Hybridization-Switching Induced Mott Transition in $AB{\mathrm{O}}_3$ Perovskites

We propose the concept of a “hybridization-switching induced Mott transition” which is relevant to a broad class of $AB{\mathrm{O}}_3$ perovskite materials including ${\mathrm{BiNiO}}_3$ and ${\mathrm{PbCrO}}_3$ that feature extended $6s$ orbitals on the $A$-site cation (Bi or Pb), and a strong $A\text{−}\mathrm{O}$ covalency induced ligand hole. Using ab initio electronic structure and slave rotor theory calculations, we show that such systems exhibit a breathing phonon driven $A$-site to oxygen hybridization-wave instability which conspires with strong correlations on the $B$-site transition metal ion (Ni or Cr) to trigger a Mott insulating state. This class of systems is shown to undergo a pressure induced insulator to metal transition accompanied by a colossal volume collapse due to ligand hybridization switching.

Figure 1

(a) DFT cohesive energy versus volume for AP (triclinic) and HP (orthorhombic) ${\mathrm{BiNiO}}_3$; the intersection points of the plotted common tangent with the two curves yields the volume change $\mathrm{\Delta }V/V\approx 3\%$ at the transition. (b) Elastic energy of the $\mathrm{Ni}\text{−}\mathrm{O}$ sublattice as a function of O-displacement for the HP (dashed) and AP (solid) volumes. Inset shows similar plot for the $\mathrm{Bi}\text{−}\mathrm{O}$ sublattice.

Figure 2

(a) $\mathrm{GGA}+U$ projected DOS of ${\mathrm{BiNiO}}_3$ in AP (top panels) and HP (bottom panels) phase. Left, middle, and right panels show projections to Bi-$s$ (displaying relevant energy ranges), Ni-$d$, and O-$p$. Zero of energy is set at $\mathrm{GGA}+U$ Fermi energies. For AP, projections to Bi1 (solid, black) and Bi2 (shaded) are shown separately. (b) Calculated energy levels of Bi-$s$, O-$p$, and Ni-$e_g$ in the AP and HP phase. Zero of energy is set at Ni-$e_g$.

Figure 3

Phase diagram varying $\kappa$ and ${\epsilon }_B-{\epsilon }_{\mathrm{O}x}$ for (a) $U/t_B=6.5$ and (b) $U/t_B=2.5$. (c) DOS $N_{\mathrm{Tot}}(\omega )$ for typical points in AP and HP phases, marked by stars in (a), as well as DOS for the metastable AP phase with imposed $\phi =0$.

Figure 4

Bond-dependent kinetic energy from the model (color bar shows magnitude) for $B\text{−}\mathrm{O}$ and $A\text{−}\mathrm{O}$ sublattices in the AP [(a) and (b)] and HP [(d) and (c)] phase, projected to the $xy$ plane. Constant amplitude surfaces of DFT O-$p$ Wannier functions for ${\mathrm{BiNiO}}_3$ at AP (e) and HP (f), superposed on ${\mathrm{NiO}}_6$ octahedra with adjacent two Bi ions. Cyan (light) and magenta (dark) colors indicate opposite signs.