Design and evaluation of SiC - based photonic crystal biosensor for amino acid detection and screening of hyperglycinemia - augmented with machine learning

Abstract

Detection of glycine and other amino acids is demonstrated using a SiC based two- dimensional photonic crystal biosensor. SiC is one of the materials suitable for developing a biosensor, which could help in the diagnosis of hyperglycinemia. To evaluate the performance of the sensor, quality factor (QF), wavelength sensitivity, figure of merit (FOM) and detection limit are calculated, and the achieved values are 522.7, 211 nm/RIU, 56.1 RIU⁻¹ and 0.0018 RIU. The FDTD (Finite difference time domain) approach is used for designing the sensor and understanding the field distribution and transmission. The sensing area of a biosensor is a crucial factor that affects the performance of the sensor. The photonic crystal rod radius in the sensing area is investigated and the suitable radius is identified. Sensor parameters are evaluated for three different values of rod radius. Amino acids such as leucine, proline, lysine and methionine are detected along with glycine. The topology of the sensor is chosen to improve the sensitivity and spectral width, which in turn increases QF. Sensitivity exhibited by the sensor is suited to detect each component and detection limit demonstrated is remarkable for biosensing. K–nearest neighbor (KNN), Random Forest (RF) and XGBoost (XGB) regression are utilized to predict intensity across analyte spectra to enhance the analysis.