SEAMM Forcefield Plug-in

A SEAMM plug-in for setting up a forcefield or EAM potentials for subsequent simulations.

This plug-in provides a graphical user interface (GUI) for choosing the forcefield or EAM potentials use by subsequent simulation steps.

• Free software: BSD license

• Documentation: https://molssi-seamm.github.io/forcefield_step/index.html

History

2026.9.20.1 – Warnings about the charges now appear in the output

• When the charges from the forcefield do not add up to the charge of the system they are adjusted to compensate. The warning about it went only to the log, which for a job in a queue is the queueing system’s output file if it is kept at all, so a change to the charges could pass unnoticed. It now appears in the step’s output and the job’s output, with the rest of what assigning a forcefield reports.

2026.9.20 – Bugfix: LigParGen parameters were put on the wrong atoms

• The utility that reads LigParGen’s output assumed that the atoms in the .key file were in the same order as those in the structure it was asked about. LigParGen perceives the molecule afresh and orders the atoms its own way, so for many molecules they are not. For fluoroethylene carbonate, this gave the fluorine two connections and a ring oxygen one – neither of which is possible – and handed the atoms of the fragment each other’s types, so a hydrogen was given the mass and charge of an oxygen, and a bond of the molecule ended up between two types that are not bonded and so has no parameters.

• Nothing complained, because reordering the types leaves the charges summing to the same total. The atoms of the .key file are now matched to those of the structure through the molecule’s connectivity, which the .key file itself describes, rather than by assuming the two orders agree. The entry written for a molecule is checked against the molecule before it is added: that every atom gets a type of its own element, that no type claims more bonds than its element can have, and that every bond of the molecule is one the LigParGen file has parameters for.

• Forcefield files generated before this will have the same problem, and should be regenerated.

2026.2.26 – Internal: moving from pkg_resources to importlib.resources

2026.1.28 – Bugfix: ligpargen failed for some molecules

• ligpargen crashed for molecules with no improper torsions.

2025.7.22 – Allowing the forcefield to be specified using a variable.

• The graphical code forced the selected forcefield to be one of the available forcefields, not allowing for the use of a variable. This has been fixed.

2025.6.23 – Bugfix: corrected handling of personal forcefield files.

2025.6.20 – Adding support for MACE PyTorch potentials

2025.5.26.1 – Corrected the references in the ReaxFF files.

• The references in the ReaxFF forcefields included with the release were not correctly formatted so were being ignored.

2025.5.26 – Updated the OPLS-AA forcefield file to allow personal versions

• Updated the OPLS-AA forcefield to support an optional local user file ‘my_oplsaa.frc’ to add to or override parameters in the main file.

• Fixed issues in the ligpargen utility due to changes in the SMILES routines in SEAMM.

2025.4.7 – Added a number of ReaxFF forcefields.

• Added the ReaxFF forcefields from LAMMPS, plus two from lithium battery work.

• Added metadata sections to the forcefield to better support the features of different forcefields.

• Improved the listing and display of available forcefields, and made it dynamic so that forcefields added locally are immediately found without needing to restart SEAMM.

2025.4.1 – Added support for ReaxFF, which does not need atom type assignment.

2025.3.18 – Added bis(fluorosulfonyl)imide anion (FSI) to the OPLS-AA and CL&P forcefields

• Added parameters for the bis(fluorosulfonyl)imide anion (FSI) to the OPLS-AA and CL&P forcefields.

• Added a test for FSI in the test suite.

2025.3.16 – Added the Dreiding forcefield

2024.12.14 – Updated for changes in SMILES handling

• Changes in the handling of SMILES in MolSystem required small changes in the tests to continue using Open Babel for SMILES.

2024.10.25 – Changed the default forcefield to OPLSAA

2024.6.30 – Bugfix: Error submitting jobs with local forcefield files.

2024.6.29 – Bugfix: factor of 2 for dihedrals and impropers in ligpargen

• The ligpargen tool was missing a factor of 2 in the dihedral and improper parameters.

• Corrected the search paths for forcefields.

• Improved the documentation.

2024.6.28 – Added customizable local forcefields and LigParGen interface.

• Added the machinery to handle local forcefield files in either ~/.seamm.d/data/Forcefields (personal) or ~/SEAMM/data/Forcefields (site).

• Added ‘ligpargen’ command to access custom parameters from the LigParGen service at Yale University, ading them to the ‘ligpargen.frc’ personal forcefield, which is automatically included in ‘oplsaa.frc’ if it exists.

2024.1.10 – Fixed PF6- issue in CL&P forcefield

• The angle parameters for PF6- in the CL&P forcefield only work if the 180º F-P-F angles are not included in the calculation. Replacing them with an equivalent periodic SHAPES-like potential almost works; however, since 0º is a valid angle and there are no 1-3 nonbonds, nothing keeps the F atoms apart. This is solved using a tabulated potential based on the SHAPES potential but with an added 1-3 repulsion large enough that the gradient is always pusing small angles apart, but not large enough to affect the minimum at 90º.

2023.9.14 – Fixed errors! And added C2mim to test.

• The units of the torsions were incorrect in the last implementation.

• Added parameters for 1-alkyl-3-methylimidazolium cations from JCP 108, 2038 (2004)

• Tested much more thoroughly.

2023.9.13 – Added parameters for TFSI to CL&P/OPLSAA

• Parameters for TFSI - bis[(trifluoromethyl)sulfonyl]imide

2023.9.8 – Added more typing for OPLS-AA

• cyclopropane -CH2-, -CHR-, and -CR2-

• hexafluorobenzene

• difluorobenzene

• bromobenzene

• iodobenzene

• thiophenol

• alkyl nitriles

• nitroalkanes

• nitrobenzene

• methylene in phenylacetonitrile

• corrections to methylene nitrile anion

2023.9.7 – Added typing in OPLS_AA for fluorobenzene

2023.9.6 – Fixed issue with PF6- geometry

• The Lennard-Jones repulsive term added to the F-P-F angle was too weak, allowing the structure to get trapped in a symmetric state with ~40º angles.

2023.8.27 – Fixed issue with angle in octahedral systems

• The SHAPES-type simple fourier potential used for octahedral complexes has a fals minimim at 0º. Added a LJ 1/R^12 repulsive term between the two end atoms of the angle to prevent small angles. This required using tabulated potentials in LAMMPS.

2023.5.1 – Fixed bug in Lithium battery forcefield

• Fixed a typo in the angle type unit line which caused a crash

2023.4.6 – Added Lithium battery forcefield

• An initial set of parameters for cathode materials, specifically LiCoO2.

2023.2.13 – Added OPLS-AA forcefield

• Added parameters for OPLS-AA along with some extra parameters for ionic liquids * PF6- * ethylene carbonate (EC) and fluoronated EC (FEC)

• Added atom-typing templates for most of OPLS-AA. Still missing a few and amino acids and DNA not yet tested.

• Added extensive, almost-complete testing, for OPLS-AA