Linear Scaling, Quantum-Accurate Interatomic Potentials with SNAP; Reaching those Hard-to-Reach Places in Classical Molecular Dynamics

Mitchell Wood, Aidan Thompson

Sandia National Laboratories

Of the many ways to improve the predictive capacity of classical Molecular dynamics (MD), progress into the fidelity of the underlying interatomic potential(IAP) seems to consistently lag behind the hardware and software advancements that expand the accessible length/time-scales. In this talk we report on the current capability of SNAP, a machine learned IAP that has been demonstrated to preserve quantum-mechanical accuracy for a number of different material systems, most notably bcc-metals. The focus of much of the SNAP development has been for ‘hard-to-reach’ problems in classical MD, here we will show examples of materials in extreme environments such as plasma facing materials in fusion reactors and extremes of temperature and pressure. A general discussion of the promise and potential pitfalls of machine learned and neural net potentials will also be provided.