Water activity and activation diameters from hygroscopicity data - Part II: Application to organic species

Abstract. A method has been developed for using particle hygroscopicity measurements made with a humidified tandem differential mobility analyzer (HTDMA) to determine water activity as a function of solute weight percent. In Part I, the method was tested for particles composed of sodium chloride and ammonium sulfate. Here, we report results for several atmospherically-relevant organic species: glutaric acid, malonic acid, oxalic acid and levoglucosan. Predicted water activities for aqueous dicarboxylic acid solutions are quite similar in some cases to published estimates and the simplified predictions of Köhler theory, while in other cases substantial differences are found, which we attribute primarily to the semivolatile nature of these compounds that makes them difficult to study with the HTDMA. In contrast, estimates of water activity for levoglucosan solutions compare very well with recently-reported measurements and with published data for aqueous glucose and fructose solutions. For all studied species, the critical dry diameters active at supersaturations between 0.2 and 1% that are computed with the HTDMA-derived water activities are generally within the experimental error (~20%) estimated in previously-published direct measurements using cloud condensation nuclei counters. For individual compounds, the variations in reported solution water activity lead to uncertainties in critical dry diameters of 5-25%, not significantly larger than the uncertainty in the direct measurements. To explore the impact of these uncertainties on modeled aerosol-cloud interactions, we incorporate the variations in estimates of solution water activities into the description of hygroscopic growth of aerosol particles in an adiabatic parcel model and examine the impact on the predicted drop number concentrations. For the limited set of initial conditions examined here, we find that the uncertainties in critical dry diameters for individual species lead to 0-21% changes in drop number concentration, with the largest effects at high aerosol number concentrations and slow updraft velocities. Ammonium sulfate, malonic acid and glutaric acid have similar activation behavior, while glutaric acid and levoglucosan are somewhat less hygroscopic and lead to lower drop number concentrations; sodium chloride is the most easily activated compound. We explain these behaviors in terms of a parameter that represents compound hygroscopicity, and conclude that this parameter must vary by more than a factor of 2 to induce more than a 15% change in activated drop number concentrations. In agreement with earlier studies, our results suggest that the number concentration of activated drops is more sensitive to changes in the input aerosol size and number concentrations and the applied updraft velocity than to modest changes in the aerosol composition and hygroscopic properties.