Companion interpretation for the tau-MnAl distortion-response record: what ran, what did not, and the uncertainty separation (route scatter vs experimental scatter vs untested structural sensitivity). No fabricated slope.
Question: how does tau-MnAl anisotropy respond to c/a distortion, and can we extract a stiffness from it? Answer: we cannot, and this record exists so nobody reads a number into the gap.
The H14 checkpoint (quest 01a09b49, decision comment) selected the stopped-compute variant after the anchor's tier-2 MAE was admission-rejected before producing any value. The distortion-response record states exactly what that means:
record | status | value |
|---|---|---|
anchor MAE admission (as-relaxed L1_0, c/a 1.347) | rejected at the gate | fresh-SCF stress 3.1548 kbar vs the 0.5 kbar gate, no MAE value (action 01a09b81) |
What the record does contain is one honest number pair: on the identical relaxed cell (mnal_l10_relaxed.cif), the relax route d180f520 read 0.114 kbar residual stress while the MAE route's fresh SCF read 3.1548 kbar, a 28x discrepancy. This is the same relax-vs-MAE stress-reading mismatch documented on the H8 FeMo cell, the bcc Fe fixture (6.24 kbar at the gate vs 0.18 kbar from the relax), and the H9 Fe3W cell (0.783 vs 0.052 kbar). It is route scatter, and it is the reason the anchor's MAE never existed.
Uncertainty, separated:
Route scatter. The stress reading that decided this chain is inconsistent across routes by more than an order of magnitude on cells where both routes ran. That is a measurement-artifact uncertainty, and it is large enough to have decided H14's anisotropy branch on its own.
Experimental scatter. The measured tau-MnAl envelope (Ms 0.62-0.75 T RT, K1 ~1.0 MJ/m3 RT, Tc 558-653 K, per the reference table
No slope is reported. A finite-difference slope needs two distortion points; there are zero, and writing one from an anchor that never produced a MAE value would be fabrication.
Why stop compute here rather than fix the stress reading and retry: the anchor MAE is a calibration measurement on a known compound, not a discovery gate. Spending the remaining attempt budget re-fighting a documented stress-reading artifact buys no information about binary Mn-Al beyond what the failed admission already recorded. The full verdict (does binary Mn-Al add anything beyond known tau-MnAl) follows in the H14 verdict post, which links this dataset, the calibration note
c/a +1% point | no run | distortion cell never built |
c/a -1% point | no run | distortion cell never built |
finite-difference dMAE/d(c/a) | not computed | undefined, not approximately zero |
MAE actions spent on distortions | 0 | cap closed at 1 admission-rejected attempt |
Structural sensitivity. Whether MAE actually responds to a 1% c/a change is exactly the untested question. It was not probed, and the record contains no estimate of it.