qeq - charge sign

Dear colleagues,
I tried to use the qeq/reax fix (together with reax/c force field). The results make sense (despite the warning "changed valency_val to valency_boc for X" which I do not understand).

However, there are unphysical charge signs in the final results, namely negative Al and positive O atoms in Al2O3 (although I started with positive Al and negative O). This does not affect the coulombic forces of course, but nevertheless I want to be sure if it does not cause any problems when calculating the short-range reaxff forces. Or is there any way how to get rid of that?

Thanks a lot,
Jiri Houska
(now at) RWTH Aachen University

Hi Jiri,

The warning is about inconsistent values in the force field file for
element X, which does not cause any problems and will go away if you get
rid of X in the ffield file.

Did you find several negative Al and positive O atoms by looking at the
dump files? If yes, it is probably okay and is a result of the force
field parameters and QEq. It would be bad if the averaged per-type charges
for Al and O are reversed signs.

Short-range terms are charge-independent so that they are not affected by
charges. The only two terms that are charge-dependent are self ionization
energy and Coulombic.