level, the margin of error for the difference between predicted and observed data is 25.4% for max and 7.4% for max Matsuura et al. (2008) reported an error range of 4.4% for max by comparing the observations and calculations for 49 dyes.
57 Simulation R esults Light a bsorption s pectra of t oluene SOA The measured and calculated UV visible absorption spectra of toluene SOA und er different NO x are shown in Figure 3 5 (a c). NO x concentrations can affect the preferred reaction pathway for toluene oxidation in the gas phase (Cao and Jang, 2008;Song et al., 2005;Ng et al., 2007b) For examp le, under low NO x condition s multifunctional carbonyls from toluene oxidation can contribute more to SOA mass For high NO x conditions, higher concentrations of organic nitrates and nitro compounds appear in SOA. When a nitro group is attached to an aro matic ring or a conjugated max and the intensity of the absorption spectrum of a molecule. Under the NO x levels of this study, the light absorbance of toluene SOA tends to become gradually hi gher as the NO x concentration increases but the effect of NO x on measured aerosol spectra is not substantial. The model simulation also supports experimental observations showing that the formation of nitrophenols as light absorbing compounds is not sens itive to NO x levels. For example, the formation rate (2.08E 12 cm 3 molecules 1 s 1 at 298K) of 3 methyl 6 nitro 1,2 benzenediol ( MNCATECH in Table 3 2 ) increases with NO x concentrations but it is also highly reactive with atmospheric oxidants such as OH ( 6. 83 E 12 cm 3 molecules 1 s 1 at 298K) and NO 3 (5.03E 12 cm 3 molecules 1 s 1 at 298K) resulting in a weak sensitivity to NO x Within the errors associated with the measured UV spectrum the model output reasonably agrees with the measured spectrum for low NO x condition but underestimates the absorbance of toluene SOA for high and middle NO x condition (Figure 3 5 ). O ne of the possible explanation s for the difference between the model s
58 predictions and the experimental measurements is that the model has fail ed to account for some light absorbing species with high molecular weights. Some oligomeric compounds formed t hrough the aerosol phase reaction s such as hydration, esterification, hemiacetal/acetal formation, and aldol condensation (Jang et al., 2002) might contain a long conjugated structure that can increase the absorption intensity in the visible range. In addition to aerosol phase reactions of organic compounds, the dissociations of dinitrocresols can influence aerosol absorption spectra. The dissociated forms of dinitrocresols usually absorb light with longer wavelength (visible range) and stronger intensity than the undissociated form s depending on aerosol s acidity. Figure 3 5 d shows spectra of individual products of toluene SOA under middle NO x condition (T2 in Table 3 1). Among toluene SOA products, MNCATECH as a nitrophenol contributes about 60% of aerosol UV light absorption and about 95% of visible light absorption at different NO x levels. MNCATECH accounts for nearly one third of the total toluene SOA mass and it has a strong absorption band tailing in the visible region (yellow) Other light absorbing species such as MNNCATCOOH, NC4MDCO2H and GLYOX polymer contribute mainly to UV light absorption. Although GLYOX polymer has a high mass percentage in toluene SOA (23~ 37%), it absorbs only UV light with low absorbance values (2~5% of aerosol UV absorption). Many laboratory studies have also reported the presence of nitrophenols in toluene SOA (Forstner et al., 1997;Jang and Kamens 2001a;Hamilton et al., 2005;Sato et al., 2007;Fang et al., 2011) For example in a recent study of characterization of toluene SOA using thermal desorption/tunable vacuum ultraviolet photoionization time
59 of flight aerosol mass Fang et al. (2011) have identified 4 nitro o cresol and its isomers as major components in toluene SOA. T he nitro substituent on a phenol ring enhances the max value and also shifts the absorption spectrum to longer wavelengths. Jacobson (1999) has pointed ou t the importance of nit ro substituted aromatic compounds in atmospheric aerosols due to their high absorption of UV light based on the literature absorption spectra. Nakayama e t al. (2010b) have also reported that nitrocresols might be a plausible source of the light abso rption for toluene SOA. Light absorption spectra of pinene SOA In a manner similar to the analysis of toluene SOA, the aerosol phase products in Table 3 3 were used to predict the light absorption spectra of pinene SOA. The calculated spectra are in good agreement with the measured spectra (Figure 3 6) In this study, n o NO x effect on li ght absorbing spectra was observed for pinene SOA although NO x can influence the product distribution through the modif ication of gas phase reactions (Ng et al., 2007a;Presto et al., 2005) Both model simulations and experimental data show that the light abs orb ing intensity of pinene SOA is much lower than that of toluene SOA. For example, the light absorbance of toluene SOA is five times higher than that of pinene SOA at 300nm, but twenty times higher at 400nm. Unlike toluene pinene SOA product s do not contain aromatic or long conjugated moieties which along with nitro functional groups can be potential chromopho r es. pinene SOA products produced under high NO x condition (A1 in Table 3 1) are shown in Figure 3 6 c. These six products contribute about 80% of the total SOA light absorption. The most frequently found functional group of these p rodu cts is the non conjugated carbonyl group (C=O), which has a major absorption near 200nm corresponding to a transition. The calculation of max using the
60 DNNO based AM1 method also shows that non conjugated SOA carbonyl products have absorptions near 20 0nm, as shown in Table 3 3. Conclusion A model to predict UV visible absorption of SOA has been developed using a SOA formation model (PHRCSOA) and semiempirical quantum chemistry (NDDO AM1) For the first time, the feasibility of semiempirical qua ntum chemistry has been explored for SOA systems with complex chemical composition. Overall, the light absorption of toluene SOA is higher than that of pinene SOA. The predicted NO x effects on toluene SOA and pinene SOA also agree with the experimental observations. There are several uncertainties in the current model approach For example, SOA products obtained from the MCM kinetic model need to be revised when an improved mechanism of SOA formation becomes available The structure of SOA oligomers needs to be experimental ly identified although the PHRCSOA model is able to predict the fraction of oligomers in the total SOA mass. In addition to th e SOA formation model, the selection of different quantum chemistry methods for spectroscopic parameters can influence the model structure and the spectrum of SOA The current quantum chemistry model of this study does not account for the spectrum shift d ue to coexistence of other organic matter in SOA, which can act as a solvent.
61 Table 3 1 Experimental conditions and the resulting SOA data for photooxidation of t pinene using the 2 m 3 Teflon film chamber a HCs No. RH % Initial NO x (ppb) Init ial HC (ppb) V seed 3 ) HC b (ppb) OM T b 3 ) Mass ratio of org to inorg Sampling time c (min) V / A d (m) Sampled SOA mass e Toluene T1 44.3 104 233.5 27 81 53 1.96 27 5964 15.2 T2 43.7 49 250 21 95 50 2.4 31 6749 17.4 T3 44.8 24 244 26 79 38 1.42 42 9171 16.8 Pinene A1 45.9 105 148.7 118 112 141 1.2 8 1682 14.3 A2 46.7 26 150.3 123 99 90 0.73 11 2293 10.8 a: Temperature: 294 298 K. b: The average consumed hydrocarbons ( ) and formed OM T during sampling. c: The SOA s ampling started at reaction time of 120min. d: V is the volume of air drawn through the filter during a given sampling time and A is the area of the sample spot. e: the collected SOA mass in the filter.
62 Table 3 2 Selected products of toluene SOA an d their mass percentages at different NO x conditions Group (i, j) k Products name a MW k max b (nm) f b F k c (%) H NO x M NO x L NO x (T1) (T2) (T3) 1, PO 1 MNCATECH 169 330, 228 0.05, 0.40 26.27 30.48 26.14 1, PO 2 MNNCATCOOH 281 270, 189 0.3, 0.31 2.95 6.36 6.47 1, PO 3 DNCRES 191 321, 222, 0.04, 0.82 2 0.88 0.32 1, H m 4 TLE MUCOOH 190 177, 165 0.04, 0.05 1.04 3.04 4.41 1, H m 5 TLEMUCNO3 190 177, 161 0.04, 0.03 2.4 1.57 1.1 2, PO 6 TLBIPEROOH 174 216, 213 0.19, 0.11 2.13 8.68 14 2, PO 7 TLBIPERNO3 174 415, 233 0.06, 0.07 3.85 3.46 2.78 2, H s 8 NC4MDCO2H 159 228, 221 0.1 5, 0.21 6.71 3.19 1.38 2, H f 9 TLEMUCCO 156 202, 148 0.35, 0.03 0.53 0.81 1.09 3, PO 10 TOL1OHNO2 153 316, 222 0.04, 0.17 0.77 0.51 0.23 3, H f 11 ACCOMEPAN 207 187, 174 0.26, 0.25 2.29 10.34 7.67 4, H m 12 MALDIALPAN 161 190, 176 0.27, 0.26 0.59 0.4 8 0.41 4, H m 13 C5COO2NO2 175 227, 184 0.59, 0.22 0.99 1.03 1.17 5, H m 14 MALANHY 98 230 0.17 2.26 1 1.01 5, H m 15 C5DICARB 98 223, 162 0.65, 0.09 0.55 0.21 0.14 5, H m 16 MGLYOX d 72 193, 171 0.11, 0.18 4.98 2.55 2.16 5, H m 17 GLYPAN 135 213, 18 9 0.13, 0.04 1.12 0.53 0.33 5, H f 18 GLYOX d 58 196, 182 0.09, 0.10 37.17 23.26 27.3 a: The names of chemicals are from MCM mechanism, only primary products which contribute more than 1% of SOA mass are shown in the table, the detailed composition can b max and f are calculated using NDDO based AM1 semiempirical quantum chemistry method. Up to two major oscillator strengths are selected. Detailed values are shown in supplementary materials. c: F k is the mass perc entage of the k th species, obtained by the mass balance of chemical compounds in toluene SOA. The sum of F k values is 98.6% for T1, 98.4% for T2 and 98.1% for T3. d: MGLYOX and GLYOX are in the form of oligomers. Their spectra were estimated using their dimer structure. The spectra of other oligomers are assumed to be as same as their monomers.
63 Table 3 3 Representative products of pinene SOA and their mass percentages in SOA at different NO x conditions Group (i, j) k Products a MW k max b (nm) f b F k c (%) H NO x L NO x (A1) (A2) 1, H s 1 C811PAN 247 183, 161 0.27, 0.11 4.59 3.98 1, H s 2 PINIC 186 181, 174 0.25, 0.06 0.02 1.98 1, H s 3 C921OOH 204 192, 186 0.09, 0.12 0.09 1.3 1, H s 4 C812OOH 190 187, 186 0.14, 0.10 0.04 0.92 1, H s 5 HOPINONIC 200 176, 171 0.25, 0.22 0.04 1.19 1, H m 6 C920PAN 261 196, 171 0.07, 0.06 8.4 3.91 1, H m 7 C98OOH 204 201, 167 0.29, 0 .14 2.72 10.43 1, H m 8 C98NO3 233 188, 183 0.18, 0.03 6.46 2.62 1, H m 9 C922OOH 220 204, 177 0.32, 0.17 0.09 1.34 1, H f 10 C7PAN3 233 205, 192 0.08, 0.35 18.43 3.54 2, H s 11 C10PAN2 245 187, 171 0.04, 0.08 16.3 5.97 2, H s 12 C97OOH 188 197, 186 0 .05, 0.05 0.49 6.13 2, H f 13 C717NO3 203 184, 182 0.08, 0.19 5.28 3.12 2, H f 14 C108OOH 216 202 0.26 4.39 14.88 3, PO 15 APINAOOH 186 160, 158 0.16, 0.13 0.07 2.32 3, PO 16 APINANO3 215 174151 0.07, 0.2 0.93 2.42 3, PO 17 APINBNO3 215 202, 153 0.05, 0.22 0.59 1.26 3, H s 18 PINONIC 184 173, 169 0.31, 0.15 0.09 0.7 3, H m 19 C89PAN 231 179, 169 0.05, 0.06 3.36 2.14 3, H m 20 C107OH 200 182, 175 0.25, 0.06 0.36 3.53 3, H m 21 C109OH 200 181, 173 0.25, 0.06 0.28 0.82 3, H m 22 C5PAN9 191 199, 184 0 .05, 0.22 2.6 0.6 4, H f 23 CO235C6CHO 156 159 0.01 2.88 2.85 4, H f 24 C109CO 182 200 0.25 0.09 0.42 5, H m 25 PINAL 168 168, 164, 0.03, 0.15 20.12 19.46 a: The names of chemicals are from MCM mechanism, only primary products which contribute more th an 1% of SOA mass are shown in the table, the detailed composition can be found in supplementary materials. max and f are calculated using NDDO based AM1 semiempirical quantum chemistry method. Up to two major oscillator strengths are given, more can be found in supplementary materials. c: F k is the mass percentage of the k th species, obtained by the mass pinene SOA. The sum of F k values is 98.7% for A1 and 97.8% for A2
64 Table 3 4 Comparisions between model predicted abs orbance and literature values. The literature absorb ance data are from NIST webbook Name By NDDO AM 1 (PECI) By literature max max f max max FWHM (nm) (M 1 cm 1 ) (nm) (M 1 cm 1 ) (nm) 9,10 Anthraquinone 303 3575 0.072 325 5048 40 2 Nitro Phenol 305 1158 0.028 346 1752 64 2,4 Hexadienal 275 44059 0.861 260 25527 32 2,3 Butanedione 465 8 0.00 02 435 9 94 3 Methyl 2 Butenoic acid 247 13673 0.33 221 12162 32 Phenol 275 1920 0.023 271 2041 20 1 Octene 202 13552 0.337 177 13335 22 Muconic acid 267 28708 0.698 259 28973 38 1 Methyl 2,4 Dinitrobenzene 211 18778 1.275 240 28642 66
65 Figure 3 1 Schematic diagram for the measurement of UV visible spectra of SOA
66 Figure 3 2 S tructure of the SOA light absorption model.
67 Figure 3 3 Comparisons of the observed and calculated spectra of testing compounds using Gaussian band shape functi on
68 Figure 3 4 FWHM for two systems: (a) toluene and ( b ) pinene
69 Figure 3 5 Comparison of the predicted absorption spectra and the measured absorption spectra of toluene SOA under three different NO x condition s : (a) high NO x (b) middle NO x (c ) and low NO x The spectra of individual products are shown in (d) for middle NO x condition. The uncertainty of the mass normalized absorbance was calculated from the uncertainties of UV visible absorbance and the aeros ol mass through the propagation of uncertainty for divisions. Dash lines with the same color indicate upper and lower range.
70 Figure 3 6 Comparison of the predicted absorption spectra and the measured spectra of pinene SOA under different NO x conditions : (a) high NO x and (b) low NO x using data A1and A2 in Table 2 1. The spectra of individual products are shown in (c) for high NO x condition. The uncertainty of the mass normalized absorbance was calculated fr om the uncertainties of UV visible absorbance and the aerosol mass through the propagation of uncertainty for divisions. Dash lines along the measured spectra indicate upper and lower range.
71 CHAPTER 4 DYNAMIC LIGHT ABSORPTION OF BIOMASS BURNING ORGANIC A EROSOL PHOCHEMICALLY AGED UNDER NATURAL SUNLIGHT Background L ight absorption of organic carbon (OC) aerosol has been poorly understood mainly due to its complex chemical composition and dynamic evolution under atmospheric conditions. OC can be either prim ary (POA) or secondary (SOA) in origin. On a global scale, approximately 69% of POA and 23% of SOA are contributed by biomass burning (Hallquist et al., 2009) Considering the large source of biomass burning O C in the atmosphere, better understanding its light absorption properties is necessary. Recently, there is growing evidence that biomass burning O C may be a nonnegligible contributor to li ght absorption in atmospheric aerosols, especially in the short wavelength visible and ultraviolet spectral regions (Hoffer et al., 2006;Kirchstetter and Thatcher, 2012;Lack et al., 2012) Kirchstetter and Thatcher (2012) analyzed the filter based wood smoke aerosol from night time ambient sampling, and estimated that O C account ed for 49% of t otal aerosol light absorption between 300nm and 400nm. Lack et. a l (2012) measured light absorption of biomass burning aerosol originating from a wildfire event using a p hoto acoustic aerosol absorption spectrometer. Their results suggested that primary OC contributed 27(15) % of the total absorption at 404 nm. Light absorption of OC is mainly determined by its chemical composition. Biomass burning OC has a yellowish to brown color, with absorption mainly at UV range, and decreasing absorption toward visible range (Kirchstetter et al., 2004;Schnaiter et al., 2005) Chen and Bond (2010) measured light absorption of solvent extracted fresh primary OC generated in a nitrogen filled furnace. They proposed that chemicals with
72 large, polar, a nd conjugated aromatic rings would be the light absorbing materials. The study of Del Vecchio and Blough (2004) suggested that hy d roxy aromatic compounds and quinoid might be responsible for OC color. In the atmosphere, the chemical composition of OC aerosol change s with aging through processes such as oxidation of organic gases, heterogeneous oxidation with atmospheric oxidants (e.g., ozone and OH radicals) and aerosol phas e reactions. Recent laboratory experiments indicate that photochemical oxidation results in the dynamic evolution of chemical and physical properties of OC aerosol The aged OC aerosol became less volatile and more oxygenated after a few hours of photoch emical oxidation (Grieshop et al., 2009) Aging processes also change the hygroscopicity The h ygroscopic water content in OC aerosol may increase or decrease depending on wood burning conditions ( Martin et al., 2012) The aging effect s on the light absorbing propert ies of OC ha ve been varied among research reports due to the different aging time s and oxidation conditions. For example, Adler et al. (2011) derive d the effective broadband refractive index of biomass burning aerosol from a wood burning event using a white light optical counter. T hey found that the aged biomass burning aerosols were less absorbing than the freshly emitted aerosols, with the imaginary refractive index decreasing from 0.04 to 0.02. However, Saleh et al. (2013) recently reported that aged OC is more absorbing than fresh one due to SOA formation. In their study, t he wood smoke aerosol was aged for only one hour in an indoor chamber. These conflicting results show that the influence of aging on light absorption of wood smoke OA needs further investigation
73 T he objectives of this study are : 1 ) to investigate the i nfluence of photochemical oxidation on light absorption of wood smoke O C ; 2 ) to explore the effects of relative humidity (RH) and NO x on light absorption of wood smoke OC over the course of the photochemical aging process and 3 ) to characteriz e the chemic al evolution of OC aerosol In order to focus on wood burning OC with minimum influence by BC wood smoke was produced from smoldering phase burning Smoldering combustion has frequently been found in biomass burning situations including wildfires presc ribed burns and agri culture burning (Alves et al., 2011;Hille and Stephens, 2005;Hays et al., 2005) It has been reported to consume over 50% of biomass in temperate and boreal fires (Bertschi et al., 2003) In this study, the resulting wood burning OC was p hotochemically oxidized using out door dual chamber s that expos ed to natural sunlight. The large outdoor chamber allowed atmospheric aging conditions to be closely mimicked and aerosol samples to be collected for longer time periods. Experimental Section Outdoor Chamber E xperimental Setu p The photooxidation of wood smoke was performed using the University of Florida Atmospheric Photochemical Outdoor Reactor (UF APHOR) dual chamber s (52 m 3 per chamber) which are located on the roof of Black Hall UF The UF APHOR dual chambers with half c ylinder design are made of FEP Teflon film attached to metal frames. A detailed description of the UF APHOR chamber s can be found elsewhere (Im et al., 2013) Prior to each experiment, the chambers were continuously flu shed with clean ambi ent air for more than 12 hours and then purged overnight with clean air from a n air cleaner ( GC Series, IQAir Inc. ) Commercia l hickory hard wood l og s w ere chopped into
74 pieces with approximate size of 3 cm 4 cm 33 cm. The wood pieces w ere burned under smoldering condition s using a wood grill stove. Before sunrise, smoke was introduced in to the chambers through a 4 inch alumina tube ( 0.5 meter in length) connected to the stove Immediately after injecting wood smoke, t he chamber air wa s mixed for five minutes using mixing fans The initial conditions for wood smoke particle concentration measured by an organic carbon (OC) analyzer are summarized in Table 1. For high NO x experiments, NO gas from the tank was introduced into the chamber after the wood smoke injection P hotochemical reactions started at sunrise The sunlight intensity, chamber temperature and relative humidity (RH) dynamically chang e d over the course of chamber experiments A typical time profile of these parameters is shown in Fig. S1. To study of the effect of humidity on photochemical aging of wood smoke, the dry condition (RH<30%) humidity was controlled one day prior to the experiment using a dehumidifier T o a chieve the wet condition (RH>60%), the chamber air wa s humidified using a steam evaporator during the experiment Sun light intensity and temperature humidity were measured continuously with an ultraviolet radiometer (model TUVR, The Eppley Labor atory) and a Temp RH sensor (CS2, Campbellscientific), respecti vely. Wood S moke C haracterization Gas and particle s were s ampl ed through manifolds that run directly to a laboratory just below the chambers. The wood smoke particles were collected on a 13mm diameter filter (Borosilicate microfibers reinforced with woven glass cloth and bonded with Teflon, Gelman Science Pallflex, Type: TX40H120 WW) at a flow rate of 17 L/min. The UV visible light absorption spectrum of aerosols collected on the filter
75 was directly monitored using a Perkin Elmer L ambda 35 UV visible spec trophotometer equipped with a Labsphere RSA PE 20 diffuse reflectance accessory. The detai led measurement procedure for light absorption spectr a of aerosols can be found in our previous study (Zhong and Jang, 2011) OC and e lemental carbon (EC) of the aerosol filter samples were analyzed with a semi continuous OC/EC analyzer (Model 4 Sunset Laboratory Inc. ). L evoglucosan obtained from the filter/acetonitrile extract was first derivatized with N,O bis(trimethylsilyl) trifluo roacetamide (BSTFA) and then measured using a gas chromatograph/ion trap mass spectromet er (GC ITMS) (CP 3800 GC, Saturn 2200 MS, Varian Inc.). The fluorescence spectr a of extrac t of the filter sample with dichloromethane were measured by a f luorescence s p ectrophotometer (F 2500 Hitachi Ltd. ). The chemical functional group of particles impacted on a silicon disk was characterized using Fourier transform infrared s pectroscop y (FTIR) (Nicolet Mag n a 560, SpectraLab Scientific Inc ). To determine the water c ontent of aerosols, the sampled silicon disk was installed in a miniature flow chamber which was located in the FTIR optical beam path. The relative humidity inside the flow chamber was controlled by passing wet air through the chamber. FTIR spectra of a erosols on the disk were obtained under varying relative humidity from 10% to 85%. Detailed description s of the method s can be found in the previous study by Jang et al. (2010) Particle concentrations and particle size distribution s w ere monitored continuously using a scanning mobility particle sizer ( Model 3080, TSI Inc.) together with a condensation nuclei counter (Model 3025A, TSI Inc.). Ozone was monitored using a photometric ozone analyzer (model 400E, Teledyne Instruments) and NO x concentrations were
76 measured using a chemiluminescence NO/ NO x an alyzer (model 200E, Teledyne Instruments). Light A bsorption of A mbient Organic Carbon In April 2012, the County Line Fire located in Pinhook Swamp Florida, 90 miles north of UF, burned more than ten thousand a cres. B iomass burning aerosol was collected at the UF sampling site on April 9 10 2012 During the sampling day, the OC concentration in the morning was as high as 92 ug m 3 (measured at the sampling site) compared to a normal day in which the OC value is 2 to 4 ug m 3 Thus wildfire emission is the main source of the ambient aerosols in UF area during this episode Fine particulate matter (PM 2.5 ) passing through a cyclone w ere collected on a 13mm diameter filter at a flow rate of 1 6.7 L/min L ight absorption spectr a of the filter sample wer e directly measured using a UV visible spectrometer equipped with an integrating sphere Results and D iscussion Light A bsorption of O C A wood burning particle co ntains a mixture of BC and OC BC and elemental carbon (EC) are treated as equivalent Due to its strong light absorption capacity, a s mall amount of BC may contribute significant spectral absorption depending upon wavelength Hence, burning condition s w ere controlled as smoldering to generate less BC The percentage of element carbon in smolder ing smoke particles was near or less than 5% as shown in Table 1. A method introduced by Kirchstetter and Thatcher (2012) was used to separate light absorption by OC from the absorption of the smoke aerosol mix ture (O C and B C) Light absor ption of O C Abs O C is obtained by subtracting the absorption of BC, Abs BC from the mixture s absorption Abs :
77 (4 1) Abs is directly measured from the filter sample. Abs BC is estimated as: (4 2) where A AE is the A bsorption Angstrom E xponent and defined as : (4 3) For Abs BC estimation, the measured AAE value was 1.03 for wavelength s between 360 and 900nm. Figure 4 1a shows absorption spectra of the fresh smoke aerosol, the decoup led OC and the decoupled BC. Over the wavelength range of 280nm to 600nm, OC contribute s 60~98% of the total sample absorption, while BC is dominate for longer wavelengths (>700nm). Light absorption of OC exhibits no distinct peaks, exponentially decreas ing with increasing wavelength. In the visible range ( 400 700nm), the average AAE of OC is 4.74 This value accords with the AAE of ambient biomass burning OC (an average value of 5) reported by Kirchstetter and Thatcher (2012) All samp les were analyzed in a similar manner as described above. T he main data reported here are mass absorption cross section of OC ( MAC OC m 2 g 1 ) which is wavelength ( ) dependent and determined by normalizing the absorption coefficient of OC ( b ab s(OC) m 1 ) with thermal OC mass concentration M OC (g m 3 ) (4 4)
78 In equation 4 4, b abs (OC) can be calculated from the measured absorbance of filter sample(abs): (4 5) where C is the correction factor with a value of 1.4845 obtained from our previous study (Zhong and Jang, 2011) A is the filter surface area sampled (7.85 10 5 m 2 in this study), and V is the volume (m 3 ) of air passing through the filter during a given sampling time Abs O C is obtained from equation 4 1 Effect o f Photochemical Aging o n Light Absorption o f OC Our chamber experimental data showed that light absorption of wood burning OC was significantly modified due to photochemical aging. Figure 4 1b shows the typical MAC OC of wood smoke exposed to sunlight using the outdoor chamber a t different daytime hours Ov erall, MAC OC increased with aging time in the morning. The total MAC OC which is estimated as the area under the MA C OC spectrum curve between 280nm and 600nm, increased by 11% ~54% (26% on average) except data from the high R H condition ( W est chamber on Oct. 24 2012) Saleh et al., (2013) have explained that the increment in aged OA absorption is caused by the formation of SOA, which absorbs more than fresh POA in the short wavel ength visible and near UV regions. Schauer et al (2001) reported that about 34% of identified organic compounds in the gas phase and 41% of organic compounds identified in POA were phenols, syringols and guaiacols. These phenols and methoxylated phenols in the gas phase or in the particle can rapidly react with atmospheric oxidants to form light brown substances (Gelencser et al., 2003;Chang and Thompson, 2010;Ofner et al., 2011)
79 Thus the enhancement in light abs orption would be caused by either oxidized POA or the SOA produced from oxidation of primary phenolic compounds. However the light absorption of O C in both the UV and visible ranges began to decrease with aging time. The total MA C OC decreased by 19% ~ 68% (41% on average) compared to the fresh POA. The possible explanation for the decreases in MA C OC is bleaching of colorant in POA and SOA by sunlight. T he previous study by Zhong and Jang (2012) reported the bleaching effect of light on light absorption of SOA. In wood smoke POA and SOA, there are abundant of chromophores, such as conj ugated aromatic rings and phenols, as well as nitro and hydroxyl groups. High energy photo n s in sunlight can excite electron s in colored molecules through or n transition s and disrupt the conjugated structure of chromophores resulting in the gradual fading of wood smoke color. Effect of RH on L ight A bsorption of O C In order to investigate the effect of humidity on MAC OC two sets of wet dry dual c hamber experiments were conducted: high RH ( 80 87% ) vs low RH ( 11 27% ) on Oct.24, 2012, and middle RH ( 56 75% ) vs low RH (12 24%) on Nov. 01 2012 Figure 4 2 illustrates the time profile of the relative MAC OC and MAC OC at 550 nm. The relative MAC OC was o btained by normalizing the total MAC OC at a certain time by the initial total MAC OC As photooxidation progressed in the afternoon, t he absorption decay at wet conditions became more rapid than the decay at dry conditions. For example, at the end of the experiment, t he relative MAC OC and MAC OC at 550nm at the low RH were higher by 43% and 45% compared to those at the high RH Unlike MAC OC values in the middle or the low RHs, no increase appeared for MAC OC values for the experiments at high RH in the morn ing. E levated values of relative humidity can increase the water
80 content of wood smoke particles. Anastasio et al. (1996) reported that illumination of aqueous phase non phenolic aromatic carbonyls in the presence of phenols can destroy phenols In addition, this aqueous phase photooxidation produces a significant amount of H 2 O 2 since both the aromatic carbonyls and phenols are major products from the combustion of wood (Rogge et al., 1998) In wet aerosols, H 2 O 2 can photodisssociate to produce OH radical s (Zellner et al., 1990;Faust, 1994) which c ould decompose chromophores and lead to bleaching of wood OA. Effect of NO X on L ight A bsorption of O C C ontrolled dual chamber experiments were also conducted to study the influence of NO x on MAC OC of wood aerosols, high NO x ( 108 ppb) vs low NO x ( 16 ppb) on Oct.30, 2012 and middle NO x (43 ppb) vs low NO x (16 ppb) on Oct. 11, 2012 Figure 4 3 shows the time profiles of the relative MAC OC values and MAC OC at 550nm under different NO x concentrations. In the presence of high NO x the MAC OC value s are slightly higher than those at low NO x concentration NO x is able to modify the reaction pathway for organic compounds in the gas phase. N itro phenols (e.g., nitrocatechols ) and nitro compounds can be produc e d through the photooxidation of phenol ic organic compounds and contribute SOA. Nitrophenols have been suggested to be strong candidates to represent light absorbing toluene SOA (Nakayama et al., 2010b;Zhong et al., 2012;Zhang et al., 2013) A higher NO x level, the concentration of these chromophoric nitro phenols can be greater, decelerating the decay of MAC OC of wood smoke OA. Chemical E volution of Organic Carbon Aerosol To investigate how the chem ical composit ion of wood smoke aerosol in fluences light absorption, the aerosol introduced into the chamber was characterized as photooxidation progressed. In this study, l evoglucosane, a major constituent in wood
81 smoke POA was monitored by GC ITMS. T he intensity of fluorescence, which is mainly due to PAHs, was measured for solvent extracted aerosol samples. To study the alter ation of the hygroscopic propert ies of wood smoke aerosol due to photooxidation, the aerosol water content was also measured by FTIR. Levog lucosan d ecay L evoglucosan is abundant in wood smoke POA contribut ing 3 ~4 9 % of total wood smoke aerosol carbon (Mazzoleni et al., 2007) Figure 4 4 a illustrates the rapid degradation of both pure levo glucosan and levoglucosan in wood smoke after wall loss correction of aerosols The experiment method for the pure levoglucosan experiment is provided in the supplementary materials ( Document C 1 ) The decay rate of the levoglucosan associated with wood smoke aerosol is similar to that of pure levoglucosan within an error range. The degradation of levoglucosan would be caused by wall loss of levoglucosan vapor and photochemical oxidation by OH radicals in the aerosol. According to Booth et al. (2011) the sub cooled liquid vapor of levoglucosan is 1.9 10 4 pa at 298K the wall loss of gas phase levoglucosan is likely. In order to identify the products produced from the photooxidation of levoglucosan in the presence of HONO aerosol filter samples were analyzed using GC ITMS after silylation of the products. The mass fragments of the tentatively identified products are shown in Figure C 2 and t he reaction pathways for the formation of the proposed products are illustra ted in Figure C 3. However the low amount of O 3 formation suggests that oxidation of l evoglucosan occurred mainly via aerosol phase reaction.
82 PAHs d ecay Although t he mass percentage of PAHs in wood smoke is less than 1% (Schauer et al., 2001) PAHs can c ontribute to the color of wood smoke OA. T he photooxidation of PAHs can therefore be examined by measuring the fluorescence as a function of time The fluorescence spectra of OA (Figure 4 4b ) show a rapid decrease in fluorescence emission when excited at 280 nm Similar to levoglucosan, such decline can be caused by both evaporation combined with increased temperature after sunrise (Figure C 1) and photochemical oxidation of PAHs PAHs can be transform ed to oxy PAHs via aerosol phase oxidation (e.g., re action with singlet oxygen produced by a photosensitizing process or reaction with OH radicals and free radical reactions) (Jang and McDow, 1995) These oxy PAHs generally have much lower fluorescence quantum yield compared to un oxidized PAHs. FTIR spectra of w ood b urning a erosol Figure 4 5 a compares the characteristic FTIR spectra of fresh and aged wood burning particles at RH of 45%. For the fresh particles, the O H stretching of alcohols is seen at 3500 3200cm 1 The C=O stretching of carbonyls (e.g., aldehydes, ketones, and carboxylic acids ) occurs at 1800 1680cm 1 and t he aromatic C=C str etching bands appear at 1610, 1517 and 1458cm 1 After photochemical aging of wood smoke OA for 5 hours the absorbance for both the O H stretch of alcohols and the aromatic C=C stretch significantly decreased The FTIR spectrum of aged particles also sh ows the formation of carboxylic acids confirmed by the broad O H stretch at 3300 2500cm 1 and the increased absor bance at the C=O stretching frequency, suggesting oxidation of wood constituents
83 H ygroscopic p ropert ies of wood burning aerosol To investigat e the hygroscopicity of wood smoke particles, the water content of wood burning OA was monitored using the FTIR equipped with a miniature flow chamber under varying RHs (Jang et al., 2010) In brief, the absor bance at 1650cm 1 which is the characteristic frequency for O H bending of water was used to esti mate the water content in the aerosol. The peak at 1650cm 1 was calibrated with the NaCl particle which has a known water content obtained by an inorganic thermodyn amic model at a given RH (e.g., AIM III model, http://www.aim.env.uea.ac.uk/aim/aim.php ) (Clegg et al., 1998) As shown in Figure 4 5 b, the fresh particle s were much more hygroscopic than the aged particle s T he water content in fresh particles exponentially increased as a function of RH. No water wa s detected when the RH was lower than 2 5%. In the morning, fresh wood smoke particles contain a high amount of the sugar compounds such as levog l ucosan which i s very hydrophilic and easily absorb s water. As described in section 3.4.1 the amount of primary sugars rapidly decay ed due to photooxidation. A lthough the carboxylic acid content increased with aging, within aged wood aerosol (FTIR data in Figure 4 5a ) the aerosol bec ame less hygroscopic Conclusion and A tmospheric I mplication The dynamic changes in light absorption of biomass burning organic aerosol were investigated using the outdoor smog chamber under natural sunlight. The diurnal change in absorpt ion of wood smoke aerosol is governed by two mechanisms: chromophore formation and sunlight bleaching. The c olored products originating from photooxidation of phenolic SVOCs can increase wood smoke OA absorption in the morning while sunlight fades the col or of both POA and SOA in the afternoon. H igher concentrations of NO x help prolong the wood smoke aerosol color (Fi g ure 4 3 ) but t he
84 h igh RH a ccelerates the degradation of aerosol color (Fig ure 4 2) The decay of both PAHs and levoglucosan indicated dyn amic changes in chemical composition of primary OC due to aging. The aged wood smoke OC became more oxidized but less h ygroscopic as shown by FITR spectra measurement The ambient data obtained during the County Line Wildfire event on April 09 and 10, 2012, were analyzed using equation ( 4 5) with a measured AAE of 0.74 for ambient BC. Figure 6 shows the light absorption of ambient biomass burning OC after subtracting background OC. Compared to the MAC OC sampled at 8:30am, the MAC OC sampled at 9:36am o n April 09 increased by 18%, but the MAC OC decreased by 28% at 11:30am. The diurnal pattern in light absorption of the ambient biomass burning OA is consistent with the result obtained in the outdoor chamber. There is increased emphasis on research pert aining to the climate forcing of OA including both experimental studies and model simulations. Considering aerosol of lifetime (about six days), the result s of both outdoor chamber studies (Fig ure 4 1 ) and ambient field data ( Fig ure 4 6 ) suggest that bio mass burning OA will absorb less light as photochemical reactions progress
85 Table 4 1 Summary of experimental conditions of fresh wood smoke for photochemical oxidation No Date Chamber RH Temp. NO x Initial OC EC/TC Comments % K ppb ug/m 3 % 1 Oct. 11, 2012 E 39 95 291 312 43 56 4.76 middle NO x 2 Oct.11, 2012 W 51 95 291 311 16 40 5.56 low NO x 3 Oct.30, 2012 E 21 87 275 300 108 58 4.17 high NO x 4 Oct.30, 2012 W 26 88 275 297 16 68 3.33 low NO x 5 Oct.24, 2012 E 10 26 291 311 5 58 3.23 low R H 6 Oct.24, 2012 W 80 87 291 310 6 58 3.57 high RH 7 Nov.01, 2012 E 12 23 281 306 5 111 4.35 low RH 8 Nov.01, 2012 W 56 75 283 304 6 77 2.78 middle RH 9 Nov. 21, 2012 W 67 87 285 305 10 144 2.50 chemical analysis
86 Figure 4 1. UV visible light absorption spectra and MAC OC of wood smoke OC. A ) The absor ption spectrum of wood smoke OC obtained by subtracting the BC absorbance from the total absorbance of wood burning particles. B ) The diurnal pattern of MAC OC of wood smoke OC photochemically age d under natural sunlig ht (Oct.11, 2012 middle NO x )
87 Figure 4 2. C omparison of light absorption of wood OA photochemically oxidized at different humidity conditions: A ) low RH vs middle RH and B ) low RH vs high RH. The open symbols are for MAC OC at 550nm. The filled symbols are for the relative MAC OC which is expressed as the total MAC OC divided by the initial total MAC OC The error associated w ith MAC OC was estimated based on the instrumental errors from OC/EC analyzer and UV visible spectromete r as well as the uncertainty from the correction factor, C (see equation 5).
88 Figure 4 3. C omparison of light absorption of wood OA photochemically oxidized at different NO x conditions : a) low NO x vs middle NO x and b) low NO x vs high NO x The open sy mbols are for MAC OC at 550nm. The filled symbols are for the relative MAC OC which is expressed as the total MAC OC divided by initial total MAC OC The error associated w ith MAC OC was estimated based on the instrumental errors from OC/EC analyzer and UV v isible spectrometer as well as the uncertainty from the correction factor, C (see equation 5).
89 Figure 4 4 Decay of levoglucosan and PAHs. A ) Time profile for the decay of pure levoglucosan ( March 27, 2013) and levoglucosan in wood smoke OA ( Nov. 21, 20 12 ). The concentrations of levoglucosane (C t ) were corrected for wall loss of particles and then normalized by the initial leveglucosan concentration (C 0 ) The associated error with the measured concentration of levoglucosan by GC MS is 20%. B ) F luoresc ence emission spectra of wood smoke particles collected at different times on Nov. 21, 2012. The excitation wavelength was 280nm. The error associated with the concentration of levoglucosan was estimated based on GC MS.
90 Figure 4 5 FTIR spectra and hygroscopic growth profile of fresh and aged wood burning particles. A ) FTIR spectra of fresh and aged wood burning particles sampled on Nov.21, 2012 The spectra were recorded at RH of 45%. B ) The w ater content of fresh particles and aged particles as a function of RH The water content was measured with decreasing RH. The error associated w ith the water fraction in aerosol was estimated based on aerosol mass and FTIR absorbance at 1650cm 1
91 Figure 4 6 Light absorption of ambient biomass burn ing OA sampled during the country line wildfire event at different date : A) April 09, 2012 and B) April 09, 2012
92 CHAPTER 5 RADIATIVE IMPACT OF ORGANIC CARBON AEROSOL Background There are few studies available for the climate forcing of org anic aerosols. Hoyle et al. (2009) calculated the climate forcing of SOA using the off line aerosol chemistry transport model Oslo CTM2 and suggested a global average radiative forcing of 0.09~ 0.06 w/m 2 This value was similar to that of POA. This study provides the first information about the anthropogenic influence of SOA climate impact. However, in their study, optical properties of SOA were taken to be similar to POA (Myh re et al., 2009) And the optical properties of POA were treated as same as those of sulfates. What is the uncertainty caused by using sulfates as the surrogate of SOA? Is POA a good surrogate for SOA? The recent studie s show that alpha pinene SOA slig htly absorb s UV light (Zhong and Jang, 2011) while POA from wood burning has much stronger absorption capability (Chen and Bond, 2010) The objective of this study is to calculate the radiative forcing of SOA and POA using the latest available measured optical data for the purpose of reducing uncertainties related to organic aerosol in the current climate model. To achieve the objective, the following steps will be followed: 1. calculate the single particle optical parameters of POA and SOA using Mie scattering model; 2. c ompute and compare radiative forcing of POA and SOA using a radiative transfer model Me thod Optic al parameters such as extinction cross ext .), single scattering albedo (SSA, were calculated usi ng Mie scattering model with input of refractive index and particle size at certain RH.
93 A simple equation was then used to estimate the radiative forcing of organic carbon aerosol The following section will describe the detailed procedure for each step. Mie Scattering Model The Lorenz electromagnetic radiation by a sphere. The Lorenz research website ( http://www.hiwater.or g/ ) calculates absorption, scattering and backscattering for spherical, coated or uncoated particles of lognormal or measured size distributions. The only required input parameters are particle size and complex refractive index (RI, n + k i). In this st udy, the dry particle size is assumed to be lognormal distributed with count median diameter of 138 nm, geometric standard deviation of 2. The size parameter is taken from a study for ambient organic aerosol (Kaul et al., 2012) The measured real part of RI (Kim and Paulson, 2013) is used for both SOA and POA. It is known that real RI is a function of wavelength. Since the measurement is only available at 550 nm, the fixed value (1.44) is used for the UV and visible wavelength range. The imaginary RI data of SOA and POA are from laboratory measurement in this work Here SOA is represented by alpha pinene SOA and POA by wood burning organic aerosol. In ad dition, sulfate droplet was simulated as reference for comparison purpose. The refractive index of sulfate droplet is taken from OPAC data base ( http://ether.ipsl.jussieu.fr/etherTypo/?id=1058 ) Using the above input parameters, the Mie calculation was performed every 10 nm between 280 and 800 nm to provide scattering, absorption and extinction cross section area. Table D 1 to D 5 summ arizes the model experiments which were carried out.
94 Eff ect o f RH o n Particle Size SOA can absorb water vapor at certain relative humidity. Consequently, the diameter of particle will depend on the relative humidity. The hygroscopic growth has i mmediate influence on the scattering of solar radiation. An empi rical model (Birmili et al., 2009) was adopted to calculate the particle size change due to RH. The model is derived from ambient a erosols from a Finnish forest with OC fraction of 30% to 50%. It is used here for 100% organic aerosol since no better model is available. Growth factor (GF) is defined as follows: (5 1) where D p is particle diameter. GF can b e estimate using an empirical equation: (5 2) where ( D p ) is a function of dry particle diameter and estimated by: (5 3) According to the model developer, this empirical model can safely be applied for all particle size although the fit is only supported by data points between 60 nm and 350 nm. Simple Radiative Efficiency Estimation A simple forcing efficiency (SFE, W/cm 3 aerosol) was used here to provide a first estimate of climate impact of organic aerosols. The calculation was originated from Chylek and Wong (1995) equation that normalizes impact by particle volume. The wav elength dependent SFE is calculated as:
95 (5 4) w here S( ) is the solar irradiance, atm is the atmospheric transmission (0.79), F C is the cloud fraction(0.6), S is the surface albedo(0 .19), is the backscatter fraction scat and abs are the scattering and absorption cross sections per cm 3 ,respectively. The matlab code for this calculation is available on the website ( http://www.hiwater.org/ ) Results and Discussion Optical Paramete rs Figure 5 1 compares thr ee optical parameters of SOA POA and sulfates extinction cross section area, aerosol asymmetry factor and single scattering albedo. The extinction cross section area is a measure of how strongly the aerosol absorbs and scatters light. The simulation result show s that extinction cross section area of the three types of aerosol is similar, decreasing with wavelength (Figure 5 1 a) Aerosol asymmetry factor is a measure of the preferred scattering direction, forward or backward. In Figure 5 1 b, the aerosol asymmetry factor is positive, indicating scattering is mainly in the forward direction. POA asymmetry factor is a little bit larger than SOA and sulfate in the UV range, but all are almost same in the visible range. Single sca ttering albedo (SSA) is the ratio of scattering to the extinction, ranging from 0 to 1. Values of SSA below 0.8 indicate that the aerosol could have a warming effect. Figure 5 1 c shows SSA of SOA is near to 1 in the wavelength range between 280 nm and 80 0nm indicating SOA is mainly scatt er ing, not active in absorbing light. SSA of POA gradually increases from 0.7 to 1, indicating the absorption of POA decrease with wavelength. This trend matches well with the trend of imaginary refractive index of POA.
96 Effect of RH on Optical Parameters Relative humidity affects the optical properties of aerosol by changing particle size and adding water into particle. Figure 5 2 compares the dry SOA and wet SOA at RH of 50%. For Mie calculation, the water can be assum ed to externally coat on the surface of aerosol. The simulation results show that higher humidity increases both extinction cross section area and aerosol asymmetry factor, but has little influence on aerosol single scattering albedo. Similar tread is al so observed for POA, but not shown here. Radiative Impact Figure 5 3 shows fo r cing of SOA, POA and sulfate as a function of wavelength. The forcing of SOA is below zero in the studied wavelength, while that of POA is positive in the short UV range and neg ative in the visible range. Due to the significant difference in forcing of SOA and POA it should be very cautious to replace POA and SOA with each other. The forcing of SOA is very similar to that of sulfate. Thus sulfate is a reasonable surrogate for SOA in terms of optical parameters. High RH increases the negative forcing for SOA and POA by adding non absorbing water. A full radiative transfer model would be necessary to determine the actual forcing of SOA. Conclusion The optical parameters of POA and SOA were estimated using measured refractive index by running the Mie scattering model. RH could change the particle size and thus the optical properties of organic aerosols. A simple estimation of radiative efficiency shows that POA has warming eff ect in the UV range, while SOA is a cooling aerosol, with similar radiative forcing to sulfate.
97 The estim ation of the actual forcing of O C requires a full radiative transfer model with O C burden and meteo rological data. O C burden can be estimated using a global model such as GEOS Chem. A radiative transfer model such as Fu Liou Gu model (Gu et al., 2006) can be tested to produce the radiat ive forcing of O C
98 Figure 5 1 Optical parameters of SOA POA and sulfate estimated using Mie code : ( a ) extinction cross section area, ( b) aerosol asymmetry factor, and ( c) single scattering albedo
99 Figure 5 2 Effec t of RH on optical parameters : ( a) extinction cross section area, ( b) aerosol asymmetry factor, and ( c) single scattering albedo
100 Figure 5 3 Comparison of r adiative efficiency of SOA, POA and sulfate as a function of wavelength
101 C HAPTER 6 CONCLUSIONS A method for measuring the mass absorption cross section ( MAC ) of organic aerosol has been developed using a conventional UV visible spectrometer equipped with an integrating sphere covering a wide range of wavelengths (280 8 00nm). Th e feasibility of the proposed method was evaluated using reference aerosol s with known absorption cross section. This method directly measures MAC of organic aerosol on a conventional filter without solvent extraction. The resulting method was applied to measure MAC of secondary organic aerosol (SOA) which was produced through photooxidation of different precursor hydrocarbons such as toluene d limonene and pinene using a 2 m 3 i ndoor Teflon film chamber. MAC value of toluene SOA (0.574 m 2 g 1 at 350 nm) was the highest compared with MAC values for pinene SOA (0.029 m 2 g 1 ) and d limonene SOA (0.038 m 2 g 1 ). When d limonene SOA or toluene SOA was internally mixed with neutral [(NH 4 ) 2 SO 4 ] or acidic inorganic seed (NH 4 HSO 4 :H 2 SO 4 = 1:1 by mole), the SOA show ed 2~3 times greater MAC values at 350 nm than the SOA with no seed Aerosol aging with a light source for this study reduced MAC values of SOA (e.g., on a verage 10% for toluene SOA and 30% for d limonene SOA within 4 hours). The new model for UV visible absorption spectrum of SOA in this work predicted that the light absorption of toluene SOA would increase with higher NO x concentration and that of alpha pinene SOA is not affected by NO x The model results w ere in reasonably good agreement with the measurement s The model also predicted that the main light absorbing material s in toluene SOA w ere nitro phenols and those in alph pinene SOA w ere pinonaldeh yde s
102 The method was also used to quantify M AC of primary organic aerosols ( P OA). POA produced under smoldering conditions w as photochemically aged under different relative humidity and NO x conditions using an outdoor chamber under natural sunlight. The measurement results suggest that MAC of POA changed under atmospheric conditions increasing in the morning and decreasing in the afternoon, due to the competition between chromophore formation and sunlight bleaching A similar trend in light absorption changes was observed in ambient smoke aerosol originating from the 2012 County Line Wildfire in Florida. We con clude that the biomass burning POA becomes less light absorbing after 8~9 hours sunlight exposure. The simple estimation of radiative efficien cy suggests that SOA is a cooling aerosol, having a similar value of radiative forcing to sulfate. Fresh POA ha s positive radiative forcing in the UV wavelength rang and negative forcing in the visible range. The net forcing is positive, thus fresh POA i s warming aerosol.
103 CHAPTER 7 FUTURE STUDIES The current study provides the experimental measured optical properties of organic carbon aerosols, contributing to reduce the uncertainties organic carbon aerosols in climate change. There are several directi on s for further study of the climate impact of aerosols: 1. E ffect of high humidity on light absorption of SOA and POA Th is study ha s shown that RH at 85% will accelerate the decay of MAC In the upper troposphere, RH usually reaches up to 100%. The st udy of optical properties of O C at high RH will benefit the model ing of O C which travel to the upper level of troposphere. 2. Light absorption of mixed aerosols. In atmosphere, organic aerosols usually mixed with other types of aerosols. For example, PO A are always emitted together with black carbon SOA can mix with dust or black carbon. Anthropogenic SOA may mix with biogenic SOA. U nderstand ing the mixing rule in light absorption of different aerosols will make it possible to predict light absorptio n of aerosol mixture in atmosphere. 3. Identification of light absorbing materials This work suggests possible light absorbing materials in SOA and POA, such as nitrophenols. However, chemical analysis for these compounds has not been conducted. GC MS o r LCMS analysis is necessary in the future to identify these light absorbing chemicals in laboratory generated aerosols and field sampled ambient aerosols. 4. H ygroscopic properties of SOA and POA Hygroscopic growth of aerosol is directly related to clou d condensation nuclei. The indirect climate effect of organic aerosols associated with cloud formation is among the least under st ood factors in
104 climate system. The study of hygroscopic properties of fresh and aged organic aerosols is necessary to provide fundamental parameters for cloud formation. 5. Evaluate the radiative forcing of O C using radiative transfer model The current study uses a simple equation to estimate the radiative efficiency of O C without considering the mass budget of O C As a furthe r study of this research, the calculation of radiative forcing of O C using radiative transfer model and global chemical transport model needs to be achieved to have better evaluation of the role of O C in climate system.
105 APPENDIX A SUPPLEMENTARY METERIALS FOR CHAPTER 2 Table A 1. Indoor Teflon film chamber experiments a involving the aerosol of known composition No. Aerosol f b Sample volume (m 3 ) L c (m) Aerosol v ol. conc. (n L /m 3 ) Density d Mass e MY n f (#/m 3 ) ln(1/ T ) g Predicted ln(1/ T ) h Measured M1 MY 1.0 0.002 19.10 67 1.47 0.15 1.59E+17 0.034 0.036 M2 MY 1.0 0.003 38.20 69 1.47 0.30 1.63E+17 0.069 0.077 M3 MY 1.0 0.005 66.85 50 1.47 0.39 1.19E+17 0.088 0.173 M4 MY 1.0 0.010 127.32 49 1.47 0.71 1.14E+17 0.162 0.217 M5 MY 1.0 0 .015 188.21 44 1.47 0.96 1.04E+17 0.217 0.363 M6 MY 1.0 0.015 190.99 70 1.47 1.54 1.65E+17 0.350 0.607 M7 MY NaCl 0.0533 0.014 178.25 238 2.11 0.38 4.30E+16 0.085 0.158 M8 MY NaCl 0.0533 0.028 358.29 244 2.11 0.77 4.41E+16 0.175 0.291 M9 MY NaCl 0.0533 0.058 742.17 237 2.11 1.56 4.28E+16 0.352 0.518 M10 MY NaCl 0.0533 0.083 1058.83 238 2.11 2.23 4.30E+16 0.506 0.665 M11 MY NaCl 0.0533 0.112 1421.78 240 2.11 3.01 4.33E+16 0.682 0.783 M12 MY NaCl 0.0533 0.160 2041.26 233 2.11 4.20 4.21E+16 0.952 0.869 a:Temperature and humidity are in the ranges of 294 298 K and 30.5 36.8%, respectively b:F is the mass fraction of MY aerosol to total sampled aerosol c:Length is obtained by sample air volume divided by the sample area of filter. d:Density of MY aeroso l is calculated by using molar weight divided by molar volume. e:The estimated MY aerosol mass for each aerosol filter sample is obtained by multiplying sample volume to vol. conc., density and f f: n is estimated by equation (7). g: Predicted ln(1/ T ) b ased on Equation ( 2 nm is 1.1210 20 m 2 h: Measured ln(1/ T nm
106 Figure A 1 M easured molar absorptivity of m etanil y ellow as a function of wavelength (280 6 80 nm).
107 Figure A 2 UV visible absorption spectra of d limonene SOA collected on the filter at different exposure time in air The SOA samples were stored in open air. SOA color turned to orange yellow in three hours corresponding to the growth of two absorption peaks at 428 nm and 505 nm.
108 APPENDIX B SUPPL EMENTARY METERIALS FOR CHAPTER 3 Table B 1. Chemical structure of toluene SOA products MCM name IUPAC name Structure MNCATECH 3 methyl 6 nitrobenzene 1,2 diol MNNCATCOOH (2R) 2 hydrope roxy 2,3 dihydroxy 1 methyl 4 nitro 6,7 dioxabicyclo[3.2.1]oct 3 en 8 yl nitrate DNCRES 2 methyl 4,6 dinitro phenol TLEMUCOOH 3 (2 hydroperoxy 1 hydroxy 3 oxo butyl) oxirane 2 carbaldehyde TLEMUCNO3 1 (3 formyloxiran 2 yl) 1 hydroxy 3 oxobutan 2 yl nitrate TLBIPEROOH (1S,4S,5S) 4 hydroperoxy 1 methyl 6,7 dioxabicyclo[3.2.1]oct 2 en 8 ol
109 Table B 1 Continued MCM name IUPAC name Structure TLBIPERNO3 (1S,2S,5S) 8 hydroxy 5 methyl 6,7 dioxabicyclo[3.2.1]oct 3 en 2 yl nitrate NC4MDCO2 H (Z) 2 methyl 3 nitro 4 oxobut 2 enoic acid TLEMUCCO 3 (1 hydroxy 2,3 dioxobutyl)oxirane 2 carbaldehyde TOL1OHNO2 2 methyl 6 nitrophenol ACCOMEPAN 3 acetoxy 3 oxopropanoic nitric peroxyanhydride MALDIALPAN nitric (E) 4 oxobut 2 enoic peroxyanh ydride C5COO2NO2 nitric (E) 4 oxopent 2 enoic peroxyanhydride
110 Table B 1 Continued MCM name IUPAC name Structure MALANHY furan 2,5 dione C5DICARB 4 Oxo pent 2 enal MGLYOX 2 oxopropanal GLYPAN nitric 2 oxoacetic peroxyanhydride GLY OX oxalaldehyde GLYOX oligomer 2 Dihydroxymethyl [1,3]dioxolane 4,5 diol MGLYOX oligomer 2 Dihydroxymethyl 2,4 dimethyl [1,3]dioxolane 4,5 diol
111 Table B 2 Representative products of toluene SOA and their mass percentages at different NO x con ditions Group ( i j ) k Products name a MW k max b (nm) f b F k c (%) H NO x (T1) M NO x (T2) L NO x (T3) 1, PO 1 MNCATECH 169 330, 228 0.05, 0.40 26.27 30.48 26.14 2 MNNCATCOOH 281 270, 199, 189, 188 0.3, 0.15, 0.31, 0.03 2.95 6.36 6.47 3 DNCRES 191 321, 229, 227, 222, 203 0.04, 0. 17, 0.40, 0.82, 0.08 2.00 0.88 0.32 1, H m 4 TLEMUCOOH 190 177, 165, 161 0.04, 0.05, 0.03 1.04 3.04 4.41 5 TLEMUCNO3 190 177, 165, 161 0.04, 0.05, 0.03 2.40 1.57 1.10 2, PO 6 TLBIPEROOH 174 216, 213, 163, 153, 143 0.19, 0.11, 0.13, 0.10, 0.05 2.13 8.68 14.00 7 TLBIPERNO3 174 415, 233 0.06, 0.07 3.85 3.46 2.78 2, H s 8 NC4MDCO2H 159 228, 226, 221, 202, 177, 172, 168 0.15, 0.08, 0.21, 0.26, 0.22, 0.19, 0.05 6.71 3.19 1.38 2, H f 9 TLEMUCCO 156 202, 148 0.35, 0.03 0.53 0.81 1.09 3, PO 10 TOL1OHNO2 153 316, 222 0.04, 0.17 0.77 0.51 0.23 3, H f 11 ACCOMEPAN 207 206, 187, 174 0.08, 0.26, 0.25 2.29 10.34 7.67 4, H m 12 MALDIALPAN 161 208, 190, 176, 158 0.07, 0.27, 0.26, 0.05 0.59 0.48 0.41 13 C5COO2NO2 175 227, 184, 176 0 .59, 0.22, 0.26 0.99 1.03 1.17 5, H m 14 MALANHY 98 230 0.17 2.26 1.00 1.01 15 C5DICARB 98 223, 166, 162, 155 0.65, 0.06, 0.09, 0.42 0.55 0.21 0.14 16 MGLYOX (oligomer) 72 193, 171, 160, 155 0.20, 0.34 4.98 2.55 2.16 17 GLYPAN 135 213, 189, 183, 168 0.13, 0.04, 0.47, 0.53 1.12 0.53 0.33 5, H f 18 GLYOX (oligomer) 58 196, 182, 163, 160 0.14, 0.06, 0.39, 0.11 37.17 23.26 27.30 max and f are calculated using NDDO based AM1 sem iempirical quantum chemistry method; c: F k is the mass percentage of the k th species, obtained by the mass balance of chemical compounds in toluene SOA.
112 Table B 3 Chemical structure of pinene SOA from MCM mechnism MCM name IUPAC name Structure C811P AN 2,2 dimethyl 3 (2 (nitroperoxy) 2 oxoethyl)cyclobutanecarboxylic acid PINIC 3 (carboxymethyl) 2,2 dimethylcyclobutanecarboxylic acid C921OOH 1 (1 hydroperoxy 3 (hydroxymethyl) 2,2 dimethylcyclobutyl) 2 hydroxyethanone C812OOH 1 hydroperoxy 3 (hydroxymethyl) 2,2 dimethylcyclobutanecarboxylic acid HOPINONIC 2 (3 (2 hydroxyacetyl) 2,2 dimethylcyclobutyl) acetic acid C920PAN 2 (3 (2 hydroxyacetyl) 2,2 dimethylcyclobutyl)acetic nitric peroxyanhydride
113 Table B 3. Continued. MCM name I UPAC name Structure C98OOH 6 hydroperoxy 5 (hydroxymethyl) 6 methylheptane 2,3 dione C98NO3 3 (hydroxymethyl) 2 methyl 5,6 dioxoheptan 2 yl nitrate C922OOH 6 hydroperoxy 1 hydroxy 5 (hydroxymethyl) 6 methylheptane 2,3 dione C7PAN3 nitric 3 ,5,6 trioxoheptanoic peroxyanhydride C10PAN2 2 (3 acetyl 2,2 dimethylcyclobutyl)acetic nitric peroxyanhydride C97OOH 1 (1 hydroperoxy 3 (hydroxymethyl) 2,2 dimethylcyclobutyl)ethanone C717NO3 1,5,6 trioxoheptan 3 yl nitrate
114 Table B 3. Con tinued. MCM name IUPAC name Structure C108OOH 3 (2 hydroperoxypropan 2 yl) 5,6 dioxoheptanal APINAOOH 2 hydroperoxy 2,6,6 trimethylbicyclo[3.1.1]heptan 3 ol APINANO3 3 hydroxy 2,6,6 trimethylbicyclo[3.1.1]heptan 2 yl nitrate APINBNO3 2 hydrox y 2,6,6 trimethylbicyclo[3.1.1]heptan 3 yl nitrate PINONIC 2 (3 acetyl 2,2 dimethylcyclobutyl)acetic acid C89PAN 2,2 dimethyl 3 (2 oxoethyl)cyclobutanecarboxylic nitric peroxyanhydride C107OH 2 (3 acetyl 3 hydroxy 2,2 dimethylcyclobutyl)acetalde hyd e
115 Table B 3. Continued. MCM name IUPAC name Structure C109OH 2 (3 (2 hydroxyacetyl) 2,2 dimethylcyclobutyl)acetaldehyd e C5PAN9 nitric 3,4 dioxopentanoic peroxyanhydride CO235C6CHO 3,5,6 trioxoheptanal C109CO 2 (2,2 dimethyl 3 (2 oxoethyl )cyclobutyl) 2 oxoacetaldehyde PINAL 2 (3 acetyl 2,2 dimethylcyclobutyl) acetaldehyde
116 Table B 4 Representative products of pinene SOA and their mass percentages in SOA at different NO x conditions Group ( i j ) k Products name a MW k max b (nm) f b F k c (%) H NO x (A1) L NO x (A2) 1, H s 1 C811PAN 247 183, 161 0.27, 0.11 4.59 3.98 2 PINIC 186 181, 174, 164, 141, 13 5 0.25, 0.06, 0.03, 0.05, 0.09 0.02 1.98 3 C921OOH 204 192, 186, 178, 167, 160, 156 0.09, 0.12,0.09, 0.08, 0.06, 0.07 0.09 1.30 4 C812OOH 190 187, 186, 169, 164, 157 0.14, 0.10, 0.09, 0.05, 0.11 0.04 0.92 5 HOPINONIC 200 176, 171, 169, 162 139 0.25, 0.22, 0.12, 0.03, 0.07 0.04 1.19 1, H m 6 C920PAN 261 196, 182, 171 0.07, 0.03, 0.06 8.40 3.91 7 C98OOH 204 201, 167 0.29, 0.14 2.72 10.43 8 C98NO3 233 188, 183 0.18, 0.03 6.46 2.62 9 C922OOH 220 204, 177 0.32, 0.17 0.09 1. 34 1, H f 10 C7PAN3 233 205, 192 0.08, 0.35 18.43 3.54 2, H s 11 C10PAN2 245 187, 171, 167 0.04, 0.08, 0.04 16.30 5.97 12 C97OOH 188 197, 186, 180, 157, 142, 135 0.05, 0.05, 0.23, 0.05, 0.06, 0.09 0.49 6.13 2, H f 13 C717NO3 203 184, 182, 177, 174, 165, 155 0.08, 0.19, 0.05, 0.06, 0.09, 0.06 5.28 3.12 14 C108OOH 216 202 0.26 4.39 14.88 3, PO 15 APINAOOH 186 160, 158, 156, 153, 139, 137 0.16, 0.13, 0.07, 0.17, 0.09, 0.10 0.07 2.32 16 APINANO3 215 174, 162, 151, 147, 146 0.07, 0.06, 0.2, 0.06, 0.08 0.93 2.42 17 APINBNO3 215 202, 168, 158, 153 0.05, 0.04, 0.06, 0.22 0.59 1.26 3, H s 18 PINONIC 184 173, 169, 165, 151 0.31, 0.04, 0.15, 0.08, 0.05 0.09 0.70 3, H m 19 C89PAN 231 179, 169 0.05, 0.06 3.36 2.14 20 C107 OH 200 182, 175, 162, 145, 141 0.25, 0.06, 0.05, 0.06, 0.14 0.36 3.53 21 C109OH 200 181, 173, 163, 140, 135 0.25, 0.06, 0.03, 0.05, 0.09 0.28 0.82 22 C5PAN9 191 199, 184, 174, 168 0.05, 0.22, 0.25, 0.54 2.60 0.60 4, H f 23 CO235C6CHO 156 159 0.01 2.88 2.85 24 C109CO 182 200 0.25 0.09 0.42 5, H m 25 PINAL 168 168, 164, 163, 157, 156, 152 0.03, 0.15, 0.06, 0.14, 0.05, 0.06 20.12 19.46 max and f are calculated using NDDO base d AM1 semiempirical quantum chemistry method; c: F k is the mass percentage of the k pinene SOA
117 Figure B 1 Comparison of model simulated and measure d concentrations of toluene, O 3 NO x and NO for experiments at different NO x levels: A) high NO x B) mid dle NO x and C) low NO x levels (T1, T2 and T3)
118 Figure B 2 Comparison of model simulated and measured concent rations of pinene, O 3 NO x and NO for experiments at different NO x levels: A) high NO x a nd B) low NO x (A1and A2)
119 Figure B 3 Comparison of the predicted OM T and the measured OM T for different systems: (a) TOL SOA and ( b ) AP SOA under different NO x conditions. T1 T3 for TOL SOA and A1 A2for AP SOA (see Table 1)
120 APPENDIX C SUPPL EMENTARY METERIALS FOR CHAPTER 4 Document C 1 Chamber operation and characterization of photooxidation of pure levoglucosan Levoglucosan ( 99% pur ity, Aldrich) solid particles was dissolved in HPLC grade water to make 0.02M aqueous solution. Levo glucosan aerosol was generated from the aqueous solution using a constant output atomizer which is connected into the outdoor chamber. After the injection of levoglucosan, chamber air was mixed for five minutes using a mixing fan. HONO was introduced to the chamber by passing the clean air through a flask in which 10 m L of 0.1 M NaNO 2 and 10 m L of 10 % H 2 SO 4 reacted to produce HONO. HONO was estimated as the difference in NO x concentration which w ere measured by a NO x analyzer with and without a base denuder ( coated using 1% Na 2 CO 3 + 1% glycerol in ethanol ) (Febo and Perrino, 1991) The initial chamber concentration of levoglucosan was 121ug/m 3 HONO 50ppb, and NO x 124ppb. A f ilter sample was collected using a 13mm Teflon coated borosilicate filter. For each filte r sample L of bornyl acetate solution (2.4mg/m L in acetonitrile) an internal standard, was added. Both Levoglucosan and oxidation products were extracted by sonicating the filter sample with 5m L acetonitrile for one hour. The extracted solution was concentra ted to 1ml and using a dry air stream and transferred to a GC vial In order to derivatize alcol, phonol, and carbo xylic acid, 35mL of N,O Bis(trimethylsilyl)trifluoroacetamide (BSTFA) solution and 15mL pyridine were added to the GC vial The solution wa s stood at 70 for 1 h our Gaseous p roducts from ox idation of levoglucosan were collected using a XAD coated denuder which was located upstream the filter. After the sample collection, the
121 denuder w as extracted using 125mL of acetonitrile. The extracted solution wa s concentrated to 1mL using a r otary e vaporator transferred to a GC vial, and derivertized with BSTFA as mention ed above. The deriv a tized products were analyzed using a gas chromatography ion trap mass spectrometer ( GC ITMS). The GC temperature profile in this study was 80 for 1min ute ramp to 100 at 5 min ute 1 ; ramp to 280 at 10 min ute 1 and hold for 8 min utes For the identification of new products, samples were analyzed i n both electron impact (EI) and chemical ionizati on (CI) modes. Acetonitrile was used as the che mical ionization reagent References Febo, A., and Perrino, C. Prediction and experimental evidence for high air concentration of nitrous acid in indoor environments, Atmospheric Environment. Part A. General Topics, 25, 1055 1061, http://dx.doi.org/10.1016/0960 1686(91)90147 Y 1991.
122 Figure C 1 Time profile of sunlight total ultra violet radiation (TUVT), temperature and relative humidity measured in the UF APHOR E ast chamber on October 30, 2012.
123 Figure C 2 M ass spectra of BSTFA derivatives of levoglucosan oxidation products in EI mode RT is retention time.
124 The mass peak at m/z = 217 and m/z=147 are the typical mass fragmentation pattern for sugar types of compounds. All products show these two peaks in mass fragmentation patterns. P1. The molecular structure was tentatively identified based on the molecular ion peak shown in the CI spectrum in Figure S3. P2. The mass peak at m/z = 243 corresponds with M 45 (~COOH) P3. The mass peak at m/z = 306 and m/z = 288 originate from M 72 [~Si(CH 3 ) 3 +1] and M 88(~OSi(CH 3 ) 3 +1), respectively. P4. The molecular structure was tentatively identified based on the molecular ion peak shown in the CI spectrum in Figure S3 P 5. The mass peak at m/z = 262 is the molecular ion peak. The mass peaks at m/z=247 and m/z=233 correspond with M 15 (~CH 3 ) and M 29 (~CHO), respectively.
125 Figure C 3 M ass spectra of BSTFA derivatives of levoglucosan oxidation products in CI mod e RT is retention time.
126 P1. The mass peak at m/z = 303 corresponds with M 89 [(~OSi(CH 3 ) 3 ]. P2. The mass peaks at m/z=289, m/z=245, and m/z=199 correspond with M 89 [~OSi(CH 3 ) 3 ], M 89 44 [~OSi(CH 3 ) 3 and ~CO 2 ] and M 89 90 [2 ~OSi(CH 3 ) 3 1 ], respectively. P3. The mass peak at m/z = 289 and m/z=199 correspond with M 89 [~OSi(CH 3 ) 3 ] and M 89 90 [2 ~OSi(CH 3 ) 3 1 ], respectively. P3 is isomer of P2. P4. The mass peak at m/z = 287 and m/z=197 correspond with M 89 [~OSi(CH 3 ) 3 ] and M 89 90 [2 ~OSi(CH 3 ) 3 1 ], resp ectively. P 5. The mass peak at m/z = 262 is the molecular ion peak
127 Figure C 4 Reaction pathways for levoglucosan decompositio n in the presence of OH radical.
128 APPENDIX D SUPPL EMENTARY METERIALS FOR CHAPTER 5 Table D 1 Input parameters in Mie code for dry SOA aerosol run lambda cmd gsd n core k core f coat n shell k shell 1 280 138 2 1.44 6.36E 03 0.000 1.33 0.000001 2 290 138 2 1.44 3.98E 03 0.000 1.33 0.000001 3 300 138 2 1.44 3.81E 03 0.000 1.33 0.000001 4 310 138 2 1.44 3.74E 03 0.000 1.33 0.000001 5 320 138 2 1.44 2.96E 03 0.000 1.33 0.000001 6 330 138 2 1.44 1.51E 03 0.000 1.33 0.000001 7 340 138 2 1.44 1.33E 03 0.000 1.33 0.000001 8 350 138 2 1.44 8.63E 04 0.000 1.33 0.000001 9 360 138 2 1.44 4.68E 04 0.000 1.33 0.000001 10 370 138 2 1.44 2.69E 04 0.000 1.33 0.000001 11 380 138 2 1.44 3.47E 04 0.000 1.33 0.000001 12 390 138 2 1.44 3.42E 04 0.000 1.33 0.000001 13 400 138 2 1.44 2.79E 04 0.000 1.33 0.000001 14 410 138 2 1.44 2.41E 04 0.000 1.33 0.000001 15 420 138 2 1.44 2.74 E 04 0.000 1.33 0.000001 16 430 138 2 1.44 1.32E 04 0.000 1.33 0.000001 17 440 138 2 1.44 1.59E 04 0.000 1.33 0.000001 18 450 138 2 1.44 1.60E 04 0.000 1.33 0.000001 19 460 138 2 1.44 1.07E 04 0.000 1.33 0.000001 20 470 138 2 1.44 1.21E 05 0.000 1.33 0.000001 21 480 138 2 1.44 7.29E 05 0.000 1.33 0.000001 22 490 138 2 1.44 1.02E 04 0.000 1.33 0.000001 23 500 138 2 1.44 1.08E 04 0.000 1.33 0.000001 24 510 138 2 1.44 1.53E 04 0.000 1.33 0.000001 25 520 138 2 1.44 1.62E 04 0.000 1.33 0.000001 26 530 138 2 1.44 2.56E 04 0.000 1.33 0.000001 27 540 138 2 1.44 2.36E 04 0.000 1.33 0.000001 28 550 138 2 1.44 2.39E 04 0.000 1.33 0.000001 29 560 138 2 1.44 3.78E 04 0.000 1.33 0.000001 30 570 138 2 1.44 3.78E 04 0.000 1.33 0.000001 31 580 138 2 1.44 3.11 E 04 0.000 1.33 0.000001 32 590 138 2 1.44 3.65E 04 0.000 1.33 0.000001 33 600 138 2 1.44 3.47E 04 0.000 1.33 0.000001 34 610 138 2 1.44 3.28E 04 0.000 1.33 0.000001 35 620 138 2 1.44 3.72E 04 0.000 1.33 0.000001 36 630 138 2 1.44 2.99E 04 0.000 1.33 0.000001 37 640 138 2 1.44 3.04E 04 0.000 1.33 0.000001 38 650 138 2 1.44 8.02E 05 0.000 1.33 0.000001 39 660 138 2 1.44 2.45E 04 0.000 1.33 0.000001 40 670 138 2 1.44 6.25E 05 0.000 1.33 0.000001 41 680 138 2 1.44 8.52E 05 0.000 1.33 0.000001 42 690 138 2 1.44 2.32E 04 0.000 1.33 0.000001 43 700 138 2 1.44 1.34E 04 0.000 1.33 0.000001
129 Table D 1 Continued run lambda cmd gsd n core k core f coat n shell k shell 44 710 138 2 1.44 1.74E 04 0.000 1.33 0.000001 45 720 138 2 1.44 2.16E 04 0.000 1.33 0.00 0001 46 730 138 2 1.44 1.53E 04 0.000 1.33 0.000001 47 740 138 2 1.44 2.45E 04 0.000 1.33 0.000001 48 750 138 2 1.44 1.10E 04 0.000 1.33 0.000001 49 760 138 2 1.44 8.65E 05 0.000 1.33 0.000001 50 770 138 2 1.44 1.51E 04 0.000 1.33 0.000001 51 780 138 2 1.44 1.94E 04 0.000 1.33 0.000001 52 790 138 2 1.44 7.15E 06 0.000 1.33 0.000001 53 800 138 2 1.44 4.96E 05 0.000 1.33 0.000001 Table D 2 Input parameters in Mie code for SOA aerosol at RH of 50% run lambda cmd gsd n core k core f coat n shell k shel l 1 280 150.872 2 1.44 0.0064 0.235 1.33 0.000001 2 290 150.872 2 1.44 0.004 0.235 1.33 0.000001 3 300 150.872 2 1.44 0.0038 0.235 1.33 0.000001 4 310 150.872 2 1.44 0.0037 0.235 1.33 0.000001 5 320 150.872 2 1.44 0.003 0.235 1.33 0.000001 6 330 150. 872 2 1.44 0.0015 0.235 1.33 0.000001 7 340 150.872 2 1.44 0.0013 0.235 1.33 0.000001 8 350 150.872 2 1.44 0.0009 0.235 1.33 0.000001 9 360 150.872 2 1.44 0.0005 0.235 1.33 0.000001 10 370 150.872 2 1.44 0.0003 0.235 1.33 0.000001 11 380 150.872 2 1.4 4 0.0003 0.235 1.33 0.000001 12 390 150.872 2 1.44 0.0003 0.235 1.33 0.000001 13 400 150.872 2 1.44 0.0003 0.235 1.33 0.000001 14 410 150.872 2 1.44 0.0002 0.235 1.33 0.000001 15 420 150.872 2 1.44 0.0003 0.235 1.33 0.000001 16 430 150.872 2 1.44 0.00 01 0.235 1.33 0.000001 17 440 150.872 2 1.44 0.0002 0.235 1.33 0.000001 18 450 150.872 2 1.44 0.0002 0.235 1.33 0.000001 19 460 150.872 2 1.44 0.0001 0.235 1.33 0.000001 20 470 150.872 2 1.44 1E 05 0.235 1.33 0.000001 21 480 150.872 2 1.44 7E 05 0.235 1.33 0.000001 22 490 150.872 2 1.44 0.0001 0.235 1.33 0.000001 23 500 150.872 2 1.44 0.0001 0.235 1.33 0.000001 24 510 150.872 2 1.44 0.0002 0.235 1.33 0.000001 25 520 150.872 2 1.44 0.0002 0.235 1.33 0.000001 26 530 150.872 2 1.44 0.0003 0.235 1.33 0.000001 27 540 150.872 2 1.44 0.0002 0.235 1.33 0.000001 28 550 150.872 2 1.44 0.0002 0.235 1.33 0.000001
130 Table D 2 Continued run lambda cmd gsd n core k core f coat n shell k shell 29 560 150.872 2 1.44 0.0004 0.235 1.33 0.000001 30 570 150.872 2 1.44 0.0004 0.235 1.33 0.000001 31 580 150.872 2 1.44 0.0003 0.235 1.33 0.000001 32 590 150.872 2 1.44 0.0004 0.235 1.33 0.000001 33 600 150.872 2 1.44 0.0003 0.235 1.33 0.000001 34 610 150.872 2 1.44 0.0003 0.235 1.33 0.000001 35 620 150.872 2 1.44 0.0004 0.235 1.33 0.000001 36 630 150.872 2 1.44 0.0003 0.235 1.33 0.000001 37 640 150.872 2 1.44 0.0003 0.235 1.33 0.000001 38 650 150.872 2 1.44 8E 05 0.235 1.33 0.000001 39 660 150.872 2 1.44 0.0002 0.235 1.33 0.000001 40 670 150.872 2 1.44 6E 05 0.235 1 .33 0.000001 41 680 150.872 2 1.44 9E 05 0.235 1.33 0.000001 42 690 150.872 2 1.44 0.0002 0.235 1.33 0.000001 43 700 150.872 2 1.44 0.0001 0.235 1.33 0.000001 44 710 150.872 2 1.44 0.0002 0.235 1.33 0.000001 45 720 150.872 2 1.44 0.0002 0.235 1.33 0.0 00001 46 730 150.872 2 1.44 0.0002 0.235 1.33 0.000001 47 740 150.872 2 1.44 0.0002 0.235 1.33 0.000001 48 750 150.872 2 1.44 0.0001 0.235 1.33 0.000001 49 760 150.872 2 1.44 9E 05 0.235 1.33 0.000001 50 770 150.872 2 1.44 0.0002 0.235 1.33 0.000001 51 780 150.872 2 1.44 0.0002 0.235 1.33 0.000001 52 790 150.872 2 1.44 7E 06 0.235 1.33 0.000001 53 800 150.872 2 1.44 5E 05 0.235 1.33 0.000001 Table D 3 Input parameters in Mie code for dry POA run lambda cmd gsd n core k core f coat n shell k shell 1 280 138 2 1.44 0.0541 0.000 1.33 0.000001 2 290 138 2 1.44 0.0554 0.000 1.33 0.000001 3 300 138 2 1.44 0.0564 0.000 1.33 0.000001 4 310 138 2 1.44 0.0597 0.000 1.33 0.000001 5 320 138 2 1.44 0.0595 0.000 1.33 0.000001 6 330 138 2 1.44 0.057 0.000 1. 33 0.000001 7 340 138 2 1.44 0.0537 0.000 1.33 0.000001 8 350 138 2 1.44 0.0518 0.000 1.33 0.000001 9 360 138 2 1.44 0.0485 0.000 1.33 0.000001 10 370 138 2 1.44 0.0443 0.000 1.33 0.000001 11 380 138 2 1.44 0.0425 0.000 1.33 0.000001 12 390 138 2 1.4 4 0.0363 0.000 1.33 0.000001
131 Table D 3 Continued run lambda cmd gsd n core k core f coat n shell k shell 13 400 138 2 1.44 0.0333 0.000 1.33 0.000001 14 410 138 2 1.44 0.0304 0.000 1.33 0.000001 15 420 138 2 1.44 0.0282 0.000 1.33 0.000001 16 430 138 2 1.44 0.0263 0.000 1.33 0.000001 17 440 138 2 1.44 0.0246 0.000 1.33 0.000001 18 450 138 2 1.44 0.0231 0.000 1.33 0.000001 19 460 138 2 1.44 0.0218 0.000 1.33 0.000001 20 470 138 2 1.44 0.0205 0.000 1.33 0.000001 21 480 138 2 1.44 0.0194 0.000 1.33 0.0 00001 22 490 138 2 1.44 0.0183 0.000 1.33 0.000001 23 500 138 2 1.44 0.0174 0.000 1.33 0.000001 24 510 138 2 1.44 0.0166 0.000 1.33 0.000001 25 520 138 2 1.44 0.0156 0.000 1.33 0.000001 26 530 138 2 1.44 0.0148 0.000 1.33 0.000001 27 540 138 2 1.44 0 .0139 0.000 1.33 0.000001 28 550 138 2 1.44 0.013 0.000 1.33 0.000001 29 560 138 2 1.44 0.0122 0.000 1.33 0.000001 30 570 138 2 1.44 0.0115 0.000 1.33 0.000001 31 580 138 2 1.44 0.0109 0.000 1.33 0.000001 32 590 138 2 1.44 0.01 0.000 1.33 0.000001 33 600 138 2 1.44 0.0093 0.000 1.33 0.000001 34 610 138 2 1.44 0.0087 0.000 1.33 0.000001 35 620 138 2 1.44 0.0081 0.000 1.33 0.000001 36 630 138 2 1.44 0.0076 0.000 1.33 0.000001 37 640 138 2 1.44 0.0071 0.000 1.33 0.000001 38 650 138 2 1.44 0.0065 0.0 00 1.33 0.000001 39 660 138 2 1.44 0.0061 0.000 1.33 0.000001 40 670 138 2 1.44 0.0057 0.000 1.33 0.000001 41 680 138 2 1.44 0.0054 0.000 1.33 0.000001 42 690 138 2 1.44 0.005 0.000 1.33 0.000001 43 700 138 2 1.44 0.0046 0.000 1.33 0.000001 44 710 13 8 2 1.44 0.0042 0.000 1.33 0.000001 45 720 138 2 1.44 0.004 0.000 1.33 0.000001 46 730 138 2 1.44 0.0036 0.000 1.33 0.000001 47 740 138 2 1.44 0.0033 0.000 1.33 0.000001 48 750 138 2 1.44 0.003 0.000 1.33 0.000001 49 760 138 2 1.44 0.0027 0.000 1.33 0 .000001 50 770 138 2 1.44 0.0024 0.000 1.33 0.000001 51 780 138 2 1.44 0.0021 0.000 1.33 0.000001 52 790 138 2 1.44 0.0019 0.000 1.33 0.000001 53 800 138 2 1.44 0.0016 0.000 1.33 0.000001
132 Table D 4 Input parameters in Mie code for POA at RH of 50% run lambda cmd gsd n core k core f coat n shell k shell 1 280 150.872 2 1.44 0.0541 0.235 1.33 0.000001 2 290 150.872 2 1.44 0.0554 0.235 1.33 0.000001 3 300 150.872 2 1.44 0.0564 0.235 1.33 0.000001 4 310 150.872 2 1.44 0.0597 0.235 1.33 0.000001 5 320 15 0.872 2 1.44 0.0595 0.235 1.33 0.000001 6 330 150.872 2 1.44 0.057 0.235 1.33 0.000001 7 340 150.872 2 1.44 0.0537 0.235 1.33 0.000001 8 350 150.872 2 1.44 0.0518 0.235 1.33 0.000001 9 360 150.872 2 1.44 0.0485 0.235 1.33 0.000001 10 370 150.872 2 1.4 4 0.0443 0.235 1.33 0.000001 11 380 150.872 2 1.44 0.0425 0.235 1.33 0.000001 12 390 150.872 2 1.44 0.0363 0.235 1.33 0.000001 13 400 150.872 2 1.44 0.0333 0.235 1.33 0.000001 14 410 150.872 2 1.44 0.0304 0.235 1.33 0.000001 15 420 150.872 2 1.44 0.02 82 0.235 1.33 0.000001 16 430 150.872 2 1.44 0.0263 0.235 1.33 0.000001 17 440 150.872 2 1.44 0.0246 0.235 1.33 0.000001 18 450 150.872 2 1.44 0.0231 0.235 1.33 0.000001 19 460 150.872 2 1.44 0.0218 0.235 1.33 0.000001 20 470 150.872 2 1.44 0.0205 0.2 35 1.33 0.000001 21 480 150.872 2 1.44 0.0194 0.235 1.33 0.000001 22 490 150.872 2 1.44 0.0183 0.235 1.33 0.000001 23 500 150.872 2 1.44 0.0174 0.235 1.33 0.000001 24 510 150.872 2 1.44 0.0166 0.235 1.33 0.000001 25 520 150.872 2 1.44 0.0156 0.235 1.3 3 0.000001 26 530 150.872 2 1.44 0.0148 0.235 1.33 0.000001 27 540 150.872 2 1.44 0.0139 0.235 1.33 0.000001 28 550 150.872 2 1.44 0.013 0.235 1.33 0.000001 29 560 150.872 2 1.44 0.0122 0.235 1.33 0.000001 30 570 150.872 2 1.44 0.0115 0.235 1.33 0.000 001 31 580 150.872 2 1.44 0.0109 0.235 1.33 0.000001 32 590 150.872 2 1.44 0.01 0.235 1.33 0.000001 33 600 150.872 2 1.44 0.0093 0.235 1.33 0.000001 34 610 150.872 2 1.44 0.0087 0.235 1.33 0.000001 35 620 150.872 2 1.44 0.0081 0.235 1.33 0.000001 36 630 150.872 2 1.44 0.0076 0.235 1.33 0.000001 37 640 150.872 2 1.44 0.0071 0.235 1.33 0.000001 38 650 150.872 2 1.44 0.0065 0.235 1.33 0.000001 39 660 150.872 2 1.44 0.0061 0.235 1.33 0.000001 40 670 150.872 2 1.44 0.0057 0.235 1.33 0.000001 41 680 15 0.872 2 1.44 0.0054 0.235 1.33 0.000001
133 Table D 4 Continued run lambda cmd gsd n core k core f coat n shell k shell 42 690 150.872 2 1.44 0.005 0.235 1.33 0.000001 43 700 150.872 2 1.44 0.0046 0.235 1.33 0.000001 44 710 150.872 2 1.44 0.0042 0.235 1.33 0. 000001 45 720 150.872 2 1.44 0.004 0.235 1.33 0.000001 46 730 150.872 2 1.44 0.0036 0.235 1.33 0.000001 47 740 150.872 2 1.44 0.0033 0.235 1.33 0.000001 48 750 150.872 2 1.44 0.003 0.235 1.33 0.000001 49 760 150.872 2 1.44 0.0027 0.235 1.33 0.000001 50 770 150.872 2 1.44 0.0024 0.235 1.33 0.000001 51 780 150.872 2 1.44 0.0021 0.235 1.33 0.000001 52 790 150.872 2 1.44 0.0019 0.235 1.33 0.000001 53 800 150.872 2 1.44 0.0016 0.235 1.33 0.000001 Table D 5 Input parameters in Mie code for dry sulfa te run lambda cmd gsd n core k core f coat n shell k shell 1 250 138 2 1.48 1E 08 0.000 1.33 0.000001 2 300 138 2 1.47 1E 08 0.000 1.33 0.000001 3 350 138 2 1.45 1E 08 0.000 1.33 0.000001 4 400 138 2 1.44 1E 08 0.000 1.33 0.000001 5 450 138 2 1.43 1E 08 0 .000 1.33 0.000001 6 500 138 2 1.43 1E 08 0.000 1.33 0.000001 7 550 138 2 1.43 1E 08 0.000 1.33 0.000001 8 600 138 2 1.43 1E 08 0.000 1.33 0.000001 9 650 138 2 1.43 2E 08 0.000 1.33 0.000001 10 700 138 2 1.43 2E 08 0.000 1.33 0.000001 11 750 138 2 1. 43 7E 08 0.000 1.33 0.000001 12 800 138 2 1.43 9E 08 0.000 1.33 0.000001
134 LIST OF REFERENCES Adler, G., Flores, J. M., Abo Riziq, A., Borrmann, S., and Rudich, Y.: Chemical, physical, and optical evolution of biomass burning aerosol s: a case study, Atmos. Chem. Phys., 11, 1491 1503, 10.5194/acp 11 1491 2011, 2011. Alexander, D. T. L., Crozier, P. A., and Anderson, J. R.: Brown carbon spheres in east Asian outflow and their optical properties, Science, 321, 833 836, 2008. Alves, C. A. Vicente, A., Monteiro, C., Gonalves, C., Evtyugina, M., and Pio, C.: Emission of trace gases and organic components in smoke particles from a wildfire in a mixed evergreen forest in Portugal, Science of The Total Environment, 409, 1466 1475, http://dx.doi.org/10.1016/j.scitotenv.2010.12.025 2011. Anastasio, C., Faust, B. C., and Rao, C. J.: Aromatic Carbonyl Compounds as Aqueous Phase Photochemical Sources of Hydrogen Peroxide in Acidic Sulfate Aerosols, Fogs, and Clouds. 1. Non Phenolic Methoxybenzaldehydes and Methoxyacetophenones with Reductants (Phenols), Environmental Science & Technology, 31, 218 232, 10.1021/es960359g, 1996. Anderson, N. M., and Sekelj, P.: Light absorbing and scat tering properties of nonhaemolysed blood, Physics in Medicine and Biology, 12, 173 184, 1967. Andreae, M., and Gelencser, A.: Black carbon or brown carbon? The nature of light absorbing carbonaceous aerosols, Atmospheric Chemistry and Physics, 6, 3131 3148 2006. Andreae, M. O., and Crutzen, P. J.: Atmospheric Aerosols: Biogeochemical Sources and Role in Atmospheric Chemistry, Science, 276, 1052 1058, 10.1126/science.276.5315.1052, 1997. Baduel, C., Voisin, D., and Jaffrezo, J. L.: Comparison of analytical methods for Humic Like Substances (HULIS) measurements in atmospheric particles, Atmospheric Chemistry and Physics, 9, 5949 5962, 2009. Barker, B. E., and Fox, M. F.: Computer resolution of overlapping electronic absorption bands, Chemical Society Reviews, 9, 1980. Bateman, A. P., Walser, M. L., Desyaterik, Y., Laskin, J., Laskin, A., and Nizkorodov, S. A.: The effect of solvent on the analysis of secondary organic aerosol using electrospray ionization mass spectrometry, Environmental Science & Technology, 42, 7341 7346, 2008. Belay, A.: Measurement of integrated absorption cross section, oscillator strength and number density of caffeine in coffee beans by integrated absorption coefficient technique, Food Chem, 121, 585 590, DOI 10.1016/j.foodchem.2009.12.0 52, 2010.
135 Bertschi, I., Yokelson, R., Ward, D., Babbitt, R., Susott, R., Goode, J., and Hao, W.: Trace gas and particle emissions from fires in large diameter and belowground biomass fuels, Journal of Geophysical Research Atmospheres, 108, 10.1029/2002JD00 2100, 2003. Birmili, W., Schwirn, K., Nowak, A., Petj, T., Joutsensaari, J., Rose, D., Wiedensohler, A., Hmeri, K., Aalto, P., Kulmala, M., and Boy, M.: Measurements of humidified particle number size distributions in a Finnish boreal forest: derivation of hygroscopic particle growth factors, Boreal Environ. Res., 14, 458 480, 2009. Bond, T. C., and Bergstromb, R. W.: Light Absorption by carbonaceous particles: an investigative review, Aerosol Science and Technology 40, 27 67, 2006. Bones, D. L., Henrick sen, D. K., Mang, S. A., Gonsior, M., Bateman, A. P., Nguyen, T. B., and Cooper, W. J.: Appearance of strong absorbers and fluorophores in limonene O 3 secondary organic aerosol due to NH 4 + mediated chemical aging over long time scales, Journal of Geophysi cal Research, 115, D05203, 10.1029/2009JD012864, 2010a. Bones, D. L., Henricksen, D. K., Mang, S. A., Gonsior, M., Bateman, A. P., Tran B. Nguyen, and Cooper, W. J.: Appearance of strong absorbers and fluorophores in limonene O 3 secondary organic aerosol d ue to NH4 + mediated chemical aging over long time scales, Journal of Geophysical Research, 115, D05203, 2010b. Booth, A. M., Montague, W. J., Barley, M. H., Topping, D. O., McFiggans, G., Garforth, A., and Percival, C. J.: Solid state and sub cooled liqui d vapour pressures of cyclic aliphatic dicarboxylic acids, Atmos. Chem. Phys., 11, 655 665, 10.5194/acp 11 655 2011, 2011. Calvert, J. G., and Pitts, J. N.: Photochemistry, John Wiley & Sons: New York, 1966. Campbell, D., Copeland, S., and Cahill, T.: Meas urement of aerosol absorption coefficient from telfon on filters using integrating plate and integrating sphere techniques, Aerosol Science and Technology, 22, 287 292, 1995. Camredon, M., Hamilton, J. F., Alam, M. S., Wyche, K. P., Carr, T., White, I. R., Monks, P. S., Rickard, A. R., and Bloss, W. J.: Distribution of gaseous and particulate pinene ozonolysis Atmospheric Chemistry and Physics, 10, 2893 2917, 2010. Cao, G., and Jang, M.: Secondary organic aerosol formation from toluene photooxidation under various NOx conditions and particle acidity, Atmos. Chem. Phys. Discuss., 8, 14467 14495, 10.5194/acpd 8 14467 2008, 2008. Cao, G., and Jang, M.: An SOA model for toluene oxidation in the presence of inorganic aerosols, E nviron Sci Technol, 44, 727 733, 10.1021/es901682r, 2010.
136 Chang, J. L., and Thompson, J. E.: Characterization of colored products formed during irradiation of aqueous solutions containing H2O2 and phenolic compounds, Atmospheric Environment, 44, 541 551, http://dx.doi.org/10.1016/j.atmosenv.2009.10.042 2010. Chen, Y., and Bond, T. C.: Light absorption by organic carbon from wood combustion, Atmos. Chem. Phys., 10, 1787, 10.5194/acp 10 1773 2 010, 2010. Chung, S., and Seinfeld, J.: Global distribution and climate forcing of carbonaceous aerosols, Journal of Geophysical Research Atmospheres, 107, 10.1029/2001JD001397, 2002. CHYLEK, P., and WONG, J.: EFFECT OF ABSORBING AEROSOLS ON GLOBAL RADIATI ON BUDGET, Geophysical Research Letters, 22, 929 931, 10.1029/95GL00800, 1995. Clark, T., Alex, A., Beck, B., Burkhardt, F., Chandrasekhar, J., Gedeck, P., Horn, A. H. C., Hutter, M., Martin, B., Rauhut, G., Sauer, W., Schindler, T., Steinke, T., and Germa ny: Vamp 8.1, University of Erlangen, Erlangen, 2002. Clegg, S. L., Brimblecombe, P., and Wexler, A. S.: Thermodynamic Model of the Physical Chemistry A, 102, 2155 2171, 10.1021/jp973043j, 1998. De Haan, D., Corrigan, A., Smith, K., Stroik, D., Turley, J., Lee, F., Tolbert, M., Jimenez, J., Cordova, K., and Ferrell, A.: Secondary organic aerosol forming reactions of glyoxal with amino acids, Environmental Science & Technology, 43, 2818 2824, 2009. Del Vecchio, R., and Blough, N. V.: On the Origi n of the Optical Properties of Humic Substances, Environmental Science & Technology, 38, 3885 3891, 10.1021/es049912h, 2004. Dewar, M. J. S., Zoebisch, E. G., Healy, E. F., and Stewart, J. J. P.: The Development and Use of Quantum Mechanical Molecular Mode ls .76. Am1 a New General Purpose Quantum Mechanical Molecular Model, Journal of the American Chemical Society, 107, 3902 3909, 1985. EPA, U. S.: Air quality criteria for particulate matter, Washington, D.C., 2004. Fabian, J., Diaz, L. A., Seifert, G., a nd Niehaus, T.: Calculation of excitation energies of organic chromophores: a critical evaluation, J Mol Struc Theochem, 594, 41 53, Pii S0166 1280(02)00322 6, 2002. Fang, W., Gong, L., Shan, X., Liu, F., Wang, Z., and Sheng, L.: Thermal Desorption/Tunable Vacuum Ultraviolet Time of Flight Photoionization Aerosol Mass Spectrometry for Investigating Secondary Organic Aerosols in Chamber Experiments, Analytical Chemistry, 83, 9024 9032, 10.1021/ac201838e, 2011.
137 Faust, B. C.: Photochemistry of Clouds, Fogs, an d Aerosols, Environmental Science & Technology, 28, 216A 222A, 10.1021/es00054a001, 1994. Febo, A., and Perrino, C.: Prediction and experimental evidence for high air concentration of nitrous acid in indoor environments, Atmospheric Environment. Part A. Ge neral Topics, 25, 1055 1061, http://dx.doi.org/10.1016/0960 1686(91)90147 Y 1991. Forster, P.: Changes in Atmospheric Constituents and in Radiative Forcing http://ipcc wg1.ucar.edu/wg1/ wg1 report.html, 2007. Forstner, H. J. L., Flagan, R. C., and Seinfeld, J. H.: Secondary organic aerosol from the photooxidation of aromatic hydrocarbons: Molecular composition, Environmental Science & Technology, 31, 1345 1358, 1997. Galloway, M. M., Chhabra, P. S., Chan, A. W. H., Surratt, J. D., Flagan, R. C., Seinfeld, J. H., and Keutsch, F. N.: Glyoxal uptake on ammonium sulphate seed aerosol: reaction products and reversibility of uptake under dark and irradi ated conditions, Atmospheric Chemistry and Physics, 9, 3331 3345, 2009. Gao, S., Ng, N. L., Keywood, M., Varutbangkul, V., Bahreini, R., Nenes, A., He, J., Yoo, K. Y., Beauchamp, J. L., Hodyss, R. P., Flagan, R. C., and Seinfeld, J. H.: Particle phase acid ity and oligomer formation in secondary organic aerosol, Environ Sci Technol, 38, 6582 6589, 2004. Gelencser, A., Hoffer, A., Kiss, G., Tombacz, E., Kurdi, R., and Bencze, L.: In situ formation of light absorbing organic matter in cloud water, Journal of A tmospheric Chemistry, 45, 25 33, 10.1023/A:1024060428172, 2003. Glasius, M., Lahaniati, M., Calogirou, A., Di Bella, D., Jensen, N. R., Hjorth, J., Kotzias, D., and Larsen, B. R.: Carboxylic acids in secondary aerosols from oxidation of cyclic monoterpenes by ozone, Environmental Science & Technology, 34, 1001 1010, 2002. Grieshop, A. P., Donahue, N. M., and Robinson, A. L.: Laboratory investigation of photochemical oxidation of organic aerosol from wood fires 2: analysis of aerosol mass spectrometer data, Atmos. Chem. Phys. 9, 2227 2240, 2009. Grosjean, D., Williams, E. L., II., Grosjean, E., Andino, J. M., and Seinfeld, J. H.: Atmospheric oxidation of biogenic hydrocarbons:reaction of ozone with beta pinene, D limonene and trans caryophyllene, Environmen tal Science & Technology, 27, 2754 2758, 1993. Gu, Y., Liou, K., Xue, Y., Mechoso, C., Li, W., and Luo, Y.: Climatic effects of different aerosol types in China simulated by the UCLA general circulation model, Journal of Geophysical Research Atmospheres, 1 11, 10.1029/2005JD006312, 2006.
138 Hallquist, M., Wenger, J. C., Baltensperger, U., Rudich, Y., Simpson, D., Claeys, M., Dommen, J., Donahue, N. M., George, C., Goldstein, A. H., Hamilton, J. F., Herrmann, H., Hoffmann, T., Iinuma, Y., Jang, M., Jenkin, M., J imenez, J. L., Kiendler Scharr, A., Maenhaut, W., McFiggans, G., Mentel, T. F., Monod, A., Prevot, A. S. H., Seinfeld, J. H., Surratt, J. D., Szmigielski, R., and Wildt, J.: The formation, properties and impact of secondary organic aerosol: current and eme rging issues, Atmos. Chem. Phys., 9, 5155 5236, 2009. Hamilton, J. F., Webb, P. J., Lewis, A. C., and Reviejo, M. M.: Quantifying small molecules in secondary organic aerosol formed during the photo oxidation of toluene with hydroxyl radicals, Atmospheric Environment, 39, 7263 7275, 10.1016/j.atmosenv.2005.09.006, 2005. Hays, M. D., Fine, P. M., Geron, C. D., Kleeman, M. J., and Gullett, B. K.: Open burning of agricultural biomass: Physical and chemical properties of particle phase emissions, Atmospheric En vironment, 39, 6747 6764, http://dx.doi.org/10.1016/j.atmosenv.2005.07.072 2005. Hecobian, A., Zhang, X., Zheng, M., Frank, N., Edgerton, E. S., and Weber, R. J.: Water Soluble Organic Aero sol material and the light absorption characteristics of aqueous extracts measured over the Southeastern United States, Atmos. Chem. Phys., 10, 5965 5977, 10.5194/acp 10 5965 2010, 2010. Hille, M. G., and Stephens, S. L.: Mixed Conifer Forest Duff Consumpt ion during Prescribed Fires: Tree Crown Impacts, Forest Science, 51, 417 424, 2005. Hoffer, A., Gelencser, A., Guyon, P., Kiss, G., Schmid, O., Frank, G. P., Artaxo, P., and Andreae, M. O.: Optical properties of humic like substances (HULIS) in biomass bu rning aerosols, Atmos. Chem. Phys., 6, 3563 3570, 10.5194/acp 6 3563 2006, 2006. Howard, P. H.: Handbook of Environmental Fate and Exposure Data for Organic Chemicals. Pesticides., Lewis Publishers, Chelsea, MI., 1991. Hoyle, C., Myhre, G., Berntsen, T., a nd Isaksen, I.: Anthropogenic influence on SOA and the resulting radiative forcing, Atmospheric Chemistry and Physics, 9, 2715 2728, 2009. Hu, D., Tolocka, M., Li, Q., and KamenS, R. M.: A kinetic mechanism for predicting secondary organic aerosol formatio n from toluene oxidation in the presence of NOx and natural sunlight, Atmospheric Environment, 41, 6478 6496, DOI 10.1016/j.atmosenv.2007.04.025, 2007. Iinuma, Y., Mller, C., Berndt, T., Bge, O., Claeys, M., and Herrmann, H.: Evidence for the Existence o Pinene Ozonolysis in Ambient Secondary Organic Aerosol, Environmental Science & Technology, 41, 6678 6683, 10.1021/es070938t, 2007.
139 Im, Y., Jang, M., and Beardsley, R. L.: Simulation of aromatic SOA formation using the lumping model integrated with explicit gas phase kinetic mechanisms and aerosol phase reactions, Atmos. Chem. Phys. Discuss., 13, 5843 5870, 10.5194/acpd 13 5843 2013, 2013. IPCC: Changes in Atmospheric Constituents and in Radiative Forcing, in Climate Change, The Phys ical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovenmental Panel on Climate Change 129, 132, 2007. Jacobson, M. Z.: Isolating nitrated and aromatic aerosols and nitrated aromatic gases as sources of ultravio let light absorption, Journal of Geophysical Research Atmospheres, 104, 3527 3542, 1999. Jang, J., Jang, M., Mui, W., Delcomyn, C. A., Henley, M. V., and Hearn, J. D.: Formation of Active Chlorine Oxidants in Saline Oxone Aerosol, Aerosol Science and Techn ology, 44, 1018 1026, 10.1080/02786826.2010.507612, 2010. Jang, M., and McDow, S. R.: Benz[a]anthracene photodegradation in the presence of known organic constituents of atmospheric aerosols, Environmental Science & Technology, 29, 2654 2660, 10.1021/es000 10a030, 1995. Jang, M., and Kamens, R. M.: Newly characterized products and composition of secondary aerosols from the reaction of a pinene with ozone, Atmospheric Environment, 33, 459 474, 1999. Jang, M., and Kamens, R. M.: Characterization of Secondary A erosol from the Photooxidation of Toluene in the Presence of NOx and 1 Propene, Environmental Science & Technology, 35, 3626 3639, 10.1021/es010676+, 2001a. Jang, M., and Kamens, R. M.: Characterization of secondary aerosol from the photooxidation of tolue ne in the presence of NOx and 1 propene, Environmental Science & Technology, 35, 3626 3639, 2001b. Jang, M., Czoschke, N. M., Northcross, A. L., Cao, G., and Shaof, D.: SOA formation from partitioning and heterogeneous reactions: model study in the presenc e of inorganic species, Environ Sci Technol, 40, 3013 3022, 2006. Jang, M. S., Czoschke, N. M., Lee, S., and Kamens, R. M.: Heterogeneous atmospheric aerosol production by acid catalyzed particle phase reactions, Science, 298, 814 817, 2002. Jaoui, M., and Kamens, R. M.: Modeling aerosol formation from alpha pinene + NOx in the presence of natural sunlight using gas phase kinetics and gas particle partitioning theory, Environ Sci Technol, 35, 1394 1405, 2001.
140 Jeffries, H. E., Gary, M. W., Kessler, M., and S exton, K. G.: Morphecule reaction mechanism, MORPHO, ALLOMORPHIC simulation software, 1998. Joback, K. G., and Reid, R. C.: ESTIMATION OF PURE COMPONENT PROPERTIES FROM GROUP CONTRIBUTIONS, Chemical Engineering Communications, 57, 233 243, 10.1080/009864 48708960487, 1987. Kalberer, M., Paulsen, D., Sax, M., Steinbacher, M., Dommen, J., Prevot, A. S. H., Fisseha, R., Weingartner, E., Frankevich, V., Zenobi, R., and Baltensperger, U.: Identification of polymers as major components of atmospheric organic aer osols, Science, 303, 1659 1662, 2004. Kaul, D. S., Gupta, T., and Tripathi, S. N.: Chemical and microphysical properties of the aerosol during foggy and nonfoggy episodes: a relationship between organic and inorganic content of the aerosol, Atmos. Chem. Ph ys. Discuss., 12, 14483 14524, 10.5194/acpd 12 14483 2012, 2012. Kim, H., and Paulson, S. E.: Real refractive indices and volatility of secondary organic pinene and toluene, Atmos. Chem. P hys. Discuss., 13, 1949 1977, 10.5194/acpd 13 1949 2013, 2013. Kirchstetter, T. W., Novakov, T., and Hobbs, P. V.: Evidence that the spectral dependence of light absorption by aerosols is affected by organic carbon, Journal of Geophysical Research: Atmosph eres, 109, n/a n/a, 10.1029/2004jd004999, 2004. Kirchstetter, T. W., and Thatcher, T. L.: Contribution of organic carbon to wood smoke particulate matter absorption of solar radiation, Atmos. Chem. Phys., 12, 6067 6072, 10.5194/acp 12 6067 2012, 2012. Lack D. A., Langridge, J. M., Bahreini, R., Cappa, C. D., Middlebrook, A. M., and Schwarz, J. P.: Brown carbon and internal mixing in biomass burning particles, Proceedings of the National Academy of Sciences, 10.1073/pnas.1206575109, 2012. Lakowicz, J. R.: P rinciples of fluorescence spectroscopy, 3rd ed., Springer, New York, 2006. Laskin, J., Laskin, A., Roach, P. J., Slysz, G., Anderson, G. A., Nizkorodov, S. A., Bones, D. L., and Nguyen, L. Q.: High resolution desorption electrospray Ionization mass spectro metry for chemical characterization of organic aerosols, Analytical Chemisty, 82, 2010a. Laskin, J., Laskin, A., Roach, P. J., Slysz, G. W., Anderson, G. A., Nizkorodov, S. A., Bones, D. L., and Nguyen, L. Q.: High resolution desorption electrospray ioniza tion mass spectrometry for chemical characterization of organic aerosols, Anal Chem, 82, 2048 2058, 10.1021/ac902801f, 2010b.
141 Leungsakul, S., Jaoui, M., and Kamens, R. M.: Kinetic mechanism for predicting secondary organic aerosol fromation from the reacti on of d limonene with ozone, Environmental Science & Technology, 39, 9583 9594, 2005. Liggio, J., Li, S. M., and McLaren, R.: Heterogeneous reactions of glyoxal on particulate matter: identification of acetals and sulfate esters, Environmental Science & Te chnology, 39, 1532 1541, 2005a. Liggio, J., Li, S. M., and Mclaren, R.: Reactive uptake of glyoxal by particulate matter, J Geophys Res Atmos, 110, Artn D10304 Doi 10.1029/2004jd005113, 2005b. Lin, C. I., Baker, M., and Charlson, R. J.: Absorption coeff cient of atmospheric aerosol: A method for measurement Applied Optics, 12, 1356 1363, 1973. Maria, S. F., Russell, L. M., Gilles, M. K., and Myneni, S. C. B.: Organic Aerosol Growth Mechanisms and Their Climate Forcing Implications, Science, 306, 1921 1924 10.1126/science.1103491, 2004. Martin, M., Tritscher, T., Jurnyi, Z., Heringa, M. F., Sierau, B., Weingartner, E., Chirico, R., Gysel, M., Prvt, A. S. H., Baltensperger, U., and Lohmann, U.: Hygroscopic properties of fresh and aged wood burning partic les, Journal of Aerosol Science, http://dx.doi.org/10.1016/j.jaerosci.2012.08.006 2012. Matsuura, A., Sato, H., Sotoyama, W., Takahashi, A., and Sakurai, M.: AM1, PM3, and PM5 calculations of the absorption maxima of basic organic dyes, Journal of Molecular Structure: THEOCHEM, 860, 119 127, 10.1016/j.theochem.2008.03.028, 2008. Mazzoleni, L. R., Zielinska, B., and Moosmller, H.: Emissions of Levoglucosan, Methoxy Phenols, and Organic Acids from Prescribed Burns, Laboratory Combustion of Wildland Fuels, and Residential Wood Combustion, Environmental Science & Technology, 41, 2115 2122, 10.1021/es061702c, 2007. McIntyre, C., and McRae, C.: Proposed guidelines for sample preparation and ESI MS analysis of humic substances to avoid self esterification, Organic Geochemistry, 36, 543 553, 2005. Menon, S., Hansen, J., Nazarenko, L., and Luo, Y.: Climate effects of black carbon aerosols in China and India, Science, 297, 2250 2253, 10.1126/science.10 75159, 2002. Moosmuller, H., Arnott, W. P., Rogers, C. F., Chow, J., Frazier, C. A., Sherman, L. E., and Dietrich, D. L.: Photoacoustic and Filter Measurements Related to Aerosol Light Absorption During the Northern Front Range Air Quality Study (Colorado 1996/1997), Journal of Geophysical Research, 103:D21 28149 28157, 1998.
142 Moosmuller, H., Chakrabarty, R. K., Ehlers, K. M., and Arnott, W. P.: Absorption Angstrom coefficient, brown carbon, and aerosols: basic concepts, bulk matter, and spherical particles, Atmospheric Chemistry and Physics, 11, 1217 1225, DOI 10.5194/acp 11 1217 2011, 2011. Myhre, C. E. L., and Nielsen, C. J.: Optical properties in the UV and visible spectral region of organic acids relevant to tropospheric aerosols, Atmospheric Chemistry a nd Physics, 4, 1759 1769, 2004. Myhre, G., Berglen, T., Johnsrud, M., Hoyle, C., Berntsen, T., Christopher, S., Fahey, D., Isaksen, I., Jones, T., Kahn, R., Loeb, N., Quinn, P., Remer, L., Schwarz, J., and Yttri, K.: Modelled radiative forcing of the direc t aerosol effect with multi observation evaluation, Atmospheric Chemistry and Physics, 9, 1365 1392, 10.5194/acp 9 1365 2009, 2009. Nakayama, T., Matsumi, Y., Sato, K., Imamura, T., Yamazaki, A., and Uchiyama, A.: Laboratory studies on optical properties o f secondary organic aerosols generated during the photooxidation of toluene and the ozonolysis of alpha pinene, J Geophys Res Atmos, 115, Artn D24204 Doi 10.1029/2010jd014387, 2010a. Nakayama, T., Matsumi, Y., Sato, K., Imamura, T., Yamazaki, A., and Uchiy ama, A.: Laboratory studies on optical properties of secondary organic aerosols generated during the photooxidation of toluene and the ozonolysis of alpha pinene, Journal of Geophysical Research Atmospheres, 115, D24204, Artn D24204 Doi 10.1029/2010jd01438 7, 2010b. Nakayama, T., Sato, K., Matsumi, Y., Imamura, T., Yamazaki, A., and Uchiyama, A.: Wavelength and NOx dependent complex refractive index of SOAs generated from the photooxidation of toluene, Atmospheric Chemistry and Physics, 13, 531 545, 10.5194/ acp 13 531 2013, 2013. Ng, N. L., Chhabra, P. S., Chan, A. W. H., Surratt, J. D., Kroll, J. H., Kwan, A. J., McCabe, D. C., Wennberg, P. O., Sorooshian, A., Murphy, S. M., Dalleska, N. F., Flagan, R. C., and Seinfeld, J. H.: Effect of NOx level on secondar y organic aerosol (SOA) formation from the photooxidation of terpenes, Atmos Chem Phys, 7, 5159 5174, 2007a. Ng, N. L., Kroll, J. H., Chan, A. W. H., Chhabra, P. S., Flagan, R. C., and Seinfeld, J. H.: Secondary organic aerosol formation from m xylene, tol uene, and benzene, Atmos. Chem. Phys. Discuss., 7, 3909 3922, 10.5194/acp 7 3909 2007, 2007b. Nozie`re, B., and Esteve W.: Light absorbing aldol condensation products in acidic aerosols: Spectra, kinetics, and contribution to the absorption index Atmosph eric Environment 41, 1150 1163, 2007.
143 Nozire, B., and Esteve, W.: Organic reactions increasing the absorption index of atmospheric sulfuric acid aerosols, Geophys. Res. Lett., 32, L03812, 10.1029/2004gl021942, 2005. Ofner, J., Krger, H. U., Grothe, H., S chmitt Kopplin, P., Whitmore, K., and Zetzsch, C.: Physico chemical characterization of SOA derived from catechol and guaiacol – a model substance for the aromatic fraction of atmospheric HULIS, Atmos. Chem. Phys., 11, 1 15, 10.5194/acp 11 1 2011, 20 11. Patterson, E. M., and Marshall, B. T.: Diffuse reflectance and diffuse transmission measurements of aerosol absorption at the First International Workshop on light absorption by aerosol particles, Applied Optics, 21, 387 393, 1982. Presto, A. A., Huff Hartz, K. E., and Donahue, N. M.: Secondary Organic Aerosol Production from Terpene Ozonolysis. 2. Effect of NOx Concentration, Environmental Science & Technology, 39, 7046 7054, 10.1021/es050400s, 2005. Putaud, J. P., Raes, F., Van Dingenen, R., Brggeman n, E., Facchini, M. C., Decesari, S., Fuzzi, S., Gehrig, R., Hglin, C., Laj, P., Lorbeer, G., Maenhaut, W., Mihalopoulos, N., Mller, K., Querol, X., Rodriguez, S., Schneider, J., Spindler, G., Brink, H. t., Trseth, K., and Wiedensohler, A.: A European a erosol phenomenology 2: chemical characteristics of particulate matter at kerbside, urban, rural and background sites in Europe, Atmospheric Environment, 38, 2579 2595, http://dx.doi.org/10.1 016/j.atmosenv.2004.01.041 2004. Pschl, U.: Atmospheric Aerosols: Composition, Transformation, Climate and Health Effects, Angewandte Chemie International Edition, 44, 7520 7540, 10.1002/anie.200501122, 2005. Randles, C., Russell, L., and Ramaswamy, V.: Hygroscopic and optical properties of organic sea salt aerosol and consequences for climate forcing, Geophysical Research Letters, 31, 10.1029/2004GL020628, 2004. Rogge, W. F., Hildemann, L. M., Mazurek, M. A., and Cass, G. R.: Sources of Fine Organic Aer osol. 9. Pine, Oak, and Synthetic Log Combustion in Residential Fireplaces, Environmental Science & Technology, 32, 13 22, 10.1021/es960930b, 1998. Rosenfeld, D., Lohmann, U., Raga, G. B., O'Dowd, C. D., Kulmala, M., Fuzzi, S., Reissell, A., and Andreae, M O.: Flood or drought: How do aerosols affect precipitation?, Science, 321, 1309 1313, DOI 10.1126/science.1160606, 2008. Runge, E., and Gross, E. K. U.: Density Functional Theory for Time Dependent Systems, Physical Review Letters, 52, 997 1000, 1984. Sa leh, R., Hennigan, C. J., McMeeking, G. R., Chuang, W. K., Robinson, E. S., Coe, H., Donahue, N. M., and Robinson, A. L.: Absorptivity of brown carbon in fresh and
144 photo chemically aged biomass burning emissions, Atmos. Chem. Phys. Discuss., 13, 11509 1153 6, 10.5194/acpd 13 11509 2013, 2013. Sareen, N., Schwier, A. N., Shapiro, E. L., Mitroo, D., and McNeill, V. F.: Secondary organic material formed by methylglyoxal in aqueous aerosol mimics, Atmospheric Chemistry and Physics, 10, 997 1016, 2010. Sato, K., Hatakeyama, S., and Imamura, T.: Secondary organic aerosol formation during the photooxidation of toluene: NOx dependence of chemical composition, J Phys Chem A, 111, 9796 9808, 10.1021/jp071419f, 2007. Schauer, J. J., Kleeman, M. J., Cass, G. R., and Simo neit, B. R. T.: Measurement of Fireplace Combustion of Wood, Environmental Science & Technology, 35, 1716 1728, 10.1021/es001331e, 2001. Schell, B., Ackermann, I. J., Hass, H., Binkowsk i, F. S., and Ebel, A.: Modeling the formation of secondary organic aerosol within a comprehensive air quality model system, Journal of Geophysical Research Atmospheres, 106, 28275 28293, 2001. Schnaiter, M., Schmid, O., Petzold, A., Fritzsche, L., Klein, K. F., Andreae, M. O., Helas, G., Thielmann, A., Gimmler, M., Mhler, O., Linke, C., and Schurath, U.: Measurement of Wavelength Resolved Light Absorption by Aerosols Utilizing a UV VIS Extinction Cell, Aerosol Science and Technology, 39, 249 260, 10.1080/ 027868290925958, 2005. Schulz, M., Textor, C., Kinne, S., Balkanski, Y., Bauer, S., Berntsen, T., Berglen, T., Boucher, O., Dentener, F., Guibert, S., Isaksen, I., Iversen, T., Koch, D., Kirkevag, A., Liu, X., Montanaro, V., Myhre, G., Penner, J., Pitari, G., Reddy, S., Seland, O., Stier, P., and Takemura, T.: Radiative forcing by aerosols as derived from the AeroCom present day and pre industrial simulations, Atmospheric Chemistry and Physics, 6, 5225 5246, 2006. Schwarzenbach, R. P., Gschwend, P. M., and Imboden, D. M.: Environmental Organic Chemistry, John Wiley&Sons Inc., New York, 1993. Seinfeld, J. H., and Pandis, S. N.: Atmospheric Chemistry and Physics, Wiley, New York, 1998. Shapiro, E. L., Szprengiel, J., Sareen, N., Jen, C. N., Giordano, M. R., an d McNeill, V. F.: Light absorbing secondary organic material formed by glyoxal in aqueous aerosol mimics, Atmos Chem Phys, 9, 2289 2300, 2009. Song, C., Na, K., and Cocker, D. R.: Impact of the Hydrocarbon to NOx Ratio on Secondary Organic Aerosol Formatio n, Environmental Science & Technology, 39, 3143 3149, 10.1021/es0493244, 2005.
145 Stein, S. E., and Brown, R. L.: Estimation of normal boiling points from group contributions, Journal of Chemical Information and Computer Sciences, 34, 581 587, 10.1021/ci00019 a016, 1994. Sun, H., Biedermann, L., and Bond, T. C.: Color of brown carbon: A model for ultraviolet and visible light absorption by organic carbon aerosol, Geophys. Res. Lett., 34, L17813, 10.1029/2007gl029797, 2007. Kleindienst, T. E., Edney, E. O., Offenberg, J. H., Lewandowski, M., Jaoui, M., Maenhaut, W., Claeys, M., Flagan, R. C., and Seinfeld, J. H.: Organosulfate Formation in Bio genic Secondary Organic Aerosol, The Journal of Physical Chemistry A, 112, 8345 8378, 10.1021/jp802310p, 2008. Tegen, I., Lacis, A., and Fung, I.: The influence on climate forcing of mineral aerosols from disturbed soils, Nature, 380, 419 422, 10.1038/3804 19a0, 1996. Tolocka, M. P., Jang, M., Ginter, J. M., Cox, F. J., Kamens, R. M., and Johnston, M. V.: Formation of oligomers in secondary organic aerosol, Environ Sci Technol, 38, 1428 1434, 2004. Zellner, R., Exner, M., and Herrmann, H.: Absolute OH quantu m yields in the laser photolysis of nitrate, nitrite and dissolved H2O2 at 308 and 351 nm in the temperature range 278 353 K, Journal of Atmospheric Chemistry, 10, 411 425, 1990. Zerner, M.: Reviews in Computational Chemistry, 2, 313, 1991. Zhang, X., Lin, Y., Surratt, J., and Weber, R.: Sources, Composition and Absorption Angstrom Exponent of Light absorbing Organic Components in Aerosol Extracts from the Los Angeles Basin, Environmental Science & Technology, 47, 3685 3693, 10.1021/es305047b, 2013. Zhao, L., Ni, N., and Yalkowsky, S. H.: A Modification of Trouton's Rule by Simple Molecular Parameters for Hydrocarbon Compounds, Industrial & Engineering Chemistry Research, 38, 324 327, 10.1021/ie9803570, 1998. Zhong, M., and Jang, M.: Light absorption coefficient measurement of SOA using a UV Visible spectrometer connected with an integrating sphere, Atmospheric Environment, 45, 4263 4271, 10.1016/j.atmosenv.2011.04.082, 2011. Zhong, M., Jang, M., Oliferenko, A., Pillai, G. G., and Katritzky, A. R.: The SOA formation model combined with semiempirical quantum chemistry for predicting UV Vis absorption of secondary organic aerosols, Physical Chemistry Chemical Physics, 14, 9058 9066, 2012.
146 BIOGRAPHICAL SKETCH Min Zhong was born in Sichuan, China in 1980 Min spent her first 22 years in China. Since high school, Min has decided to be dedicated to the environmental career. After earning the B achelor of Science degree in environmental engineering, Min went to South Korea to obtain the M aster of Science d egree in environmental engineering at the Pohang University of Science and Technology After working in an environmental engineering company, Min came to US for air pollution research in the Department of Environmental Engineering Sciences at the Universi ty of Florida. She received her Ph.D degree from the University of Florida in the summer of 2013.