Vacancy Site Minimizer

The "vacancy site minimizer" computes the energy associated with creating vacancies at various positions throughout a structure.

It is accessible in the "LAMMPS - Vacancy Site Min" tab of the simulation generator dialog:

Minimizer tab of the simulation builder dialog


Requirements

In order to build a minimizer simulation, a structure with a non-zero bounding box must already be set up in the viewer:

Bounding box example

Additionally, the sites that are to be computed must be marked as vacancies. Here, the starting vacancies is treated as a regular vacancy.

Note that only one vacancy is created at a time per simulation (see below).


General Options

The first set of options involve basic information about the simulation and structure:

Screenshot of the general options area in the dialog

Field Description
Working Directory Location to generate the files to
Secondary Name Name of simulation. This will decide the name of the output directory
Structure Name Name of structure. Used in deciding file names of simulation outputs
Mass Table Atomic masses to use in the simulation. Defaults to the masses defined in the atom styles table
Bounding Box Bounding box of the structure. Defaults to the bounding box already defined for the structure

Potentials Options

The next set of options controls the potential coefficients and potential style used in the simulation.

Screenshot of the potentials options area in the dialog

Field Description
Pair Coefficients Coefficients used in the pair potentials
Potential Style Style to use for the potential

In order for a simulation to be considered valid, you must have enough potentials for all elements:

Potential coefficient table with all required pair potentials Potential coefficient table with that's missing some required pair potentials

For convenience, both potential coefficients and styles can be saved and loaded. More information on saving/loading potential coefficients can be found here. Additionally, more information on saving/loading/creating potential styles can be found here.


Slurm Options

The next set of options controls some variables used by slurm.

Screenshot of the slurm options area in the dialog

Field Description
Memory Amount of memory, in megabytes, to allocate per job

LAMMPS Options

The next set of options are those specific to LAMMPS:

Screenshot of the LAMMPS options area in the dialog

Field Description
LAMMPS Version LAMMPS module name for slurm to load
KSpace Solver Name of long-range force solver for LAMMPS to use. For more information on kspace styles, see here
KSpace Accuracy Accuracy of kspace solver
Timestep Timestep used in simulations. Measured in picoseconds by default
Unit System Unit system used by LAMMPS. For more information on LAMMPS unit systems, see here. Note: unit systems besides metal have not been tested, use at your own risk!

Site Minimizer Options

The final set of options configures the site minimizer itself:

Screenshot of the site minimizer options area in the dialog

To achieve a high throughput, the simulator runs many simulations in parallel using slurm's job scheduler. Essentially, simulations are divided up into "divisions," within which several simulations are run in series. Then, each division is scheduled as a separate job using slurm.

Field Description
Sites Per Division Number of sites to simulate in each division
Division Count Total number of divisions to create

The remaining options configure the minimizer. These are identical to the settings used in the regular minimizer:

Field Description
Minimizer Style Method to use for minimizing. More information on minimizer styles can be found here
Energy Tolerance Minimizer terminates when the ratio between the energy change and energy magnitude is less than this value. Set to 0.0 to disable
Force Tolerance Stopping tolerance for length of global force vector. Set to 0.0 to disable
Max. Steps Maximum number of steps to run the minimizer
Max. Force Calls Maximum number of force evaluations before terminating the minimizer
Minimizer Timeout Maximum time, in seconds, to run an individual NEB before terminating. Useful for preventing non-converging simulations from hogging compute time