Development of Femtosecond Noise Correlation Spectroscopy Enabled by Digitizer-Based Statistical Operation

In magnetic materials, magnetization fluctuates continuously due to thermal fluctuations, even in equilibrium. Near the phase transition temperature, these fluctuations intensify and play a key role in driving the transition. Understanding their dynamics is thus essential for revealing the nature of phase transitions and critical phenomena [1].

Femtosecond noise correlation spectroscopy (FemNoC) enables real-time observation of magnetization fluctuations on the picosecond scale and has recently unveiled how they directly govern second-order phase transitions [2,3,4,5]. However, conventional FemNoC uses lock-in-based detection, which restricts the analysis to simple statistical operations and prevents more advanced statistical investigations.

To overcome this, we developed a digitizer-based FemNoC system that directly records every optical pulse, allowing flexible statistical processing. The sample was a canted antiferromagnet Sm_xEr_1-xFeO₃ (x = 0.72), which exhibits a second-order spin-reorientation transition near 320 K. Two-color femtosecond pulses probed the sample, transferring magnetization fluctuations to polarization fluctuations via the magneto-optical effect. The transmitted pulses were detected by two balanced amplified photodetectors, and polarization signals were digitized for statistical analysis, successfully capturing the magnetization fluctuation dynamics.