Thank you very much for your comments. Please forgive my ignorance.
Now I revised (only one ion is added into the water vapor). I still have a question: whether the correction proposed by Bogusz.(1998) can reduce artifacts of long-rang force calculation during the simulation, or it only works when the compute group/group command is used to compute the energy that I need (I don’t really care the energy). Let me put it another way, when I do not use the “compute group/group command”, can LAMMPS give a long with the correction proposed by Bogusz.(1998)? Whether the long-range contribution to the force acting on the particles is ill-defined?
Thank you very much for your comments. Please forgive my ignorance.
Now I revised my physical model (only one ion is added into the water vapor). I still have a question: whether the correction proposed by Bogusz.(1998) can reduce artifacts of long-rang force calculation during the simulation, or it only works when the compute group/group command is used to compute the energy that I need (I don’t really care the energy). Let me put it another way, when I do not use the “compute group/group command”, can LAMMPS give a long range Coulomb force with the correction proposed by Bogusz.(1998)? Whether the long-range contribution to the force acting on the particles is ill-defined?
Your questions indicate that you (still) do not understand ewald summation sufficiently. You have to change that. Then you will also be able to answer your question yourself and be able to properly argue whatever model you choose.
Also, you have to understand that in lammps compute commands are for analysis and do not affect the simulation, so your question makes little sense to begin with.
To reiterate, the case of a non-neutral system is handled by all lattice sum based methods via ignoring the non-periodic contribution to the reciprocal sum. Whether that is acceptable or not depends very much on the specifics of a given system, the required accuracy and the options to validate the results. Thus a general answer as you are asking for is simply not possible.
As for the specifics of the lammps implementation, Stan has already answered that and for more details you will have to read the source code.
The correction is only for energy, not force and yes it is applied to Ewald and PPPM (you don’t need to use compute group/group). I’d suggest doing a literature search on this topic.