Dear LAMMPS users,
I have constructed a polymer model using Materials Studio using CVFF forcefield. I now want to transfer this model to LAMMPS. I have done so using msi2lmp. a few of the potential parameters are missing in the final LAMMPS data generated by msi2lmp. I have searched a lot and I know that this msi2lmp was developed over 10 years ago and has not been updated since except for bug fixes. and I know that I might have to find the parameters myself and add them manually. My problem is this:
These are the parameters that are missing:
Unable to find angle data for o c1 c’
Unable to find angle data for c1 c’ h
Unable to find angle data for c’ c1 oh
Unable to find oop data for c1 c’ h o’
I have gone through the cvff.frc file provided by LAMMPS thoroughly. In the angle data section, the data for these two angles are provided: (c1 is equivalent to c)
quadratic_angle cvff
E = K2 * (Theta - Theta0)^2
!Ver Ref I J K Theta0 K2
1.0 1 c o c’ 109.5000 60.0000
1.0 1 c c’ o 110.0000 122.8000
as you can see the angle data for c o c’ and c c’ o are provided but the data for o c c’ is not. Is there any way we can find the angle data for o c c’ using the data for the angles given above? I am not sure if my question is pertinent at all, so I apologize in advance if it does not seem meaningful.
Is there any updated cvff.frc file or any other potentials available anywhere that can be used instead of these outdated files?
I would appreciate any thoughts on this matter.
Sincerely,
Melika
There are many known limitations of the msi2lmp command. One of them is that it can handle the equivalences sections, but not the auto_equivalences sections.
If you look through the cvff.frc file, you should see that there are two #quadratic_angle tags: one followed by cvff and another by cvff_auto.
If you look up the atom types in the #auto_equivalence table, you can see that e.g. the c1 atom type is aliased with c_ for either apex or outer angle atom. Similarly, c’ with c_ and c’_, respectively and so on. With that information, you can look up parameters in the second list with cvff_auto parameters to see, if you can find values that are missing after applying the first table and equivalences.
If that does not help you are mostly out of luck. There may be an updated cvff.frc file bundled with your copy of material studio, but I have no idea whether this is in a compatible format. Please keep in mind, that this is a proprietary force field with unpublished parameters. There is very little that we can do.
Somebody with access to materials studio should implement a new version of msi2lmp that will take care of all force field features (it would be easier to write in Python than in C, too). That way one can confirm that the force field is correctly translated to LAMMPS by running the same simulation also with MS. However, that is a rather tedious and also thankless task.