Antiferromagnetic interlayer exchange coupling and large spin Hall effect in multilayer systems with Pt/Ir/Pt and Pt/Ir layers

We investigated Pt/Ir/Pt and Pt/Ir multilayers as candidates of nonmagnetic spacer layers in synthetic antiferromagnetic (AF) layers, which are available for the systematic study on AF spintronics. In these systems, we observed (i) AF interlayer exchange coupling in Pt/Ir/Pt and Pt/Ir nonmagnetic spacer layers sandwiched by Co layers and (ii) large spin Hall conductivity in Pt/Ir multilayer heavy-metal systems which is essential to achieve low power consumption spin-orbit torque switching. We found that total nonmagnetic spacer layer thickness $\left[t_{\mathrm{total}}=t_{\mathrm{Pt}}\left(\mathrm{Pt}\mathrm{thickness}\right)+t_{\mathrm{Ir}}\left(\mathrm{Ir}\mathrm{thickness}\right)\right]$ range in which AF interlayer exchange coupling is observed is wide in Co/nonmagnetic spacer layer/Co with Pt/Ir/Pt and Pt/Ir nonmagnetic spacer layers. Moreover, the large spin Hall angle of ${\theta }_{\mathrm{SH}}=10.3\%$ and low resistivity of ${\rho }_{xx}=35\mu \mathrm{\Omega }\mathrm{cm}$ in Pt/Ir multilayer heavy metal are observed. These results indicate that Pt/Ir/Pt and Pt/Ir are nonmagnetic spacer layers allowing us to achieve the AF interlayer exchange coupling and generation of large spin-orbit torque via spin Hall effect in synthetic AF coupling layer system.

Figure 1

(a) Schematic diagram for memory cell of spin-orbit torque (SOT)-MRAM with metal-oxide-semiconductor (MOS) transistors and synthetic antiferromagnetic (AF) layer. The red arrow shows the current (I) direction. (b) Proposed synthetic AF coupling layer structure with Pt/Ir/Pt nonmagnetic spacer layer. (c) An example of sputtered film stacks for measurement of SHE and resistivity of Pt/Ir multilayer HM. (d) Out-of-plane XRD profiles for the stacks with Pt 7 nm (sample E), ${\left[\mathrm{Pt}\left(1.2\right)/\mathrm{Ir}\left(0.8\right)\right]}_4$ (sample J), and ${\left[\mathrm{Pt}\left(0.6\right)/\mathrm{Ir}\left(0.6\right)\right]}_6$ (sample K) multilayer HMs on Ta (0.5)/CoFeB (1.5) layers. Cross-sectional transmission electron microscopy images in (e) Ta(0.5)/CoFeB(1.5)/Pt(3.5)/Ta(1) and (f) $\mathrm{Ta}\left(0.5\right)/\mathrm{CoFeB}\left(1.5\right)/{\left[\mathrm{Pt}\left(1.0\right)/\mathrm{Ir}\left(0.8\right)\right]}_2/\mathrm{Ta}\left(1\right)$. The top Ta layers are completely oxidized.

Figure 2

Normalized magnetization vs field (M-H) curves for (a)–(c) sample A ($t_{\mathrm{Ir}}=0.5$, 1.0, 1.4 nm) and (d)–(f) sample B ($t_{\mathrm{Ir}}=0.5$, 1.0, 1.4 nm). The red and blue arrows in the light-pink color boxes in (a)–(f) indicate the corresponding magnetization state and the magnetization direction of each Co layer. (g) (h) Film structures (stacks) of samples A and B.

Figure 3

Magnitude of the interlayer exchange coupling ($|J_{\mathrm{ex}}|$) as a function of Ir thickness ($t_{\mathrm{Ir}}$) measured at room temperature in samples A and B. The oscillation period (Λ) of AF interlayer exchange coupling is about 0.95 nm for both films.

Figure 4

Normalized M-H curves for (a) $\mathrm{Ta}\left(3\right)/\mathrm{Pt}\left(3\right)/{\left[\mathrm{Co}\left(1.3\right)/\mathrm{Pt}\left(0.8\right)/\mathrm{Ir}\left(0.5\right)/\mathrm{Pt}\left(0.8\right)\right]}_2/\mathrm{Co}\left(1.3\right)/\mathrm{Ir}\left(1\right)/\mathrm{Ta}\left(2\right)$ [sample D ($n=2$, $t_{\mathrm{Co}}=1.3\mathrm{nm}$)] and (b) Ta(3)/Pt(3)/Co(1.1)/Pt(0.8)/Ir(0.5)/ Pt(0.8)/Co(1.1)/Ir(1)/Ta(2) [sample D ($n=1$, $t_{\mathrm{Co}}=1.1\mathrm{nm}$)]. AF interlayer exchange coupling through Pt/Ir/Pt trilayers was observed. (c) AF interlayer exchange coupling ($-J_{\mathrm{ex}}$) as a function of $t_{\mathrm{total}}=t_{\mathrm{Pt}}+t_{\mathrm{Ir}}$ for sample C with various $t_{\mathrm{Pt}}$ and $t_{\mathrm{Ir}}$ and sample D with various $t_{\mathrm{Pt}}$ and $t_{\mathrm{Co}}$. The $-J_{\mathrm{ex}}$ values of first and second peak positions ($t_{\mathrm{Ir}}=0.5$, 1.4 nm) in sample B are also plotted in (c). The black solid line and blue dotted line in (c) are the results of least-square fit using the equation $-J_{\mathrm{ex}}\propto \exp \left(a\text{–}b{t}_{\mathrm{total}}\right)$, where $a$ and $b$ are fitting parameters ($a=1.5$, $b=1.856$), and $-J_{\mathrm{ex}}\propto {t_{\mathrm{total}}}^{\text{–}2}$. The −2nd power law of the $t_{\mathrm{total}}$ is the result of fitting assuming the RKKY interaction [72, 73].

Figure 5

(a), (b) Film stacks in prepared Hall bar devices. Schematic diagram of (c) prepared device and (d) photograph of the typical device. (e) Sheet conductance ($G_{xx}$) as a function of HM thickness ($t_{\mathrm{total}}$). The solid lines in (e) are linear fits to the data and the dotted line in (e) is the least-square fit using Eq. (1). (f) Estimated resistivity (${\rho }_{xx}$) for Pt, ${\left[\mathrm{Pt}\left(0.4\right)/\mathrm{Ir}\left(0.4\right)\right]}_n$, ${\left[\mathrm{Pt}\left(0.6\right)/\mathrm{Ir}\left(0.6\right)\right]}_n$, ${\left[\mathrm{Pt}\left(0.8\right)/\mathrm{Ir}\left(0.8\right)\right]}_n$, and ${\left[\mathrm{Pt}\left(1.2\right)/\mathrm{Ir}\left(0.8\right)\right]}_n$ HMs for samples E–J.

Figure 6

Typical longitudinal resistance $R_{xx}$ vs external magnetic field H oriented along the $y$ axis (red squares) and $z$ axis (black circles) measured at 305 K for the device with (a) ${\left[\mathrm{Pt}\left(1.0\right)/\mathrm{Ir}\left(0.8\right)\right]}_1$-multilayer HM and (b) ${\left[\mathrm{Pt}\left(0.6\right)/\mathrm{Ir}\left(0.6\right)\right]}_2$-multilayer HM. Typical SMR $\mathrm{\Delta }{R}_{xx}/R_{xx}^{H=0}$ plotted against the HM layer thickness ($t_{\mathrm{total}}$) for Pt and [Pt/Ir]-multilayer HM systems (samples E, I, and K). The solid lines show the fitting results using drift diffusion model.

Figure 7

(a) Estimated magnitude of the spin Hall angle (${\theta }_{\mathrm{SH}}$), (b) spin-diffusion length (λs), and (c) spin Hall conductivity (${\sigma }_{\mathrm{SH}}$) for Pt and [Pt/Ir]-multilayer HM for samples E–J. The solid black lines in (a)–(c) are guides for the eyes.

Figure 8

SMR $\mathrm{\Delta }{R}_{xx}/R_{xx}^{H=0}$ plotted against the HM layer thickness ($t_{\mathrm{total}}$) for (a) samples L–N and (b) samples H–J with $t_{\mathrm{Pt}}=0.8$, 1.0, and 1.2 nm. The solid lines show the fitting results using drift diffusion model.

Figure 9

Pt thickness ($t_{\mathrm{Pt}}$) dependence of (a) estimated magnitude of the resistivity (${\rho }_{xx}$), (b) spin Hall angle (${\theta }_{\mathrm{SH}}$), (c) spin Hall conductivity (${\sigma }_{\mathrm{SH}}$), and (d) spin-diffusion length (λs) for $\left[\mathrm{Pt}\left(t_{\mathrm{Pt}}\right)/\mathrm{Ir}\left(0.6\right)\right]$ HMs (samples L–N) (red solid squares) and $\left[\mathrm{Pt}\left(t_{\mathrm{Pt}}\right)/\mathrm{Ir}\left(0.8\right)\right]$ HMs (samples H–J) (blue solid circles). The solid red and blue lines in (a)–(d) are guides for the eyes.