A team at Aalto University combined miniaturized hardware with intelligent algorithms to create a microscopic spectral sensor that can accurately identify a myriad of materials. The work of the researchers could make it possible for industries like health care, food safety, and transportation to implement miniaturized spectroscopy, using everyday devices, for various applications. A high-performance, miniaturized spectral sensor that fits inside a smartphone or wearable device could be used, for example, to monitor changes in an individual’s health, detect counterfeit drugs, or identify spoiled food. Autonomous vehicles could use the sensor for accurate, cost-effective object identification. Spectral sensing, which identifies the composition of materials by analyzing how they interact with light, traditionally has required bulky, expensive systems available only in laboratories and industrial applications. Unlike traditional spectral sensors that require large optical components like prisms or gratings, the miniaturized sensor achieves spectral differentiation through its electrical responses to light. This makes it ideal for integration into small devices. Researchers at Aalto University hold a tiny chip, designed to accommodate hundreds of ultracompact spectral sensors. Courtesy of Aalto University/Faisal Ahmed and Andreas Liapis. The miniaturized sensing system uses an electrically tunable, compact optoelectronic interface to enable accurate spectral identification. The optoelectronic interface allows precise control of electrical flow through voltage adjustments. The interface leverages both bias-voltage and gate-voltage tunability. The exceptional tunability of the optoelectronic interface enables the sensor to interact with light in many ways. The interface is combined with advanced algorithms that allow it to generate distinguishable photoresponses from various input spectra. During the sensor’s training, the researchers exposed the device to a wide range of light colors. The device learned how to generate unique electrical fingerprints for each color. Using an intelligent algorithm, the device can decode these electrical fingerprints, enabling the sensor to accurately identify materials and analyze their properties based on how the material interacts with light. “Our device is trained to recognize complex light signatures that are imperceptible to the human eye, achieving a level of precision comparable to the bulky sensors typically found in laboratories,” professor Zhipei Sun said. The sensor enables accurate spectral identification for both narrow-band and broadband complex spectra, within a tiny device footprint of 5 μm by 5 μm. In demonstrations, the sensor achieved a peak wavelength identification accuracy of approximately 0.19 nm in free space and exhibited an accuracy of approximately 2.45 nm in on-chip integrated spectral sensing. The spectral sensing and identification process used by the device could facilitate numerous applications beyond material identification, such as composition analysis through photoluminescence peak sensing and computing through encoding input spectra. The researchers demonstrated the capability of the sensor to identify a range of materials, including organic dyes, metals, semiconductors, and dielectrics, directly from their luminescence. “Our innovative spectral sensing approach simplifies challenges in material identification and composition analysis,” researcher Xiaoqi Cui said. The new sensor provides high-performance, cost-effective, miniaturized optical spectroscopy for both free-space and on-chip applications. The sensor’s configuration is universal and applicable to various semiconductors, allowing for flexible device design, compatible fabrication, and mass production for large-scale applications. With its versatility, tunability, and ability to recognize thousands of colors, the researchers anticipate that the microscopic sensor could bring the power of advanced spectroscopy to smart and wearable and handheld devices — the types of devices that are used every day. “This work is a major step forward in bringing spectroscopic identification to everyone’s fingertips,” researcher Fedor Nigmatulin said. “By integrating this ultracompact hardware with intelligent algorithms, we’ve taken a significant step toward miniature, portable spectrometers that could one day transform consumer electronics.” The research was published in Science Advances (www.doi.org/10.1126/sciadv.ado6886).