The Effects of Aminium and Ammonium Cations on the Ice Nucleation Activity of K‐Feldspar

Mineral dust is one of the most abundant types of ice nucleating particles in the atmosphere. During atmospheric transport, mineral dust particles can become coated with inorganic and organic solutes, which can impact their ice nucleation activity. Aminium cations formed from amines are one type of organic solute that can coat mineral dust particles in the atmosphere, but their effects on the ice nucleation activity of mineral dust have not been studied. We investigated the effects of primary, secondary, and tertiary aminium cations with methyl and ethyl groups, as well as ammonium cations, on the ice nucleation activity of K‐feldspar, an important type of mineral dust, in the immersion freezing mode at low cation concentrations (0.2–20 mM) using a droplet‐freezing apparatus. Ammonium cations substantially increased the ice nucleation activity of K‐feldspar, consistent with previous studies. In contrast, primary aminium cations significantly reduced K‐feldspar ice nucleation activity, and secondary and tertiary aminium cations had no significant effect (the effect was less than the uncertainty of our measurements). Our combined results are consistent with the following mechanisms: ammonium cations undergo ion exchange with K‐feldspar, providing exposed N–H groups for hydrogen bonding with ice; primary aminium cations undergo ion exchange with K‐feldspar, exposing a hydrophobic tail that is not effective at nucleating ice; secondary and tertiary aminium cations do not undergo ion exchange with K‐feldspar due to steric effects caused by the multiple hydrophobic groups on the cation.

Mineral dust particles have a lifetime in the atmosphere of up to several days (Huneeus et al., 2011), and can be transported over thousands of kilometers (Uno et al., 2009).During transport, mineral dust particles can be coated with inorganic and organic solutes (Tang et al., 2016;Usher et al., 2003).In the immersion mode, these solutes can decrease the ice nucleation activity of mineral dust particles by decreasing the water activity in the liquid droplets (i.e., freezing point depression) (Koop et al., 2000;Koop & Zobrist, 2009;Zobrist et al., 2008).Solutes can also affect the ice nucleation activity of mineral dust particles in the immersion mode by interacting with and/or modifying their surfaces.At low solute concentrations (<0.1 M), which are relevant for mixed-phase clouds, freezing point depression is small, but the latter mechanism can still have a large impact on the ice nucleation activity of mineral dust particles (Klumpp et al., 2022;Kumar et al., 2018Kumar et al., , 2019aKumar et al., , 2019b;;Reischel & Vali, 1975;Whale et al., 2018;Yun et al., 2020Yun et al., , 2021)).For example, laboratory studies have shown that ammonium salts at low concentrations can increase the freezing temperature of K-feldspar by up to several degrees (Kumar et al., 2018;Reischel & Vali, 1975;Whale, 2022;Whale et al., 2018;Worthy et al., 2021).This enhancement is especially notable since many other solutes at low concentrations, such as acids (Burkert-Kohn et al., 2017;Kumar et al., 2018;Yun et al., 2021) and alkali metal salts (Reischel & Vali, 1975;Whale et al., 2018;Yun et al., 2020), decrease the ice nucleation activity of K-feldspar.The mechanism for the ice nucleation activity enhancement by ammonium salts remains unclear, although several mechanisms have been suggested (Kumar et al., 2018;Whale, 2022;Whale et al., 2018).Amines, nitrogen containing organic bases, are abundant in the atmosphere.Amines can have concentrations up to 14-23% of that of ammonia, the most abundant alkali gas in the atmosphere (Sorooshian et al., 2008).In the atmosphere, amines can react with acids to form aminium salts, which have been identified in aerosol particles as well as cloud droplets (Ge et al., 2011;Lee & Wexler, 2013;Qiu & Zhang, 2013;Shen et al., 2023).The molar ratio of total aminium cations to ammonium cations in the atmosphere can range from 0.002 to 1.9 depending on particle size, season, and location (Liu et al., 2018;Place et al., 2017;VandenBoer et al., 2011;Youn et al., 2015).Methyl and ethyl aminium salts, including their primary, secondary, and tertiary forms, are the most common in the atmosphere (Facchini et al., 2008;Ge et al., 2011).
It has been shown that aminium salts contribute significantly to new particle formation and growth in the atmosphere (Smith et al., 2010).Aminium salts may also influence the volatility, hygroscopicity, and cloud condensation nucleation activities of atmospheric aerosols (Qiu & Zhang, 2013).In addition, aminium salts may form coatings on mineral dust particles and modify their ice nucleation activities.However, the effects of aminium salts on the ice nucleation activities of mineral dust particles have not previously been studied.
To better understand the effects of aminium salts on the ice nucleation activities of mineral dust in the atmosphere, we measured the ice nucleation activity of K-feldspar in the presence of aminium salts in the immersion freezing mode using a droplet-freezing apparatus (Guangzhou Institute of Geochemistry Ice Nucleation Apparatus, GIGINA) (Chen et al., 2023).The cation concentrations were 0.2, 2, and 20 mM, which are relevant for mixed-phase clouds (Aleksic et al., 2009;Herrmann, 2003;Li et al., 2017).By holding the anion (i.e., Cl − in our study) constant, we isolated the effects of the aminium cations on the ice nucleation activity of K-feldspar.Aminium cations with one, two, and three organic functional groups (referred to as primary, secondary, and tertiary aminium cations) were studied.We compared our results for the aminium cations with those for ammonium cations to better understand the mechanisms of how ammonium and aminium cations affect the ice nucleation activity of K-feldspar.In addition, we investigated the effects of ammonium and aminium cations on the dissolution of K-feldspar particles in attempt to better understand the observed changes in the ice nucleation activity of K-feldspar in the presence of these cations.

Chemicals and Sample Preparation
The K-feldspar sample used in this work was provided by the National Research Center of Testing Techniques for Building Materials (NRCTTBM, Beijing, China) and has been used in previous studies (Chen et al., 2020(Chen et al., , 2023)).The Brunauer-Emmett-Teller (BET) surface area of the K-feldspar particles was measured to be 3.96 ± 0.01 m 2 /g using a porosimetry analyzer (ASAP 2020 PLUS, Micromeritics, USA), and the average particle size was measured to be 8.25 ± 1.57 μm using a dynamic light scattering instrument (JL-1177, Jingxin Powder Technologies Inc., China).The main minerals identified in the K-feldspar via X-ray diffraction analysis were microcline (70%), albite (26%), and quartz (4%) (Chen et al., 2020).
Table 1 provides information on the ammonium and aminium salts used in our work.Salt solutions with cation concentrations of 0.4, 4, and 40 mM were prepared by serial dilution in MilliQ water (18.2MΩ cm, Simplicity UV, Millipore) and stirred for 5 min at 200 rpm to ensure uniformity.K-feldspar suspensions with a concentration of 0.2 wt % were prepared in MilliQ water and sonicated for 15 min to disperse the particles.Equal volumes of salt solution and freshly prepared K-feldspar suspension were mixed in a glass beaker which was sealed with Parafilm to prevent contamination during the exposure, and a magnetic stir plate (HSJ-4, Changzhou Aohua Instrument Co., Ltd., China) was used to stir the mixture at 60 rpm for 12 hr at room temperature in the dark.After the exposure, the mixture was immediately used for ice nucleation measurements and dissolution measurements, as ice nucleation activities of aqueous mixtures may change over time (Polen et al., 2016).

Ice Nucleation Activity Analysis
Ice nucleation activities of K-feldspar suspensions were measured using a droplet-freezing apparatus (Guangzhou Institute of Geochemistry Ice Nucleation Apparatus, GIGINA) as described in Chen et al. (2023).Briefly, a hydrophobic siliconized slide (Marienfeld Inc., Lauda-Königshofen, Germany) was placed on the cooling block of GIGINA, a round aluminum spacer with 90 round compartments was placed onto the glass slide, and 1 μL droplets were pipetted into each compartment.The spacer was covered using another glass slide to reduce the Wegener-Bergeron-Findeisen process (Budke & Koop, 2015;Murphy & Koop, 2005).The temperature of the cold stage was first reduced to 0°C with a cooling rate of 10°C/min and then reduced with a cooling rate of 1°C/min until all the droplets were frozen.The uncertainty in temperature measurements was 0.2°C based on measurements of the melting temperatures of tridecane, dodecane, and undecane in comparison with the melting temperatures reported in the literature.During freezing experiments, a CCD camera was used to image the droplets every 30 s (i.e., every 0.5°C).Phase transitions were determined from the brightness of the droplets.
The frozen fraction of droplets, ƒ ice (T), was calculated using Equation 1: where n ice (T) is the number of frozen droplets at a given temperature (T), and n tot is the total number of droplets (90 in our work).The temperature at which 50% of droplets were frozen, T 50 , was calculated from our data by fitting an equation based on the mathematical model of freezing events by Vali (1971) (Figure S1 and Section S1 in Supporting Information S1) (Vali, 1971).A minimum of three replicates were performed for each experimental condition, and from these data we calculated the average fraction frozen at each temperature (every 0.5°C) and an average T 50 with associated uncertainties (two standard deviations).The background level of GIGNA was determined to be similar to other instruments with similar configurations (Chen et al., 2018;Polen et al., 2018) based on freezing measurements of MilliQ water (T 50 = −28.1 ± 1.4°C, Figure S3 in Supporting Information S1).
To further illustrate the effects of these solutes on the ice nucleation activity of K-feldspar, we calculated the change in the median freezing temperature caused by the solutes (ΔT 50 ) using Equation 2: where T 50,solute and T 50,water are the median freezing temperatures of the droplets containing K-feldspar with and without the solutes, respectively.We also calculated ice nucleation active site densities per unit surface area as a function of temperature, n s (T) (cm −2 ) (Section S2 in Supporting Information S1).However, we focused on the ΔT 50 values of samples to illustrate the effect of different solutes on the ice nucleation activity of K-feldspar.
Solutes can decrease the freezing temperatures of INPs by lowering the water activity in the solution (i.e., freezing point depression) (Atkins & de Paula, 2009).We estimated the freezing point depression in our experiments using Equation 3 (Atkins & de Paula, 2009): where K f is the freezing depression constant (1.86°C kg/mol), m is the molality (mol/kg) of the solute, and i is the van't Hoff factor.For the solutes used in our experiments (0.2-20 mM), the calculated freezing point depression (≤0.07°C) was smaller than the uncertainty of our temperature measurements (0.2°C).Hence, no correction was applied to the heterogeneous freezing temperatures to account for freezing point depression.These small freezing point depressions are also consistent with the freezing measurements carried out using ammonium chloride and methylammonium chloride solutions (20 mM) free of K-feldspar (see Figure S3 in Supporting Information S1).

Dissolution Measurements
We investigated the effects of the ammonium and aminium cations on the dissolution of K-feldspar particles, in attempt to better understand the observed changes in the ice nucleation activity of K-feldspar in the presence of aminium cations.Dissolution of K and Al from K-feldspar was measured using Inductively Coupled Plasma Mass Spectrometry (ICP-MS, Agilent 7900) (Zhang et al., 2021).K and Al were chosen because they are main elements of K-feldspar.
After preparing the suspensions containing the solutes and K-feldspar and stirring for 12 hr, as discussed above, a 10 mL aliquot of the suspension was filtered through a 0.22 μm PTFE membrane syringe filter and then acidified with 147 μL HNO 3 (67-70%, v/v) to generate a final HNO 3 concentration of 1% v/v.The acidification step was added to prevent precipitation of the dissolved elements during the analysis.After that, ICP-MS analysis was carried out using Sc as the internal standard.Detection limits in solution were estimated to be around 1 μg/L for the two elements examined.Three measurements were carried out for each solution and average values are reported.ICP-MS analysis allowed us to determine the masses of the two elements in the solution.We report the dissolution fraction (%) of the elements, which corresponds to the ratio of the mass of an element in solution to the mass of the element in K-feldspar particles prior to dissolution.The mass of each element in K-feldspar prior to dissolution was estimated from the mass of K-feldspar particles added to the suspension and the composition of the K-feldspar provided by the supplier.

Freezing in the Presence of Ammonium Cations
Figure 1a illustrates the frozen fraction curves of K-feldspar droplets after exposure to ammonium salts (ammonium sulfate and ammonium chloride) at different cation concentrations (0.2, 2, and 20 mM).At all investigated concentrations, ammonium cations increased the freezing temperatures of K-feldspar droplets but did not have a large effect on the shapes of the freezing curves (i.e., slopes in Figure 1a).The results for ammonium chloride overlap with those for ammonium sulfate, demonstrating that the increase in the ice nucleation activity of K-feldspar was due to the cation (NH 4

+
) and not the anions (Cl − or SO 4

2−
).This observation is consistent with previous studies (Kumar et al., 2018;Whale et al., 2018) which showed that the increase in the ice nucleation activity of K-feldspar by ammonium salts at low concentrations was caused by the cation (NH 4 + ) and was independent of the anions.Our results are also consistent with previous measurements (Klumpp et al., 2022;Yun et al., 2020Yun et al., , 2021) ) that have shown that the ice nucleation activity of K-feldspar is affected by cations rather than anions in almost all cases.
Table 2 includes the median freezing temperatures of the K-feldspar droplets, T 50 , and the changes in median freezing temperature of droplets due to the solutes, ΔT 50 (Equation 2).For the ammonium cations, ΔT 50 was approximately 3.6°C at a cation concentration of 20 mM, illustrating the strong positive effect of ammonium cations on the ice nucleation activity of K-feldspar.
In Figure 2, our results for ammonium cations are compared with previous studies (Kumar et al., 2018;Whale, 2022;Whale et al., 2018;Worthy et al., 2021).Only studies with low concentrations (<100 mM) of ammonium cations are included for better comparison.Our results are consistent with both the magnitude and trend of previous results.The results shown in Figure 2 correspond to measurements using different sources of K-feldspar.The agreement with all data sets suggests the effect of ammonium cation is independent of the K-feldspar source.

Freezing in the Presence of Aminium Cations
Figure 1b shows the frozen fraction curves for K-feldspar droplets containing primary aminium cations with methyl and ethyl groups.Both primary aminium cations decreased the freezing temperature of K-feldspar droplets, and at the highest solute cation concentrations used (20 mM), they also affected the shapes of the curves (i.e., slopes).The ΔT 50 values ranged from −0.3 ± 0.7°C to −8.2 ± 1.3°C for primary aminium cation concentrations of 0.2 and 20 mM, respectively (Table 2).Furthermore, no significant difference was found between methyl and ethyl primary aminium cations (Figure 1b and Table 2).Related to our studies with primary aminium cations, Klumpp et al. (2022) investigated the effect of amino acids on the ice nucleation activity of K-feldspar in the immersion mode.In solution amino acids form zwitterions with a negative -COO − group on one end and a positive -NH 3 + group on the other end, similar to a primary aminium cation, which also contains a positive -NH 3 + group.Klumpp et al. (2022) showed that amino acids decrease the measured heterogeneous freezing onset temperatures of K-feldspar by up to 10°C, which is consistent with our measurements for primary aminium cations.
In contrast to the primary aminium cations, the effects of secondary and tertiary aminium cations on the freezing properties of K-feldspar were small (Figures 1c and 1d).The ΔT 50 values of K-feldspar after exposure to secondary and tertiary aminium cations were ≤1.0°C and in most cases were smaller than the experimental uncertainties (Table 2).Furthermore, we did not observe a clear change in ΔT 50 with an increase in solute concentration for the secondary and tertiary aminium cations (Table 2).We conclude that the effect of secondary and tertiary aminium cations on ice nucleation activity of K-feldspar is small (≤1.0°C), if any.
The ΔT 50 values for all systems studied in our work are summarized in Figure 3.For ammonium cations, ΔT 50 is positive (∼3.6°C at 20 mM), and there is a clear trend of increasing ΔT 50 with increasing cation concentration.For primary aminium cations, ΔT 50 is negative and large (around −8.0°C at 20 mM), and there is a clear trend of decreasing ΔT 50 with increasing cation concentration.On the other hand, for secondary and tertiary aminium cations ΔT 50 is small (≤1.0°C) and no trend is observed with increasing cation concentration is observed.Figure 3 clearly shows that ammonium cations have a different effect on the ice nucleation activity of K-feldspar compared to aminium cations, and primary aminium cations are drastically different from secondary and tertiary aminium cations.

Dissolution of K-Feldspar After Exposure
To determine if the results shown above are related to the dissolution of K-feldspar, we compared the dissolution of K-feldspar in the presence of ammonium cations and aminium cations and in the absence of solutes.Figure 4 and Table S1 in Supporting Information S1 compare dissolution fractions of K and Al in the presence of ammonium cations, primary aminium cations, and secondary aminium cations all at 20 mM with dissolution fractions in the absence of solutes.The dissolution fraction of Al in K-feldspar exposed to ammonium and aminium did not differ from that of K-feldspar exposed to water by more than the uncertainty in the measurements.The dissolution fraction of K was only statistically significantly different in K-feldspar exposed to diethylammonium (a secondary aminium) compared to K-feldspar exposed to water.We also did not observe a strong or statistically significant correlation between the dissolution fractions and the T 50 values (Figure S4 in Supporting Information S1): for K, the correlation coefficient, r, was 0.20 with a p value of 0.56; for Al, r was 0.13 with a p value of 0.80.fractions (p values > 0.05, Figure S5 in Supporting Information S1), indicating that the dissolution fractions did not change significantly with changes in the concentrations of ammonium and primary aminium cations despite the fact that the ice nucleation activity of K-feldspar strongly depends on the concentrations of these cations (Figure 2).
A comparison of all dissolution results (Figures 4 and 5 and Table S1 in Supporting Information S1) with the freezing results suggests that the freezing results cannot be explained by only the dissolution of K-feldspar.However, ice nucleation active sites are preferred locations for ice nucleation with minimum areas on the order of 10-50 nm 2 based on estimates using classical nucleation theory (Kaufmann et al., 2017;Vali, 2014;Vali et al., 2015), and therefore bulk analyses like the one conducted in this study may not reflect specific changes to these rare active sites.The dissolution results also show that K dissolves more than Al, which is consistent with K + undergoing ion exchange with cations in the aqueous solutions (e.g., ammonium cations, aminium cations, or H 3 O + ), while Al does not.See below for further discussion on the possible importance of ion exchange of K + .

Mechanism of Ice Nucleation on K-Feldspar
A mechanism for ice nucleation on K-feldspar surfaces that is consistent with all our freezing results (ammonium cations and primary, secondary, and tertiary aminium cations) involves ion exchange with K-feldspar.K-feldspar consists of AlO 4 − and SiO 4 tetrahedral groups with K + compensating for the negative charge on AlO 4 − in the crystal lattice of microcline.Previous measurements have shown that ion exchange is rapid when K-feldspar is immersed in aqueous solutions (Fenter et al., 2003;Lee et al., 2008).Equation 4represents ion exchange for microcline (Yun et al., 2020): where X + represents the cations in the aqueous solutions (e.g., ammonium cations, aminium cations, or H 3 O + ).
Ammonium cations are similar in size to K + and can readily exchange on the K-feldspar surface (Kumar et al., 2018).The NH 4 + -bearing surface then exposes N-H groups that can hydrogen bond with water and enhance ice nucleation (Figure 6a) (Kumar et al., 2018(Kumar et al., , 2019b;;Whale et al., 2018).For primary aminium cations, the -NH 3 + group of the aminium cation is similar in size to NH 4 + and will likely undergo ion exchange with the K-feldspar surface replacing the parent K + in the K-feldspar structure; in this case, a hydrophobic group will be exposed at the aqueous-mineral surface, which may prevent ice nucleation (Figure 6b).In contrast to primary aminium cations, secondary and tertiary aminium cations likely do not undergo ion exchange due to steric effects caused by the multiple hydrophobic groups bonded to the central N atom of the aminium cations (Figure 6c); as a result, K + may remain part of the K-feldspar lattice or exchange with H 3 O + , and thus the freezing temperature may not change significantly in the presence of secondary and tertiary aminium cations.
Previous studies (Klumpp et al., 2022;Kumar et al., 2018;Whale, 2022;Whale et al., 2018) have suggested that ammonium cations likely modify ice nucleation of K-feldspar by both ion exchange (see discussion above) and adsorption (see discussion below).All of our freezing results (ammonium cations and primary, secondary, and tertiary aminium cations) are not  consistent with an adsorption mechanism (see below); on the other hand, we do not rule out the adsorption mechanism being important for ammonium cations.
The point of zero charge for K-feldspar is around 1-2 (Kosmulski, 2009); hence, the surface of K-feldspar is negatively charged in the presence of ammonium salts.As a result, ammonium cations can adsorb onto the K-feldspar surface, and the ammonium cations will provide N-H groups at the aqueous-mineral interface that can hydrogen bond with water and enhance nucleation of ice (Klumpp et al., 2022;Kumar et al., 2018Kumar et al., , 2019bKumar et al., , 2021;;Pruppacher & Klett, 1997;Whale et al., 2018).Primary aminium cations, like ammonium cations, should adsorb onto the K-feldspar surfaces, exposing a hydrophobic group at the aqueous-mineral interface that can prevent ice nucleation.Secondary and tertiary aminium cations should also adsorb onto the particle surface, possibly even more than primary aminium cations due to their higher hydrophobicity (i.e., hydrophobic effect); adsorbed secondary and tertiary cations should also expose hydrophobic groups at the aqueous-particle interface that can prevent ice nucleation.However, a decrease in the ice nucleation activity of K-feldspar in the presence of secondary and tertiary cations was not observed.Hence, the adsorption mechanism is consistent with the freezing results for ammonium cations and primary aminium cations but not secondary and tertiary aminium cations.Overall, our results provide additional evidence that the ion exchange mechanism is likely important for ice nucleation by K-feldspar.Consistent with this conclusion, the ice nucleation surfaces on K-feldspar identified by Kiselev et al. (2017) likely participate in ion exchange.

Summary and Conclusion
We measured the ice nucleation activity of K-feldspar after exposure to aqueous solutions of ammonium and aminium cations at different cation concentrations (0.2, 2, and 20 mM) for 12 hr using a droplet-freezing apparatus.Consistent with previous studies, we found that ammonium chloride and ammonium sulfate could significantly enhance the ice nucleation activity of K-feldspar, showing a clear trend of increasing ice nucleation activity of K-feldspar with increasing concentration.Conversely, primary aminium cations significantly reduced the ice nucleation activity of K-feldspar, and the degree of inhibition increased with increasing concentration.On the other hand, secondary and tertiary aminium cations did not have a significant effect on the ice nucleation activity of K-feldspar.
Our results are consistent with the following mechanism: ammonium cations undergo exchange with K + in the lattice of K-feldspar; this exchange provides N-H groups that can hydrogen bond with water and facilitate the nucleation of ice structures.The -NH 3 + groups of primary aminium cations also undergo ion exchange with the K-feldspar surface replacing the parent K + in the K-feldspar structure and exposing a hydrophobic group at the aqueous-mineral surface, which inhibits the formation of ice structures.Secondary and tertiary aminium cations are not able to exchange with K + due to steric effects caused by the presence of multiple hydrophobic groups attached to the N atom; as a result, their influence on the ice nucleation activity of K-feldspar is limited.
Our results illustrate the importance of solute type on the immersion freezing of droplets containing K-feldspar particles.Aminium cations are often present in atmospheric particles or cloud droplets (Ge et al., 2011;Liu et al., 2018;Smith et al., 2010;Sorooshian et al., 2008) with methyl and ethyl primary aminium cations considered to be some of the most abundant aminium cations (Ge et al., 2011;Qiu & Zhang, 2013).Our results suggest that a small amount of primary aminium cations could substantially decrease the ice nucleation activity of K-feldspar with possible implications for cloud formation in the atmosphere.
Aminium and ammonium cations are often present in the same atmospheric particles and droplets.In addition, the effects of ammonium and aminium cations may not be additive when present in the same particles and droplets.Hence, future studies are needed to examine the effects of mixtures of ammonium and aminium cations on the ice nucleation activity of K-feldspar (and mineral dust particles in general).Since ammonia concentrations are often higher than amine concentrations in the atmosphere, concentrations of ammonium cations should be higher than concentrations of aminium cations in these mixtures.Furthermore, our current study focused on freezing temperatures ranging from −16 to −10°C, which correspond to rare ice nucleating active sites on the surface of K-feldspar.Further studies are needed to investigate more abundant ice nucleating active sites.In addition, our studies were carried out using mineral dust samples exposed to cations for 12 hr.Additional studies are needed to determine if the results presented here are sensitive to exposure time.

Figure 1 .
Figure 1.Frozen fractions (f ice ) of K-feldspar after exposure to aqueous solutions of (a) ammonium cations, (b) primary aminium cations, (c) secondary aminium cations, and (d) tertiary aminium cations at different cation concentrations (0.2, 2, and 20 mM) for 12 hr as a function of temperature.Frozen fractions of K-feldspar after exposure to MilliQ water for 12 hr are also included for comparison.The error bars of f ice represent two standard deviations (2σ) determined from repeating trials under the same conditions.

Figure 5
Figure 5 and Table S1 in Supporting Information S1 compare dissolution fractions of K and Al as a function of the concentrations of ammonium and primary aminium cations.T 50 values were not correlated with dissolution

Figure 2 .
Figure 2. ΔT 50 as a function of the concentration of ammonium cations reported by our work and previous studies.The data from Kumar et al. (2018) represent the change in the onset freezing temperatures as those are the values reported.The error bars for our work represent two standard deviations (2σ) determined from repeating trials under the same conditions.

Figure 3 .
Figure 3. ΔT 50 of K-feldspar after exposure to aqueous solutions of ammonium and aminium cations and MilliQ water for 12 hr as a function of cation concentration.The gray dashed line represents the ΔT 50 of K-feldspar after exposure to MilliQ water (i.e., 0°C).The shaded region represents the two standard deviations band of ΔT 50 of K-feldspar after exposure to MilliQ water.The results of ammonium represent the results of ammonium chloride.The error bars for our work represent two standard deviations (2σ) determined from repeating trials under the same conditions.

Figure 4 .
Figure 4. Dissolution fractions of K and Al of K-feldspar after exposure to aqueous solutions of 20 mM ammonium cation, primary aminium cations (methyl and ethyl), secondary aminium cations (dimethyl and diethyl), and MilliQ water for 12 hr.The error bars for our work represent two standard deviations (2σ) determined from repeating trials under the same conditions.

Figure 5 .
Figure 5. Dissolution fractions of K and Al of K-feldspar after exposure to aqueous solutions of ammonium and primary aminium cations at different cation concentrations (0, 0.2, 2, and 20 mM) for 12 hr: (a) NH 4 + ; (b) methylammonium; (c) ethylammonium.The error bars for our work represent two standard deviations (2σ) determined from repeating trials under the same conditions.

Figure 6 .
Figure 6.Mechanisms of ice nucleation on K-feldspar consistent with our ammonium and aminium cation results.(a) Ammonium cations may enhance the ice nucleation activity of K-feldspar by ion exchange; (b) primary aminium cations may decrease the ice nucleation activity of K-feldspar particles by ion exchange; (c) secondary and tertiary aminium cations may not interact significantly with K-feldspar surfaces and therefore have no obvious effect on ice nucleation activity.

Table 1
Properties and Sources of Ammonium and Aminium Salts Used in This Work

Table 2 T
50 and ΔT 50 of K-Feldspar After Exposure to Aqueous Solutions of Two Ammonium Salts, Seven Aminium Salts, and MilliQ Water at Different Cation Concentrations (0.2, 2, and 20 mM) for 12 hr