Stillinger-Weber silicon (1985)

StillingerWeberSi (PotType::SWSi1985) evaluates the silicon potential of F. H. Stillinger and T. A. Weber, Phys. Rev. B 31, 5262 (1985), https://doi.org/10.1103/PhysRevB.31.5262.

The two-body and three-body parameters are equation (2.7). Sigma is 0.20951 nm. Epsilon is the value equation (2.9) prints, 3.4723e-12 erg per atom pair. With 1 erg = 1e-7 J and the CODATA electron-volt (1.602176634e-19 J) that is 2.16723920 eV. PotType::SWSi, the Fortran SWPot kernel, is not this potential. It uses eOn’s rescaled epsilon, 2.16826 × 1.06767385 = 2.3149945 eV. The ratio of that epsilon to the published value is 1.068176741337.

The product is copied from client/potentials/SW/SWFortran.f90 in TheochemUI/eOn. The file header says “Last change: AA 3 Apr 2002”, and the same lines are present at tags svn-Dec16_2020 and v2.8.0:

EPSILON = 2.16826D0 * 1.06767385d0, &  !* 1.07d0
LAMBDA  = 21.0d0,                   &  !* 1.07d0

Lambda stays 21 and is not multiplied. Neither that file nor the rgpot port (commit 7a3b5016, per rgpot_sw.f90) derives 1.06767385. The factor is untraced.

Calls

Lengths are Angstrom and the energy is eV. The cell is periodic in each direction. forceImpl writes the force, which is minus the gradient. energyAndGradient writes the energy and dE/dx for one configuration. energyAndGradientBatch writes both for several configurations that share one cell and one species list, so a caller enters the library once per batch. The Python method energy_gradient accepts positions of shape (n_atoms, 3) or (n_config, n_atoms, 3) and uses that entry.

Every call takes its pairs from the instance’s shared neighbour layer (nlist::updateCells). The skin is that of a nlist::NeighborList passed to energyAndGradient when it is positive. Otherwise it is one fifth of the cutoff, or 0 under setNeighbourMode(nlist::Mode::Stream). The pairs are reused while every atom has moved by at most half the skin and the cell and the atom count are unchanged. With skin 0 they are reused only for identical coordinates. energyAndGradientBatch passes no nlist::NeighborList.

Stress

forceImpl writes the Cauchy stress, sets has_stress, and caps().stress is true. The convention is the one of LJPot, MorsePot and ZBLPot, set in stress.hpp. The stress is sigma = dE/dε / V for the strain that sends every position and every cell row r to r (I + ε). It is row-major and symmetric, in eV/Angstrom3. A compressed cell has negative diagonal entries, the sign ASE uses. For the LAMMPS virial W, sigma = -W / V.

The virial adds d ⊗ g for the folded bond vector d of each pair term and for both bonds of each three-body term, where g is the energy gradient with respect to d. That sum is the many-body virial of Thompson, Plimpton and Mattson, J. Chem. Phys. 131, 154107 (2009), https://doi.org/10.1063/1.3245303. energyAndGradient and the batch call skip the virial.

SWSiTest checks all nine entries against a fourth-order central difference of the energy in the strain, step 2e-4. The cells are compressed cubic diamond, a sheared diamond cell and a rattled cubic cell. The tolerance is 5e-10 eV/Angstrom3.

The Fortran SWPot writes the same stress from the folded vectors of its own neighbour table, and its caps().stress is true. Both kernels are linear in epsilon, so the SWPot stress is the StillingerWeberSi stress times 1.068176741337. FortranPotsTest checks that ratio to 1e-14 eV/Angstrom3, and the strain difference to 5e-10, on a compressed cubic cell and a sheared, rattled cell. SWPot::forceBatch, the call eOn’s engine makes, skips the virial, and its outputs have has_stress 0 and a zero tensor. FortranPotsTest checks that its energies and forces equal those of forceImpl.

In Python, evaluate on StillingerWeberSi or SWPot returns the stress as a 3 × 3 array, and caps() reports it. RgpotCalculator then gives the ASE stress for either potential.

Diamond energy

At the tetrahedral angle the three-body term is zero, and the diamond minimum puts every nearest neighbour at the minimum of the pair term, where that term equals -1. Each atom then has two bonds, so the energy per atom is -2 epsilon. Equation (2.8) prints -1.999993 epsilon for the lattice sum at the density quoted there. SWSiTest checks both figures, and checks the analytic gradient against a finite difference.