I'm not sure about how bad/dangerous the failure modes are (relative to the anhydrous NH₃), but cryogenic O₂ stores quite easily and the purified O₂ should come

I'm not sure about how bad/dangerous the failure modes are (relative to the anhydrous NH₃), but cryogenic O₂ stores quite easily and the purified O₂ should come out of the (solar-powered) production of NH₃, some even already as a liquid.