Repository logo
 

Formation and evolution of secondary organic aerosol in laboratory experiments: precursors, processes, and properties

Date

2022

Authors

He, Yicong, author
Jathar, Shantanu H., advisor
Pierce, Jeffrey R., committee member
Bond, Tami C., committee member
Volckens, John, committee member

Journal Title

Journal ISSN

Volume Title

Abstract

Secondary organic aerosol (SOA) is an important fraction of atmospheric PM2.5 which is defined as fine-mode aerosols with diameters less than 2.5 μm. SOA is ubiquitous in the atmosphere and can have considerable impacts on the climate, air quality and human health. We are limited in our ability to predict the spatial and temporal distribution of SOA and assess its environmental impacts, because current three-dimensional chemical transport models (CTM) still have large biases and relatively weak correlations with observations of SOA. One reason for the model-observation discrepancy could be that we still lack a full understanding of the precursors, chemical/physical processes, and properties of SOA that govern its formation and evolution. Therefore, there is a need to further study the precursors, processes, and properties of SOA in laboratory experiments, and to develop more accurate SOA parameterizations that can be used to update the current CTMs. In Chapter 2, I studied SOA formation from several novel precursors which were vapors from biofuels that were under development at the National Renewable Energy Laboratory (NREL) to be used as future blendstocks to gasoline, and I developed SOA parameterizations for these biofuel precursors that corrected for the influence of vapor wall loss, using a kinetic SOA model called SOM-TOMAS (Statistical Oxidation Model coupled with TwO-Moment Aerosol Sectional Model). Although vapor wall loss has been shown to significantly impact SOA formation in environmental chamber experiments, it has rarely been corrected for in the development of SOA parameters used in atmospheric models. Our parameterizations predicted that under atmospherically relevant conditions, some of the biofuels may produce similar or even more SOA than gasoline, possibly offsetting the environmental benefits they offered. In addition, the parameterizations predicted that correcting for vapor wall loss in chambers always resulted in similar or increased atmospheric SOA mass yields compared to chamber yields, highlighting the potential for vapor wall loss correction to increase SOA predictions from CTMs and to bridge the gap with observations. In Chapter 3, I demonstrated a novel technique to constrain the SOA particle bulk diffusivity (Db) in chamber experiments, using a kinetic model (i.e., SOM-TOMAS) and measurements of the particle size distribution. Db is a property that controls the gas/particle partitioning timescale of SOA, where a higher Db (i.e., liquid aerosol) means faster partitioning and a lower Db (i.e., semi-solid aerosol) means slower partitioning. Here, I showed that the measured particle size distribution in SOA formation experiments contained sufficient information to constrain Db without direct measurement of the particle phase state or viscosity. In Chapter 4, I investigated the differences in the SOA mass yields measured in environmental chambers and oxidation flow reactors. Both chambers and flow reactors can simulate the photooxidation of Volatile Organic Compounds (VOCs), but flow reactors can achieve higher aging time (>2 weeks) than chambers (<1 day) by using very high oxidant concentrations. Their photooxidation chemistry pathways have been thought to be similar, but they produce different SOA mass yields at similar photochemical ages, which remains an unsolved problem. Here, I integrally simulated vapor and particle wall loss, semi-solid phase state, heterogeneous oxidation, particle-phase oligomerization, and new particle formation in chambers and flow reactors with experimentally constrained parameters for these processes. I showed that the SOA mass yield difference could be explained by the different contribution of these processes to SOA formation and evolution in chambers and flow reactors. Furthermore, with a single set of SOA parameterizations for photooxidation, the model was able to simultaneously predict the SOA mass concentration, bulk chemical composition (O:C ratio), and size distribution in chambers and flow reactors. The results highlight that flow reactor data can be modeled consistently with chamber data, and they should be used in synergy with chamber data to develop SOA parameterizations applicable to long photochemical aging times. In Chapter 5, in collaboration with Dr. Kelsey Bilsback, we investigated a widely employed assumption for particle wall loss correction in chamber experiments, regarding the interaction between wall-deposited particles and suspended vapors. Furthermore, as a continuation of the work from Chapter 2, we developed SOA parameterizations that corrected for both vapor and particle wall loss, and integrated these updated parameterizations into a CTM to assess the impacts on atmospheric SOA predictions. Specifically, we first showed that the interaction between vapors and wall-deposited particles was negligible through kinetic modeling, and accurate particle wall loss correction should assume no interaction between the two. We then found that the wall-loss-corrected SOA parameterizations greatly enhanced SOA formation in the CTM, reducing the gap with the observations. We argue that vapor and particle wall loss should be routinely accounted for in developing SOA parameterization.

Description

Rights Access

Subject

Citation

Associated Publications