Parity and Time-Reversal Breaking Effects on Resonant X-Ray Scattering at the Fe Pre-K-Edge in Magnetite

TL;DR

This paper investigates how noncentrosymmetric and time-reversal symmetry-breaking effects influence resonant x-ray scattering in magnetite, revealing the role of hybridization and predicting magnetic field-dependent intensity variations.

Contribution

It introduces a microscopic model explaining non-vanishing RXS intensities at forbidden spots and predicts magnetic field dependence of pre-edge intensities in magnetite.

Findings

Noncentrosymmetry causes hybridization affecting RXS spectra.

Sign change in hybridization explains forbidden spot intensities.

Magnetic field direction influences pre-edge RXS intensity.

Abstract

We study noncentrosymmetric effects on resonant x-ray scattering (RXS) in magnetite. The noncentrosymmetry at A sites in spinel structure makes the 4p states strongly hybridize with the 3d states through neighboring oxygen 2p states, giving rise to the non-vanishing contribution of the dipole-quadrupole (E1-E2) process in the RXS spectra. We substantiate this observation by introducing a microscopic model of a FeO_4 cluster with multiplets and the 4p band. We show that the hybridization changes its sign between two kinds of A sites and accordingly the local amplitude from the E1-E2 process changes its sign, resulting in non-vanishing RXS intensities at forbidden spots (002) and (006) in the pre-edge region in agreement with the experiment. A large dependence of the pre-edge intensity on the direction of the applied magnetic field is predicted as a consequence of breaking both centrosymmetry and time-reversal symmetry. Furthermore we analyze the intensity difference between two opposite directions of the applied magnetic field at the allowed spot (222) in connection with the experiment. We obtain the intensity difference of a "dispersion" form, which resembles the observed spectra at the Mn pre-K-edge in MnCr_2O_4 but is quite different from the observed one in magnetite. Although the observed spectra are claimed to arise from "magnetoelectric" amplitude, we argue that this claim has no ground.