LAMMPS hybrid meam+airebo

LAMMPS LAMMPS Beginners
1

Dear All

I want to the hybrid function(meam+airebo). The C-C uses meam potential, C-H and H-H use airebo potential. the potential setting is followed:

pair_style hybrid meam airebo 3.0
pair_coeff * * airebo CH.airebo C H
pair_coeff * * meam library.meam Si C SiC.meam C NULL

mass 1 12.0107
mass 2 1.0080

Then, I test the code only for C-C, but the result is not same with that just using the meam potential.
Could anyone give some advices to solve the problem?

Best wishes,
Yanyan

2

This is a very bad idea. Both MEAM and AIREBO are manybody potentials, so you cannot divert only C-C interactions. If you would want to have a different pairwise C-C interaction, you would have to create a modified AIREBO pair style would incorporate the MEAM interaction there. But then you would - again - have an inconsistent model, since MEAM considers many-body interactions differently and that would double count those interactions unless you would only incorporate the pair-wise part of MEAM. And even if you would manage to do that, you would still have a “zombie” potential that is unverified and inconsistent.

Bottom line, for a system with C and H, just use AIREBO and your problems are gone.

3

Dear akohlmey

But when I use AIREBO to simulate the C-C bond, the result don’t meet expectation without considering H atom. The result of MEAM potential meets expectations. Moreover, I also need to add H atom to the simulation. So I want to use MEAM potential and add H atom. Could you give me some advice? Or if could I set LJ potential about C-H/H-H to the simulation?

Best wishes,
Yanyan

4

Given what I know about how AIREBO has been parameterized and validated (please read the paper(s)), I would say that it is much more likely that your validation is incorrect.

All your considerations have the same consequence: you are creating a new parameterization and are doing it without having the experience to do this. This is a recipe for disaster.

My suggestion remains the same. Use only AIREBO. Hybrid models how you describe them are flawed by construction and are ignoring the manybody nature of the AIREBO potential.