Simultaneous enhancements of polarization and magnetization in epitaxial ${\mathrm{Pb}(\mathrm{Zr}}_{0.52}{\mathrm{Ti}}_{0.48}){\mathrm{O}}_3/{\mathrm{La}}_{0.7}{\mathrm{Sr}}_{0.3}{\mathrm{MnO}}_3$ multiferroic heterostructures enabled by ultrathin $\mathrm{CoF}{\mathrm{e}}_2{\mathrm{O}}_4$ sandwich layers

Layered thin films of $\mathrm{Pb}\left(\mathrm{Z}{{\mathrm{r}}_1}_{-x}\mathrm{T}{\mathrm{i}}_x\right){\mathrm{O}}_3$ (PZT) and $\mathrm{L}{\mathrm{a}}_x\mathrm{S}{{\mathrm{r}}_1}_{-x}\mathrm{Mn}{\mathrm{O}}_3$ (LSMO) are well-known multiferroic systems that show promise for numerous applications including data storage devices and spintronics. In this work, structure-property relationships are explored in novel $\mathrm{PZT}/\mathrm{CoF}{\mathrm{e}}_2{\mathrm{O}}_4$ (CFO)/LSMO heterostructures with optimized ferroic properties. High quality, epitaxial PZT/CFO/LSMO heterostructures with the thickness of the CFO layer varying from 0 to 50 nm were grown on $\mathrm{SrTi}{\mathrm{O}}_3$ (100) substrates using an optimized pulsed laser deposition technique. An ultrathin (10–20 nm) CFO layer was found to simultaneously improve the ferromagnetic and ferroelectric (FE) characteristics of the system through distinct mechanisms. The increase in magnetization and magnetic coercivity in the CFO-containing samples was associated with a tetragonal distortion of the CFO lattice under epitaxial strain, while perpendicular anisotropy generated by the distortion stabilized an out-of-plane orientation of the easy axis of magnetization in the thinnest CFO layers. Trapped charge at the CFO/PZT interface in PZT/CFO/LSMO induced an internal built-in field in the heterostructures, resulting in the accumulation of higher switched charges during voltage cycling and enhanced polarization in the samples over PZT/LSMO. An increase in electric coercivity was also observed in the CFO-containing heterostructures, and is discussed in terms of a dielectric/FE layered capacitor model. Above a critical thickness, ∼50 nm, the presence of a CFO layer has a negative effect on both magnetization and polarization in PZT/CFO/LSMO as compared to PZT/LSMO.

Figure 1

(a) Schematic of a PZT/CFO/LSMO heterostructure. (b) XRD θ-2θ patterns for heterostructures with CFO thicknesses of 0, 10, and 50 nm [P/C(0 nm)/L, P/C(10 nm)/L, and P/C(50 nm), respectively]. (c) Representative XRD azimuthal (φ) scans performed about the PZT (101), CFO (311), and LSMO (110) crystallographic planes. (d) Rocking curves performed about the PZT (001) crystallographic plane for P/C(0 nm)/L and P/C(50 nm)/L. AFM images of (e) LSMO in P/C(50 nm)/L, (f) PZT in P/C(0 nm)/L, and (g) PZT in P/C(50 nm)/L.

Figure 2

(a) Cross-sectional TEM image of P/C(20 nm)/L. (b) and (c) HRTEM images of the PZT-CFO and the CFO-LSMO interfaces in the regions marked by green and red boxes in (a). (d) and (e) SAED patterns near the CFO-LSMO and the PZT-CFO interfaces of P/C(20 nm)/L heterostructure. The cubic unit cell is marked by a white box with an arrow pointing along the $a$ axis.

Figure 3

Asymmetric 2θ-ω XRD scans performed about the CFO (311) (left panel) and CFO (511) (right panel) crystallographic planes for (a) and (d) P/C(50 nm)/L, (b) and (e) P/C(20 nm)/L, and (c) and (f) P/C(10 nm)/L heterostructures. CFO (311) and (511) planes were chosen to avoid contributions from underlying LSMO or overlying PZT layers. Unstrained bulk CFO peak positions are indicated by dashed lines.

Figure 4

Symmetric θ-2θ XRD scans performed about the CFO (400) crystallographic plane for (a) P/C(50 nm)/L, (b) P/C(20 nm)/L, and (c) P/C(10 nm)/L heterostructures. The unstrained bulk CFO (400) peak position is indicated by a dashed line. (d) Schematic diagram illustrating the tetragonal distortion of the fcc CFO bulk lattice $(a=0.839\mathrm{nm})$ in the heterostructures. The in-plane and out-of plane lattice parameters were calculated from the asymmetric and symmetric XRD scans, respectively.

Figure 5

In-plane magnetic hysteresis $M$($H$) loops at 300 K for P/C(0–50 nm)/L heterostructures. Insets show the schematic diagram of the PZT/CFO/LSMO heterostructure and the enlarged low-field $M$($H$) curves.

Figure 6

In-plane and out-of-plane magnetic hysteresis loops at 300 K for (a) P/C(0 nm)/L, (b) P/C(10 nm)/L, (c) P/C(20 nm)/L, and (d) P/C(50 nm)/L. Insets show enlarged low-field $M$($H$) curves and the relative orientation of the magnetic anisotropies in the CFO and LSMO layers as discussed in the text.

Figure 7

(a) Unipolar (+20 kOe → −20 kOe) transverse susceptibility ratio vs magnetic field curves for P/C(10 nm)/L and P/C(50 nm)/L, with the magnetic field applied in-plane. (b) High resolution bipolar scan of the low-field region of P/C(10 nm)/L.

Figure 8

Polarization vs applied voltage $P$($V$) loops for (a) P/C(0 nm)/L, (b) P/C(10 nm)/L, (c) P/C(20 nm)/L, and (d) P/C(50 nm)/L, measured at applied voltages varying from 1 to 9 V at 1 kHz. The capacitors geometry is illustrated in the insets to (a) and (c).

Figure 9

(a) Polarization hysteresis loops at 300 K for P/C(0–50 nm)/L recorded at 1 kHz using an applied voltage of 9 V. (b) Leakage current vs voltage $J_L\left(V\right)$. Inset: Model sandwich capacitor used for the description of FE switching behavior. (c) Charging current vs voltage $J_C\left(V\right)$. Inset: Charging current and leakage densities in the positive voltage cycling for P/C(0 nm)/L and P/C(10 nm)/L heterostructures. (d) FE switching current densities vs voltage $J_S\left(V\right)$, with a Gaussian fit (solid lines). Inset: Voltage profile used during $J_C\left(V\right)$ measurements.