Microplasma Controlled\nNanogold Sensor for SERS of\nAliphatic and Aromatic Explosives with PCA-KNN Recognition

Nanogold is an emerging material for enhancing surface-enhanced\nRaman scattering (SERS), which enables the detection of hazardous\nanalytes at trace levels. This study presents a simple, single-step\nplasma synthesis method to control the size and yield of Au nanoparticles\nby using plasma-liquid redox chemistry. The pin-based argon plasma\nreduces the Au<sup>3+</sup> precursor in under 5 min, synthesizing\nAu spherical particles ranging from ∼20 nm at 0.025 mM to ∼90\nnm at 1.0 mM, in addition to plate-like particles occurring at concentrations\nof 0.25–1.0 mM. The enhanced SERS responses correlated with\nthe UV–vis absorption and reflectance profiles, which can be\nattributed to synergistic plasmonic hotspots created by the sphere–sphere,\nplate-sphere, and plate–plate nanogold interactions. This nanogold\nmixture, combined with gold-plated CPU grid pin arrays, facilitated\nthe detection of trace explosives, including aromatic (TNT, TNB, and\nTNP) and aliphatic (RDX, PETN, and HMX) compounds. We demonstrate\nthat stabler aliphatic analytes, associated with lower vapor pressure\n(10<sup>–8</sup>–10<sup>–11</sup> atm), exhibit\nsmaller signal fluctuations (RSD ∼ 6–10%) compared to\ntheir more volatile (10<sup>–5</sup> atm) aromatic (RSD ∼\n12–17%) counterparts at similar analyte concentrations. The\ncalculated limit of detection (LoD) was found to be ∼2–6\nnM and ∼600–900 pM for aromatic and aliphatic explosives,\nrespectively. Finally, we show that the poorer performance of aromatic\nexplosives under the same sensing conditions affects SERS-PCA separation,\nwhich can then be improved either by a machine learning approach (PCA\nwith k-NN classification) or by consideration of a specific NO<sub>2</sub> symmetric stretching fingerprint range.