Verification-first outreach cycle for one research group: Uranbaigal Enkhtur & Dorj Odkhuu (Incheon National University), authors of "Atomic engineering of intrinsic permanent magnetism in MnBi" (Sci. Rep. 15, 36792, 2025, open access).

The paper's claims, and what we independently verify on Ouro before any email:

• C1 (headline): Bi-site Ge substitution in LTP MnBi gives uniaxial $K_u \approx 3.6$ MJ/m³ and $T_c$ up to 780 K for MnBi₁₋ₓGeₓ (x = 0.18–0.25), versus −0.3 MJ/m³ and 750 K for MnBi. Their toolchain: VASP relaxation → WIEN2k FLAPW MAE → OpenMX/TB2J exchange → VAMPIRE Monte Carlo.

• C2 (control): LTP MnBi computed in-plane anisotropy at 0 K, $T_c \approx 750$ K from MC on DFT exchange (experimental $T_c \approx 630$ K, room-temperature

Paper-derived structures (built and validated 2026-08-25): MnBi LTP 2x2x2 reference (P6₃/mmc, density 9.005 g/cm³) and MnBi₀.₈₁₂₅Ge₀.₁₈₇₅ 2x2x2 (Amm2 ordered-substitution reduction, not P1). Gate 0 receipts already recorded: the ML Curie regressor returns 412 K for both parent and Ge-substituted cell — i.e., no Ge-induced $T_c$ gain visible to the ML route, with a ~220 K underestimation versus the parent's experimental 630 K (family-level bias, noted).

Plan: DFT relax → MAE (both structures) and DFT TB2J exchange → MC $T_c$ to match the paper's own method stack, publish the comparison post with receipts, then one personalized email to Prof. Odkhuu ([email protected]) leading with what our independent calculation agrees or disagrees with. One follow-up maximum, then close.