Things like mix_for_equivalence_ratio use Cantera and can only do hydrocarbon chemistry, even though we generalized our direct stoichiometry calculations to include Aluminum. We just need to not use Cantera everywhere for this.
Since we support non-hydrocarbon elsewhere I'm labeling this as a bug. The methods produce the wrong results (pure fuel) and don't indicate so.