Engineering a Spin-Orbital Magnetic Insulator by Tailoring Superlattices

In $5d$ Ir oxides with an interplay of spin-orbit coupling and electron correlations, we have tailored a spin-orbital magnetic insulator out of a semimetal ${\mathrm{SrIrO}}_3$ by tuning the structure through superlattices [$({\mathrm{SrIrO}}_3{)}_m$, ${\mathrm{SrTiO}}_3$] ($m=1,2,3,4$, and $\infty$). We observed the systematic decrease of the magnetic ordering temperature and the resistivity as a function of $m$. The transition from the semimetal to the insulator is found to be closely linked to the appearance of magnetism at $m\simeq 3$. Long range magnetic ordering was realized even in the $m=1$ single layer superlattice, implying that the design and realization of novel electronic phases is feasible at the level of a single atomic layer in complex Ir oxides.

Figure 1

(a) Schematics of the superlattices $[({\mathrm{SrIrO}}_3{)}_m,{\mathrm{SrTiO}}_3{]}_k/{\mathrm{SrTiO}}_3(001)$ ($m=1,2$, and $\infty$). (b) X-ray diffraction scans ($\theta –2\theta$) of the superlattices ($m=1,2,3,4$, and $\infty$). Down arrows indicate the superlattice peaks, while open circles represent the fundamental peaks. Open triangles depict peaks from the substrate. (c) Reciprocal-space mapping of the (114) x-ray diffraction peaks for the superlattice with $m=1$. (d) Atomically resolved HAADF-STEM image of the superlattice with $m=1$ along the ${\mathrm{SrTiO}}_3[100]$ direction. The schematic of the crystal structure is superposed on the image to show that the brightest spot (Ir) is atomically aligned in the (001) plane.

Figure 2

Temperature ($T$) dependence of (a) resistivity ($\rho$), (b) $-d(\ln \rho )/dT$, (c) Hall coefficient ($R_H$), and (d) in-plane magnetization ($M$) for the superlattices ($m=1,2,3,4$, and $\infty$). Arrows in each panel indicate the temperatures of the anomaly ($T_c$) in the resistivity for $m=2$ and $m=1$, defined as the peak in $-d(\ln \rho )/dT$.

Figure 3

(a) Temperature ($T$) dependence of the magnetization ($M$) measured at 0.1 T for the superlattice with $m=1$. The upturn below 50 K, observed for both magnetic field directions, is derived from magnetic impurities in the substrates. (b) Temperature dependence of the magnetic x-ray diffraction intensity for the (0.5, 0.5, 5) peak. The unit cell dimensions are $a\times a\times 2a$ ($a\sim 0.39\mathrm{nm}$). (c) Field dependence of the in-plane magnetization measured at 100 K.

Figure 4

Magnetic structures of the superlattices with (a) $m=1$, (b) $m=2$, and (c) ${\mathrm{Sr}}_2{\mathrm{IrO}}_4$ [15]. The phase relation of the ${\mathrm{IrO}}_6$ rotation between adjacent ${\mathrm{IrO}}_2$ layers is determined so as to account for the observed magnetic moments [29].

Figure 5

Calculated band structures for (a) $m=1$ and (b) $m=2$ with $U=1\mathrm{eV}$. The red and blue colors represent Ir and Ti $d$-orbitals, respectively. The grey lines dominant below $-1.5\mathrm{eV}$ indicate oxygen $p$ orbitals. The dashed line denotes the Fermi level, and the bands near the Fermi level are mostly of $J_{\text{eff}}=1/2$ character except near the $\mathrm{\Gamma }$ and $Z$ points. Band structures in the presence of magnetic order for (c) $m=1$ and (d) $m=2$ with $U=3\mathrm{eV}$. Both manifest a charge gap shaded in orange.