the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 19 Dec 2024
Ice Nucleating Properties of Glassy Organic and Organosulfate Aerosol
Abstract. The role of secondary organic aerosol (SOA) in atmospheric ice nucleation is not well understood, limiting accurate predictions of aerosol-indirect effects in global climate simulations. This article details experiments performed to characterize the ice nucleating properties of proxy SOA. Experimental techniques in conditioning aerosol to glass transition temperatures (Tg) as low as -70 °C using a pre-cooling unit are described. Ice nucleation measurements of proxy organosulfates and citric acid were performed using the SPectrometer for Ice Nucleation (SPIN) operating at conditions relevant to upper tropospheric cirrus temperatures (-45 °C, -40 °C, -35 °C) and ice supersaturations (1.0 < Sice < 1.6). Citric acid was used as a control. Methyl, ethyl, and dodecyl organosulfates did not nucleate ice, despite dodecyl organosulfate possessing a Tg higher than ambient temperature. Citric acid nucleated ice heterogeneously at -45 and -40 °C (1.2 < Sice < 1.4), but required pre-cooling temperatures of -70 °C, notably colder than the lowest published Tg. A kinetic flux model was used to numerically estimate water diffusion timescales to verify experimental observations and predict aerosol phase state. Diffusion modeling showed rapid liquefaction of glassy methyl and ethyl organosulfates due to high hygroscopicity, preventing heterogeneous ice nucleation. The modeling results suggest that citric acid nucleated ice heterogeneously via deposition freezing or immersion freezing after surface liquefaction. We conclude that Tg alone is not sufficient in predicting heterogeneous ice formation for proxy SOA using the SPIN.
Competing interests: Thomas Berkemeier and Jason Surratt are members of the editorial board of Atmospheric Chemistry and Physics.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this preprint. The responsibility to include appropriate place names lies with the authors.- Preprint
(2610 KB) - Metadata XML
-
Supplement
(252 KB) - BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on egusphere-2024-3935', Anonymous Referee #1, 22 Jan 2025
Rapp et al. performed a series of ice nucleation measurements in the cirrus temperature regime for a class of proxy components for organosulfates, which contribute to the atmospheric load of (secondary) organic aerosol particles and may play a role in heterogeneous ice formation. The measurements were carried out using the SPIN continuous flow diffusion chamber (CFDC) and involved pre-conditioning the generated aerosol particles in a custom-built pre-cooling unit to promote their transition to a glassy phase state. Control experiments with ammonium sulfate and ammonium bisulfate particles were used to train a model that uses backscatter measurements to classify the particles as aerosol, water uptake, cloud droplets, and ice. The suitability of the setup to detect a heterogeneous freezing mode was demonstrated by further control experiments with citric acid. Furthermore, a kinetic multi-layer model was used to estimate the timescales for partial and full deliquescence of the particles at SPIN T and RH conditions. I favor the publication of the results in ACP because the manuscript presents new ice nucleation data for proxy components of a potentially relevant type of organic aerosol particles, and because a number of experimental parameters that are crucial for performing such ice nucleation measurements with a CFDC are critically discussed and also useful for the community. However, there are several places in the manuscript where I could not follow or understand the authors' line of thought, and these will need to be addressed in a careful revision.
General comments:
1) In the introduction and throughout the manuscript there are a number of notations used to refer to organic aerosol, e.g. OA, SOA, pure OA, which in my opinion are not always used accurately. For example, in line 54 the authors refer to studies with "certain pure OA" - what do you mean by "pure" here - the studies cited mostly refer to secondary organic aerosol particles. Also, the recommendation by Kasparoglu et al. (2022) mentioned in line 58 refers specifically to "complex secondary organic matter (SOM)" and "dust coated with SOM", but not to "pure and mixed OA" in general as stated in lines 58/59. OA in general could also include crystallizable organic compounds already identified as active in heterogeneous ice formation, such as oxalic acid. The world of OA is therefore a complex zoo of compounds involving crystalline and amorphous solid states - certainly difficult to summarize and do justice to in a short paper introduction, but I think the authors could be a bit more careful in their notation in some places.
2) Line 61ff, reference to Froyd et al. (2010): If I recall correctly, one hypothesis in that paper was that the sulfate component in the particles could be completely neutralized by ammonia, leading to a situation/particle morphology that was later investigated by Schill et al. (2014) and Schill and Tolbert (2013) - see lines 77/78 in the introduction, where crystalline ammonium sulfate could be embedded in an aqueous organic solution and be the actual ice nucleating entity in these particles. It may be useful to point out this potential connection between these studies.
3) Line 69ff, deposition nucleation vs. pore-condensational freezing (PCF), also lines 451-453 in Sect. 5.2: Do the authors really assume that atmospheric freeze-drying took place in their PCU unit, a very complex sequence of events involving droplet activation, freezing, and subsequent sublimation of ice? I cannot really see this happening, especially in the very dry conditions of thermal generation of citric acid aerosol. The sentence in line 451, as it is currently worded, suggests that atmospheric freeze-drying is a mechanism for inducing a glassy phase state, whereas I would rather describe it as a process for changing the morphology of glassy particles from a compact to a porous form. In my opinion, atmospheric freeze-drying is not a prerequisite for the glassy aerosol to become active in heterogeneous ice formation. Why can the solidified, glassy surface itself not provide sites for the "classical" deposition ice nucleation pathway to occur? For example, this was certainly the most likely explanation for the heterogeneous ice nucleation mode of citric acid aerosol observed in the Murray et al. (2010) experiments, which did not involve atmospheric freeze-drying. Of course, atmospheric freeze-drying could be an opportunity to increase the heterogeneous ice nucleation capability by enabling a new ice nucleation mode due to the structural change of the particles, namely PCF. So please check and revise the description of these processes in the manuscript.
4) Fig. 2 and related discussion: You discuss the results for ABS in terms of homogeneous freezing (line 295), which seems reasonable given that ABS droplets are difficult to crystallize and any crystals formed would have a low deliquescence RH. However, in Fig. 2b, the ice nucleation onset at -40 and -45°C appears to be well below Sice = 1.4; in Table S2, there is even one experiment that gives an onset Sice of only 1.31 for -45°C. Are these differences compared to the homogeneous freezing lines depicted in Fig. 2b, which are much higher, still within the experimental uncertainty (in older SPIN-related studies like Wolf et al. (2000) or Li et al. (2024) (https://doi.org/10.1021/acs.est.4c06285) smaller uncertainties for RHice are given), or how do you interpret the results?
I also had some problems understanding the data in Fig. 2 at first - the colored lines (curve-fitting data) are perhaps a bit misleading, as they suggest at first glance that for the AS experiments, for example, we have a transition from aerosol to droplets to ice. But if I understand correctly, only ice forms directly on the aerosol by deposition nucleation. How are the colored lines actually calculated? How do I interpret the fact that, for example, at ice saturation ratios of about 1.25 - 1.30, these lines indicate a depolarization ratio of about 0.2 - 0.3? Is this an average for the whole aerosol population, i.e. aerosols with low and ice crystals with high depolarization ratios, or are there also smaller ice crystals at the beginning of the nucleation process that have a lower depolarization ratio? I appreciate the inclusion of some basic SPIN measurement data, but a slightly more detailed description would be warranted.
5) Line 382/383, Fig. 4: I felt a bit “left alone” with Fig. 4. It contains a long figure caption, but I would expect a bit more detail also in the manuscript text about what the authors want to show with the depicted data. There are two orange lines for full liquefaction – are these representative for the uncertainty of the relevant citric acid parameters (like glass transition temperature, water diffusivity) in the model calculations? And why only citric acid but not dodecyl-OS shows timescales favorable for being able to observe heterogeneous ice nucleation within SPIN?
6) Line 437: Here is again the idea of the atmospheric freeze-drying process, but in which step of your setup did you want to simulate this?
7) Line 447/448: I don’t quite understand the logic here; if droplets are smaller than a certain size threshold for ice, then they could be distinguished from ice by size alone. It is again mentioned in line 530/531: Characterizing water uptake and droplet formation is of course very useful, but how does this interfere with the selective counting of larger ice crystals?
Minor comments/technical corrections:
1) Line 18: Better: “… and ice saturation ratios”
2) Eq. (1), left side, add subscript “org” to indicate the organic mass fraction
3) Line 169: “radius of 0.225” – Table 1 suggests that this is the particle diameter
4) Line 201: Please indicate whether the size is radius or diameter.
5) Line 208: section 2.5
6) Line 238: Maybe: “… using a microcontroller board (Arduino Mega) …”
7) Line 252: Wegener
8) Eq. (4): Please also indicate the scattering angle for the detectors. Shouldn’t P1 and P2 be the same quantity, just measured by two different detectors (but at the same scattering angle, see Garimella et al., 2016)? Why did you use the sum of P1 and P2 in this equation?
9) Line 336: Sect. 2.7.2
10) Fig. 3: Please label the central panel as “Droplet Breakthrough”
11) Line 358: “of various liquid saturation ratios.”
12) Sect. 3.2: Please include also here a reference to Table S2 for the collected results of the organosulfate proxies.
13) Please check the sentence in line 404, I do not really understand it.
14) Line 444: “crystalline” ABS: Are you sure that the particles have crystallized temporarily?
15) Line 533ff: Here is again a mixture of OA and SOA nomenclature, see my first general comment.
Citation: https://doi.org/10.5194/egusphere-2024-3935-RC1 - AC1: 'Reply on RC1', Christopher N. Rapp, 27 Feb 2025
-
RC2: 'Comment on egusphere-2024-3935', Anonymous Referee #2, 27 Jan 2025
This manuscript presented the ice nucleation and water uptake by citric acid and proxy organosulfates at cirrus temperatures. Combination of a pre-cooling unit and SPIN, the authors investigated the potential impacts of phase state on the ice nucleation of the investigated compounds. A kinetic flux model was used to estimate the water diffusion to predict the particle phase state. The topic of this study fits the scope of this journal. This manuscript is well written. There are several issues need to be addressed before publication.
Major comments:
- L451-453, the PCU is used as pre-cooling device, but it does not necessary induce the freeze-drying processes. The freeze-drying processes require the particles should become droplets and freeze first, then sublimination after that. If I understood it right, in the PCU, the RH is very low and the RH was not controlled. It is likely that the freeze-drying processes were not occurred in the PCU.
- For methyl and ethyl-OS, the Tg, org is below -80C, which is lower than the -70C in PCU, so it is likely not in glassy state during the ice nucleation experiments. It would be great if the author can show or calcualte the Tg,w for dodecyl-OS before water uptake or droplet formation at -45 and -35C to further support the conclusion.
- I feel that the title is too generalized and it would be better to be more specific to the investigated compounds.
- Section 2.7.2, the particle classification seems a bit unclear and not easy to understand, especially in L310-317.
Other comments:
- L108-112, Please list the purities of the chemicals.
- L143-148, What are the estimated results of Tg,org for the investigated compounds based on this method? It is suggested to list the values in the text for OS.
- Table 1, what is the uncertainty of onset temperature? The two digital number after decimal point is unnecessary. For the first two lines in the Table 1, some of the information do not match to those in Table S2, for citric acid or citric acid, anhydrous? What are the main differences between these two particles types?
- Table 2, what is the unit? Seconds?
- Fig.3, water uptake panel, why the conditions of the water uptake for the same compound show such large variations, for example citric acid?
- L357, the uncertainty of aw should be 0.025?
- L455-460, how about the total particle surface areas, which may also contribute to the difference in the onset Sice?
Citation: https://doi.org/10.5194/egusphere-2024-3935-RC2 - AC2: 'Reply on RC2', Christopher N. Rapp, 27 Feb 2025
Status: closed
-
RC1: 'Comment on egusphere-2024-3935', Anonymous Referee #1, 22 Jan 2025
Rapp et al. performed a series of ice nucleation measurements in the cirrus temperature regime for a class of proxy components for organosulfates, which contribute to the atmospheric load of (secondary) organic aerosol particles and may play a role in heterogeneous ice formation. The measurements were carried out using the SPIN continuous flow diffusion chamber (CFDC) and involved pre-conditioning the generated aerosol particles in a custom-built pre-cooling unit to promote their transition to a glassy phase state. Control experiments with ammonium sulfate and ammonium bisulfate particles were used to train a model that uses backscatter measurements to classify the particles as aerosol, water uptake, cloud droplets, and ice. The suitability of the setup to detect a heterogeneous freezing mode was demonstrated by further control experiments with citric acid. Furthermore, a kinetic multi-layer model was used to estimate the timescales for partial and full deliquescence of the particles at SPIN T and RH conditions. I favor the publication of the results in ACP because the manuscript presents new ice nucleation data for proxy components of a potentially relevant type of organic aerosol particles, and because a number of experimental parameters that are crucial for performing such ice nucleation measurements with a CFDC are critically discussed and also useful for the community. However, there are several places in the manuscript where I could not follow or understand the authors' line of thought, and these will need to be addressed in a careful revision.
General comments:
1) In the introduction and throughout the manuscript there are a number of notations used to refer to organic aerosol, e.g. OA, SOA, pure OA, which in my opinion are not always used accurately. For example, in line 54 the authors refer to studies with "certain pure OA" - what do you mean by "pure" here - the studies cited mostly refer to secondary organic aerosol particles. Also, the recommendation by Kasparoglu et al. (2022) mentioned in line 58 refers specifically to "complex secondary organic matter (SOM)" and "dust coated with SOM", but not to "pure and mixed OA" in general as stated in lines 58/59. OA in general could also include crystallizable organic compounds already identified as active in heterogeneous ice formation, such as oxalic acid. The world of OA is therefore a complex zoo of compounds involving crystalline and amorphous solid states - certainly difficult to summarize and do justice to in a short paper introduction, but I think the authors could be a bit more careful in their notation in some places.
2) Line 61ff, reference to Froyd et al. (2010): If I recall correctly, one hypothesis in that paper was that the sulfate component in the particles could be completely neutralized by ammonia, leading to a situation/particle morphology that was later investigated by Schill et al. (2014) and Schill and Tolbert (2013) - see lines 77/78 in the introduction, where crystalline ammonium sulfate could be embedded in an aqueous organic solution and be the actual ice nucleating entity in these particles. It may be useful to point out this potential connection between these studies.
3) Line 69ff, deposition nucleation vs. pore-condensational freezing (PCF), also lines 451-453 in Sect. 5.2: Do the authors really assume that atmospheric freeze-drying took place in their PCU unit, a very complex sequence of events involving droplet activation, freezing, and subsequent sublimation of ice? I cannot really see this happening, especially in the very dry conditions of thermal generation of citric acid aerosol. The sentence in line 451, as it is currently worded, suggests that atmospheric freeze-drying is a mechanism for inducing a glassy phase state, whereas I would rather describe it as a process for changing the morphology of glassy particles from a compact to a porous form. In my opinion, atmospheric freeze-drying is not a prerequisite for the glassy aerosol to become active in heterogeneous ice formation. Why can the solidified, glassy surface itself not provide sites for the "classical" deposition ice nucleation pathway to occur? For example, this was certainly the most likely explanation for the heterogeneous ice nucleation mode of citric acid aerosol observed in the Murray et al. (2010) experiments, which did not involve atmospheric freeze-drying. Of course, atmospheric freeze-drying could be an opportunity to increase the heterogeneous ice nucleation capability by enabling a new ice nucleation mode due to the structural change of the particles, namely PCF. So please check and revise the description of these processes in the manuscript.
4) Fig. 2 and related discussion: You discuss the results for ABS in terms of homogeneous freezing (line 295), which seems reasonable given that ABS droplets are difficult to crystallize and any crystals formed would have a low deliquescence RH. However, in Fig. 2b, the ice nucleation onset at -40 and -45°C appears to be well below Sice = 1.4; in Table S2, there is even one experiment that gives an onset Sice of only 1.31 for -45°C. Are these differences compared to the homogeneous freezing lines depicted in Fig. 2b, which are much higher, still within the experimental uncertainty (in older SPIN-related studies like Wolf et al. (2000) or Li et al. (2024) (https://doi.org/10.1021/acs.est.4c06285) smaller uncertainties for RHice are given), or how do you interpret the results?
I also had some problems understanding the data in Fig. 2 at first - the colored lines (curve-fitting data) are perhaps a bit misleading, as they suggest at first glance that for the AS experiments, for example, we have a transition from aerosol to droplets to ice. But if I understand correctly, only ice forms directly on the aerosol by deposition nucleation. How are the colored lines actually calculated? How do I interpret the fact that, for example, at ice saturation ratios of about 1.25 - 1.30, these lines indicate a depolarization ratio of about 0.2 - 0.3? Is this an average for the whole aerosol population, i.e. aerosols with low and ice crystals with high depolarization ratios, or are there also smaller ice crystals at the beginning of the nucleation process that have a lower depolarization ratio? I appreciate the inclusion of some basic SPIN measurement data, but a slightly more detailed description would be warranted.
5) Line 382/383, Fig. 4: I felt a bit “left alone” with Fig. 4. It contains a long figure caption, but I would expect a bit more detail also in the manuscript text about what the authors want to show with the depicted data. There are two orange lines for full liquefaction – are these representative for the uncertainty of the relevant citric acid parameters (like glass transition temperature, water diffusivity) in the model calculations? And why only citric acid but not dodecyl-OS shows timescales favorable for being able to observe heterogeneous ice nucleation within SPIN?
6) Line 437: Here is again the idea of the atmospheric freeze-drying process, but in which step of your setup did you want to simulate this?
7) Line 447/448: I don’t quite understand the logic here; if droplets are smaller than a certain size threshold for ice, then they could be distinguished from ice by size alone. It is again mentioned in line 530/531: Characterizing water uptake and droplet formation is of course very useful, but how does this interfere with the selective counting of larger ice crystals?
Minor comments/technical corrections:
1) Line 18: Better: “… and ice saturation ratios”
2) Eq. (1), left side, add subscript “org” to indicate the organic mass fraction
3) Line 169: “radius of 0.225” – Table 1 suggests that this is the particle diameter
4) Line 201: Please indicate whether the size is radius or diameter.
5) Line 208: section 2.5
6) Line 238: Maybe: “… using a microcontroller board (Arduino Mega) …”
7) Line 252: Wegener
8) Eq. (4): Please also indicate the scattering angle for the detectors. Shouldn’t P1 and P2 be the same quantity, just measured by two different detectors (but at the same scattering angle, see Garimella et al., 2016)? Why did you use the sum of P1 and P2 in this equation?
9) Line 336: Sect. 2.7.2
10) Fig. 3: Please label the central panel as “Droplet Breakthrough”
11) Line 358: “of various liquid saturation ratios.”
12) Sect. 3.2: Please include also here a reference to Table S2 for the collected results of the organosulfate proxies.
13) Please check the sentence in line 404, I do not really understand it.
14) Line 444: “crystalline” ABS: Are you sure that the particles have crystallized temporarily?
15) Line 533ff: Here is again a mixture of OA and SOA nomenclature, see my first general comment.
Citation: https://doi.org/10.5194/egusphere-2024-3935-RC1 - AC1: 'Reply on RC1', Christopher N. Rapp, 27 Feb 2025
-
RC2: 'Comment on egusphere-2024-3935', Anonymous Referee #2, 27 Jan 2025
This manuscript presented the ice nucleation and water uptake by citric acid and proxy organosulfates at cirrus temperatures. Combination of a pre-cooling unit and SPIN, the authors investigated the potential impacts of phase state on the ice nucleation of the investigated compounds. A kinetic flux model was used to estimate the water diffusion to predict the particle phase state. The topic of this study fits the scope of this journal. This manuscript is well written. There are several issues need to be addressed before publication.
Major comments:
- L451-453, the PCU is used as pre-cooling device, but it does not necessary induce the freeze-drying processes. The freeze-drying processes require the particles should become droplets and freeze first, then sublimination after that. If I understood it right, in the PCU, the RH is very low and the RH was not controlled. It is likely that the freeze-drying processes were not occurred in the PCU.
- For methyl and ethyl-OS, the Tg, org is below -80C, which is lower than the -70C in PCU, so it is likely not in glassy state during the ice nucleation experiments. It would be great if the author can show or calcualte the Tg,w for dodecyl-OS before water uptake or droplet formation at -45 and -35C to further support the conclusion.
- I feel that the title is too generalized and it would be better to be more specific to the investigated compounds.
- Section 2.7.2, the particle classification seems a bit unclear and not easy to understand, especially in L310-317.
Other comments:
- L108-112, Please list the purities of the chemicals.
- L143-148, What are the estimated results of Tg,org for the investigated compounds based on this method? It is suggested to list the values in the text for OS.
- Table 1, what is the uncertainty of onset temperature? The two digital number after decimal point is unnecessary. For the first two lines in the Table 1, some of the information do not match to those in Table S2, for citric acid or citric acid, anhydrous? What are the main differences between these two particles types?
- Table 2, what is the unit? Seconds?
- Fig.3, water uptake panel, why the conditions of the water uptake for the same compound show such large variations, for example citric acid?
- L357, the uncertainty of aw should be 0.025?
- L455-460, how about the total particle surface areas, which may also contribute to the difference in the onset Sice?
Citation: https://doi.org/10.5194/egusphere-2024-3935-RC2 - AC2: 'Reply on RC2', Christopher N. Rapp, 27 Feb 2025
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 220 | 76 | 13 | 309 | 32 | 7 | 11 |
- HTML: 220
- PDF: 76
- XML: 13
- Total: 309
- Supplement: 32
- BibTeX: 7
- EndNote: 11
Viewed (geographical distribution)
| Country | # | Views | % |
|---|---|---|---|
| United States of America | 1 | 137 | 45 |
| Germany | 2 | 58 | 19 |
| China | 3 | 20 | 6 |
| United Kingdom | 4 | 14 | 4 |
| France | 5 | 10 | 3 |
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
- 137