Relatore: prof. Marco Bernasconi, Dipartimento di Scienza dei Materiali, Università di Milano-Bicocca, Milano.

Abstract: 
Chalcogenide compounds such as GeTe and GeSbTe alloys are attracting an increasing interest for their application in Phase Change Memories (PCM) [1]. This novel type of electronic non-volatile memory rests on a fast and reversible transformation between the crystalline and amorphous phase of a chalcogenide film induced by Joule heating. The two phases corresponding to the two states of the memory can be discriminated because of their large difference in electronic conductivity.
PCMs are emerging as a leading contender for the realization of the so-called storage class memories that are sought to fill the performance gap between volatile DRAM and non-volatile Flash memories.  Storage class memories are believed to usher in seminal changes in the memory and storage hierarchy for all computing platforms. Application of PCM for neuromorphic computing is also an active field of research.
In the last decade, atomistic simulations based on density functional theory (DFT) have provided useful insights on the properties of chalcogenide alloys of interest for PCMs [2]. Still, large simulation cells and long simulation times beyond the reach of DFT simulations are needed to address several key issues of relevance for PCM operation. To overcome these limitations, we have developed an interatomic potential for the prototypical phase change compound GeTe by fitting a large DFT database with a neural network (NN) scheme. Large scale (104 atoms) NN simulations allowed us to get insights on microscopic origin of the high crystallization speed of these materials [3].
In this talk, I will review the results of NN simulations on the kinetics of homogeneous and heterogeneous crystallization of GeTe and on the structural relaxations leading to the aging of the amorphous phase which is a particularly critical issue for PCM operation [4].