Orbital Engineering in Symmetry-Breaking Polar Heterostructures

We experimentally demonstrate a novel approach to substantially modify orbital occupations and symmetries in electronically correlated oxides. In contrast to methods using strain or confinement, this orbital tuning is achieved by exploiting charge transfer and inversion symmetry breaking using atomically layered heterostructures. We illustrate the technique in the ${\mathrm{LaTiO}}_3\text{−}{\mathrm{LaNiO}}_3\text{−}{\mathrm{LaAlO}}_3$ system; a combination of x-ray absorption spectroscopy and ab initio theory reveals electron transfer and concomitant polar fields, resulting in a $\sim 50\%$ change in the occupation of Ni $d$ orbitals. This change is sufficiently large to remove the orbital degeneracy of bulk ${\mathrm{LaNiO}}_3$ and creates an electronic configuration approaching a single-band Fermi surface. Furthermore, we theoretically show that such three-component heterostructuring is robust and tunable by choice of insulator in the heterostructure, providing a general method for engineering orbital configurations and designing novel electronic systems.

Figure 1

Calculated $1\times 1$ band structures for (a) bulk ${\mathrm{LaNiO}}_3$, (b) two-component, and (c) three-component superlattices near the Fermi level with projections onto Wannier states represented by bold color shading. Horizontal lines indicate centers of $3z^2-r^2$ (red dashed) and $x^2-y^2$ (blue solid) bands.

Figure 2

(a) DFT-calculated electron transfer from Ti $t_{2g}$ to Ni $e_g$ orbitals showing donor states (left) and acceptor states (right). Backbonding leads to some electron transfer from the Ni to lower-lying Ti levels. (b) XAS spectra for an MBE-grown LTNAO superlattice, bulk ${\mathrm{SrTiO}}_3$, and a thick (10 nm) ${\mathrm{LaTiO}}_3$ film grown on ${\mathrm{LaAlO}}_3$. Dashed lines show the approximate peak positions of the LTNAO spectrum, in agreement with a predominantly ${\mathrm{Ti}}^{4+}$ oxidation state. (c) Calculated atomic structure (left) and COBRA-derived electron density map (right) for a single-repeat LTNAO. Both calculated and measured structures show the atomic displacements due to the polar electric field ($\overrightarrow{P}$) that is driven by the electron transfer.

Figure 3

Measured x-ray absorption (circles) for in-plane (blue) and out-of-plane (pink) polarizations of (a) two-component and (b) three-component nickelate superlattices. The solid colored lines are double Gaussian fits used in evaluating Eq. (1).