Photochemical Aging Alters Secondary Organic Aerosol Partitioning Behavior

Organic aerosol is the largest fraction of the atmospheric non refractory aerosol mass and has a significant impact on climate, visibility, and human health. A significant portion of organic aerosol is secondary organic aerosol (SOA), which forms when volatile organic compounds (VOCs) are oxidized and the reaction products partition to the particle phase. SOA formation and aging have been described using semivolatile partitioning theory, and a typical assumption invoked is that gas-particle partitioning rapidly reaches equilibrium. However, several studies have called into question whether traditional equilibrium assumptions are valid, either due to particle-phase diffusion limitations, particle-phase reactions, or liquid/liquid phase separation and immiscibility of organic phases. Here, we investigate the ability of equilibrium partitioning to describe laboratory SOA formation experiments using SOA mass and yield as the evaluation metric. We conduct two types of experiments: (1) co-condensation experiments in which isoprene and a-pinene are simultaneously oxidized to form SOA and (2) sequential condensation experiments in which fresh isoprene SOA is formed in the presence of aged, pre-existing isoprene- or alpha-pinene-derived SOA particles. In the co-condensation experiments, equilibrium partitioning successfully predicted the time-dependent SOA concentrations, suggesting that the SOA from both precursors rapidly formed a well-mixed phase. However, in the sequential condensation experiments, equilibrium partitioning assumptions significantly overpredicted the observed SOA yield, indicating that freshly formed isoprene SOA did not rapidly partition into either the aged a-pinene or aged isoprene SOA particles to form a well-mixed phase over the 4 h experimental time scale, even at relative humidity as high as 85%. This study shows that an equilibrium partitioning assumption is accurate for freshly formed SOA but that it breaks down after SOA has been photochemically aged for modest amounts of time (15-18 h). These results have important implications for modeling SOA formation and may help to resolve some seemingly divergent conclusions regarding diffusion limitations that exist in the literature.