Skip to main content
Dryad logo

Molecular dynamics trajectories for ionic conductors in: Paradigms of frustration in superionic solid electrolytes

Citation

Wood, Brandon et al. (2020), Molecular dynamics trajectories for ionic conductors in: Paradigms of frustration in superionic solid electrolytes, Dryad, Dataset, https://doi.org/10.5061/dryad.j3tx95xc3

Abstract

Superionic solid electrolytes have widespread use in energy devices, but the fundamental motivations for fast ion conduction are often elusive. Here, we draw upon atomistic simulations of a wide range of halide, oxide, sulfide, and closo-borate superionic conductors to illustrate some of the key features that enhance local cation mobility in these solids. We classify three types of frustration that create competition between different local atomic preferences, thereby flattening the diffusive energy landscape and enhancing entropy. These include chemical frustration, which derives from competing factors in the anion-cation interaction; structural frustration, which is connected to the lack of a clear site preference for mobile ion ordering; and dynamical frustration, which is associated with temporary fluctuations in the energy landscape due to anion orientations or cation reconfigurations. For each class of frustration, we provide detailed simulation analyses of multiple materials to show how ion mobility is facilitated, resulting in stabilizing factors that are both entropic and enthalpic in origin. Implications for identifying suitable descriptors for superionic conductivity are discussed.

Methods

Ab initio molecular dynamics trajectories (xyz format, zipped) collected on Livermore Computing facilities and on the Stampede2 system at the University of Texas, Austin. All simulations used Perdew-Burke-Ernzerhof (PBE) pseudopotentials with Gamma-only k-point sampling in the NVT ensemble.

AgI, CuI, Li2B12H12, Li2B10H10, Na2B10H10, Li10GeP2S12: Trajectories are from Car-Parrinello molecular dynamics simulations using the CP module within Quantum Espresso using ultrasoft pseudopotentials from the Quantum Espresso standard pseudopotential library. Nose-Hoover chains were used to maintain temperature.  

Li3PS4: Trajectories are from Born-Oppenheimer molecular dynamics simulations using the Vienna Ab-initio Simulation Package (VASP) and projector augmented wave pseudopotentials. A Nose-Hoover thermostat was used to maintain temperature. 

Usage Notes

Li3PS4: 50 ps, 700 K, 128 atoms, lattice parameters a = 12.78 Å, b = 15.74 Å, c = 12.04 Å, time step = 1.0 fs, output every frame.

AgI (alpha phase): 34.5 ps, 700 K, 108 atoms, lattice parameter a = 15.46 Å, electronic mass = 400 a.u., time step = 0.29 fs, output every 10 frames (2.9 fs).

CuI (alpha phase): 37 ps, 700 K, 108 atoms, lattice parameter a = 14.34 Å, electronic mass = 800 a.u., time step = 0.19 fs, output every 10 frames (1.9 fs).

Li2B12H12 (beta phase, fcc): 51 ps, 800 K, sqrt(2) x sqrt(2) x 1 supercell, lattice parameters = 14.21 Å, = 10.05 Å, electronic mass = 400 a.u., time step = 0.145 fs, output every 10 frames (1.45 fs).

Li2B10H10 (fcc): 55 ps, 800 K, sqrt(2) x sqrt(2) x 1 supercell, lattice parameters = 14.51 Å, = 10.26 Å, electronic mass = 400 a.u., time step = 0.145 fs, output every 10 frames (1.45 fs).

Na2B10H10 (fcc): 60 ps, 800 K, sqrt(2) x sqrt(2) x 1 supercell, lattice parameters = 13.96 Å, = 9.87 Å, electronic mass = 400 a.u., time step = 0.145 fs, output every 10 frames (1.45 fs).

Li10GeP2S12: 47 ps, 700 K, sqrt(2) x sqrt(2) x 1 supercell, lattice parameters = 12.69 Å, c = 12.45 Å, electronic mass = 400 a.u., time step = 0.145 fs, output every 10 frames (1.45 fs).

Funding

Vehicle Technologies Office, Award: DE-AC52-07NA27344

National Science Foundation, Award: DMR-1710630