Influence of Lowering Soot‐Water Contact Angle on Ice Nucleation of Ozone‐Aged Soot

Organic‐lean and organic‐rich size‐selected soot particles were exposed to a varying O3 concentration, progressively decreasing the soot‐water contact angle (θ) to study its impact on ice nucleation (IN). The IN ability of fresh and O3‐aged soot between 218 and 233 K was observed while monitoring the particle mass and size distributions. The properties of fresh and O3‐aged bulk organic‐lean soot samples with a low and high O3‐adsorption were characterized for soot‐water θ, chemical composition, functional groups, soot‐water interaction ability and porosity. By retaining the soot porosity between aged and unaged samples, we demonstrate that a decrease in θ after O3‐aging enhances organic‐lean soot IN via pore condensation and freezing. Fresh organic‐rich soot exhibits suppressed homogeneous freezing, but after O3‐aging it freezes within uncertainty of the homogeneous freezing threshold of solution drops, because of increased hydrophilicity.

Fresh soot particles are hydrophobic and thus poor cloud condensation nuclei and ice nucleating particles (INPs) (Friedman et al., 2011;Koehler et al., 2009;Kulkarni et al., 2016).However, hydrometeor residuals have shown an enrichment in soot (Cozic et al., 2008;Hiranuma et al., 2013) and which is attributed to its long atmospheric lifetime (∼1 week) (Liu et al., 2020;Textor et al., 2006).During this time soot particles become more hydrophilic from atmospheric aging processes, and can be incorporated in cloud hydrometeors.Ozonation is one potential aging pathway, given that O 3 is an important trace gas in the troposphere (Kuang et al., 2021;Weia et al., 2001).It has been documented that O 3 -oxidation increases the functional group density on soot surfaces (Browne et al., 2015;Friebel & Mensah, 2019b;Ghio et al., 2020;Han et al., 2016), which facilitates soot-water interaction by decreasing the soot-water contact angle (θ).While most studies have focused on a change in cloud droplet activation of O 3 -aged soot at T > 233 K (Browne et al., 2015;Friebel et al., 2019;Grimonprez et al., 2018;Lambe et al., 2015;Pöschl et al., 2001), less attention has been given to the ice nucleation (IN) ability of ozonated soot.Gorbunov et al. (2001) have speculated enhanced IN of O 3 -aged soot due to oxidation for T > 253 K whereas Jahl et al. (2021) observed suppressed immersion freezing of ozonated biomass burning aerosols (including soot) for T > 240 K, due to secondary organic coating on IN active sites.Some studies reported that ozonated soot shows insignificant IN ability changes for T > 233 K (Brooks et al., 2014;Dymarska et al., 2006;Friedman et al., 2011).However, the role of changing only θ by O 3 -aging on the IN of soot in the cirrus cloud regime (T ≤ 233 K) still remains open.
Soot particles can form ice via pore condensation and freezing (PCF) during which supercooled water condenses in mesopores (2-50 nm width) below water saturation.According to the Kelvin equation, wettable soot particles (θ < 90°) are needed to promote capillary condensation, followed by freezing below the homogeneous nucleation temperature (HNT).Soot particles of higher porosity and surface wettability were reported to be more active INPs (Mahrt et al., 2018;Nichman et al., 2019) and after increasing the porosity of soot particles by compaction or removing organics, ice nucleation by PCF is enhanced (Gao, Friebel, et al., 2022;Gao & Kanji, 2022a, 2022b;Gao, Koch, et al., 2022;Mahrt et al., 2020;Testa et al., 2024).The results also suggest that PCF of an organic-lean soot (PR90) might be limited by its insufficient surface wettability (Gao, Friebel, et al., 2022).David et al. (2020) reported enhanced PCF for silica particles with a lower θ at T ≤ HNT, thus a change in soot-water θ may also influence soot-PCF.In this study, O 3 -aging was used to decrease the soot-water θ of an organic-lean soot (PR90 as used in Gao, Friebel, et al., 2022) without modifying the soot porosity, structure and surface functional groups.Using an organic-lean sample ensures we investigate only the role of θ on soot-PCF at T ≤ 233 K by avoiding any aged organics from O 3 -treatment condensing in the pores or modifying the pore structure.In addition, the IN of an O 3 -aged organic-rich aviation soot proxy (C 3 H 8 flame soot as used in Gao, Koch, et al., 2022) was tested at T ≤ 233 K for its O 3 -aging effects on PCF.

O 3 -Generation
Before feeding the corona discharge O 3 -generator (Figure S1 in Supporting Information S1, A2ZS-1GLAB, A2Z Ozone, Inc. USA), flow-rate controlled synthetic air passed through a diffusion drier followed by a highefficiency particulate air filter and a charcoal denuder at ambient T, to remove humidity, ambient particles and volatile organics.In order to produce a stable O 3 concentration (0, 2, or 20 ppmv detected by an O 3 monitor, BMT 932 Link, V1.07, BMT Messtechnik GmbH, Berlin), an in-house-built diluter was used and the dilution ratio was adjusted manually and monitored continuously while the total flow rate was maintained constant for bulk (4 L min 1 ) or aerosol sample (3 L min 1 ) experiments.

Bulk Sample Experiments
Organic-lean PR90 (Orion Engineered Carbons GmbH, OEC, Frankfurt, Main, Germany) is a commercial and highly oxidized carbon black.Fresh PR90 was used because of bulk sample availability and experiment reproducibility.To simulate short-and long-term O 3 -exposure (termed PR90ˍS and PR90ˍL respectively), a 4.0 L min 1 flow with 20 ppmv O 3 was used to flush bulk samples (2,665 mg PR90_S and 2,647 mg PR90_L) suspended by using a magnetic stirrer in a glass container for 180 s and 4 hr respectively.The exhaust was tested with a zero O 3 concentration and directed to a fiber filter trap to collect soot particles from the aerosol phase.The adsorbed O 3 mass per mg of PR90_S (PR90_L) is approximately 1.5 × 10 4 (1.4 × 10 2 ) mg.
Bulk sample properties, including θ, volatile content, functional groups, soot-water interaction ability and pore size distribution, were characterized to investigate soot O 3 -aging effects.A more detailed experiment description is provided in the supplement (Texts S1 and S2 in Supporting Information S1).(a) Sessile-drop measurements were performed to measure θ and the droplet diameter (D w ) on a soot sample disc.Averaged θ and D w values at 0 ms are reported in Table 1.(b) Thermogravimetric analysis was conducted to measure the soot mass loss as a function of T from 30°C to 1,000°C at a rate of 5°C min 1 in a N 2 atmosphere (Figure 1a).(c) Fourier transform infrared spectroscopy employed with the attenuated total reflection technique was used to detect the functional groups on the soot surface (Figure 1b).(d) Dynamic vapor sorption measurements were performed to measure the soot-water interaction.The results show the sample mass change due to soot-water uptake (loss) under a stepwise increase (decrease) in RH w (relative humidity with respect to water) from 5% to 90% (from 90% to 5%) (Figure 1c).

Aerosol Sample Experiments
The unaged PR90 powder was sufficiently agitated by a mechanical stirrer until the size-selected particle mass reached a plateau (indicating no further compaction), prior to using a Venturi nozzle to disperse the sample in a N 2 carrier gas, that is, volatile free, following the same protocol reported in Gao, Friebel, et al. (2022).Downstream, 200 and 400 nm particles were generated using a differential mobility analyzer (classifier 3080, with a 3081 column and a polonium radiation source, TSI Inc.) with a sample to sheath flow ratio of 1:13 and 1:7.3, respectively.A 1 L min 1 aerosol sample flow was mixed with a 3 L min 1 O 3 flow (0, 2, or 20 ppmv) before feeding into a 10 L continuous-flow reactor (in dark and ambient T conditions) to allow a 2.5 min (hydrodynamic residence time) O 3 -exposure for aerosol samples (PR90, PR90ˍS and PR90ˍL) (Figure S1 in Supporting Information S1).Analogous to the well-known plug flow reactor concept (Friebel & Mensah, 2019a;Kang et al., 2007), our reactor was operated with a continuous steady feed of reactants (soot and O 3 ) and withdrawal of samples (O 3 -aged soot) at equal flow rates.Thus, the aerosol particle concentration, O 3 concentration and particle aging level are in a dynamic equilibrium in the reactor so that the reactor yields a constant output.Assuming firstorder reaction kinetics, the equivalent atmospheric O 3 -aging time of a 2 or 20 ppmv O 3 concentration is approximately ∼2-4 hr or 8-20 hr respectively, considering atmospheric O 3 background levels of 20-45 ppbv (Hough & Derwent, 1990;Vingarzan, 2004).
The particle mass and size distributions were observed using a centrifugal particle mass analyzer (CPMA, Cambustion Ltd., Cambridge, UK) pumped by a condensation particle counter (CPC 3787, TSI Inc.) and a scanning mobility particle sizer (SMPS, Classifier 3082, Column 3081, CPC 3772 or 3776 low-flow mode for 200 or 400 nm particles respectively, TSI Inc.) (Figure S1 in Supporting Information S1).Simultaneously, sizeselected soot particles were subject to controlled RH and T conditions in the horizontal ice nucleation chamber (HINC) (Kanji & Abbatt, 2009;Lacher et al., 2017).The RH i ramp (RH with respect to ice, 2% min 1 ) was performed at 218, 223, 228, and 233 K.The activated fraction, defined as the ratio of soot particles growing to ice crystals >1.0 μm (observed by an optical particle counter, MetOne, GT-526S) to the total number of soot particles (monitored by a CPC 3772), was reported as a function of RH and at a fixed T.

Experiments of an Organic-Rich Aviation Proxy Soot
An organic-rich soot was produced under fuel rich conditions (fuel/air = 1.03) from a C 3 H 8 flame using a miniature combustion aerosol standard (model 4200, Jing Ltd, Zollikofen, Switzerland), termed mBr.The particles were carried by N 2 -gas using the method reported by Gao, Koch, et al. (2022) to exclude the influence of volatiles in the air on downstream O 3 -aging experiments.Using the same flow-reactor and SMPS-CPMA-HINC system described in Section 2.1 (Figure S3 in Supporting Information S1), 200 and 400 nm soot particles were exposed to 0 and 20 ppmv O 3 , thus termed mBr and mBr_L, and then tested for particle mass, size and IN ability (Figure 2).

Bulk Sample Characterization
Compared to PR90, O 3 -exposure decreases the average θ of PR90_S and PR90_L by 4.7°and 15.8°respectively (Table 1), suggesting that a higher O 3 -exposure concentration, that is, equivalent longer O 3 -aging, leads to a nonnegligible progressive decrease in the soot-water θ (beyond the measurement uncertainty).Viegas and Varandas (2012) reported that O 3 molecules may form O 3 -(H 2 O) n (n = 0-4) clusters by hydrogen bonding.Ozonation may increase the soot functional group density by oxidation (Friebel & Mensah, 2019b;Ghio et al., 2020;Han et al., 2016), however, we show that the θ decrease of O 3 -aged PR90 does not result from oxidation (Figure 1b) over the time scales used here but is caused by O 3 -H 2 O binding.In addition, we note that the calculated θ value from Sessile-drop experiment decreases with elapsed time because of droplet soaking in the sample (Figure S2c and Text S2 in Supporting Information S1).Hence, we report θ for the 0 ms time stamp as the most conservative but also representative for initial contact.According to Phillips and Riddiford (1965) 2020), water θ shows a weak increase with decreasing T for the range below saturation temperature.Considering the droplet soaking and temperature dependence as a systematic error, the Sessile-drop results still demonstrate that increasing O 3 -exposure progressively decreases θ for PR90 particles at IN experiment temperatures (218-233 K).The results of D w changes (Table 1) of O 3 -aged samples are consistent with the θ results and detailed discussions are provided in Supplement Text S2.
The sample mass loss at 30°C (∼1% in Figure 1a) results from the water vapor desorption by dry N 2 flushing (Text S1).The slightly larger mass losses for PR90_S and PR90_L compared to PR90 (by ∼0.4%) are attributed to their increased water uptake ability because of O 3 -aging.Between 30°C and 300°C , the mass fraction curves of fresh and both aged samples exhibit a comparable volatile content loss of 1.2% and 1.5%, respectively.The overlapped curves of PR90_S and PR90_L also suggest that a higher O 3 -exposure concentration does not change the sample organic content that would become volatile between 30°C and 300°C.For T > 300°C, the aged samples show larger mass losses than PR90.PR90_L also loses more mass compared to PR90_S with increasing T above 300°C and the mass loss difference reaches the maximum near T = 550°C.The total mass loss of O 3 -aged PR90_S and PR90_L is higher than that of fresh PR90 by 0.32% and 1.80%, respectively, supporting the additional mass loss comes from O 3 -desorption.In general, given that the fresh PR90 sample is already highly graphitized, thermally labile adducts produced by ozonation are unlikely in the aged samples.The higher O 3 addition to PR90_L explains its larger mass loss and lowest θ.
Figure 1b does not show characteristic peaks or bands ascribable to functional groups for PR90, PR90_S or PR90_L (Albert et al., 2011;Coates, 2006), indicating no organics or new functional group formation during O 3 -aging.The aged samples present lower absorbance intensity between 2,700 and 800 cm 1 whereas the fresh PR90 shows stronger absorbance intensity for the rest of the wavenumber range, resulting from the O element enrichment in the aged samples, that is, O 3 -addition.Furthermore, the results suggest the direct O 3 -H 2 O binding (Viegas & Varandas, 2012;Yadav et al., 2017) is responsible for the increased surface wettability of aged samples rather than the increase in covalently bonded functional group density (Table 1).Consistent with our results, Friebel and Mensah (2019b) suggested that the reaction between O 3 and soot follows a Langmuir-type mechanism by which the soot surface mainly adsorbs O 3 molecules during first several minutes whereas the oxidation needs hours.
Figure 1c shows overlapped dynamic vapor sorption isotherms for fresh and aged PR90 samples when RH w < 40%, suggesting they contain comparable active sites for water uptake (Gao, Friebel, et al., 2022;Mahrt et al., 2018).This agrees with the results in Figure 1b showing that such a short O 3 -exposure does not increase functional group active sites.Notably, PR90_S and PR90_L adsorb more water between RH w = 40% and 80%.The O 3 -H 2 O binding possibly aids water molecules adsorbing onto soot surface, which has a higher probability of forming water clusters at higher RH w (Table 1).In addition, soot mesopores induce capillary condensation at RH w > 40%, leading to a hysteresis loop between the adsorption and desorption isotherms.The hysteresis loop range and area of PR90_S and PR90_L are similar to PR90, indicating unchanged mesopore size range and accumulative mesopore volume (Gao, Friebel, et al., 2022).The overall shift of the isotherms of PR90_S and PR90_L to lower RH w conditions (for RH w > 40%) compared to PR90 is due to decreased soot-water θ.
Figure 1d presents the pore-volume based pore size distribution of fresh and aged PR90 samples.These show a similar distribution mode around 8-9 nm demonstrating that O 3 -adsorption does not change the sample mesopore abundance which would be otherwise the case if organics were formed thus reducing available pore volume due to condensation.
In general, O 3 -exposure conducted here is designed to modify soot-water θ of our sample and the short exposure time neither changes the sample functional groups by oxidation nor modifies the sample meso-porosity or surface topography (Figure S2c in Supporting Information S1).The adsorbed O 3 molecules may bind with H 2 O molecules directly and thus decrease the soot-water θ (increase soot wettability) which scales to the O 3 -exposure concentration.

Aerosol Particle Properties
The mass increase for 400 (200) nm PR90_S and PR90_L are approximately 2.9% (5.6%) and 7.1% (7.7%) as shown in Figures 2a and 2b respectively, supporting O 3 -addition increases with O 3 -exposure concentration.PR90_S and PR90_L aerosol samples adsorb more O 3 mass by percent (2.9%-7.7%)compared to the bulk (0.015%-1.4%, Section 2.1.2),suggesting that the measured θ for bulk samples may be an upper estimate for the aerosol samples which have a more effective O 3 adsorption.The mobility diameter increase for 400 nm particles is smaller than 0.8 nm whereas the size of 200 nm particles does not change, implying that O 3 -adsorption does not restructure the soot-aggregate (Fendel et al., 1995;Friebel et al., 2019).Transmission electron microscopy images provided in Figure S9 in the Supporting Information S1 of PR90 show compacted aggregates.The unchanged dynamic shape factor (Figure S10 in Supporting Information S1) also supports negligible morphology changes for PR90 after O 3 -aging.

Ice Nucleation Results
Figure 2c shows that fresh 400 nm PR90 freezes homogeneously at 233 K within uncertainty of S i,hom .For T < 228 K, 400 nm PR90 nucleates ice below S i,hom , which is attributed to PCF owing to soot mesopores and having a θ < 90°(Section 3.1.1).The PCF process is suppressed at 233 K because of the small mesopore volume which limits the homogeneous nucleation rate of supercooled pore-water given a short particle residence time in HINC (∼12s) (Gao, Friebel, et al., 2022).Compared to fresh 400 nm PR90, the onset S i values of 400 nm PR90_S and PR90_L decrease by up to 0.06 and 0.08 respectively (T = 218 K, Figure 2c), with PR90_S still freezing homogeneously at 233 K.In particular, the 400 nm PR90_L shows a decrease in onset S i beyond the measurement uncertainty of PR90 at T ≤ 223 K.The IN enhancement for both O 3 -aged samples is consistent with a decrease in their θ (Section 3.1.1).The larger θ decrease (by 15.8°) for PR90_L than for PR90_S (by 4.7°) is also consistent with the larger IN enhancement.Compared to PR90, PR90_S exhibits a smaller onset S i by 0.06 at 218 K (0.03 at 223 K) whereas PR90_L shows even lower onset S i of 0.08 at 218 K (0.06 at 223 K).This agrees with the Kelvin equation, that smaller θ, requires lower S i trigger capillary condensation and thus PCF (Marcolli, 2014;Marcolli et al., 2021).Given the mesopore abundance is similar between the aged and unaged samples (Figures 1c and 1d), the porosity cannot be the reason for enhanced IN of PR90_S and PR90_L.This further supports that lowering θ alone contributes to the IN enhancement of PR90_S and PR90_L (onset S i decrease up to 0.08 at 218 K).We note that this is indirect evidence of a decrease in θ leading to PCF enhancement, but currently a single particle-based θ measurement technique is not available (Song & Fan, 2021).Similarly, for 200 nm particles, long exposure to O 3 results in better IN ability of the PR90_L than the PR90_S at T < 228 K.Both PR90_S and PR90_L 200 nm particles freeze homogeneously more readily at 233 K compared to PR90.All 200 nm particles have higher onset S i values than 400 nm particles at the same T and O 3 -exposure conditions and follow the size dependence of IN observed previously (Gao, Friebel, et al., 2022;Mahrt et al., 2018).

Aerosol Particle Properties
The mass increase for 400 and 200 nm mBr_L are approximately 12.2% and 4.1%, corresponding to a size increase of ∼9.9 and 7.2 nm respectively, compared to fresh mBr (Figures 2a and 2b).This is comparable to the 3 nm size increase for 100 nm mBR particles reported in Friebel et al. (2019) for a lower O 3 -concentration over a longer time.The size and mass increase result from the oxidation of volatiles and the formation of secondary organics due to ozonolysis (Friebel & Mensah, 2019a;Friebel et al., 2019) of the unsaturated hydrocarbons and aromatics (Gao, Koch, et al., 2022) in the sample.Exposing mBr at a 2 ppmv O 3 -concentration should lead to a smaller increase in particle size and mass than observed for mBr_L (Figures 2a and 2b) as the size of O 3 -aged mBr increases with aging time (Friebel & Mensah, 2019b).Compared to the fresh sample, PR90_L (200 and 400 nm) and 400 nm mBr_L show a small increase in effective density (Text S7), in contrast to 200 nm mBr_L whose effective density decreases.This is because 200 nm mBr_L undergo a relatively larger size increase (3.6%) compared to the other 3 samples (<2.5%).The size increase of 200 nm mBr_L particles with originally more densified structures may be due to organic formation over the surface.Given that the density of organics formed is smaller than that of bulk soot (Khalizov et al., 2012), the density of 200 nm mBr_L particle decreases.However, open-branched 400 nm mBr_L, which is demonstrated by its larger dynamic shape factor compared to 200 nm mBr (Text S8 and Figure S10 in Supporting Information S1), should contain more intra-aggregate voids (Gao, Friebel, et al., 2022) allowing newly formed organics to fill the voids resulting in a small increase in density.

Ice Nucleation Results
O 3 -aging decreases the onset S i of 400 and 200 nm mBr_L to the homogeneous freezing threshold of Koop et al. (2000) (Figures 2c and 2d).The decrease in onset S i is similar to that of PR90_S and PR90_L.However, all PR90 samples freeze by PCF and thus the aged samples show enhanced PCF.The absence of PCF for mBr_L results from secondary organics introduced by ozonolysis which decreases the porosity of the sample (Gao, Koch, et al., 2022)

Atmospheric Implications
The short O 3 -exposure time in this study does not allow sufficient oxidation of the soot particles but rather only adsorption of O 3 (Friebel & Mensah, 2019b).Nevertheless, high concentration O 3 -exposure (2-20 ppmv) used in this study may result in high reaction rates thus simulating an extended aging time in the atmosphere (Ghio et al., 2020;Kang et al., 2007;Simonen et al., 2017).Liu et al. (2020) and Textor et al. (2006) reported soot residence times in the troposphere of up to weeks, resulting in longer O 3 -exposure and chemical reactions on soot surfaces, which not only lowers the soot-water θ but can also decrease the mesopore availability.The latter occurs because the non-volatile reaction products can fill the pore volume (Gao, Koch, et al., 2022) thus limiting soot IN via PCF.If sufficient non-volatile organics form over the particle surface as a coating, glassy organics may form, which promote heterogenous IN of soot (Tian et al., 2022).We specifically target short O 3 -exposures at high concentration to investigate the role of θ alone on organic-lean soot IN.That is, we circumvent any changes in soot particle morphology occurring with longer aging times or different species.Our results imply that organiclean soot exposed in the troposphere, where O 3 concentration is high and the other trace species are low so that their addition does not change the morphology, will show an enhanced water interaction and IN ability in the cirrus regime.Our results of 400 nm organic-lean soot also imply that long-term O 3 -exposure may even promote soot-PCF at 233 K instead of homogeneous freezing.This may be of significance to organic-lean soot (similar to our PR90 sample), from alternative aviation fuels (Braun-Unkhoff et al., 2016) which may form ice via PCF at warmer temperatures (∼233 K) in the cirrus cloud regime after O 3 -exposure.However, organic-rich and nonporous soot, (like our mBr sample) as an aviation soot surrogate (Ess & Vasilatou, 2018;Marhaba et al., 2019), will only freeze homogeneously after O 3 -aging.Our results may also have implications on the case of OH-radical exposure for soot particles (Grimonprez et al., 2018;Lambe et al., 2015) which may enhance soot-PCF if the soot has decreased θ provided that the pore volume remains available for PCF.

Conclusions
This study shows that O 3 -aging deceases the soot-water θ of organic-lean soot, which is atrributed to O 3 -H 2 O binding and enhances PCF by lowering onset S i for T < 233 K, which positively scales with the O 3 -exposure concentration.Notably, the ozonation does not influence the mesopore abundance of organic-lean soot, highlighting the sole importance of soot-water θ in organic-lean soot IN via PCF.Organic-rich soot exhibits promoted homogeneous freezing after O 3 -aging because of increased hydrophilicity but not PCF due to unavailability of relevant mesopores.We note that the current study follows work that studied the effect of changing the mesopore volume alone on soot IN (Gao, Friebel, et al., 2022).Together, these two studies unveil the dependence of soot IN on its porosity and surface wettability independently, with the increase in porosity resulting in larger IN enhancement than that from an increase in the surface wettability.

Figure 1 .
Figure 1.Bulk soot sample experimental results.(a) Soot mass fraction as a function of T from thermogravimetric analysis.Solid and dotted lines indicate two individual measurements respectively.(b) Vector normalized attenuated total reflection absorption spectra from Fourier transform infrared spectroscopy.(c) Percent mass change as a function of RH w from dynamic vapor sorption measurements.The solid lines with closed symbols and dashed lines with open symbols indicate adsorption and desorption isotherms, respectively.(d) Pore size distribution as a function of pore radius derived from N 2 desorption measurements.

Figure 2 .
Figure 2. Soot particle size, mass, effective density and ice nucleation results.Panels (a) and (b) show the log-normal mean particle mass and mobility diameter, and effective density of 400 and 200 nm soot particles respectively, derived from the particle mass and size distribution curves (Figures S5 to S8 in Supporting Information S1 including multiple charged particles).The error bar indicates one standard deviation.In panels (c) and (d), the onset ice saturation (S i ) values as a function of T for 400 and 200 nm soot particles respectively, to reach an activated fraction of 0.001 (0.1%) (Figures S11 to S14 in Supporting Information S1), are shown.For clarity, the symbols are evenly spaced within a range of ±1 K centered at the measurement T. Black and blue dashed lines denote the homogeneous freezing conditions for solution droplets (S i,hom ) at T < homogeneous nucleation temperature according to Koop et al. (2000) and Schneider et al. (2021), respectively.Gray dotted lines represent constant water saturation conditions calculated based on Murphy and Koop (2005).The error bars are the S i uncertainty arising from the T uncertainty (±0.1 K) of the IN chamber.
and Song

Table 1
The Soot-Water θ and Water Droplet Diameter D w at the 0 ms Time Stamp of the Sessile-Drop Measurement Note.The uncertainty (σ) is the standard deviation.
. The increased hydrophilicity of mBr_L (larger activated fraction values than mBr at low RH conditions in Figures S13 and S14 in Supporting Information S1) only facilitates homogeneous freezing but does not result in PCF.Given the poor IN ability of mBr_L, an experiment at low O 3 concentration was not deemed necessary and it is expected to freeze homogeneously.