KAIST

Some materials possess an inherent property called “optical anisotropy,” in which the refractive index changes depending on the direction in which light passes through. This is a decisive “optical fingerprint” that reveals the internal structure and molecular arrangement of the material. There are two types of optical anisotropy. Uniaxial anisotropy refers to the case where only one direction is special, like a pencil, while biaxial anisotropy is a more general and complex case where all three directions differ, like a brick.

Professor YongKeun Park’s research team previously developed, for the first time in the world, “Dielectric Tensor Tomography (DTT),” a technology capable of measuring this optical fingerprint in three dimensions, opening a path for 3D dielectric tensor measurement that had not previously existed (Shin et al., Nature Materials, 2022). However, conventional DTT required a precise laser interferometer, which caused noise in images, reduced accuracy, and made the system highly sensitive to external vibrations. In particular, there were technical limitations in expanding it to large-area samples such as biological tissues.

The iDTT developed by the research team performs a total of 48 independent measurements by precisely controlling the polarization and angle of light used in hospitals. Through this, it reconstructs in three dimensions the “dielectric tensor,”* which fully describes how a material responds to light in all directions.
 *Dielectric tensor: a 3×3 matrix that represents how a material responds to light, including refraction and absorption, in all directions. It mathematically describes the characteristics of materials whose optical properties vary depending on direction.

The core of iDTT lies in the introduction of an LED light source. By using LED illumination as an incoherent light source, iDTT fundamentally resolves these noise issues and greatly improves measurement stability and practicality. In fact, in a direct comparison using a sample with micrometer-scale periodic molecular alignment structures, the research team confirmed that iDTT clearly reconstructed fine structures that were almost invisible due to noise in conventional laser-based DTT.

The iDTT technology is expected to be applicable across materials science, semiconductors, pharmaceuticals, biomedicine, and displays.

The research team succeeded in making visible in three dimensions how molecules are arranged inside liquid crystal particles. They also precisely observed fibrosis, a phenomenon in which tissue hardens, in colon tissue after radiation therapy without any additional staining.

In addition, even when different crystalline materials such as quartz and calcium chloride were mixed together, the system automatically distinguished each material based solely on differences in their response to light (anisotropy), without chemical analysis.

Furthermore, in materials composed of multiple crystals, the technology non-destructively analyzed the orientation of each small crystal and whether the crystals were well aligned with each other (coherence) or misaligned (incoherence). Through this, the team confirmed that iDTT is a new analytical method capable of connecting microscopic internal structures with physical properties such as material strength.

>

Professor YongKeun Park stated, “This study suggests the possibility of replacing material anisotropy measurements that previously relied on large-scale facilities or destructive analysis with compact optical microscopy,” adding, “As stable dielectric tensor measurements are now possible using LEDs, this technology will become a new standard for non-destructive precision analysis used across various industrial fields.”

This study, with KAIST integrated master’s–PhD student Juheon Lee as the first author, was published in the world-renowned journal Nature Photonics on April 21, 2026.
 ※ Paper title: “Incoherent dielectric tensor tomography for quantitative three-dimensional measurement of biaxial anisotropy,” DOI: 10.1038/s41566-026-01897-0

This research was supported by the National Research Foundation of Korea’s Global Leader Research Program, the Korea Institute for Advancement of Technology’s International Collaborative R&D Program, and the Samsung Research Funding Center of Samsung Electronics.