Â鶹ÒùÔº

Color, as the way light's wavelength is perceived by the human eye, goes beyond a simple aesthetic element, containing important scientific information like a substance's composition or state.

Spectrometers are optical devices that analyze material properties by decomposing light into its constituent wavelengths, and they are widely used in various scientific and industrial fields, including material analysis, chemical component detection, and life science research.

Existing high-resolution spectrometers were large and complex, making them difficult for widespread daily use. However, thanks to the ultra-compact, high-resolution spectrometer developed by KAIST researchers, it is now expected that light's color information can be utilized even within smartphones or wearable devices.

Professor Mooseok Jang's research team at the Department of Bio and Brain Engineering has successfully developed a reconstruction-based spectrometer technology using double-layer disordered metasurfaces. The work is published in .

Existing high-resolution spectrometers have a large form factor, on the order of tens of centimeters, and require complex calibration processes to maintain accuracy. This fundamentally stems from the operating principle of traditional dispersive elements, such as gratings and prisms, which separate light wavelengths along the propagation direction, much like a rainbow separates colors.

Consequently, despite the potential for light's color information to be widely useful in daily life, spectroscopic technology has been limited to laboratory or industrial manufacturing environments.

The research team devised a method that departs from the conventional spectroscopic paradigm of using diffraction gratings or prisms, which establish a one-to-one correspondence between light's color information and its propagation direction, by utilizing designed disordered structures as optical components.

In this process, they employed metasurfaces, which can freely control the light propagation process using structures tens to hundreds of nanometers in size, to accurately implement "complex random patterns (speckle)".

Specifically, they developed a method that involves implementing a double-layer disordered metasurface to generate wavelength-specific speckle patterns and then reconstructing precise color information (wavelength) of the light from the random patterns measured by a camera.

As a result, they successfully developed a new concept spectrometer technology that can accurately measure light across a broad range of visible to infrared (440–1,300 nm) with a high resolution of 1 nanometer (nm) in a device smaller than a fingernail (less than 1 cm) using only a single image capture.

Dong-gu Lee, a lead author of this study, stated, "This technology is implemented in a way that is directly integrated with commercial image sensors, and we expect that it will enable easy acquisition and utilization of light's wavelength information in daily life when built into mobile devices in the future."

Professor Mooseok Jang said, "This technology overcomes the limitations of existing RGB three-color-based machine vision fields, which only distinguish and recognize three color components (red, green, blue), and has diverse applications.

"We anticipate various applied research for this technology, which expands the horizon of laboratory-level technology to daily-level machine vision technology for applications such as food component analysis, crop health diagnosis, skin health measurement, environmental pollution detection, and bio/medical diagnostics.

"Furthermore, it can be extended to various advanced optical technologies such as hyperspectral imaging, which records wavelength and spatial information simultaneously with high resolution, 3D optical trapping technology, which precisely controls light of multiple wavelengths into desired forms, and ultrafast imaging technology, which captures phenomena occurring in very short periods."

This research was collaboratively led by Dong-gu Lee (Ph.D. candidate) and Gookho Song (Ph.D. candidate) from the KAIST Department of Bio and Brain Engineering as co-first authors, with Professor Mooseok Jang as the corresponding author.