Hybrid functional study of proper and improper multiferroics

TL;DR

This study uses hybrid functional DFT calculations to accurately analyze the structural, electronic, magnetic, and ferroelectric properties of proper and improper multiferroics, showing HSE's effectiveness over standard DFT.

Contribution

It demonstrates that the HSE hybrid functional provides a more accurate first-principles description of multiferroic materials compared to local and semilocal DFT methods.

Findings

HSE accurately predicts properties of multiferroics.

Polarization decreases in HoMnO3 with HSE, increases in BiFeO3.

Effects beyond standard DFT are necessary for realistic modeling.

Abstract

We present a detailed study of the structural, electronic, magnetic and ferroelectric properties of prototypical \textit{proper} and \textit{improper} multiferroic (MF) systems such as BiFeO$_{3}$ and orthorhombic HoMnO$_{3}$, respectively, within density functional theory (DFT) and using the Heyd-Scuseria-Ernzerhof hybrid functional (HSE). By comparing our results with available experimental data as well as with state-of-the-art GW calculations, we show that the HSE formalism is able to account well for the relevant properties of these compounds and it emerges as an accurate tool for predictive first-principles investigations on multiferroic systems. We show that effects beyond local and semilocal DFT approaches (as provided by HSE) are necessary for a realistic description of MFs. For the electric polarization, a decrease is found for MFs with magnetically-induced ferroelectricity, such as HoMnO$_3$, where the calculated polarization changes from $\sim$ 6 $\mu C/cm^2$ using Perdew-Burke-Ernzerhof (PBE) to $\sim$ 2 $\mu C/cm^2$ using HSE. However, for \textit{proper} MFs, such as BiFeO$_{3}$, the polarization slightly increases upon introduction of exact exchange. Our findings therefore suggest that a general trend for the HSE correction to bare density functional cannot be extracted; rather, a specific investigation has to be carried out on each compound.