arXiv subjects

Changi Kim

Publications and source records attributed to Changi Kim.

3 recordsLinked to original sources

arXiv2026

Magneto-Optical Detection of Anisotropic Spin Currents in Altermagnetic RuO2

Altermagnets are a recently identified class of collinear antiferromagnets that host large spin-split electronic bands, offering a promising platform for efficient spin-current generation. Among proposed candidates, the metallic oxide RuO2 is predicted to exhibit strong altermagnetic spin splitting; however, whether it sustains robust magnetic order beyond the ultrathin thickness limit remains unresolved. Here, we employ optical probes to investigate charge-to-spin conversion in a 12-nm-thick (101)-oriented RuO2 film grown on sapphire. Polarization-resolved second-harmonic generation reveals nonlinear optical responses consistent with the surface symmetry and N\'eel order of RuO2. Under an applied current, both second-harmonic generation and polar magneto-optical Kerr effect measurements detect a pronounced, directionally anisotropic spin polarization, exhibiting enhanced signals for current along [010] and strongly suppressed responses for current along [-101], in agreement with the symmetry of the altermagnetic spin-splitter effect. Non-magnetic or Rashba-type mechanisms cannot explain this symmetry-selective response. Scanning transmission electron microscopy further reveals that substantial strain persists even in relatively thick films, providing a possible explanation for the observed behavior. Therefore, these results establish RuO2 as an efficient spin source and demonstrate the potential of altermagnets for field-free spintronic devices.

cond-mat.mtrl-sci↗

arXiv2026

Lattice-mismatch Moire laser with strong flatband coupling

Inter-cell and/or interlayer coupling in Moire superlattices can generate flatbands and collective eigenmodes that enable emergent physical phenomena, motivating extensive exploration of Moire-inspired photonic devices. However, the experimental validation of robust inter-cell interactions in Moire photonic structures and the modulation of flatbands for specific photonic applications remain challenging. Here, we propose a lattice-mismatch Moire cavity and demonstrate nanolasers enabled by strong flatband coupling. In contrast to a twist-angle Moire cavity, a lattice-mismatch Moire cavity provides a stable flatband frequency and a substantial enhancement in Q factor compared to an isolated single-cell cavity, as the unit-cell size decreases. The photonic band-structure measurement of the small-unit-cell Moire cavity by photoluminescence reveals pronounced flatbands. Cell-resolved spectroscopy further confirms the presence of flatbands by identifying resonant peaks that consistently emerge across unit cells in a Moire cavity with a lattice mismatch of 102 nm, but not in a larger-unit-cell Moire cavity with a mismatch of 60 nm. Furthermore, mode selection is achieved by reducing the center-hole size, thus isolating the hexapole mode from the degenerate dipole modes while maintaining strong inter-cell coupling. Consequently, we demonstrate a low-threshold hexapole flatband laser in a single mode. Therefore, the systematic modification of the relative lattice parameters of the two constituent lattices offers a promising strategy for developing Moire nanolasers and flatband nanophotonic devices.

physics.optics↗

arXiv2025

Exploring the Capabilities of Large Language Model Encoders for Image-Text Retrieval in Chest X-rays

Multimodal learning from paired medical images and clinical text is a central challenge in medical data-driven informatics, where effective cross-modal alignment is critical for scalable analysis and retrieval. In chest radiography, vision-language pretraining is constrained by heterogeneous radiology reports that contain abbreviations, impression-only notes, and institution-specific writing styles. Unlike general-domain settings, naively aggregating large collections of noisy reports can plateau or even degrade multimodal learning when reporting styles differ substantially. We propose a domain-adapted bidirectional large language model text encoder for chest radiograph reports, trained with masked token prediction and supervised contrastive learning on stylistically diverse but clinically equivalent report variants to produce robust, generalizable text embeddings. We then integrate this encoder into a dual-tower contrastive vision-language framework using parameter-efficient adaptation to improve image-text alignment. Across 1.6 million paired studies from public datasets and a de-identified hospital cohort, the proposed models improve bidirectional retrieval accuracy and external generalization, achieving GREEN scores of 0.308 on MIMIC-CXR and 0.618 on Open-I, while reducing the degradation observed when abbreviation-rich, impression-only hospital reports are added to training.

cs.CV↗