Amorphous ice melts twice

New experiments lend credence to a counterintuitive, two-liquid theory of water.

Water can be peculiar at times. Chilled near its freezing point, it expands as it’s cooled. It can become less viscous as it’s compressed. And, according to an idea proposed two decades ago by Eugene Stanley and coworkers at Boston University, it can do something that no other pure substance is known to do: form two distinct liquid phases. (See the article in Physics Today, June 2003, page 40 .) The idea isn’t far-fetched. Molecular dynamics simulations have shown water separating into high- and low-density liquid (HDL and LDL) phases below 245 K. Moreover, it’s well known that water can form high- and low-density amorphous ices (HDA and LDA) at liquid-nitrogen temperatures. (Mass-equivalent samples of HDA and LDA are pictured at left and right, respectively.) Because molecular configurations of liquids and amorphous solids closely resemble one another, some researchers speculate that the putative liquid phases can be created in the lab by melting their amorphous-ice proxies. That’s easier said than done: HDA isn’t very stable at ambient pressure, and it typically converts to LDA before reaching its melting point, or glass-transition temperature. Thomas Loerting and colleagues at the University of Innsbruck in Austria have now found a way around that problem. By subjecting HDA to special annealing protocols, they enhanced the ice’s stability, which allowed them to detect its glass transition at 116 K. Incidentally, when the resulting HDL was heated and rechilled, it converted to LDL and then to LDA. Heating the LDA revealed water’s second glass transition, from LDA to LDL, at 136 K. (K. Amann-Winkel et al., Proc. Natl. Acad. Sci. USA, in press ; images courtesy of Osamu Mishima.)—Ashley G. Smart