the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Development and characterization of the Portable Ice Nucleation Chamber 2 (PINCii)
Dimitri Castarède
Yusheng Wu
Zamin A. Kanji
Markus Hartmann
Lauri Ahonen
Merete Bilde
Markku Kulmala
Tuukka Petäjä
Jan B. C. Pettersson
Berko Sierau
Olaf Stetzer
Frank Stratmann
Birgitta Svenningsson
Erik Swietlicki
Quynh Thu Nguyen
Jonathan Duplissy
Abstract. The Portable Ice Nucleation Chamber 2 (PINCii) is a newly developed continuous flow diffusion chamber (CFDC) for measuring ice nucleating particles (INPs). PINCii is a vertically-oriented parallel plate CFDC that has been engineered to improve upon limitations of previous generations of portable CFDCs. This work presents a detailed description of the PINCii instrument and the upgrades that make it unique compared to other operational CFDCs. The PINCii design offers several possibilities for improved INP measurements. Notably, a specific icing procedure results in low background particle counts, which demonstrates the potential for PINCii to measure INPs in very low concentrations. High spatial resolution walltemperature mapping enables the identification of temperature inhomogeneities on the chamber walls. This feature is used to introduce and discuss a new method to analyze CFDC data. A temperature gradient can be maintained throughout the evaporation section in addition to the main chamber, which enables PINCii to be used to study droplet activation processes or to extend ice crystal growth. A series of both liquid droplet activation and ice nucleation experiments were conducted at temperature and saturation conditions that span the spectrum of PINCii’s operational conditions (−50 ≤ temperature ≤ −15 °C and 100 ≤ relative humidity with respect to ice ≤ 160 %) to demonstrate the capabilities of PINCii. In addition, typical sources of uncertainty in CFDCs, including particle background, particle loss, and variations in aerosol lamina temperature and relative humidity, are quantified and discussed for PINCii.
Dimitri Castarède et al.
Status: final response (author comments only)
-
RC1: 'Comment on amt-2023-51', Anonymous Referee #2, 26 Apr 2023
The comment was uploaded in the form of a supplement: https://amt.copernicus.org/preprints/amt-2023-51/amt-2023-51-RC1-supplement.pdf
-
CC1: 'Comment on amt-2023-51', G. Kulkarni, 26 Apr 2023
The paper by Castarède et al. (2023) is a nice addition of ice nucleation chamber to the community. Our recent chamber (Kulkarni et al. 2020; see below the reference) also employs long evaporation (nucleation) section ( 0.45m), and we found that this feature evaporates supercooled droplets very efficiently. Figure 5 in Castarède et al. (2023) paper indicates droplet formation at RHw= 100%. These observations are also reflected in the droplet breakthrough section (3.2) and Figure 7. Just wondering and curious, as the evaporation section is long (0.43m) in PINCii, does the evaporation section evaporate the droplets? What are the evaporation section conditions (T and RH) while performing measurements shown in Figures 5 and 7? In our study, we maintain the evaporation section at RHice = 100% and isothermal experimental temperature conditions.
Reference:
Kulkarni, G., Hiranuma, N., Möhler, O., Höhler, K., China, S., Cziczo, D. J., and DeMott, P. J.: A new method for operating a continuous-flow diffusion chamber to investigate immersion freezing: assessment and performance study, Atmos. Meas. Tech., 13, 6631–6643, https://doi.org/10.5194/amt-13-6631-2020, 2020.
Citation: https://doi.org/10.5194/amt-2023-51-CC1 -
RC2: 'Comment on amt-2023-51', Paul DeMott, 27 Apr 2023
General Comments
This is a fine paper about very good technical improvements of an ice nucleation device. It is excellent to have the new device documented in such detail, making it both useful for a variety of studies and potentially reproducible. I will say upfront (the authors may disagree) that I find some of the purported unique features of the instrument to be overplayed in places, in consideration of what is already in the literature for CFDCs that are fully as portable as this one and have even been operated in extended design and with low flow rates to approach lower operational temperatures. It is sometimes difficult to understand what is new for PINC versus what is new for all operational CFDCs. Otherwise, the paper is very well-written and I have mostly a number of minor questions, comments and editorial suggestions listed below. My recommendation is that this paper could be modestly revised and be fully acceptable for publication.
Specific Comments
Abstract
I did not understand the "new method to analyze CFDC data." Can you clarify? is it meant that the evaporation section can have a temperature change or not and thus permit monitoring droplet activation and allow additional ice growth at low temperatures? I do not see this as a new development based on references given in the paper and below.
Introduction
Lines 51-52: I never imagine uncertainties being low enough in a CFDC to serve as a CCNc, and this paper does not give me confidence that it is so (see specific comment later). Hence, I question even making this assertion without proof.
Instrument design and operation
Line 109: Please understand that I am not quibbling about the comprehensiveness of this list of instruments. Rather I want to note that aircraft CFDCs are also "portable", as are most IN instruments. The CSU CFDC has been flown in two basic configurations. In DeMott et al., (2003), a cascade refrigeration system was used, which then also transported for use in a mountaintop laboratory (Richardson et al., 2007). In later years (e.g., Barry et al., 2021), single stage refrigeration compressors have been used for mixed phase conditions. These instruments are portable in nature, being used in different lab and field scenarios on the ground or on ships as well (e.g., Knopf et al., 2021). And not to cross fine hairs, but the long version (Patnaude et al., 2021; Kasparoglu et al., 2022) has also previously been deployed to external laboratories for studies (e.g., DeMott et al. (2009). All such instruments are portable for surface-based deployments, even if one is transporting chillers and their fluid. This is in fact how the Handix commercial CFDC is constructed (Bi et al., 2019). I understand what PINCii stands for of course, but the portable distinction is a vague one and so “…operational portable CFDCs…” is really most CFDCs.
Line 122: Here I will offer a minor quibble with the idea that the extended design of PINCii is a unique feature for deployable CFDCs, given my understanding of what can easily be deployed. In fact, the operation of such long-column instruments without using the evaporation section (i.e, continuing at the same wall temperatures) and at low flow rates has nearly fully been motivated by the known slow ice crystal “growth kinetics” at low temperatures (Patnaude et al., 2021; Kasparoglu et al., 2022), elucidated at length in the Richardson studies. It is a desirable feature of any CFDC deployed for low temperature studies, and only a few have featured this capability. “Compared to all but a few CFDCs…”?
Page 6, Section 2.4: Three minor things. What is most important for making this change in reverse flow results? Is it the narrower gap or the account for heat transfer through the ice, or what exactly? Since this is a desirable outcome, and not everyone achieves 1 mm of ice thickness in their current icing protocols, this is important to understand. Secondly, I suggest repeating here or in the figure captions that all calculation shown are for one total flow rate (10 lpm)? Finally, are the vapor pressure relations used for calculations stated anywhere?
Evaluating the chamber performance
Line 168: An NaCl particle size of 200nm must represent a critical supersaturation that would be unresolvable from water saturation. Is that the point (yes based on sentence below, so can consolidate to say so right here)? What if you used 30 nm?
Lines 183-186: It seems odd to have ice formation mentioned here as describing the activation shown in Fig. 5. Would contaminants in the salt be expected to be a major fraction of all CCN? Perhaps I did not understand. But the salient question is why the response is it not a square pulse, if the instrument truly expected to act as a CCN instrument (conjecture made on line 203 that it is possible). I conject that most evidence supports that it cannot act with the resolution of a CCN instrument, struggling to more than define ~water saturation at low temperatures, because conditions across the lamina vary and the aerosols may not stay fully in the lamina. Activation may always be skewed from the true response. Were any experiments done with monodisperse particles? If not, I suggest making it clear when introducing these experiments that they were primarily to show that a liquid activation response starts around 100%.
Line 193: This is a very useful experiment, but related to my previous comment, where are the results toward 105%? Perhaps freezing interfered?
Lines 198-200: While I can see the OPC resolution issue easily resolved (with extra expense of course), should not the inability to resolve the true lamina RH and its uncertainty at these temperatures be mentioned? This was discussed at length in the Richardson work referenced, as it considered thermal differences on the wall, so could be referenced in this regard.
Line 220: Droplet breakthrough concentrations are defined, but what aerosol concentrations entered the instrument?
Lines 221-224: What is the explanation for the temperature dependence of these results. Others have not found such strong limitations at 103% at -20°C. Is it the long growth time for droplets in the PINCii, meaning that this is an important factor in design of such instruments for measurements in different T ranges? Have you investigated predicted droplet sizes such as by using a microphysical model (e.g., as done for ice)?
Line 231: Can you be more explicit about what you mean by absent kinetic limitations? It means that you did not explicitly calculate size effects and assumed equilibrium growth? Just clarifying, as this can of course make a large difference.
Lines 256-257: CFDCs are presumably designed to overcome all but the strongest consumption, depending on aerosol concentration of course.
Lines 258-260: How does the high resolution of temperature measurements enable real-time calculation of lamina conditions? Wouldn't you need a fluids model to do that? In real-time? I mean that I don’t think one can simply assume that the temperatures at one position define the conditions in the lamina at that position. Without reference to the parcel history before that? I am not a fluids expert, so perhaps I am not correct here.
Lines 265-267: This is interesting though the use of RH extremes and their calculation is not totally convincing. And can one assume that the Koop line is necessarily the target, given for example the studies of Schneider et al. (2021)?
Figure 9 and similar plots: At least for me, these require a lot of focus to interpret, as compared to a 2D plot of T vs RH with AF given as contours. I understand it is a preference.
Measurement uncertainties
Lines 314-316: It seems that uneven spots of wall cooling will be a symptom of any coolant delivery system. It might be distinct for any particular system, but it will have an impact on communicating to the aerosol lamina, which requires further analysis (Richardson, 2009).
Background discussion: Greatly appreciated this.
- Line 320: CFDCs “typically” need to be warmed, drained, etc. I say that because in aircraft scenarios with long flight hours, complete melting and evacuation is unrealistic, so commonly icing is redone at the same stable icing temperature as used in the first process. This does melt and refreeze the ice surface.
- Line 323: Can you say something about the volume flow rate of water, or perhaps the volume that is filled and how long the fill occurs? It must matter how fast the air volume is filled. The 1 mm ice thickness is more than I am accustomed to, which is why I ask.
- Line 329-331: The procedure of temperature control after icing is a bit unclear. You immediately change the wall temperatures to -5°C after water is introduced, or after it has drained? Then step-wise cooling to the operating conditions? Both walls? I suspect that many have their own procedures for achieving low backgrounds, so great to understand this clearly to be tested against others.
- Line 334: The outlet is dried. How is that done? Overpressure of air before connecting OPC?
- Line 339: Is a progressive cooling direction of sampling important for limiting background? This is also common with other CFDCs, so would be good to say.
- Line 345: It is great to be open and honest about this, but I hope you are speaking for PINCii users primarily. There are no wider CFDC community users I know who are not practically aware that ice conditions evolve. This is especially important for heterogeneous freezing experiments. It might be great to explore the different knowledge bases out there amongst CFDC users in a workshop scenario.
Lines 365-366: Can you explain this statement better? “On the other hand, the colder temperatures are within the realm of homogeneous freezing, and thus the link between INP and observed ice above water saturation is ill-defined.” It is quite possible for water saturation to be exceeded to stimulate ice nucleation at low temperatures if a particle is sufficiently small and hydrophobic, no? Not every particle will freeze at the homogeneous freezing condition for solutes.
Lines 401-404: I feel that for operations in the mixed phase regime, this might need more analysis. I understand the focus here is on lower temperatures.
Line 411: “inside the chamber” or along the chamber walls? You are not really modeling what is happening in the lamina are you? That is, the calculation is analytical.
Lines 415-417: I see the statement at the end of this paragraph that these calculations do not represent all factors in knowing the RHi uncertainty, but the values stated here based on (I think static, point-to-point measurements) seem extremely low for these low temperature conditions, low even if you were referencing RHw uncertainty. Is the uncertainty in the T measurement itself also factored in? I may be wrong, but I feel that this is not a solvable analytical problem and really requires an approach such as the Monte Carlo approach taken in Richardson (2009), as stated in Richardson et al. (2010).
Conclusions
Lines 442-445: Should you not be able to compare conditions for significant freezing ala a time dependent freezing calculation using Koop et al, say where 1% freezing is expected? I did not understand the significance of the "onset" condition. One might want to look for agreement at a range of freezing fractions for the residence time of the chamber. Why only focus on onset conditions?
Line 446: I think readers would like to understand better how the Rogers (1988) model was modified. Are there new equations to predict the temperature at the interface of the thicker ice wall and the interior of the chamber? That is, what beyond altering the lamina distance is involved? How could others take advantage of this?
Figure A3: I am curious if this figure suggests detection limit issues in dependence on aerosol size at temperatures below -50°C, if a threshold size is used for ice detection? Is ice detection size variable for low T studies?
Editorial notes
Line 79: Mobile sounds like something that propels itself, although I realize it has been used in this field for devices that can be taken outside the laboratory. Transportable?
Line 114: May not need “reproducible” here since the sentence ends with “reproduced.”
Line 116: “most other” existing chambers
Line 207: Suggest that it would be helpful to add at the end of this first sentence “to reduce droplets to unactivated sizes.”
Line 430: “most “previous generations of deployable CFDCs…”
Line 462-463: Suggest to remove “while remaining transportable.” Most any CFDC is transportable.
Line 465: Line 446: Suggestion to end sentence with “At temperature and RH conditions for which the ambient INP concentration exceeds its limit of detection.”
References (not already in paper)
Bi, K., G. R. McMeeking, D. Ding, E. J. T. Levin, P. J. DeMott, D. Zhao, F. Wang, Q. Liu, P. Tian, X. Ma, Y. Chen, M. Huang, H. Zhang, T. Gordon, and P. Chen, 2019: Measurements of ice nucleating particles in Beijing, China. Journal of Geophysical Researh: Atmospheres, 124, 8065–8075, https://doi.org/10.1029/2019JD030609.
DeMott, P. J., K. Sassen, M. Poellot, D. Baumgardner, D. C. Rogers, S. Brooks, A. J. Prenni, and S. M. Kreidenweis, 2003: African dust aerosols as atmospheric ice nuclei. Geophysical Res. Lett., 30, No. 14, 1732, doi:10.1029/2003GL017410.
DeMott, P. J., M. D. Petters, A. J. Prenni, C. M. Carrico, S. M. Kreidenweis, J. L. Collett, Jr., and H. Moosmüller, 2009: Ice nucleation behavior of biomass combustion particles at cirrus temperatures, Journal of Geophysical Researh: Atmospheres,, 114, D16205, doi:10.1029/2009JD012036.
Kasparoglu, S., Perkins, R., Ziemann, P. J., DeMott, P. J., Kreidenweis, S. M., Finewax, Z., Deming, B. L., DeVault, M. P. and Petters, M. D. (2022). Experimental determination of the relationship between organic aerosol viscosity and ice nucleation at upper free tropospheric conditions. Journal of Geophysical Researh: Atmospheres,, 127, e2021JD036296. https://doi.org/10.1029/2021JD036296.
Schneider, J., Höhler, K., Wagner, R., Saathoff, H., Schnaiter, M., Schorr, T., Steinke, I., Benz, S., Baumgartner, M., Rolf, C., Krämer, M., Leisner, T. and Möhler, O.: High homogeneous freezing onsets of sulfuric acid aerosol at cirrus temperatures, Atmos. Chem. Phys., 21(18), 14403–14425, doi:10.5194/acp-21-14403-2021, 2021.
Citation: https://doi.org/10.5194/amt-2023-51-RC2
Dimitri Castarède et al.
Dimitri Castarède et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 359 | 86 | 10 | 455 | 4 | 3 |
- HTML: 359
- PDF: 86
- XML: 10
- Total: 455
- BibTeX: 4
- EndNote: 3
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1