You can try instead to see how the critical temperature changes when nonmagnetic impurities are added using the classical work by Prof. A. Larkin
Article Vector Pairing in Superconductors of Small Dimensions
In the case of the gap equation calculated self consistently using Matsubara frequencies, the trouble is that the numerical procedure uses a normalized temperature and a normalized gap.
Generally when we fit superconducting data, we use normalized units respect to the normal state.