Silver-Programmed Dual-Optical Au Nanostructures and Machine Learning for Intelligent Biosensing

The evolution of biosensors demands synergistic improvements in signal transduction and data processing. We present a universal biosensing platform that combines dual-mode signal responses from silver-modulated gold nanorods (AuNRs) and gold-silver nanoclusters (AuAgNCs) (including localized surface plasmon resonance (LSPR) shifts and fluorescence variations) with machine learning (ML)-enhanced image analysis. Initially, AuNR was synthesized and transformed into silver-coated gold nanorods (AuNR@Ag) via silver reduction, with LSPR shifts precisely characterized. Concurrently, AuAgNCs were engineered to enhance their fluorescence through antigalvanic reactions between surface Ag(I) and Au(0) cores. The dual-mode platform leverages the silver-linked fluorescence intensity of AuAgNCs and LSPR of AuNR@Ag, as well as the increasingly enhanced inner filter effect between AuAgNCs and the evolving LSPR of AuNR@Ag, enabling simultaneous fluorescence and colorimetric readouts. The platform achieved high-precision detection of alkaline phosphatase via dual-signal correlation (R² > 0.99), demonstrating robustness in complex matrices. Furthermore, to facilitate point-of-care testing applications, an ML algorithm encompassing feature extraction, dimensionality reduction, and model validation was integrated to process bimodal signal images. The subsequent data analysis exhibited robust correlations (R² > 0.95), thereby substantiating the effectiveness of this approach in analyzing bimodal data. The ML-augmented analytics was validated for the analysis of serum samples, giving results that matched well with those from the spectra-based standard method. This work bridges nanomaterial engineering with ML-augmented analytics, offering a versatile framework for next-generation biosensors with clinical diagnostic potential.