Abstract
The magnetocrystalline anisotropy energy of Fe16N2 is calculated by the band model. The approximate energy bands of this material are formulated by Deegan's prescription and the formulas of Slater and Koster inclusive of the spin-orbit interaction and the exchange splitting, which is consistent with the experimental and calculated results of the magnetization. The anisotropy constants, K1 and K2, are evaluated through a comparison of the electronic energies with the magnetization parallel to the [001], [100], and [101] directions. The calculated anisotropy constants obtained with the exchange splitting 0.109 Ry are much the same as the experimental ones. However, the calculated results obtained by using an exchange splitting greater than 0.117 Ry do not agree with the experimental ones in the approximate number of electrons for Fe16N2. The calculated constant Ku for Fe16C2 inclusive of the exchange splitting 0.109 Ry agrees well with the experimental one.
