Variable damping and coherence in a high-density magnon gas

We report on the fast relaxation behavior of a high-density magnon gas created by a parametric amplification process. The magnon gas is probed using the technique of spin-wave packet recovery by parallel parametric pumping. Experimental results show a damping behavior that is in disagreement with both the standard model of exponential decay and with earlier observations of nonlinear damping. In particular, the inherent magnon damping is found to depend upon the presence of the parametric pumping field. A phenomenological model, which accounts for the dephasing of the earlier injected magnons, is in good agreement with the experimental data.

Figure 1

(a) Experimental setup, consisting of a single-crystal YIG waveguide, input and output antennas, and a dielectric resonator for the application of the pumping microwave field. The wave forms are the shapes of the input, pumping, and restored pulses as seen on the oscilloscope. (b) Section of the spin-wave dispersion spectrum for a thin magnetic waveguide schematically showing the hybridization of Damon-Eshbach (dashed line) wave with standing spin-wave modes.

Figure 2

Spin-wave power envelope as experimentally observed at the oscilloscope. The light red areas schematically show the time intervals when the pump is on. The continuous red line represents the results of the two pump-pulse recovery. The first pump pulse is switched off before the recovered signal saturates. After a delay time ${\tau }_{\mathrm{delay}}$ the pumping is switched on a second time the second recovered signal is observed. The dashed blue line represents the experimental result for the pumping not being interrupted.

Figure 3

(a) The amplitudes of the second pulse shown in Fig. 2a are extracted and plotted against ${\tau }_{\mathrm{delay}}$ for different durations ${\tau }_{\mathrm{pump}1}$ of the first pumping pulse. The amplitudes are normalized to the maximum observable amplitude using only one long pumping pulse. (b) The time $t_{\max }$ at which the peak amplitude of the second pulse in Fig. 2a is observed as a function of ${\tau }_{\mathrm{delay}}$.

Figure 4

(a) Left axis: measured (circles connected by a dashed line) and simulated (dashed line) time $t_{\max }$ at which the peak amplitude of the recovered pulse is observed as a function of ${\tau }_{\mathrm{delay}}$ for ${\tau }_{\mathrm{pump}1}=400$ ns. Right axis: measured (dots connected by a solid line) and simulated (solid line) peak amplitudes of the recovered pulse for ${\tau }_{\mathrm{pump}1}=400$ ns. (b)–(d) Simulated time profiles of the signal (thin solid lines) and of the internal pumping (bold dashed lines). The light red areas indicate the time intervals when the pump is on. The gray area in panel (b) indicates the input spin-wave pulse. (b) Uninterrupted pumping (${\tau }_{\mathrm{delay}}=0$). (c) ${\tau }_{\mathrm{delay}}=300$ ns. (d) ${\tau }_{\mathrm{delay}}=700$ ns. $t=0$ corresponds to switching on the second pump. Duration of the first pump pulse ${\tau }_{\mathrm{pump}1}=400$ ns.