RESEARCH

Ultrafast Light–Matter Interactions and Applications

We investigate ultrafast light–matter interactions in a wide range of nano- and quantum materials, primarily using femtosecond pulse–based time-resolved spectroscopic techniques. We focus on elucidating nonequilibrium dynamics occurring on femtosecond to picosecond timescales, including exciton dynamics and many-body interactions1–3, hot carrier dynamics4, polarization-dependent anisotropic photophysics1,3,5–8, and spatiotemporal diffusion phenomena8,9.

Beyond fundamental studies, we actively pursue application-oriented research built upon these ultrafast physical processes. In particular, we explore ultrafast all-optical switching as a platform for high-speed, high-functionality nanophotonic devices. By exploiting multiple optical control knobs—such as light polarization1,6–8, strain1, and photon-energy-dependent excitation7—we aim to establish new physical principles and methodologies that enable high-performance optical switching and advanced photonic functionalities.

References.

1.  Suk et al., Light Sci. Appl. 13, 240 (2024).

2.  Bae et al., Small 17, 2103400 (2021).

3.  Seo et al., Phys. Rev. Appl. 18, 014010 (2022).

4.  Seo et al., Adv. Opt. Mater. 13, 2403531 (2025).

5.  Seo et al., Adv. Opt. Mater. 13, 2500032 (2025).

6.  Suk et al., Laser Photonics Rev. 18, 2300680 (2024).

7.  Suk et al., Adv. Opt. Mater. 11, 2300370 (2023).

8.  Seo et al., Adv. Opt. Mater. 11, 2201544 (2023).

9.  Seo et al., ACS Photonics 9, 1783 (2022).

Nanophotonic Machine Learning

We explore photonic machine learning by leveraging ultrafast light–matter interactions in nanophotonic platforms. We recently demonstrated that ultrafast transient absorption responses of quasi-one-dimensional nanomaterials can be utilized for machine-learning tasks within the framework of extreme learning machines, establishing a new paradigm in which ultrafast optical dynamics directly enable computation1. Building on this work, we aim to develop more advanced and higher-performance photonic machine-learning platforms by incorporating richer physical degrees of freedom and introducing new computational principles rooted in nonequilibrium ultrafast dynamics. 

References.

1.  Seo et al., Adv. Opt. Mater. e02892, Early view (2025).