Three parallel algorithms for classical molecular dynamics are presented. The first assigns each processor a fixed subset of atoms; the second assigns each a fixed subset of inter-atomic forces to compute; the third assigns each a fixed spatial region. The algorithms are suitable for molecular dynamics models which can be difficult to parallelize efficiently—those with short-range forces where the neighbors of each atom change rapidly. They can be implemented on any distributed-memory parallel machine which allows for message-passing of data between independently executing processors. The algorithms are tested on a standard Lennard-Jones benchmark problem for system sizes ranging from 500 to 100,000,000 atoms on several parallel supercomputers--the nCUBE 2, Intel iPSC/860 and Paragon, and Cray T3D. Comparing the results to the fastest reported vectorized Cray Y-MP and C90 algorithm shows that the current generation of parallel machines is competitive with conventional vector supercomputers even for small problems. For large problems, the spatial algorithm achieves parallel efficiencies of 90% and a 1840-node Intel Paragon performs up to 165 faster than a single Cray C9O processor. Trade-offs between the three algorithms and guidelines for adapting them to more complex molecular dynamics simulations are also discussed.

Fast parallel algorithms for short-range molecular dynamics

Physical Review B

A grand challenge in material science is to understand the correlation between intrinsic properties and defect dynamics. Radiation tolerant materials are in great demand for safe operation and advancement of nuclear and aerospace systems. Unlike traditional approaches that rely on microstructural and nanoscale features to mitigate radiation damage, this study demonstrates enhancement of radiation tolerance with the suppression of void formation by two orders magnitude at elevated temperatures in equiatomic single-phase concentrated solid solution alloys, and more importantly, reveals its controlling mechanism through a detailed analysis of the depth distribution of defect clusters and an atomistic computer simulation. The enhanced swelling resistance is attributed to the tailored interstitial defect cluster motion in the alloys from a long-range one-dimensional mode to a short-range three-dimensional mode, which leads to enhanced point defect recombination. The results suggest design criteria for next generation radiation tolerant structural alloys.

/pdf/enhancing-radiation-tolerance-by-controlling-defect-mobility-52so1qdurm.pdf

Enhancing radiation tolerance by controlling defect mobility and migration pathways in multicomponent single-phase alloys

Using pseudopotential density functional theory within the local-density approximation, we calculate the ideal shear strengths of W for slip on {110}, {112} and {123} planes allowing for complete structural relaxation orthogonal to the applied shear. The strengths in the weak directions on all planes are found to be very nearly equal (about 18GPa, or 11% of the shear modulus G). Moreover, the shear instability occurs at approximately the same applied shear strain (17–18%). This unusual isotropy is explained in terms of the atomic configurations of high-energy saddle points reached during shear. Analysis of these saddle points may also offer a simple explanation for the prevalence of the pencil glide of dislocations on planes containing a (111) direction in bcc metals. Finally, we calculate the ideal cleavage strengths of W on {100} and compare our calculated ideal shear and cleavage strengths with experimental nanoindentation and whisker measurements. All these results can be rather simply unders...

The ideal strength of tungsten

Recent experiments have demonstrated that plastic strains in nanocrystalline aluminum and gold films with grain sizes on the order of 50 nm are partially recoverable. To reveal the mechanisms behind such strain recovery, we perform large scale molecular dynamics simulations of plastic deformation in nanocrystalline aluminum with mean grain sizes of 10, 20, and 30 nm. Our results indicate that the inhomogeneous deformation in a polycrystalline environment results in significant residual stresses in the nanocrystals. Upon unloading, these internal residual stresses cause strain recovery via competitive deformation mechanisms including dislocation reverse motion/annihilation and grain-boundary sliding/diffusion. By tracking the evolution of each individual deformation mechanism during strain recovery, we quantify the fractional contributions by grain-boundary and dislocation deformation mechanisms to the overall recovered strain. Our analysis shows that, even under strain rates as high as those in molecular dynamics simulations, grain-boundary-mediated processes play important roles in the deformation of nanocrystalline aluminum.

https://www.pnas.org/content/pnas/106/38/16108.full.pdf

Competing grain-boundary- and dislocation-mediated mechanisms in plastic strain recovery in nanocrystalline aluminum

The contributions of different mechanisms of self-diffusion in face-centered cubic metals Ni, Cu, and Al at thermodynamic equilibrium have been analyzed using the molecular dynamics method. The vacancy, divacancy, and cyclic mechanisms of self-diffusion, as well as the mechanisms involving the vacancy migration to the second coordination sphere and the formation and recombination of Frenkel pairs, have been considered. It has been shown that the second in contribution to the self-diffusion after the vacancy mechanism in the metals under consideration is the migration of divacancies. The third is the mechanism involving the formation and recombination of dynamic Frenkel pairs. The cyclic mechanisms (with simultaneous atomic displacements) and the vacancy migration directly to the second coordination sphere in face-centered cubic metals are unlikely.

Contributions of different mechanisms of self-diffusion in face-centered cubic metals under equilibrium conditions

The study of diffusion of hydrogen atoms in the crystal lattice of fcc metals Pd, Ni, Al, Ag was performed by the method of molecular dynamics. The diffusion characteristics of hydrogen impurity (activation energy of hydrogen atom migration and pre-exponential factor in the Arrhenius equation) in the considered metals were calculated. It is shown that the prevailing mechanism of the over-barrier hydrogen diffusion in fcc metals consists of successive migration through octahedral and tetrahedral pores. During migration, as a rule, the hydrogen atom is not delayed in tetrahedral pores.

Molecular dynamics simulation of hydrogen atom diffusion in crystal lattice of fcc metals

It is known that, in a molecular dynamics model of Pt3Al crystal, a discrete breather (DB) with soft type nonlinearity (DB1) can be excited, which is characterized by a high degree of localization on a light atom (Al), stationarity, as well as a frequency that lies in the gap of the phonon spectrum and decreases with increasing amplitude of the DB. In this paper, it is demonstrated that a DB with hard type nonlinearity (DB2) can be excited in a Pt3Al nanofiber; this DB is localized on several light atoms, can move along the crystal, and has a frequency that lies above the phonon spectrum and increases with the DB amplitude. It is noteworthy that the presence of free surfaces of a nanofiber does not prevent the existence of DB1 and DB2 in it. Collisions of two DBs counterpropagating with equal velocities, as well as a collision of DB2 with a standing DB1, are considered. Two colliding DBs with hard type nonlinearity are repelled almost elastically, losing only insignificant part of their energy during the interaction. DB2 is also reflected from a standing DB1; in this case, the energy of the breathers is partially scattered into the Al sublattice. The results obtained indicate that DBs can transfer energy along a crystal over large distances. During the collision of two or more DBs, the energy localized in space can be as high as a few electron-volts; this allows one to raise the question of the participation of DBs in structural transformations of the crystal.

Simulation of the interaction between discrete breathers of various types in a Pt 3 Al crystal nanofiber

By means of molecular-dynamics simulations, it is shown that discrete breathers are excited in two-dimensional model lattice of the regulated Pt3Al alloy by periodic external action of a certain frequency. The energy of Al atoms carrying a non-linear localized mode can more than ten times exceed the mean energy of Pt atoms.

Localized oscillating modes in two-dimensional model of regulated Pt 3 Al alloy

The phenomenon of dynamic collective displacements of atoms in face-centered cubic crystals has been revealed using molecular dynamics method. This phenomenon plays an important role in the vacancy mechanism of diffusion. The vacancy mechanism is provided by the collision of two regions of collective atomic displacements that move a migrating atom and a vacancy toward each other. The collective thermal atomic displacements from crystal lattice sites occur as a result of the nonuniform momentum distribution of atoms according to the Maxwellian distribution. Owing to their statistical nature, the degree of correlation of the atomic displacements depends neither on the temperature nor on the interatomic interaction potential.

Mikhail D. Starostenkov

Papers

Contributions of different mechanisms of self-diffusion in face-centered cubic metals under equilibrium conditions

Molecular dynamics simulation of hydrogen atom diffusion in crystal lattice of fcc metals

Simulation of the interaction between discrete breathers of various types in a Pt 3 Al crystal nanofiber

Localized oscillating modes in two-dimensional model of regulated Pt 3 Al alloy

Dynamic collective displacements of atoms in metals and their role in the vacancy mechanism of diffusion