Label-free graphene-based surface plasmon resonance sensor for advanced male fertility evaluation with behavior prediction via polynomial regression

Male infertility affects approximately 15 % of reproductive-age couples globally, with male factors contributing to roughly 50 % of infertility cases, creating an urgent need for advanced, accessible diagnostic technologies for semen analysis. Current sperm assessment protocols rely predominantly on conventional light microscopy and Computer-Assisted Sperm Analysis (CASA) systems, which suffer from subjective interpretation, high costs, and limited accessibility in resource-constrained settings. This study presents a simple graphene-based Surface Plasmon Resonance (SPR) biosensor featuring a simple resonator architecture optimized for ultrasensitive sperm detection through label-free, real-time analysis. The electromagnetic analysis using COMSOL Multiphysics 6.3 demonstrates exceptional sensitivity ranging from 118 GHzRIU⁻¹ to 5000 GHzRIU⁻¹ across refractive indices of 1.33–1.3461 RIU, with a maximum figure of merit of 68.493 RIU⁻¹ and detection limits as low as 0.028 RIU. Machine learning optimization using polynomial regression achieved prediction accuracies of 87–91 % (R² values of 94–100 %) across critical operational parameters including graphene chemical potential (0.1–0.9 eV), geometric variations, and angular dependencies (0–80°), validating the sensor's robust performance for clinical sperm analysis applications.