Machine learning approach to surface plasmon resonance sensor based on MXene coated PCF for malaria disease detection in RBCs

Abstract

A photonic crystal fiber (PCF)-based biosensor is proposed with surface plasmon resonance (SPR) sensing approach. This sensor is used for malaria infected red blood cells (RBCs) detection as well as haemoglobin (Hb) concentration detection. A machine learning (ML) algorithm is also implemented and analyzed for the proposed sensor. For SPR mechanism, a Ti3C2Tx thin film is coated over the gold coated PCF. The malaria stages in RBCs are detected by comparing the resonance wavelength of healthy and infected RBCs samples. Finite element method-based simulation is used for analyzing the sensing performance. The measured wavelength sensitivity for ring stage is 12,142 nm/RIU, for trophozoite stage is 9736 nm/RIU and for schizont stage is 8241 nm/RIU. In addition, the Hb concentration in RBCs is detected with maximum wavelength sensitivity of 53 nm/g/dl. Further, the resolution of 10−5 RIU is obtained for the proposed sensor. The low mean squared error of 0.01526 and less than 2% error has been obtained in sensitivity investigation with the ML technique. Due to the enhanced sensing performance and ML approach, the proposed sensor can be the best alternative to the existing malaria sensor.