Anisotropy-driven interfacial magnetism in Ru-deficient SrRuO$_3$ thin films

Abstract

While stoichiometric SrRuO$_3$ (SRO) is a metallic itinerant ferromagnet with relatively homogeneous magnetization, Ru deficiency provides a powerful route to alter its electronic transport and depth-dependent magnetic properties. Ru-deficient SRO thin films grown by radio-frequency high oxygen pressure sputtering were investigated using a combination of X-ray reflectivity, polarized neutron reflectometry, off-specular neutron scattering, scanning transmission electron microscopy with energy-dispersive X-ray spectroscopy, electrical transport, and magnetometry. Structural and compositional analyses reveal that Ru deficiency is intrinsic to the films, with an enhanced deficiency at the interfaces. As a result, coherent electronic transport is suppressed and the saturation magnetization is reduced, while the Curie temperature remains largely unaffected, placing Ru-deficient SRO in a regime consistent with ferromagnetic insulator-like behavior. Depth- and lateral-resolved magnetic measurements further show that the interfacial regions remain ferromagnetic but exhibit enhanced perpendicular magnetic anisotropy, which constrains the local magnetization to remain predominantly out-of-plane and strongly reduces its in-plane projection. Our results establish Ru deficiency as a key control parameter governing transport, magnetization, and anisotropy in SRO thin films and highlight defect and interface engineering as powerful routes to tailor interfacial magnetism in correlated oxide heterostructures.

Figures (14)

• Figure 1: Atomic force microscopy micrographs of (a) the Nb:STO substrate and (b) the SRO thin film grown on Nb:STO. (c) Magnified view of the highlighted region in (b). The magnified image highlights the concentric multiterraced structure of an individual island. (d) X-ray diffraction pattern showing clear Laue oscillations, indicating the high crystalline quality of the SRO film. (e) Reciprocal space map and (f) $\phi$-scans measured around the (103) crystallographic reflection of the SRO thin film and Nb:STO substrate. (g) Cross-sectional high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) image of the epitaxial SRO film on the Nb:STO substrate. (h) Atomic-resolution image of the interface, revealing a sharp transition with alternating RuO$_2$/SrO/TiO$_2$ planes; the white dashed line marks the position of the SRO/Nb:STO interface. The inset shows the fast Fourier transform of the atomic-resolution image. (i) X-ray reflectivity data (black symbols) and corresponding GenX simulations (red line) based on a structural model that includes reduced scattering length density (SLD) layers at both the surface and the bottom interface (SRO/Nb:STO). The inset shows the real part of the extracted SLD depth profile. The dash-dotted horizontal line indicates the nominal X-ray SLD expected for stoichiometric SRO, shown as a reference.
• Figure 2: (a) Rutherford backscattering spectroscopy spectrum of the SRO/Nb:STO thin film together with RUMP simulations assuming stoichiometric SrRuO$_3$ (blue line) and a Ru-deficient composition (SrRu$_{0.75}$O$_3$, red line). (b) Cross-sectional HAADF-STEM image of the region used for compositional mapping. (c–f) Corresponding STEM-EDS elemental maps of Ru, Ti, Sr, and O, respectively. The dashed lines in panels (c–f) are guides to the eye marking the approximate positions of the film surface and the buried SRO/Nb:STO interface. (g) Laterally integrated and normalized STEM-EDS intensity profiles extracted from the region indicated in panel (c), plotted as a function of depth $z$, where $z = 0$ corresponds to the SRO/Nb:STO interface.
• Figure 3: (a) Temperature dependence of the resistivity for SRO thin films grown on undoped STO and Nb-doped STO substrates. (b) Field-cooled magnetization measured under an applied field of 50 Oe, with the field oriented perpendicular (out-of-plane, OP) and parallel (in-plane, IP) to the film surface. The inset shows the first derivative of the magnetization, highlighting the Curie temperature ($T_\text{Curie}$) and the antiferrodistortive transition temperature ($T_\text{AFD}$) of the STO substrate. (c,d) Magnetic hysteresis loops M($H$) measured at 135 K and 120 K, respectively, showing weak perpendicular magnetic anisotropy (PMA) with nearly identical OP and IP responses. (e) M($H$) measured at 100 K, revealing the onset of a clear bifurcation between OP and IP magnetization, indicative of a crossover toward enhanced PMA. (f,g) M($H$) measured at 80 K and 5 K, respectively, displaying a pronounced enhancement of PMA, with easier saturation along the OP direction. For clarity, the hysteresis loops at 135 K and 120 K are shown up to 3 T due to increased noise at higher fields. All M($H$) were performed after field cooling the sample in 5 T.
• Figure 4: (a–c) Spin asymmetry [(R$+$) - (R$-$)]/[(R$+$) + (R$-$)] measured at (a) 100 K, (b) 80 K, and (c) 5 K after field cooling the sample in a 4.8 T in-plane magnetic field. Open symbols represent the experimental data and solid lines the fits obtained from the best-refined model. (d) Nuclear scattering length density (nSLD, left axis) and magnetization profiles (right axis) derived from the magnetic scattering length density (mSLD) extracted from the fits. The dashed line indicates the nominal nuclear SLD expected for stoichiometric bulk SrRuO$_3$. The reduced nSLD relative to the reference value indicates Ru deficiency extending throughout the film thickness. The interfacial regions at the film surface and at the SRO/STO interface, where the nSLD is further reduced and consistent with enhanced Ru deficiency, are highlighted by shaded areas. The magnetization profiles reveal finite magnetization at all temperatures, with reduced values in these interfacial regions compared to the film bulk.
• Figure 5: Polarized off-specular neutron scattering (OSS) maps measured at 100 K, 80 K, and 5 K for both spin channels R+ (a,c,e) and R- (b,d,f), after field cooling in a 4.8 T in-plane magnetic field. The intensities were background-subtracted, and the same color scale is used for all temperatures to highlight relative changes in the scattering distribution. Panels (g,h) show simulated OSS patterns calculated from the PNR-derived structural and magnetic parameters at 5 K, reproducing the main experimental features.