A class of chalcogenide alloys named phase change materials is exploited in non-volatile memories and in neuromorphic devices. These applications rely on a fast reversible transformation between the crystalline and amorphous phases upon Joule heating. The two phases correspond to the two logical states of the memory that are discriminated thanks to their large contrast in electrical conductivity. The transformation is very fast (10-100 ns) above the glass transition temperature (about 120 °C) while the two phases are stable at normal conditions. This huge change in the crystallization speed in such a short temperature range is ascribed to the fragility of the supercooled liquid, i.e. a super-Arrhenius behavior with temperature of the viscosity which keeps very low down to temperature close to the glass transition and then raises very steeply to the value expected for the glass.
Despite its importance, the direct measurement of the viscosity at the operational conditions of the memories is very difficult because of the very fast crystallization speed.
In the recent paper entitled “Atomistic Study of the Configurational Entropy and the Fragility of Supercooled Liquid GeTe” (doi: 10.1002/adfm.202314264) published in Advanced Functional Materials (Wiley, Impact Factor 19.0, 2022 Journal Impact Factor, Journal Citation Reports (Clarivate Analytics, 2023)) thanks to a collaboration between a team of the University of Rome La Sapienza and the University of Milano-Bicocca (Prof. Marco Bernasconi and Dr. Davide Campi of the Department of Materials Science), crucial information on the viscosity in the prototypical phase change compound GeTe has been provided by large scale molecular dynamics simulations.
The team exploited a machine-learning interatomic potential, devised previously by the group in Milan, to sample the potential energy landscape in the liquid and then extract the configurational entropy in the deeply supercooled regime from which the viscosity was then computed within a well-established phenomenological scheme (Adam-Gibbs theory). The simulations provided microscopic insights on the dynamical properties of the system which was indeed confirmed to be highly fragile.