Comprehensive Study of GGA+U Approach to Modeling Electronic Structure of α″–Fe₁₆N₂

Within the framework of density-functional theory (DFT), a common extension to the usual generalized-gradient approximation (GGA) exchange-correlation functional is the inclusion of Hubbard parameters U (,J) as GGA+U. In the study of the iron nitride phase α″–Fe₁₆N₂, first-principles calculations have yielded inconsistent results for its giant magnetic moment and reasonably high magnetocrystalline anisotropy (MCA). Several previous works have applied the GGA+U approach to α″–Fe₁₆N₂ to improve agreement with these experimental observations, each with unique methodology for selecting Hubbard parameters. While this approach has yielded increased magnetic moments and MCA over GGA, it typically falls short of experimental moments.

This work employs the plane-wave DFT code Quantum ESPRESSO to more comprehensively investigate the influence of Hubbard parameters U, J on the electronic structure and resultant magnetic properties of α″–Fe₁₆N₂. We survey a broad range of values for U and J, and compare to values from first principles, e.g. self-consistent linear-response calculations. MCA varies significantly with U, J including changes in sign, whereas magnetic moments remain inline with previous GGA+U calculations. These values are compared against experimental ranges to determine which U, J yield the best agreement. To elucidate these results, we consider band structures and density of states to provide a deeper understanding of the magnetic behavior in α″–Fe₁₆N2.