Chemical Change-Point Detection on a Silver Nanoplate Substrate

Surface enhanced Raman spectroscopy (SERS) has been established as a versatile molecular detection method, with unparalleled sensitivity and fingerprint like specificity. Metallic nanostructures play a crucial role in this technique, as they provide the necessary plasmonic enhancement to the Raman signal; however, these nanosensors are hard to reuse or regenerate. This limitation restricts the technique to single-use applications and does not allow for continuous monitoring of processes that vary over time. Here, we explore an alternative approach; we obtain a time series of spectra cumulatively over time from the same SERS substrate, and try to uncover chemical changes using statistical analysis. We tested our approach by analyzing the spectral time series using a controlled solution of a sample analyte. Silver nanoplates were grown on a flat substrate and placed in a microfluidic device. The substrate was exposed to continuous flow with varying analyte concentration, and with SERS spectra acquired every 2 seconds. The time series was analyzed using change-point detection, namely with the Isolate-Detect methodology, to identify the timepoints of the chemical changes in the solution.