Hi,
I tried to apply an electric field to the ReaxFF system while using “fix qeq/reaxff” to balance the charge, and I saw some articles (https://doi.org/10.1016/j.combustflame.2022.112238) that used periodic boundary conditions in all directions, but when I tried it myself, lammps got the following error message:
“Must not have electri field component in direction of periodic boundary when using charge equilibration with ReaxFF”
I would like to ask if there is a problem with my setting, and how to set this case?
Any help or reference would be appreciated
thank you
1 Like
I can see how tracking “itinerant polarization” would be a problem, to use the technical term. Consider a ten angstrom wide box with two charges of +/- 0.5e, 1 angstrom apart, on opposite sides of a periodic boundary. Is their pair dipole moment +1 electron-angstrom, or -9 electron angstrom? This matters because under an electric field the energy of their charge separation is (dipole moment) dot (electric field), and to minimise energy (which is what QEq tries to do) you need to include the field-dipole energy, and you cannot unambiguously minimise an ambiguous quantity.
Actually to me it seems that LAMMPS can treat this unambiguously if the minimum image convention is obeyed (that is, the box is larger than twice the cutoff so each pair interaction has an unambiguous nearest-neighbor assignment [note that LAMMPS does not require MIC due to its ghost particles system, other than in this very specific example]). In that case there is no periodic dipole ambiguity. But I suppose the ReaxFF authors wanted to play it safe.
Thank you very much for your answer. I think I have roughly known that this will lead to ambiguity. However, how to achieve this is to use ReaxFF force field, apply electric field, balance charge and use period boundary. Do you know this? Thank you very much
You need to ask the author of the publication. I suspect they use a modified version of LAMMPS and have not yet contributed the corresponding functionality back to the LAMMPS distribution. The paper mentions a QTPIE charge equilibration method and that is not available in LAMMPS, only QEq and ACKS2.
The error message in LAMMPS is a for a good reason and thus you cannot do what the (available) code does not support.
Of course, this brings up the question of whether people should be allowed to publish results without people being able to independently reproduce them. But discussing that is beyond the scope of this forum.
2 Likes
Thank you very much for your question and answer, I understand this now
Some discussion here: https://doi.org/10.1021/acs.jctc.1c00975, see section II.IV. External Electric Fields with Charge Models.
1 Like
QTPIE charge equilibration mentioned in the paper the OP referenced is now available in LAMMPS, see
Add fix qtpie/reaxff by navlalli · Pull Request #4312 · lammps/lammps · GitHub.