Epitaxial phases ofBiMnO3from first principles

Abstract

${\mathrm{BiMnO}}_{3}$ is the only transition-metal perovskite oxide that is insulating and shows strong ferromagnetism in bulk. This distinctive behavior would make it a promising candidate as a magnetoelectric multiferroic if it was also a polar material, but experiments have shown that bulk ${\mathrm{BiMnO}}_{3}$ has either a very small polarization (below $0.1\phantom{\rule{0.28em}{0ex}}\ensuremath{\mu}\mathrm{C}/{\mathrm{cm}}^{2}$) or, most likely, that it is a paraelectric. There is also experimental evidence that the polarization in ${\mathrm{BiMnO}}_{3}$ films grown on ${\mathrm{SrTiO}}_{3}$ can be as high as $20\phantom{\rule{0.28em}{0ex}}\ensuremath{\mu}\mathrm{C}/{\mathrm{cm}}^{2}$. Despite the interest in these behaviors, the diagram of ${\mathrm{BiMnO}}_{3}$ as a function of epitaxial strain has remained largely unexplored. Here, we use first-principles to predict that, both under enough compressive and tensile epitaxial strain, ${\mathrm{BiMnO}}_{3}$ films are ferroelectric with a giant polarization around $100\phantom{\rule{0.28em}{0ex}}\ensuremath{\mu}\mathrm{C}/{\mathrm{cm}}^{2}$. The phases displayed by the films are similar to those experimentally found for ${\mathrm{BiFeO}}_{3}$ in similar conditions---at compressive strains, the film is supertetragonal with a large component of the polarization pointing out of plane, while at tensile strains the polarization points mostly in plane. As in ${\mathrm{BiFeO}}_{3}$ films, these phases are antiferromagnetic---the orbital ordering responsible for ferromagnetism in ${\mathrm{BiMnO}}_{3}$ is absent in the polar phases. Our calculations also show that the band gap of some of these ${\mathrm{BiMnO}}_{3}$ films is substantially smaller than gaps typically found in ferroelectric oxides, suggesting it may be a suitable material for photovoltaic applications.