Environmental Scanning Electron Microscopy of Ice Crystal Nucleation and Growth

Ice crystal nucleation and growth are dual processes that can be studied uniquely through Environmental Scanning Electron Microscopy (ESEM). By utilizing differential pumping systems and a Peltier element to vary the vapor pressure and to achieve temperatures below the freezing point, respectively, it is possible to obtain supersaturated conditions relative to ice in the sample chamber of an Environmental Scanning Electron Microscope. Ice crystals were nucleated on a variety of atmospherically relevant substrates and grown in a pure water vapor environment in the chamber of a FEI-Quanta 200 ESEM. To initiate ice crystal nucleation, the Peltier element was set at a temperature between -10°C and -25°C, while the chamber water vapor pressure was adjusted to just below the frost point. Ice crystal nucleation and growth was then controlled by careful adjustments of chamber pressure and temperature, where high-magnification images of hexagonal ice crystals were acquired at nanoscale resolution. These images display prominent mesoscopic surface topography including linear strands, crevasses, islands, and steps. The surface features are seen to be ubiquitously present at all observed temperatures, at many supersaturated and subsaturated conditions, and on all crystal facets. Additionally, a pre-growth "shadow" resembling a dark spot sometimes appeared on areas of the sample stage immediately preceding ice crystal nucleation and growth. The observations represent the most highly magnified images of ice surfaces yet reported and significantly expand the range of ambient conditions where the features are conspicuous. New knowledge of the presence and characteristics of these features could transform the fundamental understanding of ice crystal growth kinetics and its physical parameterization in the context of atmospheric and cryospheric science. To the extent these observations are applicable to atmospheric ice, the results suggest that the radiative representation of ice and mixed-phase cloud properties in climate models could be markedly affected.