Toward reproducible SERS biosensing: analysing instability origins and mitigation approaches from sample preparation...

Despite its single-molecule sensitivity, surface-enhanced Raman scattering (SERS) faces critical challenges in spatial and temporal reproducibility that hinder clinical translation in biosensing. This review systematically dissects the origins of signal fluctuations (quantified by RSD) across three interconnected domains: (i) sample-level heterogeneity driven by droplet drying dynamics (coffee-ring effect, crystallization, and stochastic molecular adsorption); (ii) substrate-induced variability from random hotspot distribution, batch-to-batch fabrication drift, and laser-induced photodegradation; and (iii) instrumental artifacts including depth-of-field constraints, dynamic fluorescence backgrounds, and system instabilities. Crucially, we establish a direct mapping between physical mechanisms and mitigation strategies, such as microdroplet confinement to suppress salt crystallization, extended-depth-of-field microscopy for topography tolerance, and shifted-excitation Raman difference spectroscopy for fluorescence correction, showing how targeted interventions can substantially reduce signal fluctuations. This framework paves the way for transforming SERS from qualitative observation to quantitative diagnostics, enabling low-RSD quantitative biomarker detection in complex biological samples and accelerating standardized deployment in point-of-care testing. By distinguishing technical artifacts from genuine biological signals, we propose a comprehensive end-to-end framework spanning sample preparation to detection. Our work establishes guiding principles for engineering robust SERS platforms, addressing the core bottleneck that has long impeded the technology's transition from benchtop research to clinical reality.