In the implementations of DFT using Gaussian basis sets, it is common to set the net charge. What exactly does this do? The Gaussian (program) manual, for instance, says that it introduces a background charge distribution made of point particles. This raises more questions than it answers for me.
1) If it's made of point particles (presumably at some internally defined coordinates), how is it a background charge distribution?
2) What exactly does setting the charge do? Is it just adjusting the electronic energy? Does it impact the geometry optimizations or the spin states given a pre-specified spin multiplciity?
I'm having trouble understanding what the charge is actually doing in terms of the DFT calculation itself. It's not a setting in plane-wave calculations, so the parallels are not clear to me.