Review of “Reversible and irreversible gas-particle partitioning of dicarbonyl compounds observed in the real atmosphere”

Hu and coauthors describe a set of field experiments designed to investigate the gas/particle partitioning of glyoxal and methylglyoxal in Beijing as a function of season. Gas phase dicarbonyls were collected using DNPH-doped cartridges and particle phase dicarbonyls were collected using a filter assembly. Both reversible and irreversible uptake pathways are considered using supporting measurement data and irreversible pathways are found to be dominant for both dicarbonyls in all seasons, although reversible uptake (self-reaction, oligomerization) becomes more relevant in the winter. As expected, the field data demonstrate particle phase concentrations that are orders of magnitude higher than those expected based solely on absorptive partitioning theory. This study, however, is particularly useful in demonstrating the estimated dominance of the irreversible pathway in all seasons, and in presenting real-world SOA contributions for these dicarbonyls at this urban location (approximately 25% of Beijing SOA is assigned to glyoxal/methylglyoxal uptake processes). Overall, I find the manuscript to be well written with comprehensive consideration of the relative importance of reversible and irreversible uptake pathways and their impact on SOA production. These data should be useful for optimizing dicarbonyl uptake in SOA modeling efforts.

Comments

Line 86: How were positive artefacts from direct deposition of gas phase glyoxal/methylglyoxal on the filter surfaces accounted for? This could bias the partitioning result from Equations 1 and 3 and should be discussed in the text.

I think the manuscript would benefit from an expanded discussion of assumptions used to calculate the reversible and irreversible pathways. Calculations of the irreversible pathway involve the particle phase and gas phase monomer dicarbonyl concentrations to be known and these data are derived from analysis of the extracts. But the particle phase monomer extract concentration will also include the contributions from the reversibly formed products present in the extract. Expanding the discussion of deriving the cp term in the formulas and what exactly it represents would be useful for readers.

Figure 2c: Change the color of one of the grey traces

Figure 3: Define SNA in caption

Figures throughout: Colorscales should go from lower values in blue to higher values in red to be consistent with general uses in the literature. It is counterintuitive for the reader for these to be the other way around. Also the axis scale in the colorscale legends has numbers increasing from right to left which is also confusing.

Abstract, change last line to present tense. eg “To our knowledge, this article is the first to…”

Line 33: reorder the two references

Line 36: consider rephrasing to “lost in the gas phase by photolysis, oxidation by OH radicals, and dry deposition” as OH oxidation is a photochemical reaction

Line 40: “Although they have relatively high…”

Line 41: How relevant is adsorption to surfaces vs absorption into the bulk particle phase material? Worth discussing here

Line 90: “common”

Line 94: “time resolution”

Line 167: “soluble”

Line 170: “close to”

Line 180: “real atmosphere”

Table 4: “Particulate matter”

Line 206: Worth noting that the observed RH dependence for the reversible pathway is consistent with the Healy et al 2009 chamber study reference

Line 270: “close to”

Line 284: Define SNA

Line 285: rephrase

Line 339: rephrase

Jingcheng Hu and co-authors have measured the gas-particle partitioning of dicarbonyl compounds, especially glycol and methylglyoxal, at field sites in China. This is a highly relevant topic for the readership of ACP. As far as I can tell (being a computational chemist, not an experimentalist), the study is well carried out, and the manuscript is well written. I can thus recommend publication subject to some fairly minor revisions.

Detailed issues

-The authors spend a lot of time pointing out that the measured partitioning is much higher than what they call the “theoretical” values - the latter seem to correspond to values obtained for the partitioning coefficient (or Henry’s law constant) of pure molecular glyoxal and methylglyoxal. However, as evident from their own introduction section, it is already very well known that the partitioning of these compounds is driven mainly by various reactions. For example, hydration alone is well-known to increase the Henry’s law coefficient of glyoxal by about five orders of magnitude (as discussed e.g. in Ip et al 2009, https://doi.org/10.1029/2008GL036212, or Kampf et al 2013 cited in the manuscript). The authors contribution to separating reversible and irreversible pathways is substantial and valuable - but just reporting that partitioning is much stronger than the “theoretical” values is not really novel (or even that interesting), and this aspect of the abstract and discussion should be toned down. For example, the speculation about “misidentification” or “discrepancies” around lines 175-180 is not really warranted: we already know mechanisms which can easily explain at least most of the observed deviations from the “theoretical” pure-compound values. 

-Concerning the saturation vapour pressures discussed around line 190: are there any estimates of the relative saturation vapour pressures of the reversible vs irreversible products? Both are of course much lower than the saturation vapour pressures of the parent dicarbonyls (this is quite well-known and obvious), but how do the two product sets compare with each other? This would be a very interesting parameter to know in terms of evaluating the atmospheric impact of the "reversible vs irreversible" competition. 

-Line 208: “strong and positive dependence on particle acidity (pH)”. Please be clear here: did the concentration increase with acidity (i.e. with decreasing pH), or did it increase with pH? These are opposite things. 

Hu et al. present observations of glyoxal and methylglyoxal collected during four seasons in Beijing. The observations included gas-phase and aerosol-phase dicarbonyls. With these observations, the authors investigate the paritioning / reverisible and irreversible uptake of the dicarbonyls. They find that theoretical values underpredict the real world observations. Further, they find that irreverisible uptake dominates in all seasons, though reversible uptake becomes more important in winter time. This study provides an interesting data set and way to investigate this long standing question of the uptake of dicarbonyls to aerosol as other studies normally just have gas-phase measurements and use a steady state model to derive the first order uptake of glyoxal to aerosol.

Though this paper is of interest to the ACP community, there are some aspects of the paper the authors can improve upon to improve the overall study. With the clarifications suggested below, the manuscript would be acceptable for ACP.

1) One of the major areas that would benefit with expanded text would be the methods. Currently, there is not enough information in order to understand the measurements and discussions from the authors. The following discussions in methods should be added to improve the understanding of the paper:

1a) As the authors are collecting the gas-phase dicarbonyls onto cartridges, a discussion on the percent collected / percent lost both during the collection and extraction / analysis period.

1b) Similarly, the authors should have a discussion about the percent collected / percent lost for the dicarbonyl aerosol on filters.

1c) Another reviewer commented, and I agree, a discussion about potential artifacts for both methods, but especially the aerosol filter collection, needs to be included. This includes if there was a cyclone for size selection, is there a denuder to prevent gas-phase from being collected onto the filters, how long the filters were collected, potential lost of dicarbonyls from the filters during sampling or preparation, and potential side reactions on the filters that may have led to biases.

1d) Besides how much material is recovered for sampling, how well were these two dicarbonyls identified? E.g., as it is expected that there are other dicarbonyls, how well were the peaks separated for glyoxal and methylglyoxal (an example chromatogram in the SI would be beneificial)?

1e) The authors state the assumption that all dicarbonyls that have done reversible partitioning to the aerosol-phase are extracted as the parent compound. A discussion showing this to be true either in the methods or in the results would be be beneficial (e.g., if possible, having the reversible products on a filter, extract, and see if they come out as glyoxal/methylglyoxal in the chromatagram).

1f) The irreversible uptake calculation (page 9, line 252 - page 10, line 268) should be moved to the methods.

1g) Were blanks collected? What is the LOD for both methods?

1h) What is the uncertainty associated with the assumptions made to calculate Kp? E.g., there would be high uncertainty in activity coefficient, vapor pressure, and potentially the absorbing fraction of the total particulate matter, depending on how well the methods measured total OA.

2) As the authors state different comparisons for the values they observed / calculated, it would be beneficial to either in their current tables or in a new table compare their results with literature.

3) I agree with the other reviewers that the discussion of theory (Section 3.1.2) does not add much to the paper as this is generally already known and would advise to either reduce this discussion or potentially remove it for more room to expand upon the reversible, irreversible, and methods.

4) It is currently unclear how the authors are separating irreversible and reversible. This is especially important in the partitioning calculations, as how much could the irreversible uptake be influencing the calculated value? Further, as the reversible was 10% or less the process the dicarbonyls undergone, is that within the associated uncertainty in the calculations, indicating potentially minimal reversible lost?

5) It is currently unclear how the authors derived the values in Fig. 2 (reversible pathway with units of ng/ug). If this is from one of the equations, please state and that will help better understand where the data from this figure originated from. If something else, please describe.

6) Another concern with Fig. 2 is the fact the authors are showing trends vs pH. As they are calculating their pH from only aerosol-phase measurements, there is large inherant uncertainty in the pH values as there is no gas-phase measurements to constrain the partitioning of the semi-volatile gases (NH3, HNO3, or HCl), which can lead to large deviations in the calculated pH from real world observations. I strongly advised the authors to not use the pH as it does not add much to the results.

7) I'm assuming the values listed in Table S2 are for bulk-phase reactions instead of aerosol-phase reactions. Recent studies have shown that these bulk-phase reactions may not represent the aerosol-phase reaction rates due to the differences in the ionic strength. Therefore, for lines 210 - 222 and Fig. 2b, I would recommend the authors to be careful with those numbers in being the "definitive" product (also correct porduct to product in 2b) distribution to the potential product distribution with uncertainty due to bulk vs aerosol phase.

8) Lines 276 - 279: It's currently unclear how the author's are drawing the conclusion that methylglyoxal is exhibiting unexpected salting-in effects if they are using Eq. 4 - 7 to calculate the uptake coefficient. As these equations don't include the aerosol composition or ionic strength, further clarification on this conclusion would help this statement.

Minor

1) Since ionic strength is being calculated with the aerosol liquid water, it may be useful to look at how these parameters relate to ionic strength.

2) For Fig. 1, it is currently hard to following what is happening with the particulate-phase dicarbonyls. I would recommend including a thin-line connecting the points to better see the data and potential trends.

3) Line 287, it should be Fig. S8 instead of S7.

4) For Fig. S7, it is unclear what sequence number (x-axis) and what the grey shaded area are for.

5) For table 1, it would be useful to include the dates of the measurements.

6) For table 2, it is unclear how "theory" Henry's law constant is calculated compared to the "field" values.

The article by Hu et al. titled "Reversible and irreversible gas-particle partitioning of dicarbonyl compounds observed in the real atmosphere" discusses the importance of reversable and irreversable gas-to-particle partitioning of glyoxal and methyl glyoxal. The authors present experimental and modeling results showing how irreversible gas-to-particle partitioning dominates the two partitioning pathways and also highlighted the other reaction processes that were not taken into account in the analysis of this study. The study is relevant for the atmospheric community and can be accepted to ACP after the comments have been addressed.

Major corrections

1. Page 5, line 134: What are the other carbonyls that were measured in the gas and particle phases?

2. Based on what's written at the end of page 7 and later, the measured dicarbonyls in the particle phase are only ones that have formed products of the reversible pathways. It wasn't clear earlier when you talk about experimental partitioning coefficients that they only include reversible partitioning, point it our somewhere earlier to avoid confusion. 

3. Page 8, line 210: How exactly are the proportions of hydrates and oligomers at different RH calculated? Table S2 gives the hydration rate constants as (pseudo) first order rate constants, so the amount of water should have no effect on the equilibrium, right? Or are the experiments used in these calculations somehow? Please specify in the text.

Minor correstions

1. In the abstract "These two pathways of dicarbonyls jointly contributed to more than 25% of SOAs in the real atmosphere"

2. Page 1, line 27-28: "The α-dicarbonyl functionality increases their water solubility and reactivity more than expected" would be better if you say something like "The α-dicarbonyl functionality leads to higher water solubility and reactivity than expected." Otherwise, specify how the solubility and reactivity have increased (from what).

3. page 2, line 36-37: "however, there is still a missing sink for the two dicarbonyls" Do you mean that the known sinks listed before are not large enough to explain the loss of the dicarbonyls from the gas phase? Or that there is a specific sink mentioned by Volkamer et al. that wasn't listed here? Please specify.

4. page 3, line 72: "among key regions with relatively higher PM2.5 concentrations" Do you mean that the the key regions have relatively higher PM2.5? Or Beijing has relative higher PM2.5 concentrations than the other key regions? Please specify.

5. Page 5, line 145: define GL and MG 

6. Page 6, line 164: "lower temperature promoted the partitioning processes" do you mean gas-to-particle partitioning, or also particle-to-gas? It isn't clear by saying "partitioning processes". 

7. Page 7, line 197: "which are more reactive than their counterparts" how do you determine "more reactive"? Aren't glyoxal and methylglyoxal also reactive, because they quickly react with water to become hydrates? Or are the reactions of the hydrates even faster than the non-hydrated glyoxals?

8. Page 7, line 199: "the most thermodynamically favored oligomer reactions for glyoxal and methylglyoxal" Specify that the reactions are for the hydrates, not (only) non-hydrated glyoxal and methylglyoxal.

9. Page 8, line 208-209: "The product distribution of the reversible formation could well explain this phenomenon." How?

10. Figure 2c: The two gray colors (estimated and theoretical values) are very similar, how about using some colors for them? Also, correct "porduct" to "product" in the title and add y-axis label to Figure 2b.  

11. Page 8, line 223: "Combined with the vapor pressure of dominant products" where do you get the vapor pressures of the dominant products?

12. Figure 3: What are the lines in 3b? Also model like in 3c? Also, there are typos in the caption "(i) galyoxal and (ii) methylglyxoal"

13. Page 10, line 287: There are 2 figures in the Supplement labelled S7. In the second Fig. S7, what is the concentration unit for SNA in the ratios? Mass/mole/volume ratio?

14. Page 13, line 369: "Furthermore, we note that there may be other potential explanations for the increase in particulate concentrations and the uncertainty in the gas-particle partitioning process." Particulate concentrations of what? And which partitioning processes?

15. Page 13, line 371-372: "Other reversible pathways, like adducts formed from glyoxal with inorganic species, like sulfate and ammonia, could also promote the gas-particle partitioning process." I think you mean "such as", not "like". You used the the word "like" similarly also earlier in the manuscript so check those too.