Question on changing the pair_style after relaxing structure

Dear All,

I have equilibrated my system while using below pair_style:

pair_style lj/cut/tip4p/long 2 1 1 1 0.1546 12.0

and kspace_style pppm/tip4p 1.0e-6 as well. Then, I used write_restart command to save my atoms coordinates.

Since centroid/stress/atom is not compatible with this pair_style, I cannot use this binary restart file. So, is there any way about using this binary file?

Or should I change the pair_style into
pair_style lj/cut/tip4p/cut 2 1 1 1 0.1546 12.0
and start relaxing my structure from the beginning?

Thanks for your comments

Maybe you can create an intermediate simulation where you read the restart file and save a lammps.data file using the write_data command (write_data command — LAMMPS documentation). You dont need to actually perform an MD simulation or minimization: I think it should suffice to do a run 0 command before the write_data. So basically it is just a way to convert the restart file into the data file (from which you can delete force-field styles that may be present). And then you can run the simulation you want the initial microstate you want via read_data.

Please see the command line option -r2data that can convert the restart file into a data file: 4.2. Command-line options — LAMMPS documentation

Please note that replacing a long-range Coulomb handling with a cutoff Coulomb defeats the purpose of using compute centroid/stress/atom, since you are replacing one source of error with a different (bigger?) one.

So, do you mean it may produce untrusted results?

Don’t take my word for it. Check the published literature! That is what it is good for. Massive amounts of accumulated knowledge. There are plenty of publications comparing the merits of the different variants of Ewald summation and their grid based approximations with cutoff Coulomb and various ways of damping and shifting.