Abstract
We review our results on large-scale atomistic simulations of the phase change compound GeTe of interest for applications in nonvolatile electronic memories. The simulations are based on an interatomic potential with an accuracy close to that of the density functional theory (DFT). The potential was generated by fitting a DFT database by means of an artificial neural network method. This methodological advance allowed us to perform molecular dynamics simulations with several thousand atoms for several ns that provided useful insights on several properties of interest for the operation of phase change memories, including the crystallization kinetics, the dynamics of the supercooled liquid, the structural relaxation in the glass and the properties of nanowires.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
