Antiferromagnetic Spin Seebeck Effect

We report on the observation of the spin Seebeck effect in antiferromagnetic ${\mathrm{MnF}}_2$. A device scale on-chip heater is deposited on a bilayer of ${\mathrm{MnF}}_2$ (110) $(30\mathrm{nm})/\mathrm{Pt}$ (4 nm) grown by molecular beam epitaxy on a ${\mathrm{MgF}}_2$ (110) substrate. Using Pt as a spin detector layer, it is possible to measure the thermally generated spin current from ${\mathrm{MnF}}_2$ through the inverse spin Hall effect. The low temperature (2–80 K) and high magnetic field (up to 140 kOe) regime is explored. A clear spin-flop transition corresponding to the sudden rotation of antiferromagnetic spins out of the easy axis is observed in the spin Seebeck signal when large magnetic fields ($>9\mathrm{T}$) are applied parallel to the easy axis of the ${\mathrm{MnF}}_2$ thin film. When the magnetic field is applied perpendicular to the easy axis, the spin-flop transition is absent, as expected.

Figure 1

(a) The crystal structure of ${\mathrm{MnF}}_2$ is presented with AFM spin structure overlaid on ${\mathrm{Mn}}^{2+}$ ions. The (110) thin film crystal orientation plane is highlighted in blue. The spin-flop transition in ${\mathrm{MnF}}_2$ is presented. (b) Device schematic outlining a typical spin Seebeck device geometry.

Figure 2

Spin Seebeck voltage response curves from ${\mathrm{MnF}}_2$ are shown with magnetic field applied parallel to the $c$ axis in (a). A control experiment is performed with a bare ${\mathrm{MgF}}_2$ substrate with Cu $(2\mathrm{nm})/\mathrm{Pt}$ (4 nm) under the same conditions at 5 K showing no measurable effect (inset). By mapping the spin-flop transition from (a), a phase diagram for ${\mathrm{MnF}}_2$ is reproduced in (b). These data are compared to data from Shapira and Foner [27] using ultrasonic attenuation (US), and differential magnetization (M), and Rezende et al. [28] from antiferromagnetic resonance (AFMR).

Figure 3

Temperature dependence of the spin Seebeck voltage response at various magnetic fields. The inset represents the 140 kOe data with the 70 kOe subtracted out to judge the temperature dependence of just the spin-flopped phase.

Figure 4

(a)–(f) Spin Seebeck voltage responses on two devices, one aligned to detect spin current parallel to the $c$ axis and one aligned to detect spin current perpendicular to the $c$ axis. The spin-flop transition is only present in the parallel configuration.