Field-Free Magnetization Switching by an Acoustic Wave

Surface acoustic waves (SAWs) propagating on magnetostrictive ferromagnets can induce magnetization reversal: their weak damping and mature technology make them ideal for remote wave control of magnetic bits. Experimental demonstrations of this spectacular coupling have so far required the simultaneous application of a static magnetic field. We show here SAW-driven all-acoustic switching (AAS) over millimetric distances. It relies on the triggering of magnetization precession of a uniaxial in-plane magnetized film by acoustic ferromagnetic resonance at zero field. Moreover, the absence of a biasing field enables toggling between the two equilibrium magnetic states for over 20 consecutive acoustic pulses. This proof-of-concept opens up an alternative path to magnetic-bit manipulation.

Figure 1

(a) A schematic of the setup. For the switching experiments, Kerr microscopy images are taken over the area delimited by the dashed contour. For SAW FMR experiments (Fig. 3), $\tau$-long pulses are applied to the right transducer. The field is applied along the hard axis [110] and the opposite transducer detects the acoustic echo. (b),(c) Kerr microscopy differential images ($306\times 410\mu {\mathrm{m}}^2$) after a single 250 ns-long SAW burst at $T=100$ K ($B=0$ mT, $P_{\mathrm{rf}}=31$ dBm), and opposite initializations. The black and white arrows indicate the magnetization direction.

Figure 2

Zero-field SAW-driven magnetization switching following a 250-ns-long 987-MHz rf pulse applied to the right (a) or left (b),(c) transducer (images $306\times 410\mu {\mathrm{m}}^2$). When observing far from the excitation, the acoustic and magneto-acoustic losses need to be compensated to obtain a similar switching efficiency; hence the larger power (+35 dBm) needed for excitation through ${\mathrm{IDT}}_2$.

Figure 3

The relative transmitted SAW amplitude versus the hard-axis field, at high rf power. Inset: the field dependence of the $T=100$ K precession frequency, calculated taking into account the anisotropy constants determined by cavity FMR. In both graphs, the arrows point to the resonance conditions for which the precession frequency most closely matches $f_{\mathrm{SAW}}=990$ MHz.

Figure 4

(a)–(f) Zero-field toggling behavior under successive 250 ns-long 990 MHz SAW bursts at $T=100$ K ($B=0$ mT, $P_{\mathrm{rf}}=35$ dBm, image $306\times 410\mu {\mathrm{m}}^2$), starting from a $+{\mathbf{M}}_0$ configuration: normalized local magnetization of selected $10\times 10{\mathrm{pixel}}^2$ bins. The labels on the image locate the position of each bin with respect to the IDT. (g) The toggling count (TC, defined in the Supplemental Material) summed over the SAW wavefront, as a function of the distance to the SAW emitter.