A dual-algorithm optimized BK7/Cu/KTaO₃/GaN SPR biosensor for stage-specific malaria detection

Malaria diagnosis at low parasitaemia remains a critical challenge, and existing optical biosensing designs largely rely on single-method evaluation without algorithmic optimization or cross-method electromagnetic validation. This work proposes and optimizes a multilayer prism-coupled surface plasmon resonance (SPR) biosensor for label-free, stage-specific malaria detection, exploiting refractive index variations in Plasmodium falciparum -infected red blood cells. A comparative framework combining the Non-dominated Sorting Genetic Algorithm II (NSGA-II) and Brute-Force Search, each coupled with the Transfer Matrix Method (TMM), identified the optimal BK7/Cu/KTaO₃/GaN configuration of copper at 48 nm, KTaO₃ at 3 nm, and GaN at 4 nm that is independently validated by Finite Element Method (FEM) simulations. The optimized sensor achieves a peak sensitivity of 567.14°/RIU, quality factor of 232.44 RIU⁻¹, and detection accuracy of 0.4098 deg⁻¹ for the ring stage, with clear discrimination across all three intraerythrocytic stages is 567.14°, 395.00°, and 339.51°/RIU for ring, trophozoite, and schizont, respectively, and strong electromagnetic confinement at the Cu–KTaO₃ interface. The design is robust against fabrication tolerances, patient-to-patient refractive-index variability, and temperature fluctuations, while NSGA-II reaches the brute-force optimum at a fraction of the computational cost. A proposed GaN-surface biofunctionalization strategy further offers a pathway toward malaria-specific clinical selectivity. To our knowledge, this is the first work to jointly apply two optimization algorithms with TMM–FEM co-validation on a Cu/KTaO₃/GaN SPR architecture for malaria sensing, outperforming recently reported designs in sensitivity and quality factor. The proposed sensor is a promising candidate for label-free malaria diagnostics at the point of care in low-resource settings.