Assessment of Hygroscopicity Uncertainties Associated With Size and Thermodynamic Model: Implications for Inferring Chemical Composition of Sub-10 nm Particles

Chemistry information of newly formed particles is essential for understanding nucleation mechanisms. The hygroscopicity parameter κ is widely employed in cloud condensation nuclei size range to indirectly infer organic mass fraction (f_org) based on the κ-f_org linear relationship, but its associated uncertainties in such applications, especially for sub-10 nm particles have not been fully constrained. Here, we first discussed the uncertainties of κ values by revisiting closure studies of κ derived from chemical composition (κ_chem), hygroscopic growth factor (κ_gf), and effective supersaturation (κ_Se). We then conducted a modeling and experiment study of exemplary substances to discuss the uncertainties from the κ dependence on particle size and thermodynamic models. The model-predicted κ of four representative substances demonstrated significant size-dependence and were compared to measured values in 3–8 nm. For ammonium sulfate (AS), the deviations of κ_Se retrieved by commonly used models varied from 2.5% for 3 nm and 41.2% for 300 nm, demonstrating significant size-dependence of model prediction deviations, which may potentially introduce significant uncertainties in determining chemistry. We further evaluated mixtures of AS and cis-pinonic acid. The relative differences in f_org for AS-organic mixtures reached 20.7% (model effect) and 41.4% (size effect). The estimated κ of pure organic particle exhibited relative variations of 22.7% (model effect) and 83.3% (size effect). Our results highlight the importance of consistent thermodynamic parameterizations and size range in inferring chemical composition through hygroscopicity, and help to interpret κ values determined at different water vapor ratios and at different size ranges (especially sub-10 nm range).