Effects of transition metal spacers on spin-orbit torques, spin Hall magnetoresistance, and magnetic anisotropy of Pt/Co bilayers

We studied the effect of inserting 0.5-nm-thick spacer layers (Ti, V, Cr, Mo, W) at the Pt/Co interface on the spin-orbit torques, the Hall effect, magnetoresistance, saturation magnetization, and magnetic anisotropy. We find that the dampinglike spin-orbit torque decreases substantially for all samples with a spacer layer compared to the reference Pt/Co bilayer, consistently with the opposite sign of the atomic spin-orbit coupling constant of the spacer elements relative to Pt. The reduction of the dampinglike torque is monotonic with atomic number for the isoelectronic $3d,4d$, and $5d$ elements, with the exception of V that has a stronger effect than Cr. The fieldlike spin-orbit torque almost vanishes for all spacer layers irrespective of their composition, suggesting that this torque predominantly originates at the Pt/Co interface. The anomalous Hall effect, magnetoresistance, and saturation magnetization are also all reduced substantially, whereas the sheet resistance is increased in the presence of the spacer layer. Finally, we evidence a correlation between the amplitude of the spin-orbit torques, the spin-Hall-like magnetoresistance, and the perpendicular magnetic anisotropy. These results highlight the significant influence of ultrathin spacer layers on the magnetotransport properties of heavy-metal/ferromagnetic systems.

Figure 1

Left: Device schematic, coordinate system, and electrical connections. Right: Cross section of the sample.

Figure 2

(a) Magnetization curves of Pt/Co and Pt/Ti/Co measured by vibrating sample magnetometry as a function of in-plane magnetic field. (b) $M_{\mathrm{s}}$ of Pt/Co and Pt/$X$/Co. The two values for each element correspond to the measurements of the reference Pt/Co layers (black circles) codeposited with Pt/$X$/Co (red squares). (c) Anomalous Hall resistance $\left(R_{\omega ,\mathrm{H}}\right)$ of Pt/Co and Pt/W/Co as a function of out-of-plane field. (d) Effective perpendicular magnetic anisotropy energy $\left({\mathit{K}}^{\perp }\right)$ and (e) $R_{\mathrm{AHE}}$ for all the samples extracted from measurements similar to the ones shown in (c). (f) Sheet resistance of all the samples obtained by four-point measurements.

Figure 3

(a) First-harmonic longitudinal resistance $\left(R_{\omega ,\mathrm{L}}\right)$ of Pt/Co measured by rotating the sample in a fixed external field of 1.8 T. (b) Illustration of the rotation planes. (c)–(e) Magnetoresistance in the three planes expressed in percentage of the 240–275 Ω resistance for all the samples [note the $y$-axis breaks in (c) and (d)].

Figure 4

(a) First-$\left(R_{\omega ,\mathrm{H}}\right)$ and second-$\left(R_{2\omega ,\mathrm{H}}\right)$ harmonic Hall resistances of Pt/Co measured by rotating the sample in a fixed external field with various amplitudes. (b),(c) Second-harmonic coefficients obtained by fitting $R_{2\omega ,\mathrm{H}}$ using Eqs. (1) and (2) (see text for details). (d) $b_{\mathrm{DL}}$ and (e) $b_{\mathrm{FL}}$ normalized to $j=1\times {10}^7\mathrm{A}/\mathrm{c}{\mathrm{m}}^2$. (f) Difference between the values of $b_{\mathrm{DL}}$ and $b_{\mathrm{FL}}$ obtained in the Pt/Co and Pt/$X$/Co samples.

Figure 5

(a) Correlation between the $zy$-plane magnetoresistance and the dampinglike SOT. The first five data points (pentagons) correspond to the Pt/$X$/Co layers; the star-shaped data point is an average of all Pt/Co bilayers. (b) Correlation between ${\mathit{K}}^{\perp }$ and the $b_{\mathrm{DL}}$ obtained in Pt/Co and Pt/$X$/Co evidencing that $b_{\mathrm{DL}}$ increases in layers with larger ${\mathit{K}}^{\perp }$, however the reference Pt/Co layer does not follow the trend.