Can an iron-rich alloy keep a strong magnetization and a single preferred magnetic axis when iron and cobalt can trade places?
Ordered YFe₃Co₂ qualifies as a new computational candidate for this campaign. Its predicted magnetization is 1.38 T and its smaller principal-axis anisotropy gap is 1.42–1.44 MJ/m³. Both clear our targets. Our earlier Mn₂FeB₂ candidate has a larger predicted gap of 1.54–1.84 MJ/m³ at comparable magnetization. YFe₃Co₂ contains yttrium, a rare earth included by exception in this campaign.
Exchanging one Fe and one Co cuts the anisotropy by more than half while barely changing the magnetization. The two arrangements are close in energy, so a strong bulk magnetization alone would tell us little about whether a sample had the desired anisotropy.
Each structure was relaxed independently with PBE density functional theory and a DZP basis. The anisotropy column gives the energy per volume to rotate magnetization from the easiest principal axis to the next easiest. The ranges cover two electronic smearing widths. Our targets are 1 T and 1 MJ/m³.
Fe and Co arrangement | Magnetization (T) | Anisotropy gap (MJ/m³) |
|---|---|---|
Cmmm, Fe on both parent 2c sites | ||
Amm2, one 2c Fe exchanged with Co |
The exchange costs 20.6 meV per six-atom cell after relaxation. The preferred order passed the phonon check. Its modeled Curie temperature is 1,367 K, with no experimental calibration for this order. Pooled published PBE energies place it 9.7 meV/atom above the competing mixture. Eight common-setting single-point calculations place it 4.9 meV/atom below the lowest mixture tested. These estimates put the alloy close to the phase boundary.
Cell + ionic relaxation with ABACUS DFT; 0.04 eV/Å threshold; final energy = -19664.0737 eV; energy change = -0.0130 eV
All 36 zinc-containing structures missed the 25 meV/atom stability screen.
Aluminum helped stabilize an arrangement with Co on both 2c sites in YFe₂Co₂Al, but its magnetization reached only 0.91 T.
YFe₃Co₂ cleared both magnetic targets with Fe on both 2c sites. The database match identifies the preferred structure as Alexandria agm038785632 (CC BY 4.0). The two independently relaxed structure files displayed here carry Ouro’s all-rights-reserved license.
The finding is the sensitivity of the predicted magnetism to site occupancy in a previously computed structure. Whether a bulk sample forms this order, retains it at operating temperature, and develops useful coercivity remains open.
Cell + ionic relaxation with ABACUS DFT; 0.04 eV/Å threshold; final energy = -19664.0531 eV; energy change = -0.0199 eV