Temperature Dependence of Laser-Induced Demagnetization in Ni: A Key for Identifying the Underlying Mechanism

The microscopic mechanisms responsible for the ultrafast loss of magnetic order triggered in ferromagnetic metals by optical excitation are still under debate. One of the ongoing controversies is about the thermal origin of ultrafast demagnetization. Although different theoretical investigations support a main driving mechanism of thermal origin, alternative descriptions in terms of coherent interaction between the laser and the spin system or superdiffusive spin transport have been proposed. Another important matter of debate originates from the experimental observation of two time scales in the demagnetization dynamics of the $4f$ ferromagnet gadolinium. Here, it is still unclear whether it is necessary to invoke two distinct microscopic mechanisms to explain such behavior, or if one single mechanism is indeed sufficient. To uncover the physics behind these two unsolved issues, we explore the dependence of ultrafast-demagnetization dynamics in nickel through a survey of different laser intensities and ambient temperatures. Measurements in a large range of these external parameters are performed by means of the time-resolved magneto-optical Kerr effect and display a pronounced change in the maximum loss of magnetization and in the temporal profile of the demagnetization traces. The most striking observation is that the same material system (nickel) can show a transition from a one-step (one time scale) to a two-step (two time scales) demagnetization, occurring on increasing the ambient temperature. We find that the fluence and the temperature dependence of ultrafast demagnetization—including the transition from one-step to two-step dynamics—are reproduced theoretically assuming only a single scattering mechanism coupling the spin system to the temperature of the electronic system. This finding means that the origin of ultrafast demagnetization is thermal and that only a single microscopic channel is sufficient to describe magnetization dynamics in the $3d$ ferromagnets on all time scales.

Popular Summary

Exposing a magnetic material to ultrafast (femtosecond) laser pulses apparently leads to a change in the magnetic state of the material on extremely fast time scales in the range of 100 femtoseconds, as was already discovered 15 years ago. Not surprisingly, intense research efforts have been invested in exploring the fundamental physics underlying such a phenomenon and in exploiting it to develop a new data-storage technology, as the speed of the current magnetic-material-based read-write technology reaches its fundamental limit. Despite such intense efforts, however, there is still no consensus on a microscopic theory of magnetism on femtosecond time scales. In this paper, our team, by combining a comprehensive set of experiments with theoretical modeling, makes important findings that will take the debate a big step forward.

Two particular questions of controversy are at the focus of our effort. Is the laser-induced ultrafast loss of magnetization in ferromagnetic metals a thermal process? Does the observation of two time scales in the ultrafast demagnetization in some ferromagnetic metals mean that the microscopic theory of femtosecond magnetism must involve two distinct mechanisms? To answer these questions, we have taken nickel, one of the best-known ferromagnetic materials, and investigated systematically how its ultrafast demagnetization dynamics changes with the ambient temperature and the intensity of the laser pulses. Our most striking finding is that the demagnetization dynamics of nickel shows a transition from a one-time-scale process to a two-time-scale process upon increasing the ambient temperature. Remarkably, we are able to reproduce both this seemingly complex phenomenon and the systematic dependence of the dynamics on the laser intensity with a microscopic theory that invokes a single mechanism: When laser-heated electrons transfer their energy to the lattice vibrations, their spins may switch directions. It is this spin switching that is responsible for the observed ultrafast demagnetization, the behavior of which is ultimately related to the temperature of the hot electrons.

In other words, the ultrafast loss of magnetization in nickel is indeed thermal, related to the thermalization of the spins with the electronic system; and a single microscopic mechanism can explain both single-time-scale and two-time-scale dynamics.

Figure 1

(a) Examples of Type-I and Type-II magnetization dynamics after excitation by a laser pulse. The magnetization (relative to its value at zero temperature, $M_{\mathrm{sat}}$) is plotted as a function of delay time between the pump and probe beams. (b) Temperatures of the electron, phonon, and spin systems ($T_{\mathrm{e}}$, $T_{\mathrm{p}}$, and $T_{\mathrm{s}}$, respectively), as a function of time. For the spin system, both Type-I and Type-II traces are depicted.

Figure 2

Fluence dependence of the laser-induced magnetization dynamics of nickel at room temperature ($T_{\mathrm{ambient}}=300\mathrm{K}$). The fluence is adjusted to six levels of the pump pulse ranging from $2.5–5.0\mathrm{mJ}/{\mathrm{cm}}^2$ in steps of $0.5\mathrm{mJ}/{\mathrm{cm}}^2$. (Symbols correspond to measurements.) The magnetization is normalized to its value at negative time delays. Lines are calculations with the refined M3TM, using the microscopic parameters obtained by fitting the $3.5\mathrm{mJ}/{\mathrm{cm}}^2$ measurement.

Figure 3

Dependence of the ultrafast demagnetization of nickel on the ambient temperature $T_{\mathrm{ambient}}$ at a fluence of $P_0=3.5\mathrm{mJ}/{\mathrm{cm}}^2$. Demagnetization dynamics measured for five different ambient temperatures from 80–480 K (symbols). Lines are calculations with the refined M3TM, using the microscopic parameters obtained by fitting the 300 K measurement.

Figure 4

Two-step dynamics of nickel. (a) Experimental verification of two-step dynamics in nickel at an ambient temperature of 480 K. (The data are taken from Fig. 3.) $M_{\min }$ denotes the minimum magnetization, and the lines serve as a guide to the eye. For the fast step, we obtain a time constant of approximately $0.2\mathrm{ps}$, whereas the slow step occurs within approximately $4\mathrm{ps}$. (b) Calculated phase diagram for ultrafast demagnetization in nickel. The white dots indicate the operating conditions of the experiments in Figs. 2, 3.