Atomtypes and charge

Hi Lammps Users

This might be a weird question. Can two atomtypes have different charge values ? In a normal scenario one would not do this. In my simulation I have a system of charged polymers counterions and salt. The atom types of positive counterions is same as that of the negative salt ions but charge wise they are different. This was a stupid error but have already burnt quite a lot of resources. I can recompute all by setting appropriate charges and atomtypes.

At this point I am curious to know whether the atom types and charge settings are intrinsically tied which means the same atomtype can never have two different charge values? I am not able to think of a scenario (apart from postprocessing) where it would make a difference if not accounted carefully.

This is my thought. If charge is stored as an array of length equal to atom types I agree this will be a problem. If it’s stored as an array of length equal to number of atoms it should not matter I believe.

Thanks in advance for your response.

Vaidyanathan M S
Department of Chemical Engineering and Materials Science

University of Minnesota Twin Cities

Hi Lammps Users

This might be a weird question. Can two atomtypes have different charge values ?

yes.

In a normal scenario one would not do this.

on the contrary. this is quite common as most biomolecular force
fields are built like this. the atom type defines the atoms according
to its "chemical function" in "functional groups" (e.g. a carbonyl
group or a secondary amine) and thus define lennard-jones and bonded
interaction parameters (bonds, angles, dihedrals), but the (partial)
charges for those atom types can be different according to the
different polarity of specific groups, say an amine attached to a
saturated hydrocarbons or attached to a benzene ring.

In my simulation I have a system of charged polymers counterions and salt. The atom types of positive counterions is same as that of the negative salt ions but charge wise they are different. This was a stupid error but have already burnt quite a lot of resources. I can recompute all by setting appropriate charges and atomtypes.

At this point I am curious to know whether the atom types and charge settings are intrinsically tied which means the same atomtype can never have two different charge values? I am not able to think of a scenario (apart from postprocessing) where it would make a difference if not accounted carefully.

atom types select the non-bonded parameters that are set via
pair_coeff statements, charges are set per atom. there is no intrinsic
connection between them.

axel.

Thanks Axel.