Non-thermal electrons open the non-equilibrium pathway of the phase transition in FeRh

The optical excitation of metals initially creates short-lived non-Fermi distributions of the electrons. The electrons and holes excited far above and below the Fermi level quickly relax to hot Fermi-distributions that subsequently cool via electron-phonon scattering. Here, we show that such non-thermal charge carriers beyond the Fermi-distribution speed up the prototypical first-order antiferromagnetic-to-ferromagnetic phase transition in FeRh. In ultrafast x-ray diffraction experiments, we vary the maximum electron temperature by increasing the pump pulse duration up to 10ps. For direct optical excitation of FeRh, ferromagnetic domains nucleate within 8ps as soon as the successively deposited energy surpasses the site-specific threshold energy. In contrast, suppressing the direct optical excitation by an optically opaque Pt layer leads to a nucleation on a 50ps timescale driven by the near-equilibrium heat transport. These findings unambiguously identify the photo-excitation of non-thermal electrons and not electron-phonon non-equilibria to enable the rapid phase transition in FeRh.

This dataset contains all raw data, data evaluation and plot scripts used in the linked publication. The data.rar archive contains 3 folders, linked to the measurements, the simulations and the publication figures. The data analysis scripts as well as the modelling of the laser-induced lattice dynamics are written in Python.

金属的光激发最初会产生寿命短暂的电子非费米分布（non-Fermi distribution）。在费米能级（Fermi level）上下被激发的电子与空穴会快速弛豫为热费米分布，随后通过电子-声子散射（electron-phonon scattering）实现冷却。本研究证实，这类超出费米分布范畴的非热载流子（non-thermal charge carriers）能够加速FeRh中典型的一级反铁磁-铁磁相变。在超快X射线衍射（ultrafast X-ray diffraction）实验中，我们通过将泵浦脉冲宽度提升至10皮秒（ps），以此调控最大电子温度。针对FeRh的直接光激发实验表明，当持续沉积的能量超过位点特异性阈值能量时，铁磁畴会在8皮秒内开始成核。与之形成对比的是，通过光学不透明铂（Pt）层抑制直接光激发后，相变成核将以50皮秒的时间尺度进行，其驱动力为近平衡热输运。上述发现明确证实，驱动FeRh发生快速相变的机制是光激发产生的非热电子，而非电子-声子非平衡态。

本数据集包含该关联论文中使用的全部原始数据、数据处理脚本与绘图脚本。data.rar压缩包内含3个文件夹，分别对应实验测量、模拟计算与论文配图。数据分析脚本以及激光诱导晶格动力学建模代码均采用Python语言编写。