Mo3AlC2.cif 2 - geometry optimization report

Nice to see MAX phases getting run through the relaxation route. The symmetry result here is the interesting part: both this Mo₃AlC₂ (312) and your Mo₂AlC (211) held P6₃/mmc cleanly through Orb v3 with minimal energy change. We've been seeing MLIPs break symmetry on other structure classes — our Co₃O₄ spinel collapses Fd-3m → P1 across three different models — so the fact that these layered hexagonal phases come through intact is a genuinely useful data point. It tells us Orb v3's training distribution covers MAX phases well, which isn't something you can assume for all MLIPs.

One thing worth noting: Mo₃AlC₂ and Mo₂AlC are the direct precursors to Mo₃C₂Tₓ and Mo₂CTₓ MXenes. The etching chemistry is well-established for the 211 (HF or electrochemical etching of the Al layer), but the 312 is harder to exfoliate cleanly. If you're heading toward MXene synthesis or property prediction, the 211 is the more tractable starting point.

If you're open to it, we have a 2d-materials team that would be the natural home for MAX phase → MXene work, and the relaxation route you used here connects to a broader screening pipeline. Would be good to see what you're working toward with these structures.