Unidirectional Spin-Wave-Propagation-Induced Seebeck Voltage in a PEDOT:PSS/YIG Bilayer

We clarify the physical origin of the dc voltage generation in a bilayer of a conducting polymer film and a micrometer-thick magnetic insulator ${\mathrm{Y}}_3{\mathrm{Fe}}_5{\mathrm{O}}_{12}$ (YIG) film under ferromagnetic resonance and/or spin wave excitation conditions. The previous attributed mechanism, the inverse spin Hall effect in the polymer [Nat. Mater. 12, 622 (2013)], is excluded by two control experiments. We find an in-plane temperature gradient in YIG which has the same angular dependence with the generated voltage. Both vanish when the YIG thickness is reduced to a few nanometers. Thus, we argue that the dc voltage is governed by the Seebeck effect in the polymer, where the temperature gradient is created by the nonreciprocal magnetostatic surface spin wave propagation in YIG.

Figure 1

(a) Sketch of the experimental setup with the pure spin current along the positive $z$ axis. (b) The top panel shows the microwave absorption spectrum of the sample. The bottom panel shows the voltage signal measured in the PEDO:PSS/YIG bilayer as scanning $H$ along the positive (black curve) and negative (red curve) $y$ axis, respectively. The microwave power is 40 mW. (c) The microwave power dependence of the voltage signal and the linear fitting (red curve). (d) ${\theta }_M$ angle dependence of the voltage with $P_{\text{in}}=40\mathrm{mW}$. The red curve is the $\sin {\theta }_M$ function fitting.

Figure 2

(a) The voltage signal measured in the PEDO:PSS/YIG bilayer as scanning $H$ along the positive (black curve) and negative (red curve) $y$ axis, respectively. The microwave power is 40 mW. (b) The measured microwave power dependence of the voltage signal and the linear fitting (red curve). (c) ${\theta }_M$ angle dependence of the voltage with $P_{\text{in}}=40\mathrm{mW}$. The red curve is the $\sin {\theta }_M$ function fitting.

Figure 3

(a) Sketch of the experimental setup with the pure spin current along the negative $z$ axis. (b) The voltage signal measured in the PEDO:PSS/YIG bilayer as scanning $H$ along the positive $y$ axis for the sample measured in the experimental setup shown in Fig. 1 (black curve) and the experimental setup shown in (a) (red curve), respectively.

Figure 4

(a) Experimental setup for the temperature gradient measurement. (b) The resistance as a function of $H$ for $R1$ (top panel) and $R2$ (middle panel). The insets are the resistance of $R1$ (black curve) and $R2$ (red curve) as a function of the temperature, respectively. The bottom panel shows $\mathrm{\Delta }{T}_x$ as a function of $H$. (c) ${\theta }_M$ angle dependence of $\mathrm{\Delta }{T}_x$. The red curve is the $\sin {\theta }_M$ function fitting. The inset is the microwave power dependence of $\mathrm{\Delta }{T}_x$ and the linear fitting (red curve). (d) The voltage signal as a function of $H$ measured in the $\mathrm{Pt}/{\mathrm{YIG}}_{\text{thin}}$ and $\mathrm{PEDOT}:\mathrm{PSS}/{\mathrm{YIG}}_{\text{thin}}$ samples with $P_{\text{in}}=300\mathrm{mW}$.