Electronic and Magnetic Reconstructions in ${\mathrm{La}}_{0.7}{\mathrm{Sr}}_{0.3}{\mathrm{MnO}}_3/{\mathrm{SrTiO}}_3$ Heterostructures: A Case of Enhanced Interlayer Coupling Controlled by the Interface

We report on the magnetic coupling of ${\mathrm{La}}_{0.7}{\mathrm{Sr}}_{0.3}{\mathrm{MnO}}_3$ layers through ${\mathrm{SrTiO}}_3$ spacers in ${\mathrm{La}}_{0.7}{\mathrm{Sr}}_{0.3}{\mathrm{MnO}}_3/{\mathrm{SrTiO}}_3$ epitaxial heterostructures. Combined aberration-corrected microscopy and electron-energy-loss spectroscopy evidence charge transfer to the empty conduction band of the titanate. Ti $d$ electrons interact via superexchange with Mn, giving rise to a Ti magnetic moment as demonstrated by x-ray magnetic circular dichroism. This induced magnetic moment in the ${\mathrm{SrTiO}}_3$ controls the bulk magnetic and transport properties of the superlattices when the titanate layer thickness is below 1 nm.

Figure 1

(a) Temperature-dependent resistivity of $[{\mathrm{LSMO}}_6/{\mathrm{STO}}_{\mathrm{N}}{]}_8$ superlattices in 14 T (open symbols) and zero (solid symbols) field. (b) Temperature-dependent magnetization upon warming in 100 Oe. $N$ is indicated in the figures. (c) Sketch illustrating magnetic coupling through interfacial Ti in the case of an ultrathin STO spacer. (d) M(H) loops measured in a vibrating sample magnetometer (VSM).

Figure 2

(a) High magnification $Z$-contrast image of a $[{\mathrm{LMSO}}_6/{\mathrm{STO}}_{\mathrm{N}}]$ superlattice where the STO layer thickness is varied between 2 and 8 u.c.; thickness is indicated in the figure. (b) Top: Normalized integrated intensity of $\mathrm{Ti}-L_{2,3}$ (blue), $\mathrm{Mn}-L_{2,3}$ (red), and La $M_{4,5}$ (green) along the direction marked with a green arrow in (a). Bottom: Mn (red) and Ti (blue) oxidation states. The blue lines are guides to the eye. (c)–(f) Atomic elemental maps corresponding to the (c) Ti $L_{2,3}$, (d) Mn $L_{2,3}$, (e) La $M_{4,5}$, and (f) Sr $L_{2,3}$ signals, after fitting the background to a power law and integrating 30 eV wide windows under the remaining edges, except for (f), where a 300 eV wide window was used due to the poor signal. Data sets in (a) and (b) were obtained in a VG Microscopes HB501UX, while data sets in (c)–(f) were obtained in the Nion UltraSTEM, both operated at 100 kV and equipped with aberration correctors and Gatan Enfina spectrometers.

Figure 3

XMCD measurements in an $[{\mathrm{LMSO}}_6/{\mathrm{STO}}_2{]}_8$ superlattice. (a) Ti $L_{2,3}$ edge (red) and Mn $L_{2,3}$ edge (green). For comparison, the Ti signal is multiplied by 10. (b) Temperature-dependent Ti (green open circles) and Mn (red full circles) intensities and magnetization (black line), all normalized at 4 K. (Inset) Maximum intensity of Ti (at 466 eV) vs Mn (at 642 eV) XMCD signal for $[{\mathrm{LMSO}}_N/{\mathrm{STO}}_2{]}_8$, $N=2$, 4, and 14. The red line is a linear fit.