The paper describes laboratory measurements of gas-ice partitioning coefficients for 14 carbonyls along with an analysis of thermodynamic properties. The paper provides very useful information of these partitioning coefficients that can be applied in atmospheric chemistry models. The analysis is interesting, indicating a quasi-liquid surface layer may play a role, but also that carbonyls with lower molecular mass are more likely to be taken up into the ice crystal lattice. Scientifically, the paper is good. Its analysis brings up a number of questions, including why methyl vinyl ketone (MVK) behaves differently than the other carbonyls.  

The paper does need some improvement. In general, there is a need for better clarity: explaining the methods and analysis for those less familiar with these tools and explaining the results in relation to atmospheric chemistry implications. As noted, the MVK behavior is curious. It may be worth having a separate small section synthesizing what was learned about MVK with possible explanations as to what causes its behavior and potential future areas of investigation.

Below I list several comments that I would like the authors to address before considering the paper for publication.

Major Comments

1. The results for MVK are perplexing for both their weak response to temperature but also the MVK partitioning coefficient values are much different than those for methacrolein, which has the same molecular mass. I was curious whether the functional group(s) of the carbonyl play a role in its deposition onto ice. At first glance, the aldehydes seem to have higher partitioning coefficients than the ketones, but that is likely because the aldehydes generally have a lower molecular mass. Further, another pair, acetone and propionaldehyde, appear to have similar partitioning coefficients with each other. Nevertheless, the results in Figure 4 show partitioning coefficients varying by two orders of magnitude for the same molecular mass (e.g., 58 g/mol). Could the authors comment on the role of functional groups and suggest other properties that may be controlling the carbonyl’s ice partitioning coefficient in the manuscript?
   
   
   
   The explanation of the weak response of MVK with temperature is a bit speculative. If it is the kinetic control of transport, then could this be investigated from theoretical calculations? If doing additional calculations is beyond the scope of the current paper, then recommendations for further research should be stated.
2. The paper would benefit by revising the text so that it is more easily understood by those who are not very familiar with laboratory studies. That is, explaining the experiments in plain English as well as providing the technical details would be useful. This should be applied to the analysis approach and discussion as well.

Specific Science Comments

1. I am confused by the remark in this manuscript that states, “The gas-ice partitioning coefficients observed here are below the 10 mol m^–3 Pa^–1 threshold given by Crutzen and Lawrence (2000) to be considered a substantial atmospheric removal process.” (line 25-26 abstract and line 528). The threshold given by Crutzen and Lawrence is the trace gas solubility, which they used as an indicator for scavenging by cloud particles. However, what I learned from this current manuscript is that the solubility of the trace gas did not play a role in its direct uptake onto ice. For example, glyoxal is very soluble (much more than formaldehyde) but its ice uptake is smaller than that of formaldehyde. I do not see the relevance of making this remark.
2. The 14 carbonyls investigated in this study were never introduced individually. On line 74, it states that 14 carbonyls are investigated and then the reader does not learn which carbonyls until section 2.3 which discusses the materials used. In the Introduction, it would be good to have a paragraph listing the carbonyls providing explanations of their atmospheric chemistry relevance, especially for the upper troposphere where the impact of deposition onto ice will have its greatest effect.
3. Lines 40-44. These sentences are still quite general in explaining how carbonyls are relevant to atmospheric composition and chemistry. There should be statements about their role in ozone formation as a source of peroxy radicals when they photodissociate and a better description of how they contribute to secondary organic aerosol formation.
4. Lines 44-48. If a trace gas is being “removed from the atmosphere” by either dry or wet deposition, then would not one want to determine the resultant deposition? Further, the results of this paper show that deposition of carbonyls onto ice is greater at colder temperatures relevant to the upper troposphere. Specifically, this process would be happening in cirrus clouds and convective anvils. The carbonyls are moved from the gas phase to the ice phase. However, at some point the ice will sublimate or fall and melt. When that happens, the carbonyls will be released back to the gas phase. In my mind, this carbonyl deposition onto ice must be characterized as an effect on tropospheric gas-phase chemistry while the ice cloud is present, and then as a source to the local region for gas-phase chemistry when the ice sublimates.
5. Line 75. I would argue that 10 ppbv of a carbonyl is two orders of magnitude larger than mixing ratios in the middle to upper troposphere. For example, formaldehyde is typically 50 pptv in the background upper troposphere.
6. Lines 330… I was wondering why formaldehyde has the highest ice partitioning coefficients compared to other compounds. What is it about each of the compounds that make it more or less likely to be taken up by ice. Since this topic gets addressed later (section 3.5), I suggest adding a comment here that an explanation is given later.
7. Section 3.2. How do the results presented in this paper compare to previous literature (e.g., Winkler et al., 2002 that’s cited in the paper)?
8. Line 472 mentions that a few factors were examined but only molar mass and heat of vaporization are discussed in the section. Could the results for the other factors be displayed in the supplement?
9. Lines 496-499 are interesting and really helps to explain the process. Are there references that should be cited that have examined the incorporation of trace gases into the ice crystal lattice? If so, please cite them.
10. Lines 529-530. It states that one should divide the uptake coefficients presented in this paper by RT in order to get a coefficient that can be compared to other coefficients like the Henry’s Law coefficient. I suggest that the authors provide that information in the supplement. Indeed, it would be useful to have temperature-dependent coefficients listed so that it can easily be used by those exploring chemistry in/near ice clouds with a model.
11. Line 530… How do the uptake coefficients compare to the HNO₃ uptake? The HNO₃ uptake onto ice is a well known process, so such a comparison will give readers a number to relate to.

Organization, Clarity, Technical Comments

1. I would like to see better construction of paragraphs. That is, each paragraph should start with the main idea (a topic sentence), followed by supporting sentences providing evidence for that idea and a concluding sentence to summarize or transition to the next sentence.
   
   
   
   As an example, the paragraph on lines 40-48 begins with introducing carbonyls, giving a general explanation of their relevance to atmospheric chemistry. However, in the middle of the paragraph the topic changes to the lack of knowledge about removal of carbonyls from the atmosphere. Here, I would suggest splitting these two topics into two paragraphs and providing more detail on how carbonyls contribute to ozone and aerosol formation for their relevance, while the removal by deposition could be generalized to mention dry deposition, wet deposition via cloud water uptake, and ice deposition. 
   
   
   
   Please review the entire manuscript on paragraph construction as there are other places (e.g., lines 349-360) that have the same issue.
2. There are a lot of acronyms. Be sure each acronym is written out the first time it is written (DNPH is not). In addition, be sure the acronym is needed, as it is easy to slip into the jargon of this specific topic and become less relevant for the broader atmospheric chemistry community.
3. Lines 121-125. It seems to me that these sentences about the PTFE tubing belong with “The second stage” paragraph (lines 103-116), which is where I was wondering about wall loss within the flow tube.
4. Line 192. Why not write ACN as CH₃CN?
5. Tables: Why does the order of species listed start with MVK and then is alphabetical?
6. Line 351. It should be, “exclusively be attributed to”.
7. Line 351. What is co-depositing? Is it the individual carbonyl and water vapor, or is it multiple carbonyls depositing together?
8. Line 371. Please remind the reader why the thermodynamics results are being investigated. Try to write to the message you want the reader to learn from this section (i.e., starting with a description of Table 2 and 3 does not draw the reader into the science).
9. Lines 373-377. It seems like Table 2 should be discussed before Table 3.
10. Line 412. Should it be, “which is attributed to weakened …”?
11. Line 417. Is there photodegradation in the flow tube?
12. In the supplement, what is IKR?

Figures and Tables

1. Table 1. Please remove the banded background coloring as it makes it harder to read. It would be nice to have a little more space between rows. Another thing to think about is to present all values greater than 1 without the exponential part (e.g., HCHO at -40degC written as 103 +/- 63. That way the discussion (line 332) about net uptake is easy to see in the table.

Summary

This article discusses a study aimed at measuring gas-ice partitioning coefficients of 14 carbonyl gaseous species onto crystalline ice surfaces. Through compositional analysis of the ice crystals after gaseous carbonyl exposure, insights into uptake behavior are gained. These insights are gained via: 1. the catalysis of ice crystal growth on the walls of a flow tube apparatus at temperatures that mimic those in the troposphere (-20, -30, and -40 ̊C), 2. vapor pressure dependence studies of each carbonyl species and how it plays a key role in diffusional uptake onto the ice crystal, 3. the determination of partitioning coefficients and using Van’t Hoff analyses to calculate the entropy and enthalpy of uptake, and 4. providing intercomparison studies produced by other researchers to help draw conclusions and provide reasoning to their overall findings and shine light on knowledge gaps that need to be further explored. The authors provide data covering various factors, such as temperature and vapor pressure, and how these factors contribute to the diffusional deposition of gaseous carbonyls onto ice surfaces. Although an in-depth study was reported, I had some questions that I feel should be addressed before the article is published.

Comments:

1. Line 67: replace K_g,ss with K_l,ss
2. Were the gaseous species cooled to experimental temperatures before entering the flow tube? If flowing room temperature/warmer analytes into a chilled flow tube, wouldn’t it take time for the analyte to reach lower temperatures, skewing the vapor pressures (especially since the temperature in the flow tube varies at certain distances from the inlet)? The gas analyte would exist at the same temperature as the ice crystals in the atmosphere. Is the time it takes to chill the gaseous analyte accounted for in the flow tube? What is the residence time of the analyte within the flow tube?
3. Line 74: If I am reading this correctly, it seems that uptake experiments done using mixtures of carbonyl compounds, rather than individual components. How can the authors be sure that there are no cooperative or competitive effects impacting the uptake equilibria? That is, some compounds might adsorb more preferentially and block an adsorption site or displace another carbonyl that may bind more weakly to the ice. Likewise, could it be that some species bind forming a monolayer of organics that and now create a more favorable surface for subsequent VOCs to bind to.  Have the authors done the requisite experiments to investigate how the uptake depends on concentration?  I would expect that uptake would decrease as the concentration of VOC increases, so total concentration of the mixture or of the individual components will be important. Ideally, one would work with one organic component at a time and at low concentration so it can be assumed that interactions between a specific VOC and water in the ice are the only interactions that need to be considered.  
4. Line 225: Was there gas analyte trapped in the flow tube at the conclusion of the experiment and during transfer to the cold chamber? Would you predict that would alter the composition of the ice crystal if residual analyte was trapped for longer times?

1. Line 310: What are the % loss of carbonyls on the apparatus/non measurable surfaces? If these losses were measured, were they used to correct/account the concentrations measured?
2. Line 343. Discussion of potential loss of formadehyde seems to fit better in the conclusion section.

1. Was optimizing flow tube surface area, length, etc.. studied to increase ice crystal growth uniformity?

1. Is there a better method for studying deposition that allows for better ice crystal uniformity and collection without use of many solvents and steps? Are you concerned with analyte loss during recollection methods?

1. Page 16, Figure 2: Individual data points are difficult to identify in the plot. Is there another way to show the data and increase ease of identifying each point?

1. Page 19, Line 416: You mention photodegradation as a potential factor for weak correlation with MVK, but did you preform any studies with your system to further confirm or deny this? Do you plan to conduct photolysis studies with these systems in the future?

1. Line 484: Carbonyls can’t H-bond between themselves, but they can interact strongly via dipole-dipole interactions. How does this figure in the discussion here?
2. Line 492-500: I have questions about this paragraph, where the authors are speculating on the mechanism that is leading to the observation that uptake tendency into the ice is inversely proportional to molecular size. My first question pertains to the sentence, “in order to be taken into the ice phase, a compound must be incorporated into the ice crystal lattice structure.”  A crystal structure is comprised of molecules packed into a particular lattice. Therefore, I believe the phrasing to use is just “the ice crystal structure.”  If that is true, then are the authors suggesting that the smaller ketones are cocrystallizing with the ice?  Or are they suggesting that the molecules are deposited in grain boundaries and more adsorption occurs when defects are enhanced? It seems to me that co-crystallization is very unlikely for ketones + water as this would result in a completely new and unique crystal structure.  I believe it is more likely that the ice crystals grown from vapor deposition are mesoporous with a high surface area and that the behavior the authors have demonstrated (increased tendency to partition to ice with decreasing molecular size) describes the process of adsorption into mesoporous materials, where larger molecules are sterically inhibited from entering the small pores, while small VOCs diffuse more easily leading to much higher adsorption capacities. If the ice surfaces were exposed to a cocktail of VOCs, would there would be a preference for the smaller VOCs adsorbing into the pores first, and the filled pores would then exclude other molecules from filling in.
3. Line 526-528: It is stated that the K-values measured in this study are below the Crutzen and Lawrence threshold for being substantial atmospheric removal processes. But what about formaldehyde, which according to line 343 has a high tendency to deposit in ice (see comment above, this may be a better place to place that discussion).