Glass-formers and viscous liquid slowdown since David Turnbull: Enduring puzzles and new twists

To Turnbull's study of the kinetic problem of nucleation and growth of crystals, we add the further enquiry into what lies behind the slow nucleation kinetics of glass-formers. Our answer to this question leads to the proposal of conditions in which a pure liquid metal, monatomic and elemental, can be vitrified. Using the case of high-pressure liquid germanium, we give electron microscope evidence for the validity of our thinking. On the question of how liquids behave when crystals do not form, Turnbull pioneered the study of glass transitions in metallic alloys, measuring the heat capacity change at the glass transition T. for the first time, and developing with Cohen the free volume model for the temperature dependence of liquid transport properties approaching T-g. We extend the phenomenological picture to include networks where free volume does not play a role and reveal a pattern of behavior that provides for a classification of glass-formers (from "strong" to "fragile"). Where Turnbull studied supercooled liquid metals and P-4 to the homogeneous nucleation limit using small droplets, we studied supercooled water in capillaries and emulsions to the homogeneous nucleation limit near -40 degrees C. We discuss the puzzling divergences observed that are now seen as part of a cooperative transition that leads to very untypical glass-transition behavior at lower temperatures (when crystallization is bypassed by hyperquenching). Finally, we show how our interpretation of water behavior can be seen as a bridge between the behavior of the "strong" (network) liquids of classical glass science (e.g., SiO2) and the "fragile" behavior of typical molecular glass-formers. The link is made using a "Gaussian excitations" model by Matyushov and the author in which the spike in heat capacity for water is pushed by cooperativity (disorder stabilization of excitations) into a first-order transition to the ground state, at a temperature typically below Tg. In exceptional Gases like triphenyl phosphite, this liquid-to-glass first-order transition lies above Tg and can be studied in detail.