This paper aims to study a highly sensitive multilayer surface plasmon resonance (SPR) biosensor with machine learning assistance; a dual-layer SPR sensor is proposed for brain lesion detection.
A Kretschmann configuration-based prism SPR sensor consisting of a combination of an SF5 prism with Cu and Y2O3 is proposed. Cu is chosen as a cost-effective alternative to Au and to overcome Ag’s oxidation issues. Then, only an additional layer of Y2O3 is used. To avoid difficulties in multilayer deposition, the proposed sensor is optimized with only two layers. The effect of the Cu and Y2O3 layers on the sensing performance is properly described.
The sensor shows a wider sensing range as well as higher refractive index (RI) sensing capabilities, as the RI of brain-injured tissues ranges from 1.3333 to 1.4833. This sensor exhibits excellent performance, including an FOM of 145.09 RIU-1, a sensitivity of 320.66 °/RIU, a DA of 1.04 /° and an SNR of 11.53.
This sensor can detect RI up to 1.485, a value rarely reported in the literature. The inclusion of Y2O3 in a prism-based SPR sensor is novel to the best of our knowledge. Furthermore, Multilayer Perceptron Regressor (MLPRegressor) model is incorporated to predict the resonance angle of any unknown analyte.
