The atmospheric aqueous-phase chemistry has received increasing attention in the last decades for its non-negligible environmental significance. Yet, the insufficient experimental data on oxidative reaction rate constants (kaq) obstructs the further analysis and modeling of this system. Predictive models based on machine learning (ML) algorithms have shown potential as an effective estimation tool, however, they are restricted to the lack of training data as well. To overcome this data limitation, we developed multi-task (MT) models that could exploit the common knowledge from reactions in gas- and aqueous-phases simultaneously. Toward kaq of organic compounds with hydroxyl radical (OH), nitrate radical (NO3), and ozone (O3), the MT models showed a notably better predictive ability compared to benchmark models, while obtaining wide applicability on compounds from different chemical classes. By interpreting the models using Shapley additive explanations (SHAP), we evidenced that the MT models utilized the common knowledge in both phases and correctly identified the reaction mechanisms. This study aims to provide new insight into the estimation of necessary kinetic parameters in atmospheric aqueous-phase chemistry, as well as a reference to ML research for other predictive tasks of atmospheric interest.

Isoprene hydroxy peroxy radicals (ISOPOO), derived from isoprene oxidation by hydroxy radicals (OH), are key intermediates for ozone and secondary organic aerosol (SOA) formation in the atmosphere. Although ISOPOO-water complexes are ubiquitous, their impacts on ISOPOO chemistry remain obscure. Here the previously overlooked water effect on the bimolecular reaction kinetics of ISOPOO was investigated in an oxidative flow reactor. The major first-generation products of ISOPOO, isoprene hydroxy hydroperoxides (ISOPOOH), methacrolein (MACR), and methyl vinyl ketone (MVK), were measured simultaneously at various relative humidity (RH) with the help of a cold trap to avoid potential losses in direct gas sampling. We found that ISOPOO reactions were accelerated significantly under wet conditions, with a greater enhancement on 1,2-ISOPOO than 4,3-ISOPOO. 1,2-ISOPOOH yield appeared faster growth with RH than 4,3-ISOPOOH. MVK yield showed an upward-downward trend with RH, while MACR yield plateaued from 30% RH. To explain the enhancement in the ISOPOOH yield from 3% to 80% RH, the overall rate constants of 1,2-ISOPOO + HO2 and 4,3-ISOPOO + HO2 reactions at 80% RH should be 13 times and twice those at 3% RH, respectively. The empirical formulas were proposed for the first time to parameterize the water effect on ISOPOO + HO2 reactions. The updated kinetics of ISOPOO reactions were incorporated in a box model to simulate the RH-dependent ISOPOOH and C4 carbonyl yields under typical atmospheric conditions. High RH can enhance the ISOPOOH yield in urban, rural, and forest areas, and promote SOA formation correspondingly. Our findings shed light on the critical role of humidity in the reactions of ISOPOO and benefit evaluating the fate of isoprene and its impacts on air quality more accurately in the ambient atmosphere.

Atmospheric organic peroxides (POs) play a key role in the formation of O3 and secondary organic aerosol (SOA), impacting both air quality and human health. However, there still remain technical challenges in investigating the reactivity of POs in ambient aerosols due to the instability and lack of standards for POs, impeding accurate evaluation of their environmental impacts. In the present study, we conducted the first attempt to categorize and quantify POs in ambient PM2.5 through hydrolysis, which is an important transformation pathway for POs, thus revealing the reactivities of various POs. POs were generally categorized into hydrolyzable POs (HPO) and unhydrolyzable POs (UPO). HPO were further categorized into three groups: short-lifetime HPO (S-HPO), intermediate-lifetime HPO (I-HPO), and long-lifetime HPO (L-HPO). S-HPO and L-HPO are typically formed from Criegee intermediate (CI) and RO2 radical reactions, respectively. Results show that L-HPO are the most abundant HPO, indicating the dominant role of RO2 pathway in HPO formation. Despite their lower concentration compared to L-HPO, S-HPO make a major contribution to the HPO hydrolysis rate due to their faster rate constants. The hydrolysis of PM2.5 POs accounts for 19% of the nighttime gas-phase H2O2 growth during the summer observation, constituting a noteworthy source of gas-phase H2O2 and contributing to the atmospheric oxidation capacity. Seasonal and weather conditions significantly impact the composition of POs, with HPO concentrations in summer being significantly higher than those in winter and elevated under rainy and nighttime conditions. POs are mainly composed of HPO in summer, while in winter, POs are dominated by UPO.