Modeling secondary organic aerosol formation : from atmospheric simulation chambers to the field
Loading...
Date
2025-09-12
Authors
Ουρούτσης, Πέτρος
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Particulate matter (PM) refers to the suspended solid or liquid particles in the atmosphere. These particles can pose significant human health risks and influence the Earth's energy balance, thereby contributing to climate change. A major component of PM is organic aerosol (OA), which includes both primary organic aerosol (POA) emitted directly into the atmosphere and secondary organic aerosol (SOA) formed through gas-phase oxidation of organic compounds. This work explores SOA formation through a combination of laboratory and field observations, supported by modeling studies.
This thesis begins by investigating the gas-phase chemistry of toluene, a common anthropogenic volatile organic compound (VOC), with a focus on its photooxidation. Our analysis draws on an extensive series of photooxidation experiments conducted under varying conditions in nine atmospheric simulation chambers across Europe, all part of the EUROCHAMP-2020 (EUROpean simulation CHambers for investigating AtMospheric Processes) infrastructure. To synthesize these results, we used a model incorporating the Master Chemical Mechanism (MCM), a near-explicit mechanism for VOC gas-phase oxidation, along with an SOA formation module. The combined data from these multi-chamber toluene experiments support our hypothesis that combining different facilities and experimental conditions provides a more comprehensive understanding of the system's behavior, as each chamber probes different regions of the chemical space. However, the synthesis also reveals significant gaps in our current understanding, especially during the later stages of toluene oxidation. The MCM-based model struggled to accurately capture the concentrations of both first-generation and oxygenated products, and it consistently underpredicted SOA mass across the experiments. These findings highlight critical gaps in the representation of SOA-forming pathways in the current version of the MCM, indicating that the uncertainties are greater when simulating SOA formation.
The oxidation of VOCs produces numerous organic compounds through complex reaction networks, making detailed analysis of atmospheric simulation experiments difficult and the incorporation of the corresponding results into chemical transport models (CTMs) challenging. To manage this complexity, the one-dimensional volatility basis set (1D-VBS) framework has been developed to simulate SOA formation in CTMs. Traditionally, SOA parameterizations in the 1D-VBS have relied on fitting SOA yield data from atmospheric simulation chamber experiments. To reduce the uncertainties of this approach, we developed an algorithm that estimates the volatility distribution and effective vaporization enthalpy of SOA products. This algorithm combines SOA yield data with results from two types of evaporation techniques, thermal and isothermal dilution. We evaluated the algorithm using pseudo-experimental data from SOA systems with known properties and found it accurately reproduced SOA yields under atmospherically relevant conditions.
The new algorithm developed in this work was used to parameterize the oxidation of various compounds, including the ozonolysis of α-humulene (a sesquiterpene) and the photooxidation of four cyclohexanes and three aromatic compounds. The resulting yields and parameterizations revealed that the current scheme in the CTM PMCAMx-SR underestimates the SOA formed during the ozonolysis of sesquiterpenes by a factor of two. Additionally, the results show that larger cyclohexanes produce higher SOA yields than aromatics, while aromatic compounds generate more highly oxidized SOA. The derived VBS parameterizations are recommended for future implementation in CTMs.
In the final part of this research, we use two CTMs, PMCAMx-SR and PMCAMx-Trj, based on the 1-D and 2-D VBS, respectively, to simulate OA levels and its degree of oxidation during a field campaign conducted in a forested background area in Greece in July 2022. The 2D-VBS extends the 1D-VBS by incorporating the oxygen-to-carbon (O:C) ratio, a key metric for assessing OA oxidation degree. Both models predicted a high influence of wildfires across Europe during this period, with PMCAMx-SR attributing nearly 40% of OA to biomass burning SOA (bbSOA) in the campaign site. However, PMCAMx-SR consistently underestimated total OA, especially biogenic SOA (bSOA). PMCAMx-Trj accurately predicted O:C for total OA (0.84 vs. 0.82 observed) and biogenic OA (0.59 vs. 0.63 observed), but it underpredicted OA concentrations, indicating probably the same model limitations as in PMCAMx-SR. Despite this agreement between the predicted and observed O:C for biogenic OA, the O:C distribution of first-generation SOA products used in PMCAMx-Trj remains uncertain. To better constrain these 2D-VBS parameterizations, further smog chamber experiments are needed. Sensitivity analyses indicated that the underestimation of OA levels by the PMCAMx-SR and PMCAMx-Trj models results from several factors. Doubling bVOC emissions improved model performance (FBIAS by 90%, FERROR by 30%), though emissions alone did not fully explain temporal discrepancies. While PMCAMx-SR predicted that 90% of biogenic SOA was transported, this likely reflects incomplete aging chemistry treatment. Incorporating aging for sesquiterpene-derived SOA significantly reduced model biases (up to 100% in FBIAS and 40% in FERROR), highlighting the importance of multi-generational oxidation processes. These findings underscore the need for better representation of SOA yields and aging rates, potentially through smog chamber constraints.
Description
Keywords
Secondary organic aerosol formation, Volatility basis set, Smog chambers, Chemical transport models