Questions and problems?

Polarization issues

Is it possible to get energies of interaction between different groups of atoms (defined using group input for example) and include polarisation effects?

There's no way to get strictly pairwise interactions out of a non-additive run. These effects are intrinsically many body, so one can't assign a pair interaction to any specific pair. Note: algorithmically, global polarization is calculated first and then the energy is calculated in a pairwise manner, so one could in principle apply group analysis to the second stage, where the resulting energies would be "for the internally polarized system" or some such.