Tailoring Nanoparticle Surfaces Empowered by Multivariate Techniques for Detection of Multiple Antibiotic Fingerprints in Real Samples Using Hand-Held Colorimetric Signal Readout

Due to the complex and varied toxicological characteristics exhibited by different antibiotics, which present significant risks to both the environment and human health, there is an urgent requirement for highly efficient sensors capable of detecting antibiotics. The present study introduces a straightforward yet efficient colorimetric sensor array comprising surface-tailored nanoparticles, analyzed using different techniques including dynamic light scattering, zeta potential, Fourier transformation infrared spectroscopy (FTIR), X-ray photoelectron spectroscopy, atomic force microscopy, field emission scanning electron microscopy, and high resolution transmission electron microscopy for identifying and discriminating antibiotics. Each antibiotic exhibits distinct binding affinities toward these sensing elements, leading to varied colorimetric and UV-visible absorption response patterns, which lead to unique "fingerprints" associated with each antibiotic. These patterns are quantitatively distinguishable through linear discriminant analysis, decision tree algorithm, random forest algorithm, support vector machine, and hierarchical clustering analysis. With 100% accuracy, the sensor array successfully discriminates between eight antibiotics (amoxicillin, azithromycin, neomycin, streptomycin, chloramphenicol, ofloxacin, fluconazole, and ciprofloxacin), with detection limits ranging from 1.7 to 8.3 μM. Additionally, binary and ternary mixture ratios of various antibiotics have also been successfully discriminated. Further, for the real-time applications, we have developed prototypes for colorimetric analysis and also modified the previously developed 96-well plate reader, which functions similarly to the UV-visible absorption spectrophotometer. These on-site sensing methods, with straightforward preparation, quick response, exceptional sensitivity, and consistently stable high-throughput signal output, show significant potential for practical use in food forensics and for environmental remediation.