Summary: This is a valuable and comprehensive study that models the interactions between water and typical CCN species, that include ions and organics, using molecular dynamics simulation. The most serious reservation that I have concerns Sec. 2.3 and misuse of the terms equimolecular dividing surface (at Re) and surface tension (see below). Unfortunately, Re is used in the equations, beginning with Eq. 3, instead of the “surface of tension", which should be used. This is likely to affect the calculations that follow, especially if the interfacial structure is broad. The authors should have look at this and comment. Otherwise the paper seems important and should be published. Major points, minor points, and a few typos are listed below.

Major 

Section 2.2 System prep and MD simulations: I have some points of confusion after reading this section. The first concerns the underlying model consisting of  cubic cells with periodic boundary conditions and edge length exceeding the droplet diameter – a figure here would help the reader.  Second, what is the advantage of periodic boundaries, with so much extra space in each cell?  How is the Ewald sum applied in this model? Usually Ewald sums are applied to extended periodic structures - not to a period set of droplets with space around each one. More details here would be helpful. 

Having called attention to Ewald sums I might point out a clever test that evaluates the accuracy of intermolecular water potentials. This by comparing the computationally relaxed structures with the 3D structure parameters and densities available for ice structures from x-ray diffraction [Morse and Rice, 1981]. For what its worth, the ST2 water potential performed quite well in the test while another did poorly.  

Section 2.3. The author’s description of the Gibb’s dividing surface seems to this reviewer a misrepresentation of this important concept. Specifically, the authors use of surface tension at  the equimolar (equimolecular might be better in context of MD)  is said to “correspond to a vanishing adsorption … ensuring that the surface free energy per unit area so defined corresponds to the surface tension”. Actually the equimolecular surface does neither! It is the dividing surface located at the “surface of tension” that has these properties. As for adsorption, the Gibbs adsorption isotherm applies only at the surface of tension. Moreover, the pressure difference across the surface of tension is the only one that appears in the standard Laplace and Kelvin relations (otherwise additional terms added to these relations are required) . See [McGraw and Laaksonen, 1997] and especially the citation to Ono and Kondo, an excellent review of the subject, therein.   

Related: Eq. 7 is similar to the equation developed by Gibbs for the work to form a capillary drop from vapor. This formula can be applied even to droplets having a broadened interfacial region - provided the radius at the surface of tension is used.

Finally, a couple of comments on the  “validity of the Kelvin and Kohler theory at droplet sizes larger than 4nm under moderate salinities and organic loadings and the need to account for ion-concentration enhancement in sub-10nm particles” mentioned in the Abstract. This is an important theme that runs through and adds value to  the paper. With respect to the Kelvin relation this has been confirmed for the Kelvin (pure water) and Kelvin-Thomson (ionic solution) relations [Winkler et al., 2012]. For Kohler theory, on the other hand, this is unlikely to be the case for organics.  The latter tend to partition between the bulk and surface phases, whereas the standard Kohler and kappa-Kohler models  pertain only to fully water-soluble species. A recent extension of Kohler theory, based on analysis of droplet stability, takes into account the partitioning of both water-soluble and surface-active species in a unified way for applications to cloud activation [McGraw and Wang, 2021].  

McGraw, R. and A. Laaksonen (1997), J. Chem. Phys. 106, 5284-5287. 

Morse, M. D. and S. A. Rice (1981), J. Chem. Phys. 74, 6514-6516. 

Winkler, P. M., et. al. (2012), Phys. Rev. Letts. 108, 085701. 

McGraw, R. and J. Wang (2021), J. Chem. Phys. 154, 024707; doi: 10.1063/5.0031436 

Minor points and typos: 

The authors present a theoretical study of nanoparticle morphology and gas/droplet partitioning behavior of water using systems consisting of sodium chloride, water, and pimelic acid. The authors discover several parameters - sphericity and fractional surface coverage - that aptly describe chemical morphology as a function of composition and size regimes and variation in mass accommodation coefficients. The authors also report a threshold for the validity of continuum theories. The paper is well-written and is of interest to the Atmospheric Chemistry and Physics community, and is recommended for publication after the following general comments have been addressed.

As the authors note in Section 3.4, classical water models are known to have biases in errors in reproducing experimental surface tensions - though with SPC/E having one of the smallest errors (Vega and de Miguel, 2007). Additionally, a study (Lbadaoui-Darvas and Takahama, 2019) suggest that carboxylic acid-water dynamics are not well captured in equilibrium MD simulations and lead to deviations in predictions of water activity even above 0.95. On the other hand, the water activity calculations seem to suggest that the simulation results are in good agreement - with observations - is this due to canceling of errors (e.g., with molar volume) or the relatively small magnitude of the error in surface tension by these models? 

Many of the conclusions summarize the effect of "organic loadings" but the simulations use a specific type of organic, namely pimelic acid. Many studies on the other hand suggest the importance of alcohols in marine aerosols (e.g., Russell et al., 2010). Is there reason that the authors can justify broadening the conclusion from a particular "organic acid" to "organics" generally? Other abundant dicarboxylic acids (e.g., oxalic acid) may also exhibit different bulk/surface partitioning behavior than demarcated by the sphericity factor. The main question is whether parts of the manuscript should be more clear in what is meant by "organic loading" in this work.

References:

Lbadaoui-Darvas, Mária, and Satoshi Takahama. “Water Activity from Equilibrium Molecular Dynamics Simulations and Kirkwood-Buff Theory.” The Journal of Physical Chemistry B 123, no. 50 (December 19, 2019): 10757–68. https://doi.org/10.1021/acs.jpcb.9b06735.

Russell, L. M., L. N. Hawkins, A. A. Frossard, P. K. Quinn, and T. S. Bates. “Carbohydrate-like Composition of Submicron Atmospheric Particles and Their Production from Ocean Bubble Bursting.” Proceedings of the National Academy of Sciences of the United States of America 107, no. 15 (2010): 6652–57. https://doi.org/10.1073/pnas.0908905107.

Vega, C., and E. de Miguel. “Surface Tension of the Most Popular Models of Water by Using the Test-Area Simulation Method.” The Journal of Chemical Physics 126, no. 15 (2007): 154707. https://doi.org/10.1063/1.2715577.