Detection of Beverage Aspartame Bysurface-Enhanced Raman Scattering Spectroscopy Based on Deep Learning

In this paper, a self-developed surface-enhanced Raman scattering (SERS) substrate was used to detect aspartame (APM) in beverages by combining Raman spectroscopy with deep learning. A composite film of titanium dioxide nanocones and precious metals (gold, silver) nanoparticles was prepared on a titanium substrate by hydrothermal method and magnetron sputtering. The detection limit of Rhodamine 6G on the substrate was 10-7 M. A convolutional neural network deep learning algorithm model was compiled based on Python language to identify and analyze aspartame detected on SERS basis. The model could quickly identify whether the beverage contained aspartame with an accuracy of 99.11%, providing a rapid detection method for protecting public food safety.