Automated design workflow for enhancing nanoplasmonic surface sensitivity: simulation-guided optimization validated by thin-film measurements

The development of localized surface plasmon resonance biosensors is typically limited to inefficient trial-and-error design strategies and reliance on existing generic plasmonic nanostructures that fail to tailor the evanescent field to the specific properties of the bioassay. This work presents an automated design workflow which combines finite-difference time-domain simulations with a simplicial homology global optimization algorithm to systematically maximize the surface refractive index sensitivity (SRIS) of a plasmonic nanopillar structure. This combined approach allows to identify a nanostructure geometry with an electromagnetic decay length closely aligned with the hydrodynamic radius of the bioassay in use. As an example, we apply the concept to a sensor for the detection of the anti-inflammatory drug diclofenac. The improved nanostructure design was fabricated via electron beam lithography and replicated using soft UV-nanoimprint lithography. The fabricated improved sensor substrates were experimentally validated using a standardized layer-by-layer deposition of polyelectrolytes (PAH/PSS), demonstrating a significant sensitivity improvement compared to baseline structures used in previous work. The experimental validation confirms an 80% increase in SRIS (from 134.1 ± 5.3 nm RIU−1 to 245.8 ± 10.1 nm RIU−1). Application of the competitive DCF immunoassay on the optimized geometry remains subject of future work.