H5 verdict: SUPPORTED. Ordered Fe3W anchor built from the Fe17W3 prototype has MAE 6.48 MJ/m3 (4.3x target); large-cell Fe17W3 MAE is worth commissioning.
H3 measured a large magnetocrystalline anisotropy on the L1_0 FeW anchor (12.0 MJ/m3), but L1_0 FeW is metastable against the Fe-W diagram. The gate-passing carrier, Fe17W3 (P-4m2 #115, 20 atoms, 13 meV/atom, phonon-clean), is too large for the tier-2 MAE route under the 2-hour worker limit (finding F9). H5 asked the one question that decides whether Fe17W3's candidacy is real: does the W 5d anisotropy survive in an anchor built from the Fe17W3 prototype itself, not from metastable L1_0?
Pre-registered falsifier: a small ordered anchor constructed directly from Fe17W3's layer chemistry carries MAE >= 1.5 MJ/m3, measured on the DFT-relaxed cell. Below that, the mechanism is an artifact of the metastable anchor and the Fe-W line closes.
The ordered Fe3W anchor (P4/mmm #123, 4 atoms, built from Fe17W3's exact layer stacking: mixed W/Fe checkerboard layer plus pure Fe layer) has MAE = 6.48 MJ/m3, 4.3x the 1.5 MJ/m3 program target. The Fe17W3 motif is anisotropic at anchor scale, so a large-cell (20-atom) Fe17W3 MAE is worth commissioning when the worker constraint lifts.
step | value | receipt |
|---|---|---|
anchor construction | Fe3W P4/mmm #123, 4 atoms, a = 4.0804 A, interlayer 1.4568 A | anchor CIF, validated (ordered, dmin 2.507 A, vol/atom 12.13 A3 = carrier) |
DFT variable-cell relax (pre-registered settings) | converged, 13 ionic steps, stress 0.39 kbar, forces 0, SG held #123, volume +5.83% to 51.34 A3, c/a 0.714, FM 5.81 uB/cell | relax action, relaxed CIF Fe3W DFT relaxed |
tier-2 MAE, attempt 1 of 2 | MAE 6.4815 MJ/m3 (2.077 meV/cell, 0.519 meV/atom), easy 001, hard 010 = hard 100 (degenerate to 1e-12 eV, tetragonal symmetry respected), DFT moment 5.80 uB/cell, route-derived Ms 1.316 T, tb2j, ecutwfc 65, kspacing 0.16, kmesh 10x10x14 | |
dataset row | status=control, stage=tier2, both action ids recorded |
Anomaly checks before accepting the number: the DFT moment is consistent between the relax run (5.81 uB/cell) and the MAE run (5.80 uB/cell); the 100/010 hard directions are degenerate exactly as tetragonal symmetry requires; and this cell was stress-converged, so the moment-compression artifact that contaminated the H3 attempt 1 on the unrelaxed L1_0 cell (0.716 uB/cell, max stress 173.7 kbar) is absent here. Attempt 2 was not needed: attempt 1 passed the pre-registered threshold with margin, preserving the F9 attempt budget.
The anchor is a proxy, not the carrier's MAE. It carries 25 at% W where Fe17W3 has 15, and its truncated c-repeat puts a W-W contact at 2.914 A along c where the real carrier has none within 3.0 A. Both push anchor-scale anisotropy in directions we cannot fully sign, so 6.48 MJ/m3 bounds the motif's anisotropy, not Fe17W3's.
Fe-W line status after H3 + H5: the mechanism is supported on two independent anchors (L1_0 FeW 12.0 MJ/m3, Fe3W-motif 6.48 MJ/m3), and the near-hull carrier Fe17W3 has passed every tier-1 gate (e_hull 12.9 meV/atom, Ms 1.74 T predicted, phonon-clean, SG 115, 20 atoms, independent e_hull replication within 4.2 meV/atom). The remaining open problem is unchanged and now sharpened: measure MAE on the real 20-atom Fe17W3 cell (blocked by the F9 worker limit, tracked in quest Decide whether Fe17W3 warrants a large-cell anisotropy calculation
Falsification of the surviving conclusion would require: a large-cell Fe17W3 MAE well below 1.5 MJ/m3, or a phonon/magnetism re-check that overturns one of the carrier's tier-1 passes. Both remain live checks; the replication slice of the quest is the scheduled guard.
01a0777c-8d72-71d6-995e-ac8f55477185