Researchers at Yonsei University have developed a non-contact, AI-driven optical sensing system capable of identifying ethanol molecules in the air. By using a graphene-based Fresnel lens (an ultrathin diffractive lens that focuses light through interference), the team captures subtle light-field distortions caused by gas molecules. This method avoids traditional, direct chemical sampling, which often suffers from slow response times and material degradation. Instead, the system translates light-field patterns into data that a deep-learning model interprets to determine ethanol concentration. Findings from this work were published in the journal Opto-Electronic Advances on June 07, 2026.

The sensing approach relies on the principle that gas molecules alter the wavefront of a laser beam passing through them. While these distortions are too faint for direct observation, the specially designed lens transforms these interactions into measurable changes in the focal spot's size and shape. The researchers trained a deep-learning model to recognize these nonlinear optical patterns as unique fingerprints for specific molecular concentrations. By choosing a longer light wavelength, the team prioritized system robustness and stability over raw sensitivity, mitigating the noise typically associated with shorter, more variable wavelengths.

This integrated approach combines physical optics with computational intelligence to perform rapid, stable gas detection without consumables. Prof. Seong Chan Jun, who led the research at the School of Mechanical Engineering, emphasizes that this platform is designed to extract physically meaningful features from light for quantitative prediction. Beyond monitoring environmental pollutants, the technology shows potential for medical diagnostics, such as non-invasive breath analysis, and industrial safety applications. Because the system is compact and based on visible-light optics, it could eventually be integrated into wearable devices for real-time monitoring in complex environments.

Researchers at Yonsei University have created a new way to detect ethanol in the air without needing a physical probe. Traditional sensors often require the gas to be captured or touched, which wears out the equipment over time and makes for slow readings. Instead, this team uses a special, paper-thin material called a graphene-based Fresnel lens to watch how light behaves when it passes through gas. Think of it like a camera that doesn't just take a picture, but analyzes the way light bends around tiny gas particles to figure out exactly what they are.

When a laser beam travels through the air, ethanol molecules cause tiny, invisible distortions in the light's path. Because these changes are too subtle for human eyes to spot, the team built an AI model to act as the interpreter. The lens catches these distortions and creates a pattern, and the AI reads that pattern like a digital fingerprint to calculate the exact amount of gas present. By using longer wavelengths of light, the team ensured the system stays steady and reliable, avoiding the jittery noise that often ruins other high-tech sensors.

This breakthrough is exciting because it blends physics with smart computing to create a fast, touch-free detector that never runs out of parts. Since it works with light, the device could eventually be shrunk down to fit inside wearable tech. This could change how we monitor our environment, helping factories stay safe from gas leaks, or even allowing doctors to check a patient's health just by analyzing their breath without any invasive testing.