Title: Tunable Nanoislands Decorated Tapered Optical Fibers Reveal Concurrent Contributions in Through-Fiber SERS Detection

Abstract: Creating plasmonic nanoparticles on a tapered optical fiber tip enables a remote SERS sensing probe, ideal for challenging sampling scenarios like biological tissue, specific cells, on-site environmental monitoring, and deep brain structures. However, nanoparticle patterns fabricated from current bottom-up methods are mostly random, making geometry control difficult. Uneven statistical distribution, clustering, and multilayer deposition introduce uncertainty in correlating device performance with morphology. Here, we employ a tunable solid-state dewetting method to create densely packed monolayer Au nanoislands (NIs) with varied geometric parameters, directly contacting the silica TF surface. These patterns exhibit analyzable nanoparticle sizes, densities, and uniform distribution across the entire taper surface, enabling a systematic investigation of particle size, density, and analyte effects on the SERS performance of the through-fiber detection system. The study is focused on the SERS response of a widely employed benchmark Rhodamine 6G molecule and Serotonin, a neurotransmitter with high relevance for the neuroscience field. The numerical simulations and limit of detection (LOD) experiments on R6G show that the increase of the total near-field enhancement volume promotes the SERS sensitivity of the probe. However, for serotonin we observed a different behavior linked to its interaction with the nanoparticle's surface. The obtained LOD is as low as 10-7 M, a value not achieved so far in a through-fiber detection scheme. Therefore, we believe our work offers a strategy to design nanoparticle-based remote SERS sensing probes and provide new clues to discover and understand the intricate plasmonic-driven chemical reactions.