the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 19 May 2025
An improved Freezing Ice Nucleation Detection Analyzer (FINDA) for droplet immersion freezing measurement
Abstract. Heterogeneous ice nucleation initiated by atmospheric ice-nucleating particles (INPs) is a key microphysical process for cloud formation. Detecting the ice nucleation ability (INA) and concentration of INPs is essential for improving global climate models. Droplet freezing techniques (DFTs) are among the widely used tools for measuring the immersion freezing of INPs, which is a predominant ice nucleation process in mixed-phase clouds. To enhance the efficiency and accuracy of DFTs, we developed a Freezing Ice Nucleation Detection Analyzer at Westlake University (FINDA-WLU) with an improved hardware setup, user-friendly software, precise droplet freezing detection, and rigorous temperature calibrations. The temperature uncertainty of FINDA-WLU is about ±0.60 °C, considering both vertical heat transfer efficiency and horizontal temperature heterogeneity. The system is tested with Milli-Q ultrapure water and reference materials like Arizona Test Dust and Snomax®, and the results are consistent with previous studies. We also use the FINDA-WLU to measure INPs in precipitation samples collected in China. Overall, FINDA-WLU proved to be a reliable and precise method for measuring INA and INP concentrations.
- Preprint
(9589 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 24 Jun 2025)
-
RC1: 'Comment on egusphere-2025-1873', Anonymous Referee #1, 09 Jun 2025
reply
This manuscript presents an advancement in droplet freezing techniques (DFTs) for measuring ice-nucleating particles (INPs) via immersion freezing. The development of FINDA-WLU addresses uncertainties in temperature control, detection accuracy, and operational efficiency. While the study is methodologically sound and provides validation data, several aspects require clarification to establish the novelty and reliability of the instrument. Below are detailed comments and suggestions for improvement.
- The authors claim improvements in hardware, software, and temperature calibration, but the specific innovations need explicit articulation. Compared to prior FINDA designs, FINDA-WLU achieves ±0.60°C uncertainty. However, the manuscript should clarify how the heat transfer efficiency (vertical) and temperature homogeneity (horizontal) were optimized. While "user-friendly software" is mentioned, details on real-time monitoring, automated droplet tracking, or data processing algorithms are lacking.
- The study asserts high precision but omits comparisons with other DFTs (e.g., number of droplets processed per run, false-positive rates). It is recommended to contrast FINDA-WLU directly with existing DFTs in a table, highlighting metrics like droplet capacity, temperature resolution, and uncertainty.
- Fig. 7 reveals horizontal temperature gradients on the cold stage. While common in DFTs, this issue significantly impacts INP quantification, as ±0.6°C uncertainty may affect INP concentrations to a large extent. How do these gradients affect the freezing temperature statistics (e.g., broadening of spectra)? The original FINDA used dynamic infrared imaging for calibration; FINDA-WLU's "rigorous temperature calibrations" and final INP concentration calibration require elaboration (e.g., correction algorithms).
- Fig. 8 shows Milli-Q water freezing at −22°C to −24°C, differing from the listed literature (−13, −14°C). It should specify droplet volumes (not mentioned) and compare with previous studies. It is recommended to test water with documented ultrapure standards and add comparisons to ≥3 DFT studies, especially for studies using similar droplet volumes and numbers, and temperature ranges.
- High-concentration dust suspensions (e.g., −2°C onset) in Fig. 9 likely do not reflect atmospheric conditions (typical INP onset: <−15°C). In high concentration suspensions, multiple INPs compete, altering freezing kinetics. Which curves are similar to real atmospheric conditions? Precipitation samples are mentioned but not linked to dust or biological results. Do these samples exhibit similar freezing behavior?
- It is recommended to include error bars in INP spectra, e.g., Figs. 8–9, to reflect uncertainty.
Citation: https://doi.org/10.5194/egusphere-2025-1873-RC1 -
RC2: 'Comment on egusphere-2025-1873', Anonymous Referee #2, 16 Jun 2025
reply
General comments:
This paper describes an improved droplet freezing instrument and gives examples of its performance. The authors took care to account and correct for temperature gradients within the plate by applying a rigorous temperature calibration. Like this, they achieve a temperature uncertainty of ± 0.6°C. Moreover, they developed user-friendly software with automatic freezing detection.
To achieve the low temperature uncertainty of ± 0.6°C, the temperature of each individual well is measured with an infrared camera. These measurements show a temperature increase in two steps due to the heat release during freezing over a temperature decrease of the ethanol bath by about 1°C. The authors assign ice nucleation to the first heat release without explaining why. Yet, to achieve the high accuracy of ± 0.6°C, the correct detection of the instance of ice nucleation is crucial. If the exact instance of ice nucleation is not identified, this will add to the temperature uncertainty.
To account for horizontal temperature differences within the well plate, a temperature correction for each well is performed. Such a correction requires that the wells’ temperature deviations from the average of the thermocouples is highly reproducible. The authors need to evaluate this potential contribution to temperature uncertainty.
Moreover, the freezing temperature measured for pure water should be compared to additional instruments.
Overall, the manuscript is well written except for the introduction. Here, the strength and weaknesses of the different immersion freezing setups are not enough pointed out and discussed. The state of the art of freezing instruments does not discriminate sufficiently between different types of setups in terms of temperature range that is accessible and how the covered sample volume depends on the droplet preparation technique. Moreover, the references given in the introduction are not sufficiently balanced (see specific comments).
Specific comments:
In the title, the abstract, and in the text, the impression is given that the FINDA-WLU is based on a previous design that has been improved. Yet, no reference to the previous design is given. Please explain.
Lines 50–54: The references given for in-situ methods and laminar flow reactors are not sufficiently balanced and seem to have a bias to references from authors of the manuscript. Specific examples of ice nucleation chambers, laminar flow reactors, and droplet freezing devices should be given together with appropriate references. See Miller et al., 2021 for an overview of instruments.
Lines 61–64: “However, ice nucleation chambers and reactors are typically expensive and have higher detection limitations compared to DFTs, especially at higher temperatures (𝑇 > –20℃) where increased background noise caused by ice residues falling from chamber walls or counting statistics of low ice crystal numbers makes detecting INPs with low concentrations challenging.” What is meant by "higher detection limitations”? To my knowledge, ice-nucleation chambers do not have a problem with falling ice. Please give references for this statement.
Lines 65–69: The references cited here are mostly about measurement campaigns and do not give detailed instrument descriptions. Moreover, they are all given as one list. Instead, they should be split up into microliter and picoliter setups, and into microfluidic devices and instruments working with well plates. References about measurement campaigns need to be replaced by references describing the instrument setup.
Lines 71–74: “Typically, the sampling time, droplet volume, and aerosol suspension concentration can be adjusted, which affects the particle number within each droplet and, thereby, its freezing ability. For example, particle numbers within a droplet can be enhanced by extending the aerosol sampling time, enlarging the droplet size, or reducing the dilution ratio of aerosol suspensions with water.”: The possibilities of adjustment that are pointed out here are typically small, because most instruments can work only in a narrow volume range (within less than an order of magnitude). Variations in sampling time are also within a quite narrow range. Droplet experiments are usually performed with a cooling rate of 1 K/min because at higher cooling rates the temperature accuracy decreases and experiments at lower cooling rates become time consuming. The authors need to demonstrate the volume and cooling rate range that they can cover with their setup.
Lines 75–77: “In this way, this approach enables the quantification of low-concentration INP species in the atmosphere, which overcomes the high detection limitations of ice nucleation chambers. Due to these advantages, DFTs are widely used in current ice nucleation studies.” DFTs operating with well plates are widely used because they are rather cheap and easy to use. Instruments working with smaller volumes like microfluidic devices and continuous flow diffusion chambers are complementary to well plate setups because they can monitor ice nucleation down to the homogeneous freezing threshold while setups with well plates only deliver results down to temperatures where freezing on “pure water” impurities sets in, which is well above the homogeneous freezing threshold. The limitations of the FINDA setup should be pointed out clearly. The temperature ranges covered with the different setups should be discussed.
Lines 101–102: “FINDA-WLU detects LED light reflected by freezing of water droplets placed in a 96-well PCR plate over time.” Sentence needs to be improved.
Lines 103–104: “The camera is fixed above the PCR wells region using an adjustable camera zoom lens (12-120 mm Focal Length, Qiyun Photoelectric Co., China).” Sentence structure needs to be improved.
Lines 114–116: “These sensors are embedded and sealed within thermally conductive epoxy (Omegabond 200, Omega Engineering, Inc., USA) within tubes cut from a PCR plate, ensuring consistent heat transfer between the PCR plate and Pt100 sensors.” How are the tubes cut? Does this mean that the commercial plates are modified?
Lines 153–155: Figure 3 shows an increase in grayscale not a decrease. Please revise the text accordingly. Moreover, Fig. 3 shows that a large change in grayscale (by about 80) is always preceded by a smaller change by around 20 at about 1 K higher temperature. What makes you sure that the second larger change marks ice nucleation and not the smaller one at higher temperature? As the accuracy of the instrument is given as ± 0.6 K, it is important whether the first small or the second larger step marks nucleation. This needs to be investigated and discussed.
Lines 235–236: “This phenomenon also explains why freezing is most often triggered at the droplet bottom from our observation.” What observation do you refer to? Can you observe where freezing starts? Also, the temperature difference of just 1°C between the bottom and the top of the well is not sufficient to trigger freezing always from the bottom, especially when freezing occurs over a large temperature range.
Lines 119–120, line 245, Figure 6: The figure shows that almost 5°C are required until the temperature difference becomes linear. As samples may freeze already at around -5°C, consider to starting the ramp at 5°C so that a good linearity is achieved when temperature reaches subzero temperatures. Just one cooling ramp is shown in Fig. 6. Have the cooling ramps been repeated? What is the reproducibility?
Line 250: David et al. (2019) does not use an aluminium block.
Line 265, Figure 7: how many times has the well calibration experiment been performed? What was the variability between experiments? Has it been performed with different PCR plates? There might be additional variability introduced when the position of the plates within the block has some variability.
Line 333: The method by Agresti and Coull (1998) should be described in some sentences.
Line 346, Figure 9: Can you specify what kind of uncertainty the shaded area shows? Min-max or percentiles? How many times was a measurement repeated?
Line 355: a reference to Fig. A2 in the appendix would be helpful here.
Line 369: References to the “previous studies” should be given.
Line 373, Figure 10: the figure caption needs to be reformulated. Moreover, the references to the studies should be added.
The freezing experiments seem to have been carried out several times as uncertainty ranges are indicated in the figures. It needs to be stated how many times.
Technical comments:
Line 220: “bottom of the” instead of “bottom of”
Line 332: “Fig. 6” should be “Fig. 9”
Line 335: “overlapping” instead of “overlapped”
Line 351: “bacteria” instead of “bacteriuma”
Line 356: “scale” instead of “are scaled”
Line 370: “Caution” instead of “Cautions”
Line 380: “overlapping” instead of “overlapped”
Line 381: “who” instead of “which”
Reference:
Miller, A. J., Brennan, K. P., Mignani, C., Wieder, J., David, R. O., and Borduas-Dedekind, N.: Development of the drop Freezing Ice Nuclei Counter (FINC), intercomparison of droplet freezing techniques, and use of soluble lignin as an atmospheric ice nucleation standard, Atmos. Meas. Tech., 14, 3131–3151, https://doi.org/10.5194/amt-14-3131-2021, 2021.
Citation: https://doi.org/10.5194/egusphere-2025-1873-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 160 | 25 | 8 | 193 | 5 | 5 |
- HTML: 160
- PDF: 25
- XML: 8
- Total: 193
- BibTeX: 5
- EndNote: 5
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1