Diffusive molecular dynamics simulations of lithiation of silicon nanopillars

Abstract

We report diffusive molecular dynamics simulations concerned with the lithiation of Si nano-pillars, i. e., nano-sized Si rods held at both ends by rigid supports. The duration of the lithiation process is of the order of miliseconds, well outside the range of molecular dynamics but readily accessible to diffusive molecular dynamics. The simulations predict an alloy Li 15 Si 4 at the fully lithiated phase, exceedingly large and transient volume increments up to 300% due to the weakening of Si-Si iterations, a crystalline-to-amorphous-tolithiation phase transition governed by interface kinetics, high misfit strains and residual stresses resulting in surface cracks and severe structural degradation in the form of extensive porosity, among other effects.

MXE: A package for simulating long-term diffusive mass transport phenomena in nanoscale systems

“…During the diffusive step, we fix the atomic positions and the frequencies, whereas the atomic molar fraction is updated by integrating Equation 22. The integration time step is several orders of magnitude larger than the critical time step in MD.…”

“…This approach has been successfully used in the past [56,19,22]. and M is the total number of species in the system.…”

“…and M is the total number of species in the system. Following [19,22], we approximate the phase average x jl f (r ij ) 0 , (B.11) respectively, where x ik is the atomic molar fraction of specie k at site i. Similarly, the phase average of the embedding function is treated in a similar way as described above for pair potentials.…”

“…Unfortunately, this is not a trivial task, and previous implementations have been based on specific scripts that modified MD codes [18], or in specific, in-house codes restricted to targeted applications [20,21,22], ultimately hindering widespread usage of such techniques.…”

“…The implementation uses a recently developed model for simulating transport phenomena in discrete systems based on a Fokker-Planck equation [24]. This new approach was inspired on the kinematic mean field theory [25], and is analogous to transport laws used for mass transport in other methods [26,18,22]. The proposed kinematic law for transport is then coupled to a non-equilibrium statistical mechanics formulation for atomic systems [27,28,19], allowing the coupled thermo-chemo-mechanical simulation of heat [29,30,31,32] and mass transport at the nanoscale [20,21,22].…”