From the NovoMag database.
Testing because of high MAE, has experimental data.
Experimental MAE: 4.9 MJ/m3
NovoMag theoretical: 4.34 MJ/m3
Like we've been doing with MnAl, Co, rerunning with decreased kspacing (0.12) to compare against NovoMag and known experimental data.
Retesting with smaller kspacing did not help this one. Actually increased further away from experimental from ~8 to ~9 MJ/m3.
Likely a single-zeta orbital issue, not a convergence issue.
Experimental MAE: 4.9 MJ/m3
Our calculation: 8.2 MJ/m3
Major overestimation! But directionally correct.
{ "kmesh": [ 15, 15, 11 ], "mae_ev": 0.0013898594977302114, "method": "tb2j", "status": "success", "n_atoms": 2, "easy_axis": "001", "hard_axis": "010", "tb2j_scf_dir": "/data/calculations/20e62c98-bbe6-48ff-9fcc-71244bf39840/tb2j_scf_soc0", "mae_mj_per_m3": 8.206431993217903, "tb2j_energies": { "100": -0.14543005058390393, "001": -0.14681991008163392, "010": -0.1454300505839037 }, "tb2j_nscf_dir": "/data/calculations/20e62c98-bbe6-48ff-9fcc-71244bf39840/tb2j_nscf_soc1", "calculation_id": "20e62c98-bbe6-48ff-9fcc-71244bf39840", "input_parameters": { "method": "tb2j", "pp_dir": "/data/data/orbitals/Dojo-NC-FR/Pseudopotential", "ecutwfc": 65, "scf_thr": 0.000001, "kspacing": 0.16, "scf_nmax": 100, "basis_dir": "/data/data/orbitals/Dojo-NC-FR", "calculation_id": "20e62c98-bbe6-48ff-9fcc-71244bf39840", "smearing_sigma": 0.05, "smearing_method": "mp" }, "mae_mev_per_atom": 0.6949297488651057, "volume_angstrom3": 27.13481831868747, "magnetic_hardness_kappa": 3.4243207676747045, "calculation_time_seconds": 467.36247277259827, "total_magnetic_moment_uB": 2.18352, "saturation_magnetization_A_per_m": 746273.1485529396 }