This report summarizes the result of LDRD project 12-0395, titled "Automated Algorithms for Quantum-level Accuracy in Atomistic Simulations." During the course of this LDRD, we have developed an interatomic potential for solids and liquids called Spectral Neighbor Analysis Poten- tial (SNAP). The SNAP potential has a very general form and uses machine-learning techniques to reproduce the energies, forces, and stress tensors of a large set of small configurations of atoms, which are obtained using high-accuracy quantum electronic structure (QM) calculations. The local environment of each atom is characterized by a set of bispectrum components of the local neighbor density projected on to a basis of hyperspherical harmonics in four dimensions. The SNAP coef- ficients are determined using weighted least-squares linear regression against the full QM training set. This allows the SNAP potential to be fit in a robust, automated manner to large QM data sets using many bispectrum components. The calculation of the bispectrum components and the SNAP potential are implemented in the LAMMPS parallel molecular dynamics code. Global optimization methods in the DAKOTA software package are used to seek out good choices of hyperparameters that define the overall structure of the SNAP potential. FitSnap.py, a Python-based software pack- age interfacing to both LAMMPS and DAKOTA is used to formulate the linear regression problem, solve it, and analyze the accuracy of the resultant SNAP potential. We describe a SNAP potential for tantalum that accurately reproduces a variety of solid and liquid properties. Most significantly, in contrast to existing tantalum potentials, SNAP correctly predicts the Peierls barrier for screw dislocation motion. We also present results from SNAP potentials generated for indium phosphide (InP) and silica (SiO 2 ). We describe efficient algorithms for calculating SNAP forces and energies in molecular dynamics simulations using massively parallel computers and advanced processor ar- chitectures. Finally, we briefly describe the MSM method for efficient calculation of electrostatic interactions on massively parallel computers.
Nanocrystalline copper lms were created by both repetitive high-energy pulsed power, to produce material without internal nanotwins; and pulsed laser deposition, to produce nan- otwins. Samples of these lms were indented at ambient (298K) and cryogenic temperatures by immersion in liquid nitrogen (77K) and helium (4K). The indented samples were sectioned through the indented regions and imaged in a scanning electron microscope. Extensive grain growth was observed in the lms that contained nanotwins and were indented cryogenically. The lms that either lacked twins, or were indented under ambient conditions, were found to exhibit no substantial grain growth by visual inspection. Precession transmission elec- tron microscopy was used to con rm these ndings quantitatively, and show that 3 and 7 boundaries proliferate during grain growth, implying that these interface types play a key role in governing the extensive grain growth observed here. Molecular dynamics sim- ulations of the motion of individual grain boundaries demonstrate that speci c classes of boundaries - notably 3 and 7 - exhibit anti- or a-thermal migration, meaning that their mobilities either increase or do not change signi cantly with decreasing temperature. An in-situ cryogenic indentation capability was developed and implemented in a transmission electron microscope. Preliminary results do not show extensive cryogenic grain growth in indented copper lms. This discrepancy could arise from the signi cant di erences in con g- uration and loading of the specimen between the two approaches, and further research and development of this capability is needed.
It is well known that screw dislocation motion dominates the plastic deformation in body-centered-cubic metals at low temperatures. The nature of the nonplanar structure of screw dislocations gives rise to high lattice friction, which results in strong temperature and strain rate dependence of plastic flow. Thus the nature of the Peierls potential, which is responsible for the high lattice resistance, is an important physical property of the material. However, current empirical potentials give a complicated picture of the Peierls potential. Here, we investigate the nature of the Peierls potential using density functional theory in the bcc transition metals. The results show that the shape of the Peierls potential is sinusoidal for every material investigated. Furthermore, we show that the magnitude of the potential scales strongly with the energy per unit length of the screw dislocation in the material.