Multilayer Plasmon Hybridization in Au–Ag–Au Nanoparticles Enables Extinction Amplification in Transport-Limited One-Step Confirmatory Syphilis Biosensing

Abstract

Syphilis diagnosis requires pairing treponemal and non-treponemal antibody results, yet rapid tests combining both markers lose sensitivity at the low RPR titers (≤1:4) most critical in early infection and pregnancy. The limitation lies with the reporter: conventional gold nanoparticles (AuNPs) cannot generate high-contrast signals at low antibody concentrations without exceeding the diameter that allows capillary transport through a nitrocellulose membrane. Here, we report gold–silver–gold (Au–Ag–Au) multilayer plasmonic nanoparticles (ML-PNPs) that exploit interfacial plasmon hybridization to overcome this trade-off. Polyhedral 50–58 nm particles were prepared by seed-mediated growth followed by galvanic replacement, and their layered structure was confirmed by electron microscopy, elemental mapping, and X-ray photoelectron spectroscopy. A silver interlayer between chemically robust gold domains couples across two metal–dielectric interfaces, while the outer gold shell preserves colloidal stability and thiol-based bioconjugation at hydrodynamic diameters below 60 nm. Finite-difference time-domain (FDTD) simulations and hyperspectral imaging showed a broadened, red-shifted resonance (scattering maximum 658 nm vs 555 nm for AuNPs) with 18-fold enhanced absorption. In a dual-marker colorimetric lateral flow immunoassay (CLFIA), ML-PNPs improved the visual detection limit by 3-fold and required 6-fold fewer particles per visible test spot. Across 36 clinical specimens spanning RPR titers of 1:1 to 1:128, treponemal and nontreponemal test-line intensities increased by 19–46%, and low-titer 1:1 samples missed by AuNP strips were reliably detected, with full agreement with reference methods and no cross-reactivity with HIV-1 or HCV sera. Coupled with fingerstick sampling and on-strip plasma separation, this supports same-visit execution of the reverse testing algorithm in an instrument-free format.