Organic particulate matter represents a significant fraction of sub-micrometer atmospheric aerosol mass. However, organic aerosol (OA) consists of thousands of different organic compounds making the simulation of its concentration, chemical evolution, physical and chemical properties extremely challenging. The identity of the great majority of these compounds remains unknown. The volatility of atmospheric OA is one of its most important physical properties since it determines the partitioning of these organic compounds between the gas and particulate phases. The use of lumped compounds with averaged properties is a promising solution for the representation of OA in atmospheric chemical transport models. The two-dimensional volatility basis set (2D-VBS) is a proposed method used to describe OA distribution as a function of the volatility and oxygen content of the corresponding compounds. In the first part of the work we evaluate our ability to measure the OA volatility distribution using a thermodenuder (TD). We use a new method combining forward modeling, introduction of ‘experimental’ error and inverse modeling with error minimization for the interpretation of TD measurements. The OA volatility distribution, its effective vaporization enthalpy, the mass accommodation coefficient and the corresponding uncertainty ranges are calculated. Our results indicate that existing TD-based approaches quite often cannot estimate reliably the OA volatility distribution, leading to large uncertainties, since there are many different combinations of the three properties that can lead to similar thermograms. We propose an improved experimental approach combining TD and isothermal dilution measurements. We evaluate this experimental approach using the same model and show that it is suitable for studies of OA volatility in the lab and the field. Measurements combining a thermodenuder (TD) and a High Resolution Time-of-Flight Aerosol Mass Spectrometer (HR-ToF-AMS) took place during summer and winter in Paris, France as part of the collaborative project MEGAPOLI and during the winter of 2013 in the city of Athens. The above volatility estimation method with the uncertainty estimation algorithm is applied to these datasets in order to estimate the volatility distribution for the organic aerosol (OA) and its components during the two campaigns. The concentrations of the OA components as a function of temperature were measured combining data from the thermodenuder and the aerosol mass spectrometer (AMS) with Positive Matrix Factorization (PMF) analysis. Combining the bulk average O:C ratios and volatility distributions of the various factors, our results are placed into the two-dimensional volatility basis set (2D-VBS) framework. The OA factors cover a broad spectrum of volatilities with no direct link between the average volatility and average O:C of the OA components. An intercomparison among the OA components of both campaigns and their physical properties is also presented. The approach combining thermodenuder and isothermal dilution measurements is tested in smog chamber experiments using OA produced during meat charbroiling. The OA mass fraction remaining is measured as a function of temperature in the TD and as a function of time in the isothermal dilution chamber. These two sets of measurements are used together to estimate the volatility distribution of the OA and its effective vaporization enthalpy and accommodation coefficient. In the isothermal dilution experiments approximately 20% of the OA evaporate within 15 min. In the TD almost all the OA evaporated at approximately 200oC. The resulting volatility distributions suggest that around 60-75% of the cooking OA (COA) at concentrations around 500 μg m-3 consists of low volatility organic compounds (LVOCs), 20-30% of semi-volatile organic compounds (SVOCs) and around 10% of intermediate volatility organic compounds (IVOCs). The estimated effective vaporization enthalpy of COA is 100 ± 20 kJ mol-1 and the effective accommodation coefficient is around 0.05. The characteristics of the COA factor from the Athens campaign are compared to those of the OA produced from meat charbroiling in these experiments. In the next step, different parameterizations of the organic aerosol (OA) formation and evolution in the two-dimensional Volatility Basis Set (2D-VBS) framework are evaluated using ground and airborne measurements collected in the 2012 Pan-European Gas AeroSOls-climate-interaction Study (PEGASOS) field campaign in the Po Valley, Italy. A number of chemical schemes are examined, taking into account various functionalization and fragmentation pathways for biogenic and anthropogenic OA components. Model predictions and measurements, both at the ground and aloft, indicate a relatively oxidized OA with little average diurnal variation. Total OA concentration and O:C ratios were reproduced within experimental error by a number of chemical aging schemes. Anthropogenic SOA is predicted to contribute 15-25% of the total OA, while SOA from intermediate volatility compounds oxidation another 20-35%. Biogenic SOA contributions varied from 15 to 45% depending on the modeling scheme. The average OA and O:C diurnal variation and their vertical profiles showed a surprisingly modest sensitivity to the assumed vaporization enthalpy for all aging schemes. This can be explained by the intricate interplay between the changes in partitioning of the semivolatile compounds and their gas-phase chemical aging reactions. The same set of different parameterizations of the organic aerosol (OA) formation and evolution in the two-dimensional Volatility Basis Set (2D-VBS) framework are evaluated using ground measurements collected in the 2013 PEGASOS field campaign in the boreal forest station of Hyytiälä in Southern Finland. The most successful is the simple functionalization scheme of Murphy et al. (2012) while all seven aging schemes have satisfactory results, consistent with the ground measurements. Despite their differences, these schemes predict similar contributions of the various OA sources and formation pathways. Anthropogenic SOA is predicted to contribute 11- 18% of the total OA, while SOA from intermediate volatility compounds oxidation another 18- 27%. The highest contribution comes from biogenic SOA, as expected contributing 40 to 63% depending on the modeling scheme. The primary OA contributes 4% while the SOA resulting from the oxidation of the evaporated POA varies between 4 to 6%. Finally, 5-6% is according to the model the results of long range transport from outside the modeling domain.

Toluene is one of the most prevalent aromatic volatile organic compounds (VOCs) in the atmosphere and has large secondary organic aerosol (SOA) yields compared to many other aromatic VOCs. Recent photo-oxidation studies highlight that toluene oxidation produces more SOA than observed previously, particularly at low levels of nitrogen oxides (NOx). This study focuses on: 1.) the development of a gas-phase chemical mechanism describing toluene oxidation by hydroxyl radicals (OH); 2.) the prediction of SOA formation from toluene oxidation products; and 3.) the impact of NOx level on SOA formation. The oxidation mechanism, which includes multiple pathways after the initial OH attack, has been incorporated into the Caltech Atmospheric Chemistry Mechanism (CACM). Toluene concentrations simulated in chamber experiments by the updated CACM as a function of time are typically within 5% of observed values for most experiments. Predicted ozone and NO2 concentrations are typically within 15% of the experimental values. The gas-phase mechanism indicates the importance of bicyclic peroxy radical reactions in determining the product distribution and thus the likelihood of SOA formation. A gas-aerosol partitioning model is used in conjunction with the gas-phase mechanism to simulate SOA formation. Predicted SOA concentrations are typically within 15% of the experimental values. Under low NOx conditions, simulation shows that more than 98% of SOA mass is contributed by bicyclic products from reactions between bicyclic peroxy radicals and other peroxy radicals. Increasing NOx levels cause bicyclic peroxy radicals to react with NO or nitrate radical, leading to fragmentation products that are less likely to form SOA. SOA yield dropped from 19.26% with zero initial NOx to 13.27% with 100 ppb initial NO because of the change in the amount of toluene consumed. Composition of NOx also has an impact on SOA yield and formation, showing that NO has a greater impact on SOA yield and formation than NO2.

In the first part of this dissertation, we study the aging of soot, a representative type of primary aerosols, in the presence of OH-initiated oxidation products of toluene. Monodisperse soot particles are introduced into an environmental chamber where toluene is oxidized by OH radicals. The variations in soot particle properties are simultaneously monitored, including particle size, mass, organic mass faction, hygroscopicity, and optical properties. The changes in particle properties are found to be largely governed by the thickness of the organic coating that is closely related to reaction time and initial reactant concentrations. Derived from particle size and mass, the effective density increases while dynamic shape factor decreases as the organic coating grows, suggesting a compaction of the soot morphology. As the organic coating grows, the particles become more hygroscopic and have enhanced light scattering and absorption. The second part discusses the potential reactions between amines and some aerosol constituents and alteration of aerosol properties. The reactions between alkylamines and ammonium sulfate/bisulfate have been studied using a low-pressure fast flow reactor coupled to a mass spectrometer at 293 K. Alkylamines react with ammonium sulfate/bisulfate to form alkylaminium sulfates, suggesting the existence of alkylaminium salts in particle phase. We have extended our study to characterize the physicochemical properties of alkylaminium sulfates. The hygroscopicity, thermostability, and density of five representative alkylaminium sulfates have been measured by an integrated aerosol analytical system. All alkylaminium sulfate aerosols show monotonic size growth when exposed to increasing relative humidity. Mixing ammonium sulfate with alkylaminium sulfates lowers the deliquescence point corresponding to ammonium sulfate. Alkylaminium sulfates are thermally comparable to or more stable than ammonium sulfate. The densities of alkylaminium sulfate particles are lower than that of ammonium sulfate. Our results suggest that the organic compounds can effectively alter the composition and properties of atmospheric aerosols, considerably influencing the impacts of aerosols on air quality, climate forcing, and human health.

I advanced the understanding of particle morphology and its implications for the behavior and effects of atmospheric aerosol particles. I have developed new experimental methods for the Aerosol Optical Tweezers (AOT) system and expanded the AOT’s application into studying realistic secondary organic aerosol (SOA) particle phases. The AOT is a highly accurate system developed to study individual particles in real-time for prolonged periods of time. While previous AOT studies have focused on binary or ternary chemical systems, I have investigated complex SOA, and how they interact with other chemical phases, and the surrounding gas-phase. This work has led to new insights into liquid-liquid phase separation and the resulting particle morphology, the surface tension, solubility, and volatility of SOA, and diffusion coefficients of SOA phases. I designed a new aerosol optical tweezers chamber for delivering a uniformly mixed aerosol flow to the trapped droplet’s position. I used this chamber to determine the phase-separation morphology and resulting properties of complex mixed droplets. A series of experiments using simple compounds are presented to establish my ability to use the cavity enhanced Raman spectra to distinguish between homogenous single-phase, and phase-separated core-shell or partially-engulfed morphologies. I have developed a new algorithm for the analysis of whispering gallery modes (WGMs) present in the cavity enhanced Raman spectra retrieved from droplets trapped in the AOT. My algorithm improves the computational scaling when analyzing core-shell droplets (i.e. phase-separated or biphasic droplets) in the AOT, making it computationally practical to analyze spectra collected over many hours at a few Hz. I then demonstrate for the first time the capture and analysis of SOA on a droplet suspended in an AOT. I examined three initial chemical systems of aqueous NaCl, aqueous glycerol, and squalane at ~ 75% relative humidity. For each system I added α-pinene SOA – generated directly in the AOT chamber – to the trapped droplet. The resulting morphology was always observed to be a core of the initial droplet surrounded by a shell of the added SOA. By combining my AOT observations of particle morphology with results from SOA smog chamber experiments, I conclude that the α-pinene SOA shell creates no major diffusion limitations for water, glycerol, and squalane under humid conditions. My AOT experiments highlight the prominence of phase-separated core-shell morphologies for secondary organic aerosols interacting with a range of other chemical phases. The unique analytical capabilities of the aerosol optical tweezers provide a new approach for advancing the understanding of the chemical and physical evolution of complex atmospheric particulate matter, and the important environmental impacts of aerosols on atmospheric chemistry, air quality, human health, and climate change.

The composition, concentration, and size of submicron aerosols were measured with a time resolution of five minutes by an Aerodyne high-resolution time-of-flight aerosol mass spectrometer (HR-ToF-AMS) at a rural location northwest of the Dallas-Fort Worth, TX, area for the month of June 2011. A TSI, Inc., Model AE51 aethalometer using an optical absorption technique also was deployed to measure black carbon (BC) concentrations. The total measured PM1 mass concentration ranged between 1.0 µg/m3 and 17.1 µg/m3, with a mean and standard deviation of 4.6± 2.7 µg/m3. Significant variability is observed in the time series of total PM1 and of all four HR-ToF-AMS species, particularly between June 21 and 25. The average aerosol composition was dominated by organic matter (52.1 ± 14.8%) and sulfate (28.8 ± 11.8%). Organic aerosol concentrations were positively correlated with tracers of combustion carbon monoxide (CO) and BC, the coefficients of determination were r2=064 and r2=0.48, respectively. Because of the large influence of organics on total aerosol concentration, organic data were analyzed in the context of ΔOA/ΔCO, which typically is used to investigate the relative importance of secondary organic aerosol. The average ∆OA/∆CO for the data used was 64.0 ± 26.9 µg/ (m3 ppmv), which is typical of an aged air mass. Other metrics of age include the ratio of OOAI (more oxidized) to total oxidized organic aerosol (OOA), the ratio of sulfate to total sulfur, the ratio of its oxidation products to isoprene, and the ratio of nitrogen oxides to total reactive nitrogen. All metrics point to aged air masses, but variations in these age matrices, particularly during one period of enhanced ΔOA/ΔCO, help elucidate the contributions of various precursors and processes to organic aerosols at the site.

Epoxides have been recently identified as one of the intermediate species in the gas phase oxidation of alkenes. This study investigates the reaction of isoprene oxide and alpha-pinene oxide with sulfuric acid to identify the potential of epoxides as important secondary organic aerosol (SOA) precursors. The reaction was explored using different methods to understand the factors governing the reaction rate and the types of products formed under different conditions. Uptake experiments of epoxides on sulfuric acid using Ion drift-Chemical Ionization Mass Spectrometry (ID-CIMS) showed an irreversible uptake of epoxides at room temperature resulting in the formation of less volatile products like diols, organosulfates and acetals. However, at lower temperatures, dehydration of diols and some rearrangement was the preferred reaction pathway resulting in the formation of higher volatility compounds like hydroxy-alkenes and aldehydes. The uptake coefficients of isoprene oxide and alpha-pinene oxide at room temperature using 96% wt acid were found to be 4x10^-2 and 0.8x10^-2, respectively. Spectroscopic study using Attenuated total reflection-Fourier transform infrared technique (ATR-FTIR) revealed that for both the epoxides, diols were the major identifiable products at low acid concentrations. At higher acid concentrations, acetal formation was observed in case of isoprene oxide, while organosulfate formation was seen for alpha-pinene oxide. No products were identified under neutral conditions due to slow reaction. Bulk studies using Nuclear Magnetic Resonance (NMR) spectroscopy conducted at low acid concentrations showed the presence of 1,2- and 1,4-diols as the major products for isoprene oxide, similar to the results from the ATR-FTIR experiments. Additionally, aldehyde formation was also observed. For alpha-pinene oxide, organosulfate formation was observed in all NMR experiments, unlike ATR-FTIR results, where organosulfate formation was observed only at high acid concentrations. These observations can be attributed to the kinetic isotope effect (KIE) due to use of D2SO4/D2O in NMR experiments rather than H2SO4/H2O. The percent yield of organosulfate products was proportional to the amount of available acidic sulfate. The results from this study suggest that acid hydrolysis of epoxides can result in the formation of a wide range of products under different conditions, that can contribute to SOA growth. It proves that epoxides can be efficient SOA precursors for ambient conditions prevailing in an urban atmosphere.

We have developed novel experiments aimed at understanding whether and how quickly organic aerosols (OA) mix using single-particle mass spectrometry, as different treatments of mixing in regional models significantly affect predicted mass and composition. First, we designed experiments that separate OA formation chemistry from thermodynamics to test whether two populations of particles equilibrate with eachother through the gas phase on experimental timescales. Single-particle mass spectrometry measurements from the aerosol mass spectrometer (AMS) allowed us to quantify the extent of mixing that had occurred. We calibrated this technique using pure-component aerosols with known vapor pressure and phase state, the results of which agreed with a condensation-evaporation model. We then applied these techniques to three atmospherically-relevant situations to determine that: 1) anthropogenic secondary OA (aSOA) does not mix with a surrogate for hydrophobic primary OA (POA), 2) biogenic SOA (bSOA) does not mix with hydophobic POA, and 3) bSOA shows significant mixing with aSOA. The sum of these experiments show that these complex interactions can be measured for atmospherically important systems, a first step towards quantifying activity coefficients for complex OA mixtures. We also investigated mixing within individual particles, using mixed-particles of squalane (a surrogate for hydrophobic POA) and SOA from ↵+pinene + O3 that we determined to contain two separate phases. In these experiments, after formation of the mixed-particles, we perturbed smog chamber with a heat ramp. These data revealed that squalane is able to quickly evaporate from the mixed-particles, and that almost all of the SOA is comprised of material lower in volatility than squalane (a low-volatility constituent of pump oil). For this latter “comparative volatility analysis,” we had to correct for the highly variable collection efficiency (CE) of the mixed particles to correctly calculate the mass fraction of SOA remaining. One of the larger implications of this work is highly dependent on the particle morphology, which we were not able to determine definitively: if indeed the particles are coreshell with squalane inside a thick layer of SOA, our results show that diffusivity within SOA is not ultra-low. Lastly, we present work that furthers our understanding of single-particle CE in the AMS, a quantity especially important for experiments where particle phase is dynamic or there are two separate populations of particles. We report the particle CE of SOA, ammonium sulfate, ammonium nitrate, and squalane. We also determine that half of SOA particles that give meaningful signal, do so at a time later than would be predicted based on their optically-measured flight time through the instrument. We present convincing evidence that the nature of this delay is due to particles ricocheting around the ionization region of the instrument before vaporizing on an auxillary surface near the the vaporizer. This process affects how much mass signal comes from a particle, the particle mass spectrum, and the bulk mass distribution derived from particle time-of-flight mode. Our results also show that while there is no size dependence to CE for SOA, particles that have passed through a thermodenuder have lower CE, implicating oxidation state and/or volatility as a controller of particle bounce.

La combustion de bois, ou plus largement de la biomasse, est une source de pollution très importante en particules atmosphériques en hiver, particulièrement en France. Si les émissions primaires ont été étudiées depuis de nombreuses années, il existe de grandes incertitudes sur le devenir de ces émissions dans l’atmosphère. Le travail de thèse a donc porté sur l’étude du vieillissement en chambre de simulation atmosphérique de l’aérosol émis par 3 appareillages pour le chauffage au bois (conçus entre 2000 et 2010 et représentatifs de la politique de renouvellement mis en place par l’ADEME), avec un intérêt particulier pour le potentiel de formation d’aérosol organique secondaire (SOA) et la modification de l’empreinte chimique à l’échelle moléculaire au cours du transport atmosphérique. Les expériences ont montré un potentiel de formation de SOA très important. La concentration en OA peut ainsi être multipliée par 7 (1.5 – 7.1) entre l’émission et après un temps de résidence atmosphérique équivalent à 5 h. Cette étude met également en évidence l’impact de l’efficacité de combustion sur les facteurs d’émission et par conséquent le rôle crucial de l’opérateur. L’étude à l’échelle moléculaire a mis en évidence la formation de composés susceptibles de servir de marqueurs de combustion de biomasse âgée, principalement des nitrocatéchols. Enfin, cette étude démontre le rôle clé de la volatilisation du lévoglucosan, principal marqueur organique de la combustion de biomasse, au cours du processus de dilution dans l’atmosphère, et pose clairement la question de la pertinence des constantes cinétiques de dégradation préalablement calculées en chambre de simulation atmosphérique. / Biomass burning is in winter a main source of air pollution by particulate matter, especially in France. While primary emissions have been characterized extensively before, few studies have addressed the aging of these emissions in the atmosphere and large uncertainties remain. Therefore, the objectives of this thesis was to study in a smog chamber the aging of the aerosol emitted by 3 different woodstoves used for residential heating (fabricated from between 2000 and 2010, and representative of the policy engaged by the French environmental agency to renew the appliances across the country), with a specific focus on the Secondary Organic Aerosol (SOA) production potential and the modification of the chemical fingerprint of the emissions at the molecular level during their transport in the atmosphere. The experiments showed the SOA production potential can be significant. The OA concentration can be increased by up to a factor of 7 (1.5 – 7.1) after being aged in the smog chamber with a time equivalent to 5 hours in the atmosphere. The study also further demonstrated the influence of the combustion efficiency on the emissions and implicitly the role of the operator. The study of the composition of the aerosol at the molecular level showed the formation of compounds, likely to serve as markers for aged biomass burning, mainly nitrocatechols. Finally, the work also illustrates the influence of the volatilization of levoglucosan, main marker of biomass burning, during the dilution process occurring in the atmosphere, and challenge the pertinence of the degradation rate constant determined previously in smog chamber.

Ozone (O₃) and terpenoids react to produce secondary organic aerosol (SOA). This work explored novel ways that these reactions form SOA indoors, with five investigations, in two categories: investigations of (i) the impacts of particle controls on indoor SOA formation, and (ii) two fundamental aspects of indoor SOA formation. For category (i), two investigations examined the particle control devices of ion generators, which are air purifiers that are ineffective at removing particles and emit ozone during operation. With a terpenoid source present (an air freshener), ion generators acted as steady-state SOA generators, both in a 15 m³ chamber and 27 m³ room. The final investigation in category (i) modeled how heating, ventilating, and air-conditioning (HVAC) systems influence SOA formation. Influential HVAC parameters were flow rates, particle filtration, and indoor temperature for residential and commercial models, as well as ozone removal by particle-laden filters for the commercial model. For category (ii), the first investigation measured SOA formation from ozone reactions with single terpenoids and terpenoid mixtures in a 90 L Teflon-film chamber, at low and high ozone concentrations. For low ozone, experiments with only d-limonene yielded the largest SOA number formation, relative to other mixtures, some of which had three times the effective amount of reactive terpenoids. This trend was not observed for high ozone experiments, and these results imply that ozone-limited reactions with d-limonene form byproducts with high nucleation potential. The second investigation in category (ii) explored SOA formation from ozone reactions with surface-adsorbed terpenoids. A model framework was developed to describe SOA formation due to ozone/terpenoid surface reactions, and experiments in a 283 L chamber determined the SOA yield for ozone/d-limonene surface reactions. The observed molar yields were 0.14–0.16 over a range of relative humidities, and lower relative humidity led to higher SOA number formation from surface reactions. Building materials on which ozone/d-limonene surface reactions are predicted to lead to substantial SOA formation are those with initially low surface reactivity, such as glass, sealed materials, or metals. The results from category (ii) suggest significant, previously unexplored mechanisms of SOA number formation indoors. / text

Aerosols play important roles in atmospheric and environmental processes. Not only do they impact human health, they also affect visibility and climate. Despite recent advances made to under their sources and fate, there remains a limited understanding of the mechanisms that lead to the formation of aerosols and their ultimate fate in the atmosphere. These knowledge gaps provide the crux of the research reported herein, which has focused on identifying novel sources of atmospheric aerosol, characterizing its physical and optical properties, and rationalizing these properties using an in-depth knowledge of the molecular level mechanisms that led to its formation. Upon mowing, turfgrasses emit large amounts of green leaf volatiles which can then be oxidized by ozone to form SOA. Overall, the mowing of lawns has the potential to contribute nearly 50 µg SOA per square meter of lawn mowed. This SOA contribution is on the same order of magnitude as other predominant SOA sources (isoprene, monoterpenes, sesquiterpenes). Turfgrasses represent an interesting and potentially meaningful SOA source because they contribute to SOA and also because they cover large land areas in close proximity to oxidant sources. Another related SOA precursor is sugarcane, which is in the same family as turfgrass and is among the largest agricultural crops worldwide. Globally, the ozonolysis of sugarcane has the potential to contribute 16 Mg SOA to the atmosphere, compared to global estimates of SOA loading that range from 12-70 Tg SOA. In order to fully understand the role of atmospheric SOA on the radiative budget (and therefore climate), it is also important to understand its optical properties; its ability to absorb vs scatter light. Turfgrass and sugarcane produced SOA that was weakly absorbing while its scatter efficiency was wavelength and size-dependent. Interestingly, SOA formed under both dry (10% RH) and wet (70% RH) conditions had the same bulk chemical properties (O:C), yet significantly different optical properties, which was attributed to differences in molecular-level composition. The work presented herein represents a unique, inclusive study of SOA precursors. A complete understanding of the chemistry leading to SOA formation is used to understand its physical and optical properties and evaluate these large-scale effects of SOA from these precursors.