[newchem-cpp] Metal Chemistry Shouldn't Require SNe Yields

[mabruzzo]

In the newchem-cpp, make_consistent currently requires the SNe Yields in order to infer the proper abundances of each metal species and then adjusts each metal field accordingly. It basically uses the yields in the same manner that HydrogenFractionByMass is used (also in make_consistent) to adjust the Hydrogen densities.

I think that this is probably a mistake. Instead, I think we should just make sure that in each cell, the total nuclei density remains before grackle solves chemistry is the same as the total nuclei density afterwards (for context, Issue #234 suggests we add a similar feature for Hydrogen)