Ab Initio Molecular-Dynamics Simulation of Neuromorphic Computing in Phase-Change Memory Materials.

We present an in silico study of the neuromorphic-computing behavior of the prototypical phase-change material, Ge2Sb2Te5, using ab initio molecular-dynamics simulations. Stepwise changes in structural order in response to temperature pulses of varying length and duration are observed, and a good reproduction of the spike-timing-dependent plasticity observed in nanoelectronic synapses is demonstrated. Short above-melting pulses lead to instantaneous loss of structural and chemical order, followed by delayed partial recovery upon structural relaxation. We also investigate the link between structural order and electrical and optical properties. These results pave the way toward a first-principles understanding of phase-change physics beyond binary switching.J.M.S. gratefully acknowledges funding from an internal graduate studentship provided by Trinity College, Cambridge, and from a U.K. Engineering and Physical Sciences Research Council Programme Grant (Grant No. EP/K004956/1). This work was primarily carried out using the Cambridge HPC facility (www.hpc.cam.ac.uk), and some additional calculations were performed using the ARCHER supercomputer through membership of the U.K. HPC Materials Chemistry Consortium, which is funded by EPSRC Grant No. EP/L000202.This is the author accepted manuscript. The final version is available from ACS via http://dx.doi.org/10.1021/acsami.5b0182