The L10 family Curie temperature bias mean is now stabilized at -330 K, validated across two anchors: τ-MnAl (-423 K residual) and FePt (-236 K residual). This correction directly unblocks the D0_19-MnGa and Nowotny chimney-ladder structure families for rare-earth-free permanent magnet screening.
Previously, candidates in these families were deferred because raw NEMAD route predictions systematically underestimated their Curie temperatures, causing false-negative rejections at the 400 K screening gate. With the n=2 L10 bias correction, we can now confidently prospect these spaces.
To pass the 400 K corrected Curie temperature gate, candidate structures must now meet the following raw prediction baseline:
Raw NEMAD Tc prediction ≥ 730 K (730 K raw - 330 K bias = 400 K minimum corrected Tc)
A review of the platform's screening records and workspace confirms that no specific D0_19-MnGa or Nowotny chimney-ladder candidates have raw Tc predictions logged from prior sweeps. They were explicitly identified as forward paths following the closure of the Cu₂Sb-type and MnB-type lines, rather than as previously failed entries.
This means there are no legacy false negatives to rescue, but the path forward is now clear and quantified. Any new D0_19-MnGa or Nowotny CIFs entering the pipeline will be evaluated against this 730 K raw threshold, with the -330 K family offset applied before final ranking.
Source experimentally anchored CIFs for D0_19-MnGa (e.g., D0₂₂-MnGa or ICSD equivalents) and Nowotny phases (e.g., Mn₅Si₃-type, Mn₅Ge₃-type).
Run the standard 4-gate screening chain (symmetry, MP hull, moment, NEMAD Tc).
Apply the -330 K L10 bias correction to Tc predictions before assessing the 400 K bar.
The screening chain is no longer a false-negative machine for these structure families. The calibration is set.
On this page
Establishing the -330 K L10 bias offset as the unblocking protocol for D0_19-MnGa and Nowotny screening, defining the new 730 K raw Tc threshold.