Optical Excitation of Propagating Magnetostatic Waves in an Epitaxial Galfenol Film by Ultrafast Magnetic Anisotropy Change

Using a time-resolved optically pumped scanning-optical-microscopy technique, we demonstrate the laser-driven excitation and propagation of spin waves in a 20-nm film of a ferromagnetic metallic alloy Galfenol epitaxially grown on a $\mathrm{Ga}\mathrm{As}$ substrate. In contrast to previous all-optical studies of spin waves, we employ laser-induced thermal changes of magnetocrystalline anisotropy as an excitation mechanism. A tightly focused 70-fs laser pulse excites packets of magnetostatic surface waves with an $e^{-1}$-propagation length of $3.4\mu \mathrm{m}$, which is comparable with that of permalloy. As a result, laser-driven magnetostatic spin waves are clearly detectable at distances in excess of $10\mu \mathrm{m}$, which promotes epitaxial Galfenol films to the limited family of materials suitable for magnonic devices. A pronounced in-plane magnetocrystalline anisotropy of the Galfenol film offers an additional degree of freedom for manipulating the spin waves’ parameters. Reorientation of an in-plane external magnetic field relative to the crystallographic axes of the sample tunes the frequency, amplitude, and propagation length of the excited waves.

Figure 1

Experimental geometry. A dc magnetic field $\mathbf{H}$ is applied along the $x$ axis, while the azimuthal orientation of the sample $\phi$ can be varied. The relative pump-probe distance is changed along either the $x$ or $y$ direction to detect propagation of backward volume magnetostatic wave (BVMSW) or MSSW modes, respectively. The blue-red pattern represents SW propagation in the $x$-$y$ plane schematically. The inset shows the $x_1{x}_2{x}_3$ reference frame linked to the crystallographic axes of the sample, with the angles ${\phi }_m$, $\phi$, and $\psi$ defining the in-plane orientations of the magnetization $\mathbf{M}$, external dc magnetic field $\mathbf{H}$, and the SW wave vector $\mathbf{k}$, respectively.

Figure 2

(a),(b) Experimental spatial-temporal $\mathrm{\Delta }y\text{−}t$ maps of $\mathrm{\Delta }{\theta }_K$ obtained at $\phi ={45}^{\circ }$ (a) and $\phi ={60}^{\circ }$ (b) and $\mathrm{\Delta }x=0$, i.e., when the pump-probe distance $\mathrm{\Delta }y$ is scanned transversely to H, corresponding to the MSSW configuration. Arrows are guides to the eye showing the centers of the propagating MSSW packets. (c) Experimental (symbols) temporal evolution of $\mathrm{\Delta }{\theta }_K$ at different pump-probe distances $\mathrm{\Delta }y$, at $\phi ={45}^{\circ }$ and $\mathrm{\Delta }x=0$. Solid lines denote the fits using Eqs. (1) and (2). (d) Experimental spatial-temporal $\mathrm{\Delta }x\text{−}t$ map of $\mathrm{\Delta }{\theta }_K$ obtained at $\phi =-{30}^{\circ }$ and $\mathrm{\Delta }y=0$, i.e., when the pump-probe distance $\mathrm{\Delta }x$ is scanned along H, corresponding to the BVMSW configuration. (e) Experimental (symbols) dependence of the precession frequency $f_0$ on the strength of the magnetic field $H$ applied at $\phi ={30}^{\circ }$ and theoretical dependence (solid line) calculated using Eq. (5). (f) Calculated (symbols) temporal evolution of $\mathrm{\Delta }{\theta }_K$ at different pump-probe distances $\mathrm{\Delta }y$, at $\phi ={45}^{\circ }$ and $\mathrm{\Delta }x=0$. Solid lines denote the fits using Eqs. (1) and (2).

Figure 3

(a),(b) Azimuthal dependences of the frequency $f_0$ (a) and amplitude $A_{\mathrm{SW}}^0$ (b) of the precession observed at $\mathrm{\Delta }y=\mathrm{\Delta }x=0$. Symbols show experimental data; solid lines are the theoretical fits. (c) Normalized amplitude of MSSW packets $A_{\mathrm{SW}}(\mathrm{\Delta }y)/A_{\mathrm{SW}}(\mathrm{\Delta }y=0.5\mu \mathrm{m})$ vs distance $\mathrm{\Delta }y$ as obtained for several angles $\phi$; lines are the fits using Eq. (3). (d) Propagation length $L_{\mathrm{prop}}$ of MSSW packets vs $\phi$ as obtained from the experimental fits (symbols) and using Eq. (10) (solid line).

Figure 4

(a) Dispersion of MSSWs reconstructed via 2D FFT (real part) at $\phi ={45}^{\circ }$. (b),(c) FFT spectra at different pump-probe shifts $\mathrm{\Delta }y$ at $\phi ={45}^{\circ }$ and ${30}^{\circ }$. (d),(e) Schematic representation of the frequency changes in the heated area when H is along a hard (HA) and easy (EA) anisotropy axis, respectively. Solid lines denote dependences of $f(H)$ for a nonheated film and dashed lines $f{}^{\prime }(H)$ for a heated film. Vertical dashed lines represent the magnitude of H in the experiments.

Figure 5

Calculated SW signals in the $x$-$y$ plane at different time delays $t$ after the excitation, 0.3 ns (a),(d), 0.7 ns (b),(e), and 1 ns (c),(f) for the angles $\phi$ of $-{15}^{\circ }$ (a)–(c) and $-{45}^{\circ }$ (d)–(f).