I'm trying to analyse doping in a certain material in low concentrations. I'm approaching the problem by taking crystallographic unit cell of that material, then expanding that to a 2x2x2 supercell and then substituting one or more atoms with other element. Changing amount of atoms substituted I can control the doping level.
If I choose to substitute one atom in a 32 atom supercell, I obtain 3.125 % doping. If I substitute two atoms, I get 6.25% doping and so on. But I cannot know beforehand which lattice points will be preferable, so I have to calculate all possible configurations. For 2 atoms, it means 240 possible structures, but most of them are repeated. For a simple example, although these two configurations presented below are the same, in P1 supercell they are different. For two atoms, I'm able to generate close to full set of configurations, but for higher number of heteroatoms it gets tedious.
My question is mostly practical. I suppose I should use symmetry operations to reduce the number of configurations, but how may I do it practically?