Strong reduction of the coercivity by a surface acoustic wave in an out-of-plane magnetized epilayer

Inverse magnetostriction is the effect by which magnetization can be changed upon application of stress/strain. A strain modulation may be created electrically by exciting interdigitated transducers to generate surface acoustic waves (SAWs). Hence SAWs appear as a possible route towards induction-free undulatory magnetic data manipulation. Here we demonstrate experimentally on an out-of-plane magnetostrictive layer a reduction of the coercive field of up to $60\%$ by a SAW, over millimetric distances. A simple model shows that this spectacular effect can be partly explained by the periodic lowering of the strain-dependent domain nucleation energy by the SAW. This proof of concept was done on (Ga,Mn)(As,P), a magnetic semiconductor in which the out-of-plane magnetic anisotropy can be made very weak by epitaxial growth; it should guide material engineering for all-acoustic magnetization switching.

Figure 1

(a) Schematics of the experimental setup (not to scale). A thin ${\mathrm{SiO}}_2$ layer (not represented) is deposited on the (Ga,Mn)(As,P). The SAW assists domain nucleation in the half period during which it lowers the DW energy. (b) Time dependence of the strain (not to scale). SAWs are generated in bursts of duration ${\tau }_{\text{SAW}}=600$ ns, every $T_{\text{rep}}=20$ ms, during the entire field plateaus of length $T_f=4$ s. (c) Hysteresis cycle averaged in front of the emitting IDT without SAW ($T=10$ K) and Kerr microscopy image ($365\times 273\mu {\text{m}}^2$) of the domains during the field reversal. The magnetic part of the image has been divided by a reference image taken at remanence and the rightmost part left raw to show the transducer.

Figure 2

SAW-assisted reversal at $P_0=8.9$ W/mm ($T=30$ K): (a) Hysteresis cycles averaged in front of the SAW emitter, with and without SAW. Inset: Magnetic aftereffect experiment at $B=2$ mT under 6000 SAW bursts, corresponding to a duration of 6 min ($f_{\text{rep}}=15$ Hz), 5 min ($f_{\text{rep}}=20$ Hz), or 3.5 min ($f_{\text{rep}}=30$ Hz), intensity averaged in front of the IDT normalized to its initial value. (b)–(g) Kerr microscopy images ($690\times 924\mu {\text{m}}^2$) corresponding to the cycle under SAW in (a).

Figure 3

(a) Hysteresis cycles averaged on the SAW path ($T=10$ K), without SAW (solid line), off IDT resonance at $P_0$ (open symbols), and at the IDT resonance for different powers. (b) Spatial map of the coercive field based on the hysteresis cycle taken at $P_0$ (image size $690\times 924\mu {\text{m}}^2$). Blue (red) indicates a poor (good) SAW efficiency. Transparent bins indicate that the data were locally too poor to extract a coercive field.

Figure 4

Temperature dependence of the coercive field averaged in front of the IDT (incident rf power $P_0$).

Figure 5

Kerr microscopy images ($690\times 924\mu {\text{m}}^2$) taken at $P_0$, $T=10$ K, and $B=11$ mT. The vertical streaks in the first and last images are microscopy artifacts.

Figure 6

(a) Fraction of reversed magnetization vs field calculated by the model, with and without SAW. (b) Time dependence of the SAW-induced strain and nucleation probability at $B=7.35$ mT. Dotted lines materialize the onset of nucleation triggered by reaching a threshold compressive strain. (c) Power dependence of the coercive field at $T=10$ K, nucleation model, and data, taken as the central value of the Gaussian distribution of $B_c$ in a small area in front of the IDT.