Default q-values for spin-polarized calculations with odd number of electrons

[tfrederiksen]

Spin-polarized calculations for molecules with an odd number of electrons should reflect this difference in the populations of the two spin components, rather than operating with fractional q-values. This branch fixes this by forcing integer division.

[sofiasanz]

Agreed. Actually we should also modify the case for unpolarized systems since we face the same problem there, no? The thing is that a system with an odd number of electrons has an open-shell solution, should we then allow the code to obtain the closed-shell solution (unpolarized) for such systems?

[tfrederiksen]

The unpolarized case is fine as it is, here the total number of electrons must be equally distributed between the two spin-components.

[sofiasanz]

OK I thought that maybe it was a bit unnatural to simulate it like that, but I guess we can do such calculation. I think this branch is ready to go.

[tfrederiksen]

The unpolarized case would be treated the same way in any DFT code. The use need to understand what he/she is doing.