General comments:

In this paper, the authors present their findings on the ice-nucleating particle (INP) characteristics of the Sea Spray Aerosol (SSA), which were generated from a Marine Aerosol Reference Tank (MART). Additionally, the researchers explored the impact of atmospheric aging on these characteristics. Interestingly, they found no observable effect of the atmospheric aging. On the whole, the study is methodologically sound. However, I must express three major concerns as well as a few specific issues related to the content, which I will delve into more deeply below. Despite these concerns, the study remains intriguing and offers valuable contributions to the broader scientific community's understanding of the INP characteristics of SSA. There is no doubt that with necessary revisions, the work will be worthy of publication. Nonetheless, it is imperative to note that major revisions are required to elevate the study to its full potential.

Three major issues:

1.The freezing mechanism at temperature < 220K

The elucidation of the nucleation mechanism in SSA remains a significant and yet unresolved scientific query. The authors of the current study, intriguingly, appear to circumvent direct discussion of the low-temperature nucleation dynamics of SSAs. They opt instead to vaguely encapsulate the complex phenomena using the generic term 'heterogeneous freezing.' The data put forth in this paper, particularly as illustrated in Figure 5, presents a compelling view. It appears to document a transition from homogeneous to heterogeneous nucleation as temperatures descend towards 220 degrees. Nucleation observed under these chillier conditions within a range spanning from water to ice saturation. This behavior should ideally be defined as deposition nucleation, however, Figure 5 sheds light on the temperature interval wherein the pore condensation freezing (PCF) manifests itself. A striking alignment is observed between the nucleation occurring below 215 degrees and the PCF. This concurrence seemingly substantiates the notion that SSA nucleation under colder conditions could indeed be characterized by the PCF.

2.The phase state of the SSA

The main objective was to compare the INP characteristics of the pure and aged SSA, however, the reviewer was concerned the state of particles could influence the results. In this study, the measurement of humidity was performed before the coil cold trap, maintaining a controlled relative humidity at 10% under ambient temperature conditions. However, this level of water vapor pressure can escalate from a few thousand to tens of thousands supersaturation with respect to ice at 220K. Consequently, it is imperative for the authors to consider the dwell time within the coil cold trap and the Continuous Flow Diffusion Chamber (CFDC). Furthermore, it would be beneficial to generate estimations of the phase state prior to its entry into the CFDC.

3. The setup of the experiment

Currently, the sample air directly enters the oxidation flow reactor after exiting from the MART instrument. It is suggested that the sample air should be dehumidified before passing through the oxidation flow reactor. This is because, during liquid-phase oxidation, the crystallization of SSA solution droplets into crystals may not affect its surface structure. However, if oxidation occurs in the solid phase, the pores on the surface of the SSA particles might be filled, thereby affecting its ice nucleation properties.

Specific comments:

P1, Line 16: I suggest use the supersaturation with respect to ice instead of RH to evaluate the onset of the ice nucleating forms.

P3, Sec 2.1: The artificial seawater was filter through a TOC+HEPA filter to remove the insoluble particles, what about the natural seawater? There might be some dust and biological particles inside the natural seawater which have influence on the INP measurement.

P5, Line 151-152: I was wondering why there were so many particles during the “blank experiment” with DI water. Does that mean there were contamination of the MART and sampling tubes? Thus, I strongly suggest do “blank experiment” before and after each experiment.

P5, Line 155: TSI models 3080, remove the 3081 and 3010, which is the model of DMA and CPC.

P5, Line 160: This paragraph is confusing and very hard to understanding.

Line 195-295, when discussing the influence of organics, some studies (Ignatius, et al., 2016; Knopf et al., 2018; Tian et al., 2022) found organic aerosol (likely secondary) could be glassy (Koop et al., 2011) and efficient heterogeneous ice nuclei under the condition of low RH, which could be referenced to support the point that organics itself may serve as INP.

P7, Sec2.4: The IS show the mixed phase regime (-38–0 ℃) INP concentration which inconsistent with the main theme of this study, and the results was shown in the supplement. The author need to carefully consider whether to retain this section of content.

References:

Koop, T., Bookhold, J., Shiraiwa, M., Pöschl, U. Glass transition and phase state of organic compounds: dependency on molecular properties and implications for secondary organic aerosols in the atmosphere. Phys. Chem. Chem. Phys., 2011, 13, 19238-19255.

Knopf, Daniel A., Alpert, Peter A., Wang, B. B. The role of organic aerosol in atmospheric ice nucleationL: A review. ACS Earth Space Chem, 2018, 2, 168-202. DOI: 10.1021/acsearthspecechem.7b00120.

Tian, P., Liu, D. T. Bi, K. et al. Evidence for anthropogenic organic aerosols contributing to ice nucleation. Geophysical Research Letters, 49, e2022GL099990. https://doi.org/10.1029/2022GL099990.

Ignatius, K., Kristensen, T. B., Jarvinen, E., at al. Heterogeneous ice nucleation of viscous secondary organic aerosol produced from ozonolysis of α-pinene. Atmos. Chem. Phys., 2016, 16, 6495-6509.

Patnaude et al. use the Marine Aerosol Reference Tank in combination with CFDC measurements to investigate the effect of atmospheric aging on the ice nucleation behavior of sea spray aerosols (SSA) under cirrus conditions. Simulated aging conditions led to the formation of a new particle mode of secondary marine aerosols (including also some contribution from VOCs outgassing from the MART) and induced some changes in the size distribution of the primary sea spray aerosol particles. These changes, however, had little effect on the observed ice nucleation behavior, which was still mainly governed by the inorganic salts as in primary sea spray aerosols.

The experiments were carefully conducted and analyzed and are therefore worth publishing. My main criticism is along the same lines as the other reviewer's: I would have liked to see more discussion of the nucleation mechanism and the phase state/hygroscopic behavior of the particles involved, which I have detailed in my comments below. I also highlighted a number of points where the discussion of earlier literature data must be improved.

Specific comments

P1, line 11 – 13: Please also mention here the drying step of the SSA to induce crystallization of the salt constituents prior to the CFDC measurements, as this is mandatory for the observation of heterogeneous freezing below 220 K.

P1, line 20: “Thus, any SMA coatings on the pSSA are also unlikely to modify the ice nucleation behavior of pSSA.” I cannot understand this conclusion. For other organic aerosols like secondary organic matter from the oxidation of terpenes and aromatic precursors investigated in the study by Kasparoglu et al. (2022), the pure SOM particles also required water saturation to freeze, but a coating layer of SOM could significantly de-activate efficient INPs like mineral dust at cirrus conditions. So, if a SMA coating does not change the ice nucleation behavior of pSSA, in my opinion two conditions have to be met: 1) the coating does not inhibit the crystallization of the inorganic salts in the drying step before entering the CFDC, and 2) water diffusion through the potential organic-rich coating layer around the salt crystals is also not significantly hindered.

P2, line 38 – 40: I found it somewhat confusing that the authors here also include the process of immersion freezing under mixed-phase cloud conditions, while in line 28 they specifically address ice formation at < 235 K. I would therefore also focus the description here on immersion freezing at cirrus condition, meaning that an INP initiates freezing from inside a sufficiently dilute aqueous solution droplet below the homogeneous freezing conditions for the aqueous layer.

P2, line 51: Please cite some sources discussing this “disagreement” on the contributions of black carbon and SOA.

P2, line 53: “fractions too low to be meaningful” I find this statement too simplistic. There are other studies, such as the modeling work from Beer et al. (2022), which states in the abstract:

“On the other hand, crystalline ammonium sulfate often shows large INP concentrations, has the potential to influence ice nucleation in cirrus clouds, and should be taken into account in future model applications.”

Beer, C. G., Hendricks, J., and Righi, M.: A global climatology of ice-nucleating particles under cirrus conditions derived from model simulations with MADE3 in EMAC, Atmos. Chem. Phys., 22, 15887–15907, https://doi.org/10.5194/acp-22-15887-2022, 2022.

P3, line 77ff: I strongly suggest describing earlier literature results for pure SSA particles in more detail, e.g. with respect to their phase state and hygroscopic behavior, their strongly temperature-dependent ice nucleation behavior under cirrus conditions and the mode of ice nucleation. It is also useful to mention the efflorescence step required to induce crystallization of the particles. In addition, there are a number of studies suggesting that some salt components already dissolve at low humidity, which means that the previously dried SSA particles under ice-saturated conditions are already an internally mixed solid-liquid particle, i.e. with a brine layer around the undissolved NaCl core:

Tang, I. N., Tridico, A. C., and Fung, K. H.: Thermodynamic and optical properties of sea salt aerosols, J. Geophys. Res. (Atmos.), 102, 23269-23275, 1997.

Schill, G. P., and Tolbert, M. A.: Heterogeneous Ice Nucleation on Simulated Sea-Spray Aerosol Using Raman Microscopy, J. Phys. Chem. C, 118, 29234-29241, 2014.

Wagner, R., Kaufmann, J., Möhler, O., Saathoff, H., Schnaiter, M., Ullrich, R., and Leisner, T.: Heterogeneous Ice Nucleation Ability of NaCl and Sea Salt Aerosol Particles at Cirrus Temperatures, J. Geophys. Res. (Atmos.), 123, 2841-2860, 2018.

Even under cirrus conditions, but at temperatures above 220 K, the SSA particles “only” fully deliquesce and nucleate ice homogeneously. Heterogeneous ice formation is only observed below 220 K, and given that the particles are in an internally-mixed solid-liquid state, one could also speculate that the nucleation mode is immersion freezing.

Most of these aspects should then of course also be taken up when discussing the results, but I think it is very useful to give some more details on the IN behavior of pSSA already here in the introduction.

With regard to the potential change in the ice nucleation behavior of pure SSA when it contains a higher fraction of organic material, it is also useful to discuss the temperature ranges above 220 K and below 220 K separately. Above 220 K, the organic components could provoke an additional immersion freezing mode, while purely inorganic SSA only freezes homogeneously. Below 220 K, the question is whether and how the organic components modify the ice nucleation behavior of purely inorganic SSA. Some of this will be discussed later, but for the benefit of the reader, the important questions could be addressed already here. Similarly, you start the second paragraph of your conclusions (P9, line 341) by mentioning some open questions raised by Wagner et al. (2021). I recommend highlighting these open questions already in the introduction and describing how you intended to address them with your study.

P5, line 163ff: This paragraph contains a detailed description of the size distribution measurements, which in itself is of course very justified, but I sometimes found it difficult to follow the lines of thought in the text. For example, in line 167/168 you specifically mention the different diameters at which the aSSA+SMA distributions converge to the pSSA distributions for ASW and SW. This is then immediately followed by the statement “This is consistent …” in line 168. I had expected an explanation here for this different “convergence” behavior in this sentence, but unless I misunderstood, this is given later in line 171. Instead, the sentence with “This is consistent …”, if I interpreted it correctly, refers only to the fact that the aSSA+SMA distributions were dominated by secondary particle formation. So you might consider rearranging the text a little.

As another example, you mention in line 182 that the blank test with DI water “also” led to secondary particle formation. But in fact, in a certain diameter range between 80 to 200 nm, it even led to the highest formed particle number in all experiments (Fig. 2a). This could easily give the impression that new particle formation in the OFR is exclusively governed by impurities/organics from the acrylic outgassing and that direct gas-phase emission from SW and ASW (do you expect a lot of organics from ASW?) is of minor importance. Only later do you give the information “as a final note” (line 188) that the blank test might have been affected by the higher temperature compared to SW and ASW – I think this information should also be given earlier in the discussion. You have also included Fig. S2 in this context to show the change in size distribution with experiment hours (and a concomitant change in temperature). Unfortunately, the y-axis scale in Fig. S2a is given on a linear scale, plotting it with a log-scale as in Fig. 2a would allow a better comparison of the two panels.

P6, line 195: There are a couple of further studies investigating the temperature-dependent efflorescence behavior of NaCl, apart from Koop et al., e.g.:

Bartels-Rausch, T., Kong, X., Orlando, F., Artiglia, L., Waldner, A., Huthwelker, T., and Ammann, M.: Interfacial supercooling and the precipitation of hydrohalite in frozen NaCl solutions as seen by X-ray absorption spectroscopy, The Cryosphere, 15, 2001–2020, https://doi.org/10.5194/tc-15-2001-2021, 2021.

Peckhaus, A., Kiselev, A., Wagner, R., Duft, D., and Leisner, T.: Temperature-dependent formation of NaCl dihydrate in levitated NaCl and sea salt aerosol particles, J. Chem. Phys., 145, 244503, 2016.

Wagner, R., Möhler, O., and Schnaiter, M.: Infrared Optical Constants of Crystalline Sodium Chloride Dihydrate: Application to Study the Crystallization of Aqueous Sodium Chloride Solution Droplets at Low Temperatures, J. Phys. Chem. A, 116, 8557-8571, 2012.

So the statement “phase state and morphology are less well understood at low temperature” should also be revised.

P6, line 199: “porous glassy state”: However, the formation of such porous glassy particles requires a special process called “atmospheric freeze-drying in ice clouds” by Adler et al. (2013). I do not think this is relevant to your study.

P6, line 200-204: This behavior is not a contradiction, but can be explained by the different viscosity of the coating material at cirrus temperatures compared to mixed-phase cloud conditions, i.e., a more liquid-like behavior at higher temperatures, and more solid, glassy-like behavior at cirrus conditions, e.g. Fig. 5 in Charnawskas et al. (2017):

Charnawskas, J. C., Alpert, P. A., Lambe, A. T., Berkemeier, T., O’Brien, R. E., Massoli, P., Onasch, T. B., Shiraiwa, M., Moffet, R. C., Gilles, M. K., Davidovits, P., Worsnop, D. R., and Knopf, D. A.: Condensed-phase biogenic–anthropogenic interactions with implications for cold cloud formation, Faraday Discuss., 200, 165-194, 10.1039/C7FD00010C, 2017.

P6, sect. 2.3:

I really liked this section, a very careful analysis of how to derive the total number of ice crystals. I think it would also be useful to describe the expected hygroscopic behavior of the pSSA - I have already mentioned this in my comment on P3, line 77ff (i.e., that one would expect a small water uptake already at low RH and the full deliquescence when finally the remaining NaCl fraction is dissolved). Have you also looked at the smaller OPC size channels in Fig. 3a to see if you can detect the full deliquescence step of the pSSA particles there, like e.g. in Figs. 2 & 3 of Kong et al. 2018:

Kong, X. R., Wolf, M. J., Roesch, M., Thomson, E. S., Bartels-Rausch, T., Alpert, P. A., Ammann, M., Prisle, N. L., and Cziczo, D. J.: A continuous flow diffusion chamber study of sea salt particles acting as cloud nuclei: deliquescence and ice nucleation, Tellus B, 70, 10.1080/16000889.2018.1463806, 2018.

And what do you think is most likely the heterogeneous ice nucleation mode seen in Fig. 3b?

P8, line 274: “were broadly consistent”: For me, this is a rather insufficient comparison with earlier literature data. You should present this in more detail and include earlier data in Fig. 5.

P8, line 279: “procedure was modified”: Yes, indeed, when I looked at Fig. 4b, I wondered about the changed shape of the SSi ramps. Could you explain a little more what the purpose of the changed procedure was?

P8, line 292: You provide here a “fit” to your low-temperature ice nucleation data within the framework of the pore condensation and freezing (PCF) concept, but – as far as I can tell – never come back to this point in the following discussion. As mentioned above, you should really elaborate on what you think the possible heterogeneous nucleation mechanism might be.

P8, line 297: “range of uncertainties”: Could you please add respective error bars on a couple of data points in Figs. 5/6 and Figs. S4/S5?

P9, line 318/319: You might be a bit more specific here and say that in the case of SSA it is the “full” deliquescence RH (because these multi-component salts show a gradual deliquescence behavior, see Tang et al. (1997)).

P9, line 341ff: As noted above, the mention of these open questions comes rather abruptly. You should also address these questions in the introduction and explain how you conducted/planned your experiments to answer these questions.

P10, line 353ff: The final sentence of your abstract reads as a very general conclusion that primary SSA remain efficient INPs even after atmospheric aging. The aging conditions in your experiments predominantly led to the formation of new, less ice-active SMA particles, and only to a lesser extent to a change in the size distribution (and thus composition) of the pSSA particles. Can you rule out atmospheric scenarios (different aging conditions) where a larger amount of condensable material is transferred to the pSSA particles, so that they could form a thicker coating layer of organics, which could then affect the efflorescence behavior and possibly hinder water diffusion?

Technical corrections

P1, line 14: Maybe better state: “homogeneous freezing conditions (xx -xx %RH) have to be reached to freeze 1% of the particles.

P2, line 42: “pore condensation and freezing”

P2, line 65: Delete the point after “mass” and replace the comma by a point after the citation.

P4, line 132: A very nested sentence, could you please rephrase it?

P5, line 161: 1.9 g cm-3

P5, line 163: You might highlight here that Fig. 2 now addresses only the downstream size distribution measurements.

P7, line 246 and 261: What means “IS” filters?

P7, line 262: What do you conclude from the agreement between real SW and ASW? I am rather surprised by this.

P8, line 271: Maybe better: “beginning at 218 K and above”

P9, line 338: One could rephrase this a bit and say that heterogeneous ice formation with activated fractions above 1% only occurred at temperatures of 218 K or below.

P10, line 380: All co-authors contributed

Caption of Fig. S5 in the supplement: Check sentence in line 29, probably delete the first “difference”

Patnaude et al. characterized the formation of ice nucleating particles in primary sea spray aerosols (pSSA), secondary marine aerosols (SMA), and a mixture of oxidized SSA and SMA generated in a marine aerosol reference tank and exposed to hydroxyl radicals in an oxidation flow reactor.  They concluded that heterogenous ice nucleation occurred with pSSA, homogenous ice nucleation occurred with SMA, and that the INP formation potentials of aSSA + SMA and pSSA were similar.

Comments

1. In the abstract (L16-L18), based on results presented later in the paper, the authors state: “Similarities between freezing behaviors of the pSSA and aSSA+SMA at all temperatures suggest atmospheric aging has little effect on the heterogeneous freezing behavior of SSA at these cirrus temperatures and remains dominated by the crystalline salts.” I think this conclusion is flawed the way it is presented. It might be applicable to the specific experimental conditions used in this manuscript, but it seems unlikely to me that it can be generalized to other SSA and SMA. First, I couldn’t find any characterization of organic content in the pSSA generated here. It seems plausible to me that the MART-generated pSSA could be relatively inert simply because it has negligible organic content, whereas pSSA with higher organic content might be more reactive towards OH and therefore experience greater changes in IN formation potential. Second, no concrete evidence of SMA coating the pSSA or aSSA was provided - as the authors pointed out, the specific conditions that were used strongly favored homogenous nucleation of SMA over condensation of SMA onto SSA. They note that “the higher concentration of aSSA+SMA compared to pSSA from natural SW between 200 nm and 1 μm (Figure 2b) suggests some modification of the pSSA” but there is no way to prove that this is associated with gas-phase condensation over the other possible explanations that were provided (L172-L174). If it were due to condensation, the aSSA+SMA size distribution should be shifted towards larger particle sizes rather than a higher concentration of the same size particles. I interpreted Figure 2b to suggest that there is negligible condensed SMA on the SSA. Third, because this paper demonstrates that the IN formation efficiency of SMA is clearly lower than that of SSA, a mixed aSSA+SMA particle should at some threshold SMA:SSA ratio have lower IN formation potential than pure SSA particles. I don’t know what that ratio is, but it doesn’t seem to me that it was reached here. Overall, it seems valid to conclude that SMA doesn’t form INP as efficiently as SSA, and that the IN formation efficiency of these pSSA didn’t change appreciably following OH exposure, but I don’t think any sort of definitive conclusions can be made regarding the IN formation efficiency of mixed SSA/SMA particles or of other pSSA particles that might contain more organics.
2. L114 - If it’s important enough to mention that the “grow light was set to a realistic PAR”, please state what the radiation flux was.
3. L124 – You might consider modifying the section title to something like “pSSA, aSSA + SMA, and SMA generation and characterization”
4. L133 - There are a few commercial and many home-built OFR designs; based on ensuing details in the following sentences and the information provided in Mayer et al. (2020), this is most likely an Aerodyne Potential Aerosol Mass OFR, but this detail should be clarified in the text.
5. L135 – The text states: “The OFR generates […] O₃ and OH radicals […] using two UV lamps at wavelengths λ = 254 nm and λ =184 nm […] at a 90:10 ratio, with 90 % intensity from λ = 254 and 10 % from λ =184 nm.” This is not correct. First, the secondary Hg emission line in ozone-producing UVC lamps is centered at ~185 nm, not 184 nm. Second, and much more important, the ratio of 254 and 185 nm radiation fluxes is not 90:10. As far as I can tell, the OFR was operated with lamp type “GPH436T5L/VH/4P 90/10”, which has doped:fused quartz fractions of 90% and 10%, but emits 185 nm radiation at approximately 0.6% of the intensity of the 254 nm radiation (Rowe et al., 2020).
6. L137 – 7 ppm O₃ does not translate to 4-6 days of equivalent atmospheric OH oxidation. Please specify the corresponding OH exposure, or range of OH exposure(s), and how they were measured or calculated, along with the ambient OH concentration that was assumed to obtain the stated 4-6 days' aging time.
7. L164-L205 – “It is clear […] cannot be discounted.“ - this content seems more appropriate to put in Results & Discussion and/or supplement than in Methods.
8. L171 – I strongly disagree that these factors alone enhance nucleation. While the gas-phase oxidation rate was indeed accelerated with the use of elevated OH concentrations, gas-to-particle condensation rates could in principle also have been increased by using SSA concentrations and increasing the condensation sink. Since pSSA concentrations were only 140 cm^-3, (L306), which is far lower than even seed particle concentrations used in many environmental chamber studies with OH concentrations that are closer to atmospheric OH concentrations, it seems to be me that homogenous nucleation of SMA particles was just as likely (if not more likely) due to the very low pSSA condensation sink than the high oxidant concentrations.
9. L181 - what exactly constitutes a "blank" or "background" experiment in this context? MART air sampled through the OFR with in the absence of pSSA generation? This was not clear to me.
10. L265-L270: “In these experiments […] the cycle was repeated” – this content seems more appropriate to put in Methods. It would also be useful to explain why 10% and 0.1% frozen fractions were chosen as the benchmark conditions for ending and beginning the RH scans in the CFDC.
11. L271/Figure 4: Figure 4a is confusing - I don't understand why heterogenous freezing occurs over the first 3 CFDC scans and homogenous freezing occurs over the last 3 CFDC scans for what I assume is nominally the same pSSA. Also, the “Temperature” axis/label is too vague; if I understand Figure 4 correctly, only T_ow is being shown, whereas T_rw is not. Please update the figure axis/legend to clarify this either way.
12. L278: What exactly is the "expected homogeneous freezing threshold?"
13. L283/Figure 5: why are 1% and 5% frozen fractions shown here, whereas the 10% frozen fraction was used as the RH scan endpoint earlier (Fig. 4)? The presentation of multiple frozen fraction values without explanation for the underlying reasons comes across as confusing/arbitrary.

References

Rowe, J. P., Lambe, A. T., and Brune, W. H.: Technical Note: Effect of varying the λ = 185 and 254 nm photon flux ratio on radical generation in oxidation flow reactors, Atmos. Chem. Phys., 20, 13417–13424, https://doi.org/10.5194/acp-20-13417-2020, 2020.