Simulation Graphics

Selected results from recent accelerated molecular dynamics simulations.

TAD simulation of deposition of Cu onto Ag(100), ~ 0.4 ML to 0.5 ML.

This movie shows a temperature-accelerated dynamics (TAD) simulation of deposition of copper onto a silver (100) surface at T=77K and 0.04 monolayers per second, matching the deposition conditions of a 1989 experiment by Egelhoff and Jacob. This simulation, the work of Jim Sprague and Francesco Montalenti, took a few days of workstation time for the segment shown (roughly 0.4 monolayer to 0.5 monolayer), and corresponds to a few seconds of deposition time (a boost factor in the millions relative to direct molecular dynamics). The movie is in a perspective view, with 8 frames per event and repeated events shown only once (some actually repeated thousands of times). Time is distorted because successive depositions appear to come further apart when thermal events occur in between. In the simulation, each deposition event was performed with direct molecular dynamics, allowing steering and athermal reactive events. In the movie, the deposition "event" is simply an animation that moves the atom towards the surface site in which it came to rest in the actual MD. [J.A. Sprague, F. Montalenti, J.A. Sprague, F. Montalenti, B.P. Uberuaga, J.D. Kress, A.F. Voter, Phys. Rev. B, in press (2002).]

TAD simulation of deposition of Cu onto Ag(100), 0 ML to 1.1 ML.

As above, showing the deposition up to about 1.1 ML at 4 frames per event, perspective view. As above, time is distorted somewhat, but basically this movie runs at about the real-time speed of the experiment. Some of the thermal events occurred hundreds or thousands of times. To keep the movie from being prohibitively long, each repeated event is displayed 1 + log10(N) times, where N is the number of times it actually repeated. [J.A. Sprague, F. Montalenti, J.A. Sprague, F. Montalenti, B.P. Uberuaga, J.D. Kress, and A.F. Voter, Phys. Rev. B, in press (2002).]

Stretching a carbon nanotube with driven parallel replica.

Here, a nanotube with a single vacancy has been stretched at a strain rate of 1.d6 per second using the parallel-replica method on 40 1-GHz Pentium processors. A REBO interatomic potential has been employed and the temperature is T=300K. Basically nothing happens, not even thermal diffusion of the vacancy, until a strain of 32% (0.32 microseconds), at which point an event near the vacancy triggers an avalanche of follow-on events. This movie shows one picosecond as the excitement starts. Undercoordinated atoms are shown in red, and overcoordinated atoms are green. [B.P. Uberuaga, S.J. Stuart and A.F. Voter (2002), to be published.]

TAD simulation of a buckyball with a vacancy

Here, a buckyball with a single vacancy is evolved using temperature accelerated dynamics (TAD) for 23 ms at T=1000K. The TAD high temperature is 3000K. The vacancy is seen to reconstruct in various ways, but never actually migrates in this time at this temperature. [B.P. Uberuaga, S.J. Stuart and A.F. Voter (2002), to be published.]

Evolution of a 60-carbon fragment sliced out of a carbon nanotube.

Here, 60 atoms have been artificially (and perfectly) sliced out of a carbon nanotube. The system is described using a REBO interatomic potential and the temperature is T=1500K. Using temperature accelerated dynamics (TAD) with a high temperature of T=3000K, the system is evolved for 2 microseconds. During this time, both ends cap off, but the final structure is not a perfect buckyball (as it might be if the slice was taken from a smaller-diameter nanotube). At 14 microseconds, viewing the system as it rotates shows it is not spherical. Undercoordinated atoms are shown in red, and overcoordinated atoms are green. [B.P. Uberuaga, S.J. Stuart and A.F. Voter (2002), to be published.]

Return to Home Page for

Arthur F. Voter   |   T-12 Group   |   Theoretical Division   |   Los Alamos National Laboratory