The onset conditions for ice nucleation on H2SO4 coated, (NH4)2SO4 coated, and uncoated kaolinite particles at temperatures ranging from 233 to 246 K were studied. We define the onset conditions as the relative humidity and temperature at which the first ice nucleation event was observed. Uncoated particles were excellent ice nuclei; the onset relative humidity with respect to ice (RHi) was below 110% at all temperatures studied, consistent with previous measurements. H2SO4 coatings, however, drastically altered the ice nucleating ability of kaolinite particles, increasing the RHi required for ice nucleation by approximately 30%, similar to the recent measurements by Möhler et al. [2008b]. (NH4)2SO4 coated particles were poor ice nuclei at 245 K, but effective ice nuclei at 236 K. The differences between H2SO4 and (NH4)2SO4 coatings may be explained by the deliquescence and efflorescence properties of (NH4)2SO4. These results support the idea that emissions of SO2 and NH3 may influence the ice nucleating properties of mineral dust particles.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
 Ice nucleation can occur in the atmosphere by either homogeneous nucleation or heterogeneous nucleation. Heterogeneous nucleation typically involves solid substrates, which are often called ice nuclei. These ice nuclei have the potential to modify climate by changing the formation conditions and properties of ice and mixed-phase clouds.
 In the following, we investigate the onset conditions for ice nucleation on H2SO4 coated, (NH4)2SO4 coated, and uncoated kaolinite particles at temperatures ranging from 233 to 246 K, a temperature range relevant for the lower troposphere. Here we define the onset conditions as the relative humidity and temperature at which the first ice nucleation event was observed. Kaolinite represents a significant component of mineral dust, comprising approximately 5–10% of aerosolized mineral dust [Glaccum and Prospero, 1980]. This research should be useful when trying to determine if anthropogenic emissions of SO2 and NH3 affect climate by influencing natural ice nuclei such as mineral dust.
2.1. Ice Nucleation Measurements
 The apparatus used in these studies has been described in detail previously [Dymarska et al., 2006; Parsons et al., 2004]. It consists of an optical microscope coupled to a flow cell in which the relative humidity could be accurately controlled. Mineral dust particles (either coated or uncoated) were deposited on the bottom surface of the flow cell; the relative humidity with respect to ice (RHi) inside the cell was increased, and the conditions under which ice crystals first formed were determined with a reflected light microscope (this is defined as the onset RHi and temperature). The RHi over the particles was controlled by continuously flowing a mixture of dry and humidified He through the flow cell. The bottom surface of the flow cell, which supported the particles, consisted of a glass cover slide treated with dichlorodimethylsilane to make a hydrophobic layer, which reduced the probability of ice nucleation directly on the surface.
 Typical experimental RHi trajectories used in our ice nucleation experiments are illustrated in Figure 1. At the beginning of the experiments, the particles were exposed to a flow of dry He gas (RHi < 1%) at room temperature. Next, the temperature of the cell was rapidly lowered and the RHi was set to approximately 95%. The nucleation experiments were then conducted by steadily decreasing the temperature and thus increasing the RHi. The RHi ramp rate was approximate 1% min−1. We also carried out experiments using a ramp rate of approximately 0.5% min−1. No differences in results for the two ramp rates were obtained, suggesting the aqueous coatings were in equilibrium with the gas-phase water vapor.
 We also carried out growth rate calculations of aqueous solution droplets using the equations presented by Pruppacher and Klett  to further confirm that the aqueous coatings were in equilibrium with the gas-phase water vapor in our experiments. These calculations show that for aqueous coatings 5 μm in thickness, the coatings should be very close to equilibrium with the gas-phase water vapor under our experimental conditions. The uncertainty in our measurements due to non-equilibrium conditions is at most 3% RHi.
2.2. Sample Preparation and Thickness of the Coatings
 The uncoated kaolinite particles were prepared by first mixing kaolinite in high-purity water (composition was 1% kaolinite by mass) to create a suspension. The suspension was then placed in an ultrasonic bath for 10 minutes. To deposit the particles on the glass slide, the suspension was passed through a nebulizer using high-purity nitrogen as a carrier gas. The flow from the nebulizer was directed at a hydrophobic glass slide and droplets containing the particles were deposited on the surface of the slide upon impaction. Water then evaporated, leaving behind the kaolinite particles. The coated kaolinite particles were prepared by mixing the kaolinite and coating material in high-purity water (composition was 1% kaolinite and 0.2% coating material by mass). This suspension was then placed in the ultrasonic bath prior to nebulization. Coated particles produced by this method had an average weight fraction of H2SO4 or (NH4)2SO4 of 0.167 under dry conditions,
 The thicknesses of the coatings in our experiments were estimated based on the compositions of the starting suspensions and assuming a spherical core shell model (i.e. a kaolinite core surrounded by a uniform H2SO4 or (NH4)2SO4 coating). According to our calculations, under dry conditions (<1% RHi) a kaolinite core with a diameter of 15 μm will have a 0.7 μm thick coating, and a 5 μm core will have a coating of 0.2 μm. A coating of 0.2 μm represents at least a few hundred sulphate layers covering the surface of the particle.
 To further characterize the thickness of the coatings we monitored the change in particle size as the relative humidity with respect to water (RHw) was increased from <1% to 95%. From the change in size we estimate the total amount of water adsorbed when cycling between <1% and 95% RHw using the thermodynamic model of Clegg et al. . From this we then estimated the amount of H2SO4 in each particle and thus the thickness of the H2SO4 coating under dry conditions. Measurements made for 15 individual particles yielded an average weight fraction for the coating of 0.12 ± 0.07 under dry conditions. The uncertainty in this value derives from the uncertainty in the relative humidity measurements.
2.3. Particle Number, Particle Size and Total Surface Area
 In our experiments a typical sample held between 100 and 1000 individual particles. As mentioned above we define the onset conditions as the RHi and temperature at which the first ice nucleation event occurred. Hence, our results correspond to when 1 to 0.1% of the kaolinite particles nucleated ice. The total surface area of mineral dust deposited in any particular experiment ranged from 1 × 10−4 to 2 × 10−3 cm2. Over this narrow range, the onset results did not depend on the surface area. To illustrate this point, in the auxiliary material we present onset data for coated and uncoated kaolinite particles as a function of surface area. Since the results did not depend on surface area, for each temperature and each type of particle we combine all the data together and calculate an average and standard deviation for the onset RHi (see Figure 2). For each temperature and each type of particle, at least six measurements were carried out. Based on an analysis of all the particles studied and assuming a normal distribution of particle sizes, the mean diameters for the uncoated particles, H2SO4-coated particles and (NH4)2SO4-coated particles were 7.7 μm, 7.8 μm and 6.9 μm with standard deviations of 5.3 μm, 5.7 μm and 5.0 μm, respectively.
3. Results and Discussion
3.1. Onset Results for Uncoated Kaolinite
 The onset results for uncoated kaolinite particles are shown in Figure 2. The data show that only a small supersaturation (<110% RHi) is required for an ice nucleation event to occur in these experiments. These results, which correspond to deposition freezing [Pruppacher and Klett, 1997], are in excellent agreement with results we obtained earlier for uncoated kaolinite particles [Dymarska et al., 2006; Eastwood et al., 2008]. The method of sample preparation in the current experiments differs from the method in our previous studies: in the previous studies we used a dry dispersion technique to generate the aerosol particles. In the current study we generated the uncoated particles using the same technique used to generate the coated particles – nebulizing an aqueous suspension – in order to have the same experimental conditions in both the uncoated and coated work. For a detailed discussion on previous measurements of the ice nucleation properties of uncoated kaolinite particles see Eastwood et al. .
3.2. Onset Results for Kaolinite Particles Coated With H2SO4
 Onset results for H2SO4 coatings are also shown in Figure 2. These results correspond to either immersion freezing or condensation freezing depending on the temperature [Pruppacher and Klett, 1997]. In contrast to the uncoated results, the H2SO4 coated particles required much higher supersaturations before ice nucleation occurred. The data show that H2SO4 coatings drastically altered the ice nucleating ability of kaolinite particles, increasing the RHi required to initiate ice nucleation by approximately 30%.
 Before discussing a possible explanation for the shift in ice nucleating ability of H2SO4 coated particles, we first discuss the composition of the H2SO4 coatings during the ice nucleation experiments with the aid of Figure 1. The dot-dashed lines in Figure 1 show the experimental conditions at which the H2SO4 coatings are 10, 20, and 30 wt% H2SO4 (the remainder being H2O), calculated using the model of Clegg et al. . Figure 1 illustrates that for almost all of the experimental conditions, the H2SO4 coatings are concentrated aqueous solutions (>10 wt % H2SO4) and highly acidic (the pH of a 10 wt % H2SO4 solution is less than 0). This may help explain our freezing results (see below).
 Kaolinite is a clay mineral whose unit cell is best described as layers of aluminosilicate sheets held together by hydrogen bonds. The “top” surface of these sheets is a gibbsite-like surface with exposed hydroxyl groups. The “bottom” surface of these sheets is a siloxane surface with exposed oxygen atoms [Stumm, 1992]. The Point of Zero Charge (defined as the pH at which the net surface charge is zero) for the gibbsite-like surface is approximately 6, while the Point of Zero Charge for the edge surfaces is approximately 7–8 [Stumm, 1992]. As mentioned above, the pH of the H2SO4 coating was below 0 under most experimental conditions. As a result, the kaolinite particles should exhibit highly protonated, positively charged surfaces with a proton surface density ∼1.2 × 1014 cm−2 [Stumm, 1992]. Under these conditions sulfate anions can be strongly adsorbed to metal oxide surfaces due to electrostatic attractions or the sulfate anions can adsorb by ligand substitution [Karltun, 1997; Peak et al., 1999; Stumm, 1992; Zhang and Peak, 2007]. Strongly adsorbed sulfate anions will change the chemical and physical properties of solid mineral interface. Hence it is not surprising that we observed higher onset RHi values for ice nucleation in our experiments with H2SO4 coatings.
Knopf and Koop  studied Arizona Test Dust (ATD) particles with and without H2SO4 coatings, but they did not observe significant differences due to the coatings. A direct comparison with these previous results is difficult, however, since ATD is a complex mixture of minerals. Archuleta et al.  studied metal oxides particles, with and without H2SO4 coatings. At 213 K, they observed an increase in RHi after coating for 200 nm aluminum oxide and amorphous aluminum silicate particles, but a decrease in RHi for 200 nm iron oxide particles. In contrast, at 228 K the change in the RHi due to the coatings on 200 nm particles was less than the uncertainty in the measurements. Differences between our measurements and the measurements by Archuleta et al.  include the minerals studied and the particle size. Also, the coatings used by Archuleta et al.  were between 2.9 and 7.1 layers thick, which are much smaller than our coating thicknesses. Finally Möhler et al. [2008b] studied illite particles coated with sulfuric acid (H2SO4 mass fraction about 30%) at 210 K. They observed a significant reduction in the ice nucleation efficiency of the particles after coating. Specifically, an increase in RHi necessary for ice nucleation of approximately 35–40% was observed. Our results are similar to these previous results.
3.3. Onset Results for (NH4)2SO4 Coatings
 The onset results for (NH4)2SO4 coated particles are also shown in Figure 2. At 240 and 245 K, the onset RHi values are significantly higher than the uncoated case. However, at 236 K, the coated particles are just as efficient at nucleating ice as the uncoated kaolinite particles. Clearly, the (NH4)2SO4 coatings are influencing the ice nucleating ability of the kaolinite particles much differently than the H2SO4 coatings. A possible explanation for this difference may be related to the phase of the coatings. The H2SO4 coatings will remain liquid during our experiments, whereas the (NH4)2SO4 coatings can undergo deliquescence and efflorescence. This can complicate the situation since ice nucleation can occur directly on crystalline (NH4)2SO4 [Abbatt et al., 2006; Shilling et al., 2006; Zuberi et al., 2002].
Figure 3 illustrates the possible phase behavior of an (NH4)2SO4 coating during a typical freezing experiment. Our experiments first start at a very low RHi (<1%). Under these conditions, the (NH4)2SO4 coatings are expected to be crystalline, based on previously reported values for the efflorescence relative humidity of (NH4)2SO4 droplets containing kaolinite cores [Pant et al., 2006]. The RHi is then increased and, above 100% RHi, ice can nucleate on the crystalline (NH4)2SO4 coating. In this case, ice nucleation occurs by deposition freezing [Pruppacher and Klett, 1997]. If ice nucleation does not occur, the solid (NH4)2SO4 coating can take up water at the deliquescence relative humidity. Above this relative humidity, the (NH4)2SO4 coating is a liquid and ice nucleation can occur by immersion freezing [Pruppacher and Klett, 1997].
 Based on the discussion above, our data suggest that crystalline (NH4)2SO4 coatings are effective ice nuclei at approximately 236 K. This is consistent with previous laboratory data that show that ice nucleation can occur on supermicron (NH4)2SO4 particles at RHi-values less than 110% and temperatures less than 225 K [Abbatt et al., 2006; Shilling et al., 2006]. This is illustrated in the auxiliary material where we compare our (NH4)2SO4 data with these previous measurements.
 At 245 K, the onset results for (NH4)2SO4 are significantly higher than the uncoated case. This suggests that aqueous (NH4)2SO4 coatings also increase the RHi values required for ice nucleation, although not as much as aqueous H2SO4 coatings. This may also be because sulfate anions are adsorbed to the mineral surface when coated with aqueous (NH4)2SO4. Martin et al.  showed using spectroscopy that sulfate anions chemisorbed to a range of metal oxide surfaces even in ammonium sulfate solutions. Therefore it seems likely that the sulfate anions are also interacting with the kaolinite surface in our (NH4)2SO4 experiments.
4. Conclusions and Atmospheric Implications
 Our initial results support the idea that anthropogenic emissions of SO2 and NH3 may influence the ice nucleating properties of mineral dust particles by increasing the relative humidity required for ice nucleation. This shift in ice nucleation conditions may influence the frequency and properties of ice clouds. Calculations using a cloud parcel model are needed to explore this possibility further. One area where these results may be especially important is the Arctic region where a large fraction of the aerosol particles (including insoluble aerosols such as mineral dust) can be coated with acidic sulfate [Bigg, 1980].
 We have reported initial results for coated and uncoated kaolinite particles. Additional experiments as a function of particle size, surface area, and coating thickness are also needed. Measurements of the fractions of ice-active mineral particles as a function of temperature and humidity would also be beneficial.
 We thank D. A. Knopf for many helpful discussions. This research was supported by CFCAS and NSERC.