EFOR
EFOR = [real array]
Default: EFOR | = 3 * NIONS * 0.0 |
Description: EFOR sets external forces in eV/Å on each atom in the cartesian x-, y-, and z-directions. The order of the ions is equivalent to the order in the POSCAR file, and for each ion, x-, y-, and z-components have to be given.
The array of external forces will be added to the forces computed internally for each ionic step (ionic minimization or molecular dynamics). If an ionic minimization is performed with non-zero external forces, the converged structures will be only at a minimum accounting for the constraint of the external forces. Regardless if the convergence criterion is set to minimize total forces, or converge the total energy.
Mind: The sum of all external forces set with EFOR must be 0 to avoid drift. |
Setting external forces
EFOR can be a very long array, as 3 values need to be set for each ion in the system. The syntax is equivalent to setting the magnetic moments in the noncollinear case, so it is possible to use an N*F
syntax to indicate that the next N entries in the array should be of value F. E.g.:
EFOR = 2*0.0 1.0 188*0.0 -1.0
In this example, we have 64 ions, so we have to set 3*64=192 values in the array. Only the cartesian z components of the first and last ion are non-zero, enabling a compact INCAR line.
If more force components are non-zero, it is better for readability to utilize the multi-line option in the INCAR file using backslashes (\) to negate line breaks, e.g. for a 32 atom system with forces in the x- and y-directions on the first and last 8 atoms:
EFOR = 2.0 0.0 1.0 \ 2.0 0.0 1.0 \ 2.0 0.0 1.0 \ 2.0 0.0 1.0 \ 0.0 0.0 1.0 \ 0.0 0.0 1.0 \ 0.0 0.0 1.0 \ 0.0 0.0 1.0 \ 48*0.0 \ -1.0 0.0 -1.0 \ -1.0 0.0 -1.0 \ -1.0 0.0 -1.0 \ -1.0 0.0 -1.0 \ -1.0 0.0 -1.0 \ -1.0 0.0 -1.0 \ -1.0 0.0 -1.0 \ -1.0 0.0 -1.0
Symmetry and periodic boundary conditions
Tip: External forces in a system with periodic boundary conditions are typically only useful if there is a vacuum region in the simulation cell to break the symmetry. |
External forces during on-the-fly learning
External forces can be used during on-the-fly training for machine-learned force fields. In that case, the external forces will be only used during the ionic updates but will be removed for the training. This ensures that only forces arising from interatomic interactions will contribute to the force field and the resulting force field can be used without, or with different external forces.