A remote fiber optic Raman sensor for rapid and nondestructive foodborne pathogen detection

Abstract

Food quality and safety have been critical issues in the world. There is an urgent need for a fast, simple, selective, and inexpensive food detection method for the identification of the degree of food spoilage. As a molecular analysis tool, Raman spectroscopy has the advantages of high selectivity, accurate analysis, simple operation, and low sample consumption. This paper reports a novel remote fiber optic Raman sensor for real-time application in food spoilage detection. Eight volatile organic compounds (VOC) liquids that typically generated by corrupted food were under-tested. The proposed sensor successfully captures the back-scattered Raman spectra for all testing samples with various dilution levels. Multiple machine learning algorithms are also applied to further analyze the correlation between Raman spectra and molecules in spoiled foods by diluting chemical samples. As a result of combining with Raman spectroscopy and machine learning algorithm, the remote fiber optic Raman probe allows qualitative measurements of VOC samples at 100-fold dilution. In comparison with surface-enhanced Raman scattering (SERS), the remote fiber optic Raman sensor allows for direct Raman spectroscopy detection without sample and SERS substrate preparation, which opens a new chapter on the nondestructive and sensitive detection of food analytes.