Phase state of secondary organic aerosol in chamber photo-oxidation of mixed precursors

The phase behaviour of aerosol particles plays a profound role in atmospheric physicochemical processes, influencing their physical and optical properties and further impacting climate and air quality. However, understanding of the aerosol phase state is still incomplete, especially that of multicomponent particles which contain inorganic compounds and secondary organic aerosol (SOA) from mixed volatile organic compound (VOC) precursors. We report measurements conducted in the Manchester Aerosol Chamber (MAC) to investigate the aerosol rebounding tendency, measured as the bounce fraction, as a surrogate of the aerosol phase state during SOA formation from photo-oxidation of biogenic (alpha-pinene and isoprene) and anthropogenic (o-cresol) VOCs and their binary mixtures on deliquescent ammonium sulfate seed. Aerosol phase state is dependent on relative humidity (RH) and chemical composition (key factors determining aerosol water uptake). Liquid (bounce fraction; BF < 0.2) at RH > 80% and nonliquid behaviour (BF > 0.8) at RH < 30% were observed, with a liquid-to-nonliquid transition with decreasing RH between 30% and 80 %. This RH-dependent phase behaviour (RHBF=0.2, 0.5, 0.8) increased towards a maximum, with an increasing organic-inorganic mass ratio (MRorg/inorg) during SOA formation evolution in all investigated VOC systems. With the use of comparable initial ammonium sulfate seed concentration, the SOA production rate of the VOC systems determines the MRorg/inorg and, consequently, the change in the phase behaviour. Although less important than RH and MRorg/inorg, the SOA composition plays a second-order role, with differences in the liquid-to-nonliquid transition at moderate MRorg/ inorg of similar to 1 observed between biogenic-only (anthropogenic-free) and anthropogenic-containing VOC systems. Considering the combining role of the RH and chemical composition in aerosol phase state, the BF decreased monotonically with increasing hygroscopic growth factor (GF), and the BF was similar to 0 when GF was larger than 1.15. The real atmospheric consequences of our results are that any processes changing ambient RH or MRorg/inorg (aerosol liquid water) will influence their phase state. Where abundant anthropogenic VOCs contribute to SOA, compositional changes in SOA may influence phase behaviour at moderate organic mass fraction (similar to 50 %) compared with purely biogenic SOA. Further studies are needed on more complex and realistic atmospheric mixtures.