Role of spatiotemporal nonuniformities in laser-induced magnetization precession damping

Abstract

Laser-induced magnetization precession measurements in ferromagnets often reveal an anomalous decrease in the damping time near a field-induced second-order spin-orientation transition, a behavior that cannot be described by the linearized Landau-Lifshitz-Gilbert equation. Here we demonstrate that this anomaly is not a material property but results from interference of precessing local magnetizations within the inhomogeneously excited region. By combining pump-probe experiments, analytical modeling that accounts for the finite sizes of the pump and probe spots, and micromagnetic simulations, we show that the standard macrospin approach fails to capture the observed dynamics. The inhomogeneous relaxation of magnetic parameters within the excitation area distorts the measured precession envelope, while dipole fields give rise to a temporally non-monotonic term in its frequency. Our results highlight the critical role of excitation locality in a vicinity of critical fields.

Figures (4)

• Figure 1: Azimuthal and field-strength dependences of laser-induced precession frequency and damping time. (a) Exemplary laser-induced precession of magnetization measured in an Fe film (black solid line). Fourier windows used to obtain the frequency at different times are shown by colored Gaussians. The dashed line is a fit to extract the initial amplitude and damping time. (b) Points show the azimuthal dependence of the precession frequency at different times after excitation. Lines show the corresponding fits using the Smit-Suhl approach. Colors correspond to the time windows in panel (a). (c) Central frequencies obtained with windowed FFT using the time windows from panel (a) and model prediction (line). (d-f) Experimental damping time values and model predictions: (e) azimuthal angle $\phi$ dependence at $\mu_0 H_{ext} = 0.1$ T; (d) and (f) external field dependencies at $\phi = 53$ and 8 $^\circ$, respectively.
• Figure 2: Analysis using laterally nonuniform linearized and full LLG. Data points in (a), (b), and (c) correspond to those in Fig. \ref{['fig:uniform']} (d), (e) and (f), respectively. The results of the nonuniform linearized (solid black), full LLG solution (dashed gray), and micromagnetic simulations (solid red) are shown by lines. Shaded areas indicate the uncertainty in the damping time.
• Figure 3: Influence of signal interference from different points across the excitation area on the observed damping time and precession envelope. The simulated signal $m_{z}$ at the center of the pump area and the predicted measured signal $\tilde{m}_{z}$ obtained using Eq. (\ref{['eq:integral']}) are shown for two cases: (a) dispersion of the eigen frequency due to heating (inset shows the frequency dependence on the coordinate); (b) nonlinear shift of the frequency due to inhomogeneity of the initial amplitude (inset shows the coordinate dependence of the initial amplitude).
• Figure 4: Non-monotonous temporal evolution of the laser-induced dipole field contribution. (a) Dipole-induced frequency change $\Delta f$ from the experiment is obtained by subtracting the time‑dependent frequency of the laterally nonuniform LLG model (which accounts for finite pump and probe spot sizes but neglects dipole fields) from the experimental data (points). The contribution from micromagnetic simulations is obtained by the same subtraction from the simulation results (lines). (b) Effective laser‑induced heating as a function of delay time, derived from the experimental data (see text for details). The lines show a guide for the eyes. Inset: schematic illustrating the equivalence between a decrease in pump fluence and a shift in measurement time after excitation. Data are shown for two external field values: 0.02 (red) and 0.08 T (blue).