Thanks to the authors for this interesting study, which uses field observations and particle-resolved simulations to suggest that black carbon complexity can be modeled more simply than has previously been thought. Accurately quantifying black carbon’s radiative impacts has proved elusive, partly because global and regional models lack the complexity needed to simulate black carbon direct effects, as pointed out by Fierce et al and others. The results of the present study look promising, and may lead to more reliable black carbon radiative impact estimates in models with parameterized aerosol properties. This is of interest largely because black carbon concentrations are extensively changing, due to regulations on industrial emissions but also increasing wildfire frequency, with current climate models having little ability to reveal how these changes affect surface temperatures. I’m hoping the authors can make some clarifications on their results and – being myself a climate modeler – I’m especially interested for more information on how these results might advance climate and chemical transport models.

Main comments/questions:

- I find it quite interesting that a steady state is reached within a day by the chosen metrics, but would like more explanation for why this particular result is useful from a modeling perspective. Would I be interpreting correctly to say that the pre-steady state period is harder to model than the steady state itself, yet is sufficiently short that an accurate representation of BC properties could reasonably ignore this period and focus on estimating steady-state chi and k values? I imagine such logic might hold reasonably for aerosols averaged over large regions (as in a GCM), but may fare less well for, say, a high-resolution modeling study of an urban area wherein a larger percent of BC emissions are fresh.

- As I’m a modeler I’d be very interested if the authors can please further explain how their findings could be used to advance BC treatment in models. Could the authors envision simple parameterizations for the steady-state chi and k being put into a climate or chemical transport model, as functions of emission rates and conditions? Presumably this would depend on the ratio of emitted BC to organics, the total mass, and possibly meteorological conditions and other factors. I'm interested in the authors' thoughts here, and ideally they might expand their two sentences in the Conclusion (Lines 334-8).

- I quite like the reconciliation of this study’s results with those of Fierce et al 2016, though I feel a few lines on that study should be more accurate. For instance Line 27: “diversity in the distribution of BC coating thickness that has been documented in previous studies”, and Line 62: “the study of Fierce et al. (2016) has noted that the coating thickness of BC aerosols is non-uniform across the distributions of BC cores”. The Fierce et at study describes particle-resolved results as dissimilar to “uniform composition” and “uniform mixing”, but not uniform coating thickness, so I find this inaccurate. Further, these descriptions make it sound very much like that study and the current one directly disagree, which is not made out to be the case. Further, this is more directly indicated to be the case in Line 70 “we explain the discrepancy between […]” rather than an “apparent discrepancy” as in Line 64. 

- I’m wondering if some of the terminology related to the interpretation of results could be clearer. Is “uniform” a clear enough description of the slopes closely following a k value? This could suggest something being uniformly distributed, which is not the case here (despite Fig. 5’s legend mentioning a “uniform distribution”). Would “independent of BC core size” or just “size-independent” be more accurate? For the several comparisons between “uniform” and “non-uniform”, these aren’t innately incompatible but the wording makes it seem that this is so. Perhaps a clearer summary would just say that the coating volume fraction varies with BC core size, but coating thickness is size-independent? Maybe there’s a better description the authors have in mind?

- Related to my comment immediately above, could the authors please comment on whether the normalization used to plot the data (Fig. 3) affects interpretation of the results? 

- The results are based on field observations in Nanjing. Can the authors please comment on whether Nanjing is sufficiently representative for the key results to hold generally? Would the steady state timescale be quite different for a site with natural biomass burning black carbon, rather than industrial black carbon? I see there are some simulated cases that are variants from the Nanjing one, but I don't have a sense by how much.

- Could the authors please briefly explain their reasoning behind the statement that chi and k give a “comprehensive depiction of the BC mixing state” (Line 134)? Certainly these two parameters could be superior to mixing state index only, but as this is presumably the first study to combine these metrics there might not be anything to cite in support of this statement, so I request a little explanation.

Specific comments:

- Line 61 states that “key scientific questions remain such as determining the characteristic timescale to reach a steady state”. The characteristic timescale has been examined here. Do the authors feel there are there remaining “key questions” to address that could be worth adding to this line?

- Since Shannon entropy and particle diversity metrics are shown in Table 2, could how these enable a mixing state index please be briefly summarized in the text, which otherwise does not mention these?

- I find the Table 1 caption (“The related quantities calculated […]”) to be worded clunky. Perhaps this could more simply be described as “Metrics of particle mass” or otherwise rewritten?

- There’s an error in Table 2 where some of the instances of ‘a’ should instead be alphas (e.g. ‘Ha’,‘Da’), which gives an incorrect impression that these metrics are species-specific.

- It seems a bit odd for there to be 2 tables full of aerosol parameters without including the ‘k’ that is used extensively in this study.

- The Results would be easier for the reader to follow if it were divided into a few sub-sections.

- Lines 213-219 could be in the Methods, as they slightly distract from the flow of the results.

- In Fig. 5 the label “Ideal uniform distribution of BC aerosols” seems off, since this is the coating thickness distribution, while BC is still lognormal if I’m not mistaken.

- In the caption to Fig. 5, should “normalized by dividing each component's mass by the maximum particle mass” instead be “total particle mass” following the other normalization described below?

- The Discussion material feels like it could be a Results subsection. It might make sense to instead use the Discussion for putting the findings into the broader context of other work and explain the potential utility for modeling efforts.

The manuscript “Steady-State Mixing State of Black Carbon Aerosols from a Particle-Resolved Model” by Zhang et al., investigates the mixing state of black carbon aerosols. Their results indicate, based on both a particle-resolved model and observations in Nanjing that the mixing state of BC aerosol reaches a steady state in few hours. The results of the paper are interesting and can be useful for improving the treatment of mixing of BC with scattering compounds in atmospheric models. However, the current study does not provide yet the means to apply this finding in models. The paper in within the scope of ACP, it presents novel ideas, reaches substantial conclusions. The paper is well written and I can recommend accepting it for publication after the following issues are addressed.

My main comment is related to how to apply this knowledge in atmospheric modelling. I assume that since this holds only near emission sources, parameterizing the mixing state should be embedded in emission schemes, right? Away from the sources and higher up in the atmosphere where there are no emissions and surface removal, such steady state assumption may not hold. It would be good to add some discussion about this.

-Page 3, Line 97: “In this study, the initial gas concentration and emission rate were slightly adjusted based on Riemer et al. (2009) according to Wang et al. (2017).” The meaning of the sentence is unclear. What was the initial assumption for gas concentrations and emissions and how were they adjusted? Were these assumption tuned to match the model with observations?

- Was the motivation for the additional cases to show that for all conditions, the exponential linear distribution occurs?

Technical comment:

Please add legends or explanations for curves and points in Figures S2 and S3