Probing ultrafast spin dynamics through a magnon resonance in the antiferromagnetic multiferroic ${\text{HoMnO}}_3$

We demonstrate an approach for directly tracking antiferromagnetic (AFM) spin dynamics by measuring ultrafast changes in a magnon resonance. We test this idea on the multiferroic ${\mathrm{HoMnO}}_3$ by optically photoexciting electrons, after which changes in the spin order are probed with a THz pulse tuned to a magnon resonance. This reveals a photoinduced change in the magnon line shape that builds up over 5–12 picoseconds, which we show to be the spin-lattice thermalization time, indicating that electrons heat the spins via phonons. We compare our results to previous studies of spin-lattice thermalization in ferromagnetic manganites, giving insight into fundamental differences between the two systems. Our work sheds light on the microscopic mechanism governing spin-phonon interactions in AFMs and demonstrates a powerful approach for directly monitoring ultrafast spin dynamics.

Figure 1

(a) THz transmission spectrum, defined as $T_R\left(\nu \right)=|E_{\mathrm{no}\mathrm{pump}}\left(\nu \right)/E_{\mathrm{incident}}\left(\nu \right)|$, where $E_{\mathrm{no}\mathrm{pump}}$ is the static THz spectrum transmitted through the ${\mathrm{HoMnO}}_3$ crystal, and $E_{\mathrm{incident}}$ is the THz spectrum before the crystal. The THz magnetic field is perpendicular to the $c$ axis of the crystal. The magnon mode that we excite is illustrated at the bottom left. (b) $E_{\mathrm{no}\mathrm{pump}}\left(t\right)$ at 10 K, and the differential transmitted field $[E_{\mathrm{pumped}}\left(t\right)-E_{\mathrm{no}\mathrm{pump}}\left(t\right)]$ for $F=2.5\mathrm{mJ}/{\mathrm{cm}}^2$ and $\tau =100$ ps (green). The inset shows the incident THz pulse. (c) The differential transmission spectrum (green), $\mathrm{\Delta }{T}_R\left(\nu \right)=|E_{\mathrm{pumped}}\left(\nu \right)/E_{\mathrm{incident}}\left(\nu \right)|-T_R\left(\nu \right)$, calculated from (b), and the 10 K static transmission spectrum from (a) (blue) for comparison.

Figure 2

(a) The photoinduced transmission change $\mathrm{\Delta }{T}_R(\nu ,\tau )$ versus frequency and pump-probe delay. (b) A simulation of the data in (a), calculated from the static THz transmission spectra in Fig. 1.

Figure 3

(a) Normalized optical-pump, THz-probe signals from ${\mathrm{HoMnO}}_3$ at different temperatures and fluences. The dashed black lines are fits to the $2.5\mathrm{mJ}/{\mathrm{cm}}^2$ curves. (b) The amplitude of the signal versus temperature for $F=2.5\mathrm{mJ}/{\mathrm{cm}}^2$, taken at the maximum of the OPTP signals. The green line shows the expected lattice heating on the right $y$ axis, calculated using a fluence of $2.5\mathrm{mJ}/{\mathrm{cm}}^2$ and the steady-state specific-heat data from [31]. (c) The rise time of the pump-probe signals taken from the fits shown in (a), including a few additional temperature points. The blue dashed line is a fit to $\text{constant}\times T^Q$, where $Q=-0.5$.

Figure 4

Measured photoinduced THz transmission change (defined in the caption of Fig. 1) at $\tau =100$ ps for different (a) sample temperatures with a pump fluence of $2.5\mathrm{mJ}/{\mathrm{cm}}^2$ and (b) for different pump fluences at 10 K. (c) Simulations of heating with a fluence of $2.5\mathrm{mJ}/{\mathrm{cm}}^2$ for different initial temperatures for comparison with (a). In these simulations, the final temperature, which varies with initial temperature, was taken from the green curve in Fig. 3. (d) Simulations using $T_{\mathrm{initial}}=10$ K and different final temperatures for comparison with (b). The simulations were done by taking the difference of the equilibrium transmitted E fields with no optical pump at different temperatures from Fig. 1. The spectra were all normalized to the value of the transmission change at the highest fluence for (a) and (c), and to the 10 K transmission change for (b) and (d).