S1.0 KINETIC MECHANISM DEVELOPMENT.A kinetic mechanism is formulated to simulate the reaction conditions of these experiments.The reactions included are listed in the Appendix (Table SA2, SA3, and SA4). Table SA5 contains a list of the abbreviations used. Rate constants for most of the reactions included in the mechanism are based on recommendations from JPL 1 , IUPAC 2-3 , or MCM v3.2 4 . However, some rate constants and branching ratios are not known. For these, we use our best judgement based on available data; explanations of the assumptions on which these estimates are based are included in this section. Some branching ratios and rate constants are estimated based on the experimental results presented here. Many of these branching ratios depend on the fraction of δ-and β-isomers that form (Table 3 and 5), which will likely depend on the lifetime of the RO 2 radical (Section 4.1). Thus, the reaction products and rates presented here are most consistent with the experimental results for this study in which the overall RO 2 lifetime was ~ 30 s. The kinetic mechanism developed here represents our current level of understanding, and deviations from the experimental results highlight areas for future study.
S1.1. Basic Reactions in Kinetic Mechanism. HO 2 was constrained in the kinetic mechanism by the measured H 2 O 2 production rate. Prior to photooxidation, H 2 O 2 is predominantly formed from HO 2 + HO 2 reactions. To match the observed H 2 O 2 production rate in experiments 5, 6, and 8, we arbitrarily increased the reaction rate constant for CH 2 O + NO 3 by a factor of 2.5-3 in the kinetic mechanism above that recommended by IUPAC. Although not perfect when correcting for the missing HO 2 in this manner, the H 2 O 2 curves for the kinetic mechanism and the experimental results were fairly consistent. Under-prediction of HO 2 could be caused by other S3 missing chemistry including unaccounted for surface chemistry, later generation chemistry not incorporated into the kinetic mechanism, or many other possibilities. Here, we are confident that there is a missing source of HO 2 , but are agnostic about the mechanism responsible.Because the predominant loss of isoprene is due to reaction with NO 3 , the measured isoprene decay rate was used to constrain the amount of NO 3 present. Cantrell et al. 5 proposed that N 2 O 5would react with water present on the wall surface to form nitric acid even under dry conditions.We included a wall loss rate for N 2 O 5 (i.e., NO 3 loss rate) such that the isoprene decay in the kinetic mechanism matched with experimental results. This rate constant is chamber/experiment specific. For experiment 5 (24 m 3 , 2.2 ppm CH 2 O), 6 (24 m 3 , 4.7 ppm CH 2 O), 7 (1 m 3 , 2 ppm CH 2 O) and 8 (1 m 3 , 4 ppm CH 2 O), N 2 O 5 wall loss rate constants that best fit experimental conditions were 1.5 x 10 -4 , 12 x 10 -4 , 6 x 10 -4 and 12 x 10 -4 s -1 , respectively. We observe that the N 2 O 5 loss rate appears to be sensitive to both the mixing ratio of CH 2 O and the chamber.However, it s...