CAPS: Publications in Progress

A framework is presented and evaluated for the parameterization of smog chamber results for use in atmospheric Chemical Transport Models (CTMs). The parameterization uses an absorptive partitioning model to describe the gas-particle partitioning of semivolatile aerosols. The key points of the framework are (1) the ability to fit experiments from several types of chamber experiments, including fixed-temperature, variable-temperature, and time-resolved experiments; (2) the use of a basis set of surrogate compounds characterized by fixed effective saturation concentrations instead of the more commonly used variable saturation concentrations; (3) more realistic uncertainties for aerosol formation in the cases of extrapolation to concentrations and temperatures outside of the original experimental range; (4) quantitative determination of the enthalpy of vaporization needed to give realistic temperature dependence in CTMs. The features of this data analysis and fitting framework are demonstrated using simulated data, and using data for α-pinene ozonolysis. Representation of semivolatile and SOA partitioning using large numbers (~7) of fixed effective saturation concentrations is shown to be feasible and have advantages over more minimal parameterizations. For the α-pinene ozonolysis test case considered, a parameterization with 2 effective saturation components was also found to fit experimental data well.

This paper describes a method for measuring aerosol wall loss rates and condensation rates as a function of time in smog chamber experiments. The wall loss rate of particles is size dependent and this size dependence depends on the turbulence in the bag, the size and shape of the bag and the charges on the particles, all of which may be changing during an experiment. The model is tested using data from smog chamber experiments of ammonium sulfate aerosol and limonene oxidation. We show that the yield of secondary organic aerosol (SOA) calculated from the limonene oxidation may vary by over 20

The species and chemistry responsible for secondary organic aerosol (SOA) formation remain highly uncertain. Laboratory studies of individual precursors do not reproduce ambient organic material either quantitatively or qualitatively, and field campaigns suffer from uncertain contributions from primary material, making unambiguous identification of SOA difficult. By tracking the chemical evolution of oxidizing diesel exhaust using an aerosol mass spectrometer, we have addressed both of these issues: the SOA formed during our experiments closely resembles ambient oxidized organic particulate matter, and we have explicit knowledge of the condensed-phase mass spectrum (MS) of the primary emissions from our engine. This knowledge allows us to decompose each observed MS into contributing primary and secondary spectra throughout the experiment. The secondary MS we observed becomes increasingly oxidized as a function of time. This observation supports our hypothesis that SOA results from atmospheric oxidation of a large suite of precursors of varying vapor pressures. The lowest vapor pressure fraction of emissions requires the fewest generations of oxidation to form condensable products, and therefore, it forms relatively less oxidized SOA very quickly.

A three-dimensional chemical transport model (PMCAMx) is used to investigate changes in fine particle (PM2.5) concentrations in response to changes in SO2 and NH3 emissions during July 2001 and January 2002 in the Eastern United States. A uniform 50concentrations by 26% during July but only 6% during January. A 50% reduction of NH3 emissions leads to an average 4% and 9respectively. During the summer, the highest concentration of sulfate is in south Indiana (12.8 g m-3) and the 50sulfate decrease over this area. During winter, the SO2 emissions reduction results in a 1.5 g m-3 (29%) decrease of the peak sulfate levels (5.2 g m-3) over southeast Georgia. The maximum nitrate and ammonium concentrations are predicted to be over the Midwest (1.9 g m-3 in Ohio and 5.3 g m-3 in south Indiana respectively) in the summer while in the winter these concentrations are higher over the Northeast (3 g m-3 of nitrate in Connecticut and 2.7 g m-3 of ammonium in New York).The 50% NH3 emissions reduction is more effective for controlling nitrate, compared to SO2 reductions, resulting in a 1.1 g m-3 nitrate decrease over Ohio in July and a 1.2 g m-3 decrease over Connecticut in January. Ammonium decreases significantly when either SO2 or NH3 emissions are decreased. However the SO2 control strategy has better results in July when ammonium decreases, up to 2 g m-3 (37%), are predicted in south Indiana. The NH3 control strategy has better results in January (ammonium decreases up to 0.4 g m-3 in New York).The spatial and temporal characteristics of the effectiveness of these emission control strategies during the summer and winter seasons are discussed.

The goal of this modeling study is to determine how concentrations of ozone respond to changes in climate over the eastern USA. The sensitivities of average ozone concentrations to temperature, wind speed, absolute humidity, mixing height, cloud liquid water content and optical depth, cloudy area, precipitation rate, and precipitating area extent are investigated individually. The simulation period consists of July 12-21, 2001, during which an ozone episode occurred over the Southeast. The ozone metrics used include daily maximum 8 h average O3 concentration and number of grid cells exceeding the US EPA ambient air-quality standard. The meteorological factor that had the largest impact on both ozone metrics was temperature, which increased daily maximum 8 h average O3 by 0.34 ppb K1 on average over the simulation domain. Absolute humidity had a smaller but appreciable effect on daily maximum 8 h average O3 (0.025 ppb for each percent increase in absolute humidity). While domain-average responses to changes in wind speed, mixing height, cloud liquid water content, and optical depth were rather small, these factors did have appreciable local effects in many areas. Temperature also had the largest effect on air-quality standard exceedances; a 2.5 K temperature increase led to a 30increase in the area exceeding the EPA standard. Wind speed and mixing height also had appreciable effects on ozone air- quality standard exceedances.

Existing parameterizations tend to underpredict the a-pinene aerosol mass fraction (AMF) by a factor of 2-5 at low organic aerosol concentrations (<5 ugm-3). A wide range of smog chamber results obtained at various conditions (low/high NOx, pres- ence/absence of UV radiation, dry/humid conditions, and temperatures ranging from 15-40 C) collected by various research teams during the last decade are used to derive new parameterizations of the SOA formation from -pinene ozonolysis. Parameterizations are developed by fitting experimental data to a basis set of saturation concentrations (from 10-2 to 10(4) ugm-3) using an absorptive equilibrium partitioning model. Separate parameterizations for a-pinene SOA mass fractions are developed for: 1) Low NOx, dark, and dry conditions, 2) Low NOx, UV, and dry conditions, 3) Low NOx, dark, and high RH conditions, 4) High NOx, dark, and dry conditions, 5) High NOx, UV, and dry conditions. According to the proposed parameterizations the a-pinene SOA mass fractions in an atmosphere with 5 ugm-3 of organic aerosol range from 0.032 to 0.1 for reacted a-pinene concentrations in the 1 ppt to 5 ppb range.

An algorithm for the calculation of organic aerosol density in mixed organic- inorganic particles combining measurements by the Aerodyne Aerosol Mass Spectrometer (AMS) and the Scanning Mobility Particle Sizer (SMPS) was developed. The approach is applicable to particles with size-dependent composition. The estimated density of secondary organic aerosol (SOA) formed by a-pinene, b-pinene and d- limonene ozonolysis was in the range of 1.4-1.65 g cm-3. However, in two cases the SOA had much lower density (0.9-1.0 g cm-3) indicating that there may be changes in particle morphology depending on the conditions of SOA formation. The high estimated density for these systems suggests that SOA particles may be solid or waxy. Based on our results, SOA yields in smog chamber experiments may be a lot higher (up to 50%) than the currently assumed values. Most of the literature results have been calculated by measuring the SOA number distribution with an SMPS and then multiplying the volume concentration with a density equal to 1 or 1.2 g cm-3.

A three-dimensional chemical transport model (PMCAMx) is used to simulate particulate matter (PM) mass and composition in the eastern United States during the four seasons of the year (July 2001, October 2001, January 2002, and April 2002). The model predictions are evaluated against daily average PM2.5 measurements taken throughout the eastern United States by the IMPROVE and STN monitoring networks and the EPA Supersites program.

During the spring and summer the model reproduces the measured daily average PM2.5 concentrations with an error of less than 50The PM2.5 error is less than 30% for 43% of the measurements during these seasons. For the fall and winter the PM2.5 predictions are within 50% of the measurements for 51% of the data points and within 30% for 34% of the time. The performance of the model in reproducing sulfate, organic mass, elemental carbon, and total PM2.5 concentrations varies from average to good depending on the season. Uncertainties in ammonia emissions during the fall cause errors in the corresponding ammonium predictions, while the ammonia emissions inventory appears to be satisfactory during the other seasons. The ability of the model to reproduce the aerosol nitrate concentrations in the spring and summer is limited by difficulties in simulating the heterogeneous nightime formation rate of nitric acid. During the summer and fall the model performance in reproducing the organic PM concentrations and diurnal patterns is good. The predicted organic PM during the summer is on average 60% primary and 40% secondary. The secondary contribution to organic PM drops to around 20% during the winter. The used average EC emission rate of approximately 0.55 ktons d-1 (0.45 ktons d-1 during the weekends) is consistent with the observed EC concentrations. The nightime chemistry of NOx determines the PM nitrate concentrations during most days in the winter, spring, and fall and its mathematical description needs improvement. The good agreement between the predicted and observed temporal profiles for most species suggests a reasonable understanding and depiction in the model of the corresponding processes. Additional strengths and limitations of current modelling approaches for this modelling domain and for the different seasons are further discussed.

Ammonium sulfate particles were generated by atomization and introduced into a smog chamber where they were coated with secondary organic aerosol from ozonolysis of limonene or -pinene. These mixed particles were then sampled with a humidified Tandem-DMA system where a monodisperse aerosol population was selected, humidified, and dried to observe the relative humidity (RH) at which the particles returned to the original dry diameter. The volume fraction of secondary organic aerosol (SOA) in the mixed particles ranged from 0.59 to 0.94 for limonene SOA and 0.54 to 0.63 for a-pinene SOA. Efflorescence RHs for our mixed aerosols were in the range of 28-34%, similar to our observation of 32% ERH for pure ammonium sulfate nanoparticles. These findings indicate that the effect of SOA on the ERH of inorganic salts in the atmosphere may be negligible.

Positive Matrix Factorization (PMF) was used to analyze a large dataset of organic molecular markers to investigate the sources of organic carbon (OC) in Pittsburgh, Pennsylvania. The source-specific nature of molecular markers greatly aids the selection of the number of factors and rotations, the association of factors with specific sources, and the identification of solutions that mix emissions from a specific source across multiple factors. PMF was performed using different combinations of species to investigate the stability of the results. The exact number of factors depends on the specific combination of input species; however, all solutions had the same core set of 7 factors. Six of these factors appear related to primary emissions and one to secondary organic aerosol. The amount of OC associated with these 7 core factors was, for the most part, well constrained across 21 different PMF solutions. However comparisons with source profiles and previously published chemical mass balance (CMB) analysis indicate that PMF does not cleanly differentiate between different sources. On a seasonal average basis there is better agreement between the PMF and CMB results in the summer than the winter. Both approaches imply that secondary organic aerosol is dominant in the summer, contributing between 60% and 70% of the OC. Both PMF and CMB analysis of molecular markers data indicate that motor vehicles are a relatively minor source, contributing only 10% of the annual- average OC.

The reversibility of gas-particle partitioning of α-pinene secondary organic aerosol (SOA) was investigated via experiments in a laboratory reaction chamber. Previously, phase partitioning has been studied quantitatively via SOA production experiments or qualitatively by perturbing temperature and observing particle evaporation. In this work, two methods were used to dilute chamber SOA: an external dilution sampler and in-chamber dilution. Repartitioning after dilution took place on the time scale of tens of minutes to hours - this behavior is consistent with uptake coefficients on the order of 0.001 - 0.01. However, given sufficient time, α-pinene SOA does equilibrate reversibly, consistent with results of conventional yield experiments. Further, data from an aerosol mass spectrometer (AMS) indicate that the composition of SOA varies with partitioning. These results suggest that the oligomerization observed in laboratory SOA may be a reversible process and that the formation and evaporation of SOA is more complex than previously imagined.

This paper presents chemical mass balance analysis of organic molecular markers data to investigate the sources of organic aerosol and PM2.5 in Pittsburgh, Pennsylvania. The model accounts for emissions from eight primary source classes, including major anthropogenic sources such as motor vehicles, cooking, and biomass combustion as well as some primary biogenic emissions (leaf abrasion products). We consider uncertainty associated with selection of source profiles and fitting species, sampling artifacts, photochemical aging, and unknown sources. In the context of the overall organic carbon (OC) mass balance, the contributions of diesel, wood- smoke, debris, and coke-oven emissions are all small and well-constrained; however, estimates for the contributions of gasoline-vehicle and cooking emissions can vary by an order o magnitude. A best-estimate solution is presented that represents the vast majority of our CMB results; it indicates that primary OC only contributes 27+/-8% and 50+/-14% (average +/- standard deviation of daily estimates) of the ambient OC in the summer and winter respectively. Approximately two-thirds of the primary OC is transported into Pittsburgh as part of the regional air mass. The ambient OC that is not apportioned by the CMB model is well correlated with secondary organic aerosol (SOA) estimates based on the EC-tracer method as well as ambient concentrations of organic species associated with SOA. Therefore, SOA appears to the major component of OC, not only in the summer, but potentially in all seasons. Primary OC does dominate the OC mass balance on a small number of non-summer days with high OC concentrations; these events are associated with specific metrological conditions such as local inversions. Primary particulate emissions only contribute a small fraction of the ambient fine- particle mass, especially in the summer.

The SOA formation from β-pinene ozonolysis at modest precursor concentrations (2 ppb to 40 ppb) was investigated in the temperature range of 0^oC to 40^oC. The presence of inert seeds and high ozone concentrations is necessary to minimize losses of semi- volatile vapors to the walls of the smog chamber. -pinene secondary organic aerosol production increases significantly with decreasing temperature. An increase by a factor of 2-3, depending on the reacted -pinene concentration, was observed as the temperature decreased from 40^oC to 0^oC. This increase appears to be due mainly to the shifting of partitioning of the semivolatile SOA components towards the particulate phase and not to a change of the -pinene product distribution with temperature. The measurements are used to develop a new temperature-dependent parameterization for the four-component basis-set. The parameterization predicts much higher β-pinene ozonolysis production for typical atmospheric conditions than the values that have been suggested by previous studies.

Diluted exhaust from a diesel engine is exposed to UV light in a smog chamber to investigate secondary organic aerosol (SOA) production. Photochemical aging rapidly produces significant SOA, doubling the organic aerosol contribution of primary emissions after only several hours of processing. Only 10oxidation of known precursors, such as light aromatics. The smog chamber walls complicate determination of the total SOA production because of uncertainty in the potential losses of semivolatile vapors. Two limiting cases are explored: in one, the mass lost to the walls is assumed to be in full equilibrium with the suspended mass; in the other, the mass on the walls is assumed to be inactive in partitioning. The available evidence favors the first case, leading to higher estimates of SOA production. Aerosol Mass Spectrometer (AMS) mass spectra reveal that the organic aerosol becomes progressively more oxidized with exposure to UV light. The experiments provide strong evidence that oxidation of a wide array of precursors contribute to ambient SOA formation.