New Apparatus for Laboratory Studies of Ice Nucleation, Growth and Sublimation Rates, and Habit of Ice Crystals

Ice-crystal growth and sublimation rates, as well as their habit, significantly affect radiation and precipitation processes in clouds. Previous measurements of these rates and habits contain gaps and major inconsistencies. The inconsistencies arise from a lack of reproducibility in previous measurements (exacerbated by poor knowledge of the influence of nucleation on habit), imprecise characterization of humidity and impurity levels around the crystal, and artifacts introduced by the laboratory apparatus. For example, are the conflicting reports of primarily tabular vs primarily columnar forms below -25 ˚C due to different nucleation modes, poorly characterized conditions, or instrument artifacts? Are the discrepancies in low-supersaturation forms due to inherent variability (e.g., from defects), differences in nucleation mode, or some artifact? The issues are serious because the growth rates and habit are sensitive to temperature and supersaturation, and possibly also to nucleation mode. The solution then is to develop an apparatus with fine control over the conditions, including nucleation mode, and then follow a careful testing procedure to determine that the method does not introduce growth-altering artifacts. Also the apparatus needs the ability to grow multiple crystals simultaneously under identical conditions to understand inherent variability. We introduce here such an apparatus.

I) Problems with previously used instruments Many apparatuses used to study ice growth from the vapor grow crystals on a relatively large supporting surface such as a flat substrate or fiber1,2. Such surfaces hold the crystal steady for good imaging and temperature–supersaturation control, but they may introduce substrate effects, including growth-promotion3 and unnatural temperature4 or humidity gradients. Another problem with substrates and fibers is that they usually produce numerous crystals that grow so close together that they compete for vapor5. In contrast, cloud chambers, vertical wind tunnels, and electrodynamic balances do not have such surface issues as the crystals are free-floating, but these methods generally have relatively poor control over the temperature, supersaturation, and air pressure. Also, due to the crystal motion, these methods often have poor crystal imaging. II) How the new design overcomes these problems The new design greatly expands upon the capillary-growth method6, hereafter “Cap1”. In that method, a 0.9-mm diameter fused-silica glass capillary was pulled down to a diameter of about 5–10 µm. Water was inserted into the end, and the capillary was inserted into a small chamber containing a vapor source. The vapor source was an aqueous solution with LiCl salt, and the chamber was surrounded by temperature-controlled, transparent cooling fluid. The water in the capillary was frozen, producing a small ice sphere at the tip, which then grew under constant, uniform conditions, accurately determined by the temperature and LiCl concentration in the vapor source. Thus, the method removed the problem of crystal competition, thermal and humidity gradients, while allowing for accurate control of the conditions, sharp crystal imaging, as well as crystal manipulation (capillary can be moved up–down and rotated). The issue of growth-promotion at the glass–ice–vapor intersection edge could largely be ignored because the complete view of the capillary made the effect obvious when it occurred, and thus we could focus on crystals, or just crystal faces, that were unaffected. As a result, with Cap1, we could determine the growth and sublimation modes, measure their rates, and closely observe habit development. But with Cap1, we could not independently change temperature and supersaturation. This fault increased the time to do experiments under different conditions and limited a given crystal to a specific temperature–supersaturation curve. For the new apparatus, hereafter “Cap2”, we added two new chambers for source vapor, each isolatable from the crystal-growth chamber by a newly designed vacuum valve, and each containing a temperature-controlled ice surface that supplies vapor and controls the vapor pressure surrounding the growing crystal or crystals. The temperature control of the source ice surfaces occurs via thermoelectric cooling elements. In addition, each vapor source has a separate port from which to add or change solute (without bringing the system to room temperature and restarting with a new crystal). This allows us to control vapor pressure the same way as was done for Cap1, but without the thermoelectric elements. In either method, the supersaturation can be quickly switched between two values, independent of temperature, all for the same crystal. The same holds for temperature changes; with Cap2, we can switch between two circulating coolers at different temperatures, to quickly change the temperature of the crystal. In addition to speeding up the data-gathering, these abilities allow us to study growth forms under changing conditions. The new apparatus addresses the inherent variability of habit, while also speeding data-gathering, by having three capillary ports into the growth chamber. (Cap1 had only one.) Each port can have one capillary and one crystal, but the capillaries slide up and down along different axes, allowing us to move the crystals closer or further apart as well as rotating each crystal about the capillary axis. Their primary purpose is to allow us to grow up to three well-separated crystals at the same time. By having three, we roughly triple the rate of data acquisition, as well as identify how crystals grown in the same conditions may or may not grow similarly. In addition, the method may be used to study crystal-crystal or crystal-droplet proximity effects. To study growth at low pressure, Cap2 has a pressure port from which to control and measure the air pressure. This ability also allows us to grow or sublimate in a pure vapor, which will make it easier to measure deposition-coefficient functions. By measuring the deposition-coefficient functions for each crystal face, we can test fundamental crystal-growth models for ice. In addition, the copper block that houses the three chambers also has a reference-ice chamber, allowing us to use a differential capacitive monometer to accurately measure the supersaturation. In our presentation, we will describe the building and testing of Cap2. We will also discuss the methods used to nucleate the crystals via different nucleation modes and to further reduce any substrate artifact from the ultra-thin glass capillaries.