We develop a neuroevolution-potential (NEP) framework for generating neural network based machine-learning potentials. They are trained using an evolutionary strategy for performing large-scale molecular dynamics (MD) simulations. A descriptor of the
atomic environment is constructed based on Chebyshev and Legendre polynomials. The method is implemented in graphic processing units within the open-source GPUMD package, which can attain a computational speed over $10^7$ atom-step per second using one Nvidia Tesla V100. Furthermore, per-atom heat current is available in NEP, which paves the way for efficient and accurate MD simulations of heat transport in materials with strong phonon anharmonicity or spatial disorder, which usually cannot be accurately treated either with traditional empirical potentials or with perturbative methods.
Understanding the mechanisms of thermal conduction in graphene is a long-lasting research topic, due to its high thermal conductivity. Peierls-Boltzmann transport equation (PBTE) based studies have revealed many unique phonon transport properties in
graphene, but most previous works only considered three-phonon scatterings and relied on interatomic force constants (IFCs) extracted at 0 K. In this paper, we explore the roles of four-phonon scatterings and the temperature dependent IFCs on phonon transport in graphene through our PBTE calculations. We demonstrate that the strength of four-phonon scatterings would be severely overestimated by using the IFCs extracted at 0 K compared with those corresponding to a finite temperature, and four-phonon scatterings are found to significantly reduce the thermal conductivity of graphene even at room temperature. In order to reproduce the prediction from molecular dynamics simulations, phonon frequency broadening has to be taken into account when determining the phonon scattering rates. Our study elucidates the phonon transport properties of graphene at finite temperatures, and could be extended to other crystalline materials.
Experimentally produced graphene sheets exhibit a wide range of mobility values. Both extrinsic charged impurities and intrinsic ripples (corrugations) have been suggested to induce long-range disorder in graphene and could be a candidate for the dom
inant source of disorder. Here, using large-scale molecular dynamics and quantum transport simulations, we find that the hopping disorder and the gauge and scalar potentials induced by the ripples are short-ranged, in strong contrast with predictions by continuous models, and the transport fingerprints of the ripple disorder are very different from those of charged impurities. We conclude that charged impurities are the dominant source of disorder in most graphene samples, whereas scattering by ripples is mainly relevant in the high carrier density limit of ultraclean graphene samples (with a charged impurity concentration < 10 ppm) at room and higher temperatures.
In this paper, we develop a highly efficient molecular dynamics code fully implemented on graphics processing units for thermal conductivity calculations using the Green-Kubo formula. We compare two different schemes for force evaluation, a previousl
y used thread-scheme where a single thread is used for one particle and each thread calculates the total force for the corresponding particle, and a new block-scheme where a whole block is used for one particle and each thread in the block calculates one or several pair forces between the particle associated with the given block and its neighbor particle(s) associated with the given thread. For both schemes, two different classical potentials, namely, the Lennard-Jones potential and the rigid-ion potential are implemented. While the thread-scheme performs a little better for relatively large systems, the block-scheme performs much better for relatively small systems. The relative performance of the block-scheme over the thread-scheme also increases with the increasing of the cutoff radius. We validate the implementation by calculating lattice thermal conductivities of solid argon and lead telluride.
