Design and optimization of localized plasmon resonance sensing via square-slotted Ag-graphene-dielectric metasurfaces for dermatological cancer identification using machine learning

Abstract

Skin cancer is a dangerous, life-threatening illness impacting countless individuals globally, requiring urgent awareness, prevention, and early detection. It is one of the most common forms of cancer, often caused by excessive sun exposure or tanning, and requires early detection for effective treatment. Early detection of skin cancer is achievable through advanced sensor designs that utilize graphene material. Graphene's exceptional properties make it extremely appropriate for creating sensitive, accurate, and non-invasive diagnostic tools to identify cancer at early stages. The integration of silver (Ag), graphene, and silicon dioxide (SiO2) materials forms a highly sensitive multilayer structure, significantly enhancing the surface plasmon resonance response, which enables precise detection of skin cancer biomarkers at extremely low concentrations. An exceptional sensitivity of 1050 nm/RIU is attained, enabling efficient skin cancer detection through advanced plasmonic biosensing technology. Optimizing the biosensor design by systematically varying key physical parameters-such as layer thicknesses, slot dimensions, and material configurations-significantly enhanced its sensitivity. The optimization is also achieved by using a Machine learning algorithm. The highest R2 value of 0.99 is achieved for this research. This strategic tuning of the structural and optical characteristics enabled more accurate detection capabilities, making the sensor highly effective for early skin cancer diagnosis through plasmonic resonance.