Skip to content

Latest commit

 

History

History
30 lines (18 loc) · 2.56 KB

File metadata and controls

30 lines (18 loc) · 2.56 KB

This directory contains examples of all-atom simulations using the DREIDING force field.

As of 2020-10-19, this software has been tested for consistency between LAMMPS and Materials Studio using only two simple molecules (butane and ethylene).

This software is experimental and the force-fields and equilbration protocols have not been tested carefully. There is no gaurantee that simulations prepared using moltemplate will reproduce the behavior of Materials Studio (or other MD codes).

Please report any bugs you find. (jewett.aij @ gmail.com)

The atomic charges in these examples are not correct

Some force-fields (like COMPASS, and moltemplate's version of OPLSAA) include rules for assigning partial charges to atoms. Most force fields, including DREIDING do not. So DREIDING users will have to obtain atomic charges by some other means, perhaps by using 3rd-party tools.

WARNING: In these examples, I obtained partial charge estimates from the OPLSAA parameter file located here.) PLEASE DO NOT DO THIS! This is not how the DREIDING force field should be used. It will probably not result in accurate behavior.

Alternatively LAMMPS' fix qeq/point feature can be used to assign partial charges. QEq is one of the built-in methods used by Gaussian to assign partial charges to atoms in molecules that are modeled using the DREIDING force field. If this fix is run infrequently (or if it is run only once at the beginning of the simulation), then it should not slow the simulation down significantly.

In addition to the QEq method, the Gasteiger method (suggested in the original DREIDING paper) is also popular.

Improper angles

The style of improper interaction used by DREIDING force fields depends on an angle which depends on the order of the atoms surrounding the central atom. When multiple atoms have the same type, this creates ambiguity in atom order. Since there is no guarantee that moltemplate will choose the same atom order as other molecule builders this can lead to small unavoidable discrepancies in energies and forces computed by LAMMPS and other simulation programs (like Materials Studio). But their effect should be negligible. (Please let us know if this is not the case.)