Spin Seebeck effect in antiferromagnet nickel oxide in wide ranges of temperature and magnetic field

In the spin Seebeck effect (SSE), the spin currents generated by thermal gradients applied to a magnetic film are detected by an electric voltage along an adjacent metallic layer due to the spin to charge current conversion through the inverse spin Hall effect. The SSE has been widely studied in bilayers made of insulating ferro- or ferrimagnet insulators and metals with large spin-orbit coupling such as platinum. Few studies of the SSE have been reported in antiferromagnetic insulators despite the increasing attention these materials have been receiving in the field of antiferromagnetic spintronics. Here we report an investigation of the SSE in bilayers made of well-textured polycrystalline films of antiferromagnetic NiO with platinum in a wide range of temperatures and magnetic fields. We demonstrate that the SSE is detected only when the NiO layer is fabricated with a Py underlayer to induce a macroscopic antiferromagnetic arrangement. The results are in good qualitative agreement with a drift-diffusion magnonic theory developed for single-crystal NiO.

Figure 1

(a) X-ray diffraction (XRD) pattern of the Si/Py/NiO/Py sample. The inset shows the x-ray reflectivity (XRR) spectra of a calibration sample of NiO thin film ($t\approx 27\mathrm{nm}$). The solid (red) line across the data indicates best fitting for the thickness measurement. (b) XRD pattern of the Si/NiO/Py (100 nm) sample. (c) MOKE hysteresis loop of the Py outer layer of the Si/Py/NiO/Py sample measured with field angle $\varphi =0$. Inset: The measured exchange bias field as a function of the in-plane field angle $\varphi$. The solid line is a fit with $Ccos\varphi$. (d) MOKE hysteresis loop of the Py outer layer in the Si/NiO/Py sample measured with field angles $\varphi =0, {90}^{\circ }$, and ${180}^{\circ }$.

Figure 2

(a) Schematic illustration of the sample structure used to measure the voltages measured in the metallic layer generated by the spin Seebeck effect in NiO (200 nm). (b) Variation with magnetic field of the SSE voltage in the Si/Py/NiO/Pt sample with four values of the temperature difference $\mathrm{\Delta }T$ indicated. Positive $\mathrm{\Delta }T$ corresponds to the Pt layer warmer than the substrate. (c) SSE voltage measured in Si/Py/NiO/Py vs magnetic field at the temperature differences indicated. (d) SSE voltage measured in Si/NiO/Pt at a temperature difference of $\mathrm{\Delta }T=+10\mathrm{K}$.

Figure 3

(a) Schematic illustration of the sample structure with the Pt Hall bar and indication of the thermal gradient and measured current. (b) Sketch of the sample mount showing the heater and thermometer arrangement. (c) Variation with magnetic field $H$ of the SSE voltage measured in the Pt layer with four values of the temperature difference $\mathrm{\Delta }T$ indicated, with the copper block at $T_b=290\mathrm{K}$. (d) SSE voltage vs temperature difference measured with $H=-1.0\mathrm{kOe}$.

Figure 4

(a) Variation with magnetic field of the spin Seebeck coefficient $S_{SSE}$ measured in Si/Py/NiO/Pt for various temperatures of the sample, as indicated, with a temperature difference $\mathrm{\Delta }T=10$ K. (b) Temperature dependence of the spin Seebeck coefficient measured with three magnetic field values, as indicated. The solid lines are polynomial fits with exponents obtained from the curves of the model described in Ref. [33].