Enhanced Fuel Adulteration Detection Using Surface Plasmon Resonance Biosensor with Machine Learning Optimization in the terahertz regime

Fuel adulteration poses a significant challenge to fuel quality control, with far-reaching effects on engine performance, environmental sustainability, and regulatory compliance. This study introduces a novel approach by integrating surface plasmon resonance (SPR) biosensing with advanced machine learning techniques to enhance fuel adulteration detection. The use of the CatBoost gradient boosting algorithm achieves highly accurate sensor performance predictions, with an optimal coefficient of determination of 100%. Furthermore, the sensor exhibits exceptional versatility, including the potential for 2-bit encoding, extending its applicability beyond fuel adulteration detection. The sensor also demonstrates superior characteristics, such as a maximum sensitivity of 1773 GHzRIU⁻¹ and a figure of merit (FOM) of 20.661 RIU⁻¹, surpassing previously reported designs. This combined approach, leveraging electromagnetic simulations, cutting-edge machine learning, and multifunctional design, offers a rapid and precise solution for fuel quality assessment. Additionally, it opens new avenues for advanced sensing and encoding applications in the terahertz regime, representing a substantial advancement in SPR biosensor technology.