Driving factors for the activity coefficient of atmospheric ammonium nitrate: discrepancies among thermodynamic models and impact on nitrate pollutions

Semi-volatile NH 4 NO 3 is a major component of atmospheric aerosols, and its environmental and climate effects are largely regulated by the gas-particle partitioning. The activity coefficient of NH 4 NO 3 , γ AN , is one key parameter controlling the gas-particle partitioning of nitrate, with lower γ AN typically favouring particle-phase partitioning of nitrate. However, the γ AN dependence on meteorological condition and chemical profile remains uncertain. Here we investigated into this issue with comprehensive simulations and ambient observations, based on results of three widely-used thermodynamic models, i.e. ISORROPIA, E-AIM, and AIOMFAC. Correspondingly, AIOMFAC estimate higher particle phase nitrate fp,NO3- values. Across all models, γAN2 ranges between 10 −2 and 10 −1 , with AIOMFAC results ∼ 33 % lower than E-AIM and ISORROPIA. Correspondingly, AIOMFAC estimate higher particle phase nitrate fp,NO3- values. For all three models and all chemical profiles tested, the γAN2 correlates positively with relative humidity (RH) and temperature, and RH generally contributes larger variations. In comparison, the effect of chemical composition on γAN2 is more complex and is strongly modulated by RH, with differed dependence pattern observed at varying RH levels. Furthermore, γAN2 responds more strongly to changes of particle chemical profile in E-AIM, whereas in ISORROPIA and AIOMFAC γAN2 is more sensitive to meteorological variations. As E-AIM is typically considered as the benchmark thermodynamic model, these results suggest the potential under-representation of chemical profiles in predicting γAN2 for ISORROPIA and AIOMFAC. The corresponding influence on 3-D chemical-transport model predictions of NH 4 NO 3 are encouraged in future studies.