This manuscript systematically investigated the water uptake potentials of a large library of the inorganic salts of amines with various functional groups (alkylamines, guanidinos, and amino acids), by computing the water activities of their aqueous solutions, as well as ternary mixtures of the salts, water and atmospheric-relevant organic proxies (either water soluble or water in-soluble). The author employed the Conductor-like Screening Model for Real Solvents (COSMO-RS) program with three parameterization methods (TZVP, FINE and ELYTE), depending on the specific cation in question. Based on the computational results, the water uptake potentials of amine salts depend more strongly on their cation chemical structures (eg, more substituents and more amino groups will enhance the water uptake of amine sulfates); but the anions may also affect such difference (eg, all bisulfate salts showed little difference in their aqueous solutions). Using an isoprene derived epoxydiol IEPOX as a proxy for water-soluble organics, the author showed that water soluble organics may enhance the water uptake of amine salts significantly, depending on the chemical structure of the salt. On the contrary, water in-soluble organics may simply become phase separated from the aqueous solutions to exert appreciable effects on the hygroscopicity of the salts.

While the manuscript has presented a thorough and comprehensive study on atmospheric-relevant amine salts in settings related to atmospheric chemical processes, I noticed a few items that may clarify some questions and further improve the structure and presentation of the manuscript:

• Introduction. What are the (semi-)quantitative differences in water activity estimations using COSMOtherm program when compared with other methods (eg, the group attribution methods)? How about the differences in computational time, resource requirements, etc?
• Methodology. Clearly it was difficult to find an “one-size-fit-all” computational method to address the complex research questions discussed here, and the manuscript states that FINE parameterization method was chosen for the amine salts and ELYTE for ammonium salts. However, when examining the Figs S1&S2, one could argue that the FINE may also be suitable for smaller amine salts. How was the “proper” parameterization method chose here? I think a more convincing approach is to calculate representative small amines using both FINE and ELYTE to see if the results are converging.
• Results&Discussion. The point above also brings up another major question: what is the estimated uncertainty in the computational results presented in the manuscript? Some of the traces in the figures were quite close to each other and the lack of error bars was making it difficult to judge if the difference was within the uncertainties of the methods.
• Results&Discussion. This study was very comprehensive with a lot of results, which leads to, in my opinion, very dense figures that are difficult to read and comprehend. I would like to recommend major reorganizations of the results for better presentations and support for the flow of the discussion. For example, some panels of Fig. 1 can go into supplementary, but others can be split into less busy ones (eg, Fig.1a may be divided into three panels, one for each amine group to highlight the differences).

Reviewer comments as a pdf file in the supplement