Simultaneous detection of the spin-Hall magnetoresistance and the spin-Seebeck effect in platinum and tantalum on yttrium iron garnet

The spin-Seebeck effect (SSE) in platinum (Pt) and tantalum (Ta) on yttrium iron garnet has been investigated by both externally heating the sample (using an on-chip Pt heater on top of the device) and by current-induced heating. For SSE measurements, external heating is the most common method to obtain clear signals. Here we show that also by current-induced heating it is possible to directly observe the SSE, separate from the also present spin-Hall magnetoresistance (SMR) signal, by using a lock-in detection technique. Using this measurement technique, the presence of additional second-order signals at low applied magnetic fields and high heating currents is revealed. These signals are caused by current-induced magnetic fields (Oersted fields) generated by the used ac current, resulting in dynamic SMR signals.

Figure 1

(a) Microscope image of the device structure, consisting of a Pt or Ta Hall-bar detector (bottom layer) and a Pt heater (top layer), separated by an insulating ${\mathrm{Al}}_2{\mathrm{O}}_3$ layer. Ti/Au pads are used for contacting the device. For external heating experiments, the device is contacted as marked. The applied field direction is given by ${\alpha }_0$, as defined in the figure. (b) Second-harmonic voltage signal generated by the SSE for a fixed magnetic field direction of ${\alpha }_0=-{90}^{\circ }$ and (c) angular dependence of the SSE signal applying a magnetic field of 50 mT, in both Pt and Ta.

Figure 2

Current-induced heating experiments on the YIG/Pt sample. (a) Magnetic field dependence of the first-order resistance for ${\alpha }_0=0^{\circ }$, showing the SMR signal for applied ac currents of 4 mA and 8 mA. (b) Angular dependence of the first-order SMR signal for $I_{\mathrm{ac}}=8$ mA, for applied magnetic fields of 0.9 mT, 1.8 mT, and 100 mT. (c) and (d) show the corresponding second-harmonic voltage signals, respectively. For applied magnetic fields above 10 mT, the second-harmonic response only shows the SSE signal. For low applied magnetic fields an additional signal is observed on top of the SSE signal. The black symbols in both figures are a guide for the eye. They show equal measurement conditions comparing the results shown in both figures. The inset of (c) shows the used measurement configuration.

Figure 3

ac-current dependence of the second-harmonic voltage for an applied field of (a) 0.9 mT and (b) 50 mT. For the shown experiments, the transverse current-induced heating measurement configuration has been used. The vertical dashed lines mark the data plotted in (c), which shows the ac-current dependence of the magnitude of the signal at ${\alpha }_0=0^{\circ }$ for $B=0.9$ mT (red dots) and $B=50$ mT (black squares) and the average magnitude of the peaks (peak to peak) around ${\alpha }_0=\pm {90}^{\circ }$ for $B=0.9$ mT (blue triangles). The dashed lines are a guide for the eye.

Figure 4

Second-harmonic response of the transverse voltage for the YIG/Ta sample. (a) Magnetic field sweep for ${\alpha }_0=0^{\circ }$ and (b) angular dependence for two different applied magnetic fields (0.9 mT and 50 mT), the dashed line is a guide for the eye. The SSE signal (=signal at high field) has opposite sign compared to the YIG/Pt sample, whereas the low-field behavior is similar for both samples. The black symbols in the figures point out the equal measurement conditions comparing both figures.

Figure 5

(a) Current dependence and (b) magnetic field dependence of the calculated second-harmonic response, including the dynamic SMR [from Eq. (8)] and the SSE for the YIG/Pt system. (c) Calculated second-harmonic response for the YIG/Ta sample using a scaling factor for $\frac{d\alpha }{dI}$ of 0.5. For all calculations the SSE signal is extracted from the measurements.

Figure 6

Evaluation of Eq. (A7) for a selected set of measurements: (a) assuming $V_3=0$ and (b) including the measured values of $V_3$. Similarly for Eq. (A8): (c) assuming $V_4=0$ and (d) including the measured values of $V_4$.