Algorithm assisted comprehensive optimization of SPR sensors towards single molecule detection

Detection of single molecules is a critical aspect of biochemistry, playing a vital role in understanding molecular interactions and functions. Surface plasmon resonance (SPR) biosensors are increasingly used for this purpose due to their ability to enable label-free and rapid monitoring. However, detecting low-concentration analytes down to the single-molecule level presents a substantial challenge for SPR sensors. This work introduces a multi-objective optimization strategy to enhance the performance of label-free SPR biosensors by improving various sensing metrics, including sensitivity, figure of merit, and depth of resonant dip. These enhancements lead to a significant improvement in the single-molecule detection capabilities of SPR sensors. Additionally, the k-means clustering method is applied to identify appropriate design parameters from the optimized parameter set, thereby mitigating the effects of processing errors on sensor performance. The developed SPR single molecule biosensor has demonstrated significant improvements in refractive index S, FOM, and DFOM, with increments of 230.22 %, 110.94 %, and 90.85 %, respectively. The optimized device was tested using mouse IgG to validate the improved sensing performance. Experimental findings indicate that this approach achieves a bulk refractive index sensitivity of 24,482.86 nm/RIU, a broad linear dynamic range from femtograms per milliliter (fg/mL) to micrograms per milliliter (μg/mL), and a detection limit as low as 54 ag/mL (0.36 aM). The multi-objective particle swarm optimization method proposed herein offers an efficient, rapid, and straightforward optimization strategy for SPR biosensors, facilitating ultra-sensitive single molecule detection with low limits of detection (LOD).