Hole-induced insulator-to-metal transition in $\mathrm{L}{\mathrm{a}}_{1-x}\mathrm{S}{\mathrm{r}}_x\mathrm{Cr}{\mathrm{O}}_3$ epitaxial films

We have investigated the evolution of the electronic properties of $\mathrm{L}{\mathrm{a}}_{1-x}\mathrm{S}{\mathrm{r}}_x\mathrm{Cr}{\mathrm{O}}_3(0\le x\le 1)$ epitaxial films deposited by molecular beam epitaxy (MBE) using x-ray diffraction, x-ray photoemission spectroscopy, Rutherford backscattering spectrometry, x-ray absorption spectroscopy, electrical transport, and ab initio modeling. $\mathrm{LaCr}{\mathrm{O}}_3$ is an antiferromagnetic insulator, whereas $\mathrm{SrCr}{\mathrm{O}}_3$ is a metal. Substituting $\mathrm{S}{\mathrm{r}}^{2+}$ for $\mathrm{L}{\mathrm{a}}^{3+}$ in $\mathrm{LaCr}{\mathrm{O}}_3$ effectively dopes holes into the top of valence band, leading to $\mathrm{C}{\mathrm{r}}^{4+}$ (${3d}^2$) local electron configurations. Core-level and valence-band features monotonically shift to lower binding energy with increasing $x$, indicating downward movement of the Fermi level toward the valence band maximum. The material becomes a $p$-type semiconductor at lower doping levels and an insulator-to-metal transition is observed at $x\ge 0.65$, but only when the films are deposited with in-plane compression via lattice-mismatched heteroepitaxy. Valence-band x-ray photoemission spectroscopy reveals diminution of electronic state density at the $\mathrm{Cr}d{t}_{2g}$-derived top of the valence band, while O K-edge x-ray absorption spectroscopy shows the development of a new unoccupied state above the Fermi level as holes are doped into $\mathrm{LaCr}{\mathrm{O}}_3$. The evolution of these bands with Sr concentration is accurately captured using density functional theory (DFT) with a Hubbard U correction of 3.0 eV $(\mathrm{DFT}+U)$. Resistivity data in the semiconducting regime $(x\le 0.50)$ do not fit perfectly well to either a polaron hopping or band conduction model but are best interpreted in terms of a hybrid model. The activation energies extracted from these fits are well reproduced by $\mathrm{DFT}+U$.

Figure 1

RHEED patterns for $\mathrm{L}{\mathrm{a}}_{1-x}\mathrm{S}{\mathrm{r}}_x\mathrm{Cr}{\mathrm{O}}_3/\mathrm{LaAl}{\mathrm{O}}_3\left(001\right)$ film set in the [100] zone axis: (a) $x=0,20\mathrm{nm}$; (b) $x=0.10,80\mathrm{nm}$; (c) $x=0.25,51\mathrm{nm}$; (d) $x=0.50,62\mathrm{nm}$; (e) $x=0.66,28\mathrm{nm}$; (f) $x=0.75,54\mathrm{m}$; (g) $x=0.85,37\mathrm{nm}$; (h) $x=1.0,42\mathrm{nm}$.

Figure 2

Unit-cell volumes for the $\mathrm{L}{\mathrm{a}}_{1-x}\mathrm{S}{\mathrm{r}}_x\mathrm{Cr}{\mathrm{O}}_3/\mathrm{LaAl}{\mathrm{O}}_3\left(001\right)$ film set taken from XRD direct-space maps (see Supplemental Material [23]) along with the expected behavior for two limiting cases: (i) films with bulk volumes assuming a linear relationship between the end members, and (ii) fully coherently strained films using Poisson ratios for SCO and LCO and assuming a linear relationship. Also shown is a single data point measured for $\mathrm{SrC}{\mathrm{O}}_{2.8}$ deposited on $\mathrm{LaAl}{\mathrm{O}}_3\left(001\right)$.

Figure 3

$\mathrm{Cr}2p$, $\mathrm{O}1s$, $\mathrm{La}4d$, and $\mathrm{Sr}3d$ core-level spectra for $\mathrm{L}{\mathrm{a}}_{1-x}\mathrm{S}{\mathrm{r}}_x\mathrm{Cr}{\mathrm{O}}_3$ as a function of $x$.

Figure 4

(a) Average chemical potential shift (Δμ) deduced from $\mathrm{O}1s$, $\mathrm{La}4d$, and $\mathrm{Sr}3d$ binding energy shifts and (b) room-temperature resistivity vs $x$.

Figure 5

Cr L-edge and O K-edge XAS for the $\mathrm{L}{\mathrm{a}}_{1-x}\mathrm{S}{\mathrm{r}}_x\mathrm{Cr}{\mathrm{O}}_3$ film series.

Figure 6

(a) Valence-band XPS and O K-edge XAS spectra for the $\mathrm{L}{\mathrm{a}}_{1-x}\mathrm{S}{\mathrm{r}}_x\mathrm{Cr}{\mathrm{O}}_3$ film series; (b) analogous theoretical densities of states based on $\mathrm{PBEsol}+U$ ($U=3\mathrm{eV}$).

Figure 7

(a) $\rho \left(T\right)$ for the $\mathrm{L}{\mathrm{a}}_{1-x}\mathrm{S}{\mathrm{r}}_x\mathrm{Cr}{\mathrm{O}}_3$ film series with the exception of pure $\mathrm{LaCr}{\mathrm{O}}_3$, which was too resistive to measure (open squares) along with fits of the data to Eq. (1) for $x=0.12,0.25$, and $0.50$ (solid curves). Inset–Calculated activation energies from a linear interpolation method based on $\mathrm{PBEsol}+U$ (see text for details).