Magnetic coupling at the interface between ultrathin tetragonal CuO and ${\mathrm{La}}_{0.7}{\mathrm{Sr}}_{0.3}{\mathrm{MnO}}_3$

High-symmetry rocksalt type tetragonal CuO (T-CuO) does not exist in bulk but can be synthesized via thin film epitaxy limited to a few unit cells (3, 4) thick and above which it relaxes to its bulk tenorite structure. Direct probe into magnetic properties of T-CuO layer has been a challenge because of its ultrathin limit. Here, we demonstrate the interfacial magnetic coupling between ultrathin T-CuO and ferromagnetic (${\mathrm{La}}_{0.7}{\mathrm{Sr}}_{0.3}{\mathrm{MnO}}_3$) layers in an epitaxial $\mathrm{CuO}\text{/}{\mathrm{La}}_{0.7}{\mathrm{Sr}}_{0.3}{\mathrm{MnO}}_3$ bilayer grown on (001)-oriented ${\mathrm{SrTiO}}_3$. We observe a positive exchange bias shift of $\sim 30$ Oe at 2 K in $\mathrm{CuO}\text{/}{\mathrm{La}}_{0.7}{\mathrm{Sr}}_{0.3}{\mathrm{MnO}}_3$ bilayer. The observation of positive exchange bias indicates that there exists antiferromagnetic exchange coupling between Mn and Cu moments at the interface. Notably, the exchange bias vanishes at 5 K and it is discussed in view of the proposed spin structure revealed from low-energy muon spin rotation and x-ray magnetic circular dichroism study [Phys. Rev. B 103, 224429 (2021)]. Furthermore, an enhanced Gilbert damping, linewidth broadening and larger inhomogeneous $4\pi M_{\mathrm{eff}}$ value from in-plane ferromagnetic resonance measurements, are the direct consequence of antiferromagnetic exchange coupling at the $\mathrm{CuO}\text{/}{\mathrm{La}}_{0.7}{\mathrm{Sr}}_{0.3}{\mathrm{MnO}}_3$ interface. Combining both static and dynamic magnetic characterization, we establish an understanding of interfacial exchange coupling in $\mathrm{CuO}\text{/}{\mathrm{La}}_{0.7}{\mathrm{Sr}}_{0.3}{\mathrm{MnO}}_3$ bilayer.

Figure 1

(a) The tetragonal crystal structure of CuO. (b) Slab of T-CuO viewed along [001] face. Bright and dark cyan atoms represent two different sublattices. (c) ${\mathrm{CuO}}_2$ plane consisting of corner shared ${\mathrm{CuO}}_4$ plaquettes. (d) Each CuO plane can be viewed to consists of two interpenetrating ${\mathrm{CuO}}_2$ sublattices (The solid and dotted black squares belong to two different sublattices). The red squares represent edge shared ${\mathrm{CuO}}_4$ plaquettes spanning the T-CuO plane. (e) Sketch of CuO/LSMO bilayer grown on STO substrate viewed along [100] direction. (f) Schematic interfacial spin configuration giving rise to a positive exchange bias effect in the T-CuO/LSMO bilayer. The cooling field ($H_{\mathrm{COOL}}$) induces a net magnetization $M_{\mathrm{PIN}}$(denoted by the cyan-colored arrow) in the T-CuO layer (larger and smaller arrows denote spins along and opposite to $H_{\mathrm{COOL}}$) in the direction of $H_{\mathrm{COOL}}$, which is pinned due to defects. Due to the AFM coupling ($J_{INT}<0$) at the interface, $M_{\mathrm{PIN}}$ induces an effective positive exchange bias ($H_{\mathrm{EB}}$, denoted by the blue-colored arrow) for the ferromagnetic LSMO layer, in the direction opposite to $H_{\mathrm{COOL}}$.

Figure 2

Structural characterization by using RHEED and XRD techniques. RHEED pattern of (a) STO (001) at $650{}^{\circ }\mathrm{C}$, (b) after the growth of ultrathin T-CuO layer at $650{}^{\circ }\mathrm{C}$ and (c) CuO/LSMO bilayer after the growth of LSMO on top of T-CuO layer at $650{}^{\circ }\mathrm{C}$. (d) XRR data of bare LSMO (10 nm) and CuO (1 nm)/LSMO(10 nm) bilayer samples. (e) XRD $\theta -2\theta$ patterns of bare LSMO and CuO/LSMO bilayer samples. Inset: The FWHMs of the $\omega$ scan performed around (002) plane for LSMO (20 nm) film and STO substrate.

Figure 3

[(a) and (b)] Temperature dependence of magnetization of bare LSMO (10 nm) layer and CuO/LSMO bilayer under an in-plane magnetic field of 100 Oe. (Insets) $T_C$ is determined by $dM\text{/}dT$ of the ZFC curve. (c) Enlarged plot of $M\left(H\right)$ loop of bare LSMO (10 nm) layer and CuO/LSMO bilayer at 2 K in ZFC mode. (d) $M\left(H\right)$ loop of bare LSMO (10 nm) layer and CuO/LSMO bilayer at 2 K in ZFC mode.

Figure 4

Exchange bias measurements. Shifted $M\left(H\right)$ loops of CuO/LSMO bilayer measured at 2 K after the sample is cooled down (a) under FC mode of $+1$ and $-1\mathrm{T}$, (b) under FC mode of $-5\mathrm{T}$, and (c) enlarged $M\left(H\right)$ symmetrical loop of single LSMO (10 nm) layer under cooling field of $-1$ and $+1\mathrm{T}$.

Figure 5

Exchange shift and change in coercivity of CuO/LSMO sample from EB measurements. [(a) and (b)] Cooling field dependence of $H_{\mathrm{EB}}$ and $H_C$ field measured at 2 K. [(c) and (d)] Temperature dependence of $H_{\mathrm{EB}}$ and $H_C$ field measured in presence of $-1\mathrm{T}$ magnetic field in FC mode. The inset shows thermoremanent magnetization measured at 2 K. The solid lines are guide to the eyes.

Figure 6

Magnified view of original and inverted loops of CuO/LSMO bilayer measured at (a) 2, (b) 3, (c) 4, and (d) 5 K, after the FC process.

Figure 7

Room temperature ferromagnetic resonance measurements. FMR derivative spectra measured at different frequencies for 10-nm-thick LSMO (a) without and (b) with 1-nm ultrathin T-CuO interfacial layer, and (c) a 20-nm-thick LSMO samples. The corresponding Lorentzian fits are shown by solid lines for all the samples.

Figure 8

Frequency-dependent resonance field $H_r$ for (a) LSMO (10 nm), (b) CuO/LSMO bilayer, and (c) LSMO (20 nm) films fitted to the lines shown by solid circles. FMR linewidths as a function of frequency (d) 10-nm LSMO, (e) CuO/LSMO bilayer, and (f) 20-nm LSMO samples.